Preface

The Second National Conference on Mathematics and its Applications (SNCMA’2022)

held at Bordj Bou Arréridj University in September 17-18, 2022, during which four ses-
sions including “Partial Differential Equations", "Ordinary differential equations and Dy-
namical Systems", "Modeling, Control", "Operational research and optimization" and "Al-
gebra and Geometry" were organized in parallel for the presentation of the papers.
The aim of the conference is to bring together as many mathematics scientists as possible
to present their latest research and to provide a space for communication between young
PhD students and applied mathematics researchers for share experiences, discuss new
ideas in applied mathematics, strengthen cooperation and increase scientific exchanges
between different researchers.
The conference focuses on advancing faculty and research experts in various fields of re-
search and provides a common forum for leading applied mathematicians and scientists
active in various fields of mathematics to advance knowledge and debate about future
research challenges.
 ORGANIZING COMMITTEE

Honorary Chairs: Pr. Bouazza Boudersaya

Rector of Bordj Bou Arréridj University
Pr. Azedine Rahmoune
Dean of Mathematics and Computer Science Faculty

General Conference Chair: Dr. Rebiha Benterki

Organizing Committee Members
Dr. Belhadj Foudhil (Vice-dean of Mcs Faculty)
Mr. Hillal Touati (Bordj Bou Arréridj University)
Mr. Fares Bensaid (Bordj Bou Arréridj University)
Mr. Belkacem Nazih (Bordj Bou Arréridj University)
Mr. Abdelmalek Berrah (Bordj Bou Arréridj University)
Dr. Hanene Debbiche (Bordj Bou Arréridj University)
Dr. Khadra Dekkar (Bordj Bou Arréridj University)
Dr. Aziza Berbache (Bordj Bou Arréridj University)
Dr. Boualem Brahmi (Bordj Bou Arréridj University)
Mr. Zoubir Ramdani (Bordj Bou Arréridj University)

Scientific Committee Members

Dr. Djamila Benterki (Bordj Bou Arréridj University)
Dr. Hadjer Adimi (Bordj Bou Arréridj University)
Dr. Rebiha Zeghdane (Bordj Bou Arréridj University)
Dr. Smail Addoune (Bordj Bou Arréridj University)
Pr. Dalah Mohamed (Constantine University)
Dr. Messaoud Ghebouli (Bordj Bou Arréridj University)
Dr. Rachid Boukoucha ( Bejaia University)
Pr. Abdelbaki Merouani (Sétif University)
Dr. Hanene Amri (Annaba University)
Pr. Abdelatif Boureghda (Sétif University)
Pr. Mohammed-Salah Abdelouahab (Mila University center)
Pr. Abderahmane Bouchair (Jijel University)
Pr. Ameur Memou (M’Sila University)
Pr. Khaled Zeneir (Qassim University)
Dr. Hassane Bouremel (Barika University center)
Dr. Zoheir Chebel (Bordj Bou Arréridj University)
Dr. Ammar Derbazi (Bordj Bou Arréridj University)
Dr. Bilal Rahmoune ( Laghouat University)
Pr. Azedine Rahmoune (Bordj Bou Arréridj University)
Dr. Achour Saadi ( Laghouat University)
Pr. Salim Messaoudi (El Sharjah University, UAI)
Pr. Nasser-eddine Tatar (King Fahd University, )
Pr. Abdenacer Makhlouf ( Haute Alsace University, French)
Pr. Aider Meziane (USTHB University)
Pr. Madani Moussai (M’Sila University)
Pr. Ahmed Roubi (Hassan 1st University, Morocco)
Pr. Gen qi xu (Tianjin University China)
Pr. Moncef Aouadi (Carthage University, Tunisia)
Dr. Sami Mabrouk (Gafsa University, Tunisia)
Dr. Serkan SUTLU ( Işık University, Turkey)
Plenary Talks
CNMA’2022
Second National Conference on
Mathematics and its Applications

Combinatorial optimization, issues and

trends

Aider Meziane1 m-aider@[Link]

1 Professor of Mathematics, University of Science and Technology Houari Boumediene Algiers, Algeria.

Abstract Combinatorial optimization problems are used to model many situations in various
domains of everyday life. For instance, one can look for a shortest route between two cities on
a large road network or to scheduling exams for different courses at a university. In the last few
decades, great progress has been made in this field of mathematics, allowing the development
of efficient methods for solving complex and large-scale problems. These methods are based
on convex optimization techniques, geometry, randomization, or are inspired by nature.
In this talk, we will accurately describe this vast and rich field, discuss how to classify these
problems according to their complexities, and the nuances behind this, and indicate some gen-
eral limitations in solving most of these problems. We will also give some current trends for
this.

Rota-Baxter operators : Dualization and

Generalization

Abdenacer Makhlouf1 [Link]@[Link]

1 University of Haute Alsace, IRIMAS, Mulhouse, France.

Abstract :
In this talk, we will discuss Rota-Baxter operators and their generalization on algebras and
coalgebras. We present a dual version of T. Brzeziński’s results about Rota-Baxter systems
which appeared in 2016. Then as a generalization to bialgebras, we introduce the notion of
Rota-Baxter bisystem and construct various examples. On the other hand, we introduce a new
type of bialgebras (named mixed bialgebras) which are consisting of an associative algebra and
a coassociative coalgebra satisfying the compatible condition determined by two coderivations.
We investigate coquasitriangular mixed bialgebras and the particular case of coquasitriangular
infinitesimal bialgebras, where we give the double construction.
CNMA’2022
Second National Conference on
Mathematics and its Applications

Finite time stability in fractional differential

problems

Nasser-eddine Tatar1 tatarn@[Link]

1 Interdisciplinary Research Center for Intelligent Manufacturing and Robotics.

Abstract : In this presentation, we will explain what we mean by "finite stability" in our study.
After exhibiting few examples in the integer-order case, we shall discuss the main difficulties
encountered in the fractional case. Some suggestions are given to get around these difficulties.

Bresse-Timoshenko type systems with

thermodiffusion effects: well-possedness,
stability and numerical results

Khaled Zennir1 khaledzennir4@[Link]

1 Associate professor of Mathematics Qassim University (KSA) Saudi Arabia.

Abstract :
Bresse-Timoshenko beam model with thermal, mass diffusion and theormoelastic effects is
studied. We stated and proved the well-posedness of problem. The global existence and
uniqueness of the solution is proved by using the classical Faedo-Galerkin approximationsa-
long with two a priori estimates. We proved an exponential stability estimate for problem un-
der an unusual assumption, and by using a multiplier technique with frictional damping in the
vertical displacement. Numerically, we constructed a numerical scheme based on the P1-finite
element method for space discretization and implicit Euler scheme for time discre-tization.
Then, we showed that the discrete energy decays, later a priori error estimates areestablished.
Finally, some numerical simulations are presented.
CNMA’2022
Second National Conference on
Mathematics and its Applications

Modélisation mathématique de la
dynamique épidémique des maladies
transmissibles

Mohammed -Salah Abdelouahab1

1 Professor of Mathematics, University Center Abdelhafid Boussouf, Mila, Algeria.

Abstract :
L’émergence fréquente d’épidémies constitue un risque mondial de santé publique, qui néces-
site le développement de politiques et de stratégies de lutte contre leur propagation. La modéli-
sation mathématique est un outil incontournable pour comprendre la dynamique épidémique
et l’évolution des maladies ransmissibles, elle permet d’établir plusieurs scénarios basés sur
différentes hypothèses, différents paramètres et différentes données collectée. Un modèle
épidémiologique bien conçu peut aider les gestionnaires de la santé publique à étudier l’impact
de la maladie et concevoir des programmes efficaces de surveillance et de prévention.
Dans cette présentation on expose quelques modèles classiques de maladies infectieuses, puis
on aborde le développement et l’étude qualitative de la dynamique d’un modèle de tubercu-
lose.

Cohomologies and generalized

derivations of n-Lie algebras.

Serkan Sultu1 [Link]@[Link].

1 [Link]. of mathematics,
Faculty of Arts and Sciences Department of Mathematics, Işık University, Turkey.

Abstract :
In the present talk we shall introduce a cohomology theory for n-Lie algebras, which coincides
with the existing cohomology theory only in the case n=3. We shall then show that this new
cohomology theory is qualified to encompass the generalized derivations over n-Lie algebras
as 1-cocycles, and the inner generalized derivations as 1-coboundaries.
CNMA’2022
Second National Conference on
Mathematics and its Applications

A stability result for a nonlinear damped

suspension-bridge problem

Salim Messaoudi1 smessaoudi@[Link]

1 Professor of Mathematics, University of Sharjah, Sharjah, United Arab Emirates.

Abstract :
For the abstract you can use the following link:
[Link]
72514f4f34a4/[Link]?id=3931565

Title: Local and Global Well-Posedness of

a Coupled Westervelt-Pennes Model of
Nonlinear Ultrasonic Heating

Belkacem Said Houari1 bhouari@[Link]

1 Professor of Mathematics, University of Sharjah, Sharjah, United Arab Emirates.

Abstract :
High-Intensity Focused Ultrasound (HIFU) waves are known to induce localized heat to a tar-
geted area during medical treatments. In turn, the rise in temperature influences their speed
of propagation. This coupling affects the position of the focal region as well as the achieved
pressure and temperature values. In this work, we investigate a mathematical model of non-
linear ultrasonic heating based on the Westervelt wave equation coupled to the Pennes bioheat
equation that captures this so-called thermal lensing effect. We prove that this quasi-linear
model is well-posed locally and globally in time and does not degenerate under a smallness
assumption on the pressure data. We also proved some decay estimates of the solution.
CNMA’2022
Second National Conference on
Mathematics and its Applications

On the numerical solutions of nonlinear

quadratic integral equations of Urysohn
type on the half-line

Azedine Rahmoune1
1 Professor of Mathematics, University of Bordj Bou Arreridj, Algeria.

Abstract :
A numerical method for solving nonlinear quadratic integral equations of Urysohn type on
the half-line is presented. This approach reduces the given equation to a systematic procedure
by using a rational Legendre-collocation approximation (RLC). The rate of convergence and
error analysis are provided. Moreover, some numerical examples are carried out to verify the
spectral accuracy and the stability of the proposed method.
List of Presentations
 Contents

Partial Differential Equations 1

Existence of radial solutions for coupled elliptic systems: New Approach
Benterki Abdessalem and Rouaki Mohamed 1

Unsteady non-Newtonian fluid flows with boundary conditions of friction type: The case of
shear thickening fluids
Mahdi Boukrouche, Hanene Debbiche and Laetitia Paoli 3

SOME ESTIMATES OF GAGLIARDO-NIRENBERG TYPE IN HOMOGENEOUS FUNCTION

SPACES
Mohamed Benallia 5

Renormalized and entropy solutions for nonlinear anisotropic parabolic equations with L1 data
Abdelaziz Hellal 9

Free boundary flow problem over a depression

Beyoud Samira 12

Internal control for nonlocal wave equations

Imen Benabbas and Djamel Eddine Teniou 15

Evolution problems of second-order type and control

Soumia Saïdi 18

A new general decay rate of rotating Euler-Bernoulli beam with memory

Amirouche Berkani 20

Étude d’un problème mixte hyperbolique caractéristique fortement bien posés dans L2
Siham Brahimi et Ahmed Zerrouk Mokrane 22

Existence solution for a free boundary problem of fractional partial differential equation
Bilal Basti 24

Abstract differential equations of elliptic type with general Robin boundary conditions in
Hölder spaces: non commutative cases
Mohammed Rabah and Rabah Haoua 30

Multiple solutions for fractional Schrödinger-Poisson system with superlinear terms

Boutebba Hamza and Lakhal Hakim 33

Existence results for subcritical and critical p-fractional elliptic equations via Nehari manifold
method
Djamel Abid and Kamel Akrout 36
Multiplicity of Solutions for Kirchhoff Type Problem with Critical Exponent in R3
Hayat Benchira 39

Blow-Up for a Stochastic Wave Equation with Logarithmic Nonlinearity

Amina Benramdane and Fatna Bensaber 41

Elliptic Problem with Asymptotically linear term

Senhadji Asma and Nasri Yasmina 44

Existence of solution for elliptic problems with variable exponent and nonlinear boundary
conditions
Fareh Souraya 46

GENERAL STABILITY FOR MOORE GIBSON THOMPSON PROBLEM WITH A NONLINEAR

DAMPING TERM
Hassan Messaoudi 48

General decay of solutions for the wave equation with a delay and a boundary feedback of
memory-type
Hocine Makheloufi 51

Etude d’un problème viscoplastique avec frottement

Kendri Dalila 55

An application of Perov fixed point theorem on a coupled system in fractional Sobolev spaces
γ ,1
W0+1 (0, 1)
Noura Laksaci and Ahmed Boudaoui 59

GEOMETRIC ASPECTS FOR THE FULL VISCOUS NONLINEAR BOUSSINESQ SYSTEM

Youssouf Maafa 61

Existence and uniqueness global of bounded nonnegative weak solutions for nonlinear parabolic
evolution problem
Nabila Barrouk and Mounir Redjouh 63

Contact problem for thermoviscoelasstic material with friction

N. Bensebaa 67

s(·)
Existence of solutions for a class of PNP model in F B p(·),q
Oussama Melkemi 69

Radially symmetric solutions for a multidimensional nonlinear fractional partial differential

equations
Rabah Djemiat 71

Inverse medium problem in scattering

L. Chorfi, I. Djerrar and I. Khélifa 76

Studies on the existence of positive solutions for a second-order iterative boundary value prob-
lem via krasnoselskii’s fixed point theorem
Safa Chouaf 79
On Elliptic Differential Equations in Fréchet Spaces
Said Baghdad 83

Strongly Lorentz summing operators

Aldjia Attallah and Dahmane Achour 85

Spectral collocation method for solving integro-differential equations on the half-line

Boutarcha Sofiane and Rahmoune Azedine 87

Burgers Inviscid Equation And Its Modified Equation

Ilhem Mous Abdelhamid Laouar 89

Weakly mid-(p1 , ...pm )-summing multilinear operators

Athmane Ferradi 91

Some properties of anisotropic Roumieu Gelfand-Shilov spaces

M’Hamed Bensaid and Rachid Chaïli 94

On the Existence of Solutions for Impulsive Coupled Differential Equations

Khadidja Mebarki and Ahmed Boudaoui 97

Solvability of a dynamic frictional contact problem

Imane Ouakil, Benyattou Benabderrahmane and Yamna Boukhatem 101

Some results concerning meromorphic solutions for the Shroder equations

Bouternikh Salih and Zerzaihi Tahar 105

General Decay of a Nonlinear Viscoelastic Wave Equation

Amel Boudiaf 108

Well posedness of solution for laminated timoshinko beams and delay term
SABAH BAIBECHE 111

Null Controllability Of Nonlocal ψ-Hilfer fractional Differential Equation

Ichrak Bouacida, Mourad Kerboua and Sami Segni 113

On The Quadratic Pencil Related To Three Bounded Operators

Somia Kamouche, Mourad Ghiat and Hamza Guebbai 115

On the Resolution of Resonance and Nonresonance Systems Involving the Fractional Laplacian
Dob Sara, Hakim Lakhal and Messaoud Maouni 117

Fractional Partial Hyperbolic Differential Inclusions with Finite State-Dependent Delay

Mohamed Helal 120

Using projection methods to solve linear Fredholm intgro-differential equation

Boutheina Tair Hamza Guebbai 123

Solving non-linear Volterra integro-differential equation of second order

Sami Segni 128

Existence results for fractional Kirchhoff problem

 Yasmina NASRI 130

QUASILINEAR PARABOLIC PROBLEM WITH VARIABLE EXPONENT AND NONLOCAL BOUND-

ARY CONDITIONS
Fairouz Souilah 132

Galerkin method for the higher dimension Boussinesq equation non linear with integral con-
ditionS
Draifia Ala Eddine 134

Modeling and Control 136

A New Efficient Family of Conjugate Gradient Methods with Strong Wolfe Line Search Condi-
tions
Amina Hallal and Mohammed Belloufi 136

Asymptotic Modeling of the Effect of a Planar Porous Elastic Thin Layer

Athmane Abdallaoui 138

Estimation of the unrestricted PEXPAR(1) model with nonlinear least squares method
BECILA Sabah and MERZOUGUI Mouna 142

Enquête sur le taux de personnes vaccinées contre le coronavirus en Algérie

Yamina Khemal Bencheikh, Mona Dali and Yassmina Bouhassir 146

Characterization of some admissible trajectories for a warfare differential game problem

Benghebrid Safa, Bouremani Touffik and Benterki Djamel 149

Predictive mathematical models of the COVID-19 pandemic in obese people

Boubekeur Maroua Amel and Belhamiti Omar 152

A logistic equation with constant imigration to model the spread of COVID-19 in the first and
second waves
Boucherma Rayane Abdelouahab Mohammed salah and Boularouk Yakoub 154

A new approach for solving an optimal control problem using parametrization technique
Ikram Boukhelkhal and Rebiha Zeghdane 157

Control and stabilisation of a jerk system using the fractional Routh-Hurwitz criteria
Chettouh Besma and Menacer Tidjani 159

ESTIMATION OF GENERALIZED LINEAR FAILURE RATE DISTRIBUTION IN BAYESIAN

DJEMOUI Nour El Houda and CHADLI Assia 160

Analysis of M/D/1/N /N Retrial Queue Using Deterministic and Stochastic Petri Net
Ikhlef Lyes 162

Estimation d’un Mélange de Modéles GARCH Périodiques

Rokia Hemis, Hafida Guerbyenne and Fayçal Hamdi 166

Refined descriptive sampling with dependent variables

 Siham KEBAILI and Megdouda OURBIH-TARI 169

Characterization of a certain admissible trajectorie of Dolichobrachistochrone differential game

problem
Ghanem Aicha, Bouremani Touffik and Benterki Djamel 171

Modélisation mathématique de l’expansion géographique de la maladie de Chagas

BENARBIA Narimene and MAHDJOUB Tewfik 175

Stability of an axially moving string with a boundary disturbance

Abdelkarim Kelleche 180

Algebra and Geometry 136

Classification of generalized Weingarten hypersurfacesimmersed in space forms

Abdelmalek Mohammed and Mokhtari Kamel Eddine 136
 
Are there infinitely many n such that d n2 = d (ϕ (n))?
Amroune Zahra and Boudaoud Abdelmadjid 139

  
The equation σ (n) = τ k · ϕ n2 has finitely many solutions
Bellaouar Djamel 142

Conjugaite Hom-actions and the diffrent Sylow’s theorems for a Hom-groups

Zoheir Chebel, Hassene Bouramel and Hadjer Adimi 145

Lie-Poisson triple systems

Fatma Zouaidi 148

Other special relations for the Bernoulli and Euler numbers

Meryem Bouzeraib and Ali Boussayoud 150

Crossed module on 3-Hom-Lie-Rinehart superalgebra

Nawel Nasser 153

Prticular neutrosophic subsets on a lattice

Soheyb Milles and Lemnaouar Zedam 156

Generating functions for certain product of orthogonal polynomials with the generalized Tri-
bonacci numbers
Wissem Benamira and Ali Boussayoud 159

EDO and Dynamical Systems 162

Limit Cycle of Discontinuous Piecewise Differential Systems Separated by a Straight Line and
Formed by a Linear Center and a Quadratic Center
R. ALLAOUA, R. CHEURFA and A. BENDJEDDOU 162

On a solvable p−dimensional system of nonlinear difference equations

 Imane Zemmouri and Ahmed Ghezal 166

Adomian Decomposition Method for Solving a Class of Nonlinear Fractional Differential Equa-
tions
Lina CHETIOUI and Ali KHALOUTA 170

Solving the nonlinear fractional Logistic equation using fractional differential transform method
Fatima HATHAT and Ali KHALOUTA 173

On the Hyers-Ulam Stability of some ordinary Differential Equations

Said Fatma and Boukerrioua Khaled 177

Existence and estimates for solutions to coincidence point and to differential equations for
mappings acting in an arbitrary set
Benarab Sarra 179

Generating hidden chaotic attractors from a new 3D dynamical system with no-equilibrium
and strong chaoticity
Khaled Benkouider 182

An existence and uniqueness result for Caputo fractional differential equations

Bochra Azzaoui and Brahim Tellab 186

On the study of an integro-differential inclusion

A. Bouabsa and S. Saïdi 188

Predictive mathematical models of the COVID-19 pandemic in obese people

Boubekeur Maroua Amel and Belhamiti Omar 192

Multiple Positive Solutions for a Singular Multipoint Boundary Value Problem Via Fixed Point
Theory for the Sum of Operators
Lydia Bouchal and Karima Mebarki 194

Control and stabilisation of a jerk system using the fractional Routh-Hurwitz criteria
Chettouh Besma and Menacer Tidjani 197

Classes of polynomial diferential systems with algebraic limit cycles

Rima Chouader, Salah Benyoucef and Ahmed Bendjeddou 198

Solving Nonlinear Fredholm Integro-Differential Equations by Using Legender’s Wavelets

Youcef HENKA, Samir LEMITA, and Mohammed Zine AISSAOUI 202

Growth estimate for meromorphic solutions of ultrametric q-difference equations of Schröder

type
Houda Boughaba and Tahar Zerzaihi 204

Approximated solutions of Schrödinger equation for the generalized Cornell potential plus
Morse potential
Amal Ladjeroud and Badredine Boudjedaa 206

Global dynamics of a class of quadratic differential systems with a non-elementary singular

point
Meryem BELATTAR, Rachid CHEURFA and Ahmed BENDJEDDOU 208
Résultat d’existence des solutions pour problème d’inclusion différentielle fractionnaire avec
des conditions aux limites
Leyla Soudani and Abdelkader Amara 212

Existence and Uniqueness Results for a Neutral Delay Differential Equations with Harvesting
Term
Lynda Mezghiche 219

Inclusions with mappings acting from a metric space to a space with generalized distance
Merchela Wassim 223

Asymptotic behavior of a viscoelastic problem with long-term memory and Tresca friction law
Aissa Benseghir 225

Conditions de Robin généralisées pour une EDA à coefficients opérateurs variables dans le
cadre Lp
Rabah Haoua and Ahmed Medeghri 226

New Approach for the Asymptotic Stability of Two Dimensional Models

Kamel BENEYETTOU, Djillali BOUAGADA and Mohamed Amine GHEZZAR 228

Boundedness and exponential stability of solutions to of abstract dynamic equations on time

scales and applications
Amira Ayari 231

Stability of Incommensurate h-Nabla Fractional-Order Difference Systems

Noureddine Djenina, Adel Ouannas and Taki-Eddine Oussaeif 233

Existence Results for a System of Nonlinear Fractional Differential Equations

BENMEHIDI HAMMOU 238

The 16th Hilbert problem for the class of discontinuous piecewise differential systems sepa-
rated by irregular line and formed by linear center and cubic reversible isochronous centers
having rational first integrals
Imane Benabdallah and Rebiha Benterki 242

Well-posedness and exponential decay of some evolution problems with internal distributed
delay
Sami Loucif and Rafik Guefaifia 245

Limit cycles generated by piecewise linear Hamiltonian systems without equilibria with three
pieces
Louiza Baymout and Rebiha Benterki 249

Automorphy in a new genetic regulatory networks model with several delays

Mohammed Salah M’hamdi 250

Caputo−Hadamard random fractional differential equations in finite and infinite dimensional

Banach spaces
BEKADA Fouzia 254

Asymptotic stability analysis of Riemann-Liouville fractional stochastic neutral differential

equations
 Chahra Kechar and Abdelouaheb Ardjouni 256

Crossing limit cycles for a class of piecewise differential system separated by a straight line
Meriem Barkate and Rebiha Benterki 259

Limit cycles of polynomial differential systems with homogeneous nonlinearities of degree 8

via the averaging method
Ahlam Belfar and Rebiha Benterki 262

Synchronization between Fractional-Order Lesser Date Moth Chaotic System and Integer-Order
Chaotic System by Nonlinear Control
M. Labid and N. Hamri 265

Chaos of a new one-dimensional fractional map

Seyf El Islam Bouzeraa and Rabah Bououden 269

Dynamics study of the modified Beverton-Holt model

Chabane Bedjguelel and Hacene Gharout 272

Global dynamics of an age-structured model

Fatima Zohra HATHOUT and Tarik Mohammed TOUAOULA 275

Stability analysis of a fractional-order ecosystem with predator harvesting

Kerioui Nadjah 276

Operational research and optimization 280

Interior Point Method for Quadratic Programming
-Inverse Barrier Method-
Boutheina Fellahi and Bachir Merikhi 280

PRIMAL-DUAL PATH-FOLLOWING ALGORITHMS FOR SEMIDEFINITE PROGRAMMING

Chafia DAILI and Mohamed ACHACHE 283

Comparative Study on Multiple Chaotic Maps Incorporated Global Optimization Algorithms

Hamaizia Tayeb and Nacer Meriem 285

An adaptive version of the conjugate gradient parameter βkCD using the Newton direction for
unconstrained optimization
Naima HAMEL, Noureddine BENRABIA, Mourad GHIAT and Hamza GUEBBAI 286

An interior-point algorithm for linear programming based on modified displacement step

Linda Menniche and Djamel Benterki 288

Primal-dual feasible interior-point algorithm for monotone linear complementarity problems

Randa Chalekh and EL Amir Djeffal 291

NEW ITERATIVE CONJUGATE GRADIENT METHOD FOR NONLINEAR UNCONSTRAINED

OPTIMIZATION
Sabrina Ben Hanachi, Badreddine SellamiI and Mohammed Belloufi 295
Comparative Study on Multiple Chaotic Maps Incorporated Global Optimization Algorithms
Hamaizia Tayeb and Nacer Meriem 298

Comparison of Some Scalarization Methods in Multiobjective Optimization

Bouguern Wafa 299

A primal-dual interior-point method for convex quadratic programming based on a new kernel
function with a hyperbolic barrier term
Youssra Bouhenache and Wided Chikouche 302

Interior-point algorithm for linear programming based on a new descent direction

Zaoui Billel, Benterki Djamel and Khelladi Samia 305

A BFGS algorithm for global optimization using random perturbation

Raouf Ziadi 307
Partial Differential
Equations

1
CNMA’2022
Second National Conference on
Mathematics and its Applications

Existence of radial solutions for coupled

elliptic systems: New Approach

Benterki Abdessalem1 [Link]@[Link]

Rouaki Mohamed2
1 LMP2M Laboratory, Department of Mathematics and Computer Sciences, University of Medea, Algeria.
2 Department of Mathematics, Faculty of Sciences, University of Blida 1, Algeria..

Abstract: In this work, existence and localization results of the following coupled elliptic sys-
tem 


 ∆u + f (x, u, v) = λ x ∈ Ω


∆v + g(x, u, v) = µ x ∈ Ω (1)





 u=0=v

x ∈ ∂Ω
are established. Our proof is based upon a local version of fixed point theorem for set-valued
mappings in the setting of 0-complet partial metric spaces which is mentioned on [A. H.
Ansari, A. Benterki, and M. Rouaki. Some local fixed point results under C-class functions
with applications to coupled elliptic systems. Journal of Linear and Topological Algebra,
07(03) :169–182, 2018.]

Keywords: Existence; Radial solutions; Elliptic system; Fixed point; Green function.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
In this work, we are interested by radial solutions of the problem



 ∆u + f (x, u, v) = λ x ∈ Ω


∆v + g(x, u, v) = µ x ∈ Ω (2)





 u=0=v

x ∈ ∂Ω

where Ω denotes the unit ball in Rn (n > 1), λ, µ > 0, f and g are a continuous real functions. It is
well known that radial solutions of (2) satisfy the following EDO system
n−1 0

00
 u + u + f (r, u, v) = λ r ∈ (0, 1)



 r
 00 n − 1 0


 v + v + g(r, u, v) = µ r ∈ (0, 1) (3)
r




 u 0 (0) = 0 = v 0 (0), u(1) = 0 = v(1)



where r = |x|. Note that the radial solutions for several nonlinear elliptic systems are studied in
several works such as in [6, 7]. For one-dimensional (n = 1), the problem (3) are studied in [2] and
for f = g and λ = µ the problem is treated in [4]. Moreover, if u = v then the problem are studied

1
CNMA’2022
Second National Conference on
Mathematics and its Applications

also in [3].

In this study, we reformulated the nonlinear elliptic problem (3) as system of integral equations
with a necessities Green function which is reduce as fixed point problem. We apply our fixed
point theorem for set-valued mappings, in the framework of 0-partial metric spaces, mentioned
in [2], to guaranties the existence, localization and uniqueness of radial solutions for nonlinear
elliptic system (2).

References
[1] S. G. Matthews. Partial metric topology. Annals of the New York Academy of Sciences, 728
(1): 183–197, 1994.

[2] A. H. Ansari, A. Benterki, M. Rouaki. Some local fixed point results under C-class functions
with applications to coupled elliptic systems. Journal of Linear and Topological Algebra,
07(03) :169–182, 2018.

[3] A. Benterki, and M. Rouaki. Existence of Solutions for boundary value problems via fixed
point method. Advance Studies in Contomporary Mathematics, 28(04) :615–623, 2018.

[4] A. Benterki, M. Rouaki, A. H. Ansari. Some coupled fixed point results for set-valued map-
pings with applications. Communication in Nonlinear Analysis, 04(01) :111–120, 2018.

[5] S. Romaguera, A kirk type characterization of completeness for partial metric spaces, Fixed
Point Theory Appl. Article ID 493298, 6 pages, 2009.

[6] M. Rouaki. Nodal radial solutions for a superlinear problem, Nonlinear Anal. Real World
Appl., 8 (2): 563–571, 2007.

[7] M. Rouaki. Existence and classification of radial solutions of a nonlinear nonautonomous

Dirichlet problem. arXiv:1110.4019, 2011.

2
CNMA’2022
Second National Conference on
Mathematics and its Applications

Unsteady non-Newtonian fluid flows with

boundary conditions of friction type: The
case of shear thickening fluids

Mahdi Boukrouche1 [Link]@[Link]

Hanene Debbiche2 [Link]@[Link]
Laetitia Paoli1 [Link]@[Link]
1 Lyon University, F-42023 Saint-Etienne, Institut Camille Jordan UMR 5208, 42023 Saint-Etienne cedex 2,
France
2 Department of mathematics, Mohamed El Bachir El Ibrahimi University, Bordj Bou Arréridj, Algeria

Abstract: We consider non-stationary flow problems for general incompressible dilatant (shear
thickening) fluids in a bounded domain Ω ⊂ R3 . The conservation of mass and momentum
lead to a p-Laplacian unsteady Stokes system where the real parameter p is greater than 2. We
assume non-standard mixed boundary conditions with a given time dependent velocity on a
part of the boundary and Tresca’s friction law on the other part. From the latter condition, we
obtain that the fluid velocity and pressure satisfy a non-linear parabolic variational inequality
and belong to Banach spaces depending on p.
We prove the existence of a solution by using Schauder’s fixed point theorem, the notion of
semigroup and monotony methods. Then, we conclude by applying De Rham’s theorem to
construct the pressure term.

Keywords: Unsteady shear thickening fluid flow, p-Laplacian, Tresca’s friction law, Non-
linear variational parabolic inequality, Monotonicity methods.
2010 Mathematics Subject Classification: 76A05 35Q35, 35K87, 76M30.

Introduction
Fluid flow problems are involved in several physical phenomena and play an important role in
many industrial applications. Motivated by applications to industrial processes like lubrication
or extrusion/injection. We study p-Laplacian non-stationary Stokes system in the case p ≥ 2 (i.e.
shear thickening fluids like colloidal fluids (p > 2) or generalized Newtonian fluids (p = 2) like oils
since the viscosity still depends on the temperature, the velocity and the modulus of the strain rate
tensor).

Main results
We decompose the boundary of Ω as ∂Ω = ΓD ∪ Γ0 . We define
p
[Link] = {ϕ ∈ (W 1,p (Ω))3 ; ϕ=0 on ΓD , ϕ·n = 0 on Γ0 and div(ϕ) = 0 in Ω},
et
H = {ψ ∈ (L2 (Ω))3 ; ψ·n = 0 on ∂Ω and div(ψ) = 0 in Ω}.

3
CNMA’2022
Second National Conference on
Mathematics and its Applications

The variational formulation of the problem is given by

p 0 p
Problem (P): Find υ ∈ C([0, T ]; H) ∩ Lp (0, T ; [Link] ) with υ0 ∈ Lp (0, T ; ([Link] )0 ) satisfay
p
[υ0 , ϕ − υ] + [Aυ, ϕ − υ] + J(ϕ) − J(υ) ≥ [f , ϕ − υ], ∀ϕ ∈ Lp (0, T ; [Link] )
υ(0) = 0 in Ω,
p p
where p0 = p−1 , υ = υ − υ0 ξ, [., .] denotes the duality product between Lp (0, T ; [Link] ) and
0 p
Lp (0, T ; ([Link] )0 ),
Z T Z
J(ϕ) = k|ϕ − s̃| dx0 dt,
0 Γ0

and Z T Z
[Aυ, ϕ] = 2µ(θ, υ + υ0 ξ, |D(υ + υ0 ξ)|)|D(υ + υ0 ξ)|p−2 dij (υ + υ0 ξ)dij (ϕ)dx dt.
0 Ω
We prove the existence of a solution to Problem (P) by using a fixed point argument. Indeed, for
all u ∈ Lp (0, T ; (Lp (Ω))3 ), we consider the following problem
p 0 p
Problem (Pu ): Find υ ∈ C([0, T ]; H) ∩ Lp (0, T ; [Link] ) with υ0 ∈ Lp (0, T ; ([Link] )0 ) satisfay
p
[υ0 , ϕ − υ] + [Au υ, ϕ − υ] + J(ϕ) − J(υ) ≥ [f , ϕ − υ], ∀ϕ ∈ Lp (0, T ; [Link] ),
υ(0) = 0 in Ω,

where
Z T Z
[Au υ, ϕ] = 2µ(θ, u + υ0 ξ, |D(υ + υ0 ξ)|)|D(υ + υ0 ξ)|p−2 dij (υ + υ0 ξ)dij (ϕ)dx dt.
0 Ω

We prove the existence and uniqueness of a solution of problem (Pu ) by using the notion of semi-
group and monotony methods [3]. Then, we consider the mapping

Λ : Lp (0, T ; (Lp (Ω))3 ) → Lp (0, T ; (Lp (Ω))3 )

u 7→ υ.

By using Schauder’s fixed point theorem we prove that Λ admits a fixed point and we use De
Rham’s theorem to establish the existence of the pressure.

References
[1] M. Boukrouche, H. Debbiche, L. Paoli. Unsteady non-Newtonian fluid flows with
boundary conditions of friction type : the case of shear thickening fluids, to ap-
pear in Nonlinear Analysis: Theory, Methods and Applications, Vol. 216, 2022,
[Link]

[2] G. Duvaut, J.L. Lions. Les inéquations en mécanique et physique, Dunod, 1972.

[3] J.L. Lions. Quelques Méthodes de Résolution des Problèmes aux Limites Non Linéaires, Dunod,
Paris, 1969.

4
CNMA’2022
Second National Conference on
Mathematics and its Applications

SOME ESTIMATES OF
GAGLIARDO-NIRENBERG TYPE IN
HOMOGENEOUS FUNCTION SPACES

Mohamed Benallia 1 benalliam@[Link]

1 Department of mathematics, Higher Normal School of BouSaada,
Labo. Funct. Anal. Geom. Spaces, M. Boudiaf University of M’Sila, 28000 M’Sila, Algeria

Abstract: In this paper, we study some inequalities of the Gagliardo-Nirenberg type in some
function spaces, in particular the homogeneous spaces of Besov and Lizorkin-Triebel, and the
inhomogeneous of Besov and Lizorkin-Triebel. Also, by using the notion of realizations, we
study this estimates in the realized homogeneous spaces of Besov and Lizorkin-Triebel. Then
we deduce some embedding properties of certain realized spaces into the Lebesgue spaces.

Keywords: Besov and Triebel-Lizorkin spaces, homogeneous spaces, Gagliardo-Nirenberg

inequalities, realizations.
2010 Mathematics Subject Classification: 46E35, 47H30.

Introduction
In the Fourier theory via the Littlewood-Paley decomposition, the homogeneous Besov spaces
Ḃsp,q (Rn ) and the homogeneous Triebel-Lizorkin spaces Ḟp,q s (Rn ) are quasi-Banach, defined as spaces

of distributions modulo polynomials, in the sense that kf kḂsp,q = kf kḞp,q s = 0 if and only if, f is a
polynomial on R . This approach presents the basic definition of Ḃp,q (Rn ) and Ḟp,q
n s s (Rn ). We will

use the notation Ȧsp,q (Rn ) := Ḃsp,q (Rn ) or Ḟp,q s (Rn ), respectively, and As (Rn ) := Bs (Rn ) or F s (Rn )
p,q p,q p,q
for inhomogeneous counterparts, when we are no need to distinguish them. These spaces will be
shortened by the initials B and F, respectively. In a recent work, in collaboration with M. Moussai
[3], we are interested in the realized homogeneous spaces of Besov B ˙ s (Rn ) and Triebel-Lizorkin
p,q
e
˙F s n
ep,q (R ) (indicated here by A ˙ s n ˙
ep,q (R ) := B s n
ep,q (R ) or F ˙ s n
ep,q (R )). The following notation is also useful
for us: For 0 < p, q ≤ ∞, we denote by k · kp the quasi-norm of the Lebesgue space Lp (Rn ), and
P q 1/q
`q (Z; E) is the set of all sequences (aj )j∈Z of elements in E s.t. k(aj )j∈Z k`q (Z;E) := j∈Z kaj kE < ∞,
if E = R (or C), we note `q (Z). For brevity, we use the notation Ep,q as Ep,q := `q (Z; Lp (R )), i.e., n
P q 1/q
k(aj )j∈Z kEp,q := j∈Z ka j kp < ∞ in the B-case, and Ep,q := Lp (Rn ; `q (Z)), i.e., k(aj )j∈Z kEp,q :=
 P 1/q
j∈Z |a |q
j p
< ∞ in the F-case, with the usual modification if p = ∞ or q = ∞. S(Rn ) de-
notes the Schwartz space and S 0 (Rn ) its topological dual, the space of tempered distributions. We
denote by P∞ (Rn ) the set of all polynomials on Rn . We denote by S∞ (Rn ) the set of all ϕ ∈ S(Rn )
such that hu, ϕi = 0 for all u ∈ P∞ (Rn ) and by S∞
0 (Rn ) the its topological dual, which can identified

to the quotient space S (R )/P∞ (R ). For all f ∈ S 0 (Rn ), we denote by [f ]P the equivalence class of
0 n n

f modulo P∞ (Rn ); this notation has been given before. The mapping which takes any [f ]P to the

5
CNMA’2022
Second National Conference on
Mathematics and its Applications

restriction of f to S∞ (Rn ) turns out to be an isomorphism of S 0 (Rn )/P∞ (Rn ) onto S∞

0 (Rn ). Then

S∞0 (Rn ) is called the space of distributions modulo polynomials.

Now, the Littlewood-Paley setting, which we will recall below, presents a basic tool to define
Ȧsp,q (Rn ) or Asp,q (Rn ), see e.g. [1], [5]. Throughout the paper we denote by ρ a fixed cut of function,
i.e., a positive, C ∞ (Rn ) and radial function ρ (i.e., ρ(ξ) = ρ e ∈ D(Rn )) such that 0 ≤ ρ ≤ 1,
e(|ξ|), with ρ
with ρ(ξ) = 1 if |ξ| ≤ 1 and ρ(ξ) = 0 if |ξ| ≥ 3/2. We put γ(ξ) := ρ(ξ) − ρ(2ξ) for all ξ ∈ Rn . Then γ
is supported by the compact annulus 1/2 ≤ |ξ| ≤ 3/2, and the following identities hold
X X
γ(2j ξ) = 1 (∀ξ ∈ Rn \ {0}), ρ(2−k ξ) + γ(2−j ξ) = 1 (∀k ∈ Z, ∀ξ ∈ Rn ).
j∈Z j≥k+1

The functions ρ and γ will be fixed once and for all. We define the pseudodifferential
 operators
(Sj )j∈Z and (Qj )j∈Z by S f (ξ) := ρ 2 −j ξ fb(ξ) and Q f (ξ) := γ 2 −j ξ fb(ξ). We also define the
j j
d d
operators (Qj )j∈N0 by Q0 := S0 and Qj := Qj for j ≥ 1. The operators Sj and Qj take values in
e e e
the space of analytical functions of exponential type, see Paley-Wiener [Link] is clear that
Sj is defined on S 0 (Rn ) and that Qj is defined on S∞ 0 (Rn ) since Q f (x) = 0 if, and only if, f is
j
a polynomial. We make use of the following convention: If f ∈ S 0 (Rn ), then [f ]P ∈ S∞ 0 (Rn ). If
0 n 0 n
f ∈ S∞ (R ) we define Qj f := Qj f1 for all f1 ∈ S (R ) such that [f1 ]P = f . By Proposition 2.5 in
[4], the weak convergence of the Littlewood-Paley decomposition of a tempered distribution is
described in the following well-known statement.
Definition 1. Let s ∈ R, 0 < p, q ≤ ∞ (with p < ∞ in the F-case).
• The homogeneous spaces Ȧsp,q (Rn ) is the set of f ∈ S∞
0 (Rn ) such that

 P 1/q
jsq q
j∈Z 2 kQj f kp < ∞ , in the B-case,




kf kȦsp,q := 

 P 1/q
jsq q

< ∞ , in the F-case.
j∈Z 2 |Qj f |



p

• The inhomogeneous spaces Asp,q (Rn ) is the set of f ∈ S 0 (Rn ) such that
 P 1/q
jsq e q
j≥0 2 kQj f kp < ∞ , in the B-case,




kf kAsp,q := 

 P 1/q
jsq e q

< ∞ , in the F-case.
j≥0 2 |Qj f |



p

Remark 1. If P ∈ P∞ (Rn ), then we have kf kȦsp,q = kf + PkȦsp,q .

Definition 2. [2] Let E be a vector subspace of S∞ 0 (Rn ) endowed with a quasi-norm such that E ,→

S∞0 (Rn ) holds. A realization of E in S 0 (Rn ) is a continuous linear mapping σ : E → S 0 (Rn ) such that

[σ (f )]P = f for all f ∈ E. The image set σ (E) is called the realized space of E.

Main results
We will prove essentially the following result:
Theorem 1. Let 0 < p, q ≤ ∞ and m ≥ 0 (with p < ∞ in the F-case). Let a, b real numbers such that
0 < a < b. Let (fj )j≥0 be a sequence in S 0 (Rn ) such that fbj is suppored by the annulus a2j ≤ |ξ| ≤ b2j ,
and (2m+n/p fj )j≥0 in Ep,q . Then the series j≥0 fj converges in S 0 (Rn ) and there exists a constant c > 0
P
such that the inequality
X
fj ≤ c (m + n/v)1/q−1 2m+n/v k(2m+n/p fj )j≥0 kEp,q ,
v
j≥0

6
CNMA’2022
Second National Conference on
Mathematics and its Applications

holds, for all v ∈ [p, ∞] (with v < ∞ in the F-case). The constant c can be chosen such that c :=
max(1, pn/p ) if p < v and c := 1 if p = v.

As an immediate consequence we have the following corollary (with fj := Qj f ):

Corollary 1. Let 0 < p, q ≤ ∞ and m ≥ 0 (with p < ∞ in the F-case). There exists a constant c > 0 such
that the inequality
m+n/p
kf kv ≤ c (m + n/v)1/q−1 2m+n/v kf kAm+n/p , ∀f ∈ Ap,q (Rn ),
p,q

holds, for all v ∈ [p, ∞] (with v < ∞ in the F-case). The constant c can be chosen such that c :=
max(1, pn/p ) if p < v and c := 1 if p = v.

We now present a characterization of inhomogeneous spaces Asp,q (Rn ), homogeneous spaces Ȧsp,q (Rn )
˙s n
and realized spaces A
e p,q (R ) by estimates of Gagliardo-Nirenberg type. We begin by the following
assertion:
Theorem 2. Let 0 < q ≤ 1. Then there exists a constant c > 0 such that the inequality
1−u/v
kf kv ≤ ckf ku/v
u kf kB0 ,
∞,q

holds, for all u ∈]0, ∞[, all v ∈ [u, ∞], and all f ∈ Lu (Rn ) ∩ B0∞,q (Rn ).
Corollary 2. Let 0 < p, q ≤ ∞ (with p < ∞ in the F-case). Then there exists a constant c > 0 such that
the inequality
1−u/v
kf kv ≤ cv 1−1/q 2n/v kf ku/v
u k[f ]P k n/p ,
Ȧp,q
n/p
holds, for all u ∈]0, ∞[, all v ∈ [u, ∞[ and all f ∈ Lu (Rn ) ∩ Ap,q (Rn ).
Theorem 3. Let 0 < q ≤ 1 and 0 < p < ∞. Then there exists a constant c > 0 such that the inequality
p/v 1−p/v
kf kv ≤ ck[f ]P kḞ 0 k[f ]P kḂ0 ,
p,q ∞,q

˙ 0 (Rn ) ∩ B
holds, for all v ∈ [p, ∞] and all f ∈ F ˙ 0 (Rn ).
p,q ∞,q
e e

Theorem 4. Let 0 < q ≤ 1 and 0 < p < ∞. Then there exists a constant c > 0 such that the inequality
p/v 1−p/v
kf kv ≤ c v 1−1/q 2n/v k[f ]P kḞ 0 k[f ]P kḂ0 ,
p,q ∞,q

˙ 0 (Rn ) ∩ B
holds, for all v ∈ [p, ∞[ and all f ∈ F ˙ 0 (Rn ).
p,q ∞,q
e e

Theorem 5. [3] Let 0 < p, q < ∞ with q ≤ 1 in the B-case and p ≤ 1 in the F-case. Then there exists a
constant c > 0 such that the inequality
1−u/v
kf kv ≤ c v 1−1/q 2n/v kf ku/v
u k[f ]P k n/p1 ,
Ȧp1 ,q

˙ n/p n
holds, for all u ∈]0, ∞[, all v ∈]0, ∞[ s.t. v ≥ max(p, u), all p1 ∈ [p, ∞[ and all f ∈ Lu (Rn ) ∩ A
e p,q (R ).

Remark 2. Since, the homogeneous spaces defined in S∞ 0 (Rn ), then these type of estimates fails to hold
˙ 0 (Rn ) ∩ A˙ n/p n 0 n n/p n n ˙ n/p n n n/p n
if F
e p,q
e p,q (R ) is replaced by Ḟp,q (R ) ∩ Ȧp,q (R ) and Lu (R ) ∩ Ap,q (R ) by Lu (R ) ∩ Ȧp,q (R ).
e
Indeed, if f be a nonzero polynomial on Rn , then k[f ]P kȦsp,q = 0 and kf ku = ∞ (it is possible to fall on
a wrong choice of representative which yields a contradiction). Contrary to the homogeneous spaces,
the results cover the case of inhomogeneous ones; in other words, we can take Fp,q 0 (Rn ) ∩ An/p (Rn ) and
p,q
n n/p n ˙ 0 n ˙ n/p n n ˙ n/p n
Lu (R ) ∩ Ap,q (R ) instead of F p,q (R ) ∩ Ap,q (R ) and Lu (R ) ∩ Ap,q (R ), respectively, in both results.
e e e

7
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] H. Bahouri, J.Y. Chemin, R. Danchin, Fourier Analysis and Nonlinear Partial Differential
Equations, Grundlehren der mathematischen Wissenschaften 343, Springer-Verlag, Berlin,
2011.

[2] G. Bourdaud, Realizations of homogeneous Besov and Lizorkin-Triebel spaces, Math. Nachr.
286 5-6 (2013), 476–491.

[3] M. Benallia, M. Moussai, Inequalities of Gagliardo-Nirenberg type in realized homogeneous

Besov and Triebel-Lizorkin spaces, MATH. REPORTS. 22(72) 1 (2020), 19–39.

[4] M. Moussai, Realizations of homogeneous Besov and Triebel-Lizorkin spaces and an applica-
tion to pointwise multipliers, Anal. Appl. (Singap.). 13 2 (2015), 149–183.

[5] H. Triebel. Local function spaces, heat and Navier-Stokes equations, EMS Tracts in Mathe-
matics 20. European Mathematical Society (EMS) Publishing House, Zürich, 2013.

8
CNMA’2022
Second National Conference on
Mathematics and its Applications

Renormalized and entropy solutions for

nonlinear anisotropic parabolic equations
with L1 data

Abdelaziz Hellal 1 [Link]@[Link]

1 Laboratory of Functional Analysis and Geometry of Spaces,
Mohamed Boudiaf-M’sila, University, Algeria

Abstract: In our work we prove the well-posedness ( existence and uniqueness ) of both
renormalized solutions and entropy solutions for an anisotropic parabolic equations with
pi -growth conditions and L1 data using the penalization method. Moreover, we obtain that
entropy solutions coincide with the renormalized solutions.

Keywords: Anisotropic parabolic equations; anisotropic Sobolev spaces; penalization

method; entropy solutions; renormalized solutions.
2010 Mathematics Subject Classification: primary: 35K55; secondary: 35D05,46E35,46E39.

Introduction
This work is concerned with the study of the nonlinear Dirichlet parabolic problem

∂t u − ∆p~ (u) + γ(·, ·)|u|p0 −2 u = f in QT ,

u(x, t) = 0 on ΣT , (4)
u(x, 0) = u0 (x) in Ω.

where Ω an open bounded domain in RN (N ≥ 2) with a Lipschitz boundary denoted by ∂Ω,

the cylinder denoted by QT = Ω × (0, T ) with T > 0 a real number, the lateral surface denoted by
ΣT = ∂Ω × (0, T ), the vector →
−
p = (p0 , p1 , · · · , pN ) satisfies 1 < pi < ∞ for i = 0, 1, · · · , N and ∆→
p is an
−
→
− →
−
operator of Leray-Lions type acted from the parabolic anisotropic space L p (0, T ; W 1, p (Ω)) into its
→−0 →−0
dual L p (0, T ; W −1, p (Ω)) is given by
N
X  
p (u) :=
∆→
− D i |D i u|pi −2 D i u ,
i=1

is called the →
−p -Laplacian which generalizes the p-Laplace operator.
Here γ is a positive function in L∞ (QT ) such that there exists a constant γ0 > 0 with γ(x, t) ≥
γ0 a.e in QT , f ∈ L1 (QT ), u0 ∈ L1 (Ω).

Our strategy focus on studying and establish the well-posedness (existence and uniqueness) of
both entropy solution and renormalized solution for problem (4) using the penalization method

9
CNMA’2022
Second National Conference on
Mathematics and its Applications

which based on making an approximate problem to problem (4) and choosing some suitable test
functions to obtain a-priori-estimates for the approximate solution un under appropriate assump-
tions. Based on the strong convergence of the truncations of approximate solutions and draw a
subsequence to obtain a limit function, and prove this function is an entropy solution of problem
(4). Moreover, we obtain the equivalence of entropy solutions and renormalized solutions with
|u|p0 −2 u ∈ L1 (Ω).

Main results
Here we give an important definition of an entropy solutions which is essential to our study of the
problem (4).

Entropy solutions
1,→
−
p
Definition 3. A function u ∈ T0 (QT ) ∩ C([0, T ]; L1 (Ω)) is an entropy solution to problem (4) if the
following conditions are satisfied:

1. |u|p0 −2 u ∈ L1 (QT ),

2. Z Z Z
ϕk (u − φ)(T ) dx − ϕk (u − φ)(0) dx + ∂t φ · Tk (u − φ) dx dt
Ω Ω QT
XN Z Z
i pi −2 i i
+ |D u| D u · D Tk (u − φ) dx dt + γ(x, t)|u|p0 −2 u · Tk (u − φ) dx dt
i=1 QT QT
Z
≤ f Tk (u − φ) dt dx,
QT
→
− 1,→
−
p →
−0 →
−0
for all φ ∈ L p (0, T ; W0 (Ω)) ∩ L∞ (QT ), with ∂t φ ∈ L p (0, T ; W −1, p (Ω)) + L1 (QT ).

where
1,→
− →− 1,→
−
 
p p
T0 (QT ) = u : QT 7−→ R is measurable,Tk (u) ∈ L p (0, T ; W0 (Ω)), D i Tk (u) ∈ Lpi (QT ) .

Here we state our first main result.

Theorem 6. Assume that the conditions on the principal part of problem (4) holds with f ∈ L1 (QT ) and
u0 ∈ L1 (Ω). Then there exists at least one entropy solution for the problem (4).

Here we introduce the definition of renormalized solutions to the problem (4).

Renormalized solutions
1,→
−
p
Definition 4. A function u ∈ T0 (QT ) ∩ C([0, T ]; L1 (Ω)) is a renormalized solution to problem (4)
if the following conditions are satisfied:
PN R i u|pi
1. limh→+∞ i=1 {h≤|u|≤h+1} |D dx dt = 0,

10
CNMA’2022
Second National Conference on
Mathematics and its Applications

2.
Z N Z
∂S(u) X
ϕ dx dt + |D i u|pi −2 D i u · (S 00 (u)ϕD i u + S 0 (u)D i ϕ) dx dt
QT ∂t QT
i=1
Z
+ d(x, t)|u|p0 −2 u · S 0 (u)ϕ dx dt
QT
Z
= f S 0 (u)ϕ dx dt,
QT
→
− 1,→
−
p
for every function ϕ ∈ L p (0, T ; W0 (Ω)) ∩ L∞ (QT ), and any renormalization S(·) ∈ C ∞ (R)
such that suppS 0 (·) ⊆ [−M, M] for some constant M > 0.

Here we state our second main result,

Theorem 7. Suppose that the assumptions on the principal part of problem (4) holds with f ∈ L1 (QT )
and u0 ∈ L1 (Ω). Then the entropy solution u in Theorem (6) is also a renormalized solution for problem
(4).

Proofs of the main results

Here we list briefly the proofs of the main results which presented in the previous section.

Proofs of Theorems 6-7

The proofs of Theorems 6-7 needs several steps: First, we approximate the problem (4) with
sequence of problems having smooth solutions (un ). Then, after deriving a-priori-estimates on un
and based on the strong convergence of the truncations of approximate solutions (un ), we pass to
the limit using a compactness results as in [1, 2, 3].

References
[1] H. Abdelaziz, F. Mokhtari; Nonlinear anisotropic degenerate parabolic equations with vari-
able exponents and irregular data, J. Elliptic and Parabolic Equations (2022) 8:513-532.

[2] M. Chrif, S. EL Manouni, H. Hjiaj; On the study of strongly parabolic problems involving
anisotropic operators in L1 , Monatshefte für Mathematik,Springer-Verlag GmbH Austria
(2021).

[3] F. Mokhtari; Anisotropic parabolic problems with measur data, Differential Equations and
Applications, 2 (2010), 123-150.

[4] F. Mokhtari, R. Mecheter; Anisotropic degenerate parabolic problems in RN with variable

exponent and locally integrable data, Mediterr. J. Math., 16(3) (2019), 1-21.

[5] J. Rákosník, Some remarks to anisotropic Sobolev spaces II, Beiträge Anal. 15 (1981), 127-
140.

11
CNMA’2022
Second National Conference on
Mathematics and its Applications

Free boundary flow problem over a

depression

Beyoud Samira1 formathematic@[Link]

1 Department of mathematics, University of Sciences and Technology Houari Boumediene (USTHB), El Alia, BP
32, 16123 Bab-Ezzouar, Algiers, Algeria

Abstract: Two-dimensional steady free-surface flows over a depression is considered. The

fluid is supposed inviscid, incompressible and the flow is irrotational. The gravity is included
in the dynamic boundary conditions. Far upstream, the flow is assumed to be uniform. The
depression is lying on the bottom of an infinit channel. In this work, the fully nonlinear flow
problem is formulated to obtain as so called the fKdV equation. The resulting equation is
solved numerically. When the surface tension is neglected and the gravity is included, there
is a parameter in the problem known as the Froude number F is used but in the case F < 1.
Finally, solution graphes for all flow regimes are presented.

Keywords: The dynamic and kinematic boundary conditions, fKdV equation, the Froude
number F.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
Problems of free surface flow over an obstacle, modeling various situations arising in oceanogra-
phy and atmospheric sciences, are studied by many researchers. Various mathematical techniques
have been employed to study the mechanism of wave generation by different kinds of obstacles
situated at the bottom of a channel. At 1988, Forbes [3] presented a numerical solution for critical
free surface flow over a semi-circular obstruction attached to the bottom of a running stream. In
[1], Dias and Vanden-Broeck studied the problem involving free-surface flow past a submerged
triangular obstacle at the bottom of a channel, and solved the problem numerically by applying a
series truncation method.

Lamb [5], presented a general linearized theory for flow over stream beds of arbitrary shape. At
1984, Akylas [1] derived the forced Korteweg de Vries (fKdV) equation, to study the excitation
of long nonlinear water waves by a moving pressure distribution. Grimshaw and Smyth [4], pre-
sented a theoretical study of a stratified fluid which is flowing over a bottom topography. They
solved the problem by using weak nonlinear theory and pointed out that the flow can be described
by a forced Korteweg-de Vries equation.

In this work we consider a free surface flow problem. The flow is bidimensional, steady and irro-
tational. The fluid is inviscid and incompressible, it is perturbed by a depression placed on the
bottom of an infinite channel. The gravity is taken into account but the superficial tension effects
are neglected, we are interested in finding the equilibrium free surface of the flow. The study is

12
CNMA’2022
Second National Conference on
Mathematics and its Applications

done numerically through transforming the Bernoulli equation into the Forced Kortweg de-Vries
equation.

Position of the problem

The configuration of the flow is shown in the following figure

Figure 1: A diagram showing the physical set up of the problem

We consider the problem : find a function Φ and a function η such that

∆Φ = 0 , b(x) < y < H + η(x, y)

D
(y − H − η) = 0 on y = H + η
Dt

∂Φ 1 1
+ (∇Φ)2 + gy = (c)2 + gH on y = H + η
∂t 2 2
D
(y − f ) = 0 on y = b
Dt
Φ is the potential of the velocity field, f the bottom topography, η the free-surface elevation, H
upstream depth of the fluid, c the speed of the upstream flow and g is the gravitational accelera-
tion.

The dimensionalization of the last equations, assuming that the potential perturbion wave motion
ϕ (Φ = cx + ϕ) and η possess asymptotic expansions of the form:

ϕ = εϕ1 + ε2 ϕ2 + ε3 ϕ3 + O(ε4 )

η = εη1 + ε2 η2 + O(ε3 ),

a system of differential equations and boundary conditions for successive approximations are
obtained according to the order of ε. Then, by solving the resulting equations, the following
forced Kortweg De-vries equation (fKdV) is derived,
3 1 1
ηt + ληx − ηηx − ηxxx = fx (5)
2 6 2
where λ = F − 1
In this work we solve equation (5) in the steady case.

13
CNMA’2022
Second National Conference on
Mathematics and its Applications

Numerical results
We solve numerically the steady equation (5) for F < 1, by using the Runge-Kutta fourth-order
method. For the fixed values of b (b is the hight of the obstacle), and the Froude number F, we
have obtained the free surface flow corresponding to semi circular, sloped step and rectangular
b3
depressions. Also we have taken f (x) = − 2 .
x + b2

References
[1] T. R. Akylas, On the excitation of long nonlinear water waves by a moving pressure distribu-
tion, J. Fluid. Mech., 141 (1984), 455-466.

[2] F. Dias, J. M. Vanden-Broeck, Open channel flows with submerged obstructions. J. Fluids.
Mech., 206 (1989), 155-170.

[3] L. K. Forbes, Critical free-surface flow over a semi-circular obstruction. J. Eng. Math., 22 (
1988), 3-13.

[4] R. H. J. Grimshaw and N. Smyth, Resonant flow of a stratified fluid over topography, J. Fluid
Mech., 16 (1986), 429-464.

[5] H. Lamb, Hydrodynamics, Cambridge University Press 1945.

14
CNMA’2022
Second National Conference on
Mathematics and its Applications

Internal control for nonlocal wave

equations

Imen Benabbas 1 [Link]@[Link]

Djamel Eddine Teniou 1
1 AMNEDP Laboratory, Faculty of Mathematics, USTHB, Algiers

Abstract: This paper is concerned with the null controllability of nonlocal wave equations
involving integral spacial terms, subject to Cauchy-Ventcel boundary conditions. We prove,
in two geometric settings, that acting on a neighborhood of the boundary, we can drive our
system to equilibrium. Our results are obtained under some assumptions on the kernel
functions by using compactness-uniqueness arguments.

Keywords: Controllability, observability, wave equations, dynamic boundary conditions.

2010 Mathematics Subject Classification: Primary 35L10, 45K05, 93B07.

Introduction
Let Ω be a bounded open domain of Rn , n ≥ 2, with boundary Γ = Γ 1 ∪ Γ 2 and let T > 0. We
denote ΩT = Ω × (0, T ), ΓT1 = Γ 1 × (0, T ), ΓT2 = Γ 2 × (0, T ). Consider the following nonlocal hyper-
bolic problem
 Z
2
∂t v − ∆v + KΩ (x, y)v(y, t) dy = w1 in ΩT



Ω



 Z

2
KΓ (ξ, ζ)vΓ (ζ, t) dΓ = w2 , v = vΓ on ΓT1

∂t vΓ + ∂ν v − ∆Γ vΓ +

 1
(6)


 Γ
v = 0 on ΓT2






(v(0), v (0)) = (v 0 , v 0 ), (∂ v(0), ∂ v (0)) = (v 1 , v 1 ) in Ω × Γ 1

Γ Γ t t Γ Γ

We analyze the issue of controllability of system (6) in two different geometric settings. One is
where Ω is a bounded domain with a smooth boundary Γ = Γ 1 ∪ Γ 2 such that Γ 1 , Γ 2 are nonempty,
closed and Γ 1 ∩ Γ 2 = ∅. Second is where Ω is a rectangular domain in Rn . We denote by ∂ν
the normal derivative on Γ where ν = (ν1 , · · · , νn ) is the outward unit normal vector to Γ , and by
∆Γ the Laplace-Beltrami operator on Γ . Moreover, the kernel functions KΩ , KΓ are assumed to lie
respectively in L2 (Ω×Ω) and L2 (Γ ×Γ ), and no connection is required between these two functions.
System (6) is said to be null-controllable in time T > 0, if we can find control functions (w1 , w2 )
that will drive the solution from the initial state (v 0 , vΓ0 , v 1 , vΓ1 ) to the equilibrium

(v(T ), vΓ (T )) = (0, 0) in Ω,
(7)
(∂t v(T ), ∂t vΓ (T )) = (0, 0) on Γ 1 .

15
CNMA’2022
Second National Conference on
Mathematics and its Applications

In the framework of Hilbert uniqueness method [4], proving system (6) to be null-controllable is
reduced to proving the corresponding adjoint system to be observable in finite time T > 0
 Z
2
∂t u − ∆u + KΩ (x, y)u(x, t) dx = 0 in ΩT






 Ω Z


∂2t uΓ + ∂ν u − ∆Γ uΓ + KΓ (ξ, ζ)uΓ (ξ, t) dΓ = 0, u = uΓ on ΓT1


 1
(8)


 Γ
u = 0 on ΓT2






(u(T ), u (T )) = (u 0 , u 0 ), (∂ u(T ), ∂ u (T )) = (u 1 , u 1 ) in Ω × Γ 1 ,

Γ T T ,Γ t t Γ T T ,Γ

To this end, we first establish some auxiliary energy estimates [1, 3]. Then, using contradiction
arguments and taking account of unique continuation results for wave equations with constant
coefficients, we shall achieve two new observability results for system (8).
We introduce the functional spaces H, V , V 0
n o
H = L2 (Ω) × L2 (Γ 1 ), V = (u, v) ∈ HΓ12 (Ω) × H 1 (Γ 1 ); v = u|Γ 1
n o
where HΓ12 (Ω) = u ∈ H 1 (Ω); u|Γ 2 = 0 and V 0 is the dual space of V .
These are Hilbert spaces endowed with the norms
Z Z Z Z
2 2 2 2 2
||(u, v)||H = |u| dx + |v| dΓ , ||(u, v)||V = |∇u| dx + |∇Γ v|2 dΓ
Ω Γ1 Ω Γ1
where ∇Γ denotes the surface gradient on Γ .
Applying results from semigroups theory [2], we get that system (6) is well-posed in the following
sense
1. Given (v 0 , vΓ0 , v 1 , vΓ1 ) ∈ V × H and (w1 , w2 ) ∈ L2 (0, T ; H), there exists a unique weak solution to (6)
such that
(v, vΓ ) ∈ C(0, T ; V ) ∩ C 1 (0, T ; H).

2. Moreover, for all (v 0 , vΓ0 , v 1 , vΓ1 ) ∈ H × V 0 and (w1 , w2 ) ∈ L2 (0, T ; H), we have a unique solution to
(6)
(v, vΓ ) ∈ C(0, T ; H) ∩ C 1 (0, T ; V 0 ).
Main results
In a smooth domain

Let Ω be an open bounded domain of Rn with a smooth boundary as mentioned in the intro-
duction. We assume the following geometric condition on the boundary Γ : there exists x0 < Ω
such that 
1
(x − x0 ) · ν(x) ≤ 0 for all x ∈ Γ ,


(9)
(x − x0 ) · ν(x) ≥ 0 for all x ∈ Γ 2 .



Now, let ω ⊂ Ω be a neighborhood of Γ 2 in Ω; namely, there exits a neighborhood O ⊂ Rn of Γ 2

such that ω = O ∩ Ω. Setting T0 = max{|x − x0 |, x ∈ Ω} (see [3]), under the geometric hypothesis (9),
our first result of this section reads
Theorem 8. Let T > T0 . For all solutions to (8) associated to final data (uT0 , uT0 ,Γ ) ∈ H, (uT1 , uT1 ,Γ ) ∈ V 0 ,
we have the estimate
Z Z
2 2 2
||(u, uΓ )(0)||H + ||(∂t u, ∂t uΓ )(0)||V 0 ≤ c( |u| dxdt + |uΓ |2 dΓ dt) (10)
ωT ΓT1

16
CNMA’2022
Second National Conference on
Mathematics and its Applications

In a rectangular domain

We consider problem (8) in a two-dimensional rectangle Ω = (0, l1 ) × (0, l2 ). What follows could be
generalized to high-dimensional rectangular domains. In this case, we have Ventcel’s condition
on Γ 1 = (0, l1 )×{l2 } and Dirichlet’s condition on the remainder of the boundary Γ 2 = Γ 2,1 ∪Γ 2,2 ∪Γ 2,3
where Γ 2,1 = {0} × (0, l2 ), Γ 2,2 = (0, l1 ) × {0}, Γ 2,3 = {l1 } × (0, l2 ).
Let O ⊂ R2 be a neighborhood of the observed region Γ 2,1 ∪ Γ 2,2 ∪ {(0, l2 )}. We denote ω1 , ω2 ,
respectively, the intersections O ∩ Γ 1 , O ∩ Ω; so that ω = ω1 ∪ ω2 is a neighborhood of Γ 2,1 ∪ Γ 2,2 ∪
{(0, l2 )} in Ω ∪ Γ 1 .
√ q
Denote TR,0 = 2( 2 + 1) l12 + 4l22 (see [1]), then we have the following

Theorem 9. Let T > TR,0 . Then, there exists a constant c > 0 such that the solution to problem (8)
fulfills Z Z
2 2 2
||(u, uΓ )(0)||H + ||(∂t u, ∂t uΓ )(0)||V 0 ≤ c( |u| dxdt + |uΓ |2 dΓ dt)
ωT2 ωT1

for all final data (uT0 , uT0 ,Γ ) ∈ H, (uT1 , uT1 ,Γ ) ∈ V 0 .

References
[1] I. Benabbas, D. E. Teniou, Observability of wave equation with Ventcel dynamic condition,
EECT. 7 (2018), 545–570.

[2] K.-J. Engel, R. Nagel, One-parameter semigroups for linear evolution equations, Springer-
Verlag, New York, 2000.

[3] C. G. Gal and L. Tebou, Carleman inequalities for wave equations with oscillatory boundary
conditions and application, SIAM J. Control Optim. 55 (2017), 324–364.

[4] J.-L. Lions, Contrôlabilité exacte perturbation et stabilisation de systèmes distribués I, Masson,
Paris, 1988.

17
CNMA’2022
Second National Conference on
Mathematics and its Applications

Evolution problems of second-order type

and control

Soumia Saïdi 1 soumiasaidi44@[Link]

1 LMPA Laboratory, Department of Mathematics, Mohammed Seddik Ben Yahia University, Jijel, Algeria

Abstract: This work is a contribution on an optimization problem subject to a second-order

differential inclusion with maximal monotone operators. We aim to prove the well-posedness
theorem to the class of evolution problems under consideration, in the context of infinite
dimensional setting. Then, we show our result in optimal control theory.

Keywords: Differential inclusion, maximal monotone operator, minimization.

2010 Mathematics Subject Classification: Primary 34A60, 47J35, 34G25, 49J52, 49J53.

Introduction
Recent developments on differential inclusions of second-order evolution problems with maxi-
mal monotone operators have appeared in several papers, see e.g., [1], [2], [3], [4], [5] and the
references therein. The authors have considered such differential inclusions with their applica-
tions to the theory of fractional differential theory. In this spirit, we propose, in the current work,
a new application to control theory related to such evolution problems.

Main results
Let H be a real separable Hilbert space and let I = [0, T ] be an interval of R. Consider the evolution
problem
 du
− (t) ∈ A(t, x(t))u(t) + f (t, x(t), u(t)) a.e. t ∈ I,


dt



 Rt

x(t) = x0 + 0 u(s)ds, t ∈ I








 u(t) ∈ D(A(t, x(t)), t∈I


u(0) = u0 ∈ D(A(0, x0 )), x(0) = x0 ∈ H.



The operators A(t, y) : D(A, y) ⊂ H ⇒ H vary in the sense of the pseudo-distance for each (t, y) ∈
I × H. The single-valued perturbation f : I × H × H → H involved in the evolution problem is
measurable on I and Lipschitz continuous with respect to its second and third variables. The
existence and uniqueness are established using a discretization scheme.
Then, we minimize an objective function subject to the differential inclusion involving maximal
monotone operators above. So, we prove that our minimization problem admits an optimal solu-
tion.

18
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] C. Castaing, C. Godet-Thobie, S. Saïdi, On fractional evolution inclusion coupled with a time
and state dependent maximal monotone operator, Set-Valued Var. Anal. 30 (2022), 621–656.

[2] C. Castaing, C. Godet-Thobie and L.X. Truong, Fractional order of evolution inclusion cou-
pled with a time and state dependent maximal monotone operator, Mathematics MDPI, (2020),
1–30.

[3] C. Castaing, M.D.P. Monteiro Marques, P. Raynaud de Fitte, Second-order evolution prob-
lems with time-dependent maximal monotone operator and applications, Adv. Math. Econ. 22
(2018), 25–77.

[4] S. Saïdi, A perturbed second-order problem with time and state-dependent maximal monotone
operators, Discuss. Math., Differ. Incl. Control Optim. 41 (2021), 61–86.

[5] S. Saïdi, On a second-order functional evolution problem with time and state dependent maxi-
mal monotone operators, Evol. Equ. Control Theory, 11 (4) (2022), 1001–1035.

19
CNMA’2022
Second National Conference on
Mathematics and its Applications

A new general decay rate of rotating

Euler-Bernoulli beam with memory

Amirouche Berkani 1 aberkanid@[Link]

1 Department of Mathematics, Faculty of Sciences and Applied Sciences, University Akli Mohand Oulhadj of
Bouira, Bouira, 10000, Algeria.

Abstract: In this talk, we study the existence and asymptotic behaviour of solutions for a
viscoelastic rotating Euler-Bernoulli beam. Using the generalised Hamilton’s principle the
dynamic of the problem is described by PDEs coupled with ODEs. Based on the standard
Faedo-Galerkin method, we prove the well-posedness of the system. Under a suitable
boundary control, we prove an arbitrary decay of the energy of the system for a large class of
relaxation functions using the multiplier method.

Keywords: Euler-Bernoulli beam; PDEs; ODEs well-posedness; Faedo-Galerkin method;

general decay; relaxation function; viscoelasticity
Mathematics Subject Classification: 35L20; 93D15; 93D20.

Introduction
The aim of this talk is to study the existence and asymptotic behaviour of solutions for can-
tilevered Euler-Bernoulli beam with memory. It is fixed to a motor at one end and to a tip mass at
its free end. The governing equations of the system are (see Berkani [3])
 ZL      

 I θ̈(t) + ρx θ̈(t) + v (x, t) dx + m L θ̈(t) + v (L, t) + J θ̈(t) + v (L, t) = τ(t),
 h tt p tt p xtt



0 (11)

   Z t

 ρ xθ̈(t) + vtt (x, t) + EIvxxxx (x, t) − EI g(t − s)vxxxx (s)ds = 0,


0

for all (x, t) ∈ (0, L) × [0, ∞) together with the boundary conditions




 v(0, t) = vx (0, t) = 0,

 Zt  

 EIvxxx (L, t) − EI

g(t − s)vxxx (L, s)ds = m p L θ̈(t) + vtt (L, t) ,

 0
t

 Z  


 EIvxx (L, t) − EI g(t − s)vxx (L, s)ds = −Jp θ̈(t) + vxtt (L, t) , ∀t ∈ [0, ∞),


0

where the dot "." denotes the derivative with respect to the time t, θ(t) is the hub rotation angle,
v(x, t) is the deflection of the beam, τ(t) is the input control torque, ρ is the linear density of the
beam, L is the length of the beam, EI is the bending stiffness of the beam, Ih is the moment of
inertia of the motor and mp is the mass with rotational Jp attached at the free end of the beam.
The integral term in the second equation of (11) represents the memory term or the viscoelastic
damping term. This term appears in the constitutive relationship between the stress and the strain

20
CNMA’2022
Second National Conference on
Mathematics and its Applications

according to the Boltzmann Principle and is called a memory term (see [1, 2, 1]). The kernel g is
called the relaxation function.

References
[1] A. Berkani (2021), Exponential stability of a rotating Timoshenko beam under thermo-
viscoelastic damping; Int. J. Comput. Math, 99(3):426–445.

[2] A. Berkani and N-E. Tatar (2019), Stabilization of a viscoelastic Timoshenko beam fixed into
a moving base; Math. Model. Nat. Phenom. 14:501.

[3] A. Berkani (2018), Stabilization of a viscoelastic rotating Euler-Bernoulli beam; Math. Meth.
Appl. Sci., 41(1):2939–2960.

[4] A. Berkani, N-E. Tatar and L. Seghour (2018), Stabilisation of a viscoelastic flexible marine
riser under unknown spatiotemporally varying disturbance; Int J Control. 93(7):1547-1557.

[5] L. Meirovitch (2001), Fundamentals of vibrations; New York, NY:McGraw-Hill.

21
CNMA’2022
Second National Conference on
Mathematics and its Applications

Étude d’un problème mixte hyperbolique

caractéristique fortement bien posés dans
L2

Siham Brahimi 1 brahimisihame2016@[Link]

Ahmed Zerrouk Mokrane1 [Link]@[Link]
1 Département de mathématiques, Université de Batna–2–

Abstract: Ce travail est consacré à l’étude d’un problème mixte hyperboliques à bord carac-
téristique pour des systèmes où l’opérateur différentiel du premier ordre est symétrisables au
sens de Friedrichs à coefficients Lipschitzet satisfaisant une condition de structure minimale;
la condition de Kreiss-Lopatinskii Uniforme. En utilisant des techniques d’analyse fine comme
le calcul paradifférentiel, on montre que le problème est fortement bien posé dans L2 , au
sens où on exhibe l’existence d’une unique solution de même régularité que les données du
problème et satisfaisant à une estimation d’énergie.

Keywords: problème mixte hyperbolique, Kreiss–Lopatinskii uniforme, fortement bien

posé.
2010 Mathematics Subject Classification: Primary 35L50 , 35L40 , 35L04.

Introduction
Pour un entier d > 2, soit Ω un ouvert de Rd .
Pour simplifier on considérera le demi-espace

Ω : = {x = (y, xd ), y = (x1 , . . . , xd−1 ) ∈ Rd−1 , xd > 0},

de bord ∂Ω : = {x ∈ Rd : xd = 0}..
Nous allons étudier le problème mixte de la forme

 Xd

Lu = ∂t u + Aj (t, x)∂j u = F dans [0, T ] × Ω,




j=1

(12)



Bu = G sur [0, T ] × ∂Ω,





u =u dans Ω,

|t=0 0

Les coefficients Aj pour j = 1 . . . , d sont des fonctions définies dans R × Ω à valeurs dans l’espace
des matrices réelles N × N .
Les données F(t, x), u0 (x) ainsi que l’inconnue u(t, x) sont à valeurs vectorielles dans CN .
La matrice de bord B est de type p × N et G(t, x) est à valeurs vectorielles dans Cp .
La problématique de ce travail est d’étudier les questions d’existence locale, d’unicité pour des

22
CNMA’2022
Second National Conference on
Mathematics and its Applications

problèmes mixtes Friedrichs symétrisable, caractéristique de multiplicité constante lorsque les

coefficients de l’opérateur des équations de l’intérieur sont à régularité Lipschitz et la matrice de
bord B satisfait une condition de structure minimale pour les problèmes mixtes, dite condition de
Kreiss Lopatinski uniforme (UKL) qui est nécessaire et suffisante pour que le problème (12) soit
fortement bien posé dans L2 .
Par fortement bien posé, on entend que le problème admet une unique solution qui vérifie une
estimation d’énergie L2 sans perte de dérivées avec les termes sources F, G , et u0 en utilisant une
outil algébrique appelé ‚symétriseur de Kreiss ƒ.
Ces questions ont également obtenu des éléments de réponse dans le cas non caractéristique avec
des coefficients peu réguliers avec les travaux de Benzoni-Gavage et Serre [1], Coulombel JF [2] et
Mokrane AZ [4] ou plus récemment G.Métivier [3] .

Main results
Le premier résultat de ce travail se formule comme suit

Theorem 10. Supposons que les fonctions matricielles Aj , S0 ∈ W 1,∞ (Q), et B ∈ W 1,∞ (Σ). Supposons
que tous les hypothèses soient satisfaites. Alors pour tout T > 0, il existe des constantes γ0 ≥ 1 et C > 0
tels que pour tout F ∈ L2γ (QT ), G ∈ L2γ (ΣT ) et f ∈ L2 (Ω), le problème (12) admet une unique solution
u ∈ L2γ (QT ) telle que u|xII =0 ∈ L2γ (ΣT ). Par ailleurs, u ∈ C 0 ([0, T ], L2γ (Ω)) et pour tout 0 < t ≤ T on a:
d

!
1
γkuk2L2 (Q ) + ku|xII =0 k2L2 (Σ ) + e−2γt ku(t)k2L2 (Ω) ≤ C (kFk2L2 (Q ) + kGk2L2 (Σ ) + ku0 k2L2 (Ω) .
γ t d γ t γ γ t γ t

Le deuxième résultat de ce travail est une amélioration du résultat précédent, en supposant la

donnée des équations de l’intérieur seulement à régularité L1 ([0, T ], L2 (Ω)).

Theorem 11. Sous toutes les hypothèses du Théorème 10 et pour tout T > 0, il existe une constante
C > 0 telle que pour tout F ∈ L1 ([0, T ], L2 (Ω)), G ∈ L2 (ΣT ) et u0 ∈ L2 (Ω), le problème (12) admet une
unique solution u ∈ L2 (QT ) telle que u|xII =0 ∈ L2 (ΣT ). En outre, la solution u ∈ C 0 ([0, T ], L2 (Ω)) et
d
vérifie pour tout 0 < t ≤ T on a:
Zt !
II
ku|x =0 kL2 (Σt ) + ku(t)kL2 (Ω) ≤ C kF(s)kL2 (Ω) ds + kGkL2 (Σt ) + ku0 kL2 (Ω) .
d
0

References
[1] S. Benzoni-Gavage et D. Serre, Multidimensional Hyperbolic Partial Differential Equations,
First Order Systems and Applications. Oxford Mathematical Monographs; Oxford Univer-
sity Press, 2007.

[2] [Link], Well-posedness of hyperbolic initial boundary value problems. Journal de

Mathématiques Pures et Appliquées. 84 (2005), 786-818.

[3] G.Métivier, On the L2 well posedness of Hyperbolic Initial Boundary Value Problems, Annales
de l’Institut Fourier. 67 (2017), 1809-1863.

[4] [Link], Problèmes mixtes hyperboliques non linéaires. PhD, Université de Rennes I,
France, 1987(in French).

23
CNMA’2022
Second National Conference on
Mathematics and its Applications

Existence solution for a free boundary

problem of fractional partial differential
equation

Bilal Basti 1 bilalbasti@[Link]

1 Department of mathematics and informatics, University of Ziane Achour Djelfa

Abstract: This paper investigates the problem of existence and uniqueness of one solution
under the traveling wave form for a free boundary problem of a space-fractional wave
equation. It does so by applying the Banach’s fixed point theorem.

Keywords: Space-fractional wave equation; traveling wave form; free boundary prob-
lem; existence; uniqueness.
2010 Mathematics Subject Classification: 35R11; 35A01; 34A08; 35C06; 34K37.

Introduction
The partial differential equations (PDEs) of fractional order appear as a natural description of
observed evolution phenomena in various scientific areas. The fractional derivative operators are
non-local and this property is important in application because it allows to model the dynamics
of many problems in physics, chemistry, engineering, medicine, economics, control theory, etc.
For further reading on the subject, readers can refer to the following books (Kilbas et al. 2006 [3]).
In this work, we shall give an example of a class of well-known fractional-order’s PDEs; such the
equation which is the space-fractional wave equation and is written as follows:

∂2 u α
2∂ u
= c + v (x, t) , c ∈ R∗ , 2 ≤ α < 3, (13)
∂t 2 ∂xα
where u = u (x, t) and v (x, t) are scalar functions of space variables x ∈ [ct, X] and time t ∈ [0, T ] ,
for T > 0 and X > |c| T . With

α  ∂m u
∂ u   ∂xm ,
 α = m ∈ N,
= x m−α−1
∂xα   ∂∂xuα = Ictm−α ∂∂xmu = ct (x−τ)
α m m
R
∂


Γ (m−α) ∂τ m
u (τ, t) dτ, m − 1 < α < m ∈ N∗ .

The symbol I∗α presents the Riemann-Liouville’s fractional integral of order α.

The space-fractional wave equation (13) becomes the wave equation for α = 2 and v = 0, (see [2]).
The existence and uniqueness of solutions for fractional differential equations or fractional-order’s
PDEs have been investigated in recent years. For more on the subject, we refer the reader to the
following works [1].

24
CNMA’2022
Second National Conference on
Mathematics and its Applications

Our main goal in this work is to determine the existence, uniqueness and main properties of
solutions of the space-fractional PDE (13), under the traveling wave form:
 
u (x, t) = exp −c2 t f (x − ct) , with c ∈ R∗ , (14)

the basic profile f is not known in advance and is to be identified.

We exemplify the role of Free Boundary Problems as an important source of ideas in modern anal-
ysis. With the help of a model problem, we illustrate the use of analytical techniques to obtain
the existence and uniqueness of weak solutions via the use of the traveling wave method. This
method permits us to reduce the fractional-order’s PDE (13) to a fractional differential equation.
This approach (14) is very promising and can also bring novel results for other applications in
fractional-order’s PDEs.

Main results

Throughout the rest of this paper, we have:

2 ≤ α < 3, T > 0 and X > |c| T for some c ∈ R∗ . (15)

Lemma 1 ([3]). Assume that C D0α+ f ∈ C ([0, λ] , R) , for all α > 0, then:
m−1
X f (k) (0) k
I0α+ C D0α+ f (η) = f (η) − η , m − 1 < α ≤ m ∈ N∗ .
k!
k=0

Statement of the free boundary problem and main theorems

In this part, we first attempt to find the equivalent approximate to the following free boundary
problem of the space-fractional wave equation:

∂2 u 2 ∂α u ∗
 ∂t 2 = c ∂xα + v (x, t) , (x, t) ∈ [ct, X] × [0, T ] , 2 ≤ α < 3, c ∈ R ,



(16)
 ∂u (ct, t) = ∂2 u (ct, t) = 0, and u (x, 0) = f (x) , f ∈ C ([0, λ] , R) ,


∂x ∂x2


under the traveling wave form

 
u (x, t) = exp −c2 t f (η) , with η = x − ct . (17)

Now, we give the principal theorem of this work.

Theorem 12. Let α, c, T , X ∈ R, be the real constants given by (15) which satisfy the following inequal-
ity:
 1
0 < X + |c| T < c−2 Γ (α + 1) α .
If
α (X + |c| T )α [2 (X + |c| T ) |c| + α − 1]
< (X + |c| T )2 , (18)
Γ (α + 1) − c2 (X + |c| T )α
then the problem (16) admits a unique solution in the traveling wave form (17).

Existence and uniqueness results of the basic profile

First, we should deduce the equation satisfied by the function f in (17) and used for the
definition of traveling wave solutions.

25
CNMA’2022
Second National Conference on
Mathematics and its Applications

Theorem 13. Let (x, t) ∈ [ct, X] × [0, T ] , and ϕ ∈ C ([0, λ] , R) , be such that
 
v (x, t) = c2 exp −c2 t ϕ (η) , η ∈ [0, λ] ,
for η = x − ct and λ = X + |c| T , then the transformation (17) reduces the partial differential equation
problem of space-fractional order (16) to the ordinary differential equation of fractional order of the
form:
D0+ f (η) = c2 f (η) + 2cf 0 (η) + f 00 (η) + ϕ (η) , η ∈ [0, λ] ,
C α
(19)
with the conditions:
f 0 (0) = f 00 (0) = 0, (20)

Proof The fractional equation resulting from the substitution of expression (17) in the original
fractional-order’s PDE (13), should be reduced to the standard bilinear functional equation (see
[4]). First, for η = x − ct, we get η ∈ [0, λ] and
∂2 u 2

2
h
2 0 00
i
= c exp −c t c f (η) + 2cf (η) + f (η) . (21)
∂t 2
In another way, for ξ = τ − ct, we get:
Zx 3
∂α u 1 2−α ∂ u (τ, t)
= (η − τ) dτ
∂xα Γ (3 − α) ct ∂τ 3
 Z
exp −c2 t x
d3
= (x − τ)2−α 3 f (τ − ct) dτ
Γ (3 − α) ct dτ
 Z
exp −c t 2 η
d3
= (η − ξ)2−α 3 f (ξ) dξ
Γ (3 − α) 0 dξ
 
= exp −c2 t C D0α+ f (η) . (22)
If we replace (21) and (22) in (13), we get:
C
D0α+ f (η) = c2 f (η) + 2cf 0 (η) + f 00 (η) + ϕ (η) .
The proof is complete. 
In what follows, we present some significant lemmas to show the principal theorems.
We have:
Lemma 2. Let f , f 0 , f 00 , C D0α+ f , ϕ ∈ C ([0, λ] , R) , then the problem (19)–(20) is equivalent to the inte-
gral equation:
Zη
1  
f (η) = f (0) + (η − ξ)α−1 c2 f (ξ) + 2cf 0 (ξ) + f 00 (ξ) + ϕ (ξ) dξ. (23)
Γ (α) 0

Proof Let f , f 0 , f 00 , C D0α+ f , ϕ ∈ C ([0, λ] , R) , then by using Lemma 1, we reduce the fractional
equation (19) to an equivalent fractional integral equation. By applying I0α+ to the equation (19),
we obtain:  
I0α+ C D0α+ f (η) = I0α+ c2 f (η) + 2cf 0 (η) + f 00 (η) + ϕ (ξ) . (24)
From Lemma 1, we simply find:
I0α+ C D0α+ f (η) = f (η) − f (0) − ηf 0 (0) − η 2 f 00 (0) .
by using (20), the fractional integral equation (24) gives us:
 
f (η) = I0α+ c2 f (η) + 2cf 0 (η) + f 00 (η) + ϕ (η) + f (0) .
The proof is complete. 

26
CNMA’2022
Second National Conference on
Mathematics and its Applications

Lemma 3. Let f , C D0α+ f ∈ C ([0, λ] , R) , be such that f 0 (0) = f 00 (0) = 0, then:

    λα−2 (2λ |c| + α − 1)

2cf 0 (η) + f 00 (η) ≤ C D0α+ f , ∀η ∈ [0, λ] . (25)
Γ (α) ∞

Proof By using Lemma 1, for all C D0α+ f ∈ C ([0, λ] , R) , we get:

d α C α
I0α−1
+
C α
D0+ f (η) = I + D0+ f (η)
dη 0
= f 0 (η) − f 0 (0) − 2f 00 (0) η.

and
d2 α C α
I0α−2
+
C α
D0+ f (η) = I + D0+ f (η)
dη 2 0
= f 00 (η) − 2f 00 (0) .

Moreover; if f 0 (0) = f 00 (0) = 0, then:

I0α−1
+
C α
D0+ f (η) = f 0 (η) and I0α−2
+
C α
D0+ f (η) = f 00 (η) ,

also, we have for any η ∈ [0, λ] ,

       
2cf 0 (η) + f 00 (η) = 2 |c| I0α−1
+
C α
D0+ f (η)  + I0α−2
+
C α
D0+ f (η) 
   
2 |c| (η − ξ)α−2 C D0α+ f (ξ) (η − ξ)α−3 C D0α+ f (ξ)
Zη Zη
≤ dξ + dξ
0 Γ (α − 1) 0 Γ (α − 2)
λα−2 (2λ |c| + α − 1) C α
≤ D0 + f .
Γ (α) ∞

The proof is complete. 

 1
Theorem 14. If we put 0 < λ < c−2 Γ (α + 1) α and

αλα (2λ |c| + α − 1)

< λ2 , (26)
Γ (α + 1) − c2 λα

then the problem (19)–(20) admits a unique solution on [0, λ] .

Proof To begin the proof, we will transform the problem (19)–(20) into a fixed point problem
Af (η) = f (η) , with
Zη
1  
Af (η) = f (0) + (η − ξ)α−1 c2 f (ξ) + 2cf 0 (ξ) + f 00 (ξ) + ϕ (ξ) dξ. (27)
Γ (α) 0

We first notice that if f , C D0α+ f ∈ C ([0, λ] , R) , then Af is being an operator of a polynomial and a
primitive of continuous functions and its derivatives is indeed continuous (see (25) and the step 1
in this proof); therefore, it is an element of C ([0, λ] , R) , and is equipped with the standard norm:
 
kAf k∞ = sup Af (η) .
η∈[0,λ]

27
CNMA’2022
Second National Conference on
Mathematics and its Applications

Let f , g ∈ C ([0, λ] , R) be two functions that satisfy the problem (19)–(20), then
Zη
1
Af (η) − Ag (η) = (η − ξ)α−1 ×
Γ (α) 0
h i
c2 (f (ξ) − g (ξ)) + 2c (f 0 (ξ) − g 0 (ξ)) + (f 00 (ξ) − g 00 (ξ)) dξ.

Also Z η
  1  
Af (η) − Ag (η) ≤ (η − ξ)α−1 C D0α+ f (ξ) − C D0α+ g (ξ)  dξ. (28)
Γ (α) 0
For all η ∈ [0, λ] , we have:
   
C D0α+ f (η) − C D0α+ g (η)  = c2 (f (η) − g (η)) + 2c (f 0 (η) − g 0 (η)) + (f 00 (η) − g 00 (η))
     
≤ c2 f (η) − g (η) + 2 |c| f 0 (η) − g 0 (η) + f 00 (η) − g 00 (η) .

By using (25) from Lemma 3, we have:

λα−2 (2λ |c| + α − 1) C α
C D0α+ f − C D0α+ g ≤ c2 kf − gk∞ + D0+ f − C D0α+ g .
∞ Γ (α) ∞

As Γ (α) − λα−2 (2λ |c| + α − 1) > 0, we have:

c2 Γ (α)
C D0α+ f − C D0α+ g ≤ kf − gk∞ .
∞ Γ (α) − λα−2 (2λ |c| + α − 1)

From (28) we find:

c2 λα
kAf − Agk∞ ≤ kf − gk∞ .
Γ (α + 1) − αλα−2 (2λ |c| + α − 1)

This implies that by (26), A is a contraction operator.

As a consequence Banach’s contraction principle (see [1]), we deduce that A has a unique fixed
point which is the unique solution of the problem (19)–(20) on [0, λ] . The proof is complete. 
Proof of main theorems
In this part, we prove the existence and uniqueness of solutions of the following free boundary
problem of the space-fractional wave equation:

∂2 u 2 ∂α u ∗
 ∂t 2 = c ∂xα + v (x, t) , (x, t) ∈ [ct, X] × [0, T ] , 2 ≤ α < 3, c ∈ R ,



(29)
 ∂u ∂2 u

(ct, t) = (ct, t) = 0, and u (x, 0) = f (x) , f ∈ C ([0, λ] , ,


∂x ∂x 2 R)

under the traveling wave form:

 
u (x, t) = exp −c2 t f (η) , with η = x − ct. (30)

Proof of Theorem 12
The transformation (30) reduces the problem of the higher order space-fractional wave equation
(16) to the ordinary differential equation of fractional order of the form:
C
D0α+ f (η) = c2 f (η) + 2cf 0 (η) + f 00 (η) + ϕ (η) , (31)

with the conditions:

f 0 (0) = f 00 (0) = 0. (32)

28
CNMA’2022
Second National Conference on
Mathematics and its Applications

Let f ∈ C ([0, λ] , R) be a continuous function. By using Theorem 31, the condition (18) is equiva-
lent to (26), which is:
αλα (2λ |c| + α − 1)
< λ2 ,
Γ (α + 1) − c2 λα
We already proved the existence of a single solution of the problem (31)–(32) in Theorem 14,
provided that (26) hold space-fractional wave equation (16) under the traveling wave form (30).
The proof is complete.

References
[1] B. Basti, Y. Arioua and N. Benhamidouche, Existence results for nonlinear Katugampola
fractional differential equations with an integral condition, Acta Mathematica Universitatis
Comenianae, 89(2) (2020), 243–260.

[2] V. L. Ginzburg, Propagation of Electromagnetic Waves in Plasmas 2nd ed., Pergamon Press,
Oxford, 1970.

[3] A. A. Kilbas, H. H. Srivastava and J. J. Trujillo, Theory and Applications of Fractional Differ-
ential Equations, Elsevier Science B.V, Amsterdam, 2006.

[4] Polyanin A. D., Zaitsev V. F., Handbook of Nonlinear Partial Equation, Chap-
man&Hall/CRC, Boca Raton (2004).

29
CNMA’2022
Second National Conference on
Mathematics and its Applications

Abstract differential equations of elliptic

type with general Robin boundary
conditions in Hölder spaces: non
commutative cases

Mohammed Rabah 1 [Link]@[Link]

Rabah Haoua2 [Link]@[Link]
1 Department of mathematics , Abdelhamid Ibn Badis University, LMPA, 27000 Mostaganem, Algeria
2 Department of technical sciences, Abdelhamid Ibn Badis University, LMPA, 27000 Mostaganem, Algeria

Abstract: we prove some new results on operational second order differential equations of
elliptic type with general Robin boundary conditions in a non-commutative framework. The
study is developed in Hölder spaces under some natural assumptions generalizing those in
[1]. We give necessary and sufficient conditions on the data to obtain a unique strict solution
satisfying the maximal regularity property, see [3]. This work completes the one given in [1],
[2] and [3].

Keywords: second-order elliptic differential equations; Robin boundary conditions; ana-

lytic semigroup.

Introduction
Consider the following abstract second-order differential equation

00
u (x) + Au (x) − ωu (x) = f (x) , x ∈ ]0, 1[ , (33)
together with the abstract boundary conditions of Robin’s type


0
 u (0) − Hu (0) − µu (0) = d0 ,



 (34)
 u (1) = u1,



where A, H are closed linear operators in a complex Banach space X with domain D (A) , D (H)
respectively, d0 , u1 are given elements in X, ω, µ are complex parametres and the second membre
f belongs to C θ ([0, 1] ; X) , 0 < θ < 1. We will seek for a strict solution u to (33)−(34), that is a
fonction u such that:





 u ∈ C 2 ([0, 1] ; X) ∩ C ([0, 1] ; D (A))

  
 u (0) ∈ D Hµ .




30
CNMA’2022
Second National Conference on
Mathematics and its Applications

The methode is essentially based on Dunford calculus, interpolation spaces, the semigroup theory
and some techniques as in [1], [2] and [3]. For our study of the problem (33)−(34), we will use the
following notation: for ϕ ∈ (0, π) , we set

Sϕ = {z ∈ C {0} : |arg (z)| 6 ϕ} ∪ {0} .

We will seek for a strict solution u to this problem, that is a fonction u such that:





 u ∈ C 2 ([0, 1] ; X) ∩ C ([0, 1] ; D (A))

  
u (0) ∈ D Hµ .





We now set for ω ∈ Sϕ0 and µ ∈ C set:

Aω = A − ωI and Hµ = H + µI.
Our main assumption on the Aω operator is the following


 ∃ϕ0 ∈ (0, π) : Sϕ0 ⊂ ρ (A) and ∃CA > 0 :



CA (35)
 ∀ω ∈ Sϕ0 , (A − ωI)−1 L(X) 6


 .
1 + |ω|
It is well known that assumption (35) implies the √
same properties, √for all ω > ω0 , on the other
hand it is well known that the square roots Q = − −A and Qω = − −A + ωI are well definied
and generate analytic semigroups not strongly contiuous at zero, see Martinez [4]. For non dense
domains, note that, D (A) = D (Q).

Main results
We obtain the following theorem:
Theorem: Assume (35). Let f ∈ C θ ([0, 1] ; X), with 0 < θ < 1 and d0 , u1 ∈ X. Then, for any ω > ω1∗ ,
we have

1. Problem (33)−(34) has a unique strict solution u if and only if

 −1  

 Qω − Hµ [d0 − Qω −1 f (0)] ∈ D Q2


2 Q − H −1 [d − Q−1 f (0)] + f (0) ∈ D (Q)

  
Qω

0
 ω µ
  ω


 u ∈ D Q 2


 1
Q2 u1 + f (1) ∈ D (Q)


ω

2. Problem (33)−(34) has a unique strict solution u satisfying the maximal regularity property
u 00 , Aω u ∈ C θ ([0, 1] ; X) if and only if
 −1  

 Qω − Hµ [d0 − Qω −1 f (0)] ∈ D Q2


2 Q − H −1 [d − Q−1 f (0)] + f (0) ∈ D (θ; +∞)

  
Qω

0
 ω µ Q
  ω


 u ∈ D Q 2


 1
Q2 u1 + f (1) ∈ DQ (θ; +∞)


ω

31
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] M. Cheggag, A. Favini, R. Labbas, S. Maingot, and A. Medeghri.,Abstract differential
equations of elliptic type with general Robin boundary conditions in Hölder spaces, Applicable
Analysis, Vol. 91, No. 8 (2012), pp. 1453-1475.

[2] A. Favini, R. Labbas, S. Maingot and A. Thorel, Elliptic differential-operator with an ab-
stract Robin boundary condition containing two spectral parameters, study in a non commuta-
tive framework, To appear.

[3] R. Haoua, R. Labbas, S. Mangot and M. Medeghri, New results on abstract elliptic problems
with general Robin boundary conditions in Hölder spaces: non commutative cases, Bollettino
dell’Unione Matematica Italiana.

[4] Martinez and M. Sanz, The Theory of Fractional Powers of Operators, North Holland, Math-
ematics studies 187, 2001.

32
CNMA’2022
Second National Conference on
Mathematics and its Applications

Multiple solutions for fractional

Schrödinger-Poisson system with
superlinear terms

Boutebba Hamza 1 hamboutebba@[Link]

Lakhal Hakim1
1 Department of mathematics , University 20 August 1955

Abstract: This paper is devoted to study the existence of multiple nontrivial solutions for
nonlinear fractional Schrödinger-Poisson system involving the s-gradient operator. By using
variational techniques and the symmetric Mountain pass theorem under certain assumptions
on V (x) and K(x) , we obtain the existing result.

Keywords: Schrödinger-Poisson system, Symmetric Mountain Pass Theorem, Palais-Smale

condition, Multiple solutions, s-gradient.
2010 Mathematics Subject Classification: Primary 35J50 , 35G50, 35R11.

Introduction
In this paper, we are concerned with the following fractional Schrödinger-Poisson system involv-
ing the s-gradient operator

s s 3
−D .(D u) + V (x)u + K(x)φu = f (x, u) in R ,


(36)
−D s .(D t φ) = K(x)u 2 in R3 .



where s, t ∈ (0; 1], 2t + 4s > 3, −D s .(D s ) and−D t.(D t ) denotes the s-gradient operator. Here the
s-gradient operator D s for s ∈ (0, 1), if u ∈ Lp RN and for some 1 < p < ∞, is defined by
∂s u ∂
(D s u)i = = I ∗ u, 0 < s < 1, i = 1, ..., N ,
∂xis ∂xi 1−s
∂
where is defined in the distributional sense and Is denotes the Riesz potential of order s.
∂xi
The s-gradient (D s ) can be written in integral form for smooth functions u ([1],[4])
Z
s u(x) − u(y) x − y
D u(x) = γ(N , s) N +s |x − y|
dy,
RN |x − y|

Moreover, D s has nice properties for u ∈ Cc∞ (RN ), it coincides with the fractional Laplacian as
follows:
N
s
X ∂s ∂s
(−∆) u = − u
∂xis ∂xis
j=1
= −D .D s u.
s

33
CNMA’2022
Second National Conference on
Mathematics and its Applications

The following assumptions are needed:

(f1 ) : There existe constant C > 0, p ∈ (2; 2∗s ) such that

|f (x, t)| ≤ C(|t| + |t|p−1 ),

6
where 2∗s = 3−2s the fractional critical Sobolev exponent .
(f2 ) : f (x, −t) = −f (x, t), x ∈ R3 ,t ∈ R.
(f3 ) : lim F(x, t)/t 4 a.e x ∈ R3 , and there exists r1 > 0 such that
|t|→∞

0 ≤ F(x, t),

Zt
for |t| ≥ r1 , where F(x, t) = f (x, s)ds.
0
(f4 ) : There exists L > 0 such that

1
0 ≤ f (x, t) − F(x, t) f or |t| ≥ L.
4

(K) : K(x) ∈ L6/(4s+2t−3) (R3 ) L∞ (R3 ), s, t ∈ (0, 1) and K(x) ≥ 0, for any x ∈ R3 .
S

(V ): V ∈ C R3 , R , infx∈R3 V (x) ≥ V0 > 0, where V0 is a constant, for every M > 0
n o
meas x ∈ R3 V (x) ≤ M < +∞.

Theorem 15. Assume that the system (36) satisfies (f1 )-(f4 ),(V ) and (K). then, (36) has multiple non-
trivial solutions.

Main results
First, we recall some definition for homogeneous fractional Sobolev space and Bessel potential
space.
The homogeneous fractional Sobolev space D s,2 (RN ) for s ∈ (0, 1), is defined by
n ∗ o
D s,2 (RN ) = u ∈ L2s (RN ) : D s u ∈ L2 (RN ) .

which is the completion of Cc∞ (RN ) with respect to the norm

Z 1
2
s 2
kukD s,2 (RN ) = ( |D u| dx) .
RN

Due to the fact that the problem (36) involves the s-gradient, we will defined the Bessel potential
space Ls,2 (RN ) for s > 0 as
Ls,2 (RN ) := gs {(L2 (RN )}.
The norm of this Bessel space is

kukLs,2 (RN ) = kf kL2 (RN ) .

We can reduce (36) in the fractional Schrödinger equation

−D s .(D s u) + V (x)u + K(x)φut u = f (x, u) x ∈ R3 . (37)

34
CNMA’2022
Second National Conference on
Mathematics and its Applications

Now, we defined our working space

 

   Z 

s,2 3 s 2 2
 
E= (u ∈ L : |D u| + V (x)u )dx < +∞ .
 
 R 


 

R3

Which is a Hilbert space equipped with the norm and the inner product
Z
kukE = (|D s u|2 + V (x) |u|2 )dx.
2

R3
Z
hu, wiE = (D s u.D s w + V (x) uw)dx.
R3
 
The energy functional I : Ls,2 R3 → R associated to (36) is defined as
Z  Z Z
1 s 2 2
 1 t 2
I (u) = |D u| + V (x)u dx + K(x)φu u dx − F (x, u) dx.
2 4
R3 R3 R3

We choose {e}i an orthonormal basis of E and define Xi = Rei ,

Yk = ⊕ki=1 Xi Zk = ⊕∞
i=k Xi k ∈ Z.

Clearly, E = Yk ⊕ Zk .
To prove Theorem15, we need the following Symmetric Mountain-Pass theorem:

Theorem 16. (See [2]) Assume that E = Yk ⊕ Zk be a Banach space where Y is finite dimensional, let
I ∈ C 1 (E, R) be even, satisfies the (PS)c condition,
 if
(i) there exist constants ρ, δ > 0 suatisfaying I ∂B T Z ≥ δ,
ρ

(ii) for any finite dimensional subspace Ẽ ∈ E, there is a constant C = C(Ẽ) > 0 such that max I(u) <
u∈Ẽ,kuk≥C
0,
then, I admits an unbounded sequence of critical values.

References
[1] Di Nezza, Eleonora, Giampiero Palatucci, and Enrico Valdinoci. "Hitchhiker’s guide to the
fractional Sobolev spaces." Bulletin des sciences mathématiques 136.5 (2012): 521-573.

[2] Rabinowitz, Paul H., ed. Minimax methods in critical point theory with applications to
differential equations. No. 65. American Mathematical Soc., 1986.

[3] Shieh, Tien-Tsan, and Daniel E. Spector. "On a new class of fractional partial differential
equations." Advances in Calculus of Variations 8.4 (2015): 321-336.

[4] Shieh, Tien-Tsan, and Daniel E. Spector. "On a new class of fractional partial differential
equations II." Advances in Calculus of Variations 11.3 (2018): 289-307.

35
CNMA’2022
Second National Conference on
Mathematics and its Applications

Existence results for subcritical and critical

p-fractional elliptic equations via Nehari
manifold method

Djamel Abid 1 contact@[Link]

Kamel Akrout 1
1 Department of mathematics and computer scinces , University of laarbi tebessi tebessa

Abstract: This paper deals with the existence and multiplicity of nonnegative solutions to the
following p-fractional Laplacian problem



 (−∆)sp u(x) = λ |u|p−2 u + f (x, u) + µg(x, u) in Ω, u > 0,


(E)





 u = 0 on Rn \ Ω,


where Ω ⊂ Rn (n > ps), is a bounded smooth domain, s ∈ (0, 1), λ, µ are positive parameters, and
f , g : Ω × [0, ∞) −→ R, are continuous functions. Using variational methods, especially, fibering
maps and Nehari manifold, we obtain existence results for either, subcritical and critical cases.
The results of the present paper, extend previous works which have recently appeared in the
literature.

Keywords: Nehari manifold , fibering maps, multiplicity of solutions..

2010 Mathematics Subject Classification: 35P30, 35J35, 35J60.

Introduction
Our first result about the sub-critical and concave case is the following.

Theorem 17. Let s ∈ (0, 1). Assume that the nonlinearities f , g are continuous satisfying homogenous
conditions. If
0 < r < 1 < p < q < ps∗ − 1, and n > ps.
Then, for all λ ∈ (0, λ1 ), there exists µ∗ (λ) > 0, such that, for all µ ∈ (0, µ∗ (λ)), problem (E) has at least
two positive solutions.

The second main result of this paper is devoted to the critical case (q = ps∗ − 1). Since the embed-
∗
ding X0 ,→ Lps (Rn ), is not compact, then the energy functional does not satisfy the Palais-Smale
condition globally, but it is true for the energy functional in a suitable range related to the best
fractional critical Sobolev constant.

36
CNMA’2022
Second National Conference on
Mathematics and its Applications

Theorem 18. Assume that s ∈ (0, 1), n > ps and 0 < r < 1 < p < q = ps∗ − 1. If there exist t0 > 0 and
u0 ∈ X0 \{0}, with u0 > 0 in Rn , such that
! n
1 p ps∗ s −n
A(u0 )t0 − t0 B(u0 ) < (ps∗ γ1 ) sps∗ Sp .
sp
(38)
p n

Then, for all λ ∈ (0, λ1 ), there exists µ∗ (λ) > 0, such that, for all µ ∈ (0, µ∗ (λ)), problem (E) has at least
two positive solutions.

Main results
Proof of Theorem (17) In order to prove Theorem (17), we need to present several results.
+
Proposition 1. There exists a minimizer uλ,µ in Nλ,µ for Jλ,µ satisfying:
+
(1) Jλ,µ (uλ,µ ) = αλ,µ < 0.

(2) uλ,µ is a solution of problem (E).

−
Proposition 2. If 0 < r < 1 < q < ps∗ − 1. Then, Jλ,µ has a minimizer vλ,µ in Nλ,µ satisfying
−
(1) Jλ,µ (vλ,µ ) = αλ,µ > 0.

(2) vλ,µ is a solution of problem (E).

Proof [Proof of Theorem (17)] By Propositions (1), (2), we get that problem (E) has two solutions uλ,µ ∈
+ − + −
Nλ,µ and vλ,µ ∈ Nλ,µ on X0 . Since Nλ,µ ∩ Nλ,µ = ∅, then, uλ,µ and vλ,µ are distinct. This completes
the proof of Theorem 17. 

Proof of Theorem (18)

Proposition 3. Assume that 0 < r < 1 < q = ps∗ − 1. Then, every Palais smail sequence {uk } ⊂ X0 for Jλ,µ
at level c, with
n
!n !− r+1
s ∗ −n
∗ ps λ ps λ p−r−1 p−r−1
p
c < (ps γ1 ) Sp 1 −
sp s −M 1− µ , (39)
n λ1 λ1
has a convergent subsequence, where Sp is the Sobolev best emmbedding constant .

Proposition 4. There exist µ∗ (λ) > 0, t̃1 and ũ1 ∈ X0 , such that, for all (λ, µ) ∈ (0, λ1 ) × (0, µ∗ (λ)), we
have
!n !− r+1
s ∗ − spn∗ psn λ ps λ p−r−1 p−r−1 p
Jλ,µ (t̃1 ũ1 ) ≤ (ps γ1 ) s Sp 1 − −M 1− µ . (40)
n λ1 λ1
In particular,
n
! psn r+1
!− p−r−1
− s − n λ λ p
< (ps∗ γ1 ) sps∗ Sp 1 −
ps
αλ,µ −M 1− µ p−r−1 . (41)
n λ1 λ1

Proof of Theorem (18) By Propositions (3) and (4), there exists two sequences {uk+ } and {uk− } in X0 ,
such that

Jλ,µ (uk+ ) −→ αλ,µ

+ 0
, Jλ,µ (uk+ ) −→ 0,
and
Jλ,µ (uk− ) −
−→ αλ,µ 0
, Jλ,µ (uk− ) −→ 0.

37
CNMA’2022
Second National Conference on
Mathematics and its Applications

+
as k −→ ∞. We observe that from the analysis of fibering maps ϕu (t), we have αλ,µ < 0. Similar
to the proof of Propositions (1) and (2) and Theorem (17), problem (E) has two solutions uλ,µ ∈
+ − + −
Nλ,µ and vλ,µ ∈ Nλ,µ in X0 .Since Nλ,µ ∩Nλ,µ = ∅, then these two solutions are distinct. This finishes
the proof.

References
[1] B. Abdellaoui, E. Colorado, I. Peral; Effect of the boundary conditions in the behavior of
the optimal constant of some Caffarelli-Kohn-Nirenberg inequalities. Application to some
doubly critical nonlinear elliptic problems. Adv. Diff. Equations 11 (6)(2006), 667-720.

[2] G. Alberti, G. Bouchitte, P. Seppecher; Phase transition with the line-tension effect, Arch.
Rational Mech. Anal., 144 (1998), 1-46.

[3] A. Ambrosetti, H. Brezis, G. Cerami; Combined effects of concave and convex nonlineari-
ties in some elliptic problems. J. Funct. Anal. 122(1994), 519–543.

[4] B. Barrios, E. Colorado, R. Servadei , F. Soria; A critical fractional equation with concave-
convex power nonlinearities, Ann. Inst. H. Poincaré Anal. Non Linéaire 32 (4)(2015), 875-
900.

[5] P. W. Bates; On some nonlocal evolution equations arising in materials science, In Nonlin-
ear dynamics and evolution equations, pp. 13–52. Fields Inst. Commun. 48, Amer. Math.
Soc., Providence, RI (2006).

38
CNMA’2022
Second National Conference on
Mathematics and its Applications

Multiplicity of Solutions for Kirchhoff Type

Problem with Critical Exponent in R3

Hayat Benchira 1 benchirahayet@[Link]

1 Department of mathematics, University of Tlemcen, Laboratory of Analysis and Control of Partial Differential
Equationss, Djillali Liabes University

Abstract: In this study we deal with a Kirchhoff type problem involving critical exponent in
R3 . The main tool is variational methods, more precisely, by using the Ekeland’s variational
principle. We can find the first critical point with negative level . From the Mountain
[Link] we also obtain a critical point whose level is positive.

Keywords: , Critical Sobolev exponent, Ekeland variational principle, Kirchhoff equa-

tion, Mountain Pass. Theorem, Variational methods.
2010 Mathematics Subject Classification: Primary ,35IJ60 , 35J20.

Introduction
This work deals with the existence and multiplicity of solutions to the following Kirchhoff prob-
lem with the critical Sobolov exponent
  
2 u 5 3
 −L(kuk ) ∆u + µ |x|2 = f (x)u + λg(x) in R ,


(P ) 
 u ∈ H 1 (R3 )

,

u2
R  
where L(t) = at+b, a and b are two positive constants, λ is parameter, µ < 41 , kuk2 = R3
|∇u|2 − µ |x| 2 dx

is the norm in H 1 (R3 ), f belongs to L∞ (R3 ) and g belongs to H −1 (R3 ). H −1 (R3 ) is dual of H 1 (R3 ).
Here we give a brief sketch of the way how we get two distinct critical points of the energy func-
tional. First, we minimize the functional in a neighborhood of zero and use the Ekeland varia-
tional principal to find the first critical point which achieves a local minimum. Moreover, the level
of this local minimum is negative. Next around the zero point. using the Mountain Pass Theorem
we also obtain a critical point whose level is positive.

Main results
Our main result is the following theorem

Theorem 19. Let a < 0, b < 0, µ < 41 and g , 0 then there exists λ∗ such that problem (P ) has at least
two nontrivial solution for any λ ∈ (0, λ∗ ).

39
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] C.O. Alves. F J.S.A. Correa T.F. MaPositive solution for a quasilinear elliptic equation of kir-
choff type, Comput Math. App, 49 (2005), 85-93.

[2] [Link], P. H. Rabinowitz, Dual variational methods in critical point theory and
application, J. Funct. Anal 14.(1973) 349-381.

[3] S. G., Chena, [Link], Multiple solution for the nonhomogenous Kirchoff equation on RN,
Nonlinear Analysis: Real World Applications 14 (2013) 1477-1486.

40
CNMA’2022
Second National Conference on
Mathematics and its Applications

Blow-Up for a Stochastic Wave Equation

with Logarithmic Nonlinearity

Amina Benramdane 1 benramdane.2017@[Link] and

Fatna Bensaber 2 ensaberfatna@[Link]
1 Higher School of Management-Tlemcen,
2 Department of mathematics, University of Tlemcen,

Abstract: In the current paper, we state an initial boundary value problem of stochastic
viscoelastic wave equation with nonlinear damping and logarithmic nonlinear source terms.
We provide a result of blow-up type for the solution with decreasing kernel.

Keywords: Blow-up, Logarithmic source, Stochastic wave equation.

2010 Mathematics Subject Classification: Primary 60H15, 35L70, 35B44.

Introduction
In the late decades, stochastic partial differential equations in a separable Hilbert space have
been considered by many authors and several results on the existence, uniqueness, stability and
blow-up of solutions beside its other quantitative and qualitative properties.
In this work, we are interested in the following problem of stochastic wave equation :
 Rt


 utt − µ∆u − (λ + µ)∇(divu) + 0
h(t − s)∆u(s) ds


+|ut |q−2 u = u|u|p−2 ln|u|k + σ (x, t)Wt (x, t) in D×]0, +∞[,


 t
(42)


u(x, t) = 0 on ∂D × [0, +∞[,






u(x, 0) = u0 (x), ut (x, 0) = u1 (x) in D,



where D is a bounded domain in Rn , n ∈ N∗ , of a smooth boundary ∂D , µ, λ are the Lamé con-

stants such that µ > 0, λ + µ ≥ 0, h is a positive function, p > q ≥ 2, the constant k is a small
non-negative real number; and L2 (D) is the set of square integrable function on D equipped with
the inner product h., .i and its norm k.k2 .
W (x, t) is an infinite dimensional Wiener process, σ (x, t) is L2 (D) valued progressively measurable
and  is positive constant which measures the strength of noise.

Main results
Let (Ω, F , P ) be a complete probability space for which a filtration {Ft , t ≥ 0} of increasing sub
σ −fields Ft is given and W (x, t) be a continuous Wiener random field in this space with mean zero
and the covariance operator Q satisfying
X
T r(Q) = λi < ∞.
i≥1

41
CNMA’2022
Second National Conference on
Mathematics and its Applications

W (x, t) is defined by
∞ q
X
W (x, t) = λj βj (t)ej (t), j ∈ N∗ , t ≥ 0,
j=1

where βj (t) is a sequence of real-valued standard Brownian motions mutually independent on the
probability space (Ω, F , P ), λj are the eigenvalues of Q, and ej are the corresponding eigenvectors.
We need some assumptions.
(A1) Assume that h : R+ → R+ is a C 1 nonincreasing function satisfying
Z∞
h(0) > 0, µ − h(s)ds = l > 0
0

and there exist tow nonnegative constants ς1 and ς2 such that

−ς1 h(t) ≤ h0 (t) ≤ −ς2 h(t), t ≥ 0.

(A2) Z ∞
(p − 2)p
h(s)ds < µ .
0 (p − 1)2

(A3) p > q ≥ 2 and 

2(n−1)
 2 < p ≤ n−2 if n ≥ 3



 (43)
 2 < k ≤ +∞ if n = 1, 2.



We define the energy associated to the solution of system (42) by

Zt
1 2 1 λ+µ
 
e(t) = kut k2 + µ − h(s)ds k∇uk22 + kdivuk22
2 2 0 2
Z
1 k p 1
+ (h o ∇u)(t) + 2 kukp − |u|p ln|u|k dx
2 p p D

where Z t
(h o v)(t) = h(t − s)kv(., t) − v(., s)k2 ds.
0
We rewrite (42) as an equivalent Itô’s system




 du = vdt,

  Rt
dv = µ∆u + (λ + µ)∇(divu) − 0 h(t − s)∆u(s) ds





 

−|v|q−2 v + u|u|p−2 ln|u|k dt + σ (x, t)dWt (x, t) (44)



 in D×]0, +∞[,


u(x, t) = 0 on ∂D × [0, +∞[,








 u(x, 0) = u0 (x), v(x, 0) = u1 (x) in D,

We purpose Z ∞Z
E σ 2 (x, t)dxdt < ∞,
0 D
∞ Z tZ
2 X
G(t) = E λj ej2 (x)σ 2 (x, s)dxds,
2 0 D
j=1

42
CNMA’2022
Second National Conference on
Mathematics and its Applications

∞ ∞Z
2 X
Z
G(∞) = E λj ej2 (x)σ 2 (x, s)dxds
2 0 D
j=1 (45)
∞Z
2
Z
≤ T r(Q)c02 E σ 2 (x, s)dxds := E1 < ∞,
2 0 D

where
∞
X
T r(Q) = λj < ∞ and c0 = sup kej k∞ < ∞.
j=1 j≥1

Theorem 20. Assume (A1) and (A2) hold. Let (u, v) be a solution of system (44) with initial data
(u0 , v0 ) ∈ H01 (D) × L2 (D) satisfying
Ee(0) ≤ −(1 + β)E1 ,
where β is nonnegative constant and E1 is given (45). If p > q, then there exists a positive time T0 ∈ [0, T ]
such that

lim E(e(t)) = +∞,

t→T0−

where
1−α
T0 = α ,
αKL 1−α (0)
L(0) = H 1−α (0) + δEhu0 , u1 i > 0.

References
[1] A. Benramdane, N. Mezouar, M. S. Alqawba, S. M. Boulaaras and B. B.
Cherif, Blow-Up for a Stochastic Viscoelastic Lamé Equation with Logarithmic Non-
linearity, Journal of Function Spaces, 2021, (2021), Article ID 9943969, 10 pages.
[Link]

[2] S. Cheng, Y. Guo and Y .Tang, Stochastic viscoelastic wave equations with nonlinear damping
and source terms, J. Appl. Math, 2014, (2014).

[3] S. Kim, J. Y. Park and Y. H. Kang, Stochastic quasilinear viscoelastic wave equation with
nonlinear damping and source terms, Boundary Value Problems, 2018 No.1, (2018), 1-15.

43
CNMA’2022
Second National Conference on
Mathematics and its Applications

Elliptic Problem with Asymptotically linear

term

Senhadji Asma 1 [Link]@[Link]

Nasri Yasmina1
1 Department of mathematics ,University of Tlemcen ,BP 119 Tlemcen 13000, Algeria

Abstract: This paper is devoted to study the existence of positive solutions for an asymptoti-
cally linear problem involving the fractional Laplace using The Mountain Pass theorem.

Keywords: Fractional Laplacian, Asymptotically linear, Mountain Pass Theorem.

2010 Mathematics Subject Classification: AMS. 35R11, 35A15.

Introduction
The aim of this work is to study the following problem

s
(−∆) u = f (x, u) in Ω,


(P )
in RN \Ω,

u = 0


where Ω is a smooth bounded domain in RN with N > 2s, s ∈ (0, 1) and the function f (x, t) ∈
C(Ω̄ × R) satisfies:
(A1 ) ∀x ∈ Ω̄

f (x, t) > 0 ∀t > 0

f (x, t) ≡ 0 ∀t 6 0.
f (x,t)
(A2 ) t is nondecreasing with respect to t > 0, a.e. x ∈ Ω.
f (x, t) f (x, t)
(A3 ) lim = p(x), lim = q(x) . 0 uniformly in a.e. x ∈ Ω, where 0 6 p(x), q(x) ∈ L∞ (Ω)
t→0 t t→+∞ t
and kp(x)kL∞ < λ1 , λ1 > 0 is the first eigenvalue of ((−∆)s , H0s (Ω)).

The fractional Laplace operator is defined as

Z
s u(x) − u(y)
(−∆) u(x) := P .V . dy.
RN |x − y|N +2s

Problem (P ) has a variational nature, thus we can prove that it has a nontrivial positive solutions
as a critical points of the following energy functional
Z
1
J (u) = kuk2 − F(x, u)dx,
2

44
CNMA’2022
Second National Conference on
Mathematics and its Applications

where Z u
F(x, u) = f (x, t)dt.
0

Main results
We begin by defining the minimization problem

( Z )
2 s 2
Λ = inf kuk : u ∈ H0 (Ω), q(x)u dx = 1 ,

then Λ > 0, such that for some ψΛ ∈ H0s (Ω), ψΛ > 0 a.e. in (Ω) we have that Λ is achieved (Lemma
1).
Our main results are the following

Theorem Let f (x, t) satisfies the hypotheses (A1 ) to (A3 ), then we have
• If Λ > 1, problem (P ) has no positive solution.
• If Λ < 1, there is a positive solution to problem (P ).
• If Λ = 1, problem (P ) has a positive solution u ∈ H0s (Ω) if and only if there exists a positive
constant a such that u(x) = aψΛ (x) and f (x, u) = q(x)u a.e. x ∈ Ω.

λ1
Taking q(x) ≡ l > 0, it follows that Λ = l . Consequently, we have the following result

Theorem
Let f (x, t) satisfies the hypotheses (A1 ) to (A3 ) with q(x) ≡ l > 0. Then
• If λ1 > l, problem (P ) has no positive solution.
• If λ1 < 1, there is a positive solution to problem (P ).
• If λ1 = l, problem (P ) has a positive solution u ∈ H0s (Ω) if and only if there exists a positive con-
stant a such that u(x) = aψ1 (x) and f (x, u) = λ1 u a.e. x ∈ Ω, where ψ1 > 0 is the λ1 -eigenfunction.

References
[1] [Link], Lévy Processes. Cambridge Tracts in Mathematics. Cambridge University Press,
Cambridge 121 (1996).

[2] D. G. Costa and O. H. Miyagaki, Nontrivial solutions for perturbations of the p-Laplacian on
unbounded domains, J. Math. Anal. Appl. 193 (1995), 737-755.

[3] G. Li, Existence of positive solutions of elliptic mixed boundary value problem. Boundary Value
Problems, 2012 (2012) (1), 1-11.

[4] M. Willem, Minimax theorems. Birkhauser, (1996)

45
CNMA’2022
Second National Conference on
Mathematics and its Applications

Existence of solution for elliptic problems

with variable exponent and nonlinear
boundary conditions

Fareh Souraya1 [Link]@[Link]

1 Department of mathematics and computer science , Larbi Tebessi University-Tebessa

Abstract: In this work, we study a class of quasilinear elliptic problems involving the
p(x)-Laplacian operator and we deal with nonlinear conditions on the boundary. By using the
variational method and mountain pass lemma, we establish the existence of a nontrivial weak
solution for the problem.

Keywords: generalized Sobolev space, variable exponents, variational methods, moun-

tain pass theorem.
2010 Mathematics Subject Classification: Primary 31B30, 35J35, 35J48.

Introduction
The aim of this paper, is to study the existence of solutions for the following nonlinear boundary
value problem
  
p(x)−2


 −div a (x) |∇u| ∇u + |u|p(x)−2 u = λH (x) f (u) in Ω,


(46)





a(x) |∇u|p(x)−2 ∂u + b (x) |u|q(x)−2 u = T (x)g(u) on ∂Ω.


∂v
∂
where Ω ⊂ RN (N ≥ 2), is a bounded with Lipschitz boundary ∂Ω, ∂v
is the outer unit normal
derivative,
(−∆)p(x) u = −div(|∇u|p(x)−2 ∇u),
 
denotes the p(x)-Laplacian, p(x) ∈ C Ω , q (x) ∈ C (∂Ω) such that

1 < p− = infp (x) ≤ p (x) ≤ supp (x) = p+ ,

Ω Ω

1 < q− = infq (x) ≤ q (x) ≤ supq (x) = q+ ,

∂Ω ∂Ω

p(x), q(x) > 1, p(x) , q(y),for any x ∈ Ω, y ∈ ∂Ω, λ ≥ 0, a and b are continuous functions such that

a1 ≤ a(x) ≤ a2 , b1 ≤ b(x) ≤ b2

where a1 , a2 , b1 and b2 are positive constants, H : Ω → R, f : R → R, T : ∂Ω → R, g : R → R are

measurable functions satisfying some conditions. By using variational methods and mountain

46
CNMA’2022
Second National Conference on
Mathematics and its Applications

pass Lemma, and for some hypothesis, we prove the existence of a nontrivial weak solutions for
the above problem.
We have the mountain pass theorem which is the main tool to prove our main result
Theorem 21. (Mountain pass theorem). Let X be a Banach space, ϕ ∈ C 1 (X, R) and e ∈ X with ||e|| > r
for some r > 0. Assume that
inf ϕ(u) > ϕ(0) ≥ ϕ(e).
||u||=r

If ϕ satisfies the (P S) condition at level c, then, c is a critical value of ϕ, where

c = inf max ϕ(γ(t)), and Γ = {γ ∈ C([0, 1], X) : γ(0) = 0, γ(1) = e}.

γ∈Γ t∈[0,1]

Main results
(A1 ) There exist C1 > 0, α, S ∈ C(Ω), such that for all (x, u) ∈ Ω × R, we have
S(x)
H (x) ∈ L S(x)−α(x) (Ω) , f (u) ≤ C1 |u|α(x)−1

and
1 < α (x) < S (x) < p∗ (x) .

(A2 ) There exist C2 > 0, β, Q ∈ C(∂Ω), such that for all (x, u) ∈ ∂Ω × R, we have
Q(x)
T (x) ∈ L Q(x)−β(x) (∂Ω) , g (u) ≤ C2 |u|β(x)−1

and
1 < β (x) < Q (x) < p∗ (x) , q (x) < p∗ (x) .
(A3 ) There exist K1 > 0, θ1 > p+ such that for all x ∈ Ω, we have

0 < θ1 H (x) F (u) ≤ H (x) f (u) u, |u| ≥ K1 ,

Ru
where F (u) = 0
f (s) ds.
(A4 ) There exist K2 > 0, θ2 > p+ such that for all x ∈ ∂Ω, we have

0 < θ2 T (x) G (u) ≤ T (x) g (u) u, |u| ≥ K2 ,

Ru
where G (u) = 0
g (s) ds We give below our main result
Theorem 22. Let min(α − , β − ) > p+ , min(θ1 , θ2 ) > q+ , λ ≥ 0 and assume that hypothesis (A1 ) − (A4 ) are
satisfied, then, problem (46) has a nontrivial weak solution.

References
[1] Z. Yücedag, Existence results for Steklov problem with nonlineair boundary condition, Middle
east journal of science. 5(2)(2019), 2618-6136.

[2] Z. Wei and Z. Chen, Existence results for the p(x)-Laplacian with nonlinear boundary condi-
tion, Applied Math., Article ID 727398, doi:10.5402/2012/727398, 2012.

[3] Chammem, R., A. Ghanmi, and A. Sahbani., Existence and multiplicity of solutions for some
Styklov problem involving p (x)-Laplacian operator, Applicable Analysis 101.7 (2022): 2401-
2417

47
CNMA’2022
Second National Conference on
Mathematics and its Applications

GENERAL STABILITY FOR MOORE GIBSON

THOMPSON PROBLEM WITH A NONLINEAR
DAMPING TERM

Hassan Messaoudi1 hassanmessaoudi50conf@[Link]

1 Department of Mathematics and Informatics, Laboratory of Informatics and Mathematics (LIM). Mohamed-
Cherif Messaadia University - Souk Ahras , 41000, Algeria

Abstract: In this work, we study the asymptotic behavior of one-dimensional Moore-Gibson-

Thompson equation with a nonlinear damping term. First, we show the existence and
uniqueness of the solution by the semi-group method, more precisely by the Hille-Yosida
theorem. And by building an appropriate Lyapunov functional, we establish a general decay
result for the solutions of the system.

Keywords: Moore-Gibson-Thompson equation, Nonlinear damping, Semigroup approach,

Lyapunov functional, Energy method, General decay.
2010 Mathematics Subject Classification: 35B40, 35B35.

Introduction
In recent years, there has been a lot of interest in the Moore-Gibson-Thompson equation (abbrevi-
ated MGT), and numerous papers have been written to investigate and comprehend it . Moreover,
Moore–Gibson–Thompson (MGT) equation is based on the modeling of high amplitude sound
waves. There has been quite a bit of work in this area of research due to a wide range of ap-
plications such as medical and industrial use of high intensity ultrasound in lithotripsy, heat
therapy, ultrasonic cleaning, etc. In [6], Marchand et al. presented a detailed analysis of this
equation. Using a quasi-abstract group approach and refined spectral analysis, they establish the
well posedness of the problem and define the accumulation point of eigenvalues. Kaltenbacher
et al. [3] also studied the fully nonlinear version of the MGT equation and established the global
well posedness and the exponential decay for the nonlinear equation under consideration. While,
S. Boulaaras, A. Zarai and A. Draifia investigated the Moore–Gibson–Thompson equation with
integral condition in [4]. A. Choucha et al. Considered the following problem
 Rt
u + αu + βu − %∆u − δ∆u − γ∆u + h (t − σ ) ∆u (σ ) dσ = F (x, t) ,

tttt ttt tt t tt

0




u(x, 0) = u0 (x), ut (x, 0) = u1 (x), utt (x, 0) = u2 (x), uttt (x, 0) = u3 (x), ,



 R t R
 ∂u =

u (ξ, τ) dξdτ, x ∈ ∂Ω.


∂η 0 Ω
Rt
The convolution term 0 h (t − s) ∆u (s) ds reflects the memory effect of materials due to vicoelastic-
ity, F is a given function and h is the relaxation function. The authors, by the Galerkin method, im-
proved the existence and uniqueness of the Fourth-Order Equation of Moore–Gibson–Thompson

48
CNMA’2022
Second National Conference on
Mathematics and its Applications

Type in the source term and integral condition. In [4], I. Lasiecka et al. Considered the following
problem
Rt
τuttt + αutt − c2 Au − bAut − 0 g (t − s) Aw (s) ds = 0 ,

where τ, α, b, c2 are physicalR parameters and A is a positive self-adjoint operator on a Hilbert space
t
H. The convolution term 0 g (t − s) Aw (s) ds reflects the memory effects of materials due to vis-
coelasticity. The authors studied the exponential decay of energy of the temporally. W Chen and A
Palmieri studied the blow–up result for the semilinear Moore– Gibson–Thompson equation with
nonlinearity of derivative type in the conservative case defined as following

βuttt + utt − ∆u − β∆ut = |ut |p , x ∈ Rn , t > 0. ,

In [3], M. Pellicer et al. Considered the following problem

(u + αut )tt + L (u + βut ) = 0, with α, β > 0 , (47)

where L is a self-adjoint, strictly positive operator in a Hilbert space H with compact resolvent.
Recall that in this situation the eigenvalues µn of L are strictly positive, increasing, tending to ∞,
semi-simple and the corresponding eigenfunctions φn are an orthonormal family. A typical case
is when L = −a2 ∆ and ∆ is Laplace’s operator with Dirichlet boundary conditions in a bounded
domain Ω with a regular boundary, and in that case H = L2 (Ω). When L = −a2 ∆, this equation is
known as the Moore-Gibson-Thompson equation. The authors showed the existence of a suitable
scalar product associated to the semigroup defined by (47) for some ranges of the parameters.
In this work, we study the asymptotic behavior of Moore–Gibson–Thompson equation with non-
linear damping term. First, by using the semi-group method we show that the existence and
uniqueness of the solution. Also by using some properties of convex functions and Lyaponov
functional we obtain general stability estimates.

Main results
In this work we study the asymptotic behavior of a one-dimensional Moore-Gibson-Thompson
equation with a nonlinear damping term. We show the well-posedness of the system, and we
proved stability estimates by means of appropriate Lyapunov functions. My work is divided as
follows. In section 2, we use the semigroup method, more precisely the Hille-Yosida theorem, to
prove the well-posedness. In section 3, we state and demonstrate some technical lemmas needed
in the proof of our main result. Section 4 is devoted to the statements and proofs of our stability
results.

References
[1] S. Boulaaras, A. Zarai, A. Dhraifia, Galerkin method for nonlocal mixed boundary value
problem for the Moore-Gibson-Thompson equation with integral condition, Mathematical
Methods in the Applied Sciences. 42 (2019), 2664-2679.
[2] B. Kaltenbacher, I. Lasiecka, M. K. Pospieszalska, Well-posedness and exponential decay of
the energy in the nonlinear Jordan-Moore-Gibson-Thompson equation arising in high inten-
sity ultrasound, Mathematical Models and Methods in Applied Sciences. 22 (2012), 1250035.
[3] I. Lasiecka, X. Wang, Moore-Gibson-Thompson equation with memory, part I: exponential
decay of energy, Zeitschrift für angewandte Mathematik und Physik. 67 (2016), 1-23.

49
CNMA’2022
Second National Conference on
Mathematics and its Applications

[4] R. Marchand, T. MCDevitt, R. Triggiani, An abstract semigroup approach to the third-order

Moore-Gibson-Thompson partial differential equation arising in high-intensity ultrasound:
structural decomposition, spectral analysis, exponential stability. Mathematical Methods in
the Applied Sciences. 35 (2012), 1896-1929.

[5] M. Pellicer, J. Solà-Morales, Optimal scalar products in the Moore-Gibson-Thompson equa-

tion, Evolution Equations and Control Theory. 8 (2019), 203-220.

50
CNMA’2022
Second National Conference on
Mathematics and its Applications

General decay of solutions for the wave

equation with a delay and a boundary
feedback of memory-type

Hocine Makheloufi 1 [Link]@[Link]

1 Faculty of Exact Sciences, Mascara University

Abstract: This paper studies the asymptotic stability of a non-dissipative wave equation with
a delay and a boundary condition of a memory type. Under a suitable condition on the weight
of the delayed feedback and a wider class of relaxation function, we establish a general decay
result by introducing a suitable Lyapunov functionnal.

Keywords: Wave equation, delay term, viscoelastic damping, non-dissipative term, mul-
tiplier method.
2010 Mathematics Subject Classification: Primary 35B40; Secondary 35L70.

Introduction
In this research work we investigate the stability problem of the one dimensional wave equation

utt (x, t) − uxx (x, t) + aut (x, t − τ) + Φ(ux (x, t)) = 0, (48)

where a is a real number, (x, t) ∈ (0, ∞), t represents the time variable and τ > 0 is the time delay.
Moreover, h is a C 1 (R) function. System (48) is subjected to the following boundary conditions:

u(0, t) = 0,


 (49)
u(1, t) = −(g ∗ u)(1, t).


where g is a positive non-increasing function defined on (0, ∞) and ∗ is the usual convolution
product Zt
(ϕ ∗ ω) = ϕ(t − s)ω(s)ds
0
Also, we consider the following initial conditions

u(x, 0) = u0 (x), ut (x, 0) = u1 (x) in (0, 1),


 (50)
ut (x, t − τ) = f0 (x, t − τ)
 in (0, 1) × (0, τ),

where the initial data (u0 , u1 , f0 ) are taken in a suitable Sobolev space.

We start our investigation by introducing, as in the work [4], the new variable

z(x, ρ, t) = φt (x, t − ρτ), x ∈ [0, 1], ρ ∈ [0, 1], t > 0,

51
CNMA’2022
Second National Conference on
Mathematics and its Applications

which satisfies
τ zt (x, ρ, t) + zρ (x, ρ, t) = 0 in ([0, 1])2 × [0, ∞].
Hence, our problem (48) − (50) becomes



 utt (x, t) − uxx (x, t) + az(x, 1, t) + Φ(ux (x, t)) = 0 in ]0, 1[ ×]0, ∞[,
in (]0, 1[)2 ×]0, ∞[,

τ z (x, ρ, t) + zρ (x, ρ, t) =0


 t



u(0, t) = 0 in ]0, ∞[,


 (51)
u(1, t) = −(g ∗ u)(1, t) in ]0, ∞[,




u(x, 0) = u0 (x), ut (x, 0) = u1 (x) in ]0, 1[,





(]0, 1[)2 .

z(x, ρ, 0) = f (x, −ρτ)
0 in

Due to the condition (P )4 , we introduce the following space

 
H?1 (0, 1) = u : u ∈ H 1 (0, 1) and u(0, .) = 0 .

We now consider the following assumptions:

(A1 ) Φ : R −→ R is a C 1 function such that Φ(0) = 0 and

|Φ 0 (s)| ≤ α. (52)

(A2 ) k : [0, ∞[−→ [0, ∞[ is a C 2 non-increasing function satisfying the following conditions:

k(0) > 0, k 00 (t) ≥ −ζ(t)k 0 (t), ∀t ≥ 0. (53)

where ζ : R+ →]0, +∞[ is a C 0 function which is not necessarily monotone such that it exist two
fixed positive constants c1 and c2 satisfying

c1 ≤ ζ(t) ≤ c2 , ∀t ≥ 0. (54)

To study the stability problem, we will use the following Lemma.

Lemma 4. [5] Let L : R+ −→ R+ be a C 1 function. Assuming that there exist positive constants λ0 , λ1 ,
λ2 and a continous and bounded function ζ : R+ →]0, +∞[ which is not necessarily monotone such that
Zt !
0
L (t) ≤ −λ0 L(t) + λ1 exp − λ2 ζ(s)ds ,
0

then Z t !
L(t) ≤ C(L(0))exp − c ζ(s)ds .
0

Stability result

In this section, we investigate the asymptotic stability of our problem by the use of the energy
method. At the first, we define the modified energy functional of the problem (P ) as

1 1 2
Z " Z1 #
2 δ δ
E(t) = u + ux + ξτ z (x, ρ, t)dρ dx + k(t)u 2 (1, t) − (k 0 ◦ u)(1, t).
2 (55)
2 0 t 0 2 2

52
CNMA’2022
Second National Conference on
Mathematics and its Applications

Then, we have
δ δ δ
E 0 (t) ≤ − ut2 (1, t) − (k 00 ◦ u)(t) + k 0 (t).u 2 (1, t) − k 2 (t)|u0 (1)|2
2 2 2
Z1 (56)
|a| − ξ 1 2
Z
|a| + ξ 2
+ αE(t) + ut dx + z (x, 1)dx.
2 0 2 0

The main result of this work is:

Theorem 23. Let (u, z) be the solution of (P ). Assuming that (A1 )-(A2 ) hold and that (u0 , u1 , f0 ) ∈
H?1 (0, 1) × L2 (0, 1) × L2 (0, 1, (0, 1)). Then, for α and |a| small enough, it exist two positive constants η1
and η2 such that the solution of (P ) satisfies the following decay property
Zt !
E(t) ≤ η1 exp − η2 ζ(s)ds , ∀t ≥ 0. (57)
0

Proof The proof the stability result stated in Theorem 23 is based on the multiplier method. For,
we introduce a Lyapunov L as

L(t) = E(t) + N1 F1 (t) + N2 F2 (t),

where Z 1 Z 1Z 1
F1 (t) = xut ux ds and F2 (t) = τ z2 (x, ρ, t)dρdx.
0 0 0
and the L satisfies for some positive constants c3 , c4 , c5

c3 E(t) ≤ L(t) ≤ c4 E(t), (58)

L0 (t) ≤ −c5 E(t) + cK 2 (t)|u0 (1)|2 , (59)

Next, we consider the following two cases:

• The first case: u0 ≡ 0.
From (59) and using (53), we get
L0 (t) ≤ −cζ(t)E(t).
A simple integration over (0, t), using (58), gives (57).
• The second case: u0 , 0.
In that case we have to assume that

k 0 (t) ≤ −ζ(t)k(t), ∀t ≥ 0.

Using this assumption and (58), we obtain from (59) that

Z t
0 0
L (t) ≤ −cL (t) + cexp ζ(s)ds.
0

Thanks to Lemma 4, we get Z t !

L(t) ≤ cexp − ζ(s) ds.
0

53
CNMA’2022
Second National Conference on
Mathematics and its Applications

The use of (58) again leads us to (57). 

References
[1] M.L. Santos. Decay rates for solutions of a system of wave equations with memory., E. J.
Diff. Eqs, 38, 1-17 (2002).

[2] S. A. Messaoudia and A. Soufyane. General decay of solutions of a wave equation with a
boundary control of memory type., Nonlinear Anal, 11, 2896-2904 (2010).

[3] M.I. Mustafa. The control of Timoshenko beams by memory-type boundary condition.,
App Anal, 100(2), 290-310, (2019)

[4] K. Ammari, S. Nicaise, C. Pignotti. Feedback boundary stabilization of wave equations with
interior delay., Systems Control Lett. 59, 623-628, (2010).

[5] H. Makheloufi. Stydy of the stability of some elastic systems by a boundary feedback and
a delay term., phd thesis, 2022.

54
CNMA’2022
Second National Conference on
Mathematics and its Applications

Etude d’un problème viscoplastique avec

frottement

Kendri Dalila 1 [Link]@[Link]

1 Department of mathematics , University Batna2

Abstract: Dans ce travail on s’intéresse à l’étude d’un problème quasistatique de contact

unilatéral avec frottement d’un corps déformable avec une base rigide pour lequel on étabilra
une formulation variationelle Pv . Le problème Pv est obtenu à partir du problème P en
utilisant la formule de Green ainsi que la loi de comportement. Nous démontrons un résultat
d’existence et d’unicité de la solution en utilisant une technique de discrétisation temporelle,
une méthode de point fixe ainsi que la théorie des inéquations variationnelles elliptiques.

Keywords: corps viscoplastique, inégalités variationnelles, contact avec frottement.

2010 Mathematics Subject Classification: Primary 73C35,49J40.

Position du problème et formulation variationnelle

On considère un corps viscoplastique occupant un domaine de RN (N = 2, 3) qu’on note Ω. La
frontière Γ , supposée suffisamment régulière, est divisée en trois parties mesurables disjointes:
Γ1 , Γ2 et Γ3 . Soit T > 0, on suppose que mes Γ1 > 0 et que le champ des déplacements s’annule sur
Γ1 ×[0, T ]. On suppose aussi que des tractions g agissent sur Γ2 ×[0, T ] et que des forces volumiques
f agissent dans Ω×[0, T ].
Nous étudions le cas où la loi de Comportement est viscoplastique de la forme:

σ̇ = ξ(ε(u̇)) + G(σ , ε(u)),

les déplacements et les contraintes normaux vérifient les conditions de Signorini suivantes :

uν 6 0, σν ≤ 0, σν uν = 0.
La loi de frottement considérée est la suivante:



|στ | 6 µp(|Rσν |)

|στ | < µp(|Rσν |) ⇒ u̇τ = 0 sur Γ3 ×[0, T ] (60)




|σ | = µp(|Rσ |) ⇒ ∃λ ≥ 0

tel que στ = −λu̇τ
τ ν

dans (60) l’opérateur R représente une régularisante normale et uτ , et στ sont les déplacements et
les contraintes tangentielles et µ > 0 est le coefficient de frottement.
La nouvelle version de la loi de frottement de Coulomb, consiste à prendre la fonction p telle que:

p(r) = r(1 − αr)+

où α est un coefficient positif assez petit lié à la dureté de la surface de contact et r+ = max {0, r}.
Notre problème se formule de la manière suivante:

55
CNMA’2022
Second National Conference on
Mathematics and its Applications

Problème P: Trouver le champ des déplacement u : Ω × [0, T ] → RN et le champ des contraintes

σ : Ω × [0, T ] → SN tel que:

σ̇ = ξ(ε(u̇)) + G(σ , ε(u)) dans Ω × [0, T ]

Divσ + f = 0 dans Ω × [0, T ]

u=0 sur Γ1 × [0, T ]

σν = g sur Γ2 × [0, T ]

uν 6 0, σν ≤ 0, σν uν = 0


|σ | 6 µp(|Rσν |)
 τ



|σ | < µp(|Rσν |) ⇒ u̇τ = 0 sur Γ3 ×[0, T ]

 τ


|σ | = µp(|Rσ |) ⇒ ∃λ ≥ 0

tel que στ = −λu̇τ
τ ν

u(0) = u0 et σ (0) = σ0 dans Ω.

Pour l’étude de problème P, on considère les hypothèses suivantes:



 le tenseur ξ : Ω × SN → SN est un tenseur défini positif c’est à dire:

 ∞
 a) ξijκl ∈ L (Ω) pour tout i, j, κ, l = 1, N ;


 (61)



 b) ξσ .τ = σ .ξτ pour tout σ , τ ∈ SN ;
 c) il existe m > 0 tel que ξσ .σ ≥ m|σ 2 | pour tout σ ∈ SN






 l’opérateur G : Ω × SN × SN → SN vérifie:

a) il exist L > 0 tel que:






 |G(., σ1 , ε(u1 )) − G(., σ2 , ε(u2 ))| ≤ L(|σ1 − σ2 | + |u1 − u2 |)


 (62)
 pour tout σ1 , ε1 , σ2 , ε2 ∈ SN , p.p dans Ω




b) G(., σ , ε) est une fonction Lebesgue mesurable sur Ω pour tout σ , ε ∈ SN






 c) G(., 0, 0) ∈ H





 La fonction de frottement p : Γ3 × R+ → R+ vérifie:




 (a) il exist M > 0 tel que: |p(x, r1 ) − p(x, r2 )| ≤ M|r1 − r2 |

pour tout r1 , r2 ∈ R+ p.p sur Γ3 (63)




(b) x 7→ p(x, r) est mesurable sur Γ3 pour tout r inR+






(c) p(x, 0) = 0 p.p sur Γ3 .


On suppose aussi que:

f ∈ W 1,∞ (0, T , H), g ∈ W 1,∞ (0, T , L2 (Γ2 )N ), (64)
tandis que µ vérifie
µ ∈ L∞ (Γ3 ) µ(x) ≥ 0 p.p sur Γ3 (65)

56
CNMA’2022
Second National Conference on
Mathematics and its Applications

On considère le sous espace fermé V de H1 défini par:


V = u ∈ H1 /γµ = 0 sur Γ1 (66)

En construisant une forme équivalente des conditions aux limites et en moyennant les hypothèses
au dessus ainsi que quelques techniques de calcul, nous obtenons la formulation variationnelle
du problème mécanique P comme suit:

Problème Pv : trouver le champ des déplacement u : Ω × [0, T ] → RN et le champ des contraintes

σ : Ω × [0, T ] → H tel que:
σ̇ = ξ(ε(u̇)) + G(σ , ε(u))



 u ∈ Uad

hσ (t), ε(ν) − ε(u̇(t))iH + j(σ (t), ν) − j(σ (t), u̇(t))





≥ hl(t), ν − u̇(t)i + R σ (ν − u̇ (t)) ∀ν ∈ V t ∈ [0, T ]



 R V Γ3 ν ν ν
σ (ν − uν (t)) ≥ 0 ∀ν ∈ Uad


Γ3 ν ν




u(0) = u


0
P
La formulation Pv est établie pour le couple (u, ν) appartenant à Uad × V où Uad et ad (g, t) sont
respectivement les ensembles des déplacements et des contraintes admissibles définis par:

Uad = {z ∈ V /νν 6 0 sur Γ3 }

X ( Z )
(g, t) = τ ∈ H/ hτ, ε(ν)iH + j(g, ν) ≥ (l, ν)V + gν νν ∀r ∈ V g ∈ H1
ad Γ3

Résultats d’existence et d’unicité- Résultats d’équivalence:

Dans cette section, on donne le résultat d’existence et d’unicité pour le problème variationnel
Pv .
Theorem 24. Sous les hypothèses de (62) à (66). il existé une constante α0 > 0 dépendant de Ω, Γ , G, ξ
et p tel que: si α < α0 alors Pv ; admet une solution unique ayant la régularité u ∈ W 1,∞ (0, T , V ) et
σ ∈ W 1,∞ (0, T , H1 ).

Proof. La démonstration s’effectue en plusieurs étapes, Il est clair que P1 , n’est pas un problème
du type usuel, donc on ne peut appliquer directement les théorèmes d’existence et d’unicité des
inéquations variationnelles elliptiques, car, la fonctionnelle j dépend du terme p(|Rσν , |). Pour
cette raison, on suppose que la régularité de la contrainte sur la frontière Γ3 est donnée, on la
note g. Aussi pour tout η ∈ L∞ (0, T , H) on suppose que la partie non élastique de la contrainte est
donnée et on la note zη , tel que:
Z t
zη (t) = η(s)ds + z0 ∀t ∈ [0, T ]
0
avec
z0 = σ0 − ξε(u(0))
gη
Ainsi on obtient un problème intermédiaire Pv . Une fois qu’on démontre l’existence et l’unicité
gη
de ( de Pv on utilise la méthode du point fixe pour l’application définie de H1 dans H1 par
g → σgη , ensuite pour l’application défini de: L∞ (0, T , H) → L∞ (0, T , H) par:

η(t) → G(ε(u(t)).σ (t)).

57
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] S. Drabla , Analyse variationnelle de quelques problèmes aux limites en elasticité et viscoplas-
ticité, thèse de doctorat, Université Sétif, 1999.

[2] [Link], Etude théorique et numerique de quelques problèmes de contact, Mémoire de Mgis-
ter. Université de Sétif. 2001.

[3] [Link], M. Sofonea, and M. Shilor, A Quasistatic Viscoplastic Contact Problem with
Friction, Journal of Elasticity . (51). p. 105-126, 1998.

58
CNMA’2022
Second National Conference on
Mathematics and its Applications

An application of Perov fixed point

theorem on a coupled system in fractional
γ ,1
Sobolev spaces W0+1 (0, 1)

Noura Laksaci 1 [Link]@[Link]

Ahmed Boudaoui1 ahmedboudaoui@[Link]
1 Laboratory of Mathematics Modeling and Applications. University of Adrar, National Road No. 06, 01000,
Adrar, Algeria.

Abstract: The purpose of this paper is to investigate the existence and uniqueness of a
system of coupled fractional differential equations with a Riemann-Liouville derivative in
γ ,1 γ ,1
the Cartesian product of fractional Sobolev spaces E = W0+1 (0, 1) × W0+2 (0, 1). Our plan is
to furnish the space E with a vector-valued norm, and next we apply the Perov fixed point
theorem.

Keywords: coupled system; fractional differential equations; Riemann–Liouville deriva-

tive; generalized Banach space; fixed point theorems; convergent to zero matrix.
AMS (MOS) Subject Classifications: 34A08, 26A33, 47H10.

Introduction
The beginnings of fractional calculus can be traced back to the end of the 17th century, to a
1
d2f
debate between Leibniz and de l’Hopital about the meaning of 1 . [1]. Furthermore, various
dt 2
researches have been introduced to explore and study this significant mathematical topic, includ-
ing Liouville, Riemann, Abel, Riesz, Weyl, Hadamard, and Caputo.
Researchers use relevant fixed point theorems to secure the solution of some nonlinear problems.
The Banach contraction principle is one of these theorems. Perov extended the Banach contraction
principle to vector-valued metric spaces in 1965 by replacing the contraction factor with a conver-
gent to zero matrix. [2]. One of the key techniques for demonstrating the existence of a solution
to systems of differential equations, fractional differential equations, and integral equations in N
variables is Perov’s fixed point theorem see [4, 3], and the references cited therein.
Recently, a number of interesting papers on the solvability of mathematical problems in Sobolev
spaces W n,p (R+ ) with the help of fixed point theory have been presented. In [5], the authors uti-
γ,p
lized the Riemann–Liouville derivative to introduce the left fractional Sobolev spaces Wa+ (a, b),
where −∞ < a < b < +∞, 1 ≤ p < +∞, and n − 1 < γ ≤ n, n ∈ N.

Main results
Our work is devoted to studying the existence and the uniqueness of a coupled system of frac-

59
CNMA’2022
Second National Conference on
Mathematics and its Applications

tional differential equations of the form:

 RL α1 α −1


 D0+ %1 (t) = f1 (t, %1 (t), %2 (t),RL D0+1 %1 (t)),

RL α2 α −1
D0+ %2 (t) = f2 (t, %1 (t), %2 (t),RL D0+2 %2 (t)),





RL α1 −1 2−α
D0+ %1 (0) = %10 , RL I0+ 1 %1 (0) = %11






RL Dα2 −1 % (0) = %0 , RL I 2−α2 % (0) = %1 ,


0+ 2 2 0+ 2 2

α −1,1 α −1,1
in the generalized Banach space W0+1 (0, b) × W0+2 (0, b) where %0 , %1 ∈ R2 , 1 < α1 , α2 ≤ 2,
α
and for each i = 1, 2, RL D0+ is the Riemann–Liouville fractional derivative of order αi and fi :
i

(0, b) × R3 −→ R.

References
[1] Leibniz, G.W.; Gerhardt, C.I. Mathematische Schriften; Olms: Zurich, Switzerland, 1962.

[2] Perov, A. On the Cauchy problem for a system of ordinary differential equations, Priblijen.
Metod Res. Dif. Urav., Kiev 1964, 2, 115–134.

[3] Aghajani, A.; Pourhadi, E.; Rivero, M.; Trujillo, J. Application of Perov’s fixed point theo-
rem to Fredholm type integro-differential equations in two variables. Math. Slovaca 2016,
66, 1207–1216 .

[4] Graef, J.R.; Kadari, H.; Ouahab, A.; Oumansour, A. Existence results for systems of
second-order impulsive differential equations.

[5] Idczak, D.; Walczak, S. Fractional Sobolev spaces via Riemann-Liouville derivatives. J.
Funct. Spaces Appl. 2013, 2013, 128043.

60
CNMA’2022
Second National Conference on
Mathematics and its Applications

GEOMETRIC ASPECTS FOR THE FULL

VISCOUS NONLINEAR BOUSSINESQ SYSTEM

Youssouf Maafa 1 [Link]@[Link]

1 Department of mathematics , University of Batna -2- ,LEDPA

Abstract: This work deals with the global well-posedness in time for the full viscous nonlinear
Boussinesq equations in two dimensions in the framework of a smooth vortex patch. Further-
more, we provide the inviscid limit for the velocity and the density.

Keywords: nonlinear Boussinseq system , inviscid limit, smooth vortex patch ,Rate of
convergence ..
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
The full viscous nonlinear Boussinesq system is given by the coupled equations,

∂t vµ + vµ · ∇vµ − µ∆vµ + ∇pµ = G(θµ ) if (t, x) ∈ R+ × R2 ,






 ∂t θµ + vµ · ∇θµ − κ∆θµ = 0 if (t, x) ∈ R+ × R2 ,



 (NBµ,κ )
divvµ = 0,







 (v , θ ) 0 0
 µ µ |t=0 = (vµ , θµ ).

Where v = (v 1 , v 2 ) refers to the velocity vector filed in position R2 at a time t , the condition
divvµ = 0 means the fluid is incompressible , the scalar function θ is a positive function denotes
the temperature or the density , p represents the pressure , G is a vectoriel function. The positive
parameters µ and κ represent the viscosity and resistivity of the fluid, respectively.
In what follows to simplify our presentation we take κ = 1, therefore our system denoted (NBµ,κ )
becomes, 
∂t vµ + vµ · ∇vµ − µ∆vµ + ∇pµ = G(θµ ) if (t, x) ∈ R+ × R2 ,






 ∂t θµ + vµ · ∇θµ − κ∆θµ = 0 if (t, x) ∈ R+ × R2 ,



 (NBµ )
divv = 0,




 µ

 (vµ , θµ )|t=0 = (vµ0 , θµ 0 ).



By applying the curl operator to first equation in (NBµ ), we get

  


 ∂t ωµ + vµ · ∇ωµ − µ∆ωµ = ∇ × G(θµ ) ,


∂t θµ + vµ · ∇θµ − ∆θµ = 0, (VDµ )





 (θ , ω ) 0 0
µ |t=0 = (θµ , ωµ ).

µ

61
CNMA’2022
Second National Conference on
Mathematics and its Applications

Where ωµ0 = curl vµ0 is the initial vorticity ,we assume that ωµ0 = 1Ω0 with Ω0 is a smooth bounded
domain with C ε+1 (R2 ) regularity such that 0 < ε < 1,so that it is already verified ωµ (t) = 1Ωt ,
where Ωt , Ψµ (t, Ω0 ) and the Ψµ is the flow generated by the velocity vµ by the following equation
∂t Ψ (t, x) = v(t, Ψ (t, x)) with Ψ (0, x) = x. Then we prove the system (NBµ ) has a unique global
solution and the velocity vµ is Lipschitz informally with respect to viscosity µ , moreover the
image of the domain initial Ω0 by the flow viscous Ψµ (t, Ω0 ) has the same regularity C ε+1 (R2 ) .
Our second task is to study the inviscid limit of the system (NB) , when the viscosity µ = 0 then
we obtain the classical nonlinear stratified Euler Boussinseq system

2
 ∂t v + v · ∇v + ∇p = G(θ) if (t, x) ∈ R+ × R ,





 ∂t θ + v · ∇θ − ∆θ = 0 if (t, x) ∈ R+ × R2 ,



 (NB)
divv = 0,






 (v, θ)|t=0 = (v 0 , θ 0 ).



We study the convergence kvµ (t) − v(t)kLp (R2 ) + kθµ (t) − θ(t)kLp (R2 ) with p ∈ [2, ∞[ when µ goes to 0.

Main results
The first main result treats essentially the global and uniqueness topic for the system (NBµ ) More
precisely, we have the following theorem.
Theorem 25 (Uniform boundedness of the velocity and density). Let G ∈ C 2 (R, R2 ) such that G(0) =
4
0. Let Ω be a C 1+ε -bounded domain with 0 < ε < 1, ωµ0 = 1Ω0 and θµ0 ∈ L1 ∩ L∞ ∩ Ḃ0m,1 for 1−ε ≤m<
∞
∞, then the system (NBµ ) admits a unique global solution (vµ , θµ ) such that (vµ , θµ ) ∈ Lloc (R+ ; Lip) ×
0
L∞ 1 ∞
loc (R+ ; L ∩ L ∩ Ḃm,1 ). More precisely:
2
k∇vµ (t)kL∞ ≤ C0 eexp{C0 t log (2+t)}
.
If Ψ denotes the flow associated to velocity v. Then the domain Ω(t) , Ψ (t, Ω0 ) is C 1+ for every t ≥ 0.

Our second main result motivates by establishing the inviscid limit of the system (NBµ ) towards
(NB) when the viscosity goes to zero. In particular, we quantify the convergence rate between
velocities and densities. More precisely, we have the following theorem.
Theorem 26 (Rate of convergence). Let (vµ , θµ ), (v, θ) be the solutions of the (NBµ ), (NB), respectively
with the same initial data satisfies the conditions of Theorem [Link] ωµ0 = ω0 = 1Ω0 with Ω0 is simply
connected bounded domain. Then for all t ≥ 0, µ ∈]0, 1[, we have.
2 1
kvµ (t) − v(t)kLp + kθµ (t) − θ(t)kLp ≤ C0 eC0 t log (2+t)
(µt) p ∀p ∈ [2, ∞[.

References
[1] T. Hmidi and S. Keraani, Inviscid limit for the two-dimensional Navier-Stokes equation in a
critical Besov space, Asymp. Anal., 53(3), 125–138, 2007.

[2] [Link] and M. Zerguine: Inviscid limit for the viscous 2d Boussinesq system with
temperature-dependent [Link].07014[[Link]]

[3] O. Melkemi and M. Zerguine :Local persistence of geometric structures of the inviscid nonlin-
ear Boussinesq system. 2020. arXiv: 2005.11605 [[Link]].

[4] S. Sulaiman :Global existence and uniquness for a non linear Boussinesq system. Journal of
Mathematical Physics 51, 093103 (2010).

62
CNMA’2022
Second National Conference on
Mathematics and its Applications

Existence and uniqueness global of

bounded nonnegative weak solutions for
nonlinear parabolic evolution problem

Nabila Barrouk 1 [Link]@[Link]

Mounir Redjouh 2 redjouhmounir@[Link]
1 Faculty of Science and Technology, Department of Mathematics and Informatics, Mohamed Cherif Messaadia
University, B.P. 1553 Souk Ahras 41000, Algeria
2 Department of Mathematics and Computer Science, Barika University Center, Algeria

Abstract: Over the years, parabolic evolution problem have attracted the attention of a large
number of researchers due to their use in many physical, biological and other phenomena.
This work we study a class of nonlinear parabolic evolution problems with Neumann or
Dirichlet boundary conditions. We prove the existence and uniqueness of bounded weak
solutions to such a problem. Our investigation applied for a wide class of the terms of reaction.

Keywords: Global solution, compact semigroups, local solution, nonlinear parabolic

equation..
2010 Mathematics Subject Classification: Primary 35K57, 35K40.

Introduction
Systems of parabolic evolution equations, have long been a subject of active research. These sys-
tems have numerous applications in physics, chemistry, ecology, biology, and other disciplines.
Examples include problems in combustion, phase transitions, the formation of patterns in chem-
ical reactions and living systems, propagation of electrical signals in nerve axons and cardiac
tissue, population dynamics, cellular differentiation, and morphogenesis, to name a few. It is be-
yond the scope of this work to give a review of the theory of parabolic systems. For systematic
expositions of some aspects of the theory, numerous applications, and a comprehensive list of
literature on this subject we refer to [3, 5].
In the present work, we study the folowing mathematical model of parabolic evolution problem

∂u



 ∂t
− a∆u = f (u, v, ∇u, ∇v) , in QT ,

∂v

 ∂t − c∆u − d∆v = g (u, v, ∇u, ∇v) , in QT ,



(67)
∂u ∂v

= = 0 or u = v = 0, in Σ ,

T

∂η ∂η





 u (0, x) = u0 (x) , v (0, x) = v0 (x) , in Ω,


where Ω is an open bounded domain of class C 1 in RN , with smooth boundary ∂Ω, QT = ]0, T [ ×
Ω, ΣT = ]0, T [ × ∂Ω, T > 0, and ∆ denotes the Laplacian operator on L1 (Ω) with Dirichlet or
Neumann boundary conditions, the constants a, c, and d are supposed to be positive, satisfying

63
CNMA’2022
Second National Conference on
Mathematics and its Applications

the conditions a > d, and c2 < 4ad, which reflects the parabolicity of the system and implies at the
same time that the matrix of diffusion is positive definite.
The system (67) for special cases has been studied extensively in recent years.
For c = 0 and nonnegative initial data arise, for example, as models for the diffusion of substances
which at the same time react with each other chemically (cf. [3]). Also (67) is related to the
Rosenzweig-Mac Arthur equation in ecology (cf. [1]).
In the case where f (u, v) = −g (u, v) = −uv σ , Alikakos [1] obtained L∞ -bounds of solutions global
n+2
existence when 1 < σ < , and Masuda [8] who showed that solutions exist globally for every
n
σ ≥ 1 and, in addition, showed that the solutions converge as t goes to +∞.
Recently Alaa and Mounir [1] treat the case where c = 0. In that article the calculations were
relatively simple since the system can be regarded as a perturbation of the simple and trivial case;
for which nonnegative solutions exist globally in time.
In [9, 10], Moumeni and Barrouk obtained a global existence result of solutions for reaction-
diffusion systems with a diagonal and triangular matrix of diffusion coefficents. By combining
the compact semigroup methods and some L1 estimates, we show that global solutions exist for a
large class of the function f .
In the case where c ≥ 0, systems of the type (67) occur in many applications (cf. [4]).
For a > d > 0, c ≥ 0, c2 < 4ad, f (u, v) = −g (u, v) = −uΨ (v) , Ψ is continuously differentiable,
c
nonnegative, global bounds were proved by Kirane in [4] if v0 (x) ≥ a−d u0 (x) ≥ 0, x ∈ Ω. The
author proved also that the solution (u, v) converges to a constant vector k = (k1 , k2 ) as t → ∞,
uniformly in Ω̄. Furthermore, k1 ≥ 0, k2 ≥ 0 and k1 Ψ (k2 ) = 0.
In this present work reviews the necessary results we need to nonlinear evolution equations and
we consider the problem (67) by using a homogeneous Dirichlet or Neumann boundary condi-
tions we state and prove the main results concerning the establish of a global existence result of
the weak solution.

Main results
Let A m-dissipative operator of the dense domain in the Banach space X and S (t) a semigroup
engendered by A, F a function locally Lipchitz, so ∀u0 ∈ X it exists T (u0 ) = Tmax such that the
problem


1 (Ω) ∩ L1 0, T ; W 1,1 (Ω) ,
   



 u ∈ C [0, T ] ; L 0
 du


 − Au = F (u, ∇u) , (68)



 dt
 u (0) = u0 .


admits a unique solution u verifying

Zt
u (t) = S (t) u0 + S (t − s) F (u (s) , ∇u (s)) ds, ∀t ∈ [0, Tmax [ .
0

Compactness of the solution

In this section we will give a compactness result of operator L defining the solution of the problem

64
CNMA’2022
Second National Conference on
Mathematics and its Applications

(68) in the case where the initial value equals zero [u (0) = 0] i.e.
Zt
L (F) (t) = u (t) = S (t − s) F (u (s) , ∇u (s)) ds, ∀t ∈ [0, T ] . (69)
0

Theorem 27. If for all t > 0, the operators S (t) are compact, then L are compact of L1 ([0, T ] , X) in
L1 ([0, T ] , X) .

Remark 3. The semigroup S (t) generated by the operator ∆ is compact in L1 (Ω).

Existence
Existence of a local solution and its positivity

We convert the system (67) to an abstract first order system in the Banach space X = L1 (Ω)×L1 (Ω)
of the form 
 w0 (t) = Aw (t) + F (w, ∇w) , t > 0,


(70)

 w (0) = w = (u , v )t ∈ X.


 0 0 0

In (70), w : R+ → X
A : D∞ (A) × D∞ (A) → X,
with
Aw (t) = (d1 ∆u (t) , d2 ∆u (t) + d3 ∆v (t))t ,
   
2,1 ∂u
and D∞ (A) = u ∈ W (Ω) for all p > n, ∆u ∈ C Ω̄ , ∂η = 0 or u = 0 ,

F (w (t)) = (f (u (t) , v (t) , ∇u (t) , ∇v (t)) , g (u (t) , v (t) , ∇u (t) , ∇v (t)))t ,

Definition 5. For each t > 0 define a linear operator S (t) from X to X

! !
d2 d2
S (t) (u0 , v0 ) = Sd1 (t) u0 , Sd3 (t) v0 − u + S (t) u0 .
d1 − d3 0 d1 − d3 d 1

Proposition 5. S (t), t > 0 is an analytic semigroup of the bounded linear operator in X with infinitesi-
mal generator A

Theorem 28. For every initial data w0 ∈ X, the problem (70) admits a unique strong local solution.

Positivity of the solution of the problem (67)

Lemma 5. Let (u, v) be a classical solution of (67) such that u0 (x) ≥ 0, v0 (x) ≥ 0, x ∈ Ω, then u (t, x) ≥ 0
and v (t, x) ≥ 0.

Global existence
2
Theorem 29. Assume that  u0 , v0 ∈ L (Ω), then there exists a positive global solution (u, v) of system
2 1
(67). Moreover, u, v ∈ L 0, T ; H0 (Ω) .

Conclusion 1. The subject proposed for this work are concerned with the existence of weak solutions of
the reaction-diffusion systems with a triangular diffusion matrix.
The study concerns the generation of semigroup in L1 spaces, the existence of local solutions and positions
and its positivity, the global solutions.

65
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] N. Alaa and I. Mounir, Global existence for reaction-diffusion systems with mass control
and critical growth with respect to the gradient, J. Math. Anal. Appl. 253 (2001), 532-557.

[2] N.D. Alikakos, Lp -bounds of solutions of reaction-diffusion equations, Comm. Partial Dif-
ferential Equations 4 (1979), 827-868.

[3] N.F. Britton, Reaction-diffusion equations and their applications to Biology, Academic
Press, London, (1986).

[4] P. V. Danckwerts, Gas-Liquid Reactions, McGraw-Hill, New York, (1970).

[5] P.C. Fife, Mathematical aspects of reacting and diffusing systems, Lecture Notes in
Biomath.28, Springer-Verlag, Berlin, New York, (1979).

[6] S. L. Hollis, R. H. Martin and M. Pierre, Global existence and boundedness in reaction
diffusion systems, SIAM J. Math anal, 18, (1987), pp. 744-761.

[7] M. Kirane, Global bounds and asysmptotics for a system of reaction-diffusion equations.
Journal of Mathematical Analysis and Applications (1989), 138, 328-342.

[8] K. Masuda, On the global existence and asymptotic behaviour of solution of reaction-
diffusion equations. Hokkaido Math, J. 12: (1983), 360-370.

[9] A. Moumeni and N. Barrouk. Existence of global solutions for systems of reaction-
diffusion with compact result, IJPAM. 102(2) (2015), 169-186.

[10] A. Moumeni and N. Barrouk, Triangular reaction-diffusion systems with compact result,
GJPAM. 11(6) (2015), 4729-4747.

66
CNMA’2022
Second National Conference on
Mathematics and its Applications

Contact problem for thermoviscoelasstic

material with friction

Nadjet Bensebaa 1 bensebaa_na@[Link]

1 Departement of Mathematics, University of Setif, Setif 1900, Algeria

Abstract: We consider a quasistatic contact problem for viscoelastic material with thermal
effects. The contact is modelled with the normal compliance condition, associated to the
Coulomb’s law of dry friction. A variational formulation of the model is drived, and the
existence of a unique weak solution is proved. The proofs are based on the classical result
of nonlinear first order evolution inequalities, arguments of history-dependent variational
inequalities and the fixed point arguments.

Keywords: Thermo-viscoelastic material; quasistatic process; compliance normal; Coulomb’s

friction; evolution in equality; weak solution; fixed point.
2010 Mathematics Subject Classification: 74F05. 74H25, 74M10, 74M15.

Introduction
We study a quasistatic contact problem compiling a viscoelastic material with thermal effects and
friction. The contact is modeled with normal compliance and the associated version of Coulomb’s
law of dry friction. The frictional contact between deformable bodies can be frequently found
in industry and every life. Because of the importance in metal forming and automotive indus-
try, a considerable effort has been made with the modeling and numerical simulations of contacts
problems and the engineering literature concerning this topic are rather extensive. An excellent
reference on the field of contact problems with or without friction is [3]. Elastic or viscoelastic
friction contact problems, with thermal considerations, can be found in [1] and the references
therein. We derive a variational formulation of the problem and the existence of a unique weak
solution.

Main result

Our main existence and uniquenss result, of the suggested model, is the mechanical problem has
{u, σ , θ} a unique weak solution satisfying

u ∈ C 1 (0, T ; V ) .
σ ∈ C (0, T ; H1 ) .
1,2
θ∈W (0, T ; E 0 ) ∩ L2 (0, T ; E) ∩ C(0, T ; L2 (Ω)).

67
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] Amassad, A. kuttler, K. L., Rochdi, M., and Shillor, M. Quasistatic thermoviscoelastic
contact problem with slip dependent friction coefficient. Math. Comput. Model., 36(7-8),
839-854 (2002).

[2] W. Han, M. Sofonea, Analysis and numerical approximation of an elastic frictional contact
problem with normal compliance, Applicationes Mathematicae, 26,4 (1999), pp. 415-435.

[3] N. Kikuchi and J. T. O. den, Contact Problems in Elasticity: A Study of Variational In-
equalities and Finite Element Methods, SIAM, Philadelphia, 1988.

[4] M. Selmani, L. Selmani, Frictional contact problem for elestic-viscoplastic materials with
termal effect∗ . Appl. Math. Mech-Engl. Ed., 34(6), 761-776 (2013), DOI 10.1007/s10483-
013-1705-7, Shanghai University and Springer-Verlag, Berlin Heidelberg 2013.

68
CNMA’2022
Second National Conference on
Mathematics and its Applications

Existence of solutions for a class of PNP

s(·)
model in F B p(·),q

Oussama Melkemi 1 [Link]@[Link]

1 Department of mathematics, University Batna 2

Abstract: This work is devoted to study a class for the PNP model, where we prove the global
well-posedness of this system in variable Fourier-Besov spaces, furthermore, we show the
Gevery regularity of these solutions.

Keywords: PNP model, Gevery Regularity, Fractional dissipation.

2010 Mathematics Subject Classification: 35Q35 35B65 76D03.

Introduction
The generalized PNP system is described by the following initial value problem,

α 3
 ∂t u + (−∆) u = −div(u∇φ), (t, x) ∈ R+ × R





β
 ∂t v + (−∆) v = div(v∇φ), (t, x) ∈ R+ × R3



 (PNPγ )
(−∆)θ φ − u + v = 0, (t, x) ∈ R+ × R3







 (u, v)|t=0 = (u0 , v0 ),


where u stands for the density of the electron, v represents the density of
√ the hole in an electrolytes
and φ stands for the electric potential. The non-local operator Λ := −∆ denotes the Zygmund
operator which is defined through the Fourier transform, for γ ∈ (0, 2],

F ((−∆)γ g)(ξ) := |ξ|2γ b

g (ξ).

It is well-known that the system (PNPγ ) is a natural generalization for the classical PNP system
(called also drift-diffusion equations), which reads as follows


∂t u − ∆u = −div(u∇φ), (t, x) ∈ R+ × R3






∂t v − ∆v = div(v∇φ), (t, x) ∈ R+ × R3



(PNP)


∆φ = u − v, (t, x) ∈ R+ × R3








 (u, v)|t=0 = (u0 , v0 ).

69
CNMA’2022
Second National Conference on
Mathematics and its Applications

The mathematical study of the classical PNP system (PNP) was initially started by Mock in [2],
who proved the initial value problem and the Cauchy problem of the System (PNP). Meanwhile,
H. Gajewski[3] investigated the existence, uniqueness and asymptotic behavior of the classical
system (PNP). Karch in [4] studied the problem (PNP) in the homogeneous Besov spaces Ḃsp,∞ (Rn )
n
with s ∈ (−1, 0) and p = s+2 . In another context, the theory of function spaces with variable ex-
ponents has evolved into an interesting field of research because of its applications in harmonic
analysis and partial differential equations. In 1931’s Orlicz in his paper [1] introduced a new gen-
eralization for the classical Lebesgue spaces Lp (Rn ), the main idea of Orlicz is replacing the real
variable exponent p by a measurable function p(·). The special structure of this kind of spaces and
the absence of some properties such as Young’s convolution inequality seems to make the study
of a class of PDEs more complicated compared to the real variable function spaces.

Main results
Our aim here is to consider and investigate a class of the PNP system in the Variable Fourier-
s(·)
Besov spaces with variable exponent F B p(·),q . More precisely, we prove the global existence of
solutions for this system by employing the famous Duhamel’s formula, the mixed time-spaces
s(·)
 
with variable exponent Lr [0, T ], F B p(·),q . Then, we prove the analyticity of this solution by using
the so called Gevrey regularity.

References
[1] W. Orlicz, Uber konjugierte Expoentenfolgen, Studia Math. 3 (1931), 200–211.

[2] Mock, M.S.: An initial value problem from semiconductor device theory. SIAM J. Math. Anal.
5, 597–612 (1974).

[3] Gajewski, H.: On existence, uniqueness and asymptotic behavior of solutions of the basic equa-
tions for carrier transport in semiconductors. Z. Angew. Math. Mech. 65, 101–108 (1985).

[4] Karch, G.: Scaling in nonlinear parabolic equations. J. Math. Anal. Appl. 234, 534–558
(1999).

70
CNMA’2022
Second National Conference on
Mathematics and its Applications

Radially symmetric solutions for a

multidimensional nonlinear fractional
partial differential equations

Rabah Djemiat 1 rabahdjemiat19@[Link]

1 Laboratory of Pure and Applied Mathematics, Mohamed Boudiaf, University of M’sila, Algeria.

Abstract: This paper discusses the existence of radially symmetric solutions for a multidi-
mensional nonlinear fractional partial differential equations that enables treating vibration
and control, image processing, among others other physical phenomena. Additionally,
the application of Schauder’s and Banach’s fixed point theorems facilitates identifying the
existence and uniqueness of solutions for the selected equation.

Keywords: Multidimensional nonlinear equation; reaction-diffusion/wave; time and

space-fractional order; radially symmetric solutions; existence and uniqueness.
2010 Mathematics Subject Classification: 35R11; 35A01; 34A08; 35C06; 34K37.

Introduction
In this work, we shall give an example of a class of fractional-order’s PDEs, which helps to de-
scribe various complex phenomena; it is a multidimensional nonlinear time and space-fractional
reaction-diffusion/wave equation and is written as follows:
 β 
∂αt u − κ2 ∆u = F t, x, u, ∂t u, (−∆)s u , for 0 < s ≤ 1 < β ≤ α ≤ 2, (71)

where u = u (t, x) is a scalar function of the time t ≥ 0 and space variables x ∈ Rm , with m ∈ N∗ .
Also F : [0, ∞) × Rm × C × C × C → C is a nonlinear function, κ ∈ R∗ is a real constant and

∂n u
 ∂t n , α = n ∈ N∗ ,


α

∂t u (t, x) =  R t (t−τ)n−α−1 ∂n
 I0n−α n
+ ∂t u = u (τ, x) dτ, n − 1 < α < n.


0 Γ (n−α) ∂τ n

Equation (71) is a representation of a large class of linear and nonlinear equations. Note that, for
F ≡ 0 and α = 1 (resp. α = 2), the PDE (71) represents the standard heat equation (resp. the wave
equation). In addition to that, it becomes the Klein-Gordon equation when we choose F = κu,
|κ| = 1 and α = 2. All these equations fall under the name of the fractional reaction-diffusion/wave
equation.

Main results
2 h √ m
Let 0 < s ≤ 1, 1 < β ≤ α ≤ 2, ε, ` > 0, and Tε = `ε α be such that Ω = [0, Tε ] × ε/ m, +∞ . We

71
CNMA’2022
Second National Conference on
Mathematics and its Applications

consider: 
2 ∆u = F t, x, u, ∂β u, (−∆)s u , (t, x) ∈ Ω, κ ∈ R∗ ,
 
α
∂ u − κ


 t
 t
(72)
δ ∂u

 u (0, x) = |x| v, ∂t (0, x) = 0, δ, v ∈ C,



where F : Ω × C × C × C → C is a nonlinear function.

This paper’s contribution regards determining the existence, uniqueness, and main properties of
the general solution of stability problems obtained through replacing classical rules with frac-
tional quadrature rules of the radially symmetric solution (see [1]),
q
 2 
u (t, x) = |x|δ f |x|− α t , for |x| = x12 + · · · + xm
2
, and δ ∈ C, (73)

the basic profile f is not known in advance and is to be identified.

Taking into consideration the regularization processes, our major aim is employing of the so-
lutions’ intermediate properties for the fractional order’s PDE’s problem (72). We consider the
intermediacy of the multidimensional nonlinear reaction-diffusion equation and the wave equa-
tion.
We illustrate that using analytical techniques to obtain the existence and uniqueness of weak
solutions via the use of form (73) is promising and can also bring new results for other applications
in fractional-order’s PDEs. It permits us to reduce the fractional-order’s PDE (71) to a fractional
differential equation; the idea is well illustrated in this paper through selected examples and
explicit solutions.
For the forthcoming analysis, we impose the following hypotheses:
(hyp.1) F : Ω × C × C × C → C is a continuous function that is invariant by the change of scale (73).
It gives us:
 4κ2
!
 β s

δ−2

0 C β 2 00
F t, x, u, ∂t u, (−∆) u = |x| ϕ η, f (η) , f (η) , D0+ f (η) − 2 η f (η) , (74)
α
2
where η = |x|− α t and ϕ : [0, `] × C × C × C → C is a continuous function.
(hyp.2) There exist three positive constants ω1 , ω2 , ω3 > 0 so that the continuous function ϕ given
by (74) satisfies:
      
ϕ (η, f , g, h) − ϕ η, f˜, g̃, h̃  ≤ ω1 f − f˜ + ω2 |g − g̃| + ω3 h − h̃ ,

for any f , g, h, f˜, g̃, h̃ ∈ C.

for any f , g, h ∈ C and η ∈ [0, `] .
λ denotes the positive constant defined by

α` β−1 |`q + ω2 | + ω3
λ= ,
` β−α Γ (α − β + 1)
2
 
where q = − 4κ
α 2 αδ + α + 1 + mα
2 and

a∗ = sup a (η) , b∗ = sup b (η) , c∗ = sup c (η) and d ∗ = sup d (η) .

η∈[0,`] η∈[0,`] η∈[0,`] η∈[0,`]

Now, we give the main theorems of this work.

72
CNMA’2022
Second National Conference on
Mathematics and its Applications

Theorem 30. Assume the hypotheses (hyp.1) , (hyp.2) hold. We give

 1
2
 Γ (α + 1) (1 − λ)  α
δκ (δ + m − 2) + ω1 , 0, λ ∈ (0, 1) and K =     .
δκ2 (δ + m − 2) + ω1 

If we put
2
Tε < ε α K, (75)
then the problem (72) admits a unique solution in the radially symmetric form (73) on Ω.

Lemma 6 ([3]). Assume that C D0α+ f ∈ C ([0, `] , C) , for all α > 0, then:

n−1 (k)
X f (0)
I0α+ C D0α+ f (η) = f (η) − η k , n − 1 < α ≤ n ∈ N∗ .
k!
k=0

Our initial aim is to infer that the function f in (73) satisfies an equation that is employed in
the definition of radially symmetric solutions.

Theorem 31. Let δ, v ∈ C, α, β ∈ R, be such that 1 < β ≤ α ≤ 2 and p = δκ2 (δ + m − 2) with κ ∈ R∗ .

If the hypothesis (hyp.1) holds, the problem of time and space-fractional order (72) is reduced by the
transformation (73) to the fractional order’s ordinary differential equation of the form

C
 β 
D0α+ f (η) = ψ η, f (η) , f 0 (η) , C D0+ f (η) , η ∈ [0, `] , (76)

where  β   β 
ψ η, f (η) , f 0 (η) , C D0+ f (η) = pf (η) + qηf 0 (η) + ϕ η, f (η) , f 0 (η) , C D0+ f (η) ,
with the conditions
f (0) = v and f 0 (0) = 0. (77)

Proof Substituting expression (73) in the original PDE of fractional order (71) results in a frac-
tional equation that needs to be narrowed down to the standard bilinear functional equation
2
(check [2]). First, for η = |x|− α t, we get η ∈ [0, `] and

4 mα 4
   
∆u (t, x) = |x|δ−2 δ (δ + m − 2) f (η) − 2 αδ + α + 1 + ηf 0 (η) + 2 η 2 f 00 (η) . (78)
α 2 α
2
On the other hand, for ξ = |x|− α τ, we get:

∂α u
= |x|δ−2 C D0α+ f (η) . (79)
∂t α
If we replace (74), (78) and (79) in (72), we obtain the problem (76)–(77). 
β
Lemma 7. Let f , f 0 , C D0α+ f , C D0+ f ∈ C ([0, `] , C) , then the problem (76)–(77) is equivalent to the
integral equation:
Zη
1  β 
f (η) = v + (η − ξ)α−1 ψ ξ, f (ξ) , f 0 (ξ) , C D0+ f (ξ) dξ, ∀η ∈ [0, `] .
Γ (α) 0

Proof By applying I0α+ to equation (76) and using Lemma 6, we obtain the requisite result. 

73
CNMA’2022
Second National Conference on
Mathematics and its Applications

β
Lemma 8. Let f , C D0α+ f , C D0+ f ∈ C ([0, `] , C) , be such that f 0 (0) = 0, then ∀η ∈ [0, `] ,

  α` α−1 
β
 ` α−β
f 0 (η) ≤ C D0α+ f and C D0+ f (η) ≤ C D0α+ f . (80)
Γ (α − β + 1) ∞ Γ (α − β + 1) ∞

Proof By using Lemma 6, for all C D0α+ f ∈ C ([0, `] , C) , we get:

d α C α
I0α−1
+
C α
D0+ f (η) = I + D0+ f (η)
dη 0
= f 0 (η) − f 0 (0) ,

Moreover; if f 0 (0) = 0, then I0α−1

+
C D α f (η) = f 0 (η) . Also, for any η ∈ [0, `] , we get
0+

    ` α−1
f 0 (η) ≤ I0α−1
+
C α
D0+ f (η)  ≤ C D0α+ f .
Γ (α) ∞

We have Γ (α + 1) > Γ (α − β + 1) , for any 1 < β ≤ α ≤ 2, then

  α` α−1
f 0 (η) ≤ C D0α+ f .
Γ (α − β + 1) ∞

In another way, we have

α−β C α−β d2 2−β d

2
β
I0+ D0α+ f (η) = I0+ I02−α
+
2
f (η) = I 0 +
2
f (η) = C D0+ f (η) ,
dη dη

then  
C Dβ+ f (η) ≤ ` α−β
0 C D0α+ f .
Γ (α − β + 1) ∞

The proof is complete. 

Theorem 32. Assume the hypothesis (hyp.2) holds. We give p + ω1 , 0 and λ ∈ (0, 1) . If we put
! α1
Γ (α + 1) (1 − λ)
`< , (81)
|p + ω1 |

then the problem (76)–(77) admits a unique solution on [0, `] .

Proof Let f , g ∈ C ([0, `] , C) be two functions that satisfy (76)–(77), then we get:
Zη
1 h  β 
Af (η) − Ag (η) = (η − ξ)α−1 ψ ξ, f (ξ) , f 0 (ξ) , C D0+ f (ξ)
Γ (α) 0
 β i
− ψ ξ, g (ξ) , g 0 (ξ) , C D0+ g (ξ) dξ.

Also Z η
  1  
Af (η) − Ag (η) ≤ (η − ξ)α−1 C D0α+ f (ξ) − C D0α+ g (ξ)  dξ. (82)
Γ (α) 0
By applying the hypothesis (hyp.2) we get:
     
C D0α+ f (η) − C D0α+ g (η)  ≤ |p + ω1 | f (η) − g (η) + |`q + ω2 | f 0 (η) − g 0 (η)
 
β β
+ ω3 C D0+ f (η) − C D0+ g (η)  ,

74
CNMA’2022
Second National Conference on
Mathematics and its Applications

for any η ∈ [0, `] . Then (80) from Lemma 8 gives us:

  α` β−1 |`q + ω2 | + ω3 C α
C D0α+ f (η) − C D0α+ g (η)  ≤ |p + ω1 | kf − gk∞ + β−α D0+ f − C D0α+ g
` Γ (α − β + 1) ∞

≤ |p + ω1 | kf − gk∞ + λ C D0α+ f − C D0α+ g ∞ ,

or
|p + ω1 |
C D0α+ f − C D0α+ g ≤ kf − gk∞ .
∞ 1−λ
From (82) we find:
` α |p + ω1 |
kAf − Agk∞ ≤ kf − gk∞ .
Γ (α + 1) (1 − λ)
Thus, according to (81), A is considered a contraction operator.
Banach’s contraction principle (see [6]) helps us infer that A has only one fixed point which is
the unique solution of the problem (76)–(77) on [0, `] . 
Proof of Theorem 30
The existence and uniqueness of a radically symmetric solution to problem (72) is demonstrated
using Theorem 32, provided that the condition (75) holds true. The proof is complete.

References
[1] B. Basti and N. Benhamidouche, Existence results of self-similar solutions to the Caputo-type’s
space-fractional heat equation, Surveys in Math. and its Applications, 15 (2020), 153–168.

[2] B. Basti and N. Benhamidouche, Global existence and blow-up of generalized self-similar
solutions to nonlinear degenerate diffusion equation not in divergence form, Appl. Math. E-
Notes, 20 (2020), 367–387.

[3] A. Granas and J. Dugundji, Fixed Point Theory, Springer-Verlag, New York, 2003.

[4] A. A. Kilbas, H. H. Srivastava and J. J. Trujillo, Theory and Applications of Fractional Differ-
ential Equations, Elsevier Science B.V, Amsterdam, 2006.

75
CNMA’2022
Second National Conference on
Mathematics and its Applications

Inverse medium problem in scattering

L. Chorfi 1 l_chorfi@[Link]
I. Djerrar 1
I. Khélifa1
1 Department of mathematics , Badji Mokhtar University , Annaba

Abstract: We consider an inverse problem in scattering of plane waves by a two dimensional

radially-symmetric potential. Our purpose is to determine the scattering potential from the
far-field pattern. The inverse problem is then reduced to a nonlinear integral equation. For
the resolution we use an iterative method with least square regularization. Some numerical
examples show that the reconstruction is effective.

Keywords: Radial symmetry, Far-field data, Inverse scattering

2010 Mathematics Subject Classification: 35P25, 65R32.

Introduction
The mathematical model of our problem is given as follows, we consider the Helmholtz equa-
tion
4u + k 2 p(x)u = 0, in R2 , (83)
the total field
u = us + ui , (84)
satisfies (83), with the incident field u i (x) = exp(ik dˆ · x) and the scattered field u s satisfies the
radiation condition
√ ∂u s
lim r( − iku s ) = 0. (85)
r→+∞ ∂r
In the following, we use polar coordinates r = |x|, x = rb x = r(cos θ, sin θ).
We define the scattered potential q = 1 − p, and we assume that q is radially symmetric i.e., q(x) =
q(r) and q(r) = 0 for r ≥ a.
The problem (83)−(85) is well-posed and the scattered field u s has the asymptotic behavior ([2, 1]):
iπ
s eikr 1 e4
u (x) = γ √ [u ∞ (θ) + O( )], with γ = √ , (86)
r r 8kπ
and u ∞ (θ), θ ∈ [0, 2π[.
Our aim is to study the following inverse problem: find the potential q(r) from the measured far-
field pattern uδ∞ , assuming that kuδ∞ − u ∞ k ≤ δ. We must cite the paper ([3], and the bibliography
therein) where we studied an inverse potential problem for near field data.

Main results

76
CNMA’2022
Second National Conference on
Mathematics and its Applications

The direct problem consists to compute the far-field pattern u ∞ (θ) from the potential q(r). We
show that Z Z
2
∞
u (θ) = k e−ik x̂·y q(y)u(y)dy, (87)
|y|≤a

where u is the solution of (83) − (85).

We decompose the far field pattern in Fourier series
∞
X
u ∞ (θ) = Sn (q) cos nθ,
n=0

where Sn : q → C is an integral operator with a kernel involving Bessel functions.

The inverse problem is formulated as the nonlinear equation

F (q) = d, (88)

where F : C[0, a] → L2 (0, 2π), q 7→ u ∞ is the forward operator.

In practice we solve the system
Sn (q) = dn , n = 1, ..., N . (89)
with Sn (q) (resp. dn ) the Fourier coefficients of F (q) (resp. d).
For the resolution we use the Gauss-Newton method.
In each iteration j = 0, · · · , J we solve the linearized equation

Sn0 (qj , h) = dn − Sn (qj ), n = 1, ..., N , (90)

where Sn0 (q, h) is the Fréchet derivative of Sn (q). Then qj+1 = qj + h is a new approximation (q0
being the first approximation).

Example. To demonstrate the effectiveness of the numerical method, let us consider the example
of potential function q1 (r) = 0.2 sin(2πr), r ∈ [0, 0.5]. We remark that the reconstruction with
noise is more accurate for large wave number k. It seems that the condition number of the system
(90)) depends of the wave number k. This fact is mentioned in the papers ([3]).

a) b)

Figure 2: Exact and reconstructed potential with level noise δ = 10−4 , for a) k = 1, b) k = 3,

77
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] F. Cakoni, D. Colton, A Qualitative approach to inverse scattering theory, Springer, New
York, 2014.

[2] A. Kirsh, An Introduction to The Mathematical Theory of Inverse Problems, Springer Science
and Business Media, Vol. 120, 2011.

[3] J. Shin, E. Arhin, Determining radially symmetric potential from near-field scatter-
ing data, Journal of Applied Mathematics and Computing, 62, 511–524 (2020).
[Link]

78
CNMA’2022
Second National Conference on
Mathematics and its Applications

Studies on the existence of positive

solutions for a second-order iterative
boundary value problem via
krasnoselskii’s fixed point theorem

Safa Chouaf 1 safachouaf905@[Link]

1 LAMAHIS Lab, Faculty of Sciences, Departement de Mathematics, University of Skikda ,

Abstract: In this research study, we are primarily interested in the boundary value problem
for an iterative second -order with iterative source term and integral boundary [Link]
method used here is based on to convert our problem into an integral equation after that
with the help of krasnoselskii’s and banach contraction principle used in conjunction with
the Green function, we obtain a suitable conditions that ensure the existence, uniqueness and
continuous dependence of positive bounded solution.

Keywords: Green’s functions, Fixed point theorem, Iterative boundary value problem,
uniqueness .
2010 Mathematics Subject Classification: 34B27, 47H10, 34K10, 32H12.

Introduction
The aim of this research is to study an iterative functional differential equation, which is a spe-
cific kind of functional differential equation with delays that are dependent on both time and
the state. are only seldom there are published papers on higher-order equations since studying
them becomes more challenging when iterates are included. Our investigation of the following
second-order iterative differential equation is motivated by this

 
x00 (t) = −f x[0] (t) , x[1] (t) , x[2] (t) , ..., x[n] (t)
d  
+ g x[0] (t) , x[1] (t) , x[2] (t) , ..., x[n] (t) , (91)
dt Z
η
x (0) = 0, α x (t) dt = x (T ) , η ∈ (0, T ) , α , 0, (92)
0

where x[0] (t) = t, x[1] (t) = x(t), ..., x[n] (t) = x[n−1] (x(t)) and f , g : [0, T ]×Rn → [0, +∞) are a continuous
function with respect to its arguments.
We define a subset X of C ([0, T ] , R) as follows:
( Zη )
∗
X = x ∈ C ([0, T ] , R) : x (0) = 0, α x (s) ds = x (T ) , α ∈ R , η ∈ (0, T ) ,
0

79
CNMA’2022
Second National Conference on
Mathematics and its Applications

endowed with the supremum norm, be a Banach space, and For 0 ≤ L ≤ T , M ≥ 0, let
Ω = {x ∈ X , 0 ≤ x (t) ≤ L, |x (t2 ) − x (t1 )| ≤ M |t2 − t1 | , ∀t1 , t2 ∈ [0, T ]} ,
a closed convex and bounded subset of X .
Moreover, we assume that
  n
X
f (t, x1 , ..., xn ) − f (t, y1 , ..., yn ) ≤ ci xi − yi , (93)
i=1
  n
X
g (t, x1 , ..., xn ) − g (t, y1 , ..., yn ) ≤ ki xi − yi (94)
i=1

and we introduce the following constants:

ρ = sup |f (s, 0, 0, ..., 0)| , σ = sup |g (s, 0, 0, ..., 0)| .
s∈[0,T ] s∈[0,T ]

n
X j=i−1
X n
X j=i−1
X
j
ζ = ρ+L ci M , ω = σ +L ki Mj .
i=1 j=0 i=1 j=0

Lemma 9. Let 2T , αη 2 , then for f ∈ C ([0, T ] , [0, +∞)) and g ∈ C 1 ([0, T ] , [0, +∞)) the problem (91)-
(92) has a unique solution given by
Zη
−αt 2

[1] [2] [n]

x(t) = (η − s) f s, x (s) , x (s) , ..., x (s) ds
2T − αη 2 0
Zη
2αt 
[1] [2] [n]

+ (η − s) g s, x (s) , x (s) , ..., x (s) ds
2T − αη 2 0
ZT
2t 
[1] [2] [n]

+ (T − s) f s, x (s) , x (s) , ..., x (s) ds
2T − αη 2 0
ZT 
2t [1] [2] [n]

− g s, x (s) , x (s) , ..., x (s) ds
2T − αη 2 0
Zt  
− (t − s) f s, x[1] (s) , x[2] (s) , ..., x[n] (s) ds
0
Zt  
+ g s, x[1] (s) , x[2] (s) , ..., x[n] (s) ds
0

Main results
Existence

Now, to apply Krasnoselskii fixed point theorem, we need to construct two mappings, one of
which is completely continuous and the other is a contraction, so we define an operator B such
that B = B 1 + B2 : Ω→ X as follows:
Zη
αt 2

[1] [2] [n]

(B1 ϕ) (t) = − (η − s) f s, ϕ (s) , ϕ (s) , ..., ϕ (s) ds
2T − αη 2 0
ZT
2t 
[1] [2] [n]

+ (T − s) f s, ϕ (s) , ϕ (s) , ..., ϕ (s) ds
2T − αη 2 0
Zt  
− (t − s) f s, ϕ [1] (s) , ϕ [2] (s) , ..., ϕ [n] (s) ds. (95)
0

80
CNMA’2022
Second National Conference on
Mathematics and its Applications
Z η
2αt  
(B2 ϕ) (t) = (η − s) g s, ϕ [1] (s) , ϕ [2] (s) , ..., ϕ [n] (s) ds
2T − αη 2 0
T 
Z
2t [1] [2] [n]

− g s, ϕ (s) , ϕ (s) , ..., ϕ (s) ds
2T − αη 2 0
Zt  
+ g s, ϕ [1] (s) , ϕ [2] (s) , ..., ϕ [n] (s) ds. (96)
0

It follows that, ϕ is a solution of the boundary-value problem (91)-(92) if and only if ϕ is a fixed
point of the operator B.

Lemma 10. Let 2T , αη 2 , then operator B1 and B2 given by (95) and (96) are well defined.

Lemma 11. Suppose that condition (93) and (94) hold. If

 
 |α| η 3 + 3T 2 1  4T 2
T ζ    + T  +   ω ≤ L, (97)
3 2T − αη 2  2 2T − αη 2 

and    
 ζ 3T 2 + η 3 |α| 4T ω 
 
   + ζT +    ≤ M, (98)
 3 2T − αη 2  2T − αη 2  

then (B1 ϕ) + (B2 ψ) ∈ Ω, for all ϕ, ψ ∈ Ω .

Lemma 12. Suppose that condition (93) holds. Then the operator B1 defined by (95) is completely
continuous.

Lemma 13. Suppose that condition (94) holds and

 
 4T 2 Xn j=i−1
X 
ki M j  < 1, (99)
   
 
 2T − αη 2  
i=1 j=0

then the mapping B2 given by (96) is a contraction

Theorem 33. Suppose that conditions (93)-(94) and (97)-(98) hold. Then the problem (91)-(92) has
at least one positive bounded solution x in Ω.

Uniqueness
Theorem 34. Suppose that condition (93) and (94) hold and
   2  
   3T + |α| η 3 1  Pn

Pj=i−1 j  
 T    + T  i=1 ci j=0 M  

 3 2T − αη 2  2  < 1,

(100)
 2 Pn Pj=i−1 j  
 + 2T4T k M 
| −αη 2 | i=1 i j=0

then the problem (91)-(92) has a unique solution in Ω.

Continuous dependence
Theorem 35. Suppose that the conditions of Theorem 34 hold. The unique solution of (91)-(92) de-
pends continuously on the functions f and g.

81
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] [Link] and [Link], Positive periodic solutions for revisited Nicholson’s blowflies
equation with iterative harvesting term, J. Math. Anal. Appl.494 (2021), 124663.

[2] [Link] and [Link], Positive periodic solutions for a class of second-order differential
equations with state-dependent delays, Turk J Math. 44 (2020), 1412-1426.

[3] [Link], A. Bouakkaz and R. Khemis , On bounded solutions of asecond-order iterative

boundary value problem, Turk J Math. 46 (2022), 453-464.

[4] [Link], R. Khemis and A. Bouakkaz, Some Existence Results on Positive Solutions for an
Iterative Second-order Boundary-value Problem with Integral Boundary Conditions, Bol. Soc.
Paran. Mat. 40 (2022), 1-10.

82
CNMA’2022
Second National Conference on
Mathematics and its Applications

On Elliptic Differential Equations in Fréchet

Spaces

Said Baghdad 1 said_bg@[Link]

1 Department of mathematics , Ibn khaldoun University - Tiaret

Abstract: The objective of this paper is to present the existence and uniqueness results for
a class of partial sequential fractional differential equations in an appropriate Fréchet space
equipped with a family of measures of nocompactness applying fixed point theorems. Also we
will give an illustrative example.

Keywords: partial sequential fractional differential equations, fixed point theorems, Fréchet
space, measure of nocompactness..
2010 Mathematics Subject Classification: Primary 26A33, 35R11, 35B30, 35J60, 46E10,
47H08, 47H10.

Introduction
Partial differential equations (PDEs) appear frequently in all areas of physics and engineering.
Moreover, in recent years we have seen a dramatic increase in the use of PDEs in areas such as
biology, chemistry, computer sciences (particularly in relation to image processing and graphics)
and in economics (finance). In fact, in each area where there is an interaction between a number
of independent variables, we attempt to define functions in these variables and to model a vari-
ety of processes by constructing equations for these functions. When the value of the unknown
function(s) at a certain point depends only on what happens in the vicinity of this point, we shall,
in general, obtain a PDE. This theory was set apart as a separate scientific discipline. However,
studying partial differential equations still stays closely connected with the description - model-
ing - of physical or other phenomena [3].

Main results
∗
The space of continuous functions C1−γ (J)

We considered a functional space equipped with a family of semi-norms and we have seen that
the fractional operator is continuous on it, after, we proved that it is a complete space. Next, we
∗
defined a family of measures of nocompactness in the space C1−γ (J) in the sens of certain defini-
tion.

Global existence and uniqueness of the solutions

We presented some results regarding to the existence of the solutions for an initial value problem
under sufficient conditions by applying Tychonoff theorem. Also, we have proven the uniqueness
of the solution under additional assumptions.

83
CNMA’2022
Second National Conference on
Mathematics and its Applications

An illustrative example

Consider an initial value problem which is a special case of the studied equation and we checked
that the functions involved in this problem satisfy assumptions of our main theorem, so, this
example indicates the applicability of the study presented in this paper.

References
[1] S. Abbas, M. Benchohra, Impulsive partial hyperbolic functional differential equations
of fractional order with state-dependent delay. Fract. Calc. Appl. Anal. 13 (2010), no. 3,
225–244. DOI = [Link]

[2] [Link], [Link], Nonlinear sequential fractional differential equations in

partially ordered spaces. Filomat 32 (2018), no. 13, 4577-4586. DOI =
[Link]

[3] Y. Pinchover, J. Rubinstein, An introduction to partial differential equations. Cambridge

University Press, Cambridge, 2005.

[4] S. Samko, A. Kilbas, O. I. Marichev Fractional integrals and derivatives (Theorie and Appli-
cations) Gordon and Breach Science Publishers. Yverdon, 1993. ISBN = 978-2881248641.

[5] T. Shivaji, V. V. Kharat, H. T. Dinde, On nonlinear mixed fractional integro-differential

equations with positive constant coefficient. Filomat 33 (2019), no. 17, 5623–5638. DOI =
[Link]

84
CNMA’2022
Second National Conference on
Mathematics and its Applications

Strongly Lorentz summing operators

Aldjia Attallah 1 alattallah99@[Link]

Dahmane Achour 1
1 Department of mathematics , University of M’sila

Abstract:
We introduce and study the new Lorentz ideal of strongly Lorentz summing linear operators in
order to study the adjoints of the Lorentz summing linear operators. We also prove the related
dual result: an operator is Lorentz summing if and only if its adjoint is strongly Lorentz
summing. Some examples, counterexamples and connections with the theory of absolutely
summing operators are given.

Keywords: Absolutely summing operators, Lorentz sequence space, Operator ideals,

Strongly summing operators.
2010 Mathematics Subject Classification: 47B10, 46B45, 46G20, 46A45.

Introduction
The principal idea is to replace, in the definition of summing operators, `p sequences by ab-
stract sequences and see what happen. Since the space of Lorentz sequence spaces is a natural
generalization of `p sequences, the replacement of `p sequences by Lorentz sequence spaces in
the definitions of summing operators and strongly summing operators seems to be an interesting
avenue of investigation.

Main results
The main goal of this thesis is to introduce an approach of summability of operators; more pre-
cisely we introduce the new Lorentz ideal of Lorentz summing operators
In the first part:

Exten lz
• `p hEi `p,q hEi.

In the second part:

• D((p,q);(r,s)) (E, F) lz hFi.

Tb : `p,q (E) → `r,s

85
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] D. Achour and A. Attallah, Strongly Lorentz ((r, s) ; (p, q))-operators, Advances in Opera-
tor Theory 6.2 (2021) 1-24.

[2] A. Attallah and D. Achour, Vector valued Lorentz sequence spaces and their dual spaces,
Colloquium Mathematicum. 166, 4, (2021), 53–73.

[3] J. Diestel, H. Jarchow, A. Tonge, Absolutely summing operators, Cambridge University

Press, Cambridge (1995)

[4] K. Miyazaki, (p, q; r)-absolutely summing operators, J. Math. Soc. Japan 24, 341-354 (1972).

86
CNMA’2022
Second National Conference on
Mathematics and its Applications

Spectral collocation method for solving

integro-differential equations on the
half-line

Boutarcha Sofiane 1 bouterchasofiane06@[Link]

Rahmoune Azedine 1 [Link]@[Link]
1 Department of mathematics , University of Bordj Bou Arreridj, El Anasser,

Abstract: In this work, we give a new class of spectral collocation method for solving
liner Fredholm integro-differential equations on the half-line, this approach is based on the
rationnal scaled Laguerre functions method. At first, we describe some properties of this
functions approximation, and ther, operational matrices of integration and differentiation
are giving to reduce the problem to on algebraic liner system. Also, in order to improve
the obtained numerical solution, corrected residual error is proposed. Finally we give some
numerical examples to illustrate the efficiency and accuracy of the method.

Keywords: Fredholm integro-differential equation on the half-line, scaled Laguerre function

and operational matrix, numerical solutions Science.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
The aim of this paper is to approximate the solution of Fredholm integro-differential equation
of the second kind with the initial conditions on the half-line
Xm Z∞
r
Fr (x)u (x) = f (x) + k(x, t)u(t)dt, m ∈ N, x ≥ 0
r=0 0

With the initial boundary conditions

(λ)
u r (0) = βr , r = 0, 1, ..., m − 1, lim u(x) = 0
x→∞

where k, Fr from 0 to m, and f are known sufficiently smooth functions and u is an unknown
function to be determined.
Integro-differential equations are combinations of differential and integral equations model many
situations of many physical phenomena, including engineering, mechanics, physics, etc.,see [5].
the analytically solution of these problems are usually difficult to find it, for solve it , we need to
use spectral collocation methods.
The spectral methods have been used for solving PDEs for over thirty years, and in the last ten
years became a widely applied for determine numerical solutions of integral equations ,The main

87
CNMA’2022
Second National Conference on
Mathematics and its Applications

merit of spectral methods is the high accuracy. The idea is to write the solution of aur problem as
a sum of certain basis functions.
The paper is organized as follows. in the next section, we introduce some properties and results
of rescaled Laguerre function and a new operational matrix of derivatives and integrations in
section 2. section 3 summarizes the application of this method for solving the Fredholm integro
differential equation with initial conditions, and in section 4 we define the accuracy solution , for
the last section we give same numerical examples.

Main results
Example 1. Consider the second-order differential equation

6 2
u (2) (x) + u (1) (x) + 14u(x) = 4x2 e−x , x ∈ [0, ∞)
x

subject to
u(0) = 1, u (1) (0) = 0, u(∞) = 0.
2
The analytical solution is u(x) = e−x . If we apply scaled Laguerre function collocation method for
different degrees n with different scaling factors β, we have the following results

n β=1 β=2 β=3 β=4 β=5 β=6

4 5.44e-01 3.35e-01 8.86e-02 1.74e-01 6.93e-02 4.03e-02
8 3.08e-01 7.24e-02 4.17e-03 3.36e-03 1.18e-03 1.06e-03
16 5.04e-02 1.24e-03 2.45e-04 6.12e-05 1.62e-05 3.70e-06
32 1.27e-03 2.55e-05 3.99e-06 1.71e-07 5.05e-08 7.27e-09
64 3.66e-05 5.23e-08 6.22e-09 5.94e-11 5.97e-13 3.19e-14
128 8.39e-08 5.06e-11 1.93e-14 1.57e-15 1.25e-15 7.91e-16

Table 1: Example 1: A comparison of the discrete L2 -error for different factor β .

References
[1] A. Rahmoune, Spectral collocation method for solving Fredholm integral equations on the half-
line . Applied Mathematics and Computation,2013.

[2] N. Jewell, Spectral computation with Laguerre functions. The University of Adelaide, Aus-
tralia, 2009.

[3] S-Yuzbasi,M Sezer and B Kemanci, Numerical solution od integro differential equation and
application of population model with an improved Legendre method. applied Mathematical
Modeling ,turkey,2013.

[4] O Coulaud,D Funaro and O Uvian, Laguerre spectral Approximation of Elliptic problems in
exterior domains. Computer methods in applied memechanics and engineering,1990.

88
CNMA’2022
Second National Conference on
Mathematics and its Applications

Burgers Inviscid Equation And Its Modified

Equation

Ilhem Mous 1 mousilhem@[Link]

Abdelhamid Laouar 1
1 Department of mathematics, Badji Mokhtar University of Annaba

Abstract: In this work, we are interested to present a numerical studying of the Burgers’
inviscid equation and its corresponding modified equation. The finite difference method is
used to solve the equation, we add to a numerical dispersion due to truncation errors. Then,
we study the stability and convergence of the solution and make a comparison with some
existing results. To support the theory, numerical simulations are provided as examples.

Keywords: Burgers’ equation, finite difference method ,stability and convergence , trun-
cation errors..
2010 Mathematics Subject Classification: Primary 58J45, 76B25, 65N30.

Introduction
Burgers’ equation is nonlinear equation was initially given by Bateman in 1915. This equation
was introduced in particular by J.M. Burgers as the simplest differential model for a fluid flow,
and is therefore often called the viscous Burgers’ equation. The equation is applicable for various
fields such as gas dynamics, heat conduction, and elasticity. Its form is:

∂u ∂u ∂2 u
+u =v 2,
∂t ∂x ∂x

When v approaches to zero, it becomes the following inviscid Burgers’ equation

∂u ∂u
+u = 0,
∂t ∂x
where u = u(x, t) is the velocity of fluid, which is a prototype of equations that develop shock
waves and has important applications in physics. The modified Burgers’ equation has strong non-
linear aspects and has been used in many practical transport problems.
It often appears as a simplification of the Navier Stokes equations (NSE) and allows us to under-
stand some of the inside behavior of the latter.

In recent years, a class of powerful mathematical methods has been presented to solve the Burgers-
type equation.[8,9] Many researchers have also constructed various numerical schemes to obtain
approximate solutions.
In this work, we are interested in studying a class of problème for Burgers inviscid equation:

89
CNMA’2022
Second National Conference on
Mathematics and its Applications

For any x ∈] − M; M[ and t > 0, find u such as

∂u ∂u
+u = 0,
∂t ∂x

subject to the initial condition

u(x, 0) = φ(x),

where φ(x) is a given function at time t = 0, u = u(x, t) is the velocity of fluid.

Main results
We used the finite difference method to establish some numerical schemes for Burger’s equations
and the corresponding modified equation, then we studied the stability and the convergence. Nu-
merical simulations are obtained and commented on the results already existing.
The results obtaied are encouraging and deseve to be deepened and generalized to other types of
problems.

References
[1] A.G. Bratsos, A fourth-order numerical scheme for solving the modified Burgers equation,
Computers and Mathematics with Applications 60 (2010), 1393–1400.

[2] M. Javidi, A numerical solution of Burger’s equation based on modified extended BDF scheme,
Inter. [Link] 1, no. 32 (2006), 1565–1570.

[3] S. Kutluay, A.R. Bahadir, and A. Ozdes, Numerical solution of one-dimensional Burgers
equation: explicit and exact-explicit finite difference methods., J. of Comput. and Appl. Math.
103 (1999), 251–261.

[4] I. Mous and A. Laouar, A study of the shock wave schemes for the modified Burgers’equation.,
Journal of Mathematical Analysis, 11(1) (2020), 38–51.

90
CNMA’2022
Second National Conference on
Mathematics and its Applications

Weakly mid-(p1, ...pm)-summing multilinear

operators

Athmane Ferradi 1 ferradiathmane@[Link]

1 Department of mathematics, University of M’sila

Abstract: In this paper, we introduce the new ideal of the weakly mid-(p1 , ...pm )-summing
multilinear operators as multilinear version of weakly mid−p− summing linear operators.
Using the space of mid-p-summable sequences, we present a characterization given by
summability property. Also, we give an analogue of the Pietsch domination theorem.

Keywords: Dominated miltilinear operators, weakly mid-p-summing operators, mid p-

summable sequences, Pietsch domination theorem.
2010 Mathematics Subject Classification: 46G25, 47H60, 47B10, 47L20.

Introduction
Karn and Sinha in [4] by using the notion of operator p-summability, they introduced the class of
weakly mid-p-summing operators. There are many attempts available in the literature to study
the ideal of sequentially p-limited operators in different textures, see [2, 4]. Recently, Baweja and
Philip studied some aspects of the operator ideal of sequentially p-limited operators using the the-
ory of tensor products. This concept led them to a new space of summability between p-summable
and weakly p-summable sequeces. A bounded linear operator T : X → Y between Banach spaces
is sequentially p-limited if it sends weakly p-summable sequences to mid p-summable sequences.
Equivalently by Theorem 4.4 [4], S ◦ T : X → `p is p-summing for all bounded linear operator S
from Y to `p . The aim of this paper is to study the multilinear version of this class by considering
the (p1 , ..., pm )-dominated multilinear operators S ◦ T .
This paper is divided into three sections. In the first section, we recall some basic definitions
and notations concerning the linear and multilinear summing operators, some facts on sequence
spaces. We introduce in the second section, a multilinear version of weakly mid-p-summing op-
erators introduced in [4] for which the resulting vector space is a Banach ideal of multilinear
mappings. Finally, in last section we prove a natural analogue of Pietsch domination theorem of
this new class.
For a Banach space X, and 1 ≤ p < ∞. Let us recall some important Banach sequence spaces, and
will be used in the sequel.
The Banach space of absolutely p-summable sequences given by
 
 ∞ 
∞ ∞
 
p
    X
p
 
`p (X) :=  x ⊂ X : k x k := k x k < ∞ .
 
j j p j
j=1 j=1
 

 
 j=1 

91
CNMA’2022
Second National Conference on
Mathematics and its Applications

The Banach space of weakly p-summable sequences given by

 
 ∞    
 ∞  ∞ p
X p 
w  ∗
  
`p (X) :=  x ⊂ X : k x k := sup x x < ∞ .
 
j j w,p j
j=1 j=1
   


 ∗
x ∈BX ∗ j=1 


The Banach space of unconditionally p-summable sequences given by

 ∞  ∞ 
u w
`p (X) := xj ∈ `p (X) : limk k xj kw,p = 0 .
j=1 j=k

Definition 6. [4, 2] A sequence (xj )∞ j=1 in a Banach space X is said to be mid-p-summable,1 ≤ p < ∞,
if ((xn (xj ))j=1 )n=1 ∈ `p (`p ) whenever (xn∗ )∞
∗ ∞ ∞ w ∗
n=1 ∈ `p (X ). The space of all such sequences shall be denoted
by `pmid (X) which is a Banach space under the norm
 1
∞ X
∞     p
 ∞ X p
x∗ xj   .
  
k xj kmid,p := sup  n
j=1  
(xn∗ )∞
n=1 ∈B`p
w (X ∗ )
j=1 n=1

It can be easily seen that,

`p (X) ( `pmid (X) ( `pw (X).

Main results

The following definition was given by Karn and Sinha in [4, Definition 4.1] and [2, Definition 2.1].

Definition 7. Let T be in L(X; Y ). For 1 ≤ p < ∞, we say that T is weakly mid-p-summing, if

(T (xj ))∞ mid ∞ w
j=1 ∈ `p (Y ) whenever (xj )j=1 ∈ `p (X). The set of all weakly mid-p-summing operators from X
into Y will be denoted by Ltp (X, Y ) and ltp (T ) = sup{πp (ST ) : S ∈ L(Y , `p ), kSk ≤ 1} and (Ltp (X, Y ), ltp (.))
is a Banach operator ideal ( see [4, Proposition 4.7]).

Now, we extend the preceding definition to the case of multilinear operators.

Definition 8. Let 1 ≤ p, p1 , · · · , pm < ∞ with 1/p = 1/p1 + · · · + 1/pm . A multilinear operator T :

X1 ×· · ·×Xm → Y is weakly mid-(p1 , . . . , pm )-summing if (T (xj1 , ..., xjm ))∞ mid i ∞
j=1 ∈ `p (Y ) whenever (xj )j=1 ∈
`pwi (Xi )(1 ≤ i ≤ m). The set of all weakly mid-(p1 , . . . , pm )-summing operators from X1 × · · · × Xm into Y
will be denoted by Ltp1 ,··· ,pm (X1 , · · · , Xm ; Y ).

Lemma 14. Let (xj )∞ w

j=1 ∈ `p (X). The following are equivalent.
1- (xj )∞ mid
j=1 ∈ `p (X).
2- (S(xj ))∞
j=1 ∈ `p (`p ) for all S ∈ L(X; `p ). Moreover

k(xj )∞ ∞
j=1 kmid,p = sup{k(S((xj ))j=1 k : S ∈ L(X; `p ), kSk ≤ 1}.

The following proposition characterize the weakly mid-(p1 , . . . , pm )-summing operators.

Proposition 6. For T ∈ L(X1 , · · · , Xm ; Y ), the following statements are equivalent

1- T is weakly mid-(p1 , . . . , pm )-summing.
2- S ◦ T ∈ Πp1 ,··· ,pm (X1 , · · · , Xm ; `p ) for every S ∈ L(Y ; `p ).

92
CNMA’2022
Second National Conference on
Mathematics and its Applications

Theorem 36. For T ∈ L(X1 , · · · , Xm ; Y ), the following statements are equivalent

1- T ∈ Ltp1 ,...,pm (X1 , . . . , Xm ; Y ).
2- The induced map Te : `pw1 (X1 ) × · · · × `pwm (Xm ) → `pmid (Y ) given by

Te((xj1 )∞ m ∞ 1 m ∞
j=1 , · · · , (xj )j=1 ) = (T (xj , · · · , xj ))j=1

is a well-defined continuous multilinear operator.

3- The induced map Tb : `pu1 (X1 ) × · · · × `pum (Xm ) → `pmid (Y ) given by

Tb((xj1 )∞ m ∞ 1 m ∞
j=1 , · · · , (xj )j=1 ) = (T (xj , · · · , xj ))j=1

is a well-defined and continuous multilinear operator.

4- there is a constant C > 0 such that for any (xji )∞ w
j=1 ∈ `pi (Xi ), (1 ≤ i ≤ m) we have

  m
 ∞
T x1 , . . . , xm ∞
Y
j j
≤C xji .
j=1 mid,p
j=1 w,pi
i=1

 
Theorem 37. The class Ltp1 ,...,pm , ltp1 ,...,pm (.) is a Banach ideal of multilinear operators.

Proposition 7. For T ∈ Ltp1 ,··· ,pm (X1 , · · · , Xm ; Y ), we have ltp1 ,··· ,pm (T ) = wp1 ,··· ,pm (T ).

Using the domination theorem for the class of Πp1 ,...,pm we gave the domination theorem the class
of weakly mid-(p1 , ..., pm )-summing multilinear operators.

Theorem 38. Let T ∈ L(X1 , · · · , Xm ; Y ). The following statements are equivalent:

(a) T is weakly mid-(p1 , ..., pm )-summing
(b) There are a constant C > 0 and a regular Borel probability measures µi on BXi∗ such that,
∞ 1/p 
m Z
1/pi
X  Y   pi 
∗ 1 m p i ∗ ∗
|hyn , T (x , · · · , x )i|  ≤ C x x  dµ(x )
  
i i 
 

n=1 i=1 BX ∗
i

for all (yn∗ )∞ w ∗ i

n=1 ∈ `p (Y ), x ∈ Xi .

References
[1] G. Botelho and J.R. Campos, On the transformation of vector-valued sequences by multilinear
operators, Monatsh. Math 183 (2017), 415-435.

[2] G. Botelho, J.R. Campos and J. Santos, Operator ideals related to absolutely summing and
Cohen strongly summing operators, Pacific J. Math. 287 (2017), 1–17.

[3] J. Diestel, H. Jarchow and A. Tonge. Absolutely summing operators. Cambridge University
Press, Cambridge, (1995).

[4] A. Karn and D. Sinha, An operator summability of sequences in Banach spaces, Glasg.
Math. J. 56 (2014), no. 2, 427–437.

[5] M.C. Matos, On multilinear mappings of nuclear type. Rev. Mat. Univ. Complut. Madrid
6(1), 61–81 (1993).

93
CNMA’2022
Second National Conference on
Mathematics and its Applications

Some properties of anisotropic Roumieu

Gelfand-Shilov spaces

M’Hamed Bensaid 1 bensaidx@[Link]

Rachid Chaïli 2
1 Department of mathematics , University of Oran1
2 Department of mathematics , University of Science and technology of Oran

Abstract: The purpose of this work is to prove some properties of anisotropic Roumieu
{M}
Gelfand-Shilov spaces S{N } (Rn ), and to establish the inclusion between them.

Keywords: Gelfand-Shilov spaces, Roumieu sequences.

2010 Mathematics Subject Classification: 46E15; 46F05.

Introduction
{M}
 
The Roumieu Gelfand-Shilov spaces S{N } (Rn ) defined by sequences of positive real numbers Mp
 
and Np satisfying some generic conditions, form a general class of ultra-differentiable functions
including the Gelfand-Shilov spaces. The aim of this paper is show some properties of anisotropic
{M}
Roumieu Gelfand-Shilov spaces S{N } (Rn ). We also investigate the conditions on the sequences
defining the anisotropic Roumieu Gelfand-Shilov spaces to get a comparison in the sense of in-
clusion between these spaces.
{M}
For the definition of Roumieu Gelfand-Shilov spaces S{N } (Rn ), we will consider sequences of
 
positive real numbers Mp satisfying the following conditions:
logarithmic convexity:
M0 = 1 and Mp2 ≤ Mp−1 Mp+1 , ∀p ∈ N∗ (101)
stability under derivation and multiplication
!
p+q
∃H > 0 : Mp Mq ≤ Mp+q ≤ H p+q+1 Mp Mq , ∀p, q ∈ N (102)
q

Example 2. The sequence Mp = p!s , s ≥ 1, satisfies the conditions (101) − (102). It is called Gevrey
sequence of order s.
 
An important property of sequences Mp satisfying the conditions (101) − (102), witch we will
use in the prove of the main results is the following.
 
Proposition 8. For any sequence Mp satisfying condition (101), we have the estimate

∃L > 0 : p! ≤ Lp Mp , ∀p ∈ N (103)

94
CNMA’2022
Second National Conference on
Mathematics and its Applications

Let us
 now
 define
 and enumerate some elementary properties of Roumieu Gelfand-Shilov spaces.
Let Mp and Np be two sequences satisfying the conditions (101) - (102).
{M}
Definition 9. We call anisotropic Roumieu Gelfand-Shilov space in Rn , and we denote S{N } (Rn ), the
space of all functions u ∈ C ∞ (Rn ) such that

∃C > 0, ∀α, β ∈ Zn+ : xβ D α u L∞ (Rn ) ≤ C |α|+|β |+1 M|α| N|β |

{M}
Definition 10. We call Roumieu Gelfand-Shilov space in Rn , and we denote S{M} (Rn ), the space of all
functions u ∈ C ∞ (Rn ) such that

∃C > 0, ∀α, β ∈ Zn+ : xβ D α u L∞ (Rn ) ≤ C |α|+|β |+1 M|α| M|β |

Remark 4. We can take the norm k·kL2 (Rn ) in the above definitions instead of the norm k·kL∞ (Rn ) , accord-
ing to Sobolev imbedding theorems.
{M}
Example 3. If Mp = p!µ , µ > 0 and Np = p!ν , ν > 0 then S{N } (Rn ) is the anisotropic Gelfand-Shilov
µ
space and it is denoted Sν (Rn ).

Main results
Some elementary properties of anisotropic Roumieu Gelfand-Shilov spaces
   
Proposition 9 (Stability under multiplication). If the sequences Mp and Np satisfy (102), then
{M}
the space S{N } (Rn ) is stable by multiplication, i.e.

{M} {M}
∀f , g ∈ S{N } (Rn ) : f g ∈ S{N } (Rn )
   
Proposition 10 (Stability under derivation). If the sequences Mp and Np satisfy (102), then the
{M}
space S{N } (Rn ) is stable by derivation, i.e.

{M} {M}
∀γ ∈ Zn+ , ∀f ∈ S{N } (Rn ) : D γ f ∈ S{N } (Rn )

{M}
   
Corollary 3. If the sequences Mp and Np satisfy (102), then the space S{N } (Rn ) is a differential
subalgebra of C ∞ (Rn ).

Inclusion between Anisotropic Roumieu Gelfand-Shilov spaces

In order to study the inclusion between the spaces of Roumieu Gelfand-Shilov, we introduce the
following definition.
   
Definition 11. Let Mp and Np be two positive sequences, we denote Mp ⊂ Np if

∃L > 0, ∃C > 0 : Mp ≤ CLp Np , ∀p ∈ N

Example 4. If Mp = p!µ , Np = p!ν , with µ ≤ ν, then Mp ⊂ Np .

 
Another example is given by proposition 8: we have p! ⊂ Mp , for any sequence Mp satisfying the
condition (102).

95
CNMA’2022
Second National Conference on
Mathematics and its Applications
       
Proposition 11. Let Mp , Np and Mp , Np be sequences satisfying (101).

{M} {M}
If Mp ⊂ Mp and Np ⊂ Np then S{N } (Rn ) ⊂ S{N } (Rn )

References
[1] R. Chaïli, Systems of differential operators in anisotropic Roumieu classes, Rend. Circ. Mat.
Palermo 62 (2013), 189-198.

[2] I.M. Gelfand and G.E. Shilov,, Generalized Functions II,Academic Press, New York, 1968.

[3] H. Komatsu, Ultradistributions I, structure theorems and characterization, Journal of the

Facul ty of Science, University of Tokyo, Sect. 1A, 20, (1973), 25-105.

96
CNMA’2022
Second National Conference on
Mathematics and its Applications

On the Existence of Solutions for Impulsive

Coupled Differential Equations

Khadidja Mebarki 1 [Link]@[Link]

Ahmed Boudaoui 2 ahmedboudaoui@[Link]
1 Laboratory of Mathematics Modeling and Applications, University of Adrar,
2 note1

Abstract: In this presentation, we employ coupled fixed points to give sufficient conditions to
guarantee a solution to system of differential equations with impulse effects. The proof of the
existence theorem is based on a coupled fixed point theorem in b−metric space endowed with
directed graph.

Keywords: Impulse differential equations, Coupled fixed point, b−metric space, Directed
graph.
2010 Mathematics Subject Classification: 47H10, 34A37.

Introduction
One of the best tools in applied sciences that can be used to determine an existence solution
for such integral equation or differential equation is the fixed point theory.
In 2006, Bhaskar and Lakshmikantham [1] employed coupled fixed points to give sufficient
conditions to solve some differential equations by introducing and proving many nice results for
coupled fixed points.
We start with the notion of b-metric space.

Definition 12. [3] Given s ≥ 1. On the set X, let d : X × X → R+ be a map such that:

1. d(x, y) = 0 if and only if x = y,

2. d(x, y) = d(y, x),

3. d(x, y) ≤ s[d(x, z) + d(z, y)]

hold for all x, y, z ∈ X. Then, we refer the duo (X, d) to a b-metric space.

In 1987, Guo and Lakshmikantham [1] introduced the following notion of coupled fixed point
.

Definition 13. The pair (x, y) ∈ X × X is called a coupled fixed point of

T : X × X → X if
x = T (x, y) and y = T (y, x).

97
CNMA’2022
Second National Conference on
Mathematics and its Applications

Definition 14. [2] Endowed the complete metric space (X, d) with the direct graph G. The mapping
T : X × X → X possess the mixed G-monotone property if

(x1 , x2 ) ∈ E(G) ⇒ (T (x1 , l), T (x2 , l)) ∈ E(G),

for all x1 , x2 , l ∈ X, and

(y1 , y2 ) ∈ E(G) ⇒ (T (l, y2 ), T (x, y1 )) ∈ E(G),
for all l, y1 , y2 ∈ X.

Let (X, d) stands to a b-metric space endowed with directed graph G such that the set V (G) = X,
∆ ⊆ E(G), G is transitive ((a, b) ∈ E(G) and (b, c) ∈ E(G) implies that (a, c) ∈ E(G)) and has no parallel
edges. The mapping T : X × X → X possess the mixed G-monotone property. Further, we endow
the product space X × X by another graph denoted also by G, such that

((x, y), (u, v)) ∈ E(G) ⇔ (x, u) ∈ E(G) and (v, y) ∈ E(G),

for any (x, y), (u, v) ∈ X × X. We considered the following contraction.

Definition 15. The mapping Let T : X × X → X is called b−contraction if there exist α, β, γ ∈ [0, 1)
with
∞ !i
X
i β +γ
s <∞
1−α−β
i=0
such that
d(x, T (x, y))[1 + d(u, T (u, v))]
d(T (x, y), T (u, v)) ≤ α
1 + d(x, u)
+ β[d(x, T (x, y)) + d(u, T (u, v))] + γd(x, u),

holds for all (x, y), (u, v) ∈ X × X with ((x, y) , (u, v)) ∈ E(G).

We employ the following results to assure the solutions of some impulsive differential equa-
tions are exist under some conditions.

Theorem 39. On (X, d, G), suppose that T is continuous and b−contraction mapping. If there exist
x0 , y0 ∈ X such that ((x0 , y0 ) , (T (x0 , y0 ), T (y0 , x0 ))) ∈ E(G), then T possess a coupled fixed point.

The continuity of T in Theorem 39 can be discarded by adding some new conditions. Assume
that (X, d, G) possess the following property.

Property 1. 1. for any {xn }n∈N in X such that (xn , xn+1 ) ∈ E(G) and lim xn = x, then (xn , x) ∈ E(G),
n→+∞

and

2. for any {xn }n∈N in X such that (xn+1 , xn ) ∈ E(G) and lim xn = x then (x, xn ) ∈ E(G),
n→+∞

Theorem 40. Endowed (X, d, G) with the property1. Suppose that T is b-contraction mapping. If there
exist x0 , y0 ∈ X such that ((x0 , y0 ) , (T (x0 , y0 ), T (y0 , x0 ))) ∈ E(G), then T possess a coupled fixed point.

Main results

98
CNMA’2022
Second National Conference on
Mathematics and its Applications

Let consider the following system

x0 (t) = f (t, x(t), y(t)), y 0 (t) = f (t, y(t), x(t)), t ∈ J\{t1 , . . . , tm } (104)

x(tk+ ) − x(tk− ) = Ik (x(tk ), y(tk )), y(tk+ ) − y(tk− ) = Ik (y(tk ), x(tk )), t = 1, . . . , m (105)
x(0) = x0 , y(0) = y0 , (106)
where 0 < t < 1, J := [0, 1], f : J × R × R → R, Ik ∈ C(R × R, R). The notations x(t + ) = lim+ x(t + h) and
h→0
x(t − ) = lim+ x(t − h).
h→0

In order to define a solutions for Problem (104)–(106), consider the space of piecewise contin-
uous functions:

P C(J, R) ={y : J → R, y ∈ C(J\{tk }, R); such that y(tk− ) and y(tk+ )

exist and satisfy y(tk− ) = y(tk )}.

Definition 16. A function y ∈ P C is said to be a solution of (104)–(106), if y satisfies (104)–(106).

Define d on P C(J, R) by
d(x, y) = (sup |x(t) − y(t)|)2 .
t∈J

We study the existence of a solution to the previous system.

Assumption 1. Assume the following assertions:

1. f : J × R × R → R is continuous.

2. for all x, y, u, v ∈ P C(J, R), with x ≤ u and v ≤ y, we have

f (t, x(t), y(t)) ≤ f (t, u(t), y(t)) and Ik (x(tk ), y(tk )) ≤ Ik (u(tk ), v(tk )) ∀ t ∈ J;

∞
β+γ i
 
2i
P
3. there exist α, β, γ ∈ [0, 1) with 1−α−β < ∞ such that
i=0

α |x(t) − f (t, x(t), y(t))|2 [1 + |u(t) − f (t, u(t), v(t))|2 ]

|f (t, x(t), y(t)) − f (t, u(t), v(t)|2 ≤
2 1 + |x(t) − u(t)|2
β h i
+ |x(t) − f (t, x(t), y(t))|2 + |u(t) − f (t, u(t), v(t))|2 ,
2
and
γ 
|Ik (x(tk ), y(tk )) − Ik (u(tk ), v(tk ))|2 ≤ |x(t) − u(t)|2
2
and for each t ∈ J, x, y, u, v ∈ P C(J, R), x ≤ u and v ≤ y.

Consider on P C(J, R) × P C(J, R) the partial order relation:

(x1 , y1 ) ≤ (x2 , y2 ) ⇔ x1 (t) ≤ x2 (t) and y1 (t) ≥ y2 (t), t ∈ J,

Rt
and define for t ∈ J, T (x, y)(t) = x0 + 0
f (s, x(s), y(s))ds + Ik (x(tk ), y(tk )), t ∈ J.

99
CNMA’2022
Second National Conference on
Mathematics and its Applications

Theorem 41. Assume the Assumption 1 holds. Assume that there exists (u0 , v0 ) ∈ P C(J, R) × P C(J, R)
such that Z t
u0 (t) ≤ u0 (0) + f (s, u0 (s), v0 (s))ds + Ik (u(tk ), v(tk ))
0
and
Z t
v0 (t) ≥ v0 (0) + f (s, v0 (s), u0 (s))ds + Ik (v(tk ), u(tk )), t ∈ J.
0
Then the system (104)–(106) possess a solution.

Proof We prove that the integral system (104)–(106) has a solution by showing that the operator
T : X × X → X has a coupled fixed point in X × X. To do this, we have to show that that T satisfies
the conditions of Theorem 39 or Theorem 40. 

References
[1] [Link] and [Link]. Coupled fixed points of nonlinear operators with applica-
tions,Nonlinear analysis: theory, methods & applications, 11(5), 623–632 ,(1987).

[2] M.R. Alfuraidan and M. Khamsi. Coupled fixed points of monotone mappings in a metric
space with a graph, arXiv preprint arXiv:1801.07675 (2018).

[3] S. Czerwik. Nonlinear set-valued contraction mappings in b-metric spaces, AttiSem. Mat. Fis.
Univ. Modena, 46, 263–276 (1998).

[4] T.G. Bhaskar and V. Lakshmikantham. Fixed point theorems in partially ordered metric
spaces and applications, Nonlinear Analysis: Theory, Methods & Application, 65 (7), 1379–
1393 (2006).

100
CNMA’2022
Second National Conference on
Mathematics and its Applications

Solvability of a dynamic frictional contact

problem

Imane Ouakil 1 [Link]@[Link]

Benyattou Benabderrahmane 1
Yamna Boukhatem 2
1 Laboratory of Pure and Applied Mathematics , University of Mohamed Boudiaf, M’sila, Algeria
2 Laboratory of Pure and Applied Mathematics, University of Laghouat, Laghouat, Algeria.

Abstract: A dynamic contact problem between a viscoelastic body and an obstacle is consid-
ered. The contact is modelled with normal compliance. We derive a variational formulation of
the model. Then, we prove the unique weak solvability of solution. The proof is based on the
theory of nonlinear equations and fixed point arguments.

Keywords: Frictional contact , dynamic contact, normal compliance condition, fixed

point.
2010 Mathematics Subject Classification: 34A12, 70G45, 74M10.

Introduction
In this paper, we consider the following dynamic contact problem with normal compliance con-
dition:

Problem P . Find a displacement field u : Ω × [0, T ] → Rd and a stress field σ : Ω × [0, T ] → S d

σ = Aε(ut ) + Gε(u) in Ω × (0, T ),

ρü = Divσ + f0 in Ω × (0, T ),
u=0 on Γ1 × (0, T ),
σ ν = f2 on Γ2 × (0, T ), (107)
−σν = pν (uν ) on Γ3 × (0, T ,
−στ = pτ (uτ ) on Γ3 × (0, T ),
u(0) = u0 , u̇(0) = u1 in Ω,

where the domain Ω ⊂ Rd (d = 2, 3) with a regular surface Γ that is partitioned into three disjoint
measurable parts Γ1 , Γ2 and Γ3 such that meas Γ1 > 0. Let [0, T ] denote the time interval of interest.
We note by u the displacement vector, ut represents the velocity vector, σ represents the stress
field and ε = ε(u) is the small strain tensor. The viscoelastic constitutive law is assumed to be

σ = Aε(ut ) + Gε(u) in Ω × (0, T )

101
CNMA’2022
Second National Conference on
Mathematics and its Applications

in which A, G are nonlinear given functions .

In this paper, we consider the contact condition of the form

−σν = p (uν ) on Γ3 × (0, T ),

where σν is the normal stress, uν is the normal displacement, pν is a prescribed function which
equals to zero when its argument is negative.
we assume that the friction law is given by

−στ = pτ (uτ ) on Γ3 × (0, T ),

where στ is the tangential stress, uν is the tangential displacement, pτ is a prescribed function

which equals to zero when its argument is zero.
In the study of problem (107), we assume that the viscosity operator A, the elasticity operator G
and the contact functions pν and pτ satisfy




 (a) A : Ω × S d → S d ;





 (b) There exists LA > 0 such that

∀ε1 , ε2 ∈ S d , a.e. x ∈ Ω;

kA (x, ε1 ) − A (x, ε2 ) k ≤ LA kε1 − ε2 k






(c) There exists mA > 0 such that (108)





(A (x, ε1 ) − A (x, ε2 )) · (ε1 − ε2 ) ≥ mA kε1 − ε2 k2 ∀ε1 , ε2 ∈ S d , a.e. x ∈ Ω;










 (d) x → A(x, ε) is Lebesgue measurable on Ω;


(e) A(x, 0) belongs toH.







 (a) G : Ω × S d → S d ;





 (b) There exists LG > 0 such that

∀ε1 , ε2 ∈ S d , ∀α1 , α2 ∈ R a.e. x ∈ Ω;

kG (x, ε1 ) − G (x, ε2 ) k ≤ LG (kε1 − ε2 k+ | α1 − α2 |)






(c) There exists mG > 0 such that (109)





(G (x, ε1 ) − G (x, ε2 )) · (ε1 − ε2 ) ≥ mG kε1 − ε2 k2 ∀ε1 , ε2 ∈ S d , a.e. x ∈ Ω;






(d) ∀ ∈ S d andα ∈∈ R, x → G(x, ε, α) is Lebesgue measurable on Ω;







(e) G(x, 0, 0) belongs toH.







 (a) pν : Γ3 × R −→ R;


b) there exists a constant Lν such that






kpν (x, r1 ) − pν (x, r2 ) k ≤ Lν kr1 − r2 k ∀r1 , r2 ∈ R a.e x ∈ Γ3 ;



(110)


c) (pν (x, r1 ) − pν (x, r2 )) (r1 − r2 ) ≥ 0 ∀r1 , r2 ∈ R a.e.x ∈ Γ3 ;










 d) The mapping x −→ pν (x, r) is measurable on Γ3 for any r ∈ R;


e) pν (x, r) = 0 ∀r ≤ 0 a.e.x ∈ Γ3 .



102
CNMA’2022
Second National Conference on
Mathematics and its Applications




 (a)pτ : Γ3 × Rd → Rd ;


b) there exists a constant Lτ such that






kpτ (x, r1 ) − pτ (x, r2 ) k ≤ Lτ kr1 − r2 k ∀r1 , r2 ∈ R a.e x ∈ Γ3 ;



(111)


c) (pτ (x, r1 ) − pτ (x, r2 )) (r1 − r2 ) ≥ 0 ∀r1 , r2 ∈ R a.e.x ∈ Γ3 ;






d) The mapping x −→ pτ (x, r) is Lebesgue measurable on Γ3 for any r ∈ Rd ;







e) pτ (x, r) = 0 if r = 0Rd a.e.x ∈ Γ3 .



We also suppose that the forces and the tractions have the following regularity
 
f0 ∈ L2 (0, T ; H) , f2 ∈ L2 0, T ; L2 (Γ2 )d . (112)

Next, we define the mapping f : [0, T ] → V 0 by

Z Z
(f (t), v)V 0 ×V = f0 (t).vdx + f2 (t).vda, (113)
Ω Γ2

for all v ∈ V , t ∈ [0, T ].

It follows from assumption (112) that

f ∈ L2 (0, T ; V 0 ) . (114)

The initial data satisfy the following condition

u0 ∈ V , v0 ∈ H (115)

The study of the contact problems has attracted many researchers see for instace ([1], [2], [4]).

Main results
First, our attention is turned to a derivation of variational formulation of our contact problem
Using Green’s formula, we obtain

Problem PV Find a displacement field u : [0, T ] → V and a stress field σ : [0, T ] → H such that

σ (t) = Aε(u̇) + Gε(u(t)) (116)

(ü, v)V 0 ×V = −(σ (t), ε(t)H − j(u(t), v) + (f (t), v)V 0 ×V ; ∀v ∈ V , a.e. t ∈ (0, T ). (117)
u(0) = u0 , u̇(0) = v0 . (118)
where the functional j : V × V → R is given by
Z Z
j(u, v) = pν (uν ) vν + pτ (uτ ) vτ da, ∀u, v ∈ V , (119)
Γ3 Γ3

Next, we state our main theorem

Theorem 42. Assume that conditions (108)-(111), (112) and (115) hold. Then, Problem PV has a
unique solution (u, σ ) which satisfies

u ∈ W 1,2 (0, T ; V ) ∩ C 1 (0, T ; H), ü ∈ L2 (0, T ; V 0 ) (120)

σ ∈ L2 (0, T ; H), Div σ ∈ L2 (0, T ; V 0 )

103
CNMA’2022
Second National Conference on
Mathematics and its Applications

To prove the previous theorem, we need the following lemmas

Let η ∈ L2 (0, T ; V 0 ) be given.
We consider the following variational problem P Vη

Problem P Vη

(üη , w)V 0 ×V + (A(εu̇η ), ε(w))H + (η(t), w)V 0 ×V = (f (t), w)V 0 ×V ∀w ∈ V a.e t ∈ (0, T ),

uη (0) = u0 , u̇η (0) = v0 .

Lemma 15. There exists a unique solution to problem P Vη satisfying the regularity expressed in (120).
Moreover, if ui represents the solution of problem P Vη for η = ηi ∈ L2 (0, T ; V 0 ) , i = 1, 2, then there exists
C > 0 such that Z t Z t
ku̇1 (s) − u̇1 (s)k2V ds ≤ C kη1 (s) − η2 (s)k2V 0 ds ∀t ∈ [0, T ]
0 0

Now, we introduce the operator Λ : L2 (0, T ; V 0 ) → L2 (0, T ; V 0 ) defined by

   
(Λη(t), v)V 0 ×V = Gε uη (t) , ε(v) + j(uη , v)
H

Lemma 16. The operator Λ has a unique fixed point η ∗ ∈ L2 (0, T ; V 0 ).

References
[1] A Klarbring, A Mikelić, and M Shillor, frictional contact problems with normal compli-
ance, International Journal of Engineering Science, 26(8):811–832, 1988.

[2] M Campo, JR Fernández, W Han, and M Sofonea, a dynamic viscoelastic contact problem
with normal compliance and damage, Finite elements in analysis and design, 42(1):1–24,
2005.

[3] Soumia Latreche and Lynda Selmani, Dynamic contact problem with normal damped
response, friction and adhesion, Mediterranean Journal of Mathematics, 18(3):1–21, 2021.

[4] O Chau, A Petrov, A Heibig, and M Marques, a frictional dynamic thermal contact prob-
lem with normal compliance and damage,Nonlinear Analysis and Global Optimization,
pages 71–[Link], 2021.

104
CNMA’2022
Second National Conference on
Mathematics and its Applications

Some results concerning meromorphic

solutions for the Shroder equations

Bouternikh Salih 1 bouternikhsalah18@[Link]

Zerzaihi Tahar1
1 Department of mathematics , University of Mohamed Seddik Ben Yahia, Jijel

Abstract: Let K be an algebraically closed field, complete for an ultrametric absolute value and
let M(K) be the field of meromorphic functions in all K. In this paper, we consider the ultra-
n P (x, f (x))
Aj (x)f (qj x) = R(x, f (x)) =
P
metric functional equation of Shröder-type: , where q is
j=1 Q(x, f (x))
an element of K, A1 (x), . . . , An (x) are rational functions and P , Q are relatively prime polyno-
mials in f over the field of rational functions satisfying p = degf P , t = degf Q, d = p − t > 2.
First we prove that, if all coefficients Aj (x) j = (1, . . . , n) and all coefficients of R(x, f (x)) are
constants, then every meromorphic solution of our equation is a rational function.
Next, we study meromorphic solutions of the above equation in the case where the coefficients
are rational and we give some estimates about the order of growth of these solutions.

Keywords: Nevanlinna theory, Ultrametric meromorphic solution, Difference equations,

Order of growth.
2010 Mathematics Subject Classification: Primary 11J97, 12H10.

Introduction
Let K be an algebraically closed field, complete for an ultrametric absolute value. We denote
by A(K) the K-algebra of entire functions in K and M(K) the field of meromorphic functions in
K.
Recently, many papers focused on the properties of meromorphic solutions of difference equations
and q−difference equations and many meaningful results have been obtained about the growth of
their solutions in both complex and ultrametric cases. The main objective of this paper is to
generalize some of their results to the case of ultrametric equations of Schröder-type.
Throughout this paper, we use standard notations in the ultrametric Nevanlinna theory (see [3, 4,
3]). So, we have to recall some basic notations of this theory.
Let f (x) = n>0 an xn be an analytic function on A(d(0, R− )), for all r ∈]0, R[, the function |.|(r)
P
defined by |f |(r) = supn≥0 |an |r n is a multiplicative norm on A(d(0, R− )). This norm is extended
to M(d(0, R− )) as follows: if f ∈ M(d(0, R− )) is given by f = h/g, with h, g ∈ A(d(0, R− )), we write
|f |(r) = |h|(r)/|g|(r).
Finally, for every f ∈ M(d(0, R− )) \ {0} and every α ∈ d(0, R− ), we denote by ωα (f ) the integer iα of
Z such that f (x) = i>iα ai (x − α)i and aiα , 0.
P

We introduce some notations and properties of ultrametric Nevanlinna theory. For every r ∈]0, R[,

105
CNMA’2022
Second National Conference on
Mathematics and its Applications

let f be a non-constant meromorphic function on the disk d(0, r). Using the notation log+ a =
max{log a, 0} (where log is the real logarithm function), we define the compensation function by

m(r, f ) = log+ |f |(r) = max{log |f |(r), 0}.

Let f ∈ M(d(0, R− )) such that 0 is neither a zero nor a pole of f . For every r ∈]0, R[, Z(r, f ) will
denote the counting function of zeros of f in the disk d(0, r) counting multiplicity. It is defined by
X r
Z(r, f ) = ωα (f ) log ,
|α|
ωα (f )>0,|α|≤r

and N (r, f ) will denote the counting function of poles of f in the disk d(0, r) counting multiplicity.
It is defined by
1 r
  X
N (r, f ) = Z r, =− ωα (f ) log .
f |α|
ωα (f )<0,|α|≤r

Now, we define the characteristic function of f as follows

T (r, f ) = N (r, f ) + m(r, f ).

The function r 7→ T (r, f ) is called the Nevanlinna function or characteristic function of Nevan-
linna.
Finally, the order of growth of a meromorphic function f in K is defined by

log T (r, f )
ρ(f ) = lim sup .
r→∞ log r

In the particular case when f is an entire function in K, we have

log(log |f |(r))
ρ(f ) = lim sup .
r→∞ log r

Main results
At the beginning of this section, we will study the following equations
n Pp
X P (f (x)) Bi f (x)i
j
Aj f (q x) = R(f (x)) = = Pti=0 , (121)
Q(f (x)) i=0 Ci f (x)
i
j=1

where q ∈ K such that 0 < |q| < 1, Aj , Bi , Ci are constants and p − t > 2.

Theorem 43. If f ∈ M(K) is a meromorphic solution of equation (121), then f is a rational function
which has no poles α , 0.

We will use the following lemma which proof is easily checked.

Lemma 17. For every f ∈ M(K), every r > 0 and every n ∈ N, we have:

1. |f ◦ σqn |(r) = |f |(|q|n r),

2. m(r, f ◦ σqn ) = m(|q|n r, f ),

3. N (r, f ◦ σqn ) = N (|q|n r, f ),

106
CNMA’2022
Second National Conference on
Mathematics and its Applications

4. T (r, f ◦ σqn ) = T (|q|n r, f ),

such that σq (x) = qx and σqn = σq ◦ σq ◦ σq · · · σq n times.

In the second part of this section, we are interested in the case of non-constant coefficients in the
above equation, so we will study the following equation
n
X P (x, f (x))
Aj (x)f (qj x) = R(x, f (x)) = , (122)
Q(x, f (x))
j=1

where q ∈ K such that 0 < |q| < 1, Aj (x) are rational fractions and P , Q are relatively prime polyno-
mials in f over the field of rational functions satisfying p = degf P , t = degf Q, d = p − t > 2. Then,
we obtained the following results

Theorem 44. All transcendental meromorphic solutions f of equation (122), which have at most a
finite number of poles, satisfy
T (r, f ) = O((log r)2 ), r → ∞.

Theorem 45. All transcendental meromorphic solutions f of equation (122), which have an infinity of
poles, satisfy
(log r)2 = O(T (r, f )), r → ∞.

For the proofs of our theorems we need the following lemma.

Lemma 18. If f ∈ M(K) \ K(X) is a solution of equation (122) and if it has infinitely many poles, then
there is a constant R such that: for every r > R, f admits at least one pole in d(0, r|q|−n ) \ d(0, r).

References
[1] A. Boutabaa, Applications de la théorie de Nevanlinna p-adique, Collectanea Mathematica,
42 (1991), 75-94.

[2] A. Boutabaa and A. Escassut, Applications of the p-adic Nevanlinna theory to functional
equations, Annales de l’institut Fourier, 50 (2000), 751-766.

[3] A. Escassut, Analytic Elements in p-adic Analysis, World Scientific (1995).

[4] G. G. Gundersen, J. Heittokangas, I. Laine, J. Rieppo, and D. Yang, Meromorphic solutions

of generalized Schroder equations, Aequationes Math, 63 (20002), 110–135.

[5] P.-C. Hu and C.-C. Yang, Meromorphic functions over non-Archimedean fields, Springer Sci-
ence (2000).

107
CNMA’2022
Second National Conference on
Mathematics and its Applications

General Decay of a Nonlinear

Viscoelastic Wave Equation

Amel Boudiaf 1 [Link]@[Link]

1 Department of mathematics , University of setif, Algeria

Abstract: In this work we establish a general decay rate for a nonlinear viscoelastic wave
equation with boundary dissipation where the relaxation function satisfies g 0 (t) ≤ −ξ (t) g p (t) ,
t ≥ 0, 1 ≤ p ≤ 23 .

Keywords: Viscoelastic, General decay, Relaxation function, Dissipation, Wave equa-

tion.
2010 Mathematics Subject Classification:35B37, 35L55, 74D05, 93D15, 93D20.

Introduction
It is well known that viscoelastic materials have memory effects, which is due to the mechani-
cal response influenced by the history of the materials themselves. As these materials have a wide
application in the natural sciences, their dynamics are interesting and of great importance. From
the mathematical point of view, their memory effects are modeled by integrodifferential equa-
tions. Hence, questions related to the behavior of the solutions for the PDE system have attracted
considerable attention in recent years. In the present work, we are concerned with
Rt
utt − k0 4 u (t) + 0 g (t − s) div (a (x) ∇u (s)) ds + b (x) ut = |u|γ−2 u, in Ω × (0, ∞)
∂u R t
k0 − 0 g (t − s) (a (x) ∇u (s)) .νds + h (ut ) = 0, on Γ1 × (0, ∞)
∂ν (123)
u = 0, on Γ0 × (0, ∞)
u (x, 0) = u0 , ut (x, 0) = u1 , x ∈ Ω.

Where k0 > 0 and Ω is a bounded domain in Rn (n ≥ 1) with a smooth boundary, Γ = Γ0 ∪Γ1 . Here Γ0
and Γ1 are closed and disjoint with meas(Γ0 ) > 0, and ν is the unit outward normal to Γ . b : Ω → R+
is a function, and
2n
2 < γ≤ , n ≥ 3,
n−2
γ > 2, if n = 1, 2.

Our aim in this work is to obtain a more general and explicit energy decay formula, from which
the usual exponential and polynomial decay rates are only special cases of our result.

Main results
We state the assumptions for problem (123)) as follows.

108
CNMA’2022
Second National Conference on
Mathematics and its Applications

(A1) g : R+ → R+ is a nonincreasing differentiable function such that

Z∞
g (0) > 0, k0 − g (s) ds = l > 0.
0

(A2) There exists a nonincreasing differentiable function ξ : R+ → R+ , with ξ (0) > 0, satisfying

3
g 0 (t) ≤ −ξ (t) g p (t) , ∀t ≥ 0, 1 ≤ p < .
2
(A3) h : R → R is a nondecreasing function with

h (s) s ≥ α |s|2 , ∀s ∈ R,

|h (s)| ≤ β |s| , ∀s ∈ R
We introduce the following functionals
Zt !
1 1 1 γ
J (t) = k0 − a (x) g (s) ds k∇uk22 + (g ◦ ∇u) (t) − kukγ
2 0 2 γ
1
E (t) = J (u (t)) + kut k22 , for t ∈ [0, T )
2
,

where E (t) is the energy functional.

A direct differentiation, using (123), leads to
Z Z
1 0 1 2
0
E (t) = (g ◦ ∇u) (t) − a (x) g (t) |∇u (t)| dx − b (x) |ut (t)|2 dx ≤ 0.
2 2 Ω Ω

our main result.

 
Theorem 46. Let (u0 , u1 ) ∈ HΓ1 × L2 (Ω) be given . Assume that (A1)-(A4) are satisfied. Then, for any
0
t0 > 0, there exist two positive constants K, and λ such that the solution of (123)) satisfies, for all t ≥ t0 ,
Rt
−λ ξ(s)ds
E (t) ≤ Ke t0
, if p = 1.
 1
 2p−2
 1 
E (t) ≤ K  , p > 1. (124)
 
 1 + t ξ 2p−1 (s) ds 
R 
t 0

Moreover, if
1
Z +∞ " # 2p−2
1 3
2p−1
dt < +∞, 1 < p < , (125)
0 tξ (t) + 1 2
then 1
  p−1
 1 

E (t) ≤ K   , p > 1.

 1 + t ξ p (s) ds 
R
t 0

Simple calculations show that (124) and (125) yield

Z +∞
E (t) dt < +∞.
t0

109
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] S. Berrimi, S.A Messaoudi, Exponential decay of solutions to a viscoelastic equation with
nonlinear localized damping, Electronic Journal of Differential Equations, vol. 88, (2004) ,
1-10.

[2] M.M Cavalcanti, V.N Domingos Cavalcanti, J.A Soriano, Exponential decay for the so-
lution of semilinear viscoelastic wave equations with localized damping, Electronic Jour-
nal of Differential Equations, vol. 44, (2002) , 1–14.

[3] F. Li, N. Zhao, Uniform energy decay rates for nonlinear viscoelastic wave equation with
nonlocal boundary damping, Nonlinear Analysis: Theory, Methods & Applications, vol.
74, no. 11, (2011) , 3468–3477.—

[4] W. Shun-Tang, Hsueh-Fang Chen, Uniform decay of solutions for a nonlinear viscoelastic
wave equation with boundary dissipation, Journal of Function Spaczs and Applications,
(2012) , vol. Article ID 421847.

110
CNMA’2022
Second National Conference on
Mathematics and its Applications

Well posedness of solution for laminated

timoshinko beams and delay term

SABAH BAIBECHE 1 sabahbaibeche@[Link]

1 Department of mathematics , Laboratory of Applied Mathematics and History and Didactics of Mathematics
(LAMAHIS), University of 20 August 1955, Skikda, Algeria ,

Abstract: In this work, we consider a laminated Timoshenko beams with a delay term. We
prove well-posedness by using Faedo Galerkin method.
Keywords: laminated Timoshenko beams, Faedo Galerkin method, delay term, Hille Yosida
theorem..
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
We consider the laminated beam system where the heat flux is given by Fourier’s law with de-
lay term (See[14]). The system is written as




 ρωtt + G (ψ − ωx )x + µ1 ωt + µ2 (s)ωt (t − s) ds = 0, x ∈ (0, 1),


Iρ (3Stt − ψtt ) − G (ψ − ωx )x − D (3Sxx − ψxx ) = 0 (126)





 I S + G (ψ − ω ) + 4 γS + 4 βS − DS = 0

ρ tt x 3 3 t xx

Under the boundary conditions




 ω (0, t) = ξ (0, t) = S (0, t) = 0, t > 0


ξx (1, t) = Sx (1, t) = 0, t > 0





 3S (1, t) − ξ (1, t) ω (1, t) = 0, t > 0

x

and the initial conditions


 (ω, ξ, S)|t=0 = (ω0 , ξ0 , S0 ) , (ωt , ξt , St )|t=0 = (ω1 , ξ1 , S1 )




ωt (x, −t) = f0 (x, t) in (0, 1) × (0, τ2 )




with ξ = 3S − ψ
Here, we prove the well-posedness and stability results for problem on the following parameter,
under the assumption Z τ2
µ0 ≥ µ (s) ds. (127)
τ1

111
CNMA’2022
Second National Conference on
Mathematics and its Applications

Main results
By using Hille Yosida theorem, we state and prove the well posedness of problem (126)-(127)
and by using the perturbed energy method, we then establish the exponential result if and only if
ρ1 ρ2
G = D.

As in [24] , we introduce the new variable

z(x, ρ, s, t) = ϕt (x, t − ρs) in (0, 1) × (0, 1) × (τ1 , τ2 ) × (0, ∞).

It is straight forward to check that z satisfies

szt (x, ρ, s, t) + zρ (x, ρ, s, t) = 0 in (0, 1) × (0, 1) × (τ1 , τ2 ) × (0, ∞).

Consequently, problem (126) is equivalent to

 R τ2



 ρ1 ϕtt − a1 ϕxx − a2 ψxx + µ1 ϕt + τ1
µ2 (s)z (x, 1, s, t) ds = 0, in (0, 1) × (0, ∞)


ρ2 ψtt − a3 ψxx + a2 ϕxx + δθx = 0, in (0, 1) × (0, ∞)










 ρ3 θt − qx + δψtx = 0, in (0, 1) × (0, ∞)


τqt + βq + θx = 0, in (0, 1) × (0, ∞)






szt (x, ρ, s, t) + zρ (x, ρ, s, t) = 0, in (0, 1) × (0, 1) × (τ1 , τ2 ) × (0, ∞)





ϕ (x, 0) = ϕ0 (x) , ϕt (x, 0) = ϕ1 (x) , θ(x, 0) = θ0 (x) in (0, 1),










 ψ (x, 0) = ψ0 (x) , ψt (x, 0) = ψ1 (x) , q (x, 0) = q0 (x) in (0, 1),


ϕ(0, t) = ϕ(1, t) = ψx (0, t) = ψx (1, t) = θ(0, t) = θ(1, t) = 0 in (0, ∞),






z (x, 0, t, s) = ϕt (x, t) in (0, 1) × (0, ∞) × (τ1 , τ2 ) ,








 z (x, ρ, 0, s) = f0 (x, ρ, s) in (0, 1) × (0, 1) × (τ1 , τ2 ) ,

References
[1] Racke, R.; Instability of coupled systems with delay, Comm. Pure, Appl. Anal., 11 (5), (2012).

[2] Beuter, A.; B élair, J.; Labrie, C.; Feedback and delays in neurological diseases: a modeling
study using dynamical systems, Bull. Math. Bio., 55 (3), 525-541 (1993).

[3] Joseph, D. D.; Preziosi, L.; Heat waves, Reviews of Modern Physics 61 (1), 41â” “73 (1989).

[4] Tarabek, M. A.; On the existence of smooth solutions in one-dimensional nonlinear ther- moe-
lasticity with second sound, Quart. Appl. Math. 50 (4), 727â” “742 (1992).

112
CNMA’2022
Second National Conference on
Mathematics and its Applications

Null Controllability Of Nonlocal ψ-Hilfer

fractional Differential Equation

Ichrak Bouacida 1 ichrakbouacida@[Link]

Mourad Kerboua 1
Sami Segni1
1 Laboratoire des Mathématiques Appliquées et de Modélisation,(LMAM), Université 8 Mai 1945 Guelma

Abstract: In this work, we study exact null controllability of ψ-Hilfer fractional semilineair
differential equation with ψ- fractional nonlocal conditions in Hilbert spaces. At first, a
group of sufficient conditions is established for the existence of mild solutions without the
compactness of operator semigroup. Then the null controllability is studied. The results are
obtained with the help of semigroup theory, ψ-Hilfer fractional calculus and the Banach fixed
point theorem.

Keywords: null controllability, ψ-Hilfer fractional, mild solution, Banach fixed point
theorem

Introduction
In this work, we are concerned with the following ψ-Hilfer fractional evolution equation with
nonlocal condition:
H α,β,ψ
D x(t) = Ax(t) + f (t, x(t)) , t ∈ (0, b] , (128)
I (1−α)(1−β),ψ [x(t)]t=0 = x0 + h(x), (129)
Where H D α,β,ψ is the ψ-Hilfer fractional derivative of order 21 < α < 1 and type 0 < β < 1 and
the state x(.) take values in a Hilbert space X and J = [0, b]. The operators A is the infinitesimal
generator of C0 -semigroup {S(t) = eAt }t≥0 . The functions f : J × X → X, h : C(J, X) → X are some
suitable functions satisfying certain conditions.

Main results
• To prove the main result, we need the following hypotheses:

(H1) There exist constants L > 0, such that

kf (t, x) − f (t, y)k ≤ Lkx − yk

for any x, y ∈ X, i = 1, 2 and t ∈ J. Thereover supkf (t, 0)k := M < ∞
t∈J

(H2) is a consequence of (H1)

113
CNMA’2022
Second National Conference on
Mathematics and its Applications

(H2) There exist non negative continuous functions h1 , h2 ∈ R such that

kf (t, x)k ≤ h1 (t) + h2 (t)kxk. x ∈ X, t ∈ J.

With h∗1 = suph1 (t), h∗2 = suph2 (t).

t∈J t∈J

(H3) The function g : C(J, X) → X is continuous, for any x, y ∈ C(J, X) there exist constant C > 0;
such that
kg(x) − g(y)k ≤ Ckx − yk
For computational convenience, we use the following notations:

(ψ(λ) − ψ(0))γ+v−1
Ψ (v, λ) =
Γ (γ + v)
Λ(v, λ) = LΓ (γ)Ψ (v, λ)
A2 kBkk(L0 )−1 k
!
1−γ α 1−γ+α
∆ = A1 C + A2 Qψ (T , 0)Λ(α1 , T ) + A1 CQψ (T , 0) + A2 Γ (α)Qψ (T , 0)Λ(α, T )
α
Theorem 47. Assume assumptions (H1)–(H3) are satisfied. If

∆ < 1,

then ψ−Hilfer fractional system (128)-(129) has a mild solution on [0, b].

References
[1] K. Mourad , Approximate controllability of fractional neutral stochastic evolution equations
in Hilbert spaces with fractional Brownian motion , Stoch. Anal . Appl., 36 (2018), 209-223.

[2] A. Pazy , Semigroups of Linear Operators and Applications to Partial Differential Equa-
tions , Applied Mathematical Sciences, Springer, New York, 1983.

[3] [Link], F. Jiao , Existence of mild solutions fo fractional neutral evolution equations , Comput.
Math. Appl., 59 (2010).

[4] J. P. Dauer, P. Balasubramaniam , Null controllability of semilinear integro-differential sys-

tems in Banach space, Appl. Math. Lett., 10(6), 1997, 117-123.

[5] J. Vanterler da C. Sous and E. Capelas da Oliveira, On the ψ-Hilfer fractional derivative,
Commun. Nonlinear Sci. Numer. Simul., 60 (2018), 72 -91.

114
CNMA’2022
Second National Conference on
Mathematics and its Applications

On The Quadratic Pencil Related To Three

Bounded Operators

Somia Kamouche 1 [Link]@[Link]

Mourad Ghiat 1 mourad.ghi24@[Link]
Hamza Guebbai 1 guebaihamza@[Link]
1 Laboratoire de Mathématiques Appliquées et de Modélisation, Université 8 Mai 1945 Guelma.

Abstract: This work is devoted to dealing with the quadratic pencil associated with three
linear bounded operators. We establish a new tool to transform our problem into a classical
case known as a "standard spectral problem." This transformation allows us to show that,
under appropriate assumptions, both properties U and L are proved. The occurrence of these
properties confirms the convergence of the spectral approximation method used in our study.

Keywords: Quadratic pencil, spectral approximation, property U, property L.

2010 Mathematics Subject Classification: 47A58, 47A05, 45L05, 15A18.

Introduction

Let (B , k·kB ) be a Banach space. We denote BL the space of all linear bounded operators defined
on B into itself, which is a Banach space equipped with the following norm

∀T ∈ BL(B ) : kT k = sup kT ukB .

kukB =1

In this paper, we focus on the quadratic pencil associated with three linear bounded operators
of the following form:
Q(λ) := λ2 I + λA + B.
Where A and B in BL(B ), I is the identity operator of BL(B ), and λ ∈ C is a spectral parameter.

This kind of problem appears in the resolution of differential equations, which arises in most
fields of applied mathematics such as quantum mechanics (the position of the electron around
the atomic), chemistry (chemical reactions and radioactivities), and physics (diffusion problems),
etc.

Main results

Let A and B in BL(B ) and I is the identity operator.

115
CNMA’2022
Second National Conference on
Mathematics and its Applications

We define the quadratic resolvent set, denoted by re(I, A, B):

n o
re(I, A, B) = λ ∈ C : λ2 I + λA + B is invertible and bounded .

The quadratic spectrum set, denoted by sp(I, A, B):

sp(I, A, B) = C\R(I, A, B).

Let (An )n∈N and (Bn )n∈N be two sequences of linear bounded operators.

Our major goal is to demonstrate that both properties U and L are attainable under appropriate
hypotheses on (An )n∈N and (Bn )n∈N , that is

Property U: if λn ∈ sp(I, An , Bn ) and λn → λ then λ ∈ sp(I, A, B).

Property L: if λ is a quadratic eigenvalue of finite type isolated in sp(IA, B), there is a sequence
λn ∈ sp(I, An , Bn , ) such that λn → λ.

References
[1] E. Brian Davies, Linear oprerators and their spectra, Cambridge University Press, 2007.

[2] M. Ahues, A. Largillier, B. Limaye, Spectral Computations For Bounded Operators, CRC
Press, 2001.

[3] S. Kamouche, H. Guebbai, M. Ghiat, S. Segni, Generalized quadratic spectrum approxima-

tion in bounded and unbounded cases, Probl. Anal. Issues Anal. 10 (28)(3) (2021), 53–70
.

[4] M. di Bernardo, C. J. Budd, A. R. Champneys and P. Kowalczyk, Piecewise-Smooth Dy-

namical Systems: Theory and Applications, Appl. Math. Sci. Series 163, Springer-Verlag,
London, 2008.

[5] A. Khellaf, H. Guebbai, A Note On Generalized Spectrum Approximation, Lobachevskii

Journal of Mathematics. 39(9) (2018), 1388–1395.

116
CNMA’2022
Second National Conference on
Mathematics and its Applications

On the Resolution of Resonance and

Nonresonance Systems Involving the
Fractional Laplacian

Dob Sara 1 dobsara@[Link]

Hakim Lakhal 1
Messaoud Maouni1
1 Laboratory of Applied Mathematics and History and Didactics of Mathematics (LAMAHIS), Department of
Mathematics, University 20 August 1955, Skikda, Algeria

Abstract: This paper is devoted to study of the existence of a weak solutions for the nonlinear
fractional elliptic systems with Dirichlet boundary conditions. We use the Leray-Schauder
degree to solve a resonance and non-resonance systems.

Keywords: Topological degree, elliptic systems, fixed point, homotopy.

2010 Mathematics Subject Classification: 35J67, 37C25, 55P05.

Introduction
This work is devoted to the study of the existence of solutions to nonlocal equations involving
the fractional Laplacian, we give an application of the Leray-Schauder degree theorem to prove
the existence of a weak solution to the system



(−∆)s u(x) = f (x, u(x), v(x)) + f1 (x) in Ω,

(−∆)s v(x) = g(x, u(x), v(x)) + f2 (x) in Ω, (130)




u = v = 0 on Rn \ Ω,


with s ∈ (0, 1) on a bounded domain Ω ⊂ Rn , n ≥ 2, f , g are continuous functions defined in

Ω × R × R and h = (f1 , f2 ) ∈ (L2 (Ω))2 .
As far as we know, this result is new and represent a fractional version of a classical theorem
obtained working with Laplacian equations. Let us assume now that the nonlinear functions f , g
are of the form: 
f (x, u, v) = au + f0 (x, v)



g(x, u, v) = bv + g0 (x, u)


where a, b are real positive constants.

We recall that the fractional Laplacian (−∆)s is a defined as (see [2])
Z
s ϕ(x) − ϕ(y)
(−∆) ϕ(x) = C(n, s)P .V . n+2s
dy,
Rn |x − y|

117
CNMA’2022
Second National Conference on
Mathematics and its Applications

for ϕ ∈ C0∞ (Rn ), where s ∈ (0, 1), P .V . denotes the integral in the sense of the principal value,
C(n, s) is a positive constant of normalization defined as

2s Γ ( n+2
2 )
C(n, s) = s2 ,
Γ ( 12 )Γ (1 − s)

and its spectrum in L2 (Ω) is formed by the sequence (λk )k ∈ R∗ such that |λk | → +∞ (in the rest
of the paper we note the spectrum with sp). Throughout this paper, we denote by λ1 the first
eigenvalue of (−∆s ) and ϕ1 is the normalized eigenfunction associated a λ1 .
The purpose of the present paper is to extend the results of [3, 5] to the system (130) under the
following conditions on the functions f0 and g0 :
 − +


 β1 , β1 ∈ L2 (Ω)

and





 − +
β1 (x) ≤ f0 (x, t) ≤ β1 (x) a.e. Ω



(131)



 where
lim f0 (., t) = β1− (.) a.e. Ω






 t→−∞
 lim f0 (., t) = β1+ (.) a.e. Ω,


t→+∞

and  − +


 β2 , β2 ∈ L2 (Ω)

and





 − +
β2 (x) ≤ g0 (x, t) ≤ β2 (x) a.e. Ω



(132)



 where
lim g0 (., t) = β2− (.) a.e. Ω






 t→−∞
 lim g0 (., t) = β2+ (.) a.e. Ω.


t→+∞

Main results
Let’s now give the main result of this paper.

Theorem 48. Assume that (131) and (132) are fulfilled. Then (130) has at least one solution (u, v) ∈ U .

References
[1] C. Bucur and A. L. Karakhanyan, Potential theoretic approach to Schauder estimates for the
fractional Laplacian, Proceedings of the American Mathematical Society, 2017;145(2):637-651.

[2] M. D’Elia and M. Gunzburger, The fractional Laplacian operator on bounded domains as a
special case of the nonlocal diffusion operator, Computers & Mathematics with Applications,
2013;66(7):1245-1260.

[3] S. Dipierroa and A. Pinamonti, A geometric inequality and a symmetry result for elliptic
systems involving the fractional Laplacian, Journal of Differential Equations, 2013;255(1):85-
119.

[4] H. Lakehal, B. Khodja and W. Gharbi, Existence results of nontrivial solutions for a semi
linear elliptic system at resonance, Journal of Advanced Research in Dynamical and Control
Systems, 2013;5(3):1-12.

118
CNMA’2022
Second National Conference on
Mathematics and its Applications

[5] A. Moussaoui and B. Khodja, Existence results for a class of semilinear elliptic systems, Jour-
nal of Partial Differential Equations, 2009;22(2):111-126.

119
CNMA’2022
Second National Conference on
Mathematics and its Applications

Fractional Partial Hyperbolic Differential

Inclusions with Finite State-Dependent
Delay

Mohamed Helal 1 helalmohamed@[Link]

1 Science and Technology Faculty. Mustapha Stambouli University of Mascara, B.P. 763, 29000, Mascara, Algeria.
Laboratory of Mathematics, Djillali Liabes University of Sidi Bel-Abbès, B.P. 89, 22000, Sidi Bel-Abbès, Algeria

Abstract: In this paper we investigate the existence of solutions of initial value problem for
partial hyperbolic differential inclusions of fractional order involving Caputo’s fractional
derivative with finite state-dependent delay when the right hand side is convex valued by
using a multi-valued version of nonlinear alternative of Leray-Schauder type.

Keywords: Partial functional differential inclusion, fractional order, solution, left-sided

mixed Riemann-Liouville integral, Caputo fractional-order derivative, finite state-dependent
delay, fixed point.
2010 Mathematics Subject Classification: Primary 26A33, 34K30.

Introduction
This work concerns the existence of solutions to fractional order initial value problems (IV P for
short), for the system

(c D0r u)(t, x) ∈ F(t, x, u(ρ1 (t,x,u(t,x) ),ρ2 (t,x,u(t,x) )) ), if (t, x) ∈ J,

˜
u(t, x) = φ(t, x), if (t, x) ∈ J,
u(t, 0) = ϕ(t), u(0, x) = ψ(x), (t, x) ∈ J,

where ϕ(0) = ψ(0), J := [0, a] × [0, b], a, b, α, β > 0, J˜ := [−α, a] × [−β, b]\[0, a] × [0, b], c D0r is the stan-
dard Caputo’s fractional derivative of order r = (r1 , r2 ) ∈ (0, 1]×(0, 1], F : J ×C([−α, 0]×[−β, 0], Rn ) →
P (Rn ), is a compact valued multivalued maps, P is a family of all subsets of Rn , ρ1 : J × C →
[−α, a], ρ2 : J × C → [−β, b] are given functions, φ ∈ C([−α, 0] × [−β, 0], Rn ) is a given continuous
function with φ(t, 0) = ϕ(t), φ(0, x) = ψ(x) for each (t, x) ∈ J, ϕ : [0, a] → Rn , ψ : [0, b] → Rn are
given absolutely continuous functions.
We denote by u(t,x) the element of C([−α, 0] × [−β, 0], Rn ) defined by

u(t,x) (s, τ) = u(t + s, x + τ); (s, τ) ∈ [−α, 0] × [−β, 0],

here u(t,x) (·, ·) represents the history of the state u.

Preliminaries

120
CNMA’2022
Second National Conference on
Mathematics and its Applications

In this section, we introduce notations, definitions, and preliminary facts which are used through-
out this paper.
By L1 (J, Rn ) we denote the space of Lebesgue-integrable functions u : J → Rn with the norm
Z aZ b
kukL1 = ku(t, x)kdxdt,
0 0

where k · k denotes a suitable complete norm on Rn .

Definition 17. r = (r1 , r2 ) ∈ (0, ∞) × (0, ∞), θ = (0, 0) and u ∈ L1 (J, Rn ). The left-sided mixed Riemann-
Liouville integral of order r of u is defined by
Z tZ x
1
(Iθr u)(t, x) = (t − s)r1 −1 (x − τ)r2 −1 u(s, τ)dτds.
Γ (r1 )Γ (r2 ) 0 0

Definition 18. [5] Let r ∈ (0, 1] × (0, 1] and u ∈ L1 (J, Rn ). The mixed fractional Riemann-Liouville
derivative of order r of u is defined by the expression
2 1−r
Dθr u(t, x) = (Dtx Iθ u)(t, x)

and the Caputo fractional-order derivative of order r of u is defined by the expression

∂2
(c D0r u)(t, x) = (Iθ1−r u)(t, x).
∂t∂x

In the sequel we will make use of the following generalization of Gronwall’s lemma for two inde-
pendent variables and singular kernel.

Lemma 19. [4] Let υ : J → [0, ∞) be a real function and ω(·, ·) be a nonnegative, locally integrable
function on J. If there are constants c > 0 and 0 < r1 , r2 < 1 such that
Z tZ x
υ(s, τ)
υ(t, x) ≤ ω(t, x) + c dτds,
0 0 (t − s)r1 (x − τ)r2

then there exists a constant δ = δ(r1 , r2 ) such that

Z tZ x
ω(s, τ)
υ(t, x) ≤ ω(t, x) + δc dτds,
0 0 (t − s)r1 (x − τ)r2

for every (t, x) ∈ J.

Theorem 49. (Nonlinear alternative of Leray-Schauder type) [2] Let X be a Banach space and C a
nonempty convex subset of X. Let U a nonempty open subset of C with 0 ∈ U and T : U → P (C) an
upper semicontinuous and compact multivalued operator. Then either

(a) T has a fixed points. Or

(b) There exist u ∈ ∂U and λ ∈ [0, 1] with u ∈ λT (u).

References
[1] J. P. Aubin and A. Cellina, Differential Inclusions, Springer-Verlag, Berlin-Heidelberg, New-
York, 1984.

121
CNMA’2022
Second National Conference on
Mathematics and its Applications

[2] A. Granas and J. Dugundji, Fixed Point Theory, Springer-Verlag, New York 2003.

[3] M. Helal, Fractional Partial Hyperbolic Differential Inclusions with State-Dependent Delay,
Fractional Calculus and Applications, 10(1) (2019), 179-196.

[4] D. Henry, Geometric Theory of Semilinear Parabolic Partial Differential Equations, Springer-
Verlag, Berlin-New York, 1989.

[5] A. N. Vityuk and A. V. Golushkov, Existence of solutions of systems of partial differential

equations of fractional order, Nonlinear Oscil. 7 (3) (2004), 318-325.

122
CNMA’2022
Second National Conference on
Mathematics and its Applications

Using projection methods to solve linear

Fredholm intgro-differential equation

Boutheina Tair 1 tairboutheina2@[Link]

Hamza Guebbai1 guebaihamza@[Link]
1 Laboratoire de Mathématiques Appliquées et de Modélisation, Faculté de Mathématiques et de l’Informatique
et des Sciences de la Matière, Université 8 Mai 1945 Guelma

Abstract: In this work, we construct numerical approximations solutions to solve a linear

Fredholm integro-differential equation. But first, we start by proving the solution’s existence
and uniqueness. So, we reformulate our equation to a system of integral equations. We explain
the general framework of the projection method and we construct two projection solutions
by using the concept of colloaction and Kantorovich methods. We give theorems to show the
convergence of the constructed solutions to the exact solution. Finally, to observe the error
behaviour of the two methods, we give numerical examples.

Keywords: Fredholm integral equations, Integro-differential, System of integral equa-

tions, Projection methods.
2010 Mathematics Subject Classification: 45B05, 47G20, 45F05, 65R10.

Introduction
The first integro-differential equation is considered as the development of mathematical mod-
elling, it is the Volterra equation to model the problem of competing species [1, 2]. The integro-
differential equations are used to model several phenomena such as the system of leaking aquifers,
the processes of coagulation, the electric circuit LRC and and earthquake [5, 6]. Due to the great
importance of these equations, scientists discovering several forms and types which makes the an-
alytical resolution difficult. So, several numerical methods are constructed to find an approximate
solution of these equations.

In this paper, we focus on the numerical solution of the following Fredholm linear integro-
differential equation
Z b Z b
∀x ∈ [a, b], λu(x) = K1 (x, t)u(t) dt + K2 (x, t)u 0 (t) dt + f (x), (133)
a a

where, λ is a real or complex paramater, f is a given function in the Banach space C 1 [a, b] and the
kernels Kp for p = 1, 2 satisfied
∂Kp
(x, t) ∈ C 0 ([a, b]2 , R).
∂x

123
CNMA’2022
Second National Conference on
Mathematics and its Applications

Our main goal is to find a two-sequences approximation, an approximation of u and an ap-

proach of the derivative u 0 . Before applying our numerical treatment, we reformulate our equa-
tion to an equivalent system of integral equations. We construct a sufficient condition to demon-
strate the solution’s existence and uniqueness of the system. To illustrate our computation of
projection methods, we built two sequences {Pn u}n≥1 and {Pn u 0 }n≥1 based on piecewise linear
functions. Then, we construct two numerical solutions by using the concept of collocation and
Kantorovich methods. This two approximate solutions will lead to the resolution of block matrix
linear system . We give some theorems to prove the convergence of our approximate solutions to
the exact solution. At the end, we present the numerical test examples to compare the error be-
haviour between collocation and Kantorovich method, to establish the efficiency of Kantorovich
method and the simplicity of both.

Main results
We have already mentioned that the kernels are derivable with respect to [Link], we can derive
both sides of our equation. Then, we put u (p−1) = up for p = 1, 2 we get the following integral
equations system:
 Zb Zb

K2 (x, t)u 0 (t) dt + f1 (x),




 λu1 = K 1 (x, t)u 1 (t) dt +


 a a
(134)




 Zb Zb
 ∂K1 ∂K2
(x, t)u 0 (t) dt + f2 (x).

 λu2 = (x, t)u1 (t) dt +



a ∂x a ∂x
Let note X = C 0 [a, b] × C 0 [a, b] which is equipped with the next norm

∀(v1 , v2 ) ∈ X, ||(v1 , v2 )||X = ||v1 ||C 0 [a,b] + ||v2 ||C 0 [a,b] .

Let define AT the block operator as
AT : X −→ X
(u1 , u2 ) 7−→ AT (u1 , u2 ) = (T11 u1 + T12 u2 , T21 u1 + T22 u2 ). (135)
where, {Tpq }1≤p,q≤2 are bounded linear operator defined as for all v ∈ C 0 [a, b]
Zb
T1q v(x) = Kq (x, t)v(t) dt, q = 1, 2,
a
Z b ∂Kq
T2q v(x) = (x, t)v(t) dt, q = 1, 2.
a ∂x
We define the norm of the block operator matrix AT in the following way:
2
X
||AT || = max ||Tqp ||,
1≤q≤2
p=1

where the norm of linear and bounded operator T defined in C 0 [a, b] in itself, is given as
||T || = sup ||T v||C 0 [a,b] .
||v||C 0 [a,b]≤1

Finally, (134) yields in the next equivalent version

(λI2 − AT )(u1 , u2 ) = (f1 , f2 ), (136)
where, f F = (f , f 0 ) and I2 is the identity operator of the Banach space X.

124
CNMA’2022
Second National Conference on
Mathematics and its Applications

Now, we give the following theorem to prove the existence and uniqueness of the solution of
(136).
Theorem 50. If |λ| > ||AT || the system (136) has a unique solution in X.

Projection Concepts

After ensuring the existence and uniqueness of the solution, now we can apply our numerical
process to construct an approximate solution of (133). But in this part, we begin to define the
projection concept in order to illustrate the general framework of collocation and Kantorovich
methods [3, 4].
First, we give ∆n for all n ≥ 1 the uniform discritization of [a, b]:

∆n = {a = x0 < x1 < x2 . . . xn−2 < xn−1 < xn = b, h = xj+1 − xj , 0 ≤ j ≤ n.}

We note by Xn the sequence of finite dimension subspace of the Banach space X. The principle of
two methods is based on the application of the of the sequence projection operator {Pn }n≥1 defined
by

Pn : X −→ Xn
(u1 , u2 ) 7−→ Pn (u1 , u2 ) = (Pn,1 u1 , Pn,2 u2 ), (137)

where, {Pn,q }n≥1 for q = 1, 2 is a linear combination of haat functions, which is written as
n
X
∀n ≥ 1, ∀v ∈ C 0 [a, b], Pn,q v(x) = αq,i ei (x), q = 1, 2,
i=0

where, {ei }ni=0 are given as

|x − xi |

 1+ h , if x ∈ [xi−1 , xi+1 ],



ei (x) = 

0, else,



and {αi,q } are unknowns to be determinate. We notice that {Pn,q }n≥1 verifies the following interpo-
lation condition
∀n ≥ 1, ∀uq ∈ C 0 [a, b], Pn,q uq (xi ) = uq (xi ).

Now, we move to the collocation method. The idea of this method is based to apply the projec-
C C
tion sequence (137) on the equation (136), we get a new approximation problem: Find (u1,n , u2,n )
solution of the following equation
C C C C
λ(u1,n , u2,n ) = Pn AT (u1,n , u2,n ) + Pn (f1 , f2 ). (138)

Contrairy, the concept of Kantorovich solution is based to apply the sequence Pn (137) only on AT
on the eqaution (136). So, our approximation problem is presented as: Search for the approximate
K K
solution (u1,n , u2,n ) of the following approximate equation 4
K K
λ(u1,n , u2,n ) = Pn AT + (f1 , f2 ), (139)

and we compute an approximation solution has the next form

" #
K K 1 C C
(u1,n , u2,n ) = (v1,n , v2,n ) + (f1 , f2 ) ,
λ

125
CNMA’2022
Second National Conference on
Mathematics and its Applications

C C
where, (v1,n , v2,n ) is a collocation solution of
C C C C
λ(v1,n , v2,n ) = Pn AT (v1,n , v2,n ) + Pn (f1 , f2 ). (140)
Convergence Analysis

In this part, we follow some steps to show that our numerical solutions converge to exact so-
lution. The first, step is to prove that our projection operator {Pn }n≥1 is pointwise convergent to
the identity operator I2 , which is verified in the next theorem
Theorem 51. Let {Pn }n≥1 be a sequence of projection operator defined by (137), then
||(I2 − Pn )(u1 , u2 )||X → 0, when n → +∞.

The second step is to demonstrate that the inverse of (λI2 − Pn ) exists and bounded, so we need to
present the next theorem
Theorem 52. Let Pn AT be a block projection operator, such that AT is defined by (136) then (λI2 −
Pn AT )−1 exists and bounded.

The finale step, is in the next theorem

C C K K
Theorem 53. Let (u1,n , u2,n ) the collocation approximation verifies (138) and (u1,n , u2,n ) the Kan-
torovioch solution checks the equation (139), then
C C
||(u1,n , u2,n ) − (u1 , u2 )||X → 0, when n → +∞,
and
K K
||(u1,n , u2,n ) − (u1 , u2 )||X → 0, when n → +∞.

Numerical test

To illustrate the efficiency of our proposed methods and to prove that the Kantrovich method
is better then the collocation method, we give the following numerical test
Z1 Z1
u(t) u 0 (t)
∀x ∈ [0, 1], λu(x) = x t
dt + 2t
dt + f (x), (141)
0 e +e 0 1+x+e

where, λ = 2, u(x) = ex and

arctan( √ 1 )
e1
x 1 x x x+1
f (x) = log(e + 1) − log(e + e ) + 2e − arctan( √ )− √ .
x+1 x+1

Table 2: The error between the exact and approximation solution of equation (141)

n Collocation Kantorovich
10 6.6852e-04 1.0445e-05
100 7.1682e-06 1.1061e-07

Conclusion
Our work focuses on the numerical study of the linear integro-differential equation. We apply to
projection methods: Collocation and Kantorovich to serach for a better approximation in C 1 [a, b].
To examine the error behaviour of our methods, we give a numerical test.

126
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] V. Lakshmikantham , Theory of Integro-Differential Equations, CRC Press, 1995.

[2] V. Volterra, Theory of Functionals and of Integral and Integro-Differential Equations, Dover
Publications, 2000.

[3] K. Atkinson, W. Han , Theoretical Numerical Analysis: A Functional Analysis Framework,

Springer, New York, 2001.

[4] M. T. Nair, , Linear Operator Equations: Approximation and Regularization, Indian Institute of
Technology Madras, India, 2001. New York, 2001.

[5] B. Tair, H. Guebbai, S. Segni, M. Ghiat, An approximation solution of linear Fredholm integro-
differential equation using Collocation and Kantorovich methods, [Link]. Math. Comp. 2021.

[6] R. B. Tair, H. Guebbai, S. Segni, M. Ghiat, Solving linear Fredholm integro-differential equation
by Nyström method, [Link]. Math. Comp. Mech. Series 20, 3, 2021.

127
CNMA’2022
Second National Conference on
Mathematics and its Applications

Solving non-linear Volterra

integro-differential equation of second
order

Sami Segni 1 segnianis@[Link]

1 Laboratoire de Mathématiques Appliquées et de Modélisation, Faculté de Mathématiques et de l’Informatique
et des Sciences de la Matière, Université 8 Mai 1945 Guelma, B.P. 401, 24000 Guelma, Algeria

Abstract: Our objective is the analytical and numerical study of the nonlinear Volterra differ-
ential equation. Therefore, we construct a change of variable to create a sufficient condition
that shows the existence and uniqueness of the solution on the one hand. On the other hand,
this condition makes the analytical and numerical study consistent. Our numerical framework
is based on the Nyström method which allows us to reformulate our equation into a nonlinear
algebraic system such that our approximation is the unique solution of this system. We present
important theorems that play a necessary role in proving our result and demonstrating the
convergence of the approximated solution.

Keywords: Volterra integral equation, non-linear integral equation, integro-differential

equation
2010 Mathematics Subject Classification: Primary 45D05, 45G10, 47G20.

Introduction
The integro-differential equation was a quantum leap in the field of mathematical modelling [2].
Covide 19, cancer treatment by radiotherapy, earthquake model, telegraphy model, nuclear reac-
tor and the list goes on for all these problems where they can be expressed as linear or non-linear
Volterra integro-differential equations. The intrusion of these equations in many fields has made
them take different types and forms depending on the scientific need. What interests us in this
article is the non-linear Volterra integral-differential equation with regular kernel:
Zx
2
∀x ∈ [a, b], ∀u ∈ C [a, b], u(x) = (t − x) K(x, t, u(t), u 0 (t), u 00 (t)) dt + f (t), (142)
a

where f ∈ C 2 [a, b] is a given function, the kernel K satisfies the assumptions we will present next.
In this work, we follow two essential steps: Show the existence and uniqueness of the solution of
(142) then construct an approximate solution and show its convergence.

The Analytical Study

This section is devoted to the analytical study of the integro-differential equation (142). Then,
we suppose that the kernel K and the datum f satisfy the following hypotheses:

128
CNMA’2022
Second National Conference on
Mathematics and its Applications
2
 
(1) ∂∂t 2k ∈ C 0 [a, b]2 × R ,

(2) ∃M ∈ R, ∀t, s ∈ [a, b], ∀x, y, z ∈ R,
     2 
max k(t, s, x, y, z) ,  ∂k∂t
(t, s, x, y, z)  ,  ∂ 2k (t, s, x, y, z) ≤ M
∂t

(3) ∃a , b , c , a , b , c , a , b , c ∈ R, ∀x, y, z, x, y, z ∈ R, ∀t, s ∈ [a, b],
1 1 1 2 2 2 3 3 3
(H1)  
k(t, s, x, y, z) − k(t, s, x, y, z) ≤ a1 |x − x| + b1 |y − y| + c1 |z − z|,

| ∂k
∂t
(t, s, x, y, z) − ∂k
∂t
(t, s, x, y, z)| ≤ a2 |x − x| + b2 |y − y| + c2 |z − z|,
2 2
| ∂∂t 2k (t, s, x, y, z) − ∂∂t 2k (t, s, x, y, z)| ≤ a3 |x − x| + b3 |y − y| + c3 |z − z|,

(4) c1 < 1

This assumption allows the equation (142) to be derived twice. This allows us to pose u 00 = v by
the following change of variable:
Zx
u(x) = Av(x) = (x − s)v(s) ds + (t − a)u 0 (a) + u(a),
Zax
u 0 (x) = Bv(x) = v(s) ds + u 0 (a).
a

We obtain a new equation equivalent to (142)

Zx
∂K
∀x ∈ [a, b], v(x) = K(x, x, Av(x), Bv(x), v(x)) + 2 (x, t, Av(t), Bv(t), v(t)) dt
a ∂x
Zx
∂2 K
+ (x − t) 2 (x, t, Av(t), Bv(t), v(t)) dt + f 00 (x). (143)
a ∂ x
Theorem 54. Under the hypothesis (H1 ) the equation (??) has a unique solution.

Numerical Study: Nyström method

For all n ≥ 1, we define the uniform descritization of the interval [a, b] by: xi = a + ih such that
b−a
h= . We apply the Nyström [1] method on the equation (143) and we choose the collocation
n
points x = xi , we obtain the following non-linear algebraic system
i−1
P ∂K
∀x ∈ [a, b], vn (xi ) = K(xi , xi , Av(xi ), Bv(xi ), v(xi )) + 2 (xi , xj , Av(xj ), Bv(xj ), v(xj )) dt
j=0 ∂x
i−1 ∂2 K
(xi , xj , Av(xj ), Bv(xj ), v(xj )) dt + f 00 (xi ),
P
+ (xi − xj ) 0 ≤ i ≤ n. (144)
j=0 ∂2 x

Theorem 55. If v is a solution of (143) and vn is a solution of (144), then vn converges to v when
n → +∞.

References
[1] V. Volterra, Theory of Functionals and of Integral and Integro-Differential Equations, Dover
Publications, 2000.
[2] K. Atkinson, W. Han , Theoretical Numerical Analysis: A Functional Analysis Framework,
Springer, New York, 2001.

129
CNMA’2022
Second National Conference on
Mathematics and its Applications

Existence results for fractional Kirchhoff

problem

Yasmina NASRI 1 y_nasri@[Link]

1 Department of mathematics ,University of Tlemcen ,BP 119 Tlemcen 13000, Algeria

Abstract: In this work we consider a fractional Kirchhoff problem with changing sign data.
The main goal of this study is to get the existence of nonnegative solutions.

Keywords: Fractional Laplacian, Kirchhoff type problem, variational method.

2010 Mathematics Subject Classification: AMS. 26A33, 35A15, 35J25.

Introduction
This paper is devoted to study the existence of positive solutions of the following problem

! |u(x) − u(y)|2
 !
M RN ×RN dxdy (−∆)s u = λf (x) + |u|p−2 u in Ω,



(Ps )  |x − y| N +2s

in RN \Ω,

u = 0

where Ω is a smooth bounded domain in RN with N > 2s, s ∈ (0, 1), 1 < p < +∞, λ is a positive real
parameter, f is a changing sign data function that will be specified later.
The fractional Laplacian operator (−∆)s is defined, up to a normalization factor as follows
Z
s (u(x) − u(y))
(−∆) u(x) = 2 lim+ N +2s
dy, x ∈ RN ,
ε→0 RN \Bε (x) |x − y|

for more details about the operator see [3] .

When M is not identically constant, then for the local case s = 1, the problem becomes
 R 
2
−M Ω |∇u| dx ∆u = g (x, u) in Ω,


(P1 ) 
u = 0
 on ∂Ω,

the problem (P1 ) is well known as Kirchhoff type problem, this latter is extensively studied. (P1 )
is the stationary problem of the kirchhoff hyperbolic equation

∂2 u
Z !
2
−M |∇u (x, t)| dx ∆u = g (x, u) ,
∂t 2 Ω

it describes the vibration of stretched elastic string. In 1883, Kirchhoff proposed the study of the
problem for M (t) = αt + β where α; β are positive constants and Ω = (0, L), as an extention of the
classical d’Alembert’s wave equation.

130
CNMA’2022
Second National Conference on
Mathematics and its Applications

Alves and Corrêa [2] studied (P1 ) when M is a positive non-increasing function and g satisfies
Ambroseti-Rabinowitz condition. Azzouz and Bensedik [1] considered the local problem, they
establish the existence resulys using Galerkin approch and sub-super solution method.
For the case where s ∈ (0, 1) ,there is many interesting works we cite for example [4], [5] and the
references therein.
Motivated by the previous works, we prove existence and nonexistence results. It turns out that
the resolution of (Ps ) is closely linked to the positivity of the function M and some hypothesis on
f that assure us the existence of positive solutions.

Main results
Before we state our results we list some assumptions:
(M1) M : R+ → R+ is a continous function such that ∃m0 > 0; M (t) ≥ m0 ∀t ∈ R+ .
 
(F1) f ∈ C 1 Ω .
(F2) There exists a ε > 0 and δ > 0 such that the problem

s + −
(−∆) u = θ(f (x) − (1 + ε)f (x)) in Ω,


in RN \Ω.

u = 0


has a non-negative solution for any θ ∈ (0, δ] Our main result

  Assume that (M1) , (F1)-(F2) hold. M is non increasing function. Suppose H (t) :=
Theorem
tM t 2 is increasing on R. Then there exists a real Λ > 0 such that problem (Ps ) :
- has a positive solution for all λ ∈ (0, Λ) .
- has no positive solution for all λ > Λ.

The proof of result is based on variational methods.

References
[1] N. Azzouz, A. Bensedik, Existence results for an elliptic equation of Kirchhoff-type with
changing-sign data, Funkcial. Ekva., 55 (2012), 55-66.

[2] C. O. Alves, F. J. S. A Corrêa, T. F. Ma, Positive solutions for quasilinear elliptic equation of
Kirchhoff type, Comput. Math. Appl, 49 (2005), 85-93.

[3] Di Nezza E, Palatucci G, Valdinoci E. Hitchhiker’s guide to the fractional Sobolev spaces. Bull
Sci Math. 2012;136(5), 521–573

[4] P. Pucci, E, Valdinoci, Progress in Nonlinear Kirchhoff Problems. Nonlinear Analysis 186
(2019) ,214–228.

[5] P. Pucci, M. Xiang, B. Zhang, Existence and multiplicity of entire solutions for fractional
p-Kirchhoff equations, Adv. Nonlinear Anal. 5 (2016), 27-55.

131
CNMA’2022
Second National Conference on
Mathematics and its Applications

QUASILINEAR PARABOLIC PROBLEM WITH

VARIABLE EXPONENT AND NONLOCAL
BOUNDARY CONDITIONS

Fairouz Souilah 1 [Link]@[Link]

1 Departement of Mathematics. Laboratory of Applied Mathematics, History and Didactics of Maths
"LAMAHIS". University 20th August 1955, Skikda, Algeria. University 20th August 1955, Skikda, Algeria

Abstract: In this work we study the existence for quasilinear parabolic problem with variable
exponent and with nonlocal boundary conditions and L1 data. The main contribution of our
work is to prove the existence of a renormalized solution. The results of the problem discussed
can be applied to a variety of different fields in applied mathematics for example in elastic
mechanics, image processing and electro-rheological fluid dynamics, etc..

Keywords: Quasilinear parabolic problem; variable exponent; renormalized solution;

nonlocal boundary conditions.
2010 Mathematics Subject Classification: 35K59.

Introduction
The study of problems with variable exponent is a new and interesting topic which raises many
mathematical difficulties . One of our motivations for studying (145) comes from applications to
electrorheological fluids as an important class of non-Newtonian fluids (sometimes referred to as
smart fluids). Other important applications are related to image processing and elasticity .
As a physical motivation, problem (145) arises from the study of quasi-static thermoelasticity.
The main difficulty of this problem is related to the presence of both quasilinear term in (145)
and nonlocal boundary condition . Literatures to this type of problem are very limited. Let Ω
be a bounded domain of RN (N ≥ 2) with lipshitz boundary ∂Ω and Q = Ω × ]0, T [ for any fixed
T is a positive real number. Our aim is to prove the existence of a renormalized solution for a
quasilinear parabolic problem with variable exponent and with nonlocal boundary conditions


ut − ∆p(x) u + |u|p(x)−2 u = f (x, t, u) in Q = Ω×]0, T [,





 R
 u(x, t) = Ω k(x, y)u(y, t)dy on Γ = ∂Ω×]0, T [,



 (145)
u(t = 0) = u0 in Ω,










where p : Ω −→ [1, +∞) be a continuous real-valued function and let p− = minx∈Ω p(x) and p+ =
maxx∈Ω p(x) with 1 < p− ≤ p+ < N . The function f : Q × R → R be a Carathéodory function .
Finally the data f (x, t, u) and u0 is in L1 .
As a physical motivation, problem (145) arises from the study of quasi-static thermoelasticity.

132
CNMA’2022
Second National Conference on
Mathematics and its Applications

The main difficulty of this problem is related to the presence of both quasilinear term in (145)
and nonlocal boundary condition of problem (145). Literatures to this type of problem are very
limited. We only found in which the authors study a quasilinear parabolic equation with nonlocal
boundary conditions different from nonlocal boundary condition of problem (145).
In recent years, there are a lot of interest in the study of various mathematical problems with
variable exponent , the problems with variable exponent are interesting in applications and raise
many difficult mathematical problems, some of the models leading to these problems of this
type are the models of motion of electrorheological fluids, the mathematical models of stationary
thermo-rheological viscous fows of non-Newtonian fluids and in the mathematical description of
the processes filtration of an ideal barotropic gas through a porousmedium.
In the classical case (p(.) = 2 or p(.) = p (a constant)), to recall that the notion of renormalized
solutions was introduced by Di Perna and Lions [3] in their study of the Boltzmann equation.
For the quasilinear parabolic problem with variable exponent and nonlocal boundary conditions
of (145) the existence of renormalized solution, this result can be seen as a generalization of the
result in classical sobolev space obtained by S. Fairouz and all in [3] in the case where b(u) = u
and u0 ∈ L1 (Ω).
Let us briefly summarize the contents of this article: In Section 2, we give some preliminaries and
basic assumptions. Section 3, we give the definition of a renormalized solution of (145), and we
establish (Theorem (56)) the existence of such a solution.

Main results
The Existence of result

This section is devoted to establish the existence of a renormalized solution .

Theorem 56. Under assumptions there exists at least a renormalized solution u of Problem (145).

References
[1] Y. Akdim, J. Bennouna, M. Mekkour, H. Redwane, Existence of a Renormalised Solutions for a Class
of Nonlinear Degenerated Parabolic Problems with L1 Data, J. Part. Diff. Eq., Vol. 26, No. 1, March
2013, pp. 76-98.

[2] E. Azroula, H. Redwane, M. Rhoudaf, Existence of solutions for nonlinear parabolic systems via weak
convergence of truncations , Electronic Journal of Differential Equations, Vol. 2010(2010), No.
68, pp. 1-18.

[3] S. Fairouz, M. Messaoud, S. Kamel, Study of quasilinear parabolic problems with data L1 . Submitted.

[4] J.-L. Lions, Quelques méthodes de résolution des problémes aux limites non linéaires . Dunod
et Gauthier-Villars, 1969.

[5] S. Ouaro and A. Ouédraogo,Nonlinear parabolic equation with variable exponent and L1 -data. Elec-
tronic Journal of Differential Equations, Vol. 2017 (2017), No. 32, pp. 1-32.

133
CNMA’2022
Second National Conference on
Mathematics and its Applications

Galerkin method for the higher dimension

Boussinesq equation non linear with
integral condition

Draifia Ala Eddine 1 2 draifia1991@[Link]

1 Institute of Sciences, Department of mathematics and Informatics, University Center Aflou, Laghouat, Algeria.
2 Department of Mathematics and Computer Science, Larbi Tébessi University, 12002 Tébessa, Algeria

Abstract: This paper deals with the solvability of a higher dimension mixed non local problem
for a Boussinesq equation non linear. Galerkin’s method was the main used tool for proving
the solvability of the given non local problem.

Keywords: Boussinesq equation; non local condition; Galerkin’s Method.

2010 Mathematics Subject Classification: 35L20, 58J45.

Introduction
By applying mathematical modeling to various phenomena of physics, biology and ecology there
often arise problems with non-classical boundary conditions, which connect the values of the
unknown function on the boundary and inside of the given domain. Some times the physical
phenomena are modeled by non classical boundary value problems which involve a boundary
condition as an integral condition over the spatial domain of a function of the desired solution.
The nonlocal boundary condition arises mainly when the data on the boundary cannot be mea-
sured directly, but their average values are known. In the very recent years, nonlocal problems,
particularly those with integral constraints have received great attention. The physical signifi-
cance of nonlocal conditions such as a mean, total mass, moments, etc, has served as a funda-
mental cause for the considerably increasing interest to this kind of boundary value problems.
Nonlocal problems are generally encountered in chemical engineering, heat transmission, plasma
physics, heat transmission, thermoelsticity and underground water flow. See in this regard the
papers by Ewing and Lin [3], Choi and Chan [2]. As a special application see Bouziani [1], where
the author has considered a nonlocal problem which is proposed in the mathematical modeling of
technologic process of external elimination of gas, practices in the refining of impurities of Silicon
lamina.
In section 1, we state the problem, define some spaces and give a relevant definition of weak
solution. Section 2 is devoted to the study of existence of the weak solution of the posed problem
by applying Galerkin’s method.
In this paper, we are concerned with the following nonlocal mixed boundary value problem for
the n−dimensional Boussinesq equation non linear in a cylinder QT = Ω × (0, T ), where Ω is a

134
CNMA’2022
Second National Conference on
Mathematics and its Applications

bounded domain in Rn with smooth boundary ∂Ω.


2 2 p−2
 utt − α ∆u − β ∆utt = |u| u,





 u(x, 0) = ϕ(x), ut (x, 0) = ψ(x), (146)


 RtR
 ∂u =

u(ξ, τ)dξdτ, x ∈ ∂Ω,


∂η 0 Ω

where p > 2, ϕ(x) and ψ(x) are given functions and ∂u ∂η

designates the normal derivative.
Now let V (QT ) and W (QT ) be the set spaces defined respectively by:
n o
V (QT ) := u ∈ W21 (QT ) : ∇ut ∈ L2 (QT ), u ∈ Lp (QT ), ut ∈ Lp (QT ) ,

and
W (QT ) := {u ∈ V (QT ) : v(x, T ) = 0} . (147)

Consider the equation

 
(utt , v)L2 (QT ) − α 2 (∆u, v)L2 (QT ) − β 2 (∆utt , v)L2 (QT ) = |u|p−2 u, v . (148)
L2 (QT )

Evaluation of the inner products in (148) and use of boundary condition in (146) leads

−(ut , vt )L2 (QT ) + α 2 (∇u, ∇v)L2 (QT ) − β 2 (∇ut , ∇vt )L2 (QT )
  Z ZT Z tZ !
p−2 2
= |u| u, v 2 − (ψ(x), v(x, 0))L2 (Ω) + α v(x, t) u(ξ, τ)dξ dtdsx
L (QT )
∂Ω 0 0 Ω
Z Z T Z ! Z Z T Z !
+β 2 v(x, t) ut (ξ, t)dξ dtdsx − β 2 v(x, t) ut (ξ, 0)dξ dtdsx
∂Ω 0 Ω ∂Ω 0 Ω
2
+β (∇ψ (x) , ∇v (x, 0))L2 (Ω) , (149)

∀v ∈ W (QT ).
Definition 1.1. A function u ∈ V (QT ) is called a generalized solution of problem (146), if it satis-
fies equation (149) for each v ∈ W (QT ) and u(x, 0) = ϕ(x).

Main results
We now give the main result on the existence of solution of problem (146) and prove it by us-
ing the Galerkin method.
Theorem 57. If ϕ(x) ∈ W21 (Ω), ψ(x) ∈ Lp (Ω) and ψ(x, t) ∈ W21 (Ω), then there is at least one generalized
solution in V (QT ) to problem (146).

References
[1] A. Bouziani, Strong solution for a mixed problem with a nonlocal condition for certain pluri-
parabolic equations, Horishima. Math. J, 27 (1997), 373-390.

[2] Y.S Choi. and K.Y. Chan, A parabolic equation with nonlocal boundary conditions arising from
electro-chemestry, Nonlinear Anal, 18 (1992), 317-331.

[3] R.E. Ewing. and T. Lin, A class of parameter estimation techniques for fluid flow in porous media,
Adv. water resour, 14 (1991), 89-97.

135
Modeling and control

136
CNMA’2022
Second National Conference on
Mathematics and its Applications

A New Efficient Family of Conjugate

Gradient Methods with Strong Wolfe Line
Search Conditions

Amina Hallal 1 [Link]@[Link]

Mohammed Belloufi 1
1 Department of Computer Science and Mathematics, Laboratory Informatics and Mathematics (LiM) , Mohamed
Cherif Messaadia University

Abstract: One of the most important methods for solving nonlinear minimization problems
is the nonlinear conjugate gradient. The new hybrid conjugate gradient methods are defined
as a mix of traditional nonlinear conjugate gradient methods and the new hybrid conjugate
gradient methods. On every iteration, the search direction meets the descent criterion On
every iteration, the search direction meets the descent criterion. To acquire the stepsize, the
strong Wolfe conditions must be met. The method provided is For some assumptions, it meets
the global convergence criterion. Experiments with numbers Improve the competitiveness of
the new algorithm for some test functions.

Keywords: Unconstrained optimization; Hybrid conjugate gradient method; Strong Wolfe

conditions; Numerical results.
2010 Mathematics Subject Classification: Primary 49M07, 49M10, 65K05, 90C06, 90C30.

Introduction
The problem of minimization unconstrained can be express as:

minimize f (x), xRn . (P)

Where f : Rn → R is a smooth nonlinear function.

The conjugate gradient method is one of the most popular methods to solve the most important
nonlinear unconstrained optimization problems, especially for large sizes.
The CG method is usually written in the iterative form.

x0 Rn , xk+1 = xk + αk xk , kN

αk is the step-size and is calculated under exact line search or inexact line search methods. Inexact
line searches are usually used in convergence analysis.
We wiedly use the Wolfe line search : the weak Wolfe conditions, the strong Wolfe conditions, the
generalized Wolfe conditions.
Here the strong Wolfe conditions are used to find αk where

f (xk + αk dk ) − f (xk ) ≤ δαk ∇f (xk )T dk .

136
CNMA’2022
Second National Conference on
Mathematics and its Applications
 
∇f (xk + αk dk )T dk  ≤ −σ ∇f (xk )T dk .
0 < δ < σ < 1.
dk search directions are determinated as follows

 −∇f (x0 ) , for k = 0,



dk = 
 −∇f (x ) + β for k ≥ 1.
k−1 dk−1 ,

 k

βk R the conjugate gradient scalar, where ∇f (xk ) = gk denotes the gradient.

The different choices for βk correspond to different conjugate gradient methods.

References
[1] J. Jian and L. Han and X. Jiang, A hybrid conjugate gradient method with descent property
for unconstrained optimization, Applied Mathematical Modelling. 39 (2015), 1281–1290.

[2] P. Wolfe , Convergence conditions for ascent methods, SIAM Review. 13 (1971), 185–188.

[3] G. Zoutendijk , Nonlinear programming, computational methods, Integerand nonlinear pro-

gramming. (1970), 37–86.

137
CNMA’2022
Second National Conference on
Mathematics and its Applications

Asymptotic Modeling of the Effect of a

Planar Porous Elastic Thin Layer

Athmane Abdallaoui 1 a.abdallaoui18@[Link]

1 Laboratoire de Mathématiques et Physique Appliquées, École Normale Supérieure de Bou Saâda.

Abstract: In this paper, we derive an approximate impedance of a planar thin layer in the
framework of linear elasticity with voids by using the techniques of asymptotic expansion
with scaling. We prove also an error estimate.

Keywords: Elasticity with voids, porous thin layer, impedance operator, asymptotic ex-
pansion
2010 Mathematics Subject Classification: Primary Primary 05C38, 15A15.

Introduction
Governing equations and main motivation

This paper deals with study of a transmission model problem in the framework of linear elasticity
with voids or pores set in a fixed domain Ω− bonded with a planar thin layer Ωδ+ of thickness δ,
from a numerical point of view, the resolution of this problem can not be computed accurately
since the small thickness δ of the thin layer creates instabilities related to the parameter δ. To
avoid these numerical instabilities, we will use the concept of impedance condition which allows
us to replace the initial transmission problem by an equivalent one which doesn’t take into ac-
count any more the thin layer called impedance boundary problem. To begin with, we consider a
three-dimensional model, of linear elasticity with voids in a domain Ωδ = R2 × ]−1, δ[ consisting
of two bonded porous elastic bodies, Ω− = R2 × ]−1, 0[ and a layer Ωδ+ = R2 × ]0, δ[ , we set also
Γ− = R2 × {−1} , Σ = R2 × {0} and Γ+δ = R2 × {δ}. We assume that Ω− and Ωδ+ are homogeneous
and isotropic, we denote by the index + (resp. −) to the restriction on Ωδ+ (resp. on Ω− ) and we
consider the following transmission problem (P δ ) (see [2]):
(1) Equilibrium equations in Ω−
  
P3 δ δ
 j=1 Dj σ−ij u− , ω− = −p−i , i = 1, 2, 3



 P3     (150)

 D h ω δ − g u δ , ωδ = −q .
 j=1 j −j − − − − −

(2) Equilibrium equations in Ωδ+

  
P3 δ , ωδ = 0, i = 1, 2, 3,
D σ u


 j=1 j +ij + +

 P3     (151)

 D h ω δ − g u δ , ωδ = 0.
 j=1 j +j + + + +

138
CNMA’2022
Second National Conference on
Mathematics and its Applications

(3) Dirichlet boundary conditions on Γ−


δ
 u−i = 0, i = 1, 2, 3,




ω−δ = 0.




(4) Newmann boundary conditions on Γ+δ

  
P3 δ δ
 j=1 σ+ij u+ , ω+ νj = 0, i = 1, 2, 3,



 (152a)
D3 ω+δ = 0.




(5) Transmission conditions at the interface Σ


δ δ
u−i = u+i , , i = 1, 2, 3,






ω−δ = ω+δ ,



(153)

P3    
δ , ωδ ν = P3 σ


 σ u u δ , ωδ ν , i = 1, 2, 3,


 j=1 +ij + + j j=1 −ij − − j



 δ
α− D3 ω− = α+ D3 ω+ . δ

where v = (ν1 , ν2 , ν3 ) = (0, 0, 1) is the unit normal vector to Σ, σ±ij is the stress tensor, p− is the
body force vector, g± is the intrinsic equilibrated body force, h± is the equilibrated stress vector,
∂
q− is the extrinsic equilibrated body force and Dj = ∂x . For the sake of simplicity in the next, we
j
adopt the following notations:
   
u±δ , ω±δ = u±1 δ δ
, u±2 δ
, u±3 , ω±δ
        
σ± u−δ , ω−δ ν = σ±13 u±δ , ω±δ , σ±23 u±δ , ω±δ , σ±33 u±δ , ω±δ
           
σ± u−δ , ω−δ ν, α± D3 ω±δ = σ±13 u±δ , ω±δ , σ±23 u±δ , ω±δ , σ±33 u±δ , ω±δ , α± D3 ω±δ .

The constitutive equations for the linear isotropic elasticity with voids are defined by:

σ±ij = 2µ± e±ij + λ± e±pp δ±ij + β± ω±δ δij , i, j = 1, 2, 3

h±j = α± Dj ω±δ , j = 1, 2, 3
g± = β± e±pp + ζ± ω±δ , p = 1, 2, 3

where δij is the Kronecker delta, e±ij is the strain tensor defined by:

1 δ δ

e±ij = Di u±j + Dj u±i , i, j = 1, 2, 3
2
and µ± , α± , ζ± and β± are material constants satisfying the inequalities:

µ± > 0, α± > 0, ζ± > 0, 2µ± + 3λ± > 0, (2µ± + 3λ± ) ζ± > β±2 .

As was already pointed out, our aim in this paper is to derive an approximate impedance bound-
ary condition on the interface Σ that incorporates in an approximate way the effect of the thin
layer Ωδ+ on Ω− to reduce the transmission problem (P δ ) to an equivalent impedance boundary
value problem set in the fixed domain Ω− by using the techniques of asymptotic expansion with
scaling.

139
CNMA’2022
Second National Conference on
Mathematics and its Applications

Concept of impedance in modeling

Our goal is to reduce the transmission problem set in Ωδ = Ω− ∪ Ωδ+ to an impedance bound-
ary value problem set only on the fixed domain Ω− . The exact effect of the thin layer Ωδ+ on the
domain Ω− is given by the impedance operator Tδ defined by:
     
Tδ v δ , φδ := σ+ u+δ , ω+δ ν|Σ , α+ D3 ω+|Σ
δ

 
where u+δ , ω+δ is the unique solution of the following boundary value problem:



 Equations (151) in Ωδ+ ,


Boundary conditions on Γ+δ ,





 u δ = v δ on Σ, ωδ = φδ on Σ,

+ +

from the transmission conditions (153), it follows that:

      
σ− u−δ , ω−δ ν|Σ , α− D3 ω−|Σ
δ
= Tδ u−δ , ω−δ ,
|Σ

and the transmission problem (P δ ) is then equivalent to the following impedance problem set in
Ω− : 
 Equations (150) in Ω− ,




 u δ = 0 on Γ− ,


δ  −

(P− ) : 
 δ


 ω− = 0 on Γ− ,
       
δ

 σ− u−δ , ω−δ ν|Σ , α− D3 ω−|Σ = Tδ u−δ , ω−δ on Σ.


|Σ

Since an explicit expression of the exact impedance operator Tδ is not reachable for the general
case, we will just derive an effective approximation T∗δ of Tδ with:
   
T∗δ = δT∗ and T∗ v δ , φδ = (C1 , C2 , C3 , C4 ) v δ , φδ ,

where
  4µ+ (µ+ + λ+ ) 2 δ 2µ+ λ+
C1 v δ , φδ = D1 v+1 + µ+ D22 v+1
δ
+ D D v δ + µ+ D1 D2 v+2
δ
(λ+ + 2µ+ ) 2µ+ + λ+ 1 2 +2
2µ+ β+
+ D ϕδ ,
2µ+ + λ+ 1 +
  4µ+ (µ+ + λ+ ) 2 δ 2µ+ λ+
C2 v δ , φ δ = D2 v+2 + µ+ D12 v+2
δ
+ D D v δ + µ+ D1 D2 v+1
δ
(λ+ + 2µ+ ) 2µ+ + λ+ 1 2 +1
2µ+ β+
+ D ϕδ ,
2µ+ + λ+ 2 +
 
C3 v δ , φδ = 0,

and
    2µ+ β+   ζ+ (2µ+ + λ+ ) − β+2
C4 v δ , φδ = α+ D12 ϕ+δ + D22 ϕ+δ − δ
D1 v+1 δ
+ D2 v+2 − ϕ+δ .
2µ+ + λ+ 2µ+ + λ+

140
CNMA’2022
Second National Conference on
Mathematics and its Applications

The solution (u−δ , ω−δ ) of the transmission problem (P δ ) in Ω− is then approximated by the solution
δ , ωδ ) of the following approximate impedance problem:
(u−∗ −∗
 P  
3 δ , ωδ = −p , i = 1, 2, 3
Dj σ−ij u−∗

−i
 −∗


 j=1
 P3    
δ δ δ
 j=1 Dj h−j ω−∗ − g− u−∗ , ω−∗ = −q− .
   

δ
P−∗ :  
 δ = ωδ = 0 on Γ ,
u−∗


 −∗ −
      
 δ , ωδ ν , α D ωδ δ δ
 σ− u−∗ − 3 −∗|Σ = T∗δ u−∗ , ω−∗ on Σ,

−∗ |Σ

and we prove the following main result of the paper:

 
Theorem 1.1. The boundary value problem P−∗ δ has a unique solution in the space

 h i4 



 (v− , ϕ− ) ∈ H 1 (Ω− ) : 




 h i4 

2
 (D1 v−1 , D1 v−2 , D1 v−3 , D1 ϕ− ) ∈ L (Σ) ,

 

 
W∗ = 

2 (Σ) 4 ,
h i 
(D v , D v , D v , D ϕ ) ∈ L

 



 2 −1 2 −2 2 −3 2 − 



 


 v− = 0 on Γ− , ϕ− = 0 on Γ− . 


and the following error estimate holds

u−δ − u−∗
δ
[H (Ω− )]3
1 + ω−δ − ω−∗
δ 1
H (Ω )
≤ Cδ2 .
−

where the constant C depends only on p− , q− and the elasticity coefficients.

References
[1] A. Abdallaoui and K. Lemrabet, Mechanical impedance of a thin layer in asymmetric
elasticity. Applied Mathematics and Computation, 316:467–479, 2018.

[2] S.C. Cowin and J.W. Nunziato, Linear elastic materials with voids. J. Elasticity 13 (1983)
125.

[3] P.A. Raviart, J.M. Thomas, Introduction à l’analyse numérique des equations aux dérivées
partielles, Masson, Paris Milan Barcelone Mexico, 1988.

141
CNMA’2022
Second National Conference on
Mathematics and its Applications

Estimation of the unrestricted PEXPAR(1)

model with nonlinear least squares
method

BECILA Sabah 1 sabahbc@[Link]

MERZOUGUI Mouna 1
1 Department of mathematics , Badji Mokhtar University of Annaba

Abstract: In this paper, we study the strong consistency and asymptotic normality properties
of nonlinear least squares (NLS) estimator of the periodic EXPAR(1) model. The general
statistical literature on estimation of nonlinear models of Gallant and White[3] is used.
Simulation study and one real example are given to assess the performance of this NLS.

Keywords: Nonlinear time series, periodic exponential autoregressive model, nonlinear

least squares, asymptotic normality, consistency.
2010 Mathematics Subject Classification: 62F12; 62M10

Introduction
The problem of estimation in nonlinear context is more complicated, it has been approached by
many authors, see, for instance, Tong (1990)[5] and Bierens (1994)[2]. The nonlinear least squares
(NLS) and the quasi-maximum likelihood estimators (QMLE) are essentially used in this case.
The NLS estimator of parameter θ is the one that minimizes the sum of squared residuals or the
one that solves the first-order conditions of the minimization problem. The application requires
the strict stationarity of the model and we will use the standard tools as the compactness and the
Taylor expansion.
In this paper, we prove the consistency and the asymptotic normality of the NLS estimator for
the PEXPAR(1) model. We obtain the same results as Amendola and Francq (2009)[1] for their
QMLE for the non periodic EXPAR(1) and for the CLSE (conditional least squares estimators) of
Tj∅stheim (1986)[4] who uses slightly different conditions.

Main results
Definition of the NLS estimator for P EXP ARS (1) model

The proces {Yt ; t ∈ Z} is said to follow a periodic exponential autoregressive P EXP ARS (1), with
period S(S ≥ 2) , if it satisfies:
0 0
Yt = (αt,1 + αt,2 exp(−γt0 Yt−1
2
))Yt−1 + εt , t ∈ Z. (154)
Where {εt ; t ∈ Z} is i.i.d. process with mean 0 and finite variance σt2 and γt > 0. The autoregressive
parameters αt,1 , αt,2 ∀t ∈ Z, the nonlinear parameter,γt , and the innovation variance σt2 are peri-
odic, in time, with period S, i.e.,

142
CNMA’2022
Second National Conference on
Mathematics and its Applications

2
αt+kS,1 = αt,1 , αt+kS,2 = αt,2 , γt+kS = γt , and σt+kS = σt2 , ∀k, t ∈ Z.
Putting t = i + Sτ, i = 1, 2, . . . , S and τ ∈ Z, one can rewrite Equation (154) in the equivalente form:
0 0
Yi+Sτ = (αi,1 + αi,2 exp(−γi0 Yi+Sτ−1
2
))Yi+Sτ−1 + εi+Sτ , i = 1, . . . , S, τ ∈ Z (155)

The vector of the parameters θ = (θ10 , . . . , θS0 )0 ∈ R3S , where θi = (αi,1 , αi,2 , γi )0 , i = 1, . . . , S,
belongs to a parameter space of the form:
Θ S ⊂ (] − 1, +1[×] − 1, +1[×]0, ∞[)S .
The true value of the parameters is unknown and denoted by:
0 0
θi,o = (αi,1 , αi,2 γi0 )0 , i = 1, . . . , S, and θ0 = (θ1,0
0 0
, . . . , θS,0 )0 ∈ R3S .
Suppose that the observations Y1 , . . . , Yn constitute a realization of length n from (155) and n = mS
which means that we have m full cycle of data. The NLS estimator of h is the one that minimizes
the sum of squared residuals. Let θbn the NLS estimator which is defined as a measurable solution
of the following minimization problem with respect to θ:
bn = arg inf
θ Qn (θ).
θ∈Θ S
Where
n
Qn (θ) = n1 2
))Yt−1 )2 .
P
`t and `t (θ) = (Yt − (αt,1 + αt,2 exp(−γt Yt−1
t=1
The initial value is unknown but its choice is not important for the asymptotic behavior of the
NLS estimator so we put Y0 = 0 which defines the operational criterion Q
en (θ),
then θn = arg infθ∈Θ S Qn (θ).
b e

S m−1 S
en (θ) = 1
X1X 1Xe
2
Q (Yi+Sτ − (αi,1 + αi,2 exp(−γi Yi+Sτ−1 ))Yi+Sτ−1 )2 = Qi,m (θi ). (156)
S m S
i=1 τ=0 i=1

The existence of a solution to (156) is guaranteed under the assumption:

A1: Θ S is compact Borel set⊂ R3S .
The assumption of compactness of Θ S plays a key role this ensures that the infimum over Θ S
exists see, for instance, theorem 1.6.1 (Bierens 1994, 16)[2].
To obtain the estimators we differentiate Q ei,m (θi ) with respect to each parameter θi , i = 1, . . . , S.
The first-order condition of the NLS minimization problem is a system of 3S nonlinear equations
with 3S unknowns. For a given season i, the gradient vector of Q ei,m (θi ) is
 
m−1
 Y i+Sτ−1

ei,m (θi )
∂Q 1X
 
2  2

= −2 (Yi+Sτ − (αi,1 + αi,2 exp(−γi Yi+Sτ−1 ))Yi+Sτ−1 ) ×  Yi+Sτ−1 exp(−γi Yi+Sτ−1 )  .

∂θi m  
τ=0 −α Y 3
 2

i,2 i+Sτ−1 exp(−γi Yi+Sτ−1 )


ei,m (θi )
∂Q
Then analytical solutions cannot be found for ∂θ = 0 and the NLS estimators θ bi,m must be
i
calculated by numerical optimization such as gradient or quasi-Newton method. Under the hy-
pothesis that the expectation exists, we have
  



 Yi+Sτ−1


0 0 0 2
 
0 2

E (Yi+Sτ − (αi,1 + αi,2 exp(−γi Yi+Sτ−1 ))Yi+Sτ−1 )  Yi+Sτ−1 exp(−γi Yi+Sτ−1 )  = 0
 

  
  0 3 0 2 
−α Y exp(−γ Y
i,2 i+Sτ−1 i ) i+Sτ−1

ei,m (θi )
∂Q
which is the asymptotic version of ∂θi
= 0 at θi,0 by the ergodic theorem.

143
CNMA’2022
Second National Conference on
Mathematics and its Applications

Asymptotic behavior of the NLS estimator

We make the following assumptions to show strong consistency and asymptotic distribution.
A2: (i)Yt is periodically strictly stationary, a sufficient condition for this is;
|αi,1 | + |αi,2 | < 1 for i = 1, . . . , S.
(ii)E(εt2 ) < ∞, for any t ∈ Z.
A3 :θi,0 belongs to the interior Θ ◦ ofΘ.
A4 :E(εt6 ) < ∞, for any t ∈ Z.

Simulation results
The asymptotic properties of the NLS estimator are shown by a small simulation study. We gener-
ate time series from the P EXP AR2 (1) models with sizes n = 300 and 600. We consider N = 1000
Monte Carlo replications and report the mean value of NLS estimations and their standard devia-
tions. Table 1 gives the estimation for the parameters θ = (−0.7, 0.2, 2; 0.5, −0.4, 1)0 . The programs
are written in R using the nlm function.

N LSE α1,1 α1,2 γ1 α2,1 α2,2 γ2

n = 300 −0.7005 0.2012 2.0006 0.4975 −0.4013 1.0001
SD 0.0309 0.0307 0.0147 0.0597 0.0586 0.0147
n = 600 −0.7003 0.2000 1.9997 0.4999 −0.3989 0.9995
SD 0.0270 0.0271 0.0142 0.0436 0.0411 0.0144
Table 1: Estimation results for P EXP AR2 (1)
From Table 1, we can see that the parameters are well estimated and the standard deviation de-
creases when n increases showing that the estimates are consistent.

Application

As an example, we fit the model P EXP AR1 2(1) to the famous monthly Fraser River flow data
by the NLS method, from January 1915 to December 1990. The estimation results of the parame-
ters are given in Table 2.

i Jan Feb Mar Apr May Jun

αi,1 0.6829 0.8145 0.7652 0.7069 0.1805 0.2368
αi,2 -0.4936 -1.2685 0.2472 2.5652 0.0450 -4.6453
γi 46.6180 242.1705 16.8679 104.7684 27.9312 440.3529
i Jul Aug Sep Oct Nov Dec
αi,1 0.6748 0.7616 0.7066 0.7842 0.7406 0.7369
αi,2 0.9969 -1.4752 1.3841 0.3392 1.5638 1.6418
γi 146.5914 248.9337 153.4060 6.5568 97.6726 153.5859
Table 2: Parameter estimation of the P EXP AR12 (1) .

144
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] A. Amendola and C. Francq, Concepts and tools for nonlinear time series modelling, In
Handbook of Computational Econometrics, UK. pp. 377-427, 2009.

[2] H. J. Bierens, Estimation, testing, and specification of cross-section and time series models,
Cambridge University Press, 1994.

[3] A. R. Gallant and H. White, A Unified Theory of Estimation and Inference for Nonlinear
Dynamic Models, Basil Blackwell, New York, 1988.

[4] D. Tj∅stheim, Estimation in Nonlinear Time Series Models, Elsevier Science Publishers B.V.
North-Holland, Stochastic Processes and their Applications 21 (1986), 251-273.

[5] H. Tong, Nonlinear Time Series : a Dynamical System Approach, Oxford University Press,
Oxford, 1990.

145
CNMA’2022
Second National Conference on
Mathematics and its Applications

Enquête sur le taux de personnes

vaccinées contre le coronavirus en
Algérie

Yamina Khemal Bencheikh 1 bencheikh-00@[Link]

Mona Dali 2
Yassmina Bouhassir 2
1 Laboratoire de Mathématiques Fondamentales et Numériques LMFN Département de Mathématiques Faculté
des Sciences-Université Sétif 1 El Bez
2 Département de Mathématiques Faculté des Sciences-Université Sétif 1 El Bez

Résumé: L’échantillonnage est le processus par lequel on détermine l’échantillon et son but
suprême est l’atteinte d’une représentativité impartiale de la population à l’étude, pour que
toute estimation basée sur l’échantillon soit sans biais et inférée à la population, avec un
niveau de précision requis. L’échantillonnage a été utilisé conjointement avec les méthodes
de l’inférence statistique pour apporter quelques éléments de réponse à l’enquête que nous
avons mené dans ce travail; concernant le taux de vaccination en Algérie pour la pandémie
de coronavirus, qui a envahi le monde entier depuis la fin de l’année 2019 et jusqu’à l’heure
actuelle.

Mots clés: Statistique inférentielle, échantillonnage, estimation, test statistique, enquête,

Covid’19.

Introduction
Un vaccin contre la maladie à coronavirus 2019 entraîne et prépare le système immunitaire à
reconnaître et à combattre le coronavirus SARS-CoV-2, ce qui permet de prévenir cette maladie.
L’année 2021 a marqué le lancement des campagnes de vaccination contre la Covid-19 à travers le
monde. L’objectif de notre travail est de tester deux hypothèses annoncées par le ministre algérien
de la santé. La première hypothèse annoncée le 25 janvier 2022, sous forme d’une interview faite à
la chaine algérienne Echourouk News([Link] dans
laquelle le ministre affirme que le pourcentage de personnes ayant plus de 18 ans vaccinées (ayant
reçu exactement deux doses) est de 29 %. La deuxième hypothèse annoncée le 21 avril 2022, sous
forme d’un communiqué fait pour la chaine algérienne Enahar TV
([Link] qui confirme que le pourcentage de person-
nes ayant plus de 18 ans vaccinées est de 30 %. Ce dernier pourcentage concerne les personnes
ayant plus de 18 ans et ayant reçu au moins deux doses (2 doses et 3 doses).

Les données
Les données ont été recueillies suivant la technique boule de neige [2]. Le recueil des données

146
CNMA’2022
Second National Conference on
Mathematics and its Applications

à commencé le 14 février 2022 et s’est achevé le 24 avril 2022. Les données obtenues concernent
564 familles qui correspondent à n = 2477 personnes enquêtées.

Taille de l’échantillon
Pour le choix de la taille de notre échantillon, nous travaillons avec un risque d’erreur α égal
à 1% ; ce qui correspond à un seuil de confiance de 99% avec une longueur d’intervalle de confi-
ance ne dépassent pas les 0.05. La formule utilisant la loi normale (ou la technique du TCL [1])
p(1−p)
donnant la taille minimale de l’échantillon est n = 2 ×α [2] où p est la proportion de personnes
vaccinées dans la population et  = 0.05
2 = 0.025.

Étude statistique préliminaire

L’examen du tableau de données obtenu montre que 33.75 % ont reçu au moins 1 dose de vac-
cin. Le pourcentage de personnes décédées suite à la maladie de corona est de 0.28 % pour les
personnes vaccinées et 1.09 % pour les personnes non vaccinées. Le pourcentage de personnes
vaccinées décédées souffrant de maladies chroniques représente 28.5 % de la population des vac-
cinées décédées. Quant aux personnes non vaccinées décédées souffrant de maladies chroniques,
il est de 22.2 % de la population des non vaccinés décédées.

Estimation des paramètres utilisés

On estime ponctuellement que la proportion de personnes vaccinées (au moins une dose) représente
33.75 % dont 2.3 % des personnes ont reçu uniquement 1 dose, 27.04 % ont reçu exactement 2
doses et 4.4 % ont reçu trois doses. Quant aux pourcentages de personnes décédées suite à la
maladie de corona, il est de 1.37 %.
L’expression de l’intervalle de confiance IC1−α (p) dans le cas de l’estimation d’une proportion p
de la population avec une confiance de (1 − α) [2] est :
r r
f (f − 1) f (f − 1)
IC1−α (p) = [f − Z1− α2 f + Z1− α2 ]
n n
Z1− α2 est la valeur correspondante à une probabilité égale à (1 − α2 ) et f est la proportion dans
l’échantillon. Nous pouvons conclure avec une confiance de 99 % que :

• La proportion de personnes ayant reçu 1 dose de vaccin se situe entre 1.5 % et 3.1 %.

• La proportion de personnes ayant reçu 2 doses de vaccin se situe entre 24.75 % et 29.35%.

• La proportion de personnes ayant reçu 3 doses de vaccin se situe entre 3.3 % et 5.4 %.

• La proportion de personnes non vaccinées se situe entre 63.78 % et 68.70 %.

• La proportion de personnes vaccinées décédées se situe entre 0.007 % et 0.55 %.

• La proportion de personnes non vaccinées décédées se situe entre 0.55 % et 1.62 %.

Test des hypothèses du ministre de la santé

Test de l’hypothèse du 25 janvier 2022 : p0 = 0.29.

147
CNMA’2022
Second National Conference on
Mathematics and its Applications

Proportion du nombre de personnes ayant reçu 2 doses de vaccin dans la population. Dans le test
bilatéral [3], les hypothèses se formulent ainsi

H 0 : p = p0 ,
H 1 : p , p0

Si H0 est vraie alors la zone d’acceptation ZA est :

r r
h p0 (1 − p0 ) p0 (1 − p0 ) i
ZA = p0 − Z1− α2 p0 + Z1− α2 = [0.2664774 0.3135226]
n n
f2 = 0, 270488494 étant la proportion de personnes dans l’échantillon ayant reçu exactement deux
doses de vaccin. On remarque que f2 ∈ ZA = [0.2664774 0.3135226].

La première hypothèse du ministre de la santé est donc vraie avec un risque d’erreur de 1%.

Test de l’hypothèse du 22 avril 2022 : cette fois-ci p0 = 0.3.

Cette proportion représente le nombre de personnes ayant reçu au moins deux doses de vaccins.
r r
h p0 (1 − p0 ) p0 (1 − p0 ) i
ZA = p0 − Z1− α2 p0 + Z1− α2 = [0.2762444 0.3237556]
n n
f étant la proportion de personnes dans notre échantillon ayant reçu au moins 2 doses de vaccin.
On remarque que f = f2 + f3 = 0, 270488494 + 0, 044004845 = 0.31449333 ∈ ZA .

L’hypothèse nulle H0 est donc vraie avec un risque d’erreur de 1%. On conclue que le ministre de
la santé a raison pour cette deuxième déclaration.
Nous pouvons alors conclure avec une confiance de 99 %, que le taux de vaccination de la popu-
lation algérienne est aux environs de 30%, comme l’a annoncé le ministre de la santé algérienne,
le Professeur Benbouzid. Ce taux étant relativement faible par rapport aux taux de vaccinations
dans le reste du monde.

References
[1] Anderson D.R., Sweeney D. J, Williams T. A, Statistique pour l’économie et la gestion, 2ème
édition, deboeck supérieur, (2007)

[2] Elmarhoum A, Cours échantillonnage et estimation,Université Mohamed V Rabat. (2019).

[3] Escoffier J, Probabilités et statistiques pour le capes et l’agrégation interne, 2ème édition,
ellipses, (2006).

148
CNMA’2022
Second National Conference on
Mathematics and its Applications

Characterization of some admissible

trajectories for a warfare differential game
problem

Benghebrid Safa 1 [Link]@[Link]

Bouremani Touffik 2
Benterki Djamel 3
1 Laboratory of numerical and fundamental mathematics, Department of mathematics, University of Ferhat Ab-
bas Setif1, Algeria
2 Laboratory of applied mathematics, Faculty of Technology, University of Ferhat Abbas Setif1, Algeria
3 Laboratory of numerical and fundamental mathematics, Department of mathematics, University of Ferhat Ab-
bas Setif1, Algeria

Abstract: In this work, we use the approach of dynamic programming algorithm to charac-
terize some admissible trajectories of warfare problem formulated in [1] as well as to identify
maximal interval in which some admissible condition are satisfied. To illustrate our results,
some numerical experiments are presented.

Keywords: dynamic programming, differential games, Hamilton jacob Bellman equa-

tion.
2010 Mathematics Subject Classification: Primary 49J15, 49L20, 34A60.

Introduction
This study concerns the warfare game problem between two opposing forces in military conflicts,
formulated and studied heuristically by Isaacs [1]. Firstly, we use the theoretical dynamic pro-
gramming algorithm, described in [2, 3], to identify a certain admissible trajectories in the form
of the Hamiltonian flow as solutions of HJB equation. Secondly, we describe the maximal domain
in which some admissible conditions are verified.

Dynamic programming formulation

Problem. Given T , m1 , m2 > 0, c1 > c2 > 0. Find

min max C(y, u(.), v(.)), ∀y ∈ Y0

u(.) v(.)

149
CNMA’2022
Second National Conference on
Mathematics and its Applications

subject to
RT
C(u(.), v(.)) = g(x(T )) + f0 (x(t), u(t), v(t))dt
0
0
x (t) = f (x(t), u(t), v(t)), a.e. ([0, T ]), x (0) = y
u(t) ∈ U (x(t)), v(t) ∈ V (x(t)), a.e. ([0, T ])
x(t) ∈ Y0 ∀t ∈ [0, T ), x(T ) ∈ Y1 , T fixed
defined by the following data:

f (x, u) = (m1 − c1 vx2 , m2 − c2 ux1 , −1), f0 (x, u) = (1 − v)x2 − (1 − u)x1

U (x) = U = [0, 1] , V (x) = V = [0, 1] , g (ξ) = 0, ∀ξ ∈ Y1 ,
Y0 = R+ × R+ × [0, T ), Y1 = R+ × R+ × {0}.

Characterization of the Hamiltonian

The pseudo-Hamiltonian is given in our case by:

H(x, p, u, v) = p1 m1 + p2 m2 − p3 + x2 − x1 + (1 − c2 p2 )x1 u − (1 + c1 p1 )x2 v.

Proposition 12. The Isaacs’ Hamiltonian and the corresponding marginal multifunctions defined in
[2, 3] are given by the formulas:

H(x, p) = min max H(x, p, u, v) = max min H(x, p, u, v)

u(.) v(.) u(.) v(.)

 
1

 {0} if p2 < c2

 {0} if p1 > − c1

 
 1
 
1
Û (p) =  if p2 > and V̂ (p) =  if p1 < − c1 ; F̂(p) = Û (p) × V̂ (p)
 
 {1}
 c2  {1}
 1
 
 [0, 1] if 1  [0, 1] if 1
p2 = p1 =
 
c2 c2

The Hamiltonian H(., .) as well as its domain Z are C 1 -stratified by the stratification

SH = {Z +,+ , Z +,− , Z +,0 , Z −,+ , Z −,− , Z −,0 , Z 0,+ , Z 0,− , Z 0,0 }

defined by:


−1 1
p1 m1 + p2 (m2 − c2 x1 ) − p3 + x2 if (x, p) ∈ Z +,+ = {(x, p) ∈ z p1 > c1 , p2 > c2 }






p1 m1 + p2 m2 − p3 + x2 − x1 if (x, p) ∈ Z +,− = {(x, p) ∈ z −1
p1 > c , p2 < c1 }



1 2



p1 m1 + mc 2 − p3 + x2 − x1 Z +,0 = {(x, p) ∈ z p1 > −1 p2 = c1 }



 if (x, p) ∈ c1 ,

 2 2

Z −,+ = {(x, p) ∈ z p1 < −1 p2 > c1 }


p1 (m1 − c1 x2 ) + p2 (m2 − c2 x1 ) − p3 if (x, p) ∈ c1 ,




 2

1
H(x, p) =  p1 (m1 − c1 x2 ) + p2 m2 − p3 − x1 if (x, p) ∈ Z −,− = {(x, p) ∈ z p1 < −1
c1 , p2 < c } .


 2

p1 (m1 − c1 x2 ) + mc 2 − p3 − x1

if (x, p) ∈ Z −,0 = {(x, p) ∈ z −1
p1 < c , p2 = c1 }



2 1 2



−m1
Z 0,+ = {(x, p) ∈ z p1 = −1 p2 > c1 }



 c1 p2 (m2 − c2 x1 ) − p3 + x2 if (x, p) ∈ c1 ,

 2
−m1
Z 0,− = {(x, p) ∈ z p1 = −1 p2 < c1 }

c1 + p2 m2 − p3 + x2 − x1 if (x, p) ∈ c1 ,




 2
−m1 m2

1
c1 + c2 − p3 + x2 − x1 if (x, p) ∈ Z 0,0 = {(x, p) ∈ z p1 = −1
c1 , p2 = c }



2

150
CNMA’2022
Second National Conference on
Mathematics and its Applications

The Set of terminal transversality points Z ∗ (see[3,4]) in our case is given by :

∗
Z+,− = {((s1 , s2 , 0), (0, 0, s2 − s1 )), s1 ≥ 0, s2 ≥ 0}.

Generalized Hamiltonian and characteristic flow

The Hamiltonian system on the stratum Z +,− for which H +,− (x, p) = p1 m1 + p2 m2 − p3 + x2 − x1
is given as follow:
 


 x́1 (t) = m1 x1 (0) = s1  
 ṕ1 (t) = 1 p1 (0) = 0

 

x́2 (t) = m2 x2 (0) = s2 ;  ṕ2 (t) = −1 p2 (0) = 0
 

 

 

 x́ (t) = −1 x (0) = 0  ṕ (t) = 0 p (0) = s − s
 
3 3 3 3 2 1

The solution of this system in the form of the Hamiltonian flow X∗+,− (t, s1 , s2 ) = (X +,− (t, s1 , s2 ) , P +,− (t, s1 , s2 ))
is given by the formulas:


+,− +,−
 X (t, s1 , s2 ) = (m1 t + s1 , m2 t + s2 , −t) , t ∈ I (s1 , s2 )




 P +,− (t, s1 , s2 ) = (t, −t, s2 − s1 ) , I +,− (s1 , s2 ) = (τ +,− (s1 , s2 ) , 0)



The admissible trajectories must also satisfy the following conditions (see [2, 3]):

(X ∗ (t, s), P ∗ (t, s)) ∈ Z +,− ,∀t ∈ (τ +,− (s), 0) , s = (s1 , s2 )

on the maximal intervals I +,− (s) = (τ +,− (s), 0). In order to characterize the extremity τ +,− (s) < 0, we
present the following result.

Lemma 20. The extremity τ +,− (.) in our case is defined by

m1 m2


 − c1 if s1 > c1 and s2 > c1

 1

τ +,− (s) =  s1 < mc 1 s2 > s1mm2

 −s1

m1 if and

 1 1
 −s2 s1 > sm2
2 m1
s2 < mc 2

if and

m2 1

References
[1] [Link], Differential games, Dover Publications Inc, New York, 1964.

[2] Ş. Mirică, User’s guide on dynamic programming for autonomous differential games and opti-
mal control problems, Rev Roumaine Math. 49 (2004), 501-529.

[3] Ş. Mirică, Constructive dynamic programming in optimal control, Editura Academiei Ro-
mane, Bucureşti, 2004.

151
CNMA’2022
Second National Conference on
Mathematics and its Applications

Predictive mathematical models of the

COVID-19 pandemic in obese people

Boubekeur Maroua Amel 1 [Link]@[Link]

Belhamiti Omar1 [Link]@[Link]
1 Department of mathematics , University of Mostaganem Algeria

Abstract: In this work, we propose a mathematical model that highlights the narrow link
between obesity and the mortality due to COVID-19, as well as patients admitted to the ICU
for COVID-19 acute respiratory distress syndrome (ARDS). This model takes into account
different disease states and is represented mathematically by a nonlinear temporal system of
ordinary differential equations. An analysis of the stability of different equilibrium states is
also obtained to theoretically confirm the mathematical realism. Numerical simulations are
presented to explain the usefulness of the proposed model.

Keywords: Obesity, Covid-19, Model Validation, Stability Analysis.

Introduction
The Covid-19 epidemic was alerted by WHO in December 2019, and was declared a public health
emergency of international concern (USPPI) by the same organization on January 30, 2020. As of
October 17, 2021, more than 241 million cases had been confirmed and nearly 5 million deaths
worldwide. The global spread is very rapid, with 170 countries now reporting at least one case.
It is very important to understand the dynamics of the epidemic’s transmission early in order to
better control its evolution and assess the effectiveness of control measures [8].
Many studies have established that several factors have a surprising correlation with higher mor-
tality in individuals with Covid-19: arterial hypertension and smoking [8], obesity [4], diabetes,
cardiac and pulmonary pathology [3, 8]. Over the past two years, many mathematical modeling
studies of covid-19 associated with other chronic diseases have emerged, among these works, we
can cite the work in [5, 6].
Our aim in this work is to highlight the very negative effect of the COVID-19 pandemic on over-
weight and obese people. In this work, we analyze the dynamics of the co-infection of obesity and
COVID-19. Our objective is threefold,

1. We framed a deterministic model that describes obesity and COVID-19 co-infection.

2. We show the existence of an equilibrium and discuss its stability.

3. The results and discussions for the model are presented .

Main results

152
CNMA’2022
Second National Conference on
Mathematics and its Applications

Model formulation and mathematical analysis

The model is given by the following set of differential equations

Stability of equilibrium points

The system admits two equilibrium points, the disease-free-equilibrium point E0 and the endemic
equilibrium point E1 exists if the basic reproduction number R0 > [Link]
µα1 B1 ((µ + γ1 ) β2 + (µ + γ2 ) β1 )
R0 = .
(µ + η) (β1 + β2 + β3 + µ) (µ + γ1 ) (µ + γ2 )
Theorem 58. The disease-free equilibrium E0 of the system is locally-asymptotically stable in Ω, when-
ever
R0 < 1.
Theorem 59. The endemic equilibrium E1 of the system is locally-asymptotically stable in Ω, whenever

R0 > 1.

References
[1] [Link] et al , Optimal Control of Mathematical modeling of the spread of the
COVID-19 pandemic with highlighting the negative impact of quarantine on diabetics
people with Cost-effectiveness, Chaos, Solitons & [Link]. 145(2021), 110777.

[2] [Link] et al. The effect of metformin consumption on mortality in hospitalized

COVID-19 patients: a systematic review and meta-analysis. Diabetes & Metabolic Syn-
drome, Clinical Research & Reviews. vol. 14, (2020) , 2177-2183.

[3] [Link] et al, Risk factors for disease severity, unimprovement, and mortality in COVID-
19 patients in Wuhan, China. Clinical microbiology and infection no 6,vol. 26, (2020),767-
772.

[4] [Link] et al, Obesity as a risk factor for greater severity of COVID-19 in patients with
metabolic associated fatty liver disease, Metabolism vol. 108, (2020), 154244.

[5] S. Anusha and S. Athithan, Mathematical Modelling Co-existence of Diabetes and

COVID-19: Deterministic and Stochastic Approach. (2021).

[6] Y. Marimuthu et al, COVID-19 and tuberculosis: a mathematical model based forecasting
in Delhi, India. indian journal of tuberculosis, no 2, vol. 67,(2020) , 177-181.

[7] [Link] et al. Neutrophil-to-lymphocyte ratio as an independent risk factor for mortality in
hospitalized patients with COVID-19, Journal of Infection .vol. 81,(2020): e6-e12.

[8] [Link] et al, Risk factors of critical & mortal COVID-19 cases: A systematic literature
review and meta-analysis. Journal of infection, no 2,vol. 81,(2020),e16-e25.

[9] [Link] et al., On the definition and the computation of the basic reproduction ratio
R0 in models for infectious diseases in heterogeneous populations. Journal of Mathematical
Biology„ vol. 28 (1990), 365-382.

[10] [Link] et al, SARS-CoV-2 Transmission From People Without COVID-19 Symp-
toms. JAMA network open, vol. 4,(2021) e2035057-e2035057.

153
CNMA’2022
Second National Conference on
Mathematics and its Applications

A logistic equation with constant

imigration to model the spread of
COVID-19 in the first and second waves

Boucherma Rayane 1 ranabcm09@[Link]

Abdelouahab Mohammed salah 1
Boularouk Yakoub 1
1 Department of Mathematics and Computer Sciences, Abdelhafid Boussouf University , Mila , Algeria

Abstract: Some recent papers have modeled the growth dynamics of COVID-19 in several
countries for the first wave using the logistic equation and getting agreeable results. However,
we got poor results when using it to model the other waves. To overcome this flaw, we used
a logistic equation involving constant immigration. The study was conducted in several
countries, including Italy, Russia, Japan, Algeria, Iraq, and Qatar, using data from the Center
of Systems Science and Engineering (CSSE) and WHO reports. One has estimated the model
parameters for the second wave through linear regression (after performing a suitable variable
change that transforms the nonlinear model into a linear one). According to a comparative
study between our approach and the existing one, we conclude that our approach has higher
accuracy for both the first and the second waves.

Keywords: Logistic equation with constant immigration; Logistic equation; COVID-19;

Mathematical modeling; epidemic..
2010 Mathematics Subject Classification : 37N25, 62J05, 62J12 .

Introduction
Recent Coronavirus outbreaks, which have spread across all countries, have aroused considerable
interest in mathematical models that can describe epidemic dynamics and predict their progres-
sion [1, 2, 3, 5]. In [1] Pelinovsky et al. have modelled the growth dynamics of COVID-19 in
different countries for the first wave using the logistic equation given by (157), which was intro-
duced first by Verhulst [4] in 1838.

dN N
= rN (1 − ), (157)
dt N∞

Where
r: The infected rate.
N: The current number of infected.
N∞ : The total number of infected persons .

154
CNMA’2022
Second National Conference on
Mathematics and its Applications

The solution of this equation with constant coefficients can be easily found in the form :

N0 N∞ exp(rt)
N (t) = ,
N∞ + N0 [exp(rt) − 1]

Logistic equation with constant immigration

In this part we presents the logistic equation with constant immigration rate given by (158) and
estimate its coefficients using linear regression technique after performing a suitable variable
change that transforms the original nonlinear model into a linear one, for the six countries: Italy,
Japan, Algeria, Iraq, and Qatar, using data from the Center of Systems Science and Engineering
(CSSE) and WHO reports.

dN N
= rN (1 − ) + b, (158)
dt N∞
Where
r: The infected rate.
b: The number of infected remainings from the previous wave.
N: The current number of infected.
N∞ : The total number of infected persons .

Due to the fact that medical statistics works with cases per day, it is important to introduce the
difference logistic equation

dN N
Kn = Nn+1 − Nn = = rNn (1 − n ) + b,
dt N∞
We get a straightforward relationship between the number of cases each day (K) and the total
number of cases N (t) after removing the index n.

dN N
K= = rN (1 − ) + b, (159)
dt N∞

In order to estimate its coefficients using linear regression technique we perform the variable
change X = N (1 − NN ), so the nonlinear equation (159) is transformed to the following linear
∞
equation
K = rX + b,
and the parameters r and b are determined by

cov(X, K)
r= ,
var(X)

b = K̄ − r X̄,

The table.3 and table.4 illustrate the results of calculating the model parameters in the first and
second waves of Coronavirus infection for six countries: Italy, Russia, Japan, Algeria, Iraq, and
Qatar, using the logistic equation and the logistic equation with constant immigration.
The coefficient of determination R2 is calculated using the equation
Pn
(Ki − K̃i )2
R = 1 − Pi=1
2
n 2
,
i=1 (Ki − K̄i )

155
CNMA’2022
Second National Conference on
Mathematics and its Applications

Model logistique equation Logistic equation with constant immigration

N country N∞ r R2 r c R2
1 Italy 215855 0.1 0.74 0.083 732.4 0.80
2 Russia 895253 0.046 0.4 0.03 2771 0.68
3 Japan 15306 0.142 0.83 0.13 30.8 0.84
4 Algeria 53379 0.037 0.70 0.034 29.59 0.71
5 Iraq 588454 0.031 0.72 0.023 848.4 0.86
6 Qatar 110845 0.064 0.92 0.06 92.92 0.94

Table 3: Model and determination coefficients for two considred approximation wave 1

Model logistique equation Logistic equation with constant immigration

N country N∞ r R2 r c R2
1 Italy 1730948 0.082 0.67 0.061 6448 0.79
2 Russia 3309388 0.035 0.63 0.023 7336,07 0.95
3 Japan 319099 0.077 0.55 0.053 1434 0.76
4 Algeria 41382 0.106 0.78 0.084 165.1 0.86
5 Iraq 518770 0.058 0.16 0.033 2422.3 0.56
6 Qatar 43899 0.095 0.33 0.056 322.9 0.92

Table 4: Model and determination coefficients for two considred approximation wave 2

Where
Ki : is the number of infected persons on the day i.
K̃i : is the model value.
K̄i : is the average value.

References
[1] E. Pelinovsky, A. Kurkin , O. Kurkina, M. Kokoulina , A. Epifanova. Logistic equation and
COVID-19. Chaos, Solitons and Fractals 2020;140:110241.

[2] E. Pelinovsky, M. Kokoulina, A. Epifanova, A. Kurkin , O. Kurkina, [Link], [Link],

M. Kieillin. Gompertz model in COVID-19 spreading simulation. Chaos, Solitons and
Fractals (2021).

[3] M.V. Kokoulina, A. Epifanova, E. Pelinovsky, O. Kurkina, A. Kurkin. Analysis of coro-

navirus dynamics using the generalized logestic model. Transactions NNSTU n.a. R.E.
Alekseev 2020;vol. 3, pp. 28–41.

[4] P. F. Verhulst. Notice sur la loi que la population suit son acroissement . Correspondence
Mathematique et Physique (Ghent)(1838),Vol. 10, pp. 113-121.

[5] L. Kaihao. Mathematical model of infection kinetics and its analysis for COVID-19, SARS
and MERS. Infect Genet Evol (2020).

156
CNMA’2022
Second National Conference on
Mathematics and its Applications

A new approach for solving an optimal

control problem using parametrization
technique

Ikram Boukhelkhal 1 [Link]@[Link]

Rebiha Zeghdane 1 [Link]@[Link]
1 Department of mathematics , Bordj Bou-Arreridj University

Abstract: In practice, many optimal control problems are subject to constraints in state and/or
control variables. In direct methods, the optimal solution is obtained by direct minimization
of the performance index, subject to constraints. To identify the best solution to optimal
control problems, many numerical methods have been introduced. Direct methods are based
on the transformation of the original optimal control problem into a nonlinear programming
problem (NLP) by discretizing or parametrizing the state and/or control variables and then
solving the resulting NLP problem. They can be classified into three different approaches.
The first approach is based on state parameterization only. The second approach is control
parametrization and its idea is to approximate the control variables and obtain the state
variables by integrating the state equations. The third approach is based on state and control
variable parametrization. In order to obtain a solution to an optimal control problem, a
numerical technique based on state-control parametrization method is presented. This
method can be facilitated by the computation of performance index and state equation via
approximating the state variable as a function of time by using some basis polynomials. The
convergence of this method is inverstigated. Several numerical examples are presented to
confirm the analytical findings and illustrate the efficiency of the proposed method.

Keywords: Optimal control, State-control parametrization, Basis polynomials, numeri-

cal method.

2010 Mathematics Subject Classification: Primary 90CXX, 65KXX, 49NXX

Introduction
Optimal control problem can be considered as a generalization of the classical calculus of vari-
ation. The essential parts of an optimal control problem are, a mathematical system to be con-
trolled, a desired output of the system, a set of admissible inputs and a performance index or a
cost functional that measures the effectiveness of a given control operation. State parametrization
converts the problem to a non-linear optimization problem and finds (n + 1) unknown polyno-
mial coefficients of degree, at most, n. In this work, an efficient iterative algorithm is obtained. In
this way, only one unknown coefficient is calculated for finding a suitable approximation; further
iterations leads to favorable accuracy . In addition, by the proposed algorithm, the control and
state variables can be approximated as a function of time.

157
CNMA’2022
Second National Conference on
Mathematics and its Applications

Let given the following optimal control problem

U (s) = F(s, X(s), Ẋ(s)), (160)

subject the boundary conditions

X(t0 ) = x0 , X(t1 ) = x1 ,
where the function F is assumed to be continuously differentiable, the vectors X(.) : [t0 , t1 ] −→
R and U (.) : [t0 , t1 ] −→ R given in equation (160) are called the optimal trajectory and optimal
control, respectively. The aim here is to find an optimal control U (t) such that the cost functional
J in the following equation (161) is minimum
Z t1
J= L(s, X(s), U (s))ds. (161)
t0

Main results
The principle idea of state parametrisation is to approximate only the state variable of the sys-
tem by a sequence of given functions with unknown parameters, as :
n
X
Xn (s) = ai Φi (s), n = 1, 2, · · ·,
i=0

which can be used with different basis functions. By this method the problem of optimal control
is converted into a mathematical optimization problem.
The convergence of the parametrization technique is based on Weierstrass theorem of approxi-
[Link] can also see the effeciency and accuracy of our algorithm by giving some numerical
examples. The proposed algorithm is compared with some numerical tehniques existing in liter-
ature.

References
[1] Jaddu, Hussein M., Numerical Methods for solving optimal control problems using chebyshev
polynomials,Thesis, School of Information Science, Japan Advanced Institute of Science
and Technology, (1998) 122.

[2] Kafash, Behzad and Delavarkhalafi, Ali and Karbassi, Seyed-Mehdi., Application of
Chebyshev polynomials to derive efficient algorithms for the solution of optimal control prob-
lems,Scientia Iranica,Elsevier 19 (3) (2012),795–805.

[3] Kafash, Behzad and Delavarkhalafi, Ali and Karbassi, Seyed-Mehdi. , A numerical ap-
proach for solving optimal control problems using the Boubaker polynomials expansion scheme,
J. Interpolat. Approx. Sci. Comput, 3, (2014) 1–18.

[4] Ouda, Eman Hassan., The Efficient Generalized Laguerre Parameterization for Quadratic Op-
timal Control Problem, Journal of College of Education, Al-Mustansyriah University, 3(
1812-0380),(2014) 263–276.

158
CNMA’2022
Second National Conference on
Mathematics and its Applications

Control and stabilisation of a jerk system

using the fractional Routh-Hurwitz criteria

Chettouh Besma 1 chettouhbesma0707@[Link]

Menacer Tidjani 1
1 Department of mathematics , University of Mohamed khidher Biskra

Abstract: Many dynamic systems are better characterized by a dynamic fractional order
model, generally based on the notion of differentiation or integration of integer-order. In this
work we choose "The Jerk System" to apply the theories of stability and control using the
generalized Routh-Hurwitz criterion to fractional order.

Keywords: dynamic systems, fractional order, Routh-Hurwitz criterion , The Jerk Sys-
tem, stability, control.

Introduction
The fractional calculus is more than 300 years old with the first written note dated to 1695 [1].
Several physical phenomena can be described more accurately by fractional differential equations
rather than integer-order models.

Main results
The results obtained in this work show the effect of the fractional order on the control, which
proves the effectiveness of the method applied to distinguish the fractional case and that of the
whole case and to underline the importance of the control of the fractional systems, those systems
that have proven to be more accurate than its whole order counterparts.

References
[1] DUBOIS, Francois, GALUCIO, Ana Cristina, et POINT, Nelly. Introduction à la dériva-
tion fractionnaire-Théorie et Applications. 2010 .

[2] [Link], Tarek. Analyse du chaos dans un système d’équations différentielles frac-
[Link]èse de doctorat .[Link]. 2014 .

[3] TIDJANI, Menacer. Synchronisation des systèmes dynamiques chaotiques ‘a dérivées

[Link]èse de doctorat .[Link]. Constantine ,2013.

159
CNMA’2022
Second National Conference on
Mathematics and its Applications

ESTIMATION OF GENERALIZED LINEAR

FAILURE RATE DISTRIBUTION IN BAYESIAN

DJEMOUI Nour El Houda 1 hou17da@[Link]

CHADLI Assia 1
1 Department of mathematics , Badji Mokhtar Annaba

Abstract: The exponential and Rayleigh distributions are the two that are most frequently
used to evaluate lifetime data. These distributions have a lot of compelling physical explana-
tions and favorable properties. Unfortunately, compared to the exponential distribution, the
Rayleigh distribution only has a constant failure rate. We look at the dependability charac-
teristics of the generalized linear failure rate distribution and the estimate of the unknown
parameters. The Bayes estimators and associated hazards are produced using a gamma prior
distribution, complete data, and multiple loss functions. A simulation research was carried
out to ascertain all the results. In order to illustrate our methods, we conclude by analyzing a
set of data from the real world.

Keywords: : Reliability, Estimation, Bayesian analysis, GLFR distribution, Posterior dis-

tribution.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
The exponential, Rayleigh, linear failure rate, and modified exponential distributions are fre-
quently used in lifetime data analysis. In contrast to Rayleigh, linear failure rate, and generalized
exponential distribution, which can only have monotone (increasing in the case of Rayleigh or
linear failure rate and increasing/decreasing in the case of generalized exponential distribution)
hazard functions, it is well known that exponential distributions can only have constant hazard
functions. Unfortunately, non-monotonic functions like bathtub-shaped hazards must frequently
be taken into account in practice, for example, see Lai et al. In this study, we provide a novel sim-
ple distribution that generalizes a number of well-known distributions, such as the conventional
linear failure rate distribution, and may have a bathtub-shaped hazard function.

Methodology and survival properties

Let X be a random variable with the following CDF for α > 0, β > 0 and θ > 0 as follows:

!#θ
x2
"
F(x) = 1 − exp −(αx + β )
2

Here θ is shape parameter. The distribution of this form is said to be a generalized linear failure
rate distribution with parameters a, b, θ and will be denoted by GLFRD(α, β, θ). The PDF and the

160
CNMA’2022
Second National Conference on
Mathematics and its Applications

reliability function of GLFRD(α, β, θ) will be:

!#θ−1
x2 x2
!"
f (x) = θ (α + βx) exp −(αx + β ) 1 − exp −(αx + β )
2 2

And !#θ
x2
"
R(x) = 1- 1 − exp −(αx + β )
2

It is observed, see Gupta and Gupta, that the reversed hazard function plays an important role in
the reliability analysis. The hazard function of the GLFRD (α, β, θ) is:

 2
h  2
iθ−1
θ (α + βx) exp −(αx + β x2 ) 1 − exp −(αx + β x2 )
H(x) = h  iθ (162)
2
1- 1 − exp −(αx + β x2 )

It is well known that the hazard function or the reversed hazard function uniquely detrmines the
corresponding probabililty density function. From (162) it is clear that the GLFRD(α, β, θ) is a
proprtional reversed hazard family. It may be mentioned that the reversed hazard function is a
decreasing function.

References
[1] M. V. Aarset, How to identify bathtub hazard rate, IEEE Transactions on Reliability. 36
(1987), 106 –108.

[2] L. J. Bain, Analysis for the Linear Failure-Rate Life-Testing Distribution , Technometrics, 16
(1974), 551–559.

[3] R. E. Barlow, R. Campo, Total time on test processes and applications to failure data anal-
ysis,Reliability and Fault Tree Analysis (Barlow, Fussell, Singpurwalla, Eds,(1975), 451–
481.

[4] B. Bergman, B. Klefsjo, A graphical method applicable to agereplacement problems, IEEE

Transactions on Reliability,31, (1982), 478–481.

[5] B. Bergman, B. Klefsjo, The total time on test concept and its use in reliability theory, Oper-
ations Research, 32,(1984),596–606.

[6] I.W. Burr, Cumulative frequency function, Annals of Mathematical Statistics, 13,(1942),
215–232.

[7] M.E. Ghitany, Reliability properties of extended linear failure-rate distributions, Probability
in the Engineering and Information Sciences, 21,(2006), 441–450.

161
CNMA’2022
Second National Conference on
Mathematics and its Applications

Analysis of M/D/1/N /N Retrial Queue

Using Deterministic and Stochastic Petri
Net

Ikhlef Lyes 1 [Link]@[Link]

1 University of Algiers, Faculty of Sciences
Research Unit LaMOS (Modeling and Optimization of Systems)

Abstract: We study a finite source retrial queue with deterministic service times using
an approach based on the theory of Markov Regenerative Process. A detailed Deterministic
Stochastic Petri Net (DSP N ) model which copes with the complexity of this queue is given. For
the steady state of this model, we construct the one step transition probability matrix of em-
bedded Markov chain and the conversion matrix. As an example the retrial system M/D/1/2/2
is detailed. We establish an algorithm in Matlab environment based on the theoretic results
obtained in order to compute efficiently various performance measures and to study the effect
of system parameter’s on the characteristics of the DSP N models the retial queue M/D/1/N /N .

Keywords: Retrial Systems, Deterministic and Stochastic Petri Nets, Embedded Markov
Chain, Steady State.

Introduction
The Standard queueing systems with deterministic service times are widely found in literature
see Bunday [3],... . Brun and Garcia [4] give an analytical solution of the system M/D/1/K. Franx
et al. study the multi-server system M/D/c. Madan and Saleh study the queue M/D/1 with
general vacations. Choi et al. investigate the transient and sensitivity analysis of Determinis-
tic and Stochastic Petri Nets (DSP N ), as application, they detailed the analysis of the classical
queue M/D/1/2/2 and M/D/1/2/2 with vacation. However, little attention has been paid to re-
trial queues with deterministic service times. Wu and Ke, consider a infinite single server retrial
queueing system in which each customer (primary or retrial customer) has discrete service times.
For bibliographies on retrial queues, see [2] and the references therein. The DSP N class intro-
duced by Ajmone and Chiola. An analytic method based on the Markov Regenerative Process
(MRP ) theory discussed by Choi in 1985 [5]. The underlying stochastic process of a DSP N is a
MRP , with the restriction that at most one deterministic timed transition is enabled in each mark-
ing. In this paper, we give a model and performances analysis of finite source retrial system with
deterministic service times M/D/1/N /N by using DSP N tool.

Model description
We study the M/D/1/N /N retrial queue, in which primary customers arrive according to a Poisson
process with rate λ. If an arriving customer finds the server idle, he obtains service immediately
and joins the source after service completion. Otherwise, if the server is occupied, the arriving

162
CNMA’2022
Second National Conference on
Mathematics and its Applications

primary customer enters to the orbit. The policy of access from the orbit to the server is governed
by an exponential law with rate kγ, where k is the number of customers in the orbit. Each cus-
tomer has a deterministic (constant) service times of length τ > 0. The Fig. 9 shows the DSP N
model describing the M/D/1/N /N retrial queue.

Figure 3: DSP N models retrial queue M/D/1/N /N with classical retrial policy.

Embedded Markov Chain

We give the steady state analaysis of the DSP N models the retrial queue M/D/1/N /N . The states
space of the DSP N depicted in Fig. 9 is given by

Ω = {M2k = (k, 0), M2k+1 = (k, 1) : 0 ≤ k ≤ N − 1}.

Let Mi , Mj ∈ Ω, the one step transition probability matrix P = [PMi Mj ] is given by:




 1, if i = 0 and j = 1;






j−i+1

j−i+1 j+2

CN −2 i+1 (1 − e−λτ ) (e−λτ )N −

,

 2 2



 2

if 0 ≤ k ≤ N − 1, i = 2k + 1 and i − 1 ≤ j ≤ 2N − 2, j = 2k;











 i
2γ

,


 i i
PMi Mj = 2 γ+(N − 2 )λ


if 1 ≤ k ≤ N − 1, i = 2k + 1 and j = i − 1;











(N − 2i )λ


,

 i i
2 γ+(N − 2 )λ





if 1 ≤ k ≤ N − 1, i = 2k + 1 and j = i + 1;














 0, otherwise.

163
CNMA’2022
Second National Conference on
Mathematics and its Applications

and the conversion matrix C = [cMi Mj ] is given by:





 1, if i = j = 0;






j−i

j−i j+1
1 τ
 R
−λt ) 2 (e−λt )N − 2 dt,
 2
τ 0 CN − i+1 (1 − e





 2

if 0 ≤ k < N − 1, i = 2k + 1 and i ≤ j ≤ 2N − 1, j = 2k + 1;


cMi Mj =








1, if 1 ≤ k ≤ N − 1, i = 2k, and j = i;













0, otherwise.



Numerical results

In this section, we give some numerical results concern the DSP N associated to M/D/1/N /N
using the algorithm that we establish in Matlab environment. In Tab. 5, the DSP N model pro-
posed for the queue M/D/1/N /N with classical retrial policy, is validated by the exact numerical
results given in [1]. We see that the performance indices corresponding the DSP N associated to
M/D/1/N /N queue are close to those obtained in [1].

Table 5: Comparison of stationary distributions of the model M/D/1/N /N retrial queue given in
[1], with the DSP N model ,“N = 11, λ = 0.01, γ = 5.2, and τ = 15".

i p(0,i) p(1,i) π(0,i) π(1,i)

0 0.01010 0.03869 0.0101041 0.0386973

1 0.00074 0.09112 0.0007442 0.0911271
2 0.00078 0.15838 0.0007886 0.1583811
3 0.00081 0.21048 0.0008122 0.2104897
4 0.00070 0.21147 0.00070839 0.2114737
5 0.00048 0.15663 0.00048802 0.1566328
6 0.00025 0.08251 0.00025101 0.0825136
7 0.00009 0.02935 0.00009067 0.0293532
8 0,00002 0.00650 0.00002117 0.0065053
9 0.27 × 10−5 0.00077 0.27 × 10−5 0.0007788
10 0.14 × 10−6 0.00003 0, 14 × 10−6 0.0000362

164
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] J. R. Artalejo, A. Gomez-Corral, Information theoretic analysis for queueing systems with
quasi-random input. Mathematical and Computer Modelling, vol. 22, pp. 65-76, (1995).

[2] J. R. Artalejo, Accessible bibliography on retrial queues: Progress in 2000-2009.

[3] B. D. Bunday, Basic queueing theory, Edward Arnold, Australia, (1986).

[4] O. Brun, & J. Garcia, Analytical solutions of finite capacity M/D/1 queues, J. Appl. Probab.,
vol. 37, pp. 1092-1098, (2000).

[5] H. Choi, V. G. Kulkarni, K. S Trivedi, Transient analysis of deterministic and stochas-

tic Petri nets, The 14th International Conference on Application and Theory of Petri Nets,
chicago, U.S.A., pp. 21-25, (1993).

[6] Ikhlef, L., Lekadir, O., and Aissani, D. (2016). MRSPN analysis of semi-markovian finite
source retrial queues. Ann. Oper. Res., 247, 141-167.

[7] T. Yang, J. G. C. Templeton, A survey on retrial queue, Queueing Systems, vol. 2, pp.
201-233, (1987).

165
CNMA’2022
Second National Conference on
Mathematics and its Applications

Estimation d’un Mélange de Modéles

GARCH Périodiques

Rokia Hemis 1 hemisrokia@[Link]

Hafida Guerbyenne 2 hguerbyenne@[Link]
Fayçal Hamdi 2 fhamdi@[Link]
1 Faculté des Sciences, Université de Sétif
2 Faculté de Mathématiques, Université USTHB

Abstract: Nous nous intéressons dans notre travail à une nouvelle classe de modèle des séries
chronologiques les modèles mélange GARCH périodiques qui est une extension des modèles
mélange ARCH périodiques. Dans la première partie de notre travail nous présentons cette
nouvelle classe de modèles ainsi ses propriétés probabilistes. Dans la deuxième partie nous
nous intéressons au problème de l’estimation des paramètres de ces modèles nous proposons
l’algorithme du griddy Gibbs où on détermine le noyau de la loi a posteriori conditionnelle de
chaque paramètre du modèle nécessaire pour l’application de l’algorithme.

Keywords: Séries chronologiques, Analyse bayesienne, Modèles GARCH, Modèles péri-

odiques, Algorithme du Griddy Gibbs.
2010 Mathematics Subject Classification: 62.

Introduction
Il est bien connu que les modèles les plus populaires et les plus utilisés dans la modélisation de la
volatilité instantanée dans les séries chronologiques financiéres sont les modéles Autorégressifs
Conditionnellement Hétéroscédastiques ARCH, introduit par Engle [5] et leur éxtention général-
isés de Bollerslev [3] et en suite les modéles GARCH périodiques une classe de modéles introduite
par Bollerslev et Ghysels (1996) qui ont montré une grande capacité à capturer la périodicité dans
la variance conditionnelle. Divers modèles ont été proposés afin de capturer différentes caracter-
istiques telles que la longue mémoire, le changement de regime et périodicité dans la variance
conditionnelle.
Notre but est de proposer un modéle qui peut presenter des series chronologiques avec une struc-
ture d’autocorrrélation périodique aussi bien que d’autres caractéristiques (telles que la multi-
modalité, le changement de régime) nous proposons la classe de mélange des modèles GARCH
périoqiques notée: MP GARCH.

Mélange de modéles GARCH périodiques

Définition 1 :
Un processus stochastique {yt , t ∈ Z} suit un mélange de K modèles autorégressifs condition-
nellement hétéroscédastiques périodiques, de période S et d’ordres p1 , p2 , . . . , pK ; q1 , q2 , . . . , qK , noté
MP GARCHS (K; p1 , p2 , . . . , pK ; q1 , q2 , . . . , qK ), s’il est défini par:

166
CNMA’2022
Second National Conference on
Mathematics and its Applications

  
y 
  
F (yt |Ft−1 ) = K
 P
λk Φ  q t  , t ∈ Z,
 
k=1


 (k)

 ht
(163)

(k) (k) P qk (k) 2 Ppk (k)
ht = ωt + i=1 + j=1 k = 1, ...., K.




 αt,i yt−i βt,j ht−j
(k)

 ht = K λk ht

 P
k=1

où Φ (.) et F (.|Ft−1 ) sont réspéctivement, la distribution cumulative de la loi normale centrée

reduite et de la distribution cumulative conditionnelle de yt , sachant les valeurs passées du pro-
(k) (k) (k)
cessus jusqu’au temps t − 1. Les paramétres ωt , αt,i et βt,j sont périodiques en t, de période S,
(k) (k) (k) (k) (k) (k)
i.e., ωt+Sn = ωt , αt+Sn,i = αt,i , i = 1, ..., qk et βt+Sn,j = βt,j , j = 0, ..., pk et t ∈ Z. Pour exclure la pos-
sibilité d’avoir des variances conditionnelles négatives ou nulles, les parametres doivent verifiés
(k) (k) (k)
les conditions suivantes: ωt > 0, αt,i ≥ 0, i = 0, ..., qk , k = 1, ..., K et t ∈ Z , et βt,j ≥ 0, j = 0, ..., pk ,
k = 1, ..., K et t ∈ Z . les constantes λk , k = 1, ..., K sont des constantes réelles strictement positives
tel que K
P
k=1 λk = 1, avec λk > 0. Nous pouvons écrire également (163) en fonction de sa fonction
de densité.

Propriétés Probabilistes
Dans cette partie de notre travail, nous allons donner la condition de stationnarité stricte du
modèle MP GARCH. D’apres (Zhang et al. [9], Boshnakov [4]) nous réecrivons notre modèle sous
la forme markovienne, ce qui nous permettra d’etudier la stationarité stricte et au second ordre et
l’existence des moments d’ordres supérieurs.

Yt = At Yt−1 + Bt , t∈Z

La stationnarité de ce modèle découle de l’existence d’une solution strictement stationnaire. L’outil

principal pour l’étude de la stationnarité stricte est le concept d’exposant de lyapounov, le plus
grand exposant de Lyapunov

1
γ S (A) = inf E [log kAnS AnS−1 · · · A1 k]
n∈N N

La condition de stationnarité périodique au second ordre est:

 S−1 
Y ⊗2 
ρ  AS−s  < 1
s=0

Estimation
Pour procéder à l’estimation par la méthode bayesienne des paramétres d’un modèle MP GARCHS ,
via l’algorithme de Griddy-Gibbs (Ritter and Tanner [6]). Cette méthode consiste à déterminer
les noyaux des lois a posteriori de chaque paramétre du modèle considéré, en utilisant l’analyse
bayesienne nous avons ainsi P (θ|y) ∝ L (θ|y) · P (θ) ,où P (θ|y), P (θ) et L (θ|y) sont réspectivement
les lois a posteriori, a priori et la fonction de vraisemblance du paramétre θ.
Dans notre travail les lois a priori des paramétres sont choisies comme suit:

• pour le parametre λ (les proporions du mélange) on propose la loi de Direchlet D(α).

167
CNMA’2022
Second National Conference on
Mathematics and its Applications

• pour les autre parametres les lois a priori sont choisies comme étant des lois uniormes sur
des intervalles qui verifient les conditions de stationnarité du processus {yt , t ∈ Z} .

Comme il est nécessaire de donner l’expression de la fonction de vraisemblance dans le cas des
modèles MP GARCH.
Fonction de vraisemblance

• y = (y1 , y2 , ..., yN S , Z1 , ..., ZN S ) et Z = (Z1 , ..., ZN S ) le vecteur contenant, respectivement,

 0
(1) (K)
les données complétes et les données manquantes, avec Zt = zt , ..., zt .
0  0
(k) (k) (k) (k) (k)

0 0 0 0 0 0
• Θ = λ0 , Θ1,1 , · · · , ΘS,1 , Θ1,2 , · · · , ΘS,2 , · · · ,Θ1,K , · · · , ΘS,K où Θs,k = ωs , αs,1 , ..., αs,q , βs,1 , ..., βs,p
et λ = (λ1 , ..., λK )0 .

(k)
S Y
K " − 21 ( 2
)#zs+Sn
 ys+Sn 0
(k)
Y
L (Θ) ∝ λk · hs+Sn0 exp − (k)
0
2hs+Sn
s=s0 k=1 0

(k)
S N K "
−1 Y  − 21 ( )#zs+Sn
2
(k) ys+Sn
Y Y
× λk · hs+Sn exp − (k)
2hs+Sn
s=1 n=n0 +1 k=1

(k)
soit τs+Sn la probabilité conditionnelle que l’observation yt soit generée par la k eme composante.

References
[1] M. C. Ausin, P. Galeano, Bayesian estimation of the Gaussian mixture GARCH model, Com-
put. [Link] Anal, 51, 2007, 2636-2652.

[2] M. Bentarzi , F. Hamdi, Mixture Periodic Autoregressive conditional Heteroscedastic Models

Comput. [Link] Anal, 53, 2008,1-16.

[3] T. Bollerslev, Generalized autoregressive heterscedasticity, J. Econometrics, 51, 1986, 307-

327.

[4] G. Boshnakov, On first and Secend Order Stationarity of Random Coefficient Model, Linear
Algebra Appl, 434, 2011, 415-423.

[5] R. F. Engle, Autoregressive Conditional Heteroskedasticity with estimates of variance of the

U. K inflation, Econometrica, 50, 1982, 987-1008.

[6] C. Ritter, M.A Tanner, Facilitating the Gibbs sampler: the Gibbs stopper and the Griddy–
Gibbs sampler J. Amer. Statist. Assoc, 87, 1992, 861–868.

[7] Q. Shao, Mixture periodic autoregressive time series models, Statist. Proba. Lett, 76, 2006,
609-618.

[8] C.S. Wong, W.K. Li, On a mixture autoregressive conditional heteroscedastic model., J. Amer.
[Link], 96, 2001, 982- 995.

[9] Z. Zhang, W.K. Li, K.C. Yuen, On a mixture GARCH time series model, J. Time ser. Anal.
27, 2006, 577-597.

168
CNMA’2022
Second National Conference on
Mathematics and its Applications

Refined descriptive sampling with

dependent variables

Siham KEBAILI 1 [Link]@[Link]

Megdouda OURBIH-TARI 2 ourbihmeg@[Link]
1 Laboratoire de Mathématiques appliquées, Faculté des Sciences Exactes, Université de Bejaia, Bejaia, 06000,
Algeria
2 Institut des Sciences, Centre Universitaire Morsli Abdellah de Tipaza, Tipaza, 42020, Algeria

Abstract: This paper deals with Monte Carlo simulation in case of dependent input random
variables. We propose an algorithm to generate refined descriptive samples from dependent
random variables for estimation of expectations of functions of output variables using the
Iman and Conover algorithm to transform the dependent variables to independent ones.
Therefore, such estimates obtained through a chosen mathematical model are compared with
those obtained using the simple random sampling method, which proved that the former are
the most efficient. Besides, using already published work on independent input variables, we
can deduce in case of dependent input random variables, that asymptotically the variance of
the RDS estimator is less than that of SRS estimator for any simulation function having finite
second moment.

Keywords: Simulation, Monte Carlo Methods, Variance reduction Iman and Conover
method.
2010 Mathematics Subject Classification: Primary 11K45, 60B12, 60G50.

Introduction
A mathematical model for the device is developed from which we can simulate the behavior of the
device on a computer. So experiments are carried out on the model built and unknown parameter
θ of the output random variable Y of interest denoted as the unknown but observable univariate
transformation of X given by the function Y = h(X) is estimated. Thus, we have the problem of
approximating θ. Since h(X) may be difficult to compute for each new value of X, it is important
to pick a sampling scheme that allows us to estimate h(X) well while keeping N , the number of
replication, to a minimum. There exist several procedures for choosing X, X2 , .., XN . The simplest
is Simple Random Sampling (SRS) also known as Monte Carlo (MC) is usually used for high-
dimensional problems. That is, N values of the input random vector X, X2 , .., XN are generated in
some manner such that the parameter θ = E(g(Y )) can be estimated by

N
1X
TSRS = T (X1 , X2 , .., XN ) = g(Yj ).
N
j=1

Main results

169
CNMA’2022
Second National Conference on
Mathematics and its Applications

Simulation of a problem which can be described by dependent variables but trans-

formed to independent ones through the IC method

Although we can use any simulation problem to compare RDS with SRS, we prefer to use a simple
problem already used by (Saliby, 1997) for such comparison such as g(Y ) = Y : The study of the
response variable

1
Y = (X12 + X22 ) 2 .

The behavior of the selected model depends on a random vector X = (X1 , X2 ) having the following
joint distribution
2e−x1 x2

if x1 ≤ 1 and x2 ≥ 0



x12

f (x1 , x2 ) = 




 0 otherwise
Our purpose is to simulate E(Y ) and Var(Y), incidentally, we have that

E(Y ) = 1.2533

and that
var(Y ) = 0.1842.
To simulate X, we first simulate the random variable X1 and then given the x1 observation, we
simulate the random variable X2 /X1 = x1 .

References
[1] Baiche, L. and Ourbih-Tari, M., Large sample variance of simulation using refined descrip-
tive sampling: Case of independent variables, Communications in Statistics - Theory and
Methods, 46 (2017), 510–519.

[2] Iman R. L. and Conover W. J., Distribution-free approach to inducing rank correlation among
input variables, Communications in Statistics – Computation and Simulation, 11 (1982),
311–334.

[3] Tari, M. and Dahmani, A., Refined descriptive sampling: a better approach to monte carlo
simulatin, Simulation Modeling Practice and Theory, 14 (2006), 143-160.

170
CNMA’2022
Second National Conference on
Mathematics and its Applications

Characterization of a certain admissible

trajectorie of Dolichobrachistochrone
differential game problem

Ghanem Aicha 1 [Link]@[Link]

Bouremani Touffik 2
Benterki Djamel 1
1 Laboratory of Fundamental and Numerical Mathematics, Ferhat Abbas University, Setif 1, Algeria
2 Laboratory of applied Mathematics , Ferhat Abbas University, Setif 1, Algeria

Abstract: We apply step by step manner, the theoretical dynamic programming algorithm
described in [3,4], to characterize a certain admissible trajectorie of Dolichobrachistochrone
problem [1,2,5].
We use a certain extension of Cauchy’s Method of characteristics for stratified Hamilton-Jacobi
equations to combine it with numerical procedures.

Keywords: Optimal control, differential inclusion, Hamiltonian flow, Dynamic program-

ming.
2010 Mathematics Subject Classification: Primary 49J15, 49L20, 34A60.

Introduction
The aim of this work is to apply step by step manner the dynamic programming algorithm in
[3,4], to characterize a certain admissible trajectorie as solution in the form of Hamiltonian maxi-
mal flow as well as to identify maximal interval in which some admissible conditions are satisfied.
"Dolichobrachistochrone" differential game formulated and studied by Isaacs [1] and studied in
the same rather heuristical way by Basar and Olsder [5] and Chigir [2],...,etc. We use a certain
extension of Cauchy’s method of characteristics (see [4]) for stratified Hamilton-Jacobi equations
to describe a large set of admissible trajectories as solutions in the form of maximal flows. Also,
due to the complexity of the involved functions, we will combine these results with numerical
procedures.

Main results
Dynamic programming Formulation

Problem 1. Given w > 0. Find:

infsup C(y, u(.), v(.)), ∀y ∈ Y0 ,

u(.) v(.)

171
CNMA’2022
Second National Conference on
Mathematics and its Applications

subject to:
RT
C(y, u(.), v(.)) = g(x(T )) + f0 (x(t), u(t), v(t))dt,
0
x0 (t) = f (x(t), u(t), v(t)), x(0) = y, a.e.([0, T ]),
u(t) ∈ U (x(t)) a.e.([0, T ]), v(t) ∈ V (x(t)) a.e.([0, T ]),
x(t) ∈ Y0 , ∀t ∈ [0, T ), x(T ) ∈ Y1 , T fixed.

Defined by the following data:

√ √ 
f (x, u, v) = x2 u1 + w2 (v + 1), x2 u1 + w2 (v − 1) , f0 (x, u, v) = 1, g(ξ) = 0, ∀ξ ∈ Y1 ,
n o
U = u ∈ R2 ; kuk = 1 , V = [−1, 1]
Y0 = (0, +∞)2 , Y1 = {0} × (0, +∞)

Characterization of the Hamiltonian and the set of extremal points.

The “pseudo-Hamiltonian” is given in our case by:
√
H(x, p, u, v) = x2 hp, ui + w2 [(p1 + p2 )v + p1 − p2 ] + 1, ∀p ∈ R2

Proposition 13. The Hamiltonian and the corresponding marginal multifunction (see [3,4]) are given
by the formulas: ∀(x, p) ∈ dom(H(., .)) = Z :
√
H + (x, p) = H − (x, p) = − x2 kpk + w2 [|p1 + p2 | + p1 − p2 ] + 1,
 
p




 − kpk , 1 , if h(x, p) = p1 + p2 > 0,

  p
 
F̂(x, p) = F̂+ (x, p) = F̂− (x, p) =  − , −1 , if h(x, p) < 0,
n kpk


 − p ×V

 o

kpk
if h(x, p) = 0,

The Hamiltonian and Z are C1 -stratified by the stratification SH = {Z+ , Z− , Z0 } defined by:

Z+ = {(x, p) ∈ Z; h(x, p) > 0} ,

Z− = {(x, p) ∈ Z; h(x, p) < 0} ,
Z0 = {(x, p) ∈ Z; h(x, p) = 0} .

If we denote by: H± (., .) = H(., .) |Z± , H0 (., .) = H(., .) |Z0 then it follows:

√
H+ (x, p) = − x2 kpk + wp1 + 1, if (x, p) ∈ Z+
√
H− (x, p) = − x2 kpk − wp2 + 1, if (x, p) ∈ Z−
√
H0 (x, p) = − 2x2 |p1 | + wp1 + 1, if (x, p) ∈ Z0

Set of terminal transversality points Z ∗ (see [3,4]) in our case is given by:

1
  
Z+∗ = (0, s2 ), (q1 , 0) ; q1 = ,s > w ,
s−w

Generalized Hamiltonian and characteristic flow

172
CNMA’2022
Second National Conference on
Mathematics and its Applications

The Hamiltonian system on the stratum Z+ (for which h(x, p) = p1 + p2 > 0) :

 √ p



 x10 = − x2 kpk1 + w, x1 (0) = 0,
 √ p
x20 = − x2 kpk2 , x2 (0) = s2 ,




(164)

1
p10 = 0,

p1 (0) = s−w ,





kpk

p20 =


 √ ,
2 x2
p2 (0) = 0,

The solution of system (164) in the form of maximal flows is given by the formulas:
 2
 
X1+ (t, s) = − s2 sin st + t(w − 2s ),





  
 X2+ (t, s) = s2 cos 2st ,



1
p1+ (t, s) = s−w

, s > w,





  
 + 1
 p2 (t, s) = s−w tan 2s ,
 t

−πs 2πw
Lemma 21. if τ0 (s) = 2 ,s ≥ s0 = π+2 then :

1. X1+ (t, s) ∈ Z+ , ∀t ∈ (τ0 (s), 0) ,

2. X1+ (τ0 (s), s0 ) = 0,

3. X1+ (τ0 (s), s) > 0, ∀s > s0

Lemma 22. The set

G = X + (τ0 (s), s) = (X1+ (τ0 (s), X2+ (τ0 (s)), s ∈ [s0 , +∞)


is a parabolic curve.

Numerical results

We present two examples to illustrate the efficiency of the dynamic model in the representation
of trajectories and the positivity of the corresponding guiding function h+ (., .). For the implemen-
tation we use MATLAB 2010 executed on a Core i3-380M PC (2.53 GHz).

173
CNMA’2022
Second National Conference on
Mathematics and its Applications

1. The admissible trajectories. 2. The guiding function h+ .

2πw
Figure 4: The data: w = 1, s = 3 so s0 = π+2 = 1.222, (s > s0 )

1. The admissible trajectories. 2. The guiding function h+ .

2πw
Figure 5: The data: w = 2, s = 5 so s0 = π+2 = 2.444, (s > s0 )

Concluding Remarks

Isaacs assumed that the line y = w2 is a barrier of trajectories but our approach based on dy-
namic method finds another points and extends the curve paths under the line y = w2 to complete
the coverage of the region Z+ .

References
[1] R. Isaacs, Differential games, Dover Publications Inc, New York, 1964.

[2] S.A. Chigir, The game problem of the dolichobrachistochrone, PPM. 40 (1976), 1003–1013

[3] Şt. Mirica, User’s guide on dynamic programming for autonomous differential games and op-
timal control problems, Rev Roumaine Math. 49 (2004), 501–529.

[4] Şt. Mirica, Constructive dynamic programming in optimal control, Editura Academiei Ro-
mane, Bucureşti, 2004.

[5] T. Basar, G.J. Olsder, , Dynamic noncooperative game theory, Academic Press Inc, London,
1982.

174
CNMA’2022
Second National Conference on
Mathematics and its Applications

Modélisation mathématique de
l’expansion géographique de la maladie
de Chagas

BENARBIA Narimene 1 [Link]@[Link]

MAHDJOUB Tewfik 1 [Link]@[Link]
1 Département de Mathématiques , Université ABOU BAKER BELKAID, Tlemcen

Résumé: L’expansion géographique de la maladie de Chagas est due essentiellement à ses

vecteurs qui sont des punaises de la famille des triatomines. Les processus démographique et
de dispersion spatiale de ces derniers sont décrits par un système de réaction-diffusion sur un
domaine composé de deux sous-domaines contigus où le premier est habité par les individus
sains et le second par les individus infectés. Par un changement de variables approprié, le
système est réduit à une équation différentielle abstraite. L’objectif est d’étudier l’existence et
l’unicité de la solution du système obtenu (problème bien posé ).

Mots-Clefs: Maladie de Chagas, Equations de réaction-diffusion, Equations d’assymétrie.

2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
La maladie de Chagas est un problème de santé publique dans les pays d’Amérique Latine où
elle touche, selon les statistiques de l’OMS de l’année 2010, près de 6 millions de personnes avec
une prévalence pour 100 habitants de 1.055 et 30 000 nouveaux cas estimés chaque année.
En raison de l’intense mobilité humaine, son expansion géographique a atteint les pays d’Amérique
du nord, d’Europe, d’Australie et d’Asie, ramenant ainsi le nombre d’infectés à quelque 10 mil-
lions à travers le monde. Pour des raisons liées à la colonisation, c’est l’Espagne qui en compte le
plus grand nombre avec 17000 infectés estimés. Aucune statistique sur la prévalence de la mal-
adie n’est donnée pour le continent africain. Cependant les échanges massifs entre l’Espagne et
ses proches voisins d’Afrique du nord permettent de douter que la maladie est bien présente dans
ces pays sans être toutefois diagnostiquée.
La maladie de Chagas ou trypanosomiase américaine est une maladie à transmission vectorielle
causée par le parasite protozoaire flagellé Trypanosoma cruzi (T. cruzi). Ce parasite peut infecter
un large éventail d’hôtes mammifères (dont l’homme ) ou oiseaux domestiques ou sauvages.
Le principale mode de transmission est le contact avec des insectes vecteurs appelées triatominae
ou de punaises. Lorsque ces dernièrs vivent dans des abris voisins des habitations humaines, on
dit que la transmission du parasite est domestique. Si elles vivent dans les nids de mammifères
ou d’oiseaux non domestiqués, la transmission est sylvatique.
Au cours du cycle d’infection typique, une triatomine infectée prélève le repas de sang d’un hôte
et libère le parasite dans ses selles près du site de l’infection.

175
CNMA’2022
Second National Conference on
Mathematics and its Applications

Le parasite pénètre dans l’hôte par une plaie ou une muqueuse intacte, telle que la conjonctive.
C’est ce qu’on appelle la transmission stérocorarienne. Un hôte infecté transmet également le
parasite à un vecteur sain pendant les repas de sang et le parasite réside alors dans l’intestin du
vecteur.
La transmission chez les hôtes peut également se produire par voie congénitale (de la mère infec-
tée à l’enfant) et par voie orale.
La lutte antivectorielle est restée la principale stratégie pour contrer la propagation de la mal-
adie. Elle a été basée sur la pulvérisation régulière d’insecticide dans les villages atteints. Cette
lutte menée conjointement par plusieurs pays d’Amérique Latine dans le cadre de projets inter-
gouvernementaux coûtaient annuellement 30 millions de US $ sans pouvoir atteindre les objectifs
fixés par l’OMS. En effet, dans des zones où la présence du vecteur a été drastiquement réduite
ont vu l’installation d’un processus de réinfestation par les triatomines.
L’objectif de la modélisation mathématique est donc d’essayer d’expliquer ce phénomène puis de
déterminer les principaux paramètres démographiques qui font qu’une invasion de triatomines
réussisse afin de permettre aux décideurs de contrôler ces paramètres.

Approche de modélisation
Le cycle biologique des Triatominae se compose de sept stades de développement : un stade œuf,
cinq stades larvaires et un stade adulte. Pour simplifier l’étude, le développement de l’œuf au
cinquiéme stade larvaire est considéré comme un seul stade qui sera appelé stade juvénile [2]. On
note:

• J(t, x, y) : La densité des juvéniles au temps t au point (x, y).

• A(t, x, y) : La densité des adultes.

• H(t, x, y) : La densité des hôtes.

On considère deux états sanitaires:

• Les individus sains, non porteurs du parasite [Link], de densité JS , AS , HS .

• Les individus infectés, porteurs du parasite, de densité JI , AI , HI .

Démographie

Entre t et t + dt, les juvéniles ayant survécu jusqu’à t avec une probabilité σJ passeront au stade
adulte avec une probabilité τJ ou resteront juvéniles avec une probabilité (1 − τJ ). Les adultes qui
ont survécu avec une probabilité σA vont pondre des œufs avec un taux de fécondité fA .
Les mammifières hôtes ont un taux de survie σH et une fécondité fH . (figure 6)
Les paramètres démographiques sont supposés dépendre du temps mais pas de l’espace.

Transmission verticale de [Link]

La maladie n’est pas génétiquement héréditaire chez les vecteurs, c’est-à-dire que les triatomines
adultes infectés donneront des juvéniles susceptibles. Cependant, il a été constaté que chez les
mammifères hôtes, il existe un certain taux d’hérédité que nous noterons ν.

176
CNMA’2022
Second National Conference on
Mathematics and its Applications

Figure 6: Cycles de vie des vecteurs (triatomines)

et des hôtes (mammifères)

Transmission de [Link] par contacts

Les taux de transmission de [Link]:

βHJ , βHA : de l’hôte vers le vecteur juvénile, adulte.
βJH , βAH : du vecteur juvénile, adulte vers l’hôte.
Alors, les coefficient de contact sont:
ΛJ = βHJ HI
ΛA = βHA HI
Sous l’hypothèse que les taux de transmission chez les hôtes, s’ajoutent transmissions des vecteurs
juvéniles et adultes, nous avons alors:

ΛH = βJH JI + βAH AI

Domaine d’habitat des vecteurs et de leurs hôtes

L’habitat des vecteurs et de leurs hôtes est une partie de la forêt, que nous représentons par le
domaine Ω. Sur ce domaine les vecteurs juvéniles, adultes et leurs hôtes diffusent avec des coeffi-
cients de diffusion dJ , dA et dH .
En considérant que la population est dans un état d’équilibre, nous avons:

JS + JI = J∗
AS + AI = A∗
HS + HI = H∗

Nous supposons que le domaine Ω se partage en deux sous-domaines contigus Ωi où vivent les
individus infectés et Ωs les individus sains.( figure 7)

177
CNMA’2022
Second National Conference on
Mathematics and its Applications

Figure 7: Domaine d’habitat des vecteurs et

de leurs hôtes

Cette hypothèse dépend de la condition nécessaire qu’un mouvement des individus au voisinage
de la frontière commune Γ existe: les sains rejoignent Ωs et les infectés Ωi . En suivant Cantrell et
Cosner [1], le flux de part et d’autre de Γ est décrit par les équations d’assymétrie:

∂JI ∂JS
P J dJ . = (1 − PJ )dJ
∂x ∂x
∂AI ∂AS
P A dA . = (1 − PA )dA
∂x ∂x
∂HI ∂HS
P H dH . = (1 − PH )dH
∂x ∂x
Où PJ (resp. PA , PH ) sont les probabilités pour qu’un juvénile (resp. un adulte, un hôte) se trouvant
dans ΩI traverse la frontiére Γ pour se trouver dans ΩS .

Système à étudier
Les équation de réaction- diffusion d’un tel système biologique s’écrivent:

∂JS
= dJ .∆JS + [(1 − τ(t))σJ (t)JS + fA (t)σA (t)AS + fA (t)σA (t)AI ](1 − ΛJ )
∂t
∂AS
= dA .∆AS + [τ(t)σJ (t)JS + σA (t)AS ](1 − ΛJ )
∂t
∂HS
= dH .∆HS + [σH (t)HS + fH (t)σH (t)HS + (1 − ν)fH (t)σH (t)HI ].(1 − ΛH )
∂t
∂JI
= dJ .∆JI + [(1 − τ(t))σJ (t)JS + fA (t)σA (t)AS + fA (t)σA (t)AI ].ΛJ + (1 − τ(t))σJ (t)JI
∂t
∂AI
= dA .∆AI + [τ(t)σJ (t)JS + σA (t)AS ].ΛJ + τ(t)σJ (t)JI + σA (t)AI
∂t
∂HI
= dH .∆HI + [σH (t)HS + fH (t)σH (t)HS + (1 − ν)fH (t)σH (t)HI ].ΛH + σH (t)HI + νfH (t)σH (t)HI
∂t
Avec les conditions initiales:

Ji (0, x, y) = Ji0 (x, y). Js (0, x, y) = Js0 (x, y).

Ai (0, x, y) = A0i (x, y); (x, y) ∈ Ωi As (0, x, y) = A0s (x, y); (x, y) ∈ Ωs
Hi (0, x, y) = Hi0 (x, y). Hs (0, x, y) = Hs0 (x, y).

178
CNMA’2022
Second National Conference on
Mathematics and its Applications

Et les équations d’assymétrie:

∂JI ∂JS
P J dJ . = (1 − PJ )dJ
∂x ∂x
∂AI ∂AS
P A dA . = (1 − PA )dA
∂x ∂x
∂HI ∂HS
P H dH . = (1 − PH )dH
∂x ∂x

Conclusion
Notre objectif est de résoudre le système obtenu par application des outils de l’analyse fonction-
nelle.

References
[1] Cantrell, Robert Stephen and Cosner, Chris, Spatial ecology via reaction-diffusion equa-
tions, John Wiley & Sons,2004.

[2] Menu, Frédéric and Ginoux, Marine and Rajon, Etienne and Lazzari, Claudio R and
Rabinovich, Jorge E , Adaptive developmental delay in Chagas disease vectors: an evolution-
ary ecology approach, PLoS neglected tropical diseases, 4 (2010).

179
CNMA’2022
Second National Conference on
Mathematics and its Applications

Stability of an axially moving string with a

boundary disturbance

Abdelkarim Kelleche 1 [Link]@[Link]

1 Faculté des Sciences et de la Technologie, Université Djilali Bounâama, Route Theniet El Had, Soufay 44225
Khemis Miliana

Abstract: This work deals with the stabilization problem of an axially moving string with
a tip mass attached at the free end and subject to an external disturbance. The disturbance
here is not uniformly bounded, and it is assumed to be exponentially increasing. First, the
tip mass equation is designed under a boundary controller. By using this equation, the active
disturbance rejection control technique (ADRC) is applied to design a disturbance observer,
and it is shown that the observer can be estimated exponentially. Then the closed loop
system is formulated and the well-posedness of the model is proved in the framework of the
semigroup theory. The stability of the closed loop system is then proved by means of the
multiplier technique, where the energy system converges to equilibrium with an exponential
manner. The efficiency of the obtained results is verified through numerical simulations.

Keywords: Axially moving string, Boundary distribuance, ADRC technique, exponential

stability..
2010 Mathematics Subject Classification: Primary 35L20; 65N30; 74G25; 74G30; 93D15;
93D20.

Introduction
Investigations in the control of axially moving systems have considerably grown in recent years.
These systems include in general: threads, belts, wires, cables, magnetic tapes, chains, see [1]. One
of the main factors that provokes instability and produces vibrations is the external disturbance.
The disturbance is produced for variety of reasons, especially from the external environment, and
it is not necessarily uniformly bounded. In order to reduce or to put end to this phenomena, many
works has been realized in this direction, see e.g., [2] and references therein.
Our concern, throughout this paper, is to discuss the stabilization of an axially moving string with
a tip mass, which is exhibited to unknown disturbances at the tip mass extremity, namely

2
 wtt + 2vwxt − (1 − v )wxx = 0, x ∈ (0, 1) , t > 0,





 w(0, t) = 0, t ≥ 0,



 (165)
mw (1, t) = v (wt + vwx ) (1, t) − wx (1, t) + µ(t) + d(t), t ≥ 0




 tt


 w(x, 0) = w0 (x), wt (x, 0) = w1 (x), x ∈ (0, 1),


where
w : denotes the transversal displacement of the string,
v : is the axial speed of the string, which is assumed constant such that 0 < v < 1,

180
CNMA’2022
Second National Conference on
Mathematics and its Applications

µ(t) : is the control force applied at the tip mass extremity,

d(t) : represents an unknown external disturbance, which is not necessarily bounded and might
be exponentially increasing,
w0 and w1 : are the initial displacement and the initial velocity, respectively.
This system can be used to model the dynamics of cables, marine riser, elevator cables, robotic
arm, aerial tramway, crane and transmission belts. As shown in Figure 1, a moving string system
consists of a cable and a tip mass attached at the end of the cable. A control mechanism consists
of hydraulic actuator is implemented at this end to generates the control force µ(t).
The stationary state, that is when v = 0 in the system (165), has been investigated by Xie and Xu
in [5], under the following requirements on the disturbance d(t). They assumed d ∈ C 1 (R+ ) and
that there exist two positive constants K and k such that
 
˙  ≤ Kekt , t ≥ 0.
|d(t)|, d(t)

The authors adopted the same approach ADRC to design a disturbance observer to estimate the
disturbance exponentially. Then, they employed the estimate term as negative feedback to cancel
the effect of the disturbance, and to drive the system to the equilibrium with an exponential
fashion.
System (165) without disturbance, that is with d(t) = 0, and similar systems have been studied in
several research works, e.g., [3].
Our objective throughout this study is to extend the result obtained by Xie and Xu in [5] to the
case of an axial movement. We adopt the following conditions on the disturbance d(t). We assume
that d ∈ C 1 (R+ ) and there exist a positive constant Cd such that
 
˙  ≤ Cd |d(t)|, t ≥ 0.
d(t)

The outline of the paper is as follows. In Section 2, we design boundary feedback control and
the tip mass equation. Then, we adopt the active disturbance rejection control (ADRC) technique
to estimate the disturbance, and we show that the observer is of exponential type, by means of
the Lyapunov function. In Section 3, we show that the closed-loop system is well-posed in the
framework of the semigroup theory. Then, in section 4, we prove that the closed loop system is
exponentially stable. In section 5, we give numerical simulations to validate the obtained results.
Finally, in section 6, we give some concluding remarks.

Main results
Our main result is as follows

Theorem 60. Assume that hypotheses (H1) and (H2) hold

1. If limt→∞ |d(t)|
r(t)
= 0, then limt→∞ E(t) = 0 and limt→∞ η(t) = 0.
|d(t)|
2. If r(t)
≤ Ke−kt , then there exist positive constants B and γ such that

E(t) + η 2 (t) ≤ Be−γt , t ≥ 0.

References
[1] S. Abrate, Vibration of belts and belt drives. Mech. Mach. Theory, 27 (1992), 645-659.

181
CNMA’2022
Second National Conference on
Mathematics and its Applications

[2] K.S. Hong and P.T. Pham, Control of axially moving systems: A review. Int. J. Control Autom.
Syst., 17 (2019), 2610-2623.

[3] C.W. Kim, K.S Hong and H. Park, Boundary control of an axially moving string: Actuator
dynamics included. J. Mech. Sci. Tech. 19 (2005), 40-50.

[4] A. Kelleche, N-e. Tatar and A. Khemmoudj, Uniform stabilization of an axially moving Kirch-
hoff string by a boundary control of memory type. J. Dyn. Control Syst., 23 (2016), 237-247.

[5] Y. R. Xie and G. Q. Xu, Stabilization of a wave equation with a tip mass based on disturbance
observer of time-varying gain. J. Dyn. Control Syst., 23 (2017), 667–677.

182
Algebra and Geometry

136
CNMA’2022
Second National Conference on
Mathematics and its Applications

Classification of generalized Weingarten

hypersurfacesimmersed in space forms

Abdelmalek Mohammed 1 [Link]@[Link]

Mokhtari Kamel Eddine2 abdelmalekmhd@[Link]
1 Higher School of Management of Tlemcen
2 Department of mathematics , Djilali Liabes University

Abstract: In this work we consider the class of compact generalized Weingarten hypersurfaces
(or (r,s)-Weingarten hypersurfaces) immersed in the Euclidean space Rn , the hyperbolic space
Hn or the open half sphere Sn+ . That is an hypersurface whose some of the k-mean curvatures
are lineary related. ie : for some integers s and r satisfying the inequality 0 ≤ r ≤ s ≤ n, we have:

as Hs + ... + ar Hr = 0

We prove that a closed generalized Weingarten hypersurface embedded in the Euclidean space
Rn , the hyperbolic space Hn or the open half sphere Sn+ must be a round sphere.
Keywords: Generalized Weingarten hypersurfaces, higher order mean curvatures, Minkowski
formulae.
2010 Mathematics Subject Classification: Primary 53A10, 53C42, 53C24.

Introduction
The Alexandrov’s sphere theorem [1] states that the only closed hyersurface embedded in Rn+1 are
the round sphere.
Notice that the above result is not true if the hypersurfaces is immersed and not embedded.
[16,10]. Ros [13] later prove the above result for hypersurfaces of constant Hk for k > 1, embedded
in Euledean space. The result was generalized by Montiel and Ros [11] for hypersurfaces with
constant Hk embedded in H n+1 and S+n+1 .
Hk
Koh [8] and Koh-Lee [9] later gived an analogue for constant Hl hypersurfaces.
In a recent work de Lima [5] gived a gneralization of the Alexandrov theorem for linear Wein-
garten hypersurfaces embedded in Euclidean space. That is an hypersurface where Hk and H are
lineary related. this means that for a ≥ 0 and b > 0, Hk = aH + b.
In this work we consider a compact generalized Weingarten hypersurfaces (or (r, s) −Weingarten
hypersurface) embedded in the Euclidean space Rn+1 , the hyperbolic space H n+1 or the half space
S+n+1 . That is an hypersurface whose some of the k th mean curvatures Hk are lineary related. ie :
for 0 ≤ s ≤ r ≤ n, the relation :
as Hs + .... + ar Hr = b
holds, where b > 0 and ai ≥ 0 with (as , ..., ar ) , (0, 0, ..., 0).

Main results

136
CNMA’2022
Second National Conference on
Mathematics and its Applications

We proof the following result :

Theorem 61. Let M n be a closed, oriented (r, s) −Weingarten hypersurface embedded in the Euclidean
space Rn+1 , the hyperbolic space H n+1 or the half space S+n+1 with non vanishing k−mean curvature Hk .
If we have one of the following cases :
(i) For a given integres r and s such that 0 ≤ s ≤ r ≤ n − 1, the linear relation
as H1 + .... + ar Hr = b
holds, where b > 0 and ai ≥ 0 with (as , ..., ar ) , (0, 0, ..., 0).
(ii) For some integer r where 0 ≤ r ≤ n − 1, the relation
Hr = a1 H1 + .... + ar−1 Hr−1
holds, with ai ≥ 0 and (as , ..., ar ) , (0, 0, ..., 0)..
Then M n is the geodesic hypersphere.

References
[1] A. D. Aleksandrov, A characteristic property of spheres, Ann. Mat. Pura Appl.58 (1962) 303–
315.
[2] [Link]ías, S. de Lira, J.M. Malacarne : Constant higher-order mean curvature hypersurfaces
in Riemannian spaces. Journal of the Inst. of Math. Jussieu 5(4), 527–562 (2006).
[3] L. J. Alías, J. M. Malacarne, Constant scalar curvature hypersurfaces with spherical boundary
in Euclidean space , Rev. Mat. Ibero. 18 (2002), 431-442.
[4] C. Aquino, H. de Lima, and M. Velasquez, A new characterization of complete linear Wein-
garten hypersurfaces in real space forms, Pacific J. Math. 261 (2013), no. 1, 33–43.
[5] E.L. de Lima, A note on compact Weingarten hypersurfaces embedded in Rn+1, Arch. Math.
December 2018, 111(6) (2018), 669–672.
[6] C.C. Hsiung, Some integral formulas for closed hypersurfaces. Math. Scand. 2 (1954), 286–
294.
[7] N. J. Korevaar, Sphere theorems via Alexandrov constant Weingarten curvature hypersur-
faces: appendix to a note of A. Ros, J. Differential Geom. 27 (1988), 221-223.
[8] S.E. Koh, Sphere theorem by means of the ratio of mean curvature functions, Glasgow Math.
J. 42(1) (2000), 91–95.
[9] S.E. Koh and S.-W. Lee, Addendum to the paper: Sphere theorem by means of the ratio of
mean curvature functions, Glasgow Math. J. 43(2) (2001), 275–276.
[10] H. Z. Li, Y. J. Suh, and G. X. Wei, Linear Weingarten hypersurfaces in a unit sphere, Bull.
Korean Math. Soc. 46 (2009), no. 2, 321–329.
[11] S. Montiel and A. Ros, Compact hypersurfaces: The Alexandrov theorem for higher order
mean curvatures, in Differential Geometry, eds. B. Lawson and K. Tonenblat, Pitman Mono-
graphs & Surveys in Pure & Applied Mathematics, Vol. 52 (Longman Higher Education,
1991), 279–297.

137
CNMA’2022
Second National Conference on
Mathematics and its Applications

[12] R.C. Reilly, Variational properties of functions of the mean curvature for hypersurfaces in
space forms, J. Differential Geom. 8 (1973), 465–477.

[13] A. Ros, Compact hypersurfaces with constant higher order mean curvatures, Rev. Mat.
Iberoamericana 3 (1987), 447–453.

[14] H. Rosenberg, Hypersurfaces of constant curvature in space forms, Bull. Sc. Math., 117
(1993), 211–239

[15] S. C. Shu, Linear Weingarten hypersurfaces in a real space form, Glasg. Math. J. 52 (2010),
no. 3, 635–648.

[16] D. Yang, Linear Weingarten spacelike hypersurfaces in locally symmetric Lorentz space, Bull.
Korean Math. Soc. 49 (2012), no. 2, 271–284.

138
CNMA’2022
Second National Conference on
Mathematics and its Applications

Are there infinitely many n such that

 
d n2 = d (ϕ (n))?

Amroune Zahra 1 [Link]@[Link]

Boudaoud Abdelmadjid 1
1 Laboratory of Pure and Applied Mathematics (LMPA), University of M’sila, B.P. 166, Ichbilia, 28000 Mísila,
Algeria., Algeria

Abstract: Abstract: For any positive integer n let d (n) and ϕ (n) be the number of divisors
of n and the Euler’s
 phi
 function of n, respectively. In this paper we present some notes
on the equation d n2 = d (ϕ (n)). In fact, we characterize a class of solutions that have at
most
 two distinct prime factors. Moreover, we show that Dickson’s conjecture implies that
d n2 = d (ϕ (n)) infinitely often.

Keywords: Diophantine equations, Euler’s phi function, divisor function.

2010 Mathematics Subject Classification: Primary 11A25, 11A41, 11D99.

Introduction
Let d (n) be the divisor function, which counts the number of positive divisors of n, i.e., if n
a a a
has the prime factorization n = q11 q22 ...qkk with distinct primes q1 , q2 , ..., qk and positive integers
a1 , a2 , ..., ak , then
d (n) = (a1 + 1) (a2 + 1) ... (ak + 1) .
Let ϕ (n) be the Euler function, which counts the number of positive integers m ≤ n with (m, n) = 1.
It is well-known that
a −1 a −1 a −1
ϕ (n) = q11 (q1 − 1)q22 (q2 − 1)...qkk (qk − 1) .
Recall that various diophantine equations involving the divisor function and Euler’s phi function
were investigated by many authors. For example, see [1], [5] and [4].
In [3, Problem 705, page 78], it is shown that ϕ (d (n)) = d (ϕ (n)) has infinitely many solutions;
while in [5, pages 110-111], it is shown that d (n) = ϕ (n) has the only solutions 1, 3, 8, 10, 24 and
30, where d (n) < ϕ (n) for n ≥ 31. Using these multiplicative functions, we are interested here
in problems involving the number of positive divisors of ϕ (n). In fact, in the present work, we
compare the value of the divisor function to its value at Euler’s functions. More precisely, we aim
to prove that the diophantine equation
 
d n2 = d (ϕ (n)) (166)
has infinitely many integer solutions as well as we identify large families of solutions. The first
few terms are:
1, 5, 57, 74, 202, 292, 394, 514, 652, 1354, 2114, 2125, ....
For this purpose, define n   o
S := n ∈ N : d n2 = d (ϕ (n)) .

139
CNMA’2022
Second National Conference on
Mathematics and its Applications

In this paper, we characterize the elements of S that have at most three distinct prime factors.
The problem is interesting because it can force us to solve some diophantine equations involving
prime numbers. Note also that the proofs are all on the elementary side and depend on long case
by case analysis type arguments.

Recall that the Fermat numbers are the sequence (Fn ) of positive integers defined by
n
Fn = 22 + 1, n = 0, 1, ...

If a particular Fm is prime it is called a Fermat prime. The only known Fermat primes are
F0 , F1 , F2 , F3 and F4 and it has been conjectured that there are only finitely many. On the other
hand, if p = 2k + 1 is a prime then k = 2n for some n and p is a Fermat prime.
It is well-known that d(n) = 2 if and only if n is prime and that d(n) is prime if and only if n = pq−1 ,
where p and q are both prime. Note also that if n is a prime power, namely n = pa with p ≥ 2 is
prime and a ≥ 1, then n ∈ S implies (2a + 1) = d (p − 1) a. But the last equation is only true for a = 1
and p = 5. Hence, n = 5. Observe first of all that there is a connection between Fermat primes and
the solutions of the equation (166), where F1 is the unique prime solution.
Assume that n = q1a q2b ∈ S, where q1 , q2 are distinct primes with 2 ≤ q1 < q2 and a, b ≥ 1. Since
((q1 − 1) (q2 − 1) , q2 ) = 1, we obtain
 
(2a + 1) (2b + 1) = d (q1 − 1) (q2 − 1) q1a−1 b.

Results
We have the following results:

Proposition 14. The only square-free solutions of the form q1 q2 are:

i) n = 3 · 19.

ii) n = 2F3 , where F3 = 257.

 
iii) n = 2 4p2 + 1 , where p and 4p2 + 1 are simultaneously prime.

Next, assume that n is odd. We have the following results:

Proposition 15. Let n = q1a q2 , where 3 ≤ q1 < q2 and a ≥ 2. If n ∈ S, then n is one of the numbers:

• n = F13 · F2 .
 
• n = 35t−3 23 · 3t + 1 , where t ≥ 2 and 23 · 3t + 1 is prime.
 
• n = 55t−3 22 · 5t + 1 , where t ≥ 2 and 22 · 5t + 1 is prime.
 
• n = 3t−1 2 · 3t + 1 , where t ≥ 4 and 2 · 3t + 1 is prime.

Proposition 16. n = F1 · F23 is the only solution of the form q1 q2b , where 3 ≤ q1 < q2 and b ≥ 2.
 b
Theorem 62. Let n = q1a q2b , where 3 ≤ q1 < q2 and a, b ≥ 2. If n ∈ S, then n = 3a 2 · 3t + 1 , where
2 · 3t + 1 is prime and ab + 2a + 2b + 1 = 3bt.

140
CNMA’2022
Second National Conference on
Mathematics and its Applications

Now, assume that n ∈ S is even. We also have the following notes.

Proposition 17. Let n = 2a q2 , where q2 ≥ 3 and a ≥ 2. If n ∈ S, then n is one of the numbers:

 
• n = 25t−3 2t · p4 + 1 , where p and 2t · p4 + 1 are simultaneously prime.
 
• n = 2t−1 2t · p2 + 1 , where p and 2t · p2 + 1 are simultaneously prime.
i
• n = 2(2 −3)/5 Fi , where i ≡ 3(mod 4) and Fi is a Fermat prime.

Proposition 18. Let n = 2q2b , where q2 is odd prime and b ≥ 2. Then n < S.

Proposition 19. Let n = 2a q2b , where q2 ≥ 3 and a, b ≥ 2. If n ∈ S, then n is one of the numbers:

 b
• n = 2a 2s · p2 + 1 , where p and 2s · p2 + 1 are simultaneously prime with ab + 2a + 2b + 1 = 3bs,
 b
• n = 2a 2(3ab+2a+2b+1)/b + 1 , where b divides 2a + 1 and 2(3ab+2a+2b+1)/b + 1 is prime.

Theorem 63. Assuming Dickson’s conjecture, there exist infinitely many primes p such that 4p + 1 and
16p + 1 are primes.

Corollary 4. There exist infinitely many positive integers n such that n ∈ S.

References
[1] Bellaouar, D.: Notes on certain arithmetic inequalities involving two consecutive primes,
Malays.J. Math. Sci., 10, 253–268 (2016).

[2] De Koninck, J. M., Mercier, A.: 1001 problems in classical number theory, Providence, RI:
American Mathematical Society, 2007.

[3] Iannucci, D. E.: On the equation σ (n) = n + ϕ (n), J. Integer Seq., 20, Article 17.6.2. (2017).

[4] Guy, R. K.: Unsolved problems in number theory, Springer-Verlag, New York, 2 edition,
1994.

[5] Sándor, J.: Geometric theorems, Diophantine equations, and arithmetic functions. Amer-
ican Research Press. Rehoboth, 2002.

141
CNMA’2022
Second National Conference on
Mathematics and its Applications

  
The equation σ (n) = τ k · ϕ n2 has finitely
many solutions

Bellaouar Djamel 1 [Link]@[Link]

1 University 08 Mai 1945 Guelma, Department of Mathematics, B.P. 401 Guelma 24000, Algeria

Abstract: For any positive integer n let σ (n), τ (n) and ϕ(n) stand for the divisor sum function
of n, the number of positive divisors of n and the Euler function of n, respectively. In
the present
 paper,
  we show that if k is any positive integer then the diophantine equation
σ (n) = τ k · ϕ n2 has only finitely many solutions. We also find all solutions of the above
equation and related inequalities when k = 1 and k is prime.

Keywords: Diophantine equations, divisor sum function, divisor function, Euler’s phi
function.
2010 Mathematics Subject Classification: Primary 11A25, 11A41, 11D99.

Introduction
P
Let σ (n) be the sum of the natural number divisors of n, so that σ (n) = d|n d, where d runs over
α α α
the positive divisors of n including 1 and itself. If n has the prime factorization n = q1 1 q2 2 ...qs s
with distinct primes q1 , q2 , ..., qs and positive integers a1 , a2 , ..., as , then
s α +1
Y q i −1
i
σ (n) = .
qi − 1
i=1

Let τ (n) be the divisor function, which counts the number of positive divisors of n, i.e.,

τ (n) = (α1 + 1) (α2 + 1) ... (αs + 1) .

In our main results, we will use the following well-known inequalities (see; eg. [3],[7]):

• For any positive integer n, we have √

τ (n) ≤ 2 n.

• For all positive integers m and n, we have

τ (mn) ≤ τ (m) τ (n) .

• If (x, y) > 1, then

τ (x) τ (y) > τ (xy)
Moreover, if x divides y, then
τ (y) ≥ τ (x) .

142
CNMA’2022
Second National Conference on
Mathematics and its Applications

Let ϕ (n) be the Euler function, which counts the number of positive integers m ≤ n with (m, n) = 1.
It is well-known that
α −1 α −1 α −1
ϕ (n) = q1 1 (q1 − 1)q2 2 (q2 − 1)...qs s (qs − 1) .

Several authors have treated diophantine equations involving the sum of divisors function and
Euler’s function. For example, in [6], it shown that n = 2, 3, 4 and 5 are the only solutions of
the equation 2σ (n!) = m!, while in [5], it shown that n = 2 is the only known solution of σ (n) =
n + ϕ (n). Other similar problems have been discussed in publications such as Guy [4]; e.g., §B-38
σ (n) = ϕ (m), and §B-42, σ (ϕ (n)) = ϕ (σ (n)), ϕ (σ (n)) = n, ϕ (σ (n)) = ϕ (n).
The present work is a continuation of the author’s articles [1],[2]. We first define for any positive
integer k the following sets:
n   o
Ek : = n ∈ N : σ (n) = τ k · ϕ n2 ,
n   o
Lk : = n ∈ N : σ (n) < τ k · ϕ n2 ,
n   o
Gk : = n ∈ N : σ (n) > τ k · ϕ n2 .

The
 main focus of this paper is to examine the set Ek of solutions n of the equation σ (n) =
τ k · ϕ n2 . In fact, we characterize the elements of E1 , E2 and E3 , and then we deduce that
Ek and Lk are finite, while Gk is infinite (k ≥ 1). Moreover, we prove that if p is prime with p ≥ 13
and p , 31, then Ep = {6, 14, 31, 33, 77} and if p ≥ 23, then

Lp = {1, 2, 3, 4, 5, 7, 9, 11, 13, 15, 17, 19, 21, 35} .

Main results
  
Theorem 64. E2 = {1, 3, 5, 7} and L2 = {∅}. That is, the only solutions of σ (n) = τ ϕ n2 are 1, 3, 5
  
and 7. Moreover, for every n , 1, 3, 5, 7 we have σ (n) > τ ϕ n2 .

Proposition 20. Let s ≥ 2 and let n = q1 q2 ...qs be a square-free integer. Then n ∈ G1 .

For the proof we need the following lemma.

Lemma 23. Let s ≥ 2 and let q1 , q2 , ..., qs be distinct primes. Then

(1 + q1 )2 (1 + q2 )2 ... (1 + qs )2 > 22s+2 (q1 − 1) (q2 − 1) ... (qs − 1) .

Proposition 21. Let s, k ≥ 2 and let n1 , n2 , ..., ns be relatively prime positive integers with ni ≥ 2 for
i = 1, 2, ..., s. If n1 , n2 , ..., ns ∈ Ek ∪ Gk , then n1 n2 ...ns ∈ Gk .

Proposition 22. E2 = {2, 11} and L2 = {1, 3, 5, 7, 13}.

Theorem 65. Let k ≥ 2. Then the sets Lk and Ek are finite, while Gk is infinite.

Proposition 23. We have ∩ Lk = L2 .

k≥2

Proposition 24. Let k ≥ 2. If r divides s, then Lr ⊂ Ls .

Corollary 5. There are infinitely many k such that Lk ⊂ Lk+1 .

143
CNMA’2022
Second National Conference on
Mathematics and its Applications

Theorem 66. If k ∈ Ek , then k = 1, 2, 6, 11 and 14.

In the following result we characterize the sets Ep and Lp , where p is prime.

Theorem 67. Let p ≥ 13 be a prime number with p , 31. Then Ep = {6, 14, 31, 33, 77}.

Conclusion and open problems

It seems that the most diophantine equations of the form σ (n) = f (n) has finitely many solutions,
where f is a multiplicative function formed by τ, ϕ and many others. As we have already seen
in [5], the author conjectured that n = 2 is the only known solution of  σ (n) = 
n + ϕ (n). Similarly,
in the present paper for any positive integer k, the equation σ (n) = τ k · ϕ n 2 has finitely many
solutions. Even though it may seem that any finite set of consecutive integers includes in Lk for
some k. For example, {1, 2, ..., 11} ⊂ L12 . In the same context and for further research we propose
the following question: When does {1, 2, ..., k} ⊂ Lk+1 ? There are also questions on the set Ek , we
ask whether Ek includes one element for infinitely many k or not. For example, we can prove that
E21 = {51}, E32 = {10}, E38 = {23}, E46 = {95} and E48 = {47}. In particular, when does Ek = {p} with
p is prime? On the other hand, does Ek = {∅} infinitely often?

References
[1] Bellaouar, D., Boudaoud, A., & Özer, Ö. (2019). On a sequence formed by iterating a
divisor operator, Czech. Math. J., 69 (144), 1177–1196.

[2] Bellaouar, D., Boudaoud, A., & Jakimczuk, R. (Accepted) Notes on the equation d(n) =
d(ϕ(n)) and related inequalities, Math. Slovaca.

[3] De Koninck, J. M., & Mercier, A. (2007). 1001 problems in classical number theory, Provi-
dence, RI: American Mathematical Society.

[4] Guy, R. K. (1994). Unsolved problems in number theory, Springer-Verlag, New York, 2 edi-
tion.

[5] Iannucci, D. E. (2017). On the equation σ (n) = n + ϕ(n), J. Integer Seq., 20, Article 17.6.2.

[6] Luca, F. (2000). Equations involving arithmetic functions of factorials, Divulg. Mat., 8,
15–23.

[7] J. Sándor. J. (2002), Geometric theorems, Diophantine equations, and arithmetic functions.
American Research Press. Rehoboth.

144
CNMA’2022
Second National Conference on
Mathematics and its Applications

Conjugaite Hom-actions and the diffrent

Sylow’s theorems for a Hom-groups

Zoheir Chebel 1 zoheir_chebel1@[Link]

Hassene Bouramel 2
Hadjer Adimi 3
1 ETA Laboratory, Electronics Department, University Mohamed El Bachir El Ibrahimi of Bordj Bou Arreridj,
Algeria,
2 Departement of Mathemetics university Barika, Algeria
3 Department of mathematics, University Mohamed El Bachir El Ibrahimi of Bordj Bou Arreridj, Algeria

Abstract: The notion of hom-groups is defined as a generalization of a non-associative

group. They can be obtained by twisting the associative operation with a compatible bijection
mapping. In this presentation, we give some constructions by twisting and also give properties
related to Hom-groups. We introduce the different notions of actions conserning a Hom-
groups. And as an application, the validity of the tree Sylow’s theorems for a Hom-group are
anonced and proved in the general case.

Keywords: Hom-groups; Hom-subgroups; Hom-quotient groups; Isomorphism; Hom-

group actions; First, Second and Third Sylow’s theorems.

2010 Mathematics Subject Classification: Primary 17A99, 17B61, 20B99, 20D20, 20N05.

Introduction
During the study of deformations of the Witt and the Virasoro algebras, the notion of Hom-Lie
algebra were appeared. For this the Jacobi identity is twisted by a linear mapping called in this
case by Hom-Lie algebra in [1, 2, 3]. Firstly, Hartwig, Larsson, and Silvestrov were introduced
this concept in [9]. More people are paying specially attention to those algebraic structure. The
corresponding associative algebras called Hom-associative algebras, were introduced in [13]. The
pierces in many algebraic constructions. The notion of Hom-group was introduced first in [6, 7, 8]
as a non-associative structure analogue of a group in [11], where the authors given a new construc-
tion of the universal enveloping algebra that is different from the one in [15]. Many authors have
developed some notions in Hom-groups for examples, normality, commutator, three isomorphism
of Hom- groups, left action and first Sylow’s theorem in [2, 2]. This new construction leads to a
Hom-Hopf algebra structure on the universal enveloping algebra of a Hom-Lie algebra. In addi-
tion, one can associate a Hom-group to any Hom-Lie algebra by considering group-like elements
in its universal enveloping algebra. The main of this work is to translate this language to groups
theory and looking for a different changing.
The presentation is organized as follows. In section 2, we give the origin and the motivation to
interesting to the Hom-group. In section 3, we present the definition illustrated with example
concern the Hom- group. The construction by twisting map and some important proprieties are
exposed. In Section 4, we discuss the fundamental notions of Hom-groups and Hom-subgroups.

145
CNMA’2022
Second National Conference on
Mathematics and its Applications

In Section 4, we introduce the different notions of Hom-group actions and study some properties
derived from. We prove the first, second and third Sylow’s Theorems.

Preliminaries
This section serves an introductory purpose. We recall some concepts and facts used in this paper,
we give the definition of hom-groups and example is given. Further information on Hom- groups
can be found in [2, 6, 7, 8, 2].

Main results
Sylow’s Theorems for a Hom-group

In this final subsection, we prove the first, Second and Third Sylow’s Theorems by using the
Hom-actions with some applications.

Definition 19. Let (G, α) be a finite Hom-group such that |G| = pn m with p∧m = 1 and n, m are positive
integers. A Hom-subgroup H of G is called a p-Sylow Hom-subgroup of G if and only if |H| = pn .

Lemma 24. [10] Let p be a prime number and k ≥ 0 and m ≥ 1 be integers. Then

pk m
!
≡m mod p.
pk

Theorem 68. (First Sylow’s Theorem)

Let (G, α) be a finite regualr Hom-group such that |G| = pn m with p ∧ m = 1 and n, m are positive
integers. Then G has at least one p-Sylow Hom-subgroup.

Theorem 69. (Second Sylow’s Theorem)

Let (G, α) be a finite regular Hom-group and H be a p-Hom-subgroup of G, then

i) H is contained in a p-Sylow Hom-subgroup of the Hom-group G.

ii) Every p-Sylow Hom-subgroup of G is α-conjugate.

Theorem 70. (Third Sylow’s Theorem)

Any p-Sylow Hom-subgroups of a finite regular Hom-group G is α-conjugate. Moreover, if np is a
number of distinct p-Sylow Hom-subgroups of G, then np divides |G| and np = pt + 1 for some integer t.

References
[1] N. Aizawa and H. Sato, q-deformation of the Virasoro algebra with central extension, Phys.
Lett. B 256 (1991), 185-190.

[2] I. Basdouri., S. Chouaibi., A. Makhlouf., E. Peyghan. Free Hom-groups, Hom-rings and

Semisimple modules, arXiv:2101.03333v1, (2020).

[3] M. Chaichian, P. Kulish, and J. Lukierski, q-deformed Jacobi identity, q-oscillators and q-
deformed infinite-dimensional algebras, Phys. Lett. B 237 (1990), 401-406.

[4] T. L. Curtright and C. K. Zachos, Deforming maps for quantum algebras, [Link]. B 243
(1990), 237-244.

146
CNMA’2022
Second National Conference on
Mathematics and its Applications

[5] L. Chen,. T. Feng., Yao Ma., R. Saha.,H. Zhang., On Hom-Groups and Hom-Group actions.
[Link].

[6] M. Hassanzadeh. Hom-groups, Representations and homological algebra, Colloq. Math. 158
(2019), no. 1,21–38.

[7] M. Hassanzadeh. Lagrange’s theorem For Hom-Groups, Rocky Mountain J. Math. 49 (2019),
no. 3, 773–787.

[8] M. Hassanzadeh, On Antipodes of Hom-Hopf Algeras, [Link]

(2018).

[9] J. T. Hartwig, D. Larsson and S. D. Silvestrov, Deformations of Lie algebras using σ -

derivations, J. Algebra 295 (2006), no. 2, 314-361.

[10] I. Martin Isaacs, Finite group theory, Volume 92 of Graduate studies in mathematics, Ameri-
can Mathematical Soc., 2008.

[11] C. Laurent-Gengoux, A. Makhlouf, and J. Teles, Universal algebra of a Hom-Lie algebra and
grouplike elements, Journal of Pure and Applied Algebra, Volume 222, Issue 5, (2018), P.
1139–1163.

[12] Jun Jiang, Satyendra Kumar Mishra, Yunhe Sheng, Hom-Lie algebras and Hom-Lie groups,
integration and differentiation, arXiv:1904.06515.

[13] A. Makhlouf and S. Silvestrov, Hom-algebra structures, J. Gen. Lie Theory Appl. 2(2), 51–64
(2008).

[14] A. Makhlouf A. and S. Silvestrov, Notes on formal deformations of Hom-associative and

Hom-Lie algebras, Forum Math. 22(4), 715–759 (2021).

[15] D. Yau, Enveloping algebras of Hom-Lie algebras, J. Gen. Lie Theory Appl. 2 (2008), no. 2,
95–108

147
CNMA’2022
Second National Conference on
Mathematics and its Applications

Lie-Poisson triple systems

Fatma Zouaidi 1 zouaidifatmazouaidi@[Link]

1 Department of mathematics , University of Sfax

Abstract: The goal of this work is to introduce the notion of Lie-Poisson triple systems . We
explore the construction of Lie-Poisson triple systems using Poisson algebras. Then We give the
derivation and representation of Lie-Poisson triple systems and some basics results. Moreover,
we give the Constructions of Lie-Poisson triple system using the Rota-Baxter operator.

Introduction
The concept of Lie triple system was introduced first by Jacobson. The present formulation is due
to Yamguti . Moreover, it appeared in Cartan’s work on Riemannian Geometry and was strongly
developed for Symmetric spaces and related spaces. Indeed, the tangent space of a symmetric
space is a Lie triple system. It turns out that they have important applications in physics, in par-
ticular, in elementary particle theory and the theory of quantum mechanics, as well as numerical
analysis of differential equations. They have become an interesting subject in mathematics, their
structure have been studied first by Lister.
The concept of Rota-Baxter operators on Lie algebras appeared first, in a paper of B. A. Kupersh-
midt as an operator analogue of classical r-matrices and Poisson structures. However, Rota-Baxter
operators were introduced by G. Baxter in his study of fluctuation theory in probability. Then
developed by G.-C. Rota in Combinatorics. They have important applications in the algebraic as-
pects of the renormalization in quantum field theory. Rota-Baxter operators on Lie algebras (resp.
associative algebras) are also related to the splitting of algebraic structures. R. Bai, L. Guo, J. Li
and Y. Wu introduced the notion of a Rota-Baxter operator on a 3-Lie algebra. Recently, in order
to construct solutions of the classical 3-Lie Yang-Baxter equation, they introduced a more general
notion called relative Rota-Baxter operator on a 3-Lie algebra with respect to a representation.

Main results
Lie triple systems (L.t.s) is a vector space L endowed with a ternary bracket [·, ·, ·] : L ⊗ L ⊗ L → L
satisfying:

[x, x, z] = 0
[x, y, z] + [y, z, x] + [z, x, y] = 0,
[x, y, [z, t, e]] = [[x, y, z], t, e] + [z, [x, y, t], e] + [z, t, [x, y, e]],
0 0
for any x, y, z, t, e ∈ L. A morphism f : (L, [·, ·, ·]) → (L , [·, ·, ·] ) of L.t.s is a linear map satisfying
0
f ([x, y, z]) = [f (x), f (y), f (z)] , ∀x, y, z ∈ L.

An isomorphism is a bijective morphism.

148
CNMA’2022
Second National Conference on
Mathematics and its Applications

As usual, we have a characterisation of a representation by a semi-direct product i.e. (M, θ) is a

representation of a L.t.s L if and only if L ⊕ M is a L.t.s under the following bracket
 
[x + u, y + v, z + w]L⊕M = [x, y, z], θ(y, z)u − θ(x, z)v + D(x, y)w ,

for any x, y, z ∈ L and u, v, w ∈ M.

Let (L, [·, ·]) be a Lie algebra. we define a L-module as a vector space V together with a linear map
ρ : L → End(V ) such that
ρ([x, y]) = ρ(x)ρ(y) − ρ(y)ρ(x) ∀x, y ∈ L
Note that a LieRep pair (L, [·, ·], ρ) is a Lie algebra (L, [·, ·]) with a L-module.
Theorem 71. Let (L, [·, ·]; ρ) be a LieRep pair. Then (L, [·, ·, ·]; θρ ) is a [Link] pair, where
[x, y, z] = [[x, y], z], ∀ x, y, z ∈ L,
θρ (x, y) = ρ(y)ρ(x), ∀ x, y ∈ L.
Definition 20. Let (L, [·, ·, ·], θ) be [Link] pair. A linear operator T : M → L is said be a O-operator
with respect to θ if it satisfies forall u, v, w ∈ M.
 
[T u, T v, T w] = T D(T u, T v)w + θ(T v, T w)u − θ(T u, T w)v .

Theorem 72. Let T : M → L be a H-twisted Rota-Baxter operator on [Link] pair (L, [·, ·, ·]; θ). Then
there is a L.t.s structure [·, ·, ·]T on M given by
[u, v, w]T = θ(T v, T w)u − θ(T u, T w)v + D(T u, T v)w + H(T u, T v, T w)
for all u, v, w ∈ M. Furthermore, T is a morphism from the L.t.s (M, [·, ·, ·]T ) to (L, [·, ·, ·]).

Let (L, [·, ·]) be a Lie algebra. we define [·, ·, ·]T such that
[x, y, z]T = [[x, y], z], ∀x, y, z ∈ L
Definition 21. Let (L, [·, ·, ·]) be L.t.s. A linear mup R : L → L is said be a Rota-Baxter operator if it
satisfies forall u, v, w ∈ M.
Definition 22. A Lie-Poisson triple system is a triple (A, {·, ·, ·}, ·) where:

1. (A, {·, ·, ·}) is a Lie triple system.

2. (A, ·) is a commutative associative algebra.
3. Compatibility:
{x, y, z · t} = {x, y, z} · t + z · {x, y, t}; ∀x, y, z ∈ A.
Theorem 73. Let (L, [·, ·], ·) be a Poisson algebra. Then (L, [·, ·, ·], ·) is a Lie-Poisson algebras, where
[x, y, z] = [[x, y], z], ∀ x, y, z ∈ L,

References
[1] Ait Ben Haddou, M., Benayadi, S., Boulmane, S. (2016). Malcev–Poisson–Jordan algebras.
Journal of Algebra and its Applications, 15(09), 1650159.

[2] S. Mabrouk, Pre-Lie triple system structures and generalized derivations, preprint. (2021).

[3] Chtioui, T., Hajjaji, A., Mabrouk, S., Makhlouf, A. (2022). O-operators on Lie triple sys-
tems. arXiv preprint arXiv:2204.01853.

149
CNMA’2022
Second National Conference on
Mathematics and its Applications

Other special relations for the Bernoulli

and Euler numbers

Meryem Bouzeraib 1 meryembouzeraib@[Link]

Ali Boussayoud 1
1 Department of Mathematics, Mohamed Seddik Ben Yahia University, Jijel, Algeria

Abstract: In this work, We establish some new combinatorial identities involving Bernoulli
(Euler) numbers and symmetric functions. The derivations use elementary techniques and
are based on functional equations for the respective generating functions. We give several
interesting identities involving them k-Fibonacci, k-Lucas numbers.

Keywords: symmetric functions, exponential generating functions, Euler numbers, Bernoulli

numbers.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
In the literature, we have seen several articles and research that are interested in studying the
other k-numbers, for example Falcon and Plaza in [5] defined and studied the k-Fibonacci num-
bers which is defined by:

Fk,n = kFk,n−1 + Fk,n−2 , with Fk,0 = 0 and Fk,1 = 1.

After that, Falcon in [4] presented some results of the k-Lucas numbers which is defined as:

Lk,n = kLk,n−1 + Lk,n−2 , with Lk,0 = 2 and Lk,1 = k.

The Bernoulli and Euler numbers {Bn }n>0 and {En }n>0 are respectively defined by the exponential
generating functions as:
X zn z
Bn = ,
n! exp(z) − 1
n>0
X zn 2 exp(z)
En = .
n! exp(2z) + 1
n>0

Definition 23. [1] Let A and E be any two alphabets. We define Sn (A − E) by the following form:
Q
(1 − ez) X ∞
e∈E
Q = Sn (A − E)zn ,
(1 − az)
a∈A n=0

with the condition Sn (A − E) = 0 for n < 0.

150
CNMA’2022
Second National Conference on
Mathematics and its Applications

Definition 24. Let n be positive integer and E = {e1 , e2 } are set of given variables. Then, the nth
symmetric function Sn (e1 + e2 ) is defined by:
e1n+1 − e2n+1
Sn (E) = Sn (e1 + e2 ) = , for n ≥ 2,
e1 − e2
with
S0 (E) = S0 (e1 + e2 ) = 1,
S1 (E) = S1 (e1 + e2 ) = e1 + e2 ,
S2 (E) = S2 (e1 + e2 ) = e12 + e1 e2 + e22 ,
..
.

Main results
In this part, we are now in a position to provide three new theorems.
Theorem 74. Given an alphabet E = {e1 , e2 }, then for any positive integer n, we have:
n !
X n
(e − e )l S (e + e )B = ne2n−1 ,
l 1 2 n−l−1 1 2 l
l=0
n !
X n
(e − e )l (2l − 1) (Sn−l (e1 + e2 ) − e1 e2 Sn−l−2 (e1 + e2 )) Bl = −(e1 − e2 )ne2n−1 .
l 1 2
l=0
Theorem 75. Given an alphabet E = {e1 , e2 }, then for any positive integer n, we have :
bn/2c
X n! n(Sn−1 (e1 + e2 ) − e1 e2 Sn−3 (e1 + e2 ))
(e1 − e2 )2l Sn−2l−1 (e1 + e2 )B2l = ,
2l 2
l=0
bn/2c
(e − e )2 nSn−2 (e1 + e2 )
!
X n
(e1 − e2 )2l (22l − 1) (Sn−2l (e1 + e2 ) − e1 e2 Sn−2l−2 (e1 + e2 )) B2l = 1 2 .
2l 2
l=0
Theorem 76. Given an alphabet E = {e1 , e2 }, then for any positive integer n, we have:
bn/2c
X n!
2(4)n−2k−1 (e1 − e2 )2l+1 Sn−2l−1 (e1 + e2 )E2l = (3e1 + e2 )n − (e1 + 3e2 )n ,
2l
l=0
bn/2c ! 2l
X n e1 − e2
(Sn−2l (e1 + e2 ) − e1 e2 Sn−2l−2 (e1 + e2 )) E2l = 21−n (e1 + e2 )n .
2l 2
l=0
√ √
2 2
We let e1 = k+ 2k +4 and e2 = k− 2k +4 , then we have the following relations of Bernoulli and Euler
numbers with k-Fibonacci and k-Lucas numbers.
Corollary 6. Let n be an positive integer, we have:
n !
X n √ 2 l
k + 4 Fn−l,k Bl = ne2n−1 ,
l
l=0

or, equivalently,
bn/2c !
X n  2 l nLn−1,k
k + 4 Fn−2l,k B2l = .
2l 2
l=0

151
CNMA’2022
Second National Conference on
Mathematics and its Applications

Corollary 7. Let n be an positive integer, we have:

n ! √
X n √ 2 l
k + 4 (2l − 1)Ln−l,k Bl = − k 2 + 4ne2n−1
l
l=0

or, equivalently,
bn/2c
(k 2 + 4)nFn−1,k
!
X n  2 l
k + 4 (2l − 1)Ln−2l,k B2l = .
2l 2
l=0

Corollary 8. Let n be an positive integer, we have:

bn/2c ! √ √ √
X n 2l+1
2(4n−2l−1 ) l 2 + 4 Fk,n−2 E2l = (2k + k 2 + 4)n − (2k − k 2 + 4)n
2l
l=0

and
bn/2c !l
k2 + 4
!
X n
Lk,n−2l E2l = 21−n (k)n .
2l 4
l=0

• If we put k = 1 in the Corollaries 6, 7 and 8, we get the following results (see [2, 3]): For each
n > 0,
n ! bn/2c !
X n l X n l nL
5 2 Fn−l Bl = ne2n−1 or,equivalently 5 Fn−2l B2l = n−1 ,
l 2l 2
l=0 l=0
n bn/2c
√
! X n!
X n l l n−1 5nFn−1
5 2 (2 − 1)Ln−l Bl = − 5ne2 or,equivalently (20l − 5l )Ln−2l B2l = ,
l 2l 2
l=0 l=0
bn/2c
√ 2l+1 √ √
!
X n
2(4n−4l−1 ) 5 Fn−2l E2l = (2 + 5)n − (2 − 5)n ,
2l
l=0
bn/2c
n 5r
X !
L E = 2n−1 .
2l 4 n−2l 2l
l=0

References
[1] A. Abderrezzak., Généralisation de la transformation d’Euler d’une série formelle, Adv. Math.
103, 180-195, 1994.

[2] P. F. Byrd, New relations between Fibonacci and Bernoulli numbers,,Fibonacci Quart. 13
(1975), 59–69.

[3] P. F. Byrd, Relations between Euler and Lucas numbers, Fibonacci Quart. 13 (1975), 111–114.

[4] S. Falcon, On the k-Lucas numbers of arithmetic indexes, Appl. Math., 3, 1202-1206, 2012.

[5] S. Falcon, A. Plaza, The k-Fibonacci sequence and the Pascal 2-triangle, Chaos Solitons Frac-
tals, 33, 38-49, 2007.

152
CNMA’2022
Second National Conference on
Mathematics and its Applications

Crossed module on 3-Hom-Lie-Rinehart

superalgebras

Nawel Nasser 1 nassernawel01@[Link]

1 Department of mathematics , University of Sfax

Abstract: The aim of my talk is to study the concept of 3-Hom-Lie-Rinehart superalgebra,

which is a generalization of 3-Lie-Rinehart superalgebra, and systematically describe their rep-
[Link], we study the relationships between a Hom-Lie-Rinehart superalge-
bra and its induced 3-Hom-Lie-Rinehart superalgebra. Finally, we introduce crossed modules
of 3-Hom-Lie-Rinehart superalgebra to gain a better understanding of their third cohomology
groups.

Introduction
The notion of Lie-Rinehart algebras was introduced by J. Herz and further developed in [6].
G. Rinehart developed in [6] a formalism of differential forms for general commutative algebras
which relies on the notion of (K, A)-Lie algebra where K is a commutative ring with unit and A
a commutative K-algebra. Lie-Rinehart algebra includes an abstract algebraic characterization of
the algebraic structure which underlies a Lie algebroid. Thus a Lie-Rinehart algebra can be seen
as an algebraic generalization of the notion of a Lie algebroid: the space of sections of a vector
bundle is replaced by a module over a ring, a vector field by a derivation of the ring. Lie-Rinehart
structures have been the subject of extensive studies, in relations to symplectic geometry, Pois-
son structures, Lie groupoids and algebroids and other kind of quantizations. In [2], the authors
introduced a notion of cross modules of Lie-Rinehart algebras. Hom-Lie-Rinehart algebras and
their extensions in the small dimension cohomology space was introduced and studied in [5].
Hom-Rinehart algebras have close relations with Hom-Gerstenhaber Algebras and Hom-Lie Al-
gebroids. The ternary case of (Hom-)Lie Rinehart (super)algebras was developed in [4].

Main results
Let M = M0 ⊕ M1 be a Z2 -graded vector space. If m ∈ M is a homogenous element, then its degree
will be denoted by m, where m ∈ Z2 = {0, 1}. Denoted by H(M) the set of homogeneous elements
of M. Let End(M) be the Z2 -graded vector space of endomorphisms of a Z2 -graded vector space
M = M0 ⊕ M1 . The graded binary commutator [a, b] = a ◦ b − (−1)ab b ◦ a induces the structure of
Lie superalgebra in End(M) where a ◦ b is the composition of two endomorphisms a and b. In [1]
the authors introduces the notion of Hom-Lie superalgebra given by triple (L, [·, ·], α) consisting
of a superspace L, an even bilinear map [·, ·] : L × L → L and an even homomorphism α : L → L
satisfying
[x, y] = −(−1)xy [y, x],
(−1)xz [α(x), [y, z]] + (−1)zy [α(z), [x, y]] + (−1)yx [α(y), [z, x]] = 0
for all homogeneous element x, y, z in L.

153
CNMA’2022
Second National Conference on
Mathematics and its Applications

Let (L, [·, ·], , α) and (L0 , [·, ·]0 , α 0 ) be two Hom-Lie superalgebras. An homomorphism f : L → L0 is
said to be a morphism of Hom-Lie superalgebras if

f ◦ α =α 0 ◦ f ,
[f (x), f (y)]0 =f ([x, y]), ∀x, y ∈ L.

Hom-subalgebras of Hom-Lie superalgebra (L, [·, ·], α) are defined as Z2 -graded subspaces I ⊆ L
closed under both α and [·, ·], that is α(I) ⊆ I and [I, I] ⊆ I. Hom-subalgebras I is called a Hom-
ideal of Hom-Lie superalgebra L, if [I, L] ⊆ I.
A Z2 -graded vector space L = L0 ⊕ L1 is said to be a 3-Hom-Lie superalgebra, if it is endowed
with an even trilinear map (bracket) [·, ·, ·] : L × L × L → L and an even homomorphism α : L → L
satisfying the following conditions:

[x, y, z] = −(−1)xy [y, x, z], [x, y, z] = −(−1)yz [x, z, y],

[α(x), α(y), [z, u, v]] = [[x, y, z], α(u), α(v)] + (−1)z(x+y) [α(z), [x, y, u], α(v)]
+ (−1)(z+u)(x+y) [α(z), α(u), [x, y, v]],

where x, y, z, u, v ∈ L are homogeneous elements.

Definition 25. A Hom-Lie-Rinehart superalgebra over (A, φ) is a tuple (L, A, [·, ·], α, φ, ρ), where A is
an associative commutative superalgebra, L is an A-module, [·, ·] : L × L → L is an even skew-symmetric
bilinear map, φ : A → A is a superalgebra homomorphism, α : L → L is an even linear map, and the
K-map ρ : L → Derφ (A) such that following conditions hold.

1. The triplet (L, [·, ·], α) is a multiplicative Hom-Lie superalgebra.

2. α(ax) = φ(a)α(x).

3. (ρ, φ) is a representation of (L, [·, ·], α) on A.

4. ρ(ax) = φ(a)ρ(x).

5. The compatibility condition:

[x, ay] = ρ(x)(a)α(y) + (−1)ax φ(a)[x, y]

for all a ∈ H(A), x, y ∈ H(L). A Hom-Lie-Rinehart superalgebra (L, A, [·, ·], α, φ, ρ) is said to be regular if
the maps φ, α are bijective.

Definition 26. A 3-Hom-Lie-Rinehart superalgebra over (A, φ) is a sextuple (L, A, [·, ·, ·], α, φ, ρ), where
A is an associative supercommutative superalgebra, L is an A-module, [·, ·, ·] : L × L × L → L is an even
super skew-symmetric trilinear map, φ : A → A is a superalgebra homomorphism, α : L → L is an even
linear map, and the map ρ : L × L → Derφ (A) such that the following conditions hold:

1. α(ax) = φ(a)α(x).

2. (L, [·, ·, ·], α) is a multiplicative 3-Hom-Lie superalgebra.

3. (ρ, φ) is a representation of (L, [·, ·, ·], α) on A.

4. For all x, y ∈ H(L), a ∈ H(A),

ρ(ax, y) = (−1)ax ρ(x, ay) = φ(a)ρ(x, y).

154
CNMA’2022
Second National Conference on
Mathematics and its Applications

5. Compatibility condition:

[x, y, az] = (−1)a(x+y) φ(a)[x, y, z] + ρ(x, y)(a)α(z)

for all x, y, z ∈ H(L), a ∈ H(A).

Theorem 77. Let (L, A, [·, ·], α, φ, ρ) be a Hom-Lie-Rinehart superalgebra and τ is a supertrace. If the
conditions
τ(ax)y = φ(a)τ(x)y,

are satisfied for all x, y ∈ H(L), a ∈ H(A), then (L, A, [·, ·, ·]τ , α, φ, ρτ ) is a 3-Hom-Lie-Rinehart superal-
gebra, where [·, ·, ·]τ and ρτ are defined by :

[x1 , x2 , x3 ]τ = τ(x1 )[x2 , x3 ] − (−1)x1 x2 τ(x2 )[x1 , x3 ] + (−1)x3 (x1 +x2 ) τ(x3 )[x1 , x2 ],

and
ρτ (x, y) = τ(x)ρ(y) − (−1)xy τ(y)ρ(x), ∀x, y ∈ H(L).
We say that (L, A, [·, ·, ·]τ , α, φ, ρτ ) is the 3-Hom-Lie-Rinehart superalgebra induced by (L, A, [·, ·], α, φ, ρ)
and is denoted by Lτ .

Conversely, we can construct a Hom-Lie-Rinehart superalgebra structure from a given 3-Hom-

Lie-Rinehart superalgebra.

Proposition 25. Let (L, A, [·, ·, ·], α, φ, ρ) be a 3-Hom-Lie-Rinehart superalgebra. Let x0 ∈ L0 such that
α(x0 ) = x0 . Define the bracket [·, ·]x0 = [x0 , ·, ·] and ρx0 (x)(v) = ρ(x0 , x)v. Then (L, A, [·, ·]x0 , α, φ, ρx0 ) is a
Hom-Lie-Rinehart superalgebra.

References
[1] A. Makhlouf, F. Ammar, Hom-Lie superalgebras and Hom-Lie admissible superalgebras.
[Link] 324(2010) 1513–1528.

[2] Casas, J. M., Ladra, M., Pirashvili, T., Crossed modules for Lie-Rinehart algebras, J. Alge-
bra, 274(1), 192-201 (2004)

[3] A. [Link], T. Chtioui, M. Elhamdadi, S. Mabrouk, (2021). Extensions and crossed mod-
ules of n-Lie-Rinehart algebras. arXiv preprint arXiv:2103.15006.

[4] [Link] Hassine, [Link], S. Mabrouk, S. Silvestrov, Structure and cohomology of 3-Lie-
Rinehart superalgebras. Communications in Algebra, 49(11), 4883–4904(2021).

[5] Mandal, A., Mishra, S. K., Hom-Lie-Rinehart algebras, Comm. Algebra, 46 (9), 3722-3744
(2016).

[6] Rinehart, G., Differential forms on general commutative algebras, Trans. Amer. Math. Soc.
108, 195-222 (1963).

155
CNMA’2022
Second National Conference on
Mathematics and its Applications

Prticular neutrosophic subsets on a lattice

Soheyb Milles 1 [Link]@[Link]

Lemnaouar Zedam2
1 Department of mathematics, University Center of Barika
2 Department of Data Analysis and Mathematical Modelling
Ghent University, Coupure links 653

Abstract: In this paper, we aim to deepen the study of ideals and filters notions on a given
lattice in the neutrosophic setting. We show their various properties and characterizations,
in particular, we pay attention to their characterizations based on of the lattice min and max
operations. In addition, we study the notion of prime single-valued neutrosophic ideal (resp.
filter) as interesting kind and we discuss some its set-operations, complement and associate
sets.

Keywords: Lattice, Ideal, Filter, neutrosophic set.

2010 Mathematics Subject Classification: Primary 03B52, 06B10.

Introduction
The notions of ideals and filters are well known in many algebraic structures (e.g., semi-groups,
rings, MV-algebras, lattices, et cetera). They have been applied in different subjects of mathe-
matics, see e.g., topological spaces, metric spaces and congruence relations. They have been used
as tools in the representations of Boolean algebra and distributive lattice. Also, in the theory of
Lukasiewicz and Post algebras, as they are the kernels of the homomorphisms into the power set
subalgebras. In ring theory, ideals generalize certain subsets of the integers, such as the even
numbers or the multiples.

Similar studies of the notions of ideals and filters in neutrosophic context have been done by sev-
eral authors. For instance, Salama considered the notion of filters via neutrosophic crisp set and
investigated several relations between different neutrosophic filters and neutrosophic topologies.
Salama introduced the notion of filters on a neutrosophic set as a generalization of the notion of
fuzzy filters. Recently, Hamidi studied the notion of single-valued neutrosophic filters on EQ-
algebras and its relationship with filters on these kind algebras.

The present study is motivated by the work of Arockiarani, in which they have considered the
notions of lattice, ideal and filter in neutrosophic setting as single-valued neutrosophic sets on a
given crisp lattice. More specifically, we deepen the study of these important notions by providing
their various characterizations and properties. We pay particular attention to their characteriza-
tions based on the lattice min and max operations. Furthermore, the notion of prime single-valued
neutrosophic ideal (resp. filter) as interesting kinds is investigated.

Main results

156
CNMA’2022
Second National Conference on
Mathematics and its Applications

Proposition 26. The following equivalences hold for any A ∈ SV N (L) and x, y ∈ L:
(i) (TA (x f y) ≥ TA (x) ∨ TA (y))⇔ (x 6 y ⇒ TA (x) ≥ TA (y)),
(ii) (TA (x g y) ≥ TA (x) ∨ TA (y)) ⇔ (x 6 y ⇒ TA (x) ≤ TA (y)),
(iii) (IA (x f y) ≥ IA (x) ∨ IA (y)) ⇔ (x 6 y ⇒ IA (x) ≥ IA (y)),
(iv) (IA (x g y) ≥ IA (x) ∨ IA (y)) ⇔ (x 6 y ⇒ IA (x) ≤ IA (y)),
(v)(FA (x f y) ≤ FA (x) ∧ FA (y)) ⇔ (x 6 y ⇒ FA (x) ≤ FA (y)),
(vi)(FA (x g y) ≤ FA (x) ∧ FA (y)) ⇔ (x 6 y ⇒ FA (x) ≥ FA (y)).

The following two theorems provide characterizations of SVN-ideal (resp. SVN-filter) on a lattice.
Theorem 78. I is a SVN-ideal on L if and only if for any x, y ∈ L, the following conditions are satisfied:
(i) TI (x g y) = TI (x) ∧ TI (y),
(ii) II (x g y) = II (x) ∧ II (y),
(iii) FI (x g y) = FI (x) ∨ FI (y).
Theorem 79. F is a SVN-filter on L if and only if for any x, y ∈ L, the following conditions are satisfied:
(i) TF (x f y) = TF (x) ∧ TF (y),
(ii) IF (x f y) = IF (x) ∧ IF (y),
(ii) FF (x f y) = FF (x) ∨ FF (y).

The following corollaries characterize crisp (fuzzy) ideals and intuitionstic fuzzy ideals (resp.
crisp (fuzzy) filters and intuitionstic fuzzy filters) on a given lattice.
Corollary 9. For any crisp (fuzzy) sets I and F on L, the following equivalences hold:
(i) I is a crisp (fuzzy) ideal on L if and only if TI (x g y) = TI (x) ∧ TI (y), for any x, y ∈ L,
(ii) F is a crisp (fuzzy) filter on L if and only if TF (x f y) = TF (x) ∧ TF (y), for any x, y ∈ L.
Corollary 10. For any IFSs I and F on L, the following equivalences hold:
(i) I is an IF-ideal on L if and only if for any x, y ∈ L, the following two conditions are satisfied: (a)
TI (x g y) = TI (x) ∧ TI (y), (b) FI (x g y) = FI (x) ∨ FI (y).
(ii) F is an IF-filter on L if and only if for any x, y ∈ L, the following two conditions are satisfied:(a)
TF (x f y) = TF (x) ∧ TF (y), (b) FF (x f y) = FF (x) ∨ FF (y).

Given a SVN-ideal I (resp. SVN-filter F) on a lattice L.

I is called a prime SVN-ideal if, for any x, y ∈ L, it holds that
(i) TI (x f y) ≤ TI (x) ∨ TI (y),
(ii) II (x f y) ≤ II (x) ∨ II (y),
(iii) FI (x f y) ≥ FI (x) ∧ FI (y) .
Dually, F is called a prime SVN-filter if, for any x, y ∈ L, it holds that
(i) TF (x g y) ≤ TF (x) ∨ TF (y),
(ii) IF (x g y) ≤ IF (x) ∨ IF (y),
(iii) FF (x g y) ≥ FF (x) ∧ FF (y) .
The above Theorem 78 lead to the following result which characterize prime single-valued neu-
trosophic ideals.
Proposition 27. I is a prime SVN-ideal on L if and only if for any x, y ∈ L, the following conditions
hold:
(i) TI (x g y) = TI (x) ∧ TI (y),
(ii) TI (x f y) = TI (x) ∨ TI (y),
(iii) II (x g y) = II (x) ∧ II (y),
(iv) II (x f y) = II (x) ∨ II (y),

157
CNMA’2022
Second National Conference on
Mathematics and its Applications

(v) FI (x g y) = FI (x) ∨ FI (y),

(vi) FI (x f y) = FI (x) ∧ FI (y).

Similarly, Theorem 79 lead to the following result which characterize prime SVN-filters.

Proposition 28. F is a prime SVN-filter on L if and only if for any x, y ∈ L, the following conditions
hold:
(i) TF (x g y) = TF (x) ∨ TF (y),
(ii) TF (x f y) = TF (x) ∧ TF (y),
(iii) IF (x g y) = IF (x) ∨ IF (y),
(iv) IF (x f y) = IF (x) ∧ IF (y),
(v) FF (x g y) = FF (x) ∧ FF (y),
(vi) FF (x f y) = FF (x) ∨ FF (y).

References
[1] S. Boudaoud, L. Zedam, S. Milles, Principal intuitionistic fuzzy ideals and filters on a lattice,
Discussiones Mathematicae General Algebra and Applications, 40, (2020), 75–88.

[2] I. Mezzomo, B.C. Bedregal and R.H.N. Santiago, Types of fuzzy ideals in fuzzy lattices, Jour-
nal of Intelligent and Fuzzy Systems 28, (2015), 929–945.

[3] S. Milles, L. Zedam, E. Rak, Characterizations of intuitionistic fuzzy ideals and filters based
on lattice operations, J. Fuzzy Set Valued Anal., 2, (2017), 143–159.

[4] H. Wang, F. Smarandache, Y.Q. Zhang, R. Sunderraman, Single valued neutrosophic sets,
Multispace Multistruct, 4, (2010), 410-413.

[5] L.A. Zadeh, Fuzzy sets, Information and Control, 8, (1965), 331–352.

158
CNMA’2022
Second National Conference on
Mathematics and its Applications

Generating functions for certain product

of orthogonal polynomials with the
generalized Tribonacci numbers

Wissem Benamira 1 wissembenamira18@[Link]

Ali Boussayoud 1
1 Department of mathematics , University Mohamed seddik ben yahia jijel

Abstract: We will deal with the calculation of certain new generating functions of the product
of generalized Tribonacci numbers with Chebyshev polynomials of first, second, third and
fourth kind and Fibonacci polynomials.

Keywords: Generating functions, Symmetric functions, Orthogonal polynomials, Gener-

alized Tribonacci numbers.
2010 Mathematics Subject Classification: Primary 05E05; Secondary 11B39.

Introduction
Recently, many studies have been conducted on numbers known as generalized Tribonacci num-
bers defined by:


 wn = awn−1 + bwn−2 + cwn−3



 , ∀n ≥ 3,
w0 = α, w1 = β, w2 = γ




with a, b, c ∈ R, α, β, γ ∈ Z.
Many other famous numbers like Tribonacci, Narayana, Padovan, Peren, and Jacobstalthe num-
bers can be defined by these third order recurrence relations.
In this work, we will adopt on the theory of symmetric functions to calculate on one hand new
generating functions of the product of generalized Tribonacci numbers with Chebyshev polyno-
mials of the four kinds and on the other hand the product of generalized Tribonacci numbers with
Fibonacci polynomials.

Some Definition
We consider in the following two positive integers n, k and λ1 , ..., λn different roots (real or com-
plex) of an algebraic equation of degree n :

159
CNMA’2022
Second National Conference on
Mathematics and its Applications

Definition 27. We define the elementary symmetric function of the order k by:
(n)
X
i i
ek = λ11 ...λnn , i1 , ..., in ∈ {0, 1} ,
i1 +...+in =k
(n)
ek = 0, k  n, k ≺ 0.

Definition 28. We define the complete homogeneous symmetric function of the order k by:
(n)
X
i i
hk = λ11 ...λnn , i1 , ..., in ≥ 0,
i1 +...+in =k
(n)
hk = 0, k ≺ 0.

Definition 29. Consider the alphabets P = {p1 , p2 } , A = {a1 , a2 , a3 }, we define the following symmetric
functions :

p1n+1 − p2n+1
Sn (P ) = Sn (p1 + p2 ) = , n ∈ N,
p1 − p2
a S (a + a ) − a3 Sn (a1 + a3 )
Sn (A) = Sn (a1 + a2 + a3 ) = 2 n 1 2 , n ∈ N.
a2 − a3

Definition 30. The symmetric operator is defined by:

p1k f (p1 ) − p2k f (p2 )

δpk1 ,p2 (f ) = , n ∈ N.
p1 − p2

Main Results

Theorem 80. For each natural number n, the new generating function of the product of generalized
Tribonacci numbers with the Chebyshev polynomials of the second kind is given by the following relation:
  
α + 2 (β − αa) xt + b − a (β − αa) + (γ − αb − βa) 4x2 − 1 t 2 +
+∞
X 2x (αc − a (γ − αb − βa)) t 3 + (c (β − αa) − b (γ − αb − βa)) t 4
n
wn un (x) t =     .
n=0 1 − 2axt + −4bx2 + a2 + 2b t 2 − 2 4cx2 − (ab + 3c) xt 3
 
+ b2 − 2ac + 4cx2 t 4 + 2bcxt 5 + c2 t 6

Theorem 81. For each natural number n, the new generating function of the product of generalized
Tribonacci numbers with the Chebyshev polynomials of the first kind is given by the following relation:
   
α + (β − 2αa) xt + αb − a (β − αa) + (γ − αb − βa) 4x2 − 1 − 2x2 (γ − βa) t 2
   
+ 2αc − a (γ − αb − βa) − αc 4x2 − 1 − b (β − αa) − b xt 3
   
+∞
X − c (β − αa) 2x2 − 1 + b (γ − αb − βa) t 4 + c (γ − αb − βa) xt 5
wn Tn (x) t n =     .
1 − 2axt + −4bx 2 + a2 + 2b t 2 − 2 4cx2 − (ab + 3c) xt 3
n=0
 
+ b2 − 2ac + 4cx2 t 4 + 2bcxt 5 + c2 t 6

160
CNMA’2022
Second National Conference on
Mathematics and its Applications

Theorem 82. For each natural number n, the new generating function of the product of generalized
Tribonacci numbers with the Chebyshev polynomials of the third kind is given by the following relation:
   
α + (2x (β − αa) − β) t + αb − a (β − αa) + (γ − αb − βa) 4x2 − 1 − 2x (γ − βa) t 2
   
+ 2x (αc − a (γ − αb − βa)) − αc 4x2 − 1 − b (β − αa) − b t 3
+∞
X + (c (β − αa) − b (γ − αb − βa) − 2cx (β − αa)) t 4 − c (γ − αb − βa) t 5
n
wn Vn (x) t =     .
n=0 1 − 2axt + −4bx2 + a2 + 2b t 2 − 2 4cx2 − (ab + 3c) xt 3
 
+ b2 − 2ac + 4cx2 t 4 + 2bcxt 5 + c2 t 6

Theorem 83. For each natural number n, the new generating function of the product of generalized
Tribonacci numbers with the Chebyshev polynomials of the Fourth kind is given by the following relation:
   
α + (2x (β − αa) + β) t + αb − a (β − αa) + (γ − αb − βa) 4x2 − 1 − 2x (γ − βa) t 2
   
+ 2x (αc − a (γ − αb − βa)) + αc 4x2 − 1 + b (β − αa) − b t 3 .
+∞
X + (c (β − αa) − b (γ − αb − βa) + 2cx (β − αa)) t 4 + c (γ − αb − βa) t 5
wn Wn (x) t n =    
n=0 1 − 2axt + −4bx2 + a2 + 2b t 2 − 2 4cx2 − (ab + 3c) xt 3
 
+ b2 − 2ac + 4cx2 t 4 + 2bcxt 5 + c2 t 6

Theorem 84. For each natural number n, the new generating function of the product of generalized
Tribonacci numbers with Fibonacci polynomials is given by the following relation:
  
α + (β − αa) xt + −αb − a (β − αa) + (γ − αb − βa) x2 + 1 t 2 +
+∞
X x (−αc + a (γ − αb − βa)) t 3 + (c (β − αa) − b (γ − αb − βa)) t 4
wn Fn (x) t n =     .
n=0 1 − axt − bx2 + b (b + 2) t 2 − cx3 + (ab + 3c) xt 3
 
+ b2 − 2ac − acx2 t 4 + bcxt 5 − c2 t 6

References
[1] A. Boussayoud And N. Harrouche, Complete symmetric functions and k-Fibonacci numbers,
Commun. Appl. Anal. 20 (2016), 457–465.

[2] A. Boussayoud, S. Boughaba, M. Kerada, S. Araci And M. Acikgoz, Symmetric Functions

of Binary Products of k-Fibonacci and Orthogonal Polynomials, Rev. R. Acad. Cienc. Exactas
Fis. Nat. 113 (2019), 2575–2586.

[3] [Link], A generalization of the symmetry between complete and elementary symmetric
functions, Indian [Link] Appl. Math. 45 (2014), 75–89.

[4] M. Chelgham And A. Boussayoud, Construction of Symmetric functions of generalized Tri-

bonacci numbers, J. Sci. Arts. 50 (2020), 65–74.

161
EDO and Dynamical
Systems

162
CNMA’2022
Second National Conference on
Mathematics and its Applications

Limit Cycle of Discontinuous Piecewise

Differential Systems Separated by a
Straight Line and Formed by a Linear
Center and a Quadratic Center

R. ALLAOUA 1 [Link]@[Link]
R. CHEURFA 1 rcheurfa@[Link]
A. BENDJEDDOU 1 Bendjeddou@[Link]
1 Laboratory of Applied Mathematics, Department of Mathematics, Faculty of Sciences, University of Setif1 Fer-
hat Abbas , Algeria

Abstract: This research investigates the existence and number of limit cycles that can be
exhibited by planar piecewise differential systems separated by one straight line y=0, formed
by a linear center, and quadratic center. Using the first integrales, we prove this class of
piecewise differential systems can have at most one explicit limit cycle.

Keywords: Piecewise differential systems, Linear center, Quadratic center, First integral,
Limit cycle. irreducible cubic curves.
2010 Mathematics Subject Classification: 34A30, 34C05, 34C25, 34C07, 37G15.

Introduction and statement of the main results

The study of the existence of limit cycles and their number is one of the main concerns in the
dynamics of differential systems on the plan. The famous 16th Hilbert’s problem is restricted to
polynomial differential systems, see more details in [1, Hilbert,1900]. A limit cycle is a periodic
orbit of a differential system that is isolated from of its possible periodic orbits.
In the 1920s Andronov, Vitt, and Khaikin were interested in different differential systems defined
in distinct regions separated by a curve, and later on, Filippov, in 1988, provided the theoretical
bases for these kinds of differential systems; for instance, see the book of [2]. As it appeared
in the characterization of many real operations such as switches in electronic circuits. Recently
the study of piecewise discontinuous differential systems, has found several applications in other
branches of science.
There are many studies of the limit cycles of continuous and discontinuous piecewise differential
systems in R2 with two pieces separated by a straight line. In general these differential systems are
linear. From an other side, several papers concerned by the study of the existence and the number
of the limit cycles for some classes of discontinuous piecewise differential systems separated by
a straight line and formed by quadratic system have appeared recently see for instance ([3], [4],
[5]). In the last two papers, the authors are interested by the maximal number of limit cycles

162
CNMA’2022
Second National Conference on
Mathematics and its Applications

obtained after perturbation of a linear center. The technique used is the averaging method, Our
objective in this work is to study of the existence, number of limit cycles for these systems, their
existence is proved by using the first Integral.
Let p ∈ R2 be a singularity which is a center for the linear part of the system at p. Then p is either a
center or it is a focus in which case p is called a weak focus. A theorem of Poincaré says that p is a
center if and only if the system has a nonconstant analytic first integral in the neighborhood of p.
In [Mazzi & Sabatini, 1988] it is shown that an isolated singular point p of an analytic differential
system is a center if and only if there exists a first integral of class C ∞ with an isolated minimum
at the singular point p.
More precisely we study the limit cycles, which passes through the intersection of two isolated
points with the straight line. We define two regions separated by a straight line Σ for the class of
piecewise differential systems
n in R2 . We can
o consider without loss of generality that the disconti-
2
nuity straight line is Σ = (x, y) ∈ R : y = 0 . It separates the plane into two regions, namely
n o n o
Σ− = (x, y) ∈ R2 : y < 0 and Σ+ = (x, y) ∈ R2 : y > 0 .

Our main concern is to discuss the existence of periodic orbits for such discontinuous piecewise
differential systems assuming that both differential systems, the one is quadratic systems defined
in y < 0 and the other is linear center defined in y > 0. We can write such a discontinuous piecewise
differential system as

2
+δ2
ẋ = −αx + βy + γ, ẏ = − α β x + αy + ξ, if (x, y) ∈ Σ+ ,
ẋ = y + ax2 + bxy + 2ay 2 , ẏ = −x − 2bx2 − axy − by 2 , if (x, y) ∈ Σ− .

Our main result is the following theorem.

Theorem 85. The discontinuous piecewise differential systems in the plane with two pieces separated
by one straighy line and having in one piece a linear center and in the other piece a quadratic center
have at most one limit cycle.

The next proposition shows that there are discontinuous piecewise differential systems separated
by a straight line Σ, one of which is a linear center and the other a quadratic center, with one limit
cycle.
Proposition 29. Consider the class of discontinuous piecewise differential systems separated by the
straight line y = 0 and formed by a linear center and a quadratic center. Then these differential systems
can have at most one limit cycle. Moreover, the discontinuous piecewise differential system in this class
formed by the differential system

ẋ = − 61 x − 3y + 1, ẏ = 34 x + 16 y + ξ. (167)

in y > 0, and by the differential system

1 2 1
ẋ = y + 10 x + 10 xy + 15 y 2 , ẏ = −x − 51 x2 − 10
1 1 2
xy − 10 y . (168)

in y < 0, has one limit cycle, reaching the maximum upper bound. See Figure 8.

Preliminaries
A differential system
ẋ = P (x, y) , ẏ = Q (x, y) , (169)

163
CNMA’2022
Second National Conference on
Mathematics and its Applications

Figure 8: One limit cycle of systems (167)+(168) with ξ = 0.47947.

with P and Q is called a polynomial differential system of degree n (when n = 1 system (169) is
called linear system, and when n = 2 system (169) is called quadratic system).
Let U be an open subset of R2 and H : U → R be a C 1 function. H is a first integral of a differential
system (169) if and only if

dH dH dH
=P +Q = 0 in all the points of U ,
dt dx dy

if the function H on the orbits of system (169) contained in U is constant.

The following lemma provides a normal form for a linear differential system having a center.

Lemma 25. A linear differential system having a center can be written as

2
+δ2
ẋ = −αx + βy + γ, ẏ = − α β x + αy + ξ,

where α, β, γ, δ and ξ are real constants with β < 0. Moreover, this system has the first integral

  (αγ − βξ)2 + γ 2 δ2
H1 (x, y) = α 2 + δ2 x2 − 2αβxy + β 2 y 2 − 2β (xξ − γy) + .
δ2

The quadratic centers where characterized in the following lemma

Lemma 26. Any quadratic system candidate to have a center can be written in the form

ẋ = y + ax2 + bxy + 2ay 2 , ẏ = −x − 2bx2 − axy − by 2 ,

where a and b are real constants. This system has a center at the origin and has the first integral
 
H2 (x, y) = (ay + bx + 1)2 x2 + y 2 .

164
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] D. Hilbert, [1900] “Mathematische probleme,” Lecture, Second Int. Congr. Math., Paris,
Nachr. Ges. Wiss.¨ ottingen Math. Phys. Kl. 3, pp. 253–297; English transl. [1902] Bull. Amer.
Math. Soc. 8, 437–479; [2000] Bull. (New Series) Amer. Math. Soc., 37, 407–436.

[2] D.J.W. Simpson, Bifurcations in Piecewise-Smooth Continuous Systems, World Scientific Se-
ries on Nonlinear Science Series A; World Scientific: Singapore 69 (2010).

[3] J. Llibre, Y. Tang, Limit cycles of disconyinuous piecewise quadratic and cubic polynomial
perturbations of a linear center, Mathematic: Dynamical Systems 1 (2017).

[4] M. Esteban, J. Llibre, and C. Valls, The 16th Hilbert problem for discontinuous piecewise
isochronous centers of degree one or two separated by a straight line, Chaos: An Interdisci-
plinary Journal of Nonlinear Science 31 (2021), 043112.

[5] X. Cen, L. Yang and M. Zhang, Limit cycles by perturbing quadratic isochronous centers
inside piecewise smooth polynomial differential systems, Mathematics: Classical Analysis and
ODEs 2 (2017), 13–22.

165
CNMA’2022
Second National Conference on
Mathematics and its Applications

On a solvable p−dimensional system of

nonlinear difference equations

Imane Zemmouri 1 imanezemmouri25@[Link]

Ahmed Ghezal 2 ahmedghezal$25$@[Link]
1 Department of Mathematics, University of Annaba, Elhadjar 23, Annaba, Algeria
2 Department of Mathematics and Computer Sciences, University Center of Mila, Algeria

Abstract: In this paper, we investigate the solutions of the following system of p−nonlinear
difference equations
(i+1) mod(p) (i+1) mod(p)
(i) a(i) xn xn−2
xn+1 = (i) (i+1) mod(p)
, n ∈ N0 , p ∈ N,i ∈ {1, ..., p} ,
b(i) xn−1 + c(i) xn−2
     
where N0 = N ∪ {0} , the sequences a(i) , b(i) , c(i) , are non-zero real numbers and initial
(i)
values x−j , j ∈ {0, 1, 2}, i ∈ {1, ..., p}.

Keywords: Riccati difference equation, periodicity, general solution, system of difference

equations, Fibonacci sequence.
2010 Mathematics Subject Classification: 39A05, 39A10.

Introduction
In the recent years, there has been a lot of interest in studying nonlinear difference equations
and systems. Not surprisingly therefore, several studies have been published on this topic (see,
e.g., [1] − [5], and the related references therein). Besides their theoretical value, most of the re-
cent applications have appeared in many scientific areas such as biology (population dynamics in
particular), ecology, physics, engineering and economics (see, e.g. [2], [3] ). It is very worthy to
find systems belonging to solvable nonlinear difference equations systems in closed-form.
In Abo-Zeid et al. [1] the authors presented the solutions of the one-dimensional system of non-
linear difference equations which reduced to the Riccati difference equation under appropriate
transformations,
xn xn−2
xn+1 = , n ∈ N0 . (170)
±xn−1 ∓ xn−2

But, two (resp. three)−dimensional system of difference equations in (170) was extended to the
following two (resp. three)−dimensional system of difference equations with constant coefficients
yn yn−2 xn xn−2
xn+1 = ,y = , n ∈ N0 , (171)
bxn−1 + ayn−2 n+1 dyn−1 + cxn−2
(resp.
yn yn−2 zn zn−2 xn xn−2
xn+1 = , yn+1 = , zn+1 = , n ∈ N0 ), (172)
bxn−1 + ayn−2 dyn−1 + czn−2 f zn−1 + exn−2

166
CNMA’2022
Second National Conference on
Mathematics and its Applications

and system (171) (resp. (172)) was solved using convenient transformations in [5] (resp. [4]). Its
extension with constant coefficients and p−dimensional is a system of a huge interest. For this
reason, another extension of system (172) is the following system of p−dimensional nonlinear
difference equations
(i+1) mod(p) (i+1) mod(p)
(i) a(i) xn xn−2
xn+1 = (i) (i+1) mod(p)
, n ∈ N0 , p ∈ N,i ∈ {1, ..., p} . (173)
b(i) xn−1 + c(i) xn−2

Now, we consider system (173) in the case when a a(i) , 0 for all i ∈ {1, ..., p} . Noticing that in this
case, system (173) can be written in the form
(i+1) mod(p) (i+1) mod(p)
(i) xn xn−2
xn+1 = (i) (i+1) mod(p)
, n ∈ N0 , p ∈ N,i ∈ {1, ..., p} .
b(i) xn−1 + e
e c(i) xn−2

b(i) (i)
(i) = c , for all i ∈ {1, ..., p}, we see that we may assume that a(i) = 1, for all
b(i) =
where e and c
e
a(i) a(i)
i ∈ {1, ..., p}. Hence we consider, without loss of generality, the system
(i+1) mod(p) (i+1) mod(p)
(i) xn xn−2
xn+1 = (i) (i+1) mod(p)
, n ∈ N0 , p ∈ N,i ∈ {1, ..., p} . (174)
b(i) xn−1 + c(i) xn−2

using the same notation for coefficients as in (173) except for the coefficients a(i) , assuming that
a(i) = 1, for all i ∈ {1, ..., p}.

Main results
 
(1) (2) (p) (i)
Let xn , xn , ..., xn be a solution of system (174). If at least one of the initial values x−j ,
n≥−2
j ∈ {0, 1, 2}, i ∈ {1, ..., p}, is equal to zero, then the solutions of system (174) is not defined. For
(i )
example, if xn00 = 0 for some n0 ≥ −2, i0 ∈ {1, ..., p}. Then from the system (174) it follows that
(i ) (i ) (i +1) mod(p) (i )
xn00+1 = 0, and consequently b(i0 ) xn00+1 + c(i0 ) xn00
= 0, from which it follows that xn00+3 is not
p
(i)
Y
defined. Thus, for every well-defined solution of system (174), we get that xn , 0, n ≥ −2, if
i=1
p
(i)
Y
and only if x−j , 0, j ∈ {0, 1, 2}. Note that the system (174) can be written in the form
i=1

(i) (i+1) mod(p)

(i) xn−1 (i) xn
b (i+1) mod(p)
+c = (i)
, n ∈ N0 , p ∈ N, i ∈ {1, ..., p} . (175)
xn−2 xn+1

The next theorem is one of the main results of this paper

167
CNMA’2022
Second National Conference on
Mathematics and its Applications
 
(1) (2) (p)
Theorem 86. Let xn , xn , ..., xn be a solution of system (174). Then for n ≥ p,
n≥−2
  −1  −1






 [ 2k ]
Y






 Y k 



 
  
(i+k−2j) mod(p) 
  (i−k+2j−1) mod(p) 

 (i+k) mod(p)
y y x0
  
2j

 
 
 
 2(k−j)+1 


 
 
 
 


  j=1

  j=[ 2k ]+1

  

if p is even



 h p−2 i h p−2 i 
n−1 

  2






 2




 Y 
 Y 
 Y 

  (i+2j+t) mod(p)  (i+2j+1−t) mod(p) 
y2(h (l,k,p,n)−j)  y2(h (l,k,p,n)+t−j)−1 

   
1 1
 
  
 

 
 

 

l=0   j=0  j=0
  
 
  
(i)  
xpn+k = ,
 −1  h −1
p−2+2h
 −1 i
k n 





 [ 2 ] 

 
 k 
 [ ]
2  Y


 2




 Y
  
  Y 
 Y  

  (i+k−2j) mod(p)  
  (i−k+2j−1) mod(p)   (i+k) mod(p)  (i+2j+1−h) mod(p) 
y y x y
   

 

 2j 
 


 2(k−j)+1 

 0 

 2(h3 (l,k,p,n)−j−1+h)  

      
 j=1
 j=0

 
  j=[ k ]+1


 
 l=1  




2

 


 h
p−2h
i h
p−2+2h
i   p+2t−2t2
h i  if p is odd
n−[ n−2 [ n2 ]  Y
Y2 ] Y

 
 2 
 
 
 

 

 
 Y 2 

Y 2 


     
 (i+2j+h) mod(p) 
 (i+2j+1−h) mod(p)   (i+2j+h) mod(p) 
y y y
     

 

 2(h2 (l,k,p,n)+t∧t2 −j) 




 2(h 2 (l,k,p,n)+h−j)−1 




 2(h3 (l,k,p,n)−j+t∧t 2 )−1 


      


 l=0   j=0
 

  j=0
  l=1 

  j=0




  

h i h i
k ∈ {0, 1, ..., p − 1} , i ∈ {1, ..., p} , p ∈ N, where t = k − 2 2k ∈ {0, 1} , t2 = n − 2 n2 ∈ {0, 1} , h = t ∨ t2 − t ∧ t2 ,
hpi h i h i hpi h i
h1 (l, k, p, n) = 2 (n − l)+ 2k , h2 (l, k, p, n) = h1 (2l, k, p, n)−l+ n2 , h3 (l, k, p, n) = 2 (2l + t2 − 1)+ 2k +l,
[x] is integral part of x and
(i) (i) (i) (i) (i) (i)
(i) b(i) sm + y−1 sm+1 (i) b(i) sm + y0 sm+1
y2m+1 = (i) (i) (i)
, y2(m+1) = (i) (i) (i)
, m ≥ −1, i ∈ {1, ..., p} , p ∈ N,
b(i) sm−1 + y−1 sm b(i) sm−1 + y0 sm
(i) (i) (i) (i) (i)
with (sm )m≥−1,i∈{1,...,p} be the solution to system um+1 = c(i) um + b(i) um−1 , m ∈ N0 , such that s−1 = 0 and
(i)
s0 = 1, for i ∈ {1, ..., p} .
 
(1) (2p)
Corollary 11. Let xn , ..., xn be a well-defined solution to the following system,
n≥−2

(i+1) mod(2p) (i+1) mod(2p)

(i) xn xn−2
xn+1 = (i) (i+1) mod(2p)
, n ∈ N0 ,i ∈ {1, ..., 2p} , p ∈ N.
xn−1 + xn−2

Then

 [ 2k ]
n  (i+k−2j) mod(2p) (i+k−2j+1) mod(2p)
(i) (i+k) mod(2p)  Y x0
 Y  Fpl+j−2 + x−1 Fpl+j−1
x2pn+k = x0
 h k−2p+2 i x(i+k−2j) mod(2p) F (i+k−2j+1) mod(2p)


l=1 
j= 0 pl+j−1 + x−1 Fpl+j
2
h k+2p−2 i 
2 (i−k+2j+1) mod(2p) (i−k+2j+2) mod(2p) 
x−1 Fpl+k−j−2 + x−2 Fpl+k−j−1 
Y 

×

(i−k+2j+1) mod(2p) (i−k+2j+2) mod(2p)


k
j=[ 2 ] x −1 F pl+k−j−1 + x −2 F pl+k−j



[ 2k ] (i+k−2j) mod(2p) (i+k−2j+1) mod(2p)
Y x 0 Fj−2 + x
−1 Fj−1
× (i+k−2j) mod(2p) (i+k−2j+1) mod(2p)
j=1 x0 Fj−1 + x−1 Fj
k (i−k+2j−1) mod(2p) (i−k+2j) mod(2p)
Y x−1 Fk−j−1 + x−2 Fk−j
× (i−k+2j−1) mod(2p) (i−k+2j) mod(2p)
,
j=[ k ]+1 x−1 Fk−j + x−2 Fk−j+1
2

168
CNMA’2022
Second National Conference on
Mathematics and its Applications

for n ∈ N0 , k ∈ {0, 1, ..., 2p − 1} , i ∈ {1, ..., 2p} , p ∈ N, where (Fn )n≥−1 is the solution to the following
difference equation
Fn+1 = Fn + Fn−1 , n ∈ N0 ,
satisfying the initial conditions F−1 = 0, F0 = 1. The sequence (Fn )n≥−1 is called the well-known Fi-
bonacci sequence in literature.

References
[1] R. Abo-Zeid., H. Kamal. (2019). Global behavior of two rational third order difference equa-
tions. Universal Journal of Mathematics and Applications, 2(4), 212 − 217.

[2] Q. Din. (2014). Global stability of a population model. Chaos, Solitons & Fractals 59, 119 −
128.

[3] Q. Din. (2015). Global behavior of a plant-herbivore model. Advances in Difference Equa-
tions, 2015(1), 119, 12.

[4] M. Kara., Y. Yazlik. (2022). On the solutions of three-dimensional system of difference equa-
tions via recursive relations of order two and applications. Journal of Applied Analysis and
Computation,12 (2), 736 − 753.

[5] S. Stević. (2018). On a two-dimensional solvable system of difference equations. Electronic

Journal of Qualitative Theory of Differential Equations, 2018 (104), 1 − 18.

169
CNMA’2022
Second National Conference on
Mathematics and its Applications

Adomian Decomposition Method for

Solving a Class of Nonlinear Fractional
Differential Equations

Lina CHETIOUI1 [Link]@[Link]

Ali KHALOUTA1 [Link]@[Link]
1 Laboratory of Fundamental and Numerical Mathematics,Department of Mathematics, Faculty of Sciences, Fer-
hat Abbas Sétif University 1, 19000 Sétif, Algeria

Abstract: The purpose of this work is to present a new kind of analytical methods, the
so-called Adomian decompostion method (ADM), to represent the solutions of a class of
nonlinear fractional differential equations, in particular the nonlinear fractional Bratu-type
equation. Fractional derivatives are described in the Caputo type. Illustrative examples are
presented to demonstrate the accuracy and effectiveness of the proposed method. The results
reveal that the ADM is very efficient, simple and powerful to formulate these solutions.

Keywords: fractional Bratu-type equation, Caputo fracional derivative, Adomian de-

compostion method, approximate solution
2010 Mathematics Subject Classification: Primary 35L05, 35R11, 35A22, 26A33.

Introduction
It is well known that the classical Bratu-type equation is of the form

u 00 (t) + λeu(t) = 0, 0 < t < 1.

The exact solution is [4]   

 cosh t − 21 θ2 

u(t) = −2 ln   ,
cosh( θ4 )


where θ satisfies
√ θ
 
θ = 2λ cosh .
4

The problem has zero, one or two solutions when λ > λc , λ = λc , and λ < λc , respectively, where
the critical value λc satisfies the equation

1p θ
 
1= 2λc sinh c ,
4 4
where
λc = 3.513830719.

170
CNMA’2022
Second National Conference on
Mathematics and its Applications

The classical Bratu-type equation was used in a large variety of applications such as the fuel
ignition model of the thermal combustion theory, the model of the thermal reaction process, the
Chandrasekhar model of the expansion of the universe, questions in geometry and relativity about
the Chandrasekhar model, chemical reaction theory, radiative heat transfer and nanotechnology
[1, 2, 4].
Several authors have proposed various methods to obtain the approximate solution of the classical
Bratu-type equation. The motivation of this work is to propose a powerful method called Admian
decomposition method (ADM) to get an approximate analytical solution of nonlinear fractional
Bratu-type equations in the form

D α u(t) + λeu(t) = 0, 0 < t < 1, λ ∈ R,

with the initial conditions

u(0) = u0 , u 0 (0) = u1 ,
where D α is the fractional derivative operator in the sense of Caputo of order α with 1 < α ≤ 2.

Review of Fractional Calculus Theory

In this section, we present some necessary definitions and essentials results from fractional calcu-
lus theory

Definition 31. [4] A real function u(t), t > 0, is considered to be in the space Cµ , µ ∈ R if there exists
a real number p > µ, so that u(t) = t p h(t), where h(t) ∈ C ([0, ∞[), and it is said to be in the space Cnµ if
u (n) ∈ Cµ , n ∈ N.

Definition 32. [4] The Riemann-Liouville fractional integral operator I α of order α for a function
u ∈ Cµ , µ ≥ −1 is defined as follows

1 Rt

 α−1
 Γ (α) (t − ξ) u(ξ)dξ, α > 0, t > 0,


α

I u(t) = 
 0

u(t), α = 0,



where Γ (.) is the well-known Gamma function.

Definition 33. [4] The fractional derivative of u(t) in the Caputo sense is defined as follows

Zt
1
α
D u(t) = I n−α n
D u(t) = (t − ξ)n−α−1 u (n) (ξ)dξ, t > 0,
Γ (n − α)
0

where n − 1 < α ≤ n, n ∈ N, f ∈ Cn−1 .

For the Riemann-Liouville fractional integral and Caputo fractional derivative, we have the fol-
lowing relation
n−1
X tk
α α
I D u(t) = u(t) − u (k) (0+ ) , t > 0.
k!
k=0

Main Results
In this section, we present the methodology of the ADM

171
CNMA’2022
Second National Conference on
Mathematics and its Applications

Theorem 87. Consider the following nonlinear fractional Bratu-type equation

D α u(t) + λeu(t) = 0, (176)

with the initial conditions

u(0) = u0 , u 0 (0) = u1 , (177)
where D α is the fractional derivative operator in the sense of Caputo of order α with 1 < α ≤ 2.
Then, by ADM, the approximate analytical solution of equations (176)-(177) is given in the form of
infinite series which converges very rapidly to the exact solution as follows
∞
X
u(t) = un (t),
n=0

where
un (t) = u0 + u1 t − λI α (An−1 ) ,
and An are Adomian polynomials [5] of the nonlinear term eu(t) and it can be calculated by formula
given below  ∞ 
1 d n  X i 
An = N  ϕ ui  , n = 0, 1, 2, ...
n! dϕ n  
i=0 ϕ=0

References
[1] R. Buckmire, Application of a Mickens finite-difference scheme to the cylindrical Bratu–
Gelfand problem, Numer. Methods Partial Differential Equations 20(3) (2004), 327–337.

[2] A. Khalouta, A novel representation of numerical solution for fractional Bratu-type equation,
Advanced Studies: Euro-Tbilisi Mathematical Journal, 15(1) (2022), 93–109.

[3] A.A. Kilbas, H.M. Srivastava and J.J. Trujillo, Theory and Application of Fractional Differ-
ential Equations, Elsevier, North-Holland, 2006.

[4] M.I. Syam and A. Hamdan, An efficient method for solving Bratu equations, Appl. Math.
Comput. 176(2) (2006), 704–713.

[5] Y. Zhu, Q. Chang and S. Wu, A new algorithm for calculating Adomian polynomials, Appl.
Math. Comput. 169 (2005), 402–416.

172
CNMA’2022
Second National Conference on
Mathematics and its Applications

Solving the nonlinear fractional Logistic

equation using fractional differential
transform method

Fatima HATHAT 1 [Link]@[Link]

Ali KHALOUTA 1 [Link]@[Link]
1 Laboratory of Fundamental and Numerical Mathematics, Department of Mathematics, Ferhat Abbas Sétif Uni-
versity 1, 19000 Sétif, Algeria

Abstract: In this work, the fractional differential transform method (FDTM) is applied to
solve the nonlinear fractional Logistic equation. The fractional derivatives are described in
the Caputo sense. In addition, examples that illustrate the pertinent features of this method
are presented and the results of the study are discussed. The results show that the introduced
method is a powerful tool for solving the nonlinear fractional differential equations.

Keywords: fractional Logistic equation, Caputo fracional derivative, fractional differen-

tial transform method, numerical solution
2010 Mathematics Subject Classification: Primary 35L05, 35R11, 35A22, 26A33.

Introduction
The fractional differential equations have attracted much more interest of mathematicians and
physicists which provides an efficiency for the description of many practical dynamical arising
in engineering and scientific disciplines such as, physics, biology, electrochemistry, chemistry,
economy, electromagnetic, control theory and viscoelasticity [1, 2, 4]. Recently,various numerical
and analytical methods have been applied for the approximate solutions of fractional differential
equations.
The main objective of this work is to present the fractonal differential transform methd (FDTM)
to construct the approximate solutions of the following nonlinear fractional Logistic equation

D α u(t) = ru(t)(1 − u(t)), t > 0,

with the initial condition

u(0) = u0 ,
where D α is the fractional derivative operator in the sense of Caputo of order α with 0 < α ≤ 1.
When α = 1, this equation becomes a classical Logistic equation. The important application of
the logistic equation is that it is a model of population growth. The population size at time t is
denoted with u(t) and the constant r > 0 defines the growth rate. Another application of Logistic
equation is in medicine, where the logistic differential equation is used to model the growth of
tumors.

173
CNMA’2022
Second National Conference on
Mathematics and its Applications

Basic definitions
In this section, we give some definitions and properties of the fractional calculus and fractional
differential transform method.
Definition 34. [4] A real function u(t), t > 0, is considered to be in the space Cµ , µ ∈ R if there exists
a real number p > µ, so that u(t) = t p h(t), where h(t) ∈ C ([0, ∞[), and it is said to be in the space Cnµ if
u (n) ∈ Cµ , n ∈ N.
Definition 35. [4] The Riemann-Liouville fractional integral operator I α of order α for a function
u ∈ Cµ , µ ≥ −1 is defined as follows

1 Rt

 α−1
 Γ (α) (t − ξ) u(ξ)dξ, α > 0, t > 0,


I α u(t) = 

 0

u(t), α = 0,



where Γ (.) is the well-known Gamma function.

Definition 36. [4] The fractional derivative of u(t) in the Caputo sense is defined as follows
Zt
1
α
D u(t) = I n−α n
D u(t) = (t − ξ)n−α−1 u (n) (ξ)dξ, t > 0,
Γ (n − α)
0

where n − 1 < α ≤ n, n ∈ N, f ∈ Cn−1 .

For the Riemann-Liouville fractional integral and Caputo fractional derivative, we have the fol-
lowing relation
n−1
X tk
α α
I D u(t) = u(t) − u (k) (0+ ) , t > 0.
k!
k=0

Definition 37. [5] Let u(t) is analytic and differentiated continuously function with regard to time t,
in the domain of interest, then the differential transform of u(t) is given by
∞
1 dk
X " #
U (k) = u(t) ,
k! dt k t=t0
k=0

where u(t) is the original function and U (k) is the transformed function
Definition 38. [5] The fractional differential transform of the function u(x) is defined as
∞ " kα #
X 1 d
U (k) = u(t) , (178)
Γ (kα + 1) dt kα t=t0
k=0

where α is a parameter describing the order of fractional derivative in the Caputo sense, u(t) is the
original function and U (k) the transformed function.
Definition 39. [5] The inverse fractional differential transform of U (k) is defined as
∞
X
u(t) = U (k)(t − t0 )kα . (179)
k=0

Combining equations (178) and (179), we get

174
CNMA’2022
Second National Conference on
Mathematics and its Applications

∞ " kα #
X 1 d
u(t) = u(t) (t − t0 )kα . (180)
Γ (kα + 1) dt kα t=t0
k=0

In particular, for t0 = 0, equation (180) becomes

∞ " kα #
X 1 d
u(t) = u(t) t kα .
Γ (kα + 1) dt kα t=0
k=0

From the above definitions, the fundamental operations of the FDTM are given by the following
theorems.
Theorem 88. [5] Let U (k), V (k) and W (k) be the fractional differential transforms of the functions
u(t), v(t) and w(t) respectively, then

(1) if
w(t) = λu(t) + µv(t), λ, µ ∈ R,
then
W (k) = λU (k) + µV (k).

(2) if
w(t) = u(t)v(t),
then
k
X
W (k) = U (r)V (k − r).
r=0

(3) if
d nα
w(t) = u(t),
dt nα
then
Γ ((k + n) α + 1)
W (k) = U (k + n), n = 1, 2, ....
Γ (kα + 1)

Main Results
In this section, we present the FDTM for solving nonlinear fractional Logistic equation
Theorem 89. Consider the following nonlinear fractional Logistic equation

D α u(t) = ru(t)(1 − u(t)), (181)

with the initial conditions

u(0) = u0 , (182)
where D α is the fractional derivative operator in the sense of Caputo of order α with 0 < α ≤ 1.
The FDTM gives the solution of equations (181) and (182) in the form of infinite series which converges
rapidly to the exact solution as follows
∞
X
u(t) = U (k)t kα ,
k=0

where U (k) is the fractional differential transformed function of u(t).

175
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] A. Khalouta, A novel representation of numerical solution for fractional Bratu-type equation,
Advanced Studies: Euro-Tbilisi Mathematical Journal, 15(1) (2022), 93–109.

[2] A. Khalouta, Closed-Form Solutions to Some Nonlinear Fractional Partial Differential Equa-
tions Arising in Mathematical Sciences, Palestine Journal of Mathematics, 11 (2022), 113–
126.

[3] H. Khan, M. Alipour, R. Ali Khan, H. Tajadodi, A. Khan On Approximate Solution Of

Fractional Order Logistic Equations By Operational Matrices Of Bernstein Polynomials, Jour-
nal of Mathematics and Computer Science, 14 (2015), 222–232.

[4] A.A. Kilbas, H.M. Srivastava and J.J. Trujillo, Theory and Application of Fractional Differ-
ential Equations, Elsevier, North-Holland, 2006.

[5] J.K. Zhou, Differential Transformation and Its Applications for Electrical Circuits, Huazhong
University Press, Wuhan, China, 1986.

176
CNMA’2022
Second National Conference on
Mathematics and its Applications

On the Hyers-Ulam Stability of some

ordinary Differential Equations

Said Fatma 1 saidfatma2323@[Link]

Boukerrioua Khaled 1
1 Department of Mathematics, Badji Mokhtar-Annaba University, Annaba, Algeria

Abstract: In this work, we study the Ulam-Hyers stability of some ordinary differential equa-
tions based on Gronwall lemmas.
Keywords: Cauchy problem, Differential equation, Integral equation, Gronwall lemma, Ulam-
Hyers stability.
2010 Mathematics Subject Classification: 34K05, 37C75, 45D05.

Introduction
In [3, 4, 5] Rus has obtained some results regarding Ulam stability of differential and integral
equations, using Gronwall inequalities method and weak Picard operators technique. In this
work, based on the results presented in [3], we study the Ulam-Hyers stability of some ordinary
differential equations by Gronwall lemma techniques.

Preliminaries
We consider the following problems:
1) the differential equation

.
x(t) = f (t, x(t)), t ∈ I, (183)
with its solution set denoted by S0 ;
2) the differential inequation
.
| x(t) − f (t, x(t)) |≤ ε, t ∈ I, (184)

with the solution set (for each ε > 0) denoted by Sε . Notice that a function y ∈ C 1 (I, B) is a solution
of (184) if and only if there exists a function g ∈ C(I, B) (which may depend on y) such that

(i) | g(t) |≤ ε, for all t ∈ I,

.
(ii) y(t) = f (t, y(t)) + g(t), for every t ∈ I,

A solution to inequation (184) is called an ε-solution of equation (183).

Now, we mention several foundational definitions and results

177
CNMA’2022
Second National Conference on
Mathematics and its Applications

Definition 40. The differential equation (183) is said to be Ulam-Hyers stable if there exists a real
number Cf > 0 such that, for each ε > 0, there exists a retraction rε : Sε → S0 satisfying

| y(t) − rε (y)(t) |≤ Cf ε, for each solution,y ∈ C 1 (I, B) of (184) and every t ∈ I.

Definition 41. . Differential equation (183) is said to be generalized Ulam-Hyers stable if there exists
a function θf ∈ C(R+ , R+ ) with θ(0) = 0 such that, for each ε > 0, there exists a retraction rε : Sε → S0
satisfying
| y(t) − rε (y)(t) |≤ θf ε, for each solution,y ∈ C 1 (I, B) of (2) and every t ∈ I.
Lemma 27. (Gronwall Lemma I.) Let J be an interval of the real axis, t0 ∈ J and the functions h, k, u ∈
C(J, R+ ). If
Zt
u(t) ≤ h(t)+ | k(s)u(s)ds |, for all t ∈ J.
t0

then
Zt Rt
k(σ )dσ
u(t) ≤ h(t)+ | h(s)k(s)e s ds |, for all t ∈ J.
t0

If, in addition, h is increasing, then (J := [a, b[)

Rt
k(s)ds
u(t) ≤ h(t)e a ds, t ∈ [a, b[.

Lemma 28. (Gronwall Lemma II.) Let h ∈ C([a, b], R+ ) and β > 0 with β(b − a) < 1. If u ∈ C([a, b], R+ )
satisfies the relation
Zb
u(t) ≤ h(t) + β u(s)ds, for t ∈ [a, b].
a

then
Zb
β
u(t) ≤ h(t) + h(s)ds, for t ∈ [a, b].
1 − β(b − a)
a

References
[1] D. Bainov, P. Simeonov, Integral Inequalities and Applications, Springer, Netherlands, 1992.

[2] I.A. Rus , Fixed points, upper and lower fixed points: abstract Gronwall lemmas, Carpathian
J. Math . 20 (2014), 125–134.

[3] I.A. Rus , Gronwall lemma approach to the Hyers-Ulam-Rassias stability of an integral equa-
tion, In: Nonlin ear Analysis and Variational Problems, P. Pardalos, Th.M. Rassias and A.A.
Khan , Springer Carpathian . (2009), 147–152.

[4] I.A. Rus, Ulam stability of ordinary differential equations, Studia Univ. Babes¸-Bolyai Math.
54 (2009), 125–133.

[5] I.A. Rus, Remarks on Ulam stability of the operatorial equations, Fixed Point Theory. 10
(2009), 305–320.

178
CNMA’2022
Second National Conference on
Mathematics and its Applications

Existence and estimates for solutions to

coincidence point and to differential
equations for mappings acting in an
arbitrary set

Benarab Sarra 1 [Link]@[Link]

1 Department of Functional Analysis, Derzhavin Tambov State University, Russia

Abstract: Sufficient conditions are obtained for the existence of a coincidence point for
two mappings acting from a partially ordered space into an arbitrary set (obviously, such
mappings cannot be monotone), an estimate for such a point is found. The established results
are applied to the study of differential equations. An assertions on the existence and estimates
of the solution of the Cauchy problem are obtained.

Keywords: partially ordered space, fixed point, coincidence point, existence of solu-
tion

Introduction
The authors of [1], [2] generalize the notion of a covering for mappings that act from a partially
ordered space to a set, where some reflexive binary relation is given (this relation is not necessarily
antisymmetric or transitive, i.e., it does not represent an order). In [1], one proves a certain asser-
tion on a coincidence point of two maps, while in [2], one establishes a Chaplygin-type theorem
on the operator inequality. In this paper, we demonstrate that assumptions of coincidence points
theorems proved in [3], [4] can be weakened, namely, that is suffices to require that some binary
relation should be given in the set of values of considered mappings. Numerous extensions of the
Chaplygin theorem to systems of equations and to various functional-differential equations also
make use of an auxiliary assumption on the monotonicity of functions and mappings generating
the equations with respect to appropriate argument (see [[5], Sec. 1]). The proofs of such asser-
tions are, as a rule, based on the results about fixed points of monotone mappings in partially
ordered spaces.

Main results
COINCIDENCE POINTS IN PARTIALLY ORDERED SPACES

Let (X, ) be a partially

n ordered
o space. For its elements x, u ∈ X we write x ≺ u if x  u and x , u.
Denote OX (u) := x ∈ X : x  u . Suppose we are given a set Y , ∅, and let mappings ψ, ϕ : X → Y

179
CNMA’2022
Second National Conference on
Mathematics and its Applications

be defined. Consider the problem on the existence of an element x ∈ X satisfying the equation

ψ(x, x) = ϕ(x, x). (185)

Suppose we are given a nonempty set X ⊂ X. Let us define the set Ξ(X , ψ, ϕ) of chains S in the
space X such that S ⊂ X and one has the relation

for each u ∈ S there exists an x ∈ X such that x  u and ψ(x, u) = ϕ(x, u).
Theorem 90. Let the following conditions be satisfied:

(a) There exist u0 ∈ X and x0 ∈ X such that x0  u0 and ψ(x0 , u0 ) = ϕ(x0 , u0 ).

(b) For any u ∈ X and x ∈ X such that x ≺ u and ψ(x, u) = ϕ(x, u), there exist elements v ∈ X and
w ∈ X for which w  v ≺ u and ψ(w, v) = ϕ(w, v).

(c) For an arbitrary infinite chain S ∈ Ξ(X , ψ, ϕ) there exist elements ve ∈ X and w
e ∈ X satisfying the
relations
e ve) and w
e ve) = ϕ(w,
ψ(w, e  ve  u for each u ∈ S.
Then there exists a solution x = ξ ∈ X of Eq. (185) such that ξ  u0 .

EXISTENCE AND ESTIMATES FOR SOLUTIONS TO DIFFERENTIAL EQUATIONS

By M n we denote the space of (Lebesgue) measurable functions x : [0, 1] → Rn with the ”usual”
ordering: for x = (x1 , ..., xn ) ∈ M n and u = (u1 , ..., un ) ∈ M n we set x ≤ u if xi (t) ≤ ui (t), i = 1, ..., n,
for almost all t ∈ [0, 1]. By Ln we denote the space of n-dimensional functions that are (Lebesgue)
integrable on [0, 1]. This space is a subspace of the partially ordered space M n . By AC n we denote
a subspace of the space Ln consisting of absolutely continuous n-dimensional functions. Suppose
we are given functions fi : [0, 1] × Rn × Rn × R → R, i = 1, n. Consider the system of differential
equations
fi (t, x, ẋ, ẋi ) = 0, t ∈ [0, 1], i = 1, n. (186)
unsolved for the solution derivatives. A solution to system (186) is a function x ∈ AC n satisfying
all equations in this system for almost all t ∈ [0, 1]. In this section, for system (186) we will produce
assertions on the existence and estimates for solutions to the Cauchy problem.
Suppose that the following condition is satisfied for all i = 1, ..., n.

(F↓) For almost all t ∈ [0, 1], any x, v ∈ Rn and any yi ∈ R the function fi (·, x, v, yi ) : [0, 1] → R is
measurable, the function fi (t, ·, ·, yi ) : Rn × Rn → R is decreasing and right continuous in each of
the arguments x1 , . . . , xn and v1 , . . . , vn , and the function fi (t, x, v, ·) : R → R is continuous.

Under this assumption, we have the following assertion.

Theorem 91. Suppose that for some functions ν, η ∈ AC n such that ν(0) ≥ η(0) and ν̇ ≥ η̇, one has the
inequalities
fi (t, ν(t), ν̇(t), ν̇i (t)) ≥ 0, fi (t, η(t), η̇(t), η̇i (t)) ≤ 0, i = 1, n.
Then for each A ∈ Rn , such that η(0) ≤ A ≤ ν(0), there exists a solution x ∈ AC n to the Cauchy problem
for system (186) with the initial condition
x(0) = A,
satisfying the inequalities
η̇ ≤ ẋ ≤ ν̇.

180
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] S. Benarab, E. S. Zhukovskiy, On the Conditions of Existence of Coincidence Points for Map-
pings in Partially Ordered Spaces, Tambov University Reports. Series: Natural and Techni-
cal Sciences 23:121 (2018), 10–16.

[2] S. Benarab, E. S. Zhukovskiy, About Covering Mappings with Values in the Space with a Re-
flexive Binary Relation, Tambov University Reports. Series: Natural and Technical Sciences
23:122 (2018), 210–215.

[3] A. V. Arutyunov, E. S. Zhukovskiy, S. E. Zhukovskiy, Coincidence Points Principle for Map-

pings in Partially Ordered Spaces, Topology and its Appl. 179:1 (2015), 13–33.

[4] A. V. Arutyunov, E. S. Zhukovskiy, S. E. Zhukovskiy, Coincidence Points of Mappings in

Partially Ordered Spaces, Dokl. Ross. Akad. Nauk 453:5 (2013), 475–478.

[5] Izbrannye trudy N. V. Azbeleva (Selected Works by N. V. Azbelev), V. P. Maksimova,

L. F. Rakhmatullina, Theory of Oscillations, Eds., MoscowIzhevsk: Inst. Kompyut. Issled.,
2012.

181
CNMA’2022
Second National Conference on
Mathematics and its Applications

Generating hidden chaotic attractors

from a new 3D dynamical system with
no-equilibrium and strong chaoticity

Khaled Benkouider 1 [Link]@[Link]

1 Non Destructive Testing Laboratory , Department of automatic, University of Jijel

Abstract: Because of its theoretical and practical significance, constructing chaotic systems
with hidden attractors is an exciting area of research. In this paper, a new 3D dynamical
system is presented. It can exhibit hidden chaotic attractors with a large positive Lyapunov
exponent making it much stronger than at least 51 reported systems.

Keywords: Dynamical systems, Differential equations, chaos, bifurcation, electronic cir-

cuit.
2010 Mathematics Subject Classification: Primary 65P20, 34H10, 34C23, 34D08, 28A80,
37G35.

Introduction
Due to their tremendous complexity, chaotic systems are used in a variety of technical and sci-
entific fields. Recently, many chaotic systems have been reported in the literature and published
in high-impact journals [3]-[4]. Moreover, there has been increasing attention to chaotic systems
with hidden attractors wherein their basins of attraction are not connected to any equilibrium.
After this idea was introduced, many chaotic systems with hidden attractors has been studied
[3]-[4]. In this work, we describe a new 3D chaotic system with hidden chaotic attractors. We
show that the positive Lyapunov exponent of the proposed system is larger than seven, we ob-
serve that the new system is highly chaotic. In addition, we provide a comprehensive dynamic
analysis, including bifurcation diagrams and Lyapunov exponents. We introduce a comparative
section which shows that our system is much stronger than at least 51 reported systems. Finally,
we construct an electrical circuit for the implementation of the new system using Multisim.

A new 3D chaotic system

In this work, we present a novel 3D chaotic system described by:



 ẋ = a(y − x) + yz


ẏ = x(b − z) − 1 (187)





 ż = x2 + xz − cz


Where x, y, z are the state variables and a, b, c are positive constant parameters.
When we take the initial state as [1, 1, 1]T and parameters as a = 20, b = 50, c = 10, system (187) is

182
CNMA’2022
Second National Conference on
Mathematics and its Applications

chaotic with no-equilibrium points. It is implied that system (187) under current parameters has
hidden chaotic attractors. It can be seen from Figure 9 the chaotic nature of all hidden attractors.

Figure 9: Hidden chaotic attractors of the new system (187).

When we choose the previous values of system (187)’s parameters, the corresponding Lyapunov
exponents are calculated using Wolf’s algorithm incorporated on MATLAB as:



 LE1 = 7.196


LE2 = 0 (188)





 LE = −37.197

3

From (188), the sum of Lyapunov exponents is negative confirming that the new 3D system (187)
is dissipative. Since the value of maximum Lyapunov exponent (MLE) in Eq. (188) is large (Over
7), we conclude that the 3D system is highly chaotic and this property is very useful for applica-
tions of the chaotic system in cryptosystems and secure communications.
The Kaplan-Yorke dimension of the system (187) is calculated by using the following formulate:

LE1 + LE2
DKY = 2 + = 2.193 (189)
LE3
From (189) we see that the Kaplan-Yorke dimension is fractional. Hence, the new 3D system (187)
generates complex chaotic attractors.

Table 6 shows the maximum Lyapunov exponents (MLE), the Kaplan-Yorke dimension and the
nature of attractors of the new system (187) and those of Sign and Roy system [3]. Singh and Roy
showed in their paper published in Nonlinear Dynamics Journal (IF=5.741) that their system has
a larger MLE and a highest DKY than those of 50 reported systems. It can be seen from Table 6

System MLE DKY Nature of attractors

Sing and Roy[3] 5.578 2.16 Self-exited
System (187) 7.196 2.19 Hidden

Table 6: MLE, DKY and nature of attractors of system (187) and system[3].

that the new system (187) has Larger MLE and higher DKY than those of Sign and Roy system.
Hence, our system is much stronger than at least 51 reported chaotic systems (See Singh and Roy
paper [3]). It is a big addition to the chaotic systems family.

Dynamical analysis
The dynamical features of system (187) versus its parameters are explored in this part by uti-
lizing Lyapunov exponents spectrums and bifurcation diagrams plotted in Figure 10

183
CNMA’2022
Second National Conference on
Mathematics and its Applications

Figure 10: Bifurcation diagrams and LEs spectrums of system (187) for specifics intervals of pa-
rameters.

Figure 11: The schematic of the circuit and Multisim outputs.

Fix the parameters b and c as b = 50 and c = 10. When 10 < a < 14, system (187) has no positive
Lyapunov exponents which means that it exhibits periodic behavior. The corresponding LEs are:
LE1 = 0, LE2 = −5.248, LE3 = −16.756 for a = 12. When 14 < a < 20, system (187) exhibits chaotic
behavior with one positive LE. the corresponding LEs are: LE1 = 6.358, LE2 = 0, LE3 = −34.756 for
a = 18.
Fix the parameters a and c as a = 20 and c = 10. When 10 < b < 13, system (187) has no positive
Lyapunov exponents which means that it exhibits periodic [Link] corresponding LEs are:
LE1 = 0, LE2 = −1.562, LE3 = −28.440 for b = 12, . When 13 < b < 50, system (187) exhibits chaotic
behavior with one positive LE. The corresponding LEs are: LE1 = 6.526, LE2 = 0, LE3 = −36.528
for b = 45.
Fix a and b as a = 20 and b = 50. When 10 < c < 18, system (187) exhibits chaotic behavior with
the following values of LEs are: LE1 = 6.473, LE2 = 0, LE3 = −38.473 for c = 12. When 18 < c < 20,
system (187) exhibits periodic behavior. The corresponding LEs are: LE1 = 0, LE2 = −10.288, LE3 =
−28.712 for c = 19.
Circuit design
The analog circuit of system (187) is implemented using Multisim as given in Figure 11. By
applying Kirchhoff’s laws to this circuit, the corresponding circuital equations of the system (187)
are the following:



 ẋ = − R 1C x + R 1C y + R 1C yz

 1 1 2 1 3 1

1 1 1
ẏ = x xz R6 C2 V1


 R4 C2 − R5 C2 −


 ż = 1 2 1 1
R C x + R C xz − R C z

7 3 8 3 9 3

C1 = C2 = C3 = 1nf






R1 = R2 = 20kΩ, R4 = R9 = 40kΩ





R3 = R5 = R6 = R7 = R8 = 400kΩ








 R10 = R11 = R12 = R13 = 100kΩ
MultiSim outputs are shown in Figure 11, which agree with the MATLAB outputs shown in Figure
9. Results confirm the physical feasibility of our proposed mathematical model.

184
CNMA’2022
Second National Conference on
Mathematics and its Applications

Conclusion
In this work, we described the model of a new no-equilibrium system which exhibits much
stronger hidden chaos than at least 51 reported systems. We studied basic properties of the model
by means of phase portraits, Lyapunov exponents, Kaplan-York dimension, and bifurcation dia-
grams. Then, we confirmed the feasibility of the new system by designing its electronic circuit
using Multisim.

References
[1] J. P. Singh, and B.K. Roy, A more chaotic and easily hardware implementable new 3-D chaotic
system in comparison with 50 reported systems, Nonlinear Dynamics, 93.3 (2018), 1121-
1148.

[2] V. F. Signing, J. Kengne, and L. K. Kana, Dynamic analysis and multistability of a novel
four-wing chaotic system with smooth piecewise quadratic nonlinearity, Chaos, Solitons and
Fractals, 113 (2018), 263-274.

[3] S. Nag Chowdhury, and D. Ghosh, Hidden attractors: A new chaotic system without equi-
libria, The European Physical Journal Special Topics, 229.6 (2020), 1299-1308.

[4] Y. Xie, J. Yu, S. Guo, Q. Ding, and E. Wang, Image encryption scheme with compressed
sensing based on new three-dimensional chaotic system, Entropy, 21.9 (2019), 819.

185
CNMA’2022
Second National Conference on
Mathematics and its Applications

An existence and uniqueness result for

Caputo fractional differential equations

Bochra Azzaoui 1 bouchraa740@[Link]

Brahim Tellab 1
1 Department of mathematics , Kasdi Merbah University, Ouargla

Abstract: We investigate the existence and uniqueness of solutions of boundary value

problems of nonlinear fractional integro-differential equations involving Caputo fractional
derivative. Some new results are obtained by using Banach contraction principle and Kras-
noselskii’s fixed point theorem.

Keywords: Fractional integro-di fferential equations, fractional calculus,

fixed point.
2010 Mathematics Subject Classification:34A08, 34B15, 35R11

Introduction
Fractional derivative theory is a subject almost as old as classical calculus as we know it today,
its origins going back to the end of the 17th century [6], the time when Newton and Leibniz devel-
oped the foundations of differential and integral calculus. In particular, Leibniz introduced the
dnf
symbol dt n to designate the neme derived from a function f . When he announced in a letter to
the Hospital (apparently with the implicit assumption that n ∈ N ), the Hospital replied : What
does d n f dt n si n = 21 . This letter from the Hospital, written in 1695, is now accepted as the first
incident of what we call fractional derivation, and the fact that the Hospital asked specifically for
n = 12 , i.e. a fraction (rational number) actually gave rise to the name of this part of mathematics.
The recent development in the theory, methods and applications of fractional calculus has con-
tributed towards the popularity and importance of the subject. The tools of fractional calculus
have been effectively applied in the modeling of many physical and engineering phenomena. Ex-
amples include physics, chemistry, biology, etc. [2],[1]. For some recent work on the topic, we
refer to [3] [4] [5] and the references therein.
In this work, we study the existence of positive solutions for the following boundary value prob-
lem for a differential equation involving the Caputo fractional order derivative:
Zt
c α
D u(t) = f (t, u(t), g(s, u(s))ds, t ∈ [0, 1] (190)
0
u(0) = 0, u(1) = au(ξ), 0<ξ ≤1 (191)

Where c D α u(t) denotes the Caputo fractional derivative of order α, f and g are given continuous
function, and a is a positive real constant.

Main results

186
CNMA’2022
Second National Conference on
Mathematics and its Applications

Lemma 29. Let h ∈ C[0, 1] and aξ , 1. Then, the problem

c
D α u(t) = h(t),
u(0) = 0, u(1) = au(ξ), 0 < ξ ≤ 1

has the solution

Z 1
t
u(t) = − (1 − s)α−1 h(s)ds
(1 − aξ)Γ (α) 0
Zξ
at
+ (ξ − s)α−1 h(s)ds
(1 − aξ)Γ (α) 0
Zt
1
+ (t − s)α−1 h(s)ds,
Γ (α) 0

Theorem 92. Assume that (H1), (H2), (H3) and (H4) holds. If

1 + |a| ξ α
!
K
+1 < 1.
|1 − aξ| Γ (α + 1)

Then the Fractional boundary value problem (190)-(191) has a unique solution on [0, 1]

Theorem 93. Suppose that the condition (H1) and (H3) is satisfied and there exists a positive function
λ. if
 Zt 
f (t, u(t), g(t, u(t))dt) ≤ λ(t)kuk
0
and
1 + |a|ξ α
!
µ= λ < 1,
|1 − aξ|Γ (α + 1)
holds, Then the problem (190), (191) has at least one solution on
[0, 1]

References
[1] Podlubny, I, Fractional Differential Equations. Academic Press, San Diego 1999

[2] Sabatier, J, Agrawal, OP, Machado,Advances in Fractional Calculus Theoretical Develop-

ments and Applications in Physics and Engineering. Springer, Dordrecht 2007.

[3] Baleanu, D, Mustafa, OG,On the global existence of solutions to a class of fractional differ-
ential equations. Comput. Math. Appl. 59 (2010), 1835-1841

[4] Akyildiz, FT, Bellout, H, Vajravelu, K, Van Gorder, RA, Existence results for third order
nonlinear boundary value problems arising in nano boundary layer fluid flows over stretching
surfaces. Nonlinear Anal., Real World Appl. 12 (2011) 2919-2930

[5] Agarwal, RP, O’Regan, D, Stanek, S, Positive solutions for mixed problems of singular frac-
tional differential equations. Math. Nachr. 285 (2012) 27-41

[6] B. Ross, Fractional Calculus and its Applications, Springer-Verlag, Berlin, 1975

187
CNMA’2022
Second National Conference on
Mathematics and its Applications

On the study of an integro-differential

inclusion

A. Bouabsa 1 ayabouabsa670@[Link]
S. Saïdi1
1 LMPA Laboratory, Department of Mathematics, Mohammed Seddik Ben Yahia University, Jijel, Algeria

Abstract: The aim of the present work is to study an evolution problem driven by time-
dependent maximal monotone operators with integral perturbation in a suitable sense of
bounded variation.

Keywords: Integro-differential inclusion, maximal monotone operator, pseudo distance,

bounded variation.
2010 Mathematics Subject Classification: 34A60, 47J35, 34G25, 49J52, 49J53.

Introduction
The current work deals, in the context of a Hilbert space H, with the integro-differential inclusion
of the form
 Rt
dx
 − dr (t) ∈ A(t)x(t) + 0 f (t, s, x(s))ds dr-a.e. t ∈ I,




 x(0) = x0 ∈ D(A(0)),



where A(t) : D(A(t)) ⊂ H ⇒ H is a maximal monotone operator for all t ∈ I, and D(A(t)) stands for
the domain of operator A(t). The map r : I → [0, +∞[ is bounded variation continuous, and f is a
Carathéodory mapping that satisfies suitable conditions.

We are motivated by the recent results [2] and [3] involving integro-differential sweeping pro-
cesses. We contribute on this subject by considering evolution problems involving maximal mono-
tone operators with integral perturbations, in the bounded variation continuous case. For the
proof of the existence and uniqueness result, we use a discretization method. The well-posedness
result provides remarkable applications such as minimization and relaxation problems.

After this brief introduction, we provide useful notation and necessary preliminaries. Then, we
state our main results concerning the integro-differential inclusion above.

Preliminaries
Throughout the paper, let I := [0, T ] denotes an interval of R and H be a real Hilbert space whose
inner product is denoted by h·, ·i and the associated norm by k · k. We denote by B̄[x, r] the closed
ball of center x and radius r on H, and by B̄ the closed unit ball. For any interval I of R, we denote
by L(I) (resp. B(H)) the σ -algebra of measurable sets of I (resp. Borel σ -algebra of measurable
sets of H). On the space CH (I) of continuous maps x : I → H, we consider the norm of uniform

188
CNMA’2022
Second National Conference on
Mathematics and its Applications

convergence on I, kxk∞ = supt∈I kx(t)k.

Let µ be a positive measure on I. We denote

R by Lp (I, H, µ) for p ∈ [1, +∞[ (resp. p = +∞), the space
of measurable maps x : I → H such that I kx(t)kp dµ(t) < +∞ (resp. which are essentially bounded)
R 1
endowed with the usual norm kxkLp (I) = ( I kx(t)kp dµ(t)) p , 1 ≤ p < +∞ (resp. endowed with the
H
usual essential supremum norm k · kL∞H (I)
).
Let us recall some properties of maximal monotone operators needed in our development, see [1].
Let A : D(A) ⊂ H ⇒ H be a set-valued operator whose domain, range and graph are defined by

D(A) = {x ∈ H : Ax , ∅},
R(A) = {y ∈ H : ∃ x ∈ D(A), y ∈ Ax} = ∪{Ax : x ∈ D(A)},
Gr(A) = {(x, y) ∈ H × H : x ∈ D(A), y ∈ Ax}.

The operator A : D(A) ⊂ H ⇒ H is said to be monotone, if hy1 − y2 , x1 − x2 i ≥ 0 whenever (xi , yi ) ∈

Gr(A), i = 1, 2. It is maximal monotone, if its graph could not be contained strictly in the graph
of any other monotone operator, in this case, for all λ > 0, R(IH + λA) = H, where IH denotes the
identity map of H.
If A is a maximal monotone operator then, for every x ∈ D(A), Ax is nonempty, closed and convex.
Then, the projection of the origin into Ax, A0 (x), exists and is unique.
For λ > 0, we define the resolvent and the Yosida approximation of A respectively by, JλA =
 
(IH + λA)−1 and Aλ = λ1 IH − JλA . These operators are both single-valued and defined on the whole
space H, and we have

JλA x ∈ D(A) and Aλ (x) ∈ A(JλA x), for every x ∈ H,

kAλ (x)k ≤ kA0 (x)k, for every x ∈ D(A).

Let A : D(A) ⊂ H ⇒ H and B : D(B) ⊂ H ⇒ H be two maximal monotone operators, then we denote
by dis(A, B) (see [5]) the pseudo-distance between A and B defined by

hy − y 0 , x0 − xi
( )
0 0
dis(A, B) = sup : (x, y) ∈ Gr(A), (x , y ) ∈ Gr(B). .
1 + ||y|| + ||y 0 ||

Clearly, dis(A, B) ∈ [0, +∞], dis(A, B) = dis(B, A) and dis(A, B) = 0 iff A = B.

To prove our main results, we need the following lemmas (see [4]).
Lemma 30. Let A be a maximal monotone operator of H. If x ∈ D(A) and y ∈ H are such that

hA0 (z) − y, z − xi ≥ 0 ∀z ∈ D(A),

then x ∈ D(A) and y ∈ A(x).

Lemma 31. Let An (n ∈ N), A be maximal monotone operators of H such that dis(An , A) → 0. Suppose
also that xn ∈ D(An ) with xn → x and yn ∈ An (xn ) with yn → y weakly for some x, y ∈ H. Then x ∈ D(A)
and y ∈ A(x).
Lemma 32. Let A, B be maximal monotone operators of H. Then, one has
(1) for λ > 0 and x ∈ D(A)
q  
kx − JλB (x)k ≤ λkA0 (x)k + dis(A, B) + λ 1 + kA0 (x)k dis(A, B).

(2) For λ > 0 and x, x0 ∈ H

kJλA (x) − JλA (x0 )k ≤ kx − x0 k.

189
CNMA’2022
Second National Conference on
Mathematics and its Applications

Lemma 33. Let An (n ∈ N), A be maximal monotone operators of H such that dis(An , A) → 0 and
kA0n (x)k ≤ c(1 + kxk) for some c > 0, all n ∈ N and x ∈ D(An ). Then for every z ∈ D(A) there exists a
sequence (zn ) such that
zn ∈ D(An ), zn → z and A0n (zn ) → A0 (z).

The next is descrete version of Gronwall’s lemma (see [4])

Lemma 34. Let (αi ), (βi ), (γi ) and (ai ) be sequences of non-negative real numbers such that

ai+1 ≤ αi + βi (a0 + a1 + · · · + ai−1 ) + (1 + γi )ai for i ∈ N.

Then,
k−1
X k−1
X
ak ≤ (a0 + αj ) exp( (jβj + γj )) for j ∈ N.
j=0 j=0

We end this section by recalling the Gronwall-like differential inequality proved in [2].

Lemma 35. Let y : I → R be a non-negative absolutely continuous function and let h1 , h2 , g : I → R+

be non-negative integrable functions. Suppose for some ε > 0
Z t
1 1
ẏ(t) ≤ g(t) + ε + h1 (t)y(t) + h2 (t)(y(t)) 2 (y(s)) 2 ds a.e. t ∈ I.
0

Then, for all t ∈ I, one has

t1 Z
ε2 t
Z Zt ! !
1 1
(y(t)) ≤ (y(0) + ε) exp
2 (h(s) + 1)ds +
2 exp (h(r) + 1)dr ds
0 2 0 s
" Zt ! 21 Zt !#
1
+2 g(s)ds + ε − ε exp2 (h(r) + 1)dr
0 0
Z t ! Z t ! Z s ! 21
+2 h(s) + 1 exp (h(r) + 1)dr g(r)dr + ε ds,
0 s 0
!
h1 (t) h2 (t)
where h(t) = max 2 , 2 a.e. t ∈ I.

Main result
We stand our assumptions and establish our main results concerning the evolution problem
 Rt
dx
 − dr (t) ∈ A(t)x(t) + 0 f (t, s, x(s))ds dr-a.e. t ∈ I,




 x(0) = x0 ∈ D(A(0)).



190
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] H. Brézis, Opérateurs maximaux monotones et semi-groupes de

[2] A. Bouach, T. Haddad, L. Thibault, Nonconvex integro-differential sweeping process with

applications, arXiv:2102.11987v1.

[3] A. Bouach, T. Haddad, L. Thibault, On the discretization of truncated integro-

differential sweeping process and optimal control, J. Optim. Theory Appl.
[Link]

[4] M. Kunze and M.D.P. Monteiro Marques, BV solutions to evolution problems with time-
dependent domains, Set-Valued Anal. 5 (1997), 57-72.

[5] A.A. Vladimirov, Nonstationary dissipative evolution equations in Hilbert space, Nonlin-
ear Anal. 17(1991), 499-518.

191
CNMA’2022
Second National Conference on
Mathematics and its Applications

Predictive mathematical models of the

COVID-19 pandemic in obese people

Boubekeur Maroua Amel 1 [Link]@[Link]

Belhamiti Omar 1 [Link]@[Link]
1 Department of mathematics, University of Mostaganem Algeria

Abstract: In this work, we propose a mathematical model that highlights the narrow link
between obesity and the mortality due to COVID-19, as well as patients admitted to the ICU
for COVID-19 acute respiratory distress syndrome (ARDS). This model takes into account
different disease states and is represented mathematically by a nonlinear temporal system of
ordinary differential equations. An analysis of the stability of different equilibrium states is
also obtained to theoretically confirm the mathematical realism. Numerical simulations are
presented to explain the usefulness of the proposed model.

Keywords: Obesity, Covid-19, Model Validation, Stability Analysis.

Introduction
The Covid-19 epidemic was alerted by WHO in December 2019, and was declared a public health
emergency of international concern (USPPI) by the same organization on January 30, 2020. As of
October 17, 2021, more than 241 million cases had been confirmed and nearly 5 million deaths
worldwide. The global spread is very rapid, with 170 countries now reporting at least one case.
It is very important to understand the dynamics of the epidemic’s transmission early in order to
better control its evolution and assess the effectiveness of control measures [8].
Many studies have established that several factors have a surprising correlation with higher mor-
tality in individuals with Covid-19: arterial hypertension and smoking [8], obesity [4], diabetes,
cardiac and pulmonary pathology [3, 8]. Over the past two years, many mathematical modeling
studies of covid-19 associated with other chronic diseases have emerged, among these works, we
can cite the work in [5, 6].
Our aim in this work is to highlight the very negative effect of the COVID-19 pandemic on over-
weight and obese people. In this work, we analyze the dynamics of the co-infection of obesity and
COVID-19. Our objective is threefold,

1. We framed a deterministic model that describes obesity and COVID-19 co-infection.

2. We show the existence of an equilibrium and discuss its stability.

3. The results and discussions for the model are presented .

Main results

192
CNMA’2022
Second National Conference on
Mathematics and its Applications

Model formulation and mathematical analysis

The model is given by the following set of differential equations

Stability of equilibrium points

The system admits two equilibrium points, the disease-free-equilibrium point E0 and the endemic
equilibrium point E1 exists if the basic reproduction number R0 > [Link]
µα1 B1 ((µ + γ1 ) β2 + (µ + γ2 ) β1 )
R0 = .
(µ + η) (β1 + β2 + β3 + µ) (µ + γ1 ) (µ + γ2 )

Theorem 1: The disease-free equilibrium E0 of the system is locally-asymptotically stable in Ω,

whenever
R0 < 1.

Theorem 2: The endemic equilibrium E1 of the system is locally-asymptotically stable in Ω, when-

ever
R0 > 1.

References
[1] [Link] et al , Optimal Control of Mathematical modeling of the spread of the
COVID-19 pandemic with highlighting the negative impact of quarantine on diabetics
people with Cost-effectiveness, Chaos, Solitons & [Link]. 145(2021), 110777.

[2] [Link] et al. The effect of metformin consumption on mortality in hospitalized

COVID-19 patients: a systematic review and meta-analysis. Diabetes & Metabolic Syn-
drome, Clinical Research & Reviews. vol. 14, (2020) , 2177-2183.

[3] [Link] et al, Risk factors for disease severity, unimprovement, and mortality in COVID-
19 patients in Wuhan, China. Clinical microbiology and infection no 6,vol. 26, (2020),767-
772.

[4] [Link] et al, Obesity as a risk factor for greater severity of COVID-19 in patients with
metabolic associated fatty liver disease, Metabolism vol. 108, (2020), 154244.

[5] S. Anusha and S. Athithan, Mathematical Modelling Co-existence of Diabetes and

COVID-19: Deterministic and Stochastic Approach. (2021).

[6] Y. Marimuthu et al, COVID-19 and tuberculosis: a mathematical model based forecasting
in Delhi, India. indian journal of tuberculosis, no 2, vol. 67,(2020) , 177-181.

[7] [Link] et al. Neutrophil-to-lymphocyte ratio as an independent risk factor for mortality in
hospitalized patients with COVID-19, Journal of Infection .vol. 81,(2020): e6-e12.

[8] [Link] et al, Risk factors of critical & mortal COVID-19 cases: A systematic literature
review and meta-analysis. Journal of infection, no 2,vol. 81,(2020),e16-e25.

[9] [Link] et al., On the definition and the computation of the basic reproduction ratio
R0 in models for infectious diseases in heterogeneous populations. Journal of Mathematical
Biology„ vol. 28 (1990), 365-382.

[10] [Link] et al, SARS-CoV-2 Transmission From People Without COVID-19 Symp-
toms. JAMA network open, vol. 4,(2021) e2035057-e2035057.

193
CNMA’2022
Second National Conference on
Mathematics and its Applications

Multiple Positive Solutions for a Singular

Multipoint Boundary Value Problem Via
Fixed Point Theory for the Sum of Operators

Lydia Bouchal 1 [Link]@[Link]

Karima Mebarki 2
1 Laboratory of Applied Mathematics, Faculty of Exact Sciences, Bejaia University
2 note1

Abstract: In this work, we present a new existence criteria for existence of multiple positive
solutions for the following singular generalized Sturm-Liouville multipoint boundary value
problem
−u 00 (t) = h(t)f (t, u(t), u 0 (t)), 0 < t < 1,

m−2
au(0) − bu 0 (0) =
P
ai u(ξi ), (192)
i=1
m−2
cu(1) + du 0 (1)
P
= bi u(ξi ),
i=1

where a, b, c, d ∈ [0, ∞), 0 < ξ1 < ξ2 < ... < ξm−2 < 1 (m ≥ 3), ai , bi ∈ [0, ∞) are constants for i =
1, 2, . . . , m − 2 and ρ = ac + ad + bc > 0.
Our result is based upon a recent multiple fixed point theorems for the sum of two operators
on translate for cones of Banach spaces.

Keywords: fixed point, sum of operators, cone, Sturm-Liouville BVP, multiple positive
solutions.
2010 Mathematics Subject Classification: 47H10, 34B10, 34B24.

Introduction
One of the most important tools dealing with boundary value problems is fixed point theorems.
Recently the authors in [2] open a new direction of research in the theory of fixed point in ordered
Banach spaces for the sum of operators, Several fixed point theorems, including Krasnosel’skii
type theorems in cones, have being established for a sum of two operators.

In this work, by making use of our new theoretical results for the sum T + S where (I − T ) is a
Lipschitz invertible mapping and S is a k-set contraction on translate of cone of a Banach space,
we investigate the existence of multiple positive solutions solutions to the generalized Sturm Li-

194
 CNMA’2022
Second National Conference on
Mathematics and its Applications

ouville boundary value problem (192).

By a positive solution, it means a function u ∈ C 1 ([0, 1]) ∩ C 2 ((0, 1)) such that u(t) ≥ 0 on [0, 1] and
u satisfies (192).
Suppose the following hypotheses:

Description 1.

(H1 ) f ∈ C([0, 1] × [0, ∞) × (−∞, ∞), (−∞, ∞)),

|f (t, u, v)| ≤ k1 |u|p1 + k2 |v|p2 + k3 , t ∈ [0, 1], u, v ∈ R,

k1 , k2 , k3 , p1 , p2 are positive constants.

(H2 ) h ∈ C((0, 1), R) may be singular at t = 0 and/or t = 1

R1
and 0 G(s, s)h(s) ds < ∞.

m−2
P m−2
P
(H3 ) ∆ < 0, ρ − ai y(ξi ) > 0, ρ − bi x(ξi ) > 0.
i=1 i=1

(H4 ) g ∈ C([0, 1], (0, ∞)) be such that

Z 1
((1 − s)2 + 2(1 − s) + 2) g(s) ds ≤ A1 ,
0

for some constant A1 > 0.

Main results
Assume that the constant A1 which appears in (H4 ) satisfies the following inequality:
  p p  
A1 M (1 + (a + b)A + (c + d)B) k1 R11 + k2 R12 + k3 + R1 < 2L1 , (193)

where L1 , R1 are such that r1 < L1 < R1 with r1 a positive constant.

The main existence criteria is the following:

Theorem 94. If the assumptions (H1 )-(H4 ) and the inequality (193) are satisfied, the problem (192)
has at least three positive solutions u1 , u2 , u3 ∈ C 1 ([0, 1]) ∩ C 2 ((0, 1)) that satisfy

0 ≤ max{ max |u1 (t)|, max |u10 (t)|} ≤ r1 ,

t∈[0,1] t∈[0,1]

r1 < max{ max |u2 (t)|, max |u20 (t)|} < L1 ,

t∈[0,1] t∈[0,1]

L1 < max{ max |u3 (t)|, max |u30 (t)|} ≤ R1 .

t∈[0,1] t∈[0,1]

References
[1] S. Djebali, K. Mebarki, Fixed point on Translates of Cones and Applications, Nonlinear Studies,
21 (2014), no. 4, 579–589.

195
CNMA’2022
Second National Conference on
Mathematics and its Applications

[2] S. Djebali, K. Mebarki, Fixed Point Index Theory for Perturbation of Expansive Mappings by
k-set Contractions, Top. Meth. Nonli. Anal. 54 (2019), no. 2 , 613–640.

[3] S. Djebali, K. Mebarki, Fixed point theory for sums of operators, Jour. of Nonl. and Convex
Analy. 19 (2018), no. 6, 1029–1040.

[4] R. Ma, Multiple positive solutions for nonlinear m-point boundary value problem,
[Link]. 148 (2004), no. 1, 249–262.

[5] Y. W. Zhang, A multiplicity result for a singular generalized Sturm-Liouville boundary value
problem, Mathematical and Computer Modelling, 50 (2009), no. 1-2, 132–140.

196
CNMA’2022
Second National Conference on
Mathematics and its Applications

Control and stabilisation of a jerk system

using the fractional Routh-Hurwitz criteria

Chettouh Besma 1 chettouhbesma0707@[Link]

Menacer Tidjani 1
1 Department of mathematics , University of Mohamed khidher Biskra

Abstract: Many dynamic systems are better characterized by a dynamic fractional order
model, generally based on the notion of differentiation or integration of integer-order. In this
work we choose "The Jerk System" to apply the theories of stability and control using the
generalized Routh-Hurwitz criterion to fractional order.

Keywords: Dynamic systems, fractional order, Routh-Hurwitz criterion , The Jerk Sys-
tem, stability, control.

Introduction
The fractional calculus is more than 300 years old with the first written note dated to 1695 [1].
Several physical phenomena can be described more accurately by fractional differential equations
rather than integer-order models.

Main results
The results obtained in this work show the effect of the fractional order on the control, which
proves the effectiveness of the method applied to distinguish the fractional case and that of the
whole case and to underline the importance of the control of the fractional systems, those systems
that have proven to be more accurate than its whole order counterparts.

References
[1] DUBOIS, Francois, GALUCIO, Ana Cristina, et POINT, Nelly. Introduction à la déri-
vation fractionnaire-Théorie et Applications. 2010 ..

[2] [Link], Tarek. Analyse du chaos dans un système d’équations différentielles

[Link]èse de doctorat .[Link]. 2014 .

[3] TIDJANI, Menacer. Synchronisation des systèmes dynamiques chaotiques ‘a dérivées

[Link]èse de doctorat .[Link]. Constantine ,2013 .

197
CNMA’2022
Second National Conference on
Mathematics and its Applications

Classes of polynomial diferential systems

with algebraic limit cycles

Rima Chouader 1 [Link]@[Link]

Salah Benyoucef1 saben21@[Link]
Ahmed Bendjeddou 1 Bendjeddou@[Link]
1 Laboratory of Applied Mathematics, Department of Mathematics, Faculty of sciences, University of Setif 1,
19000, Algeria.

Abstract: For a given algebraic curve of degree n, we exhibit classes of differential systems of
degree n+1. Satisfing certain conditions on the parameters, these systems admit precisely the
bounded components of the curve as limit cycles. Examples exhibiting the applicability of our
result are introduced.

Keywords: Sixteenth problem of Hilbert, planar differential system, Invariant curve ,

Periodic solution , hyperbolic limit cycle.
2022 Mathematics Subject Classification: C00L03UN190120220004.

Introduction
The aim of the second part of sixteenth problem of Hilbert is to find the maximum number
of limit cycles of the differential system:


dx
 ẋ = = P (x, y),


dt

 dy
(194)
 ẏ = = Q(x, y),


dt

where P and Q are polynomials. This problem still persists as a research area.
Several articles and books have been published on the analysis of the existence, number and sta-
bility of limit cycles of system (194) (see for instance [Link] and [Link][3], [Link],
[Link] and [Link]é [4]).
Generally, the exact analytical expressions of limit cycles for a given differential system are
unknown, except in specific cases.
This paper is a contribution in the direction of determining the number of limit cycles and
giving their explicit form. Motivated by some research papers exhibiting planar polynomial sys-
tems with one or more algebraic limit cycles analytically given, ([Link] and [Link][1]),
and mainly based on the papers of [Link] [Theo 3.2][2], we will extend the same concept to
some classes of differential systems, where just by choosing the components of the system satisfy-
ing certain conditions, we can conclude directly the number and the explicit form of limit cycles.
Let us recall some useful notions. For U ∈ R [x, y] , the algebraic curve U = 0 is called an invariant

198
CNMA’2022
Second National Conference on
Mathematics and its Applications

curve of the polynomial system (194), if for some polynomial K ∈ R [x, y] called the cofactor of the
algebraic curve, we have

∂U ∂U
P (x, y) + Q(x, y) = KU . (195)
∂x ∂y

Simple analysis of equation (195) shows that when max(deg P , deg Q) = n, the degree of the cofac-
tor K is at most n − 1 and that the curve U = 0 is formed by trajectories of the system (194).
n o
The curve Γ = (x, y) ∈ R2 , U (x, y) = 0 is non-singular of system (194), if the equilibrium points of
the system that satisfy


 P (x, y) = 0,



 (196)
 Q(x, y) = 0,



are not contained on the curve Γ .

If the curve Γ is non-singular of system (194), the equilibrium points of the system are contained
either in its unbounded components or are located on the curve K(x, y) = 0.
A limit cycle γ = {(x (t) , y (t)) , t ∈ [0, T ]} , is a T −periodic solution isolated with respect to all
other possible periodic solutions of the system.
RT
A T -periodic solution γ is an hyperbolic limit cycle if 0 div(γ)dt is different from zero.

The main result

n o
Let Γ = (x, y) ∈ R2 , U (x, y) = 0 , a curve of degree n.
We consider a polynomial differential system of degree n + 1

 ẋ = R(x, y)U + Φ(x, y)Uy ,



 (197)
 ẏ = S(x, y)U − Φ(x, y)U ,


x

where R(x, y) and S(x, y) are polynomials of first degree and Φ(x, y) a polynomial of second degree.
Our contribution consist to show that the system (195) admits all the bounded components of Γ
as hyperbolic limit cycles if certain conditions on the parameters are satisfied.

Theorem 95. Let U = 0 be a non-singular algebraic curve of degree n,

and the curve axy + b = 0 lies outside all bounded components of U = 0. a, b, α, β are non zero real, and
α + β , 0,
then the differential system of degree n + 1

 ẋ = αxU + (axy + b) Uy ,



 (198)
 ẏ = βyU − (axy + b) U ,


x

admits all the bounded components of U = 0 as hyperbolic limit cycles.

Proof Let Γ the curve of U = 0.

199
CNMA’2022
Second National Conference on
Mathematics and its Applications

Note that Γ is non singular of system (198) and the curve axy + b = 0 lies outside all bounded
components of Γ .
To show that all the bounded components of Γ are hyperbolic limit cycles of system (198) , we will
RT
prove that Γ is an invariant curve of the system (198), and 0 div(Γ )dt , 0.
i) Γ is an invariant curve of system (198):
dU
U̇ = dt = Ux (ẋ) + Uy (ẏ)
 
= Ux αxU + (axy + b) Uy + Uy (βyU − (axy + b) Ux )
 
= αxUx + βyUy U
The cofactor is K(x, y) = αxUx + βyUy
RT
ii) 0 div(Γ )dt , 0 :
Note that Z T Z T
div(Γ )dt = K(x(t), y(t))dt,
0 0

RT H αxUx
H βyUy
0
K(x(t), y(t))dt = − Γ (axy+b)Ux
dy + Γ (axy+b)U dx
y
H βy H
αx
= Γ (axy+b)
dx − Γ (axy+b)
dy.

By applying Green’s formula, we obtain

βy
   !
αx
H βy H
αx
RR ∂ (axy+b) ∂ (axy+b)
Γ (axy+b)
dx − Γ (axy+b)
dy = int(Γ ) ∂y
+ ∂x
dxdy
 
RR β(axy+b)−βaxy α(axy+b)−αaxy
= + dxdy
int(Γ ) (axy+b)2 (axy+b)2
RR b(α+β)
= int(Γ ) (b+axy)2
dxdy.

int(Γ ) denotes the interior of Γ .

RT
As b , 0, and α + β , 0, then 0 K(x(t), y(t))dt , 0. 

Example 5. A quintic differential system with three limit cycles.

Let a = b = α = 1, β = 3.
The differential system
  2   

 ẋ = x x 2 + y 2 − 5 − xy 3 + x 2 y + 1 + (xy + 3) 4x 2 y + x2 − 3xy 2 + 4y 3 − 20y ,


  2   
 ẏ = y x2 + y 2 − 5 − xy + x2 y + 1 − (xy + 3) 4x3 + 4xy 2 + 2xy − 20x − y 3



admits three limit cycles represented by the curve

 2
x2 + y 2 − 5 − xy 3 + x2 y + 1 = 0. See Figure 12.

200
CNMA’2022
Second National Conference on
Mathematics and its Applications

Figure 12: The phase portrait in the Poincare disk of the polynomial differential system

References
[1] A. Bendjeddou and R. Cheurfa, Cubic and planar di erential systems with exact algebraic
limit cycles, Elect. J. of Di. Equ, No 15,1-12 (2011).

[2] S. Benyoucef, Polynomial differential systems with hyperbolic algebraic limit cyles, Elect. [Link]
Qual. Theo. of Di. Equ, No 34, 1-7 (2020).

[3] [Link] and [Link], The number of limite cycles of certain polynomial differential
equations, proceeding of royal society of edinburgh, section A mathematics 98/issue 3-
4/January 1984, pp 215-239.

[4] [Link], H. Giacomini and J. Gine, On a new type of limit cycles for a planar cub
systems, [Link], 36, 139-149 (1999).

201
CNMA’2022
Second National Conference on
Mathematics and its Applications

Solving Nonlinear Fredholm

Integro-Differential Equations by Using
Legender’s Wavelets

Youcef HENKA 1 [Link]@[Link]

Samir LEMITA 2
Mohammed Zine AISSAOUI 1
1 Department of mathematics, University of 08 May 1945 Guelma
2 Department of mathematics, Ecole normale supérieure de Ouergla, Algeria

Abstract: In this work, we provide a numerical study to approach solutions of the nonlinear
Fredholm integro-differential equations with initial conditions. The approximate solution
is obtained by using the Legender’s wavelets basis and its operational matrix. However, the
efficiency of our proposed method is checked through some illustrative examples.

Keywords: Fredholm integro-differential equation, Nonlinear equation, Legender’s wavelets,

Galerkin method.
2010 Mathematics Subject Classification: Primary 45J05, 65T60, 65N30, 34A34.

Introduction
Integro-differential equations have many applications in different fields of sciences such as elec-
trostatics, fluid dynamics, scattering, engineering, biology, and medicine, for this reason there are
many publications that examined the analytical and numerical study for this kind of eqautions.

However, it is diffcult to find the exact solution for such equations, so we apply various numerical
ways to find at least an approximate solution, one of those the numerical methods the projection
method which we focus on.

Main results
we are interesting in numerical study for Fredhom Integro-Differential Equations by applying
projection method which have the following position
 Z1
F(t, s, u(s), u 0 (s))ds

 u(t) = g(t) +




 0 (199)

 u(0) = 0


where F, ∂F
∂t
∈ C([0, 1]2 × R2 ), g(t) ∈ H 1 ([0, 1]) and u(t) ∈ H 1 [0, 1]

202
CNMA’2022
Second National Conference on
Mathematics and its Applications

We mention that this equation has a special and important form because the unknown u(t) and
its derivative appear inside of non linear kernel F of the integral equation.

To get the numerical solution for equation (199), We use the Legender wavelets by applying
Galerkin method that reduces the equation into a nonlinear algebraic system, then the approx-
imate solution is given by solving the nonlinear algebraic system.

References
[1] M. Fathy, M. El-Gamel, M. S. El-Azab, Legendre-Galerkin method for the linear Fredholm
integro-differential equations. Applied Mathematics and Computation, 243, (2014) 789-800.

[2] S. Kumbinarasaiah, R. A. Mundewadi, The new operational matrix of integration for the
numerical solution of integro-differential equations via Hermite wavelet. SeMA Journal, 78(3),
(2021) 367-384.

[3] R. Jalilian, T. Tahernezhad, Exponential spline method for approximation solution of Fred-
holm integro-differential equation. International Journal of Computer Mathematics, 97(4),
(2020) 791-801.

[4] S. Touati, M. Z. AISSAOUI, S. Lemita, H. Guebbai, Investigation approach for a nonlin-

ear singular Fredholm integro-differential equation. Boletim da Sociedade Paranaense de
Matemática, 40, (2022) 1-11.

203
CNMA’2022
Second National Conference on
Mathematics and its Applications

Growth estimate for meromorphic

solutions of ultrametric q-difference
equations of Schröder type

Houda Boughaba 1 houdaboughaba5@[Link]

Tahar Zerzaihi1
1 Laboratoire de Mathématiques Pures et Appliquées(LMPA), University of Mohamed Seddik Ben Yahia, Jijel

Abstract: Let K be an algebraically closed field complete for an ultrametric absolute value.
We denote by M(K) the field of meromorphic functions in K and A(K) the K-algebra of
entire function in K. By the ultrametric Nevanlinna theory, we investigate the growth of
transcendantal meromorphic solutions of some ultrametric q-difference equations. We give
also somme characterizations of the order of growth for transcendantal meromorphic solutions
of these equations.

Keywords: q-Difference equation, Ultrametric meromorphic function, Nevanlinna the-

ory, The growth order.
2010 Mathematics Subject Classification: 12J25, 32A22, 39A13.

Introduction
Let K be an algebraically closed field complete for an ultrametric absolute value. We denote
by M(K) the field of meromorphic functions in K and A(K) the K-algebra of entire function in K.
In this work, we cosider the following ultrametric functional equations of the form
n
X
Aj (x)f (qj x) = ep(x) , (200)
j=0

where A0 (x), ..., An (x) are rational functions in all K, p(x) is a polynomial of the degre d ∈ N and
q ∈ K such that 0 < |q| < 1.
and
R1 (qx, f (qx)) = R2 (x, f (x)), (201)
where R1 (x, y), R2 (x, y) are rational functions with meromorphic coefficients such that R1 (x, y), R2 (x, y)
are irreducible in y and q ∈ K. We put a = degf R1 , b = degf R2 .

Preliminary and notations

Let f (x) =
P n be an entire function. For every r > 0 we define by |f |(r) = sup|an |r n , the
n>0 an x
n>0
maximum modulus of f wich is a multiplicative norm on A(K).

204
CNMA’2022
Second National Conference on
Mathematics and its Applications

Let f be a meromorphic function in K. For every r > 0, we put log+ x = max(0, log x), where log is
the real logarithm function. So, we define the compensation function of f by

m(r, f ) = log+ |f |(r).

Let f ∈ M(K) such that 0 is neither a zero nor a pole of f , we denote by Z(r, f ) the counting
P r
function of zeros of f , counting multiplicity, we put Z(r, f ) = ωα (f ) log |α| . In the same way,
ωα (f )>0
|α|6r
1
 
we put N (r, f ) = Z r, is denoted the counting function of poles of f , counting multiplicity. We
f
finally set
T (r, f ) = N (r, f ) + m(r, f ).
The function r 7−→ T (r, f ) is called the Nevanlinna function or characteristic function of f .
Finally, similarly to complex analysis, we can define the order of growth as follows

log T (r, f )
ρ(f ) = lim sup ,
r→+∞ log r

where f ∈ M(K), and as a particular case when f is an entire function, we have

log log |f |(r)

ρ(f ) = lim sup .
r→+∞ log r

Main results

Theorem 96. Suppose that in (200), the coefficients A0 (x), ..., An (x) are constants. Then for every entire
solution f of (200), we have ρ(f ) = d.

Theorem 97. Let f be a transcendantal meromorphic solution of equation(201). If a ≤ b and |q| , 1,

then
log b − log a
ρ(f ) ≥   .
 log |q|

References
[1] S. Bourourou, A. Boutabaa And T. Zerzaihi, On the growth of solutions of difference equa-
tions in ultrametric fields, Indag. Math., New Ser. 27 (2016), 112–123 .

[2] A. Boutabaa, Théorie de Nevanlinna p-adique, Manuscr. Math. 67 (1990), 251–269.

[3] B.Q. Chen, Z.X. Chen And S. Li, Properties on solutions of some q-difference equations, Acta
Mathematica Sinica, English Series. 26 (2010), 1877–1886.

[4] G.G. Gundersen, J. Heittokangas, I. Laine, J. Rieppo and D. Yang, Meromorphic solutions
of generalized schröder equations, Aequationes Math. 63 (2002), 110–135.

205
CNMA’2022
Second National Conference on
Mathematics and its Applications

Approximated solutions of Schrödinger

equation for the generalized Cornell
potential plus Morse potential

Amal Ladjeroud 12 [Link]@[Link]

Badredine Boudjedaa 31
1 Département de mathématiques et informatique, Centre Universitaire Abdelhafid Boussouf, Mila
2 Laboratoire des Maths Appliquées et Didactique (MAD), E. N. S. Assia Djebar, Constantin
3 Laboratoire des Sciences Naturelles et Matériaux, Centre Universitaire Abdelhafid Boussouf, Mila

Abstract: In this work, we solve the radial Schrödinger equation for the generalized Cornell
potential plus the Morse potential. Using a specific approximation scheme, approximated
solutions are computed explicitly by solving the biconfluent Heun equation.

Keywords: Schrödinger equation, Generalized Cornell potential, Morse potential, Bicon-

fluent Heun equation, Approximated solutions.
2010 Mathematics Subject Classification: 81Q05, 34B30.

Introduction
Since its early days, Schrödinger equation is the cornerstone of the non-relativistic quantum me-
chanics, and obtaining its solutions is with considerable interest in many branches of physics. In
general, for a better description that concerned the physical system, a combination of potentials
are used in which they provide an important amount of information. For complex potentials,
resolving Schrödinger equation is a challenging task. It is well known that its exactly solvabil-
ity is allowed only for a restricted class of potentials, such as the harmonic oscillator, Coulomb,
Morse. For the last few decades a new class of potentials appeared in which it permitted to solve
Schrödinger equation, but only a finite portion of the eigensolutions can be obtained in a closed
form, the so-called the quasi exactly solvable potentials [1]. For central potentials, the presence of
the centrifugal term makes the solvability of the radial equation not trivial, in a such situation an
approximation scheme is applied, in order to transform it to some known and solvable equations
[2, 3].

Main results
The approximated solutions, of the following radial Schrödinger equation , are calculated in the
atomic unit (~ = 1)   "  #
2 l l +1
1 d Rr
− + V (r) + R(r) = ER(r), (202)
2M dr 2 2Mr 2
such that l is a positive integer, M is the mass of the particle and E is the energy of the system and

206
CNMA’2022
Second National Conference on
Mathematics and its Applications

V(r) is the generalised Cornell potential plus the Morse potential which takes the form
2
X
V (r) = ai r i + V0 e−2λr + V1 e−λr
−2

where V0 , V1 , and λ are real parameters.

Using the following approximation scheme [5]

1 1
≈
λr 1 − e−λr
where the right side is a good approximation of the left one as long as the values of the parameter
λ are small.
The radial equation (202) is transformed to the following equation

 
d 2R r
" #
A B 2
+ 2 + + C + Dr − Fr R(r) = 0,
dr 2 r r
where A, B, C, D and F depend on the potential V , which can be transformed into the biconfluent
Heun equation [4].
So by solving the resulting biconfluent Heun’s equation we will determine the approximate solu-
tions of the equation (202).

References
[1] A.G. Ushveridze, Quasi-exactly solvable models in quantum mechanics, CRC Press, 2017.

[2] C.S. Jia, J.Y. Liu, P.Q. Wang, Phys. Lett. A. 372(2008)4779-4782.

[3] W.C. Qiang, S.H. Dong, Phys. Lett. A. 368(2007)13-17.

[4] A. Ronveaux, Heun’s differential equations, Oxford University Press, 1995.

[5] R.L. Greene, C. Aldrich, Phys. Rev. A. 14(1976).

207
CNMA’2022
Second National Conference on
Mathematics and its Applications

Global dynamics of a class of quadratic

differential systems with a non-elementary
singular point

Meryem BELATTAR 1 [Link]@[Link]

Rachid CHEURFA 1 rcheurfa@[Link]
Ahmed BENDJEDDOU 1 Bendjeddou@[Link]
1 Department of mathematics, Laboratory of Applied Mathematics, Ferhat Abbas University, Sétif 1, Faculty of
sciences, 19 000, Sétif, Algeria

Abstract: In this paper, we classify the global phase portraits in the Poincaré disc of a class of
quadratic differential systems with a non-elementary singular point and exhibiting a straight
line as an invariant algebraic curve. As a result, we have obtained exactly 7 topologically
different phase portraits.

Keywords: invariant algebraic curve, non-elementary singular point, phase portrait, Poincaré
disc.
2010 Mathematics Subject Classification: Primary 34A34, 34C05.

Introduction
We consider the planar quadratic differential systems of the form

dx dy
ẋ = = P (x, y), ẏ = = Q(x, y), (203)
dt dt
where P and Q are real polynomials of degree two in the variables x, y.
In the study of differential systems (203), the determination of the explicit solutions is a difficult
task. So, we resort to the geometric approach to seek information about the behavior of these
solutions. This approach is based on characterization of the global phase portraits of systems
(203) in the Poincaré disc. For some related works, we refer the reader to [3, 4, 5, 1] and references
therein.
In this work, we are interested in studying the phase portraits of a class of quadratic differential
systems, given by

2
ẋ = −4a − 4a(2 + b)x + 4y − 8abx + 4bxy,


(204)
ẏ = 4ax + (1 + 16a2 + 4a2 b2 − 16a2 b + 4ab)x2 + 8a(b − 2)xy + 4y 2 ,



where a and b are real parameters.

This system has the invariant straight line U (x, y) = 1 + bx = 0, b , 0.

Main results

208
 CNMA’2022
Second National Conference on
Mathematics and its Applications

The following theorem and propositions summarise the qualitative behavior of system (204).

Theorem 98. Consider the quadratic differential system (204). Then the global phase portrait of system
(204) is topologically equivalent to one of the 7 phase portraits of Figure 13.
More precisely, the global phase portrait in the Poincaré disc of system (204) is topologically equivalent
to the phase portrait
(1) of Figure 13 if b < 1;
(2) of Figure 13 if b > 1;
(3) of Figure 13 if b = 1, a = − 21 ;
(4) of Figure 13 if b = 1, a ∈ R − {− 12 };
(5) or (6) of Figure 13 if 1 + 2ab = 0, b < 1, a ∈ R − {− 12 , 0}.
(7) of Figure 13 if 1 + 2ab = 0, b > 1, a ∈ R − {− 21 , 0}.
Our results have been checked with the program P4 where is a stable node, is an unstable node,
is a saddle, is a semi-hyperbolic singular point, is a center or a weak focus, × is a non-elementary
singular point. For more details on this program, see the Chapters 9 and 10 of [2].

Proposition 30. The differential system (204) has one finite non-elementary  singularity, a nilpotent
a+2a2 b−4a2

a
singular point topologically equivalent to the cusp at q = −2 1+2ab , 1+2ab if ab , − 12 .

Proposition 31. In the local chart U2 , the origin of system (204) is

Description 2.

(a) a saddle if b > 1;

(b) a stable node if b < 1.

(c) a semi-hyperbolic singular point topologically equivalent to the unstable node if b = 1 and a ∈ R −
{− 12 }.

Proposition 32. In the local chart U1 , system (204) has

Description 3.
   
(1) two singular points, an unstable, a stable nodes at q1 = 2a − |1+2ab|
√ , 0 , q2 = 2a + |1+2ab|
√ , 0 , respec-
2 b−1 2 b−1
tively, if b > 1;

(2) one singular point in the following cases

Description 4.

a- a linearly zero (non-elementary) singular point topologically equivalent to the cusp, or the saddle-node
at q3 = (2a, 0) if 1 + 2ab = 0, b < 1 and a ∈ R − {− 12 , 0};

b- a linearly zero (non-elementary) singular point topologically equivalent to the union of one hyperbolic
and one elliptic sectors at q3 = (2a, 0) if 1 + 2ab = 0, b > 1 and a ∈ R − {− 12 , 0};

c- a center or a weak focus at q3 = (−1, 0) if b = 1 and a = − 21 , with the line of singularities at v = 0;

(3) no infinite singular points if b < 1, or b = 1 and a ∈ R − {− 21 }.

209
CNMA’2022
Second National Conference on
Mathematics and its Applications

(1) (2) (3)

(4) (5) (6)

(7)

Figure 13: Phase portraits of quadratic differential system (204) in the Poincaré disc. An orbit of
system (204) is drawn in black. The green line is the line of singularities.

References
[1] R. Benterki and J. Llibre, Phase portraits of quadratic polynomial differential systems having
as solution some classical planar algebraic curves of degree 4, Electronic Journal of Differential
Equations 2019 (2019), 1–25.

[2] F. Dumortier, J. Llibre and J. C. Artés, Qualitative theory of planar differential systems,
SpringerVerlag, Berlin, Germany, 2006.

[3] J. Llibre, R. Oliveira and C.A.B. Rodrigues, Quadratic systems with an invariant algebraic

210
CNMA’2022
Second National Conference on
Mathematics and its Applications

curve of degree 3 and a Darboux invariant, Electronic Journal of Differential Equations 2021
(2021), 1–52.

[4] J. Llibre and C. Valls, Global dynamics of a system coming from the study of a static star,
Differential Equations and Dynamical Systems (2022), 1–11.

[5] L. Tao and J. Llibre, Phase portraits of separable quadratic systems and a bibliographical survey
on quadratic systems, Expositiones Mathematicae 39 (2021), 540–565.

211
CNMA’2022
Second National Conference on
Mathematics and its Applications

Résultat d’existence des solutions pour

problème d’inclusion différentielle
fractionnaire avec des conditions aux
limites

Leyla Soudani 1 Soudanileyla301@[Link]

Abdelkader Amara 1
1 Département de mathématiques, Université de Kasdi Merbah, 3000 Ouargla, Algérie

Abstract: L’objectif de ce travail est de donner de résultat d’existence pour problème inclusion
différentielle fractionnaire de type Caputo d’ordre α avec n − 1 ≤ α < n(n ≥ 2) de valeurs
aux limites dans l’espace de Banach. Le résultat obtenu est basé sur le théorème de Leray -
Schauder pour les applications multivoque.

Keywords: inclusion différentielle , dérivation fractionnaire de type Caputo , théorème

du point fixe, existence des solutions.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
La théorie des équations différentielles est une partie importante des mathématiques pures et
appliquées. Le calcul fractionnaire est une théorie des intégrales et des dérivées d’ordre réel arbi-
traire ou même complexe, il est utilisé dans plusieurs domaines comme viscoélasticité , biologie,
equation de diffusions, physique .
Les inclusions différentielles représentent une généralisation des équations différentielles.
En 2014, A. Bashir et S . K. Ntouyas étudient l’existence de solution pour une équation différen-
tielle fractionnaire séquentielle d’ordre arbitraire[2]

  
C D α D + λ x(t) = f (t, x(t)), α ∈ (n − 1, n), n ≥ 2




x(0) = 0 t ∈ [0, 1]






x(1) = ξx(σ )

x0 (0) = ... = x(n−1) (0) = 0


où f : [0, 1] × R → R est une fonction .

On va étudier le problème généralisé pour les fonctions multivoques .

212
CNMA’2022
Second National Conference on
Mathematics and its Applications

On considère le problème aux limites pour inclusion différentielle fractionnaire suivant :

  
C D α D + λ x(t) ∈ F(t, x(t)), α ∈ (n − 1, n), n ≥ 2




x(0) = 0 t ∈ [0, 1]


 (205)



x(1) = ξx(σ )

x0 (0) = ... = x(n−1) (0) = 0


d
Où C D(.) est la dérivée fractionnaire de Caputo et D = dt .
F : [0, 1] × R → P (R) est une fonction multivoque et λ > 0, β ∈ R, σ ∈ (0, 1).

Préliminaires
On considère C([0, 1], R) l’ espace de Banach des fonctions continues f : [0, 1] → R , muni de
la norme :
kf k = sup{|f (t)|; t ∈ [0, 1]}
Soit L1 ([0, 1], R) l’espace de Banach des fonctions mesurables f : [0, 1] → R , muni de la norme
Z 1
kf kL1 = |f (t)|dt
0

On présente quelques définitions et quelques théorèmes importantes de calcul fractionnaire et

l’analyse multivoque qui utilisés dans ce travail.
Calcul fractionnaire
Définition 1. Soit f ∈ AC n ([0, ∞), R). La dérivée fractionnaire de Caputo d’ ordre α > 0 de f définie
par
Zt
1
C α
D f (t) = (t − s)n−α−1 f (n) (s)ds
Γ (n − α) 0
Où n = [α] + 1.

Définition 2. la fonction Gamma Γ (.) définie par l’intégrale

Z∞
Γ (α) = e−t t α−1 dt
0
.

Notions d’analyse multivoque

Pour un espace normé (X, kk) , on note

- P (X) est une famille de sous ensembles non vide de X .

- Pbnd (X) la famille de sous ensembles bornés non vides de X.

- Pcmp,cvx (X) la famille de sous ensembles compacts et convexes non vides de X.

Définition 3. (Application multivoque)

Soient X, Y deux ensembles non vides. Une application multivoque F définit de X dans Y est une
application qui à chaque x ∈ X associe à sous ensemble non vide Fx de Y . On le note par F : X → P (Y )

213
CNMA’2022
Second National Conference on
Mathematics and its Applications

Définition 4. • Une fonction multivoque F : X → P (X) est convexe (fermée) si F(x) est convexe
(fermée) pour tout x ∈ X.

• La fonction F est bornée sur des ensembles bornés si supx∈B {sup |y| : y ∈ F(x)} < +∞ pour tout
B ∈ Pbnd (X) .

• F est appelée semi continue supérieure sur X si pour chaque x0 ∈ X, F(x0 ) est un sous ensemble
fermé non vide de X et pour chaque ensemble ouvert U de X contenant F(x0 ), il existe un voisinage
ouvert V de x0 tel que F(V ) ⊆ U .

• F est dite complètement continue si F(B) est relativement compact pour chaque B ∈ Pbnd (X).

• Si la fonction multivoque F est complètement continue avec des valeurs compactes non vides alors
F est semi continue supérieure si et seulement si F a un graphe fermé, c-à-d : xn → x∗ et yn → y∗
on a yn ∈ F(xn ) implique que y∗ ∈ F(x∗ ).

• F a un point fixe s’il existe x ∈ X tel que x ∈ F(x). L’ensemble des points fixes de l’ opérateur
multivoque F sera noté F I X (F).

Définition 5. Une fonction multivoque F : [0, 1] × R → P (R) est dite Carathéodory si les conditions
suivantes sont satisfaites :

i. t 7→ F(t, x) est mesurable pour chaque x ∈ R .

ii. x 7→ F(t, x) est semi continue supérieure pour tout t ∈ [0, 1].
De plus, une fonction Carathéodory F est dite L1 − Carathéodory si

iii. pour chaque r > 0, il existe P ∈ L1 ([0, 1], R+ ) telle que pour tout x ∈ R avec kxk ≤ r

kF(t, x)k = sup{|v|, v ∈ F(t, x(t))}

≤ Pr (t)

Maintenant, pour chaque x ∈ C([0, 1], R), on définit l’ensemble de sélections de F par :
SELF,x = {v ∈ L1 ([0, 1], R) : v(t) ∈ F(t, x(t))pour p.p t ∈ [0, 1]} .

Théorème 1. [3] (Alternative Non Linéaire de Leray - Schauder) Soient X un espace de Banach , C un
ensemble convexe dans X. On suppose qu’il existe un ouvert U dans C, 0 ∈ U et F : U → Pcmp,cvx (C)
une multifonction semi continue supérieurement. Alors

a. F admet un point fixe dans U , ou bien

b. il existe un point u ∈ ∂U tel que u ∈ µF(u) pour µ ∈ (0, 1).

Lemme 1. [1] Soit X un espace de Banach et soit F : [0, 1] × R → Pcmp,cvx (X) une fonction multivoque
L1 − Carathéodory et soit Θ : L1 ([0, 1], X) → C([0, 1], X) une application linéaire continue . Alors
l’opérateur
Θ ◦ SELF : C([0, 1], X) → Pcmp,cvx (C([0, 1], X)) (206)
pour tout x ∈ C([0, 1], X) on a (Θ ◦ SELF )(x) = Θ(SELF,x ) est un opérateur de graphe fermé dans
C([0, 1], X) × C([0, 1], X).

Définition 6. [1] On suppose que la condition suivante vérifiant

P(1) − ξP(σ ) , 0, 0 < σ < 1. (207)

214
CNMA’2022
Second National Conference on
Mathematics and its Applications

Une fonction x ∈ AC n−1 ([0, 1], R) est dite solution du problème (205) si x(0) = 0; x0 (0) = ... = x(n−1) (0) =
0; x(1) = ξx(σ ) et il existe une fonction v ∈ L1 ([0, 1], R) sur [0, 1] et telle que
Zt Zs
(s − u)α−1
!
−λ(t−s) P(t)
x(t) = e v(u)du ds +
0 0 Γ (α) P(1) − ξP(σ )
σ s
(s − u)α−1
" Z Z !
−λ(σ −s)
× ξ e v(u)du ds
0 0 Γ (α)
1 s
(s − u)α−1
Z Z ! #
−λ(1−s)
− e v(u)du ds
0 0 Γ (α)

Avec

i. Si n est impaire

t n−1 (n − 1)t n−2 (n − 1)(n − 2)t n−3 (n − 1)!t

P(t) = Po (t) = − 2
+ 3
− ... −
λ λ λ λn−1
(n − 1)!
+ (1 − e−λt )
λn

ii. Si n est paire

t n−1 (n − 1)t n−2 (n − 1)(n − 2)t n−3 (n − 1)!t

P(t) = Pe (t) = − 2
+ 3
− ... +
λ λ λ λn−1
(n − 1)!
− (1 − e−λt )
λn

Main results
On s’intéresse à l’existence des solutions pour le problème (205) est basée sur l’alternative non
linéaire de type Leray-Schauder.

Théorème 2. On suppose que la condition (207) et les hypothèses suivantes sont vérifiées

(H1) F : [0, 1] × R → Pcmp,cvx (R) est une fonction multivoque Carathéodory.

(H2) Il existe une fonction continue croissante ψ : [0, ∞) → (0, ∞)et une fonction P ∈ L1 ([0, 1], R+ )
telles que :

kF(t, x)k = sup{|y|, y ∈ F(t, x(t))}

≤ P (t)ψ(kxk)

pour chaque (t, x) ∈ [0, 1] × R

M
(H3) Il existe un nombre M > 0 tel que : ψ(M)
>1
Γ (α)
[(1 + (1 + ξ)P1 kP kL1 ]
 
P(t)
où P1 = maxt∈[0,1]  .
P(1) + P(t) 

215
CNMA’2022
Second National Conference on
Mathematics and its Applications

Alors le problème (205) admet au moins une solution sur [0, 1].

Pour la démonstration on transforme le problème (205) en un problème de point fixe . On con-

sidère l’opérateur N : C([0, 1], R) → P (C([0, 1], R)) défini par

N (x) = h ∈ C([0, 1], R) :

t s
(s − u)α−1
Z Z !
−λ(t−s) P(t)
h(t) = e v(u)du ds +
0 0 Γ (α) P(1) − ξP(σ )
σ s Z1 Zs
(s − u)α−1 (s − u)α−1
" Z Z ! ! #
−λ(σ −s) −λ(1−s)
× ξ e v(u)du ds − e v(u)du ds
0 0 Γ (α) 0 0 Γ (α)

pour tout v ∈ SELF,x .

Remarque 1. Les solutions du problème (205) sont les points fixes de l’opérateurN c’est a dire Si x est
un point fixe de l’opérateur N , alors x est une solution du problème (205) .

La preuve sera constituée de plusieurs étapes. alors Montrons que

Étape 1 N est à valeurs convexes

En effet, soient x ∈ C([0, 1], R), h1 , h2 ∈ N (x), alors il existent v1 , v2 ∈ SELF,x .Soit 0 ≤ d ≤ 1. Alors
pour tout t ∈ [0, 1],on a

t s
(s − u)α−1
Z Z !
−λ(t−s) P(t)
(dh1 + (1 − d)h2 )(t) = e [(dv1 (u) + (1 − d)v2 )(u)]du ds +
0 0 Γ (α) P(1) − ξP(σ )
σ s
(s − u)α−1
" Z Z !
−λ(σ −s)
× ξ e [(dv1 (u) + (1 − d)v2 )(u)]du ds
0 0 Γ (α)
1 s
(s − u)α−1
Z Z ! #
−λ(1−s)
− e [(dv1 (u) + (1 − d)v2 )(u)]du ds
0 0 Γ (α)

Comme l’ensemble SELF,x est convexe (car F est à valeurs convexes), on aura dh1 +(1−d)h2 ∈ N (x).
Étape 2 N est semi continue supérieurement
On va montrer que

⊕ N est complètement continue, on utilise le théorème d’Ascoli-Arzelà i.e il suffit de prouver

que

1. N est uniformément borné dans C([0, 1], R).

En effet, N applique tout ensemble borné en un ensemble borné dans C([0, 1], R).
soit r > 0 , pour cela on considère l’ensemble borné Br = x ∈ C([0, 1], R) : kxk ≤ r
pour tout x ∈ Br ,il existe l une constante positive tels que kN (x)k ≤ l D’après (H2) , pour

216
CNMA’2022
Second National Conference on
Mathematics and its Applications

tout h ∈ N (x), x ∈ Br , il existe v ∈ SELF,x on a

Zt Zs
(s − u)α−1
!  
−λ(t−s) P(t)
|h(t)| ≤ e |v(u)|du ds +  
0 0 Γ (α) P(1) − ξP(σ ) 

σ s Z1 Zs
(s − u)α−1 (s − u)α−1
" Z Z ! ! #
−λ(σ −s) −λ(1−s)
× ξ e |v(u)|du ds − e |v(u)|du ds
0 0 Γ (α) 0 0 Γ (α)
" Z1 Zσ #
ψ(|x|) −λ(1−s) −λ(σ −s)
≤ (1 + P1 e P (s)ds + ξP1 e P (s)ds = l
Γ (α) 0 0

D’où khk ≤ l.
2. N est équicontinue dans C([0, 1], R) .
applique tout ensemble borné en un ensemble équicontinu dans C([0, 1], R). Soient
t1 , t2 ∈ [0, 1] tel que t1 < t2 et soit x ∈ 2Br . Alors nous avons
 Z t2 Zs Z t1 Zs
(s − u)α−1 (s − u)α−1
! ! 

−λ(t −s) −λ(t −s)
|h(t2 ) − h(t1 )| ≤  e 2 v(u)du ds − e 1 v(u)du ds
0 0 Γ (α) 0 0 Γ (α)
  Zσ Zs
(s − u)α−1
 !
P(t2 ) − P(t1 )   −λ(σ −s)
+   × ξ e v(u)du ds
P(1) − ξP(σ )   0 0 Γ (α)
1 s
(s − u)α−1
Z Z ! 

−λ(1−s)
+ e v(u)du ds
0 0 Γ (α)
t1 s Z t2
(s − u)α−1
Z Z !
−λ(t2 −s) −λ(t1 −s)
≤ [e −e ] |v(u)|du ds + e−λ(t2 −s)
0 0 Γ (α) t1

s
(s − u)α−1
Z !  
P(t2 ) − P(t1 ) 
|v(u)|du ds + 
 
0 Γ (α) P(1) − ξP(σ ) 
σ s Z1 Zs
(s − u)α−1 (s − u)α−1
" Z Z ! ! #
−λ(σ −s) −λ(1−s)
× ξ e |v(u)|du ds + e |v(u)|du ds
0 0 Γ (α) 0 0 Γ (α)

D’après (H2) , et lorsque t1 → t2 , le membre droit de la dernière inégalité tend vers

zéro,d’où N est équicontinu.

on déduit que N est complètement continu.

⊕ Il reste de montrer que l’opérateur N a un graphe fermé.

Soient xn → x∗ ; hn ∈ N (xn ) et hn → h∗ . Il suffit de montrer que h∗ ∈ N (x∗ ).
On a hn ∈ N (xn ), il existe vn ∈ SELF,x telle que pour chaque t ∈ [0, 1] ,
Zt Zs
(s − u)α−1
!
−λ(t−s) P(t)
hn (t) = e vn (u)du ds +
0 0 Γ (α) P(1) − ξP(σ )
σ s Z1 Zs
(s − u)α−1 (s − u)α−1
" Z Z ! ! #
−λ(σ −s) −λ(1−s)
× ξ e vn (u)du ds − e vn (u)du ds
0 0 Γ (α) 0 0 Γ (α)

On considère l’opérateur linéaire Θ : L1 ([0, 1], R) → C([0, 1], R) donné par

217
CNMA’2022
Second National Conference on
Mathematics and its Applications

t s
(s − u)α−1
Z Z !
−λ(t−s) P(t)
Θ(v)(t) = e v(u)du ds +
0 0 Γ (α) P(1) − ξP(σ )
σ s Z1 Zs
(s − u)α−1 (s − u)α−1
" Z Z ! ! #
−λ(σ −s) −λ(1−s)
× ξ e v(u)du ds − e v(u)du ds
0 0 Γ (α) 0 0 Γ (α)

Par conséquent, khn (t) − h∗ (t)k → 0 quand n → ∞. Ainsi, d’après le lemme (1) ,Θ ◦ SELF,xn
est un opérateur de graphe fermé. De plus, on ahn (t) ∈ Θ(SELF,xn ) puisque xn → x∗ . donc N
a un graphe fermé (et a des valeurs fermées). par conséquent, N est compacte.

L’ opérateurs N satisfis toutes les conditions de théorème (1), alors l’une des conditions (a) ou
(b) du théorème (1) est vérifiée. Pour cela il suffit montrer que la condition (b) ne peut pas être
vérifiée. Si x ∈ µN x pour µ ∈ (0, 1) il existe v ∈ (SELF,x ) tel que

Z t Zs
(s − u)α−1
!
−λ(t−s) P(t)
x(t) = µ e v(u)du ds +
0 0 Γ (α) P(1) − ξP(σ )
σ s Z1 Zs
(s − u)α−1 (s − u)α−1
" Z Z ! ! #
−λ(σ −s) −λ(1−s)
× ξ e v(u)du ds − e v(u)du ds
0 0 Γ (α) 0 0 Γ (α)

ψ(kxk) R1 Rσ
D’après (H2), pour t ∈ [0, 1], on a |x(t)| ≤ Γ (α) [(1 + p1 ) 0 e8λ(1−s) p(s)ds + ξp1 0 e8λ(σ −s) p(s)ds]
Si la condition (b) du théorème (1) est vérifiée, il existe alors µ ∈ (0, 1) et x ∈ ∂U avec x ∈ µN (x) .
Alors, x est une solution de (??) avec kxk = M .
Donc, l’inégalité précédent implique

M
ψ(M)
≤1
Γ (α)
[(1 + (1 + ξ)P1 kP kL1 ]

ce qui contredit (H3). Donc, N a un point fixe dans [0, 1] par conséquent, le problème aux limite
(205) admet aux moins une solution dans [0, 1].

References
[1] A. Bashir and S.K. Ntouyas , On higher-order sequential fractional differential inclusions
with nonlocal three-point boundary conditions , Abstract and Applied Analysis . 2014
(2014), 13–22.

[2] A. Bashir and S.K. Ntouyas , A higher-order nonlocal three-point boundary value problem
of sequential fractional differential equations , Miskolc Mathematical Notes. 15 (2014), 265–
278.

[3] M. BENCHOHRA, J. HENDERSON, and S. NTOUYAS, Impulsive differential equations and

inclusions , New York,Hindawi, 2006.

218
CNMA’2022
Second National Conference on
Mathematics and its Applications

Existence and Uniqueness Results for a

Neutral Delay Differential Equations with
Harvesting Term

Lynda Mezghiche 1 linomezg3@[Link]

1 Department of mathematics , University of 20 August 1955, Skikda, Algeria

Abstract: This work is devoted to study the existence and uniqueness of positive periodic so-
lutions for a class of first-order neutral delay differential equations with harvesting term. Our
technique is based on Krasnoselskii’s fixed point theorem and Banach contraction principle.
Our results are new and complement previous studies.

Keywords: existence, neutral differential equation, fixed point theorem, periodic solu-
tion.
2010 Mathematics Subject Classification: 47H10, 65M80, 92C50, 34C25.

Introduction
Consider the following class of first-order neutral delay differential equations with harvesting
term:

d
[x (t) − cx (t − τ (t))] = −a (t) x (t) + f (t, x (t − τ (t))) − E (t, x (t − τ (t))) , (208)
dt
where c ∈ ]0, 1[ , a (t) , τ (t) : R → ]0, ∞[ are T −periodic continuous fonctions and f , E : R2 → ]0, ∞[
are T −periodic continuous functions.
Equation (208) can describe many phenomena such as models for respiration and the dynamics
single species population growth and can model many biological and ecological equations such
as: neutral Mackey-Glass models with harvesting, neutral Nicholson’s blowflies model with har-
vesting and neutral houseflies model with harvesting. where x (t) is the total population size and
a (t) is the mortality rate, f stands for the recruitment function and E is the harvesting term.

Main results
For m > 0 and L, M ≥ 0, let
PT = {x ∈ C(R, R), x(t + T ) = x(t)} ,

equipped with the norm

kxk = sup |x(t)| = sup |x(t)| ,
t∈R t∈[0,T ]

219
CNMA’2022
Second National Conference on
Mathematics and its Applications

and
PT (L, m, M) = {x ∈ PT , m ≤ x ≤ M,
|x(t2 ) − x(t1 )| ≤ L |t2 − t1 | , ∀t1 , t2 ∈ [0, T ]} .
Then (PT , k·k) is a Banach space and PT (L, m, M) is a closed convex and bounded subset of PT .
In this paper, we will assume that:
(H1 ) There exists f0 > 0 such that:
f (t, x) ≥ f0 , ∀t ∈ [0, T ] , ∀x ∈ R. (209)
(H2 ) The functions f (t, x) and E (t, x) are globally Lipschitz in x, i.e. there exist positive constants
k and l such that:  
f (t, x) − f (t, y) ≤ k x − y , (210)
and  
E (t, x) − E (t, y) ≤ l x − y . (211)

To simplify notations, we introduce the following constants:

a0 = inf a (t) , a1 = sup a (t) ,
t∈[0,T ] t∈[0,T ]
f1 = max |f (t, 0)| , E1 = max |E (t, 0)| ,
t∈[0,T ] t∈[0,T ]
 R  R 
T T
exp − 0 a (u) du exp 0 a (u) du
η1 = R  , η2 = R  .
T T
exp 0 a (u) du − 1 exp 0 a (u) du − 1

Lemma 36. If x ∈ PT (L, m, M), then x is a solution of (208) if and only if

Z t+T
x (t) = G (t, s) [f (s, x (s − τ (s))) − E (s, x (s − τ (s)))
t
−ca (s) x (s − τ (s))] ds + cx (t − τ (t)) , (212)
where R s 
exp t a (u) du
G (t, s) = R  . (213)
T
exp 0 a (u) du − 1

Remark 5. The Green’s function G satisfies the following properties:

G (t + T , s + T ) = G (t, s) , ∀t, s ∈ R, (214)
and
η1 ≤ G (t, s) ≤ η2 . (215)

By virtue of Lemma 36, we define an operator N : PT (L, m, M) → PT as follows:

(N x) (t) = (Ax) (t) + (Bx) (t) ,

where A, B : PT (L, m, M) → PT such that

Z t+T
(Ax) (t) = G (t, s) [f (s, x (s − τ (s))) − E (s, x (s − τ (s)))
t
−ca (s) x (s − τ (s))] ds, (216)

220
CNMA’2022
Second National Conference on
Mathematics and its Applications

and
(Bx) (t) = cx (t − τ (t)) . (217)

Existence

Now, we will use Krasnoselskii’s fixed point theorem to prove the existence of at least one pe-
riodic solution of equation (208).

Lemma 37. Suppose that conditions (210) and (211) hold. Then the operator A : PT (L, m, M) → PT is
continuous and compact.

Lemma 38. The operator B : PT (L, m, M) → PT is a contraction.

Lemma 39. If (209) holds and

η2 T (kM + f1 ) + cM ≤ M, (218)
as well as
η1 T f0 − η2 T (lM + E1 ) − cT η2 a1 M + cm ≥ m. (219)
Then
m ≤ (Ax) (t) + (By) (t) ≤ M,
for all y ∈ Ω.

Remark 6. If t1 , t2 ∈ [0, w] with t1 ≤ t2 then

Z t1 +T  Zs ! Zs !
 
exp a (u) du − exp a (u) du  ds
t1 t2 t1 
ZT !
≤ T a1 exp a (u) du |t2 − t1 | . (220)
0

Lemma 40. Let a ∈ PT (L, m, M). If

η2 (2 + a1 T ) (f1 + E1 + M (k + l + ca1 ))
+ L (1 + L) c ≤ L. (221)

then  
((Ax) + (By)) (t2 ) − ((Ax) + (By)) (t1 ) ≤ L |t2 − t1 | ,
for all t1 , t2 ∈ R.

Theorem 99. Suppose that conditions (209)-(211) and (218)-(221) hold, then equation (208) has at
last one solution x ∈ PT (L, m, M).

Existence and uniqueness

Theorem 100. Besides the assumptions of Theorem 99. If

η2 T (k + l + ca1 ) + c < 1, (222)

then equation (208) has a unique solution x ∈ PT (L, m, M).

221
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] A. Bouakkaz, A. Ardjouni and A. Djoudi, Periodic solutions for a second order nonlinear
functional differential equation with iterative terms by Schauder fixed point theorem, Acta
Math. Univ. Comen. 87 (2018), 223–235.

[2] A. Bouakkaz, A. Ardjouni, R. Khemis and A. Djoudi, Periodic solutions of a class of third-
order functional differential equations with iterative source terms, Bol. Soc. Mat. Mex. 26
(2020), 443–458.

[3] A. Bouakkaz and R. Khemis, Positive periodic solutions for revisited Nicholson’s blowflies
equation with iterative harvesting term, J. Math. Anal. Appl. 494 (2021), 124663.

[4] T. Candan, Existence of positive periodic solutions of first order neutral differential equations
with variable coefficients, Appl. Math. Lett. 52 (2016), 142–148.

[5] R. Khemis AND A. Ardjouni, A. Bouakkaz and A. Djoudi, Periodic solutions of a class
of third-order differential equations with two delays depending on time and state, Comment.
Math. Univ. Carolin. 60 (2019), 379–399.

[6] L. Mezghiche, R. Khemis and A. Bouakkaz, Positive periodic solutions for a neutral differen-
tial equation with iterative terms arising in biology and population dynamics, Int. J. Nonlinear
Anal. Appl. 13 (2022), 1041–1051.

222
CNMA’2022
Second National Conference on
Mathematics and its Applications

Inclusions with mappings acting from a

metric space to a space with generalized
distance

Merchela Wassim 1 [Link]@[Link]

1 Department of Functional Analysis, Derzhavin Tambov State University, Russia

Abstract: This paper deals with an inclusion in which a multivalued mapping acts from
a metric space (X, ρ) into a set Y with generalized distance d. The generalized distance d
satisfies only the first axiom of the distance: d(y1 , y2 ) is equal to zero if and only if y1 = y2 .
The generalized distance does not have to be symmetric or to satisfy the triangle inequality.
For a multivalued mapping G : X ⇒ Y , the sets of covering, Lipschitz and closedness are
introduced. Theorem on solvability of the inclusion F(x, x) 3 b y is formulated, and an estimate
for the deviation in the space (X, ρ) of the set of solutions from a given element x0 ∈ X is given.
The main conditions of the obtained statement are the following: for any x from some ball, the
pair (x,by ) belongs to the α-covering set of the mapping F(·, x) and to the β-Lipschitz set of the
mapping F(x, ·), where α > β.

Keywords: distance, generalized distance, inclusion, existence of solution, covering multival-

ued mapping
2010 Mathematics Subject Classification: 47J22, 47H04, 47H04.

Introduction
The results on operator inclusions (operator equations) with multivalued mappings (mappings)
acting in metric spaces are widely used for studying various functional inclusions (functional
equations). In particular, the results on covering mappings of metric spaces allowed one to con-
sider some classes of integral inclusions (equations). In recent studies [1], [2], the results on
covering mappings were generalized for the spaces, in which classical properties of metrics were
weakened. In works [3], [4], the notion of the covering set was extended to mappings acting from
a metric space into a set equipped with a distance obeying just identity axiom.

Main results
Let X — metric space with distance ρ : X × X → R+ , where R+ = [0, +∞]. We denote by BX (x0 , r) =
{x ∈ X : ρ(x0 , x) ≤ r} — closed ball in X centered at a point x0 ∈ X of a radius r ∈ R+ .
We also suppose that a non-empty set Y is given, on which a generalized distance is defined,
which is a mapping d : Y × Y → R+ obeying the condition
∀y1 , y2 ∈ Y d(y1 , y2 ) = 0 ⇔ y1 = y2 , (223)
it is important that the mapping d may not possess the other properties of metrics. In the space
Y we define the notion of convergence of a sequence yi ⊂ Y to an element y ∈ Y as i → ∞ by the

223
CNMA’2022
Second National Conference on
Mathematics and its Applications

relation
yi → y ⇔ d(y, yi ) → 0.
We observe that under such convergence, the limit y is not necessary unique and a symmetric
scalar sequence d(yi , y) not necessary converge to zero. We define the generalized distance in Y
from the element y ∈ Y to the set V ⊂ Y by the formula

dist(y, V ) = inf d(y, v).

v∈V

We are going to define weakened properties of closedness, covering and Lipschitz property for a
multivalued mapping G : X ⇒ Y . Given a set U ⊂ X and numbers α > 0, β ≥ 0. We define sets:
n o
Cl[G; U ] = (x, y) ∈ X ×Y : ∀{xn } ⊂ U xn → x, ∀yn ∈ G(xn ) yn → y ⇒ y ∈ G(x) ,
n d(y, z) o
Covα [G; U ] = (x, y) ∈ X ×Y : ∀z ∈ G(x) ∃u ∈ U y ∈ G(u), ρ(x, u) ≤ , ρ(x, u) < ∞ ,
α
n o
Lipβ [G; U ] = (x, y) ∈ X ×Y : ∀u ∈ U y ∈ G(u) ⇒ ∃z ∈ G(x) d(y, z) ≤ βρ(x, u) ,

Given a multivalued mapping F : X × X ⇒ Y and an element b

y ∈ Y . We consider the inclusion

G(x) := F(x, x) 3 b
y (224)

with an unknown x ∈ X. We formulate a statement on solvability of inclusion (224).

Theorem 101. Let a metric  space X be complete, and suppose that we are given α > β ≥ 0, ε > 0, x0 ∈ X
such that dist b
y , F(x0 , x0 ) < ∞. We define:

1    
R := dist b
y , F(x0 , x0 ) < ∞, U := BX x0 , (1 + ε)R .
α−β

Assume that for each x ∈ U the embeddings hold:

(x,b
y ) ∈ Covα [F(·, x); X], (x,b
y ) ∈ Lipβ [F(x, ·); U ], (x,b
y ) ∈ Cl[G; U ].

Then in the ball U there exists a solution to inclusion (224).

References
[1] A. V. Arutyunov, A. V. Greshnov, Theory of (q1 , q2 )-quasimetric spaces and coincidence
points, Doklady Mathematics 94:1 (2016), 434–437.

[2] W. Merchela, On Arutyunov theorem of coincidence point for two mapping in metric spaces,
Tambov University Reports. Series: Natural and Technical Sciences 23:121 (2018), 65–73.

[3] E. S. Zhukovskii, W. Merchela, On covering mappings in generalized metric spaces in study-

ing implicit differential equations, Ufa Mathematical Journal 12:4 (2020), 42–55.

[4] S. Benarab, E. S. Zhukovskii, W. Merchela, Theorems on perturbations of covering map-

pings in spaces with a distance and in spaces with a binary relation, Trudy instituta matem-
atiki i mekhaniki uro ran 25:4 (2019), 52–63.

224
CNMA’2022
Second National Conference on
Mathematics and its Applications

Asymptotic behavior of a viscoelastic

problem with long-term memory and
Tresca friction law

Aissa Benseghir 1 [Link]@[Link]

1 Department of mathematics Applied mathematics laboratory Lama , University of Setif1

Abstract: This paper examines the asymptotic behavior of solutions of the three dimensional
viscoelastic problem with long-term memory and Tresca friction law in a thin domain Σ .
We study the asymptotic behavior of this problem when the thickness  tends to zero and
we prove a convergence theorem for the displacement and velocity in appropriate functional
spaces. Besides, the limit problem with the limit of Tresca free boundary conditions and a
specific Reynolds limit equation is obtained.

Keywords: viscoelastic problem, Tresca friction law, long-term memory, weak solution,
thin domain.
2010 Mathematics Subject Classification: Primary 35C20, 35D10.

References
[1] G. Bayada and M. Boukrouche, On a free boundary problem for Reynolds equation de-
rived from the Stokes system with Tresca boundary conditions, J. Math. Anal. Appl., 382
(2003), pp.212–231.

[2] G. Bayada and K. Lhalouani, Asymptotic and numerical analysis for unilateral contact
problem with Coulomb’s friction between an elastic body and a thin elastic soft layer,
Asymptot. Anal. 25 (2001), 329-362.

[3] A. Benseghir, H. Benseridi and, M. Dilmi, On the asymptotic study of transmission prob-
lem in a thin domain, J. Inverse Ill-Posed Probl. [Link]

[4] H. Benseridi and M. Dilmi, Some inequalities and asymptotic behavior of dynamic prob-
lem of linear elasticity, Georgian Math. J., 20(1) (2013), pp. 25–41, ISSN (Online) 1572-
9176, ISSN (Print) 1072-947X, DOI10.1515/gmj-2013-0004, March 2013.

[5] D. Benterki, H. Benseridi, and M. Dilmi, Asymptotic study of a boundary value problem
governed by the elasticity operator with nonlinear term, Adv. Appl. Math. Mech., 6 (2014),
pp.191-202.

[6] M.M. Cavalcanti, V.N. Domingos Cavalcanti, T.F. Ma, and J. [Link], Global exisience
and asymptotic stability for viscoelastic problems, Differential and Integral Equations
volume 15, Number 6, June 2002, Pages 731-748.

225
CNMA’2022
Second National Conference on
Mathematics and its Applications

Conditions de Robin généralisées pour

une EDA à coefficients opérateurs
variables dans le cadre Lp

Rabah Haoua 1 [Link]@[Link]

Ahmed Medeghri1 [Link]@[Link]
1 Laboratory of Pure and applied Mathematics of Mostaganem, 27000 Mostaganem, Algeria

Abstract: In this paper we study an abstract second order differential equation of elliptic type
with variable operator coefficients and general Robin boundary conditions, in the framework
of UMD spaces. These problems presents for example the linearized stationary case of a model
describing information diffusion in online social networks. Existence and regularity results
are obtained when the Labbas-Terreni assumption is fulfilled using semi-groups theory and
interpolation spaces.

Keywords: Differential equation, Robin boundary conditions, analytic semigroup, maxi-

mal regularity, Dore-Venni theorem and UMD spaces.

Introduction and hypotheses

This paper is devoted to study the following general problem




 u 00 (x) + A (x) u (x) − ωu (x) = f (x), x ∈ (0, 1)


 0

 u (0) − Hu (0) = d0 (225)


 u (1) = u ,

1

with f ∈ Lp (0, 1, E), 1 < p < +∞, where E is a complex Banach space, d0 , u1 are given elements in
E and (A (x))x∈[0,1] is a family of closed linear operators whose domains D (A (x)) are dense in E. H
is a closed linear operator in E, ω is a positive real number. The results proved here in the Lp case
complete our recent paper concerning the hölderian case, see [2].
For all x ∈ [0, 1], set:

Aω (x) = A (x) − ωI.

226
CNMA’2022
Second National Conference on
Mathematics and its Applications

We will seek for a classical solution u to (225), i.e. a function u such that

a.e x ∈ (0, 1), u (x) ∈ D (A (x)) and






 x 7→ A (x) u (x) ∈ Lp (0, 1; E)




u ∈ W 2,p (0, 1; E)







 u (0) ∈ D (H) ,


The method is essentially based on Dunford calculus, interpolation spaces, the semigroup theory
and some techniques as in [3], [2].
We will assume that
E is a U MD space. (226)
We suppose that:

 −1
∃ω0 > 0, ∃C > 0 : ∀x ∈ [0, 1], ∀z ≥ 0, Aω0 (x) − zI ∈ L (E) and

Aω (x) − zI −1
  C
0
≤ ; (227)
L(E) 1+z

and setting Qω (x) = − (−Aω (x))1/2 ( see [1]), we suppose also that:
∃C, α, µ > 0 : ∀x, τ ∈ [0, 1] , ∀ω ≥ ω0 :
C |x − τ|α
  
 −1 −1 −1
Q (x) (Q (x) − zI) Q (x) − Q (τ) ≤

ω ω ω ω

|z + ω|µ

L(E) (228)



 with α + µ − 2 > 0;



this hypothesis is well known as Labbas-Terreni assumption.

We obtain the following theorem.
Theorem 102. Assume (226)∼(228). Let f ∈ Lp (0, 1; E), 1 < p < +∞ and
(Qω (0) − H)−1 d0 ∈ (D(A(0)), E) 1 ,p , u1 ∈ (D(A(1)), E) 1 ,p .
2p 2p

Then there exists ω∗ > 0 such that for all ω ≥ ω∗ , the problem (225) has a unique solution w (·) =
Qω (·)2 u (·) verifying

1. Qω (·)2 u (·) ∈ Lp (0, 1; E).

2. u 00 ∈ W 2,p (0, 1; E).

References
[1] A. V. Balakrishnan, Fractional Powers of Closed Operators and the Semigroups Generated by
them, Pacific J. Math., 10 (1960), 419-437.

[2] R. Haoua and A. Medeghri:. Robin boundary value problems for elliptic operational differ-
ential equations with variable operators, Electronic Journal of Differential Equations. Vol.
2015.

[3] R. Labbas:. Problèmes aux limites pour une equation différentielle abstraite de type elliptique,
Thèse d’état, Université de Nice (1987).

227
CNMA’2022
Second National Conference on
Mathematics and its Applications

New Approach for the Asymptotic

Stability of Two Dimensional Models

Kamel BENEYETTOU 1 [Link]@[Link]

Djillali BOUAGADA 1 [Link]@[Link]
Mohamed Amine GHEZZAR 1 [Link]@[Link]
1 Department ofMathematics and Computer Science, ACSY Team-Laboratory of Pure and AppliedMathematics,
Abdelhamid Ibn Badis University of Mostaganem, [Link] 227/118 University ofMostaganem, 27000.

Abstract: The aim of this work is to present a new sufficient conditions of asymptotic stability
for the general continuous time-system in term of linear matrix inequalities LMI s. The LMI
approach is used to produce very significant new results on the asymptotic stability analysis.
Many optimization problems in control theory, system identification and signal processing
can be formulated using the concept of LMI s. A numerical example is proposed to show the
applicability and accuracy of the proposed method.

Keywords: Two dimensional systems, Stability tests, Singular systems, Linear algebra,
Partial differential equation.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
In this work we will study the stability problem of the general 2D-continuous model, where the
linear matrix inequalities (LMI’s) approach is applied to extract new sufficient conditions for the
asymptotic stability. Many of their applications are found in control theory, modern circuit de-
sign and digital image processing, seismology, geographic data processing, energy transmission
[1, 2, 3, 4, 5]. Stability test is the most important and fundamental problem for system design and
analysis.

Main results
The purpose of this paper is to present a sufficient condition of stability for the considered system
       
 E11 E12   ∂xh (t1 ,t2 )   A11 A12   xh (t1 , t2 )   Bh 
∂t1  (t
 =   +   u 1 , t2 ) (229)
        
∂xv (t1 ,t2 )
    
 E   x (t , t )
21 E22
  A
21 A22
   B 
∂t2 v 1 2 v

where xh (t1 , t2 ) ∈ Rn1 , xv (t1 , t2 ) ∈ Rn2 are respectively the horizontal and the vertical state vectors,
E ∈ Rn×n is the singular matrix of the system with 0 ≤ rank(E) ≤ n, u (t1 , t2 ) ∈ Rm is the input vector,
Bh ∈ Rn1 ×m , Bv ∈ Rn2 ×m , A11 ∈ Rn1 ×n1 , A12 ∈ Rn1 ×n2 , A21 ∈ Rn2 ×n1 A22 ∈ Rn2 ×n2 and n = n1 + n2 . The
boundary conditions are xv (t1 , 0) and xh (0, t2 ).

228
CNMA’2022
Second National Conference on
Mathematics and its Applications

To simplify the computations and notations in the following result we denote

   
 E11 E12   A11 A12 
Ẽ =   and à = 
   

 E
21 E22 A21 A22 
 

Definition 42. The characteristic polynomial of the system (229) is defined by

G(s1 , s2 ) = det(ẼK − Ã)

where K is defined by the expression  

 s1 In
1
0 
K = 
 

 0 s2 In2 

Theorem 103. The two dimensional model described by (229) is said to be asymptotically stable if there
exist a positive definite and hermitian matrix X0 , X1 , X2 satisfying the following LMI s:

(Ã − Ẽ2T X1 Ẽ1 − Ẽ1T X1 (Ã − Ẽ2 )  0

 
 ÃT X2 Ẽ2 − Ẽ T X2 Ã X0 
2
  0
 

 Ẽ T X Ẽ − Ẽ T X Ẽ 0 
1 2 2 2 2 1

with
Ẽ1 = Ẽdiag(In1 , 0)
and
Ẽ2 = Ẽdiag(0, In2 )

Example 6. Let us consider the system (229) with u(t1 , t2 ) = 0 and the system matrices

 
 1 0 
E = 
 

 0 0 

 
 0.7 0.6 
A = 
 

 0.4 0.3 

By the use of our method we find that the LMI s in Theorem 103 are feasible, and a feasible solution is
as follows

 
 1.1571 0.5190 
X0 = 
 

 ”0.5190 1.1571 

 
 0.0000 0.0000 
X1 = 
 

 0.0000 1.3877 

 
 1.0000 0.0000 
X2 = 
 

 0.0000 1.0000 

229
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] D. Bouagada and P. Van Dooren, LMI Conditions for the Stability of 2D State-Space Models,
Numerical Linear Algebra with Applications, 20(2), (2013), 198-207.

[2] S.F Chen, Analysis and Stabilization of 2-D Singular Roesser Models, Applied Mathematics
and Computation. 250 (2015), 779–791.

[3] M.E. Valcher, On the internal stability and asymptotic behavior of 2D positive systems, IEEE
Transactions On Circuits and Systems, 44(7) (1997), 602–613.

[4] T. Kaczorek, Asymptotic Stability of Positive 2D Linear Systems, Proc. 13th Scientific Conf.
on Computer Applications in Electrical Engineering„ Poznan, Poland, 2008.

[5] M. A Ghezzar, D. Bouagada , [Link], [Link], P. Van Dooren, [Link], P. Van

Dooren, On The Stability of 2D General Roesser Lyapunov Systems, Mathematica Cluj, 63(86)
(2021), 85–97.

230
CNMA’2022
Second National Conference on
Mathematics and its Applications

Boundedness and exponential stability of

solutions to of abstract dynamic equations
on time scales and applications

Amira Ayari 1 ayari.amira1995@[Link]

1 Department of mathematics , Badji Mokhtar-Annaba University, Annaba, Algeria

Abstract: In this work, we use the Lyapunov’s method to obtain new sufficient conditions for
many types of stability like exponential stability, uniform exponential stability, of some classes
of nonlinear dynamic equations. Some examples are also presented at the end of this work.

Keywords: Lyapunov stability theory, dynamic equations, time scales

2010 AMS Mathematics Subject Classification: 34N05.

Introduction
One of the most important and useful tools for investigating the behavior of solutions of dynamic
equations on a general time scale is Lyapunov’s(Lyapunov’s direct method), which was introduced
by Lyapunov in 1892. Many studies used the Lyapunov technique to investigate various types of
stability for the systems of dynamic equations on time scales; for instance, see [3, 4, 5]. Ben Nasser
et al. [4] established some sufficient conditions for the existence of the quadratic Lyapunov func-
tion that ensure the desired asymptotic convergence of trajectories. The difficulty of the Lyapunov
technique is to construct a Lyapunov function. For equations with solutions with values in the Eu-
clidean space Rn , the situation is simpler. The Lyapunov function is usually chosen to be

V (t, x) = xT P (t)x
where P (t) is an n × n matrix and xT is the transpose of x ∈ Rn . In the Hilbert space setting, the
Lyapunov function is chosen to be

V (t, x) =< P (t)x, x >,

where P (t) is a bounded linear operator on a Hilbert space and < ., . > is its inner product.
The paper starts with the investigation of sufficient conditions for the boundedness of solutions
and the exponential stability, uniform exponential stability, h-stability, and uniform h-stability of
the abstract dynamic equation

x∆ (t) = F(t, x), x(τ) = xτ ∈ X, t ∈ Tτ+ := [τ, ∞)T , (230)

where F : T × X → X is rd-continuous in the first argument with F(t, 0) = 0. Here, T is a time

scale and X is a Banach space. Thereafter, we construct a Lyapunov function and make use of this
function to study the stability of the abstract homogeneous equation

231
CNMA’2022
Second National Conference on
Mathematics and its Applications

x∆ (t) = A(t)x(t), t ∈ Tτ+ ,

and its perturbed equation of the form

x∆ (t) = A(t)x(t) + f (t, x), t ∈ Tτ+ ,

where A(·) ∈ Crd (T, L(X)) and f : T × X → X is rd-continuous in the first argument with f (t, 0) = 0.
Also, we establish sufficient conditions for the nonhomogeneous particular dynamic equation

x∆ (t) = A(t)x(t) + f (t), t ∈ Tτ+ ,

where f ∈ Crd (T , X), to be uniformly exponentially stable.

Main results
In this section our aim is to ensure that the solutions fo Eq (230) are bounded.

Theorem 104. Let p and s be positive constants. Assume there exists a positive definite function V ∈
1
Crd (T × X, R+ ) that satisfies the following conditions:
(i) λ(t) k x kp ≤ V (t, x), for some positive nondecreasing function λ;
(ii) V ∆ (t, x) ≤ −b(t)V s (t, x) + l(t), for some positive function b with −b ∈ R+ Crd and l ∈ Crd ;
(iii) V (t, x) − V s (t, x) ≤ γ, for some γ ≥ 0;
Rt
(iv) l(u)e (−ω) (σ (u), t)∆u ≤ L, for some nonnegative constant L, where ω := inf b(t) > 0.
τ t∈T

Then all solutions of Eq. (230) are bounded.

References
[1] Bohner M, Peterson A., Dynamic Equations on Time Scales: An Introduction with Applica-
tions, MBasel, Switzerland: Birkhauser, 2001.

[2] Bohner M, Peterson A., Advances in Dynamic Equations on Time Scales, Basel, Switzer-
land: Birkhauser, 2003.

[3] Nasser BB, Boukerriona K, Hammami MA, On stability and stabilization of perturbed time
scale systems with Gronwall inequalities, J Math Phys Anal Geo 11 (2015), 207-253.

[4] Peterson AC, Tisdell CC, Boundedness and uniqueness of solutions to dynamic equations on
time scales, J Differ Equ Appl 10 (2004), 1295-1306.

[5] Raffoul YN., Boundedness and exponential asymptotic stability in dynamical systems with
applications to nonlinear differential equations with unbounded terms, Advances in Dynami-
cal Systems and Applications 2 (2007), 107-121.

232
CNMA’2022
Second National Conference on
Mathematics and its Applications

Stability of Incommensurate h-Nabla

Fractional-Order Difference Systems

Noureddine Djenina 1 noureddinedjenina1996@[Link]

Adel Ouannas 12 [Link]@[Link]
Taki-Eddine Oussaeif 1 taki_maths@[Link]
1 Department of Mathematics and Computer Science, University of Larbi Ben M’hidi, Oum El Bouaghi 04000,
Algeria
2 Nonlinear Dynamics Research Center (NDRC), Ajman University, Ajman 20550, United Arab Emirates

Abstract: In this paper we announce some results for stability of incommensurate fractional
order backward difference systems. These results are verified numerically via illustrative
numerical example that show the stabilities of the solutions of systems at hand.

Keywords: the h -nabla fractional-order sum operator, incommensurate fractional-order

difference systems, stability analysis.
2010 Mathematics Subject Classification: 37D05, 39A05, 39A13, 39A30, 39A70.

Introduction
Undoubtedly, it has been demonstrated, over the past few decades, that the non-integer calcu-
lus is a forceful mathematical argument for providing much and more dynamics lots of ancient
as well as modern models. This paper presents some simple applicable conditions for judging
the stability of such system by first converting it into another equivalent form includes FoDEs of
Volterra convolution-type as well as by using the properties of the Z -transform method. However,
this paper is organized in the following order. Section 2 introduces some primary preliminaries
associated with discrete fractional calculus, while Section 3 discusses some recently established
results in [1] that have handled the stability of incommensurate FoDS. Section 4 exhibits an ex-
ample to verify all findings, followed by the last section that summarizes the achievements of the
whole work.

Preliminaries
This section briefly introduces some basic definitions and preliminaries associated with discrete
fractional calculus. In all of the definitions below, the function f is defined on Na,h = {a, a + h, a + 2h, · · · },
for a ∈ R.
Definition 43. [2] For a function f : Na,h → R, the nabla left h−fractional sum of order α > 0 is defined
by
t/h
1 X
−α
a ∇h f (t) := (t − sh + h)α−1
h f (sh)h, for t ∈ Na,h ,
Γ (α)
s=a/h+1
Γ ( ht +α)
where Γ (.) is the Euler’s gamma function and thα = hα Γ ( ht )
.

233
CNMA’2022
Second National Conference on
Mathematics and its Applications

Definition 44. [2] The nabla R-L left h−fractional difference of order 0 < α ≤ 1 (starting from a) is
defined by

t/h
1
 
−(1−α)
X
α
∇
a h f (t) := ∇ ∇
ha h f (t) = ∇h (t − sh + h)−α
h f (sh)h, for t ∈ Na+h,h ,
Γ (1 − α)
s=a/h+1

f (t)−f (t−h)
where ∇h f (t) = h .

Definition 45. [2] Assume that 0 < α ≤ 1, 0 < h ≤ 1, a ∈ R, and f is defined on Na,h . Then the left
h−Caputo fractional difference of order α starting at a is defined by

t/h
1
 
−(1−α)
X
C α
a ∇h f (t) := a ∇h ∇h f (t) = (t − sh + h)−α
h (∇h f ) (sh)h,
Γ (1 − α)
s=a/h+1

for t ∈ Na+h,h .

Lemma 41. [4] Let 0 < α < 1. Then,

t/h ! !
1X α 1 t/h−1 α − 1
(C α
0 ∇h f )(t) = α (−1)t/h−s t f (sh) + α (−1) t f (0), t ∈ Nh,h .
h h −s h h
s=0

α
α(α−1)(α−2)···(α−β+1)
where β = Γ (β+1)
, is the general binomial coefficient.

Lemma 42. [5] Suppose that αi ’s are rational numbers between 0 and 1, for i = 1, 2, ..., n. Let M be the
lowest common multiple (LCM) of the denominators ui of αi ’s, where αi = uvi , (ui , vi ) = 1, ui , vi ∈ Z+ , i =
i
1
1, 2, ..., n, and set γ = M . Then the following statements are equivalent
(i) All roots (z) of the following characteristic equation:

1 1 1
   
det diag (1 − )α1 , (1 − )α2 , ..., (1 − )αn − HA = 0,
z z z
lie in the interior (exterior) the unit disk.
(ii) All roots (λ) of the following characteristic equation
   
det diag λMα1 , λMα2 , ..., λMαn − HJ = 0,

lie in the exterior (interior) the set

( !γ )
γ arg z γπ
K = z ∈ C : |z| ≤ 2 cos and |arg z| ≤ .
γ 2

Main results
Consider the following incommensurate FoDS:

C ᾰ
a ∇h x(t) = f (x(t)), t ∈ Na+h,h ,
ᾰ α α α
where x(t) = (x1 (t), x2 (t), ..., xn (t))T ∈ Rn , C C 1 C 2 C n T
a ∇h x(t) = (a ∇h x1 (k),a ∇h x2 (k), ...,a ∇h xn (k)) , 0 < αi <
T n n
1, for i = 1, 2, · · · , n. f = (f1 , f2 , · · · , fn ) : R → R constantly differentiable twice function. To
simplify, we choose a = 0.

234
CNMA’2022
Second National Conference on
Mathematics and its Applications

Using Lemma 42 we get for 1 ≤ i ≤ n:

t/h−1 ! !
1 1 X t/h−s+1 αi 1 t/h αi − 1
x (t) = α (−1) xi (sh) + α (−1) xi (0) + fi (x(t)), t ∈ Nh,h ,
hαi i h i t
h −s
h i t
h
s=0

we simplify, we find that this system is written in compact form as follows

k
X
y(k + 1) = B(k − s)y(s) + C(k)y(0) + Hf (y(k + 1)), k = 0, 1, 2, · · · ,
s=0

where ! ! !!
α1 k k α2 k αn
B(k) = diag (−1) , (−1) , · · · , (−1) ,
k+1 k+1 k+1
! ! !!
k+1 α1 − 1 k+1 α2 − 1 k+1 αn − 1
C(k) = diag (−1) , (−1) , · · · , (−1) ,
k+1 k+1 k+1
and
H = diag(hα1 , hα2 , · · · , hαn ).

We note that system (8) is equivalent to system (7). From now on, we will study the system (8)
Linear system

We assume that the function f is linear, this means that there is a matrix A ∈ Rn×n , so that the
system (7) is written as follows
C ᾰ
a ∇h x(t) = Ax(t), t ∈ Na+h,h .

Theorem 105. Let det(I − HA) , 0. Then (12) has a unique solution for any initial vector x0 ∈ Rn .
Moreover
• If all roots of the following characteristic equation:
1 1 1
   
det diag (1 − )α1 , (1 − )α2 , ..., (1 − )αn − HA = 0,
z z z
lie inside the unit disk, then the zero solution of system (12) is asymptotically stable.
• If there exists a zero, say z∗ of (14) such that |z∗ | > 1, then the zero solution of system (12) is not stable.

Proof The proof is based on the Z-transformations properties and the Volterra difference equa-
tions of convolution type. For more details, see [1]. 
Corollary 12. Suppose that αi ’s are rational numbers between 0 and 1, for i = 1, 2, ..., n. Let M be the
lowest common multiple (LCM) of the denominators ui of αi ’s, where αi = uvi , (ui , vi ) = 1, ui , vi ∈ Z+ , i =
i
1
1, 2, ..., n, and set γ = M . Then the zero solution of system (12) with initial value x0 = x(0) is
• asymptotically stable if and only if any zero solution of the polynomial
   
det diag λMα1 , λMα2 , ..., λMαn − HA = 0,

lie inside the set ( !γ )

γ arg z γπ
K = z ∈ C : |z| > 2 cos or |arg z| > .
γ 2

• Furthermore, If there is a zero λ of (26) with λ ∈ IntK γ , the zero solution of (12) is not stable.

235
CNMA’2022
Second National Conference on
Mathematics and its Applications

Proof The proof came immediately from the equivalence between the coditions of Theorem 105
and conditions of Corollary 12 from Lemma 41. 
Non-linear system
We will now study the non-linear system (7).

Theorem 106. Let 0 be an equilibrium point of (7). If all roots of the characteristic equation

1 1 1
   
det diag (1 − )α1 , (1 − )α2 , ..., (1 − )αn − HJ = 0,
z z z
where J is the jacobian matrix of f at 0, lie inside the unit disk, then (7) has a unique solution for all
initial vectors close enough to 0 and, moreover, 0 is asymptotically stable.

Proof The proof is based on the Z-transformations properties and the Volterra difference equa-
tions of convolution type. For more details, see [1]. 

Corollary 13. Suppose that αi ’s are rational numbers between 0 and 1, for i = 1, 2, ..., n. Let M be the
lowest common multiple (LCM) of the denominators ui of αi ’s, where αi = uvi , (ui , vi ) = 1, ui , vi ∈ Z+ , i =
i
1
1, 2, ..., n, and set γ = M . Then the zero solution of system (7) with initial value x0 = x(0) is locally
asymptotically stable if any zero solution of the polynomial equation
   
det diag λMα1 , λMα2 , ..., λMαn − HJ = 0,

lie inside the set ( !γ )

γ arg z γπ
K = z ∈ C : |z| > 2 cos or |arg z| > .
γ 2
and J is the jacobian matrix of f at 0.

Proof The proof came immediately from the equivalence between the coditions of Theorem 8
and conditions of Corollary 9 from Lemma 5. 
Nnumerical example
Example 7. Consider the nonlinear incommensurate FoDS:
 1
C 2
∇1 x1 (k) = −0.96 sin (x1 (k)) − 0.1x3 (k),



 0
 C 31




 0 ∇1 x2 (k) = 2x1 (k) − 0.98x2 (k) + 0.4x3 (k),
 2
 C ∇ 3 x3 (k) = 0.8 sin (x1 (k)) + x2 (k) − 1.2 sin (x3 (k)) ,


0 1 2

In order to handle this system, it is of course necessary to realize that the origin (0, 0, 0) is an equilibrium
point. Besides, the Jacobian matrix J is of the form:
 
 −0.96
 0 −0.1 

 
J =  2 −0.98 0.4  .

 

0.8 0 −1.2 

236
CNMA’2022
Second National Conference on
Mathematics and its Applications

Observe that M = 6, and so

   
 λ3 0 0   −0.96 0 −0.1 
   
   
det  0 λ2 0  −  2
  −0.98 0.4  = 0,

   
0 0 λ4
   0.8 0 −1.2 

⇔
λ9 + 0.98λ7 + 0.96λ6 + 1. 2λ5 + 0.940 8λ4 + 1. 176λ3 + 1. 232λ2 + 1. 207 4 = 0.
Accordingly, the solution of (19) will be as follows:

λ1 = −0.998 82
λ2 = 0.756 80 − 0.729 86i
λ3 = 0.756 80 + 0.729 86i
λ4 = −0.743 18 + 0.748 24i
λ5 = −0.743 18 − 0.748 24i
λ6 = −4. 687 8 × 10−6 + 0.989 95i
λ7 = −4. 687 8 × 10−6 − 0.989 95i
λ8 = 0.485 79 + 0.875 95i
λ9 = 0.485 79 − 0.875 95i
1
Obviously, one can deduce that λi ∈ K 6 , 1 ≤ i ≤ 9. Hence, in view of Corollary 9, one can deduce that
the trivial solution of system (19) is locally asymptotically stable.

Conclusion
In the present work, some simple applicable conditions for judging the stability of non-linear
incommensurate Fractional-order Difference Systems to have been reported. All results of this
work are applicable to be implemented in lots of difference systems. Such investigation together
with studying the dynamics of the linear incommensurate FoDSs will be some several targets that
left for future consideration.

References
[1] N Djenina, A Ouannas, T-E Oussaeif, G Grassi, I M Batiha, S Momani, R B Albadarneh.;
On the Stability of Incommensurate h-Nabla Fractional-Order Difference Systems. Fractal
and Fractional 06-00158 2022.

[2] Iyad Suwan, Shahd Owies, Thabet Abdeljawad. Monotonicity results for h-discrete frac-
tional operators and application. AdvancesinDifferenceEquations (2018).

[3] Elaydi. S, Murakami. S., Asymptotic stability versus exponential stability in linear volterra
difference equations of convolution type. J. Differ. Equ. Appl. 1996, 940, 35–46.

[4] Jan. Čermák, Luděk. Nechvátal., On a problem of linearized stability for fractional differ-
ence equations. Nonlinear Dyn (2021) 104:1253–1267.

[5] N. Djenina, A. Ouannas, I.M. Batiha, G. Grassi, V.-T. Pham, On the Stability of Linear
Incommensurate Fractional-Order Difference Systems, Mathematics 8 (2020) 1754.

237
CNMA’2022
Second National Conference on
Mathematics and its Applications

Existence Results for a System of Nonlinear

Fractional Differential Equations

BENMEHIDI HAMMOU 1 [Link]@[Link]

1 Department of nature and life sciences , University of Tiaret

Abstract: We present new ideas by establishing some results for the existence and uniqueness
of solutions for a class of nonlinear sequential fractional differential systems that involve
Hadamard, Riemann-Liouville and Caputo operators with initial conditions.

Keywords: Riemann-Liouville integral, Hadamard derivative, Sequential differential

equation.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
The domain of differential equations is regarded as an essential branch of mathematics, specif-
ically differential equations of fractional order, since the spread of these equations in other math-
ematical areas. And it has various applications in physical science, electrochemistry, biomathe-
matics, viscoelasticity, electromagnetic. Several authors have discussed the existence and unique-
ness of solutions of nonlinear sequential fractional equations. For example, we refer to the papers
[1, 2] .
Recently, in [2] some existence and the uniqueness of solutions results are given for the new class
of nonlinear sequential Riemann-Liouville and Caputo fractional differential equations
 h i

 RL D q C D r x (t) − g (t, x (t)) = f (t, x (t)) , 0 ≤ t ≤ T ,








x (ζ) = ϕ (x) , I p x (T ) = h(x)



where RL D q ,C D r denote the Riemann-Liouville and Caputo fractional derivatives of orders q and
r respectively with 0 < p, q < 1, I p is the Riemann-Liouville fractional integral of order p > 0.
Motivateed by the above results, in this paper, we are concerned with the study of the following
problem of sequential fractional type:

238
CNMA’2022
Second National Conference on
Mathematics and its Applications

 h i  
RL D β1 C D α1 x (t) − g = f1 t, x (t) ,H D δ1 y (t) , 1 ≤ t ≤ T ,
1 (t, x (t) , y (t))










 h i  
RL D β2 C D α2 y (t) − g (t, x (t) , y (t)) = f2 t, x (t) ,H D δ2 y (t) , 1 ≤ t ≤ T ,



 2



(231)





x (ζ) = ϕ1 (x) , I γ1 x (T ) = ψ1 (x) , 1<ζ<T












y (ζ) = ϕ2 (y) , I γ2 y (T ) = ψ1 (y)

1<ζ<T



 
where RL D βi ,C D αi and H D δi i = 1, 2 denote the Riemann-Liouville, Caputo fractional deriva-
tives
 and Hadamard
 fractional derivatives of orders βi , αi and δi respectively
 with
 0 < βi , α
 i , δi < 1,

γ
I i i = 1, 2 is Riemann-Liouville fractional integral of order 0 < γi < 1 i = 1, 2 and ϕi , ψi i = 1, 2 :
 
C ([1, T ] , R) −→ R are two given functionals and fi , gi i = 1, 2 : [1; T ] × R2 −→ R, are givens func-
tions.

Main results
In order to prove the existence of solutions we need some assumptions for used in the sequel,
thus we put :

T α1 Γ (β1 ) T α1 +β1 −1 T γ1 ζ α1 T α1 +γ1

!
Λ1 : = + +
Γ (α1 + 1) Γ (β1 + α1 ) |Ω1 | Γ (γ1 + 1) Γ (α1 + 1) Γ (α1 + γ1 + 1)

Γ (β1 ) T α1 +γ1 ζ α1 +β1 −1 ζ α1 T α1 +β1 +γ1 −1

!
1 Γ (β1 )
+ + ,
|Ω1 | Γ (β1 + 1) Γ (α1 + γ1 + 1) Γ (α1 + β1 + γ1 ) Γ (α1 + 1)

T α1 +β1 Γ (β1 ) T α1 +β1 −1 T γ1 ζ α1 +γ1 T α1 +β1 +γ1

!
Λ2 : = + +
Γ (α1 + β1 + 1) Γ (α1 + β1 ) |Ω1 | Γ (γ1 + 1) Γ (α1 + γ1 + 1) Γ (α1 + β1 + γ1 + 1)

Γ (β1 ) T α1 +β1 +γ1 ζ α1 +β1 −1 ζ α1 +β1 T α1 +β1 +γ1 −1

!
1 Γ (β1 )
+ ,
|Ω1 | Γ (β1 + 1) Γ (α1 + β1 + γ1 + 1) Γ (β1 + γ1 + 1) Γ (α1 + β1 + 1)

Γ (β1 ) T α1 +β1 +γ1 −1

!
1 Γ (β1 ) α1 +β1 +γ1 −1
Λ3 : = + T ,
|Ω1 | Γ (α1 + β1 ) Γ (γ1 + 1) Γ (β1 + γ1 + 1)
!
1 Γ (β1 ) α1 +β1 +γ1 −1 Γ (β1 ) α1 +β1 −1
Λ4 : = T + ζ ,
|Ω1 | Γ (α1 + β1 ) Γ (β1 + 1)

239
CNMA’2022
Second National Conference on
Mathematics and its Applications

and
T α2 Γ (β2 ) T α2 +β2 −1 T γ2 ζ α2 T α2 +γ2
!
0
Λ1 : = + +
Γ (α2 + 1) Γ (α2 + β2 ) |Ω2 | Γ (γ2 + 1) Γ (α2 + 1) Γ (α2 + γ2 + 1)

Γ (β2 ) T α2 +γ2 ζ α2 +β2 −1 ζ α2 T α2 +β2 +γ2 −1

!
1 Γ (β2 )
+ + ,
|Ω2 | Γ (β2 + 1) Γ (α2 + γ2 + 1) Γ (α2 + β2 + γ2 ) Γ (α2 + 1)

T α2 +β2 Γ (β2 ) T α2 +β2 −1 T γ2 ζ α2 +γ2 T α2 +β2 +γ2

!
0
Λ2 : = + +
Γ (α2 + β2 + 1) Γ (α2 + β2 ) |Ω2 | Γ (γ2 + 1) Γ (α2 + γ2 + 1) Γ (α2 + β2 + γ2 + 1)

Γ (β2 ) T α2 +β2 +γ2 ζ α2 +β2 −1 ζ q2 +p2 T r2 +q2 +p2 −1

!
1 Γ (β2 )
+ ,
|Ω2 | Γ (β2 + 1) Γ (α2 + β2 + γ2 + 1) Γ (β2 + γ2 + 1) Γ (α2 + β2 + 1)

Γ (β2 ) T α2 +β2 +γ2 −1

!
0 1 Γ (β2 ) α2 +β2 +γ2 −1
Λ3 : = + T ,
|Ω2 | Γ (α2 + β2 ) Γ (γ2 + 1) Γ (β2 + γ2 + 1)
!
0 1 Γ (β2 ) r2 +q2 +p2 −1 Γ (β2 ) α2 +β2 −1
Λ4 : = T + ζ .
|Ω2 | Γ (α2 + β2 ) Γ (β2 + 1)

(H1) :
  
Υ1 : = max |g1 (t, u (t) , v (t))| , Υ2 := max f1 t, u (t) ,H D δ1 v (t) 
1≤t≤T 1≤t≤T
0 0
  
Υ1 : = max |g2 (t, u (t) , v (t))| , Υ2 := max f2 t, u (t) ,H D δ2 v (t) 
1≤t≤T 1≤t≤T
  0 0  
Υ3 : = |ϕ1 (x)| , Υ4 := ψ1 (x) , Υ3 := |ϕ2 (x)| , Υ4 := ψ2 (x)

0
 
(H2) : There exists constants Ki , Ki > 0 i = 1, 4 , such that

|g1 (t, u1 , v1 ) − g1 (t, u2 , v2 )| ≤ K1 (|u1 − u2 | + |v1 − v2 |)

|f1 (t, u1 , v1 ) − f1 (t, u2 , v2 )| ≤ K2 (|u1 − u2 | + |v1 − v2 |)
0
|g2 (t, u1 , v1 ) − g2 (t, u2 , v2 )| ≤ K1 (|u1 − u2 | + |v1 − v2 |)
0
|f2 (t, u1 , v1 ) − f2 (t, u2 , v2 )| ≤ K2 (|u1 − u2 | + |v1 − v2 |) ,

and
   
ϕ1 (x) − ϕ1 (y) ≤ K3 x − y 
   
ψ1 (x) − ψ1 (y) ≤ K4 x − y 
  0  
ϕ2 (x) − ϕ2 (y) ≤ K3 x − y 
  0  
2ψ (x) − ψ (y)
2 ≤ K x − y 
4

where  
ui , vi i = 1, 2 are reals numbers, x, y ∈ C ([1, T ] , R) and t ∈ [1, T ] .

Our first result of the existence of the solution is based on Banach contraction pricipal theory.

Theorem 107. If we have the property (H2) and

0 0
P  P 
(1 + Γ (1 − δ1 )) 4i=1 Ki Λi (1 + Γ (1 − δ2 )) 4i=1 Ki Λi
+ <1
Γ (1 − δ1 ) Γ (1 − δ2 )

240
CNMA’2022
Second National Conference on
Mathematics and its Applications

Our second result is based on Shaefer’s fixed point theorem.

 
Theorem 108. If fi , gi i = 1, 2 : [1; T ] × R2 −→ R, are continuous functions and the assumption (H1)
is satisfied, then the problem (231) has at least one solution on [1, T ] .

References
[1] S. Asawasamrit, S. Ntouyas, [Link] and W. Nithiarayaphaks, Coupled systems of se-
quential caputo and hadamard fractional differential equations with coupled separated bound-
ary conditions, Symmetry, 10 (2018), 701.

[2] Asawasamrit, S. Phuangthong, N. Ntouyas, S. K. and J. Tariboon, Nonlinear sequential

Riemann-Liouville and Caputo fractional differential equations with nonlocal and integral
boundary conditions, International Journal of Analysis and Applications, 17(1),(2019), 47-
63.

[3] A.A. Kilbas, H.M. Srivastava and J.J. Trujillo, Theory and Applications of Fractional Dif-
ferential Equations, Elsevier, Amsterdam , 2006.

[4] K.S. Miller and [Link], An Introduction to the Fractional Calculus and Fractional Differen-
tial Equations, Wileyand Sons, New York , 1993.

241
CNMA’2022
Second National Conference on
Mathematics and its Applications

The 16th Hilbert problem for the class of

discontinuous piecewise differential
systems separated by irregular line and
formed by linear center and cubic
reversible isochronous centers having
rational first integrals

Imane Benabdallah 1 [Link]@[Link]

Rebiha Benterki 1 [Link]@[Link]
1 Department of Mathematics, University Mohamed El Bachir El Ibrahimi of Bordj Bou Arréridj 34000, El
Anasser, Algeria

Abstract: In this work we solve the second part of the sixteenth Hilbert’s problem for the
discontinuous piecewise differential systems separated by irregular line and formed by an
arbitrary linear center and an arbitrary differential cubic reversible isochronous center having
rational first integral, i.e we provide the maximum number of limit cycles of these classes of
discontinuous piecewise differential systems. We have also reinforced our results by giving
examples of discontinuous piecewise differential systems exhibiting the exact upper bound of
limit cycles of these systems.

Keywords: limit cycles, discontinuous piecewise differential systems, isochronous cen-

ter.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
The discontinuous piecewise differential system is a system of the form
  T

 F− (x, y) = F − (x, y), F − (x, y) (x, y) ∈ Σ− ,
1 2


(ẋ, ẏ) = F(x, y) = 

 T
 F (x, y) = F1+ (x, y), F2+ (x, y)

 +
 (x, y) ∈ Σ+ ,

where Σ is the separation line of the plane, Σ− and Σ+ are two regions in the plane.
Andronov, Vitt and Khaikin started the study of discontinuous piecewise linear differential sys-
tems in the plane [1]. Subsequently, these systems became a subject of wide attention in the
mathematical community because of their use in modeling real phenomena, see for example the
references cited therein.

242
CNMA’2022
Second National Conference on
Mathematics and its Applications

Finding out the non-existence, the existence of limit cycles and their number is one of the big
problems of the qualitative theory of planar differential systems, and in particular discontinuous
piecewise linear differential systems separated by a curve Σ. In this study, we consider that a
crossing limit cycle is an isolated periodic orbit in the set of periodic orbits of the system that has
exactly two points on the discontinuity curve Σ. The question of what is the upper bound on the
maximum number of limit cycles that a family of linear differential systems in the plane separated
by a straight line can have has been answered by many authors recently, see for example [2].
In 1990, Lum and Chua found that found that continuous (but not smooth) piecewise linear sys-
tems in the plane separated by a straight line have at most one limit cycle. This conjecture was
proved by Freire et al in 1998.
Han and Zhang in 2010 believe that linear discontinuous differential systems in the plane sepa-
rated by a straight line have at most one limit [Link], they have found that linear dis-
continuous differential systems in the plane separated by a straight line have at most two crossing
limit cycles.
In [5], Llibre and Ponce showed analytically the existence of these three limit cycles. Today, it
remains an open problem to know if three is the maximum number of crossing limit cycles that
this class of systems can have.

In this paper, we consider the discontinuous piecewise differential system formed by an arbi-
trary linear center and one of an arbirary cubic reversible isochronous center having rational first
integrals separated by the an irregular line Σ = Σ1 ∪ Σ2 where Σ1 = {(x, y) : x = 0 and y ≥ 0},
and Σ2 = {(x, y) : x ≥ 0 and y = 0}. This irregular line separate the plane into two regions,
Σ+ = {(x, y) : x > 0, y > 0} and Σ− = {(x, y) : x ≥ 0, y < 0} ∪ {(x, y) : x < 0}. Therefore in the first
region we consider one of the three classes of cubic reversible isochronous centers having rational
first integrals and in the second one we consider an arbitrary linear differential center.

The normal forme of a linear differential center is giving by this lemma.

Lemma 43. By doing a linear change of variables and a rescaling of the independent variable every
linear center can be written as

ẋ = −y(A2 + ω2 ) − Ax + B, ẏ = Ay + x + C, with ω > 0,

with the first integral

H(x, y) = (Ay + x)2 + 2(Cx − By) + ω2 y 2 . (232)

The normal forms of the three cubic reversible isochronous centers with a rational first integral
are given in the following theorem.
Theorem 109. After an affine change of variables and a rescaling of the independent variable the three
cubic reversible isochronous centers with rational first integrals can be expressed as one of the following
three differential systems.

(C1 ) ẋ = y(−1 + 2ax + 2bx2 ), ẏ = x + a(y 2 − x2 ) + 2bxy 2 ,

(C2 ) ẋ = −y(1 − x)(1 − 2x), ẏ = x − 2x2 + y 2 + 2x3 ,

 8 32  4
(C3 ) ẋ = y − 1 + x − y 2 , ẏ = x − y 2 .
3 9 3

243
CNMA’2022
Second National Conference on
Mathematics and its Applications

Main results

Theorem 110. The maximum number of crossing limit cycles with Conf 2 for the classes of discontin-
uous piecewise differential systems separated by Σ, and formed by the linear center (232) and

(I) the cubic reversible isochronous center (C̃1 ) is four, there are systems with exactly four limit cycles;

(II) the cubic reversible isochronous center (C̃2 ) is six, there are systems with exactly five limit cycles;

(III) the cubic reversible isochronous center (C̃3 ) is four, there are systems with exactly four limit cycles.

Theorem 111. The maximum number of crossing limit cycles with Conf 3 for the classes of discontin-
uous piecewise differential systems separated by Σ, and formed by the linear center (232) and

(I) the cubic reversible isochronous center (C̃1 ) is five, there are systems with exactly five limit cycles;

(II) the cubic reversible isochronous center (C̃2 ) is eight, there are systems with exactly eight limit
cycles;

(III) the cubic reversible isochronous center (C̃3 ) is five, there are systems with exactly five limit cycles.

References
[1] A. Andronov, A. Vitt and S. Khaikin, Theory of Oscillations, Pergamon Press, Oxford,
1966.

[2] J.C. Artés, J. Llibre, J.C. Medrado and M.A. Teixeira, Piecewise linear differential systems
with two real saddles, Math. Comput. Simul. 95 (2013), 13–22.

[3] M. di Bernardo, C. J. Budd, A. R. Champneys and P. Kowalczyk, Piecewise-Smooth Dy-

namical Systems: Theory and Applications, Appl. Math. Sci. Series 163, Springer-Verlag,
London, 2008.

[4] R. Benterki and J. Llibre, The limit cycles of discontinuous piecewise linear differential sys-
tems formed by centers and separated by irreducible cubic curves I, to appear in Dynamics of
Continuous, Discrete and Impulsive Systems-Series A, 2020.

[5] J. Llibre and E. Ponce, Three nested limit cycles in discontinuous piecewise linear differential
systems with two zones, Dyn. Contin. Discr. Impul. Syst., Ser. B. 19 (2012), 325–335.

244
CNMA’2022
Second National Conference on
Mathematics and its Applications

Well-posedness and exponential decay

of some evolution problems with internal
distributed delay

Sami Loucif 1
Rafik Guefaifia 2
1 Laboratory of Mathematics, Informatics and Systems (LAMIS), Larbi Tebessi University, Tebessa
2 Department of Mathematics and Computer Science, Larbi Tebessi University, Tebessa

Abstract: In this work, we consider one-dimensional system of piezoelectric beams with

distributed delay on the mechanical equation. By using Semigroups theory, we prove that
this system accepts only one solution. Next, we find the energy expression related to this
system, and by using technique of Lyapunov functional we demonstrate that this system is
exponentially stable irrespective of any condition on the coefficients of the system.

Keywords: Distributed delay, semigroup theory, Lyapunov functional, exponential sta-

bility.
2010 Mathematics Subject Classification: 35B35, 35B40.

Introduction
Piezoelectric materials have the property of converting from mechanical energy to electro-magnetic
energy, or of generating an internal electrical charge from applied mechanical pressure. The
brothers, Pierre and Jacques Curie, first demonstrated the direct piezoelectric effect in 1880 [2],
These piezoelectric materials are used in various industries of which, manufacturing, medical de-
vice industry, telecommunications and information technology. In [1], Morris and Özer they used
a variational approach to derive the differential equations and boundary conditions that model a
single piezoelectric beam with magnetic effects. Applying a Legendre transformation they obtain
Z T
L̃ = [K − (P + E) + B + W ] dt,
0

where K, P + E, B and W denote the (mechanical) kinetic energy, total stored energy, magnetic en-
ergy (electrical kinetic energy) of the beam and the work done by the external forces, respectively.
For a beam of length L and thickness h, they found

h L h2 2
Z " ! #
2 2
P +E = α vx + wxx − 2γβvx px + βpx dx,
2 0 12
Z L
µh
B= pt2 dx,
2 0

245
CNMA’2022
Second National Conference on
Mathematics and its Applications

L
h2 2
Z !
ρh
K= vt2 + 2
w + wt dx,
2 0 12 xt
and Z L
W= −px V (t) dx,
0
where V (t) denote the voltage applied at the electrodes. And by application of Hamilton’s prin-
ciple, setting the variation of admissible displacements {v, w, p} of L to zero, yields two sets of
equations one for stretching and one for bending with associated boundary conditions.
As the applied voltage V (t) affects only the stretching motion, so they neglected the equation of
bending, and they studied the stretching equations

 ρvtt − αvxx + γβpxx = 0,




 µptt − βpxx + γβvxx = 0,



with the boundary and initial conditions


V (t)
 v (0) = p (0) = αvx (L) − γβpx (L) = 0, βpx (L) − γβvx (L) = − h ,



  
(v, p, v , p ) = v 0 , p0 , v 1 , p1 .


 t t

Finally, by using only an electrical feedback controller (the current flowing through the elec-
trodes), they show that the closed-loop system is strongly stable in the energy space. Ramos et al.
[2] studied the well-posedness of solution for piezoelectric beams with magnetic effect

 ρvtt − αvxx + γβpxx = 0 in (0, L) × (0, T ) ,




 µptt − βpxx + γβvxx = 0 in (0, L) × (0, T ) ,



with the following conditions


v (L,t)
v (0, t) = αvx (L, t) − γβpx (L, t) + ξ1 t h = 0, 0<t <T,





 pt (L,t)
 p (0, t) = βpx (L, t) − γβvx (L, t) + ξ2 h = 0, 0<t <T,




v (x, 0) = v0 (x) , vt (x, 0) = v1 (x) , ∀x ∈ (0, L) ,








 p (x, 0) = p0 (x) , pt (x, 0) = p1 (x) , ∀x ∈ (0, L) .

In the case ξ1 = ξ2 = 0 in system (??)-(??), they obtain the following one-dimensional conservative
system 
 ρutt − αuxx + γβzxx = 0 in (0, L) × (0, ∞) ,




 µztt − βzxx + γβuxx = 0 in (0, L) × (0, ∞) ,



with the boundary and initial conditions


u (0, t) = αux (L, t) − γβzx (L, t) = 0, ∀t > 0,






 z (0, t) = zx (L, t) − γux (L, t) = 0, ∀t > 0,




u (x, 0) = u0 (x) , ut (x, 0) = u1 (x) , ∀x ∈ (0, L) ,







 z (x, 0) = z0 (x) , zt (x, 0) = z1 (x) ,
 ∀x ∈ (0, L) ,

246
CNMA’2022
Second National Conference on
Mathematics and its Applications

and by using multiplicative techniques, they obtain observability inequality of a conservative

system. Also by using the auxiliary problem given by

 ρφtt − αφxx + γβψxx = 0 in (0, L) × (0, T ) ,




 µψ − βψ + γβφ = 0 in (0, L) × (0, T ) ,

 tt xx xx

where p = ψ + z, v = φ + u and (u, z) solution of conservative problem.

With the boundary and initial conditions


v (L,t)
φ (0, t) = αφx (L, t) − γβψx (L, t) + ξ1 t h = 0, ∀t > 0,





 pt (L,t)
 ψ (0, t) = βψx (L, t) − γβφx (L, t) + ξ2 h = 0, ∀t > 0,



 (233)
φ (x, 0) = φ0 (x) = φt (x, 0) = φ1 (x) = 0, ∀x ∈ (0, L) ,








 ψ (x, 0) = ψ0 (x) = ψt (x, 0) = ψ1 (x) = 0, ∀x ∈ (0, L) ,
and by using some lemmas, they prove equivalence between stabilization and observability. Re-
cently, Ramos et al. [3] they prove the exponential stability for system of piezoelectric beams with
delayed 
 ρvtt − αvxx + γβpxx + ξ1 vt + ξ2 vt (x, t − τ) = 0, in (0, L) × (0, +∞) ,


(234)


 µptt − βpxx + γβvxx = 0, in (0, L) × (0, +∞) ,



where ξ2 vt (x, t − τ) is the time of delay on vertical displacement, τ > 0 is the respective retardation
time, where they proved this stability under the conditions ξ1 > ξ2 .

Main results
Motivated by the above works, in the present work, we consider the following problem
 R τ2



 ρvtt − αvxx + γβpxx + µ1 vt + τ1
µ2 (s) vt (x, t − s) ds = 0, in (0, L) × (0, ∞) ,





 µptt − βpxx + γβvxx = 0, in (0, L) × (0, ∞) ,


v (0, t) = αvx (L, t) − γβpx (L, t) = 0, t ≥ 0,






t ≥ 0, (235)


 p (0, t) = px (L, t) − γvx (L, t) = 0,


x ∈ (0, L) ,




 v (x, 0) = v0 (x) , vt (x, 0) = v1 (x) ,





 p (x, 0) = p0 (x) , pt (x, 0) = p1 (x) , x ∈ (0, L) ,


vt (x, −t) = f0 (x, t) , (x, t) ∈ (0, L) × (0, τ2 ) ,



where the parameters ρ, α, γ, µ, β and L represent respectively, the mass density, elastic stiffness,
piezoelectric coefficient, magnetic permeability, water resistance coefficient of the beam and the
length of the beam. In addition, the following condition is considered

α1 = α − γ 2 β > 0, (236)

and τ1 , τ2 , µ1 are positive numbers where τ2 ≥ τ1 and µ2 : [τ1 , τ2 ] → R is a bounded function verify
the following assumption Z τ2 
µ (s) ds ≤ µ . (237)
2 1
τ1

247
CNMA’2022
Second National Conference on
Mathematics and its Applications

This presentation is organized as follows: In Section 2, by using Hille-Yosida Theorem (see [4],[5])
we prove the well-posedness for the system (235). In Section 3, we construct the Lyapunov func-
tionals and by exploiting the conditions (236),(237) we establish an exponential stability of the
system (235).

References
[1] K. A. Morris, A. Ö. Özer, Strong stabilization of piezoelectric beams with magnetic effects, in
Proc. of 52nd IEEE Conference on Decision and Control, 2013, 3014-3019.

[2] K. A. Morris, A. Ö. Özer, Equivalence between exponential stabilization and boundary ob-
servability for piezoelectric beams with magnetic effect, Z. Angew. Math. Phys. 70(60) (2019),
1-14.

[3] A. J. A. Ramos, A. Ö. Özer, M. M. Freitas, D. S. Almeida Júnior, J. D. Martins, Exponential

stabilization of fully dynamic and electrostatic piezoelectric beams with delayed distributed
damping feedback, Z. Angew. Math. Phys.72(1) (2021), 1-15.

[4] A. Pazy, Semigroups of Linear Operator and Applications to Partial Differential Equations,
Applied Mathematical Sciences. Springer-Verlag, New York, 1983.

[5] S. Zheng, Nonlinear evolution equations, CRC Press, 2004.

248
CNMA’2022
Second National Conference on
Mathematics and its Applications

Limit cycles generated by piecewise

linear Hamiltonian systems without
equilibria with three pieces

Louiza Baymout 1 [Link]@[Link]

Rebiha Benterki 1

1 Department of Mathematics, University of Mohamed El Bachir El Ibrahimi of Bordj Bou Arréridj 34000, El
Anasser, Algeria

Abstract: The importance of studying piecewise linear differential systems has grown in
recent years, due to their applications. Like we can see the appearance of this kind of system
in modeling many natural phenomena, as in physics, biology, economics, etc. It is well known,
that the limit cycles play a main role in the study of qualitative theory of piecewise differential
systems. In most of the published papers that studied the limit cycles of piecewise differential
systems formed by linear systems consider only two pieces. In this paper we investigate the
maximum number of limit cycles for a family of piecewise linear differential systems formed
by linear Hamiltonian differential systems without equilibria, where the separation curve
splits the plane into three pieces and that made a big difference. First we prove that the
systems that intersected the separation curve in three points exhibit 1, 2 or 3 limit cycles.
Second we prove that the systems that intersected the separation curve in four points exhibit
1 limit cycle.

Keywords: Hamiltonian systems, equilibrium points, limit cycles, discontinuous piece-

wise differential systems, separation curve.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

References
[1] A. Andronov, A. Vitt and S. Khaikin, Theory of Oscillations, Pergamon Press, Oxford,
1966.

[2] J.C. Artés, J. Llibre, J.C. Medrado and M.A. Teixeira, Piecewise linear differential systems
with two real saddles, Math. Comput. Simul. 95 (2013), 13–22.

[3] J. Llibre, D.D. Novaes, M.A. Teixeira, Maximum number of limit cycles for certain piecewise
linear dynamical systems, Nonlin. Dyn. 82 (2015), 1159—1175.

[4] R. Benterki and J. Llibre, The limit cycles of discontinuous piecewise linear differential sys-
tems formed by centers and separated by irreducible cubic curves I, to appear in Dynamics of
Continuous, Discrete and Impulsive Systems-Series A, 2020.

[5] F. Mañnosas and J. Villadelprat, Mémoire sur les courbes définies par une équation differ-
entielle, J. Differential Equations. 179 (2002), 625–646.

249
CNMA’2022
Second National Conference on
Mathematics and its Applications

Automorphy in a new genetic regulatory

networks model with several delays

Mohammed Salah M’hamdi 1 [Link]@[Link]

1 Research unit LaMOS, Department of Mechanical Engineering, University of Bejaia, 06000 Bejaia, Algeria

Abstract: In this paper, we consider a new class of delayed genetic regulatory networks
(DGRNs) belonging to the category of delayed differential equations (DDEs), which is more
general than the recent genetic regulatory networks with time-varying delays : bounded or
unbounded (proportional) delays. In this work, under some delay-independent sufficient
conditions, the existence and the uniqueness of pseudo almost automorphic solution for a
new genetic regulatory networks (GRNs) with time-varying coefficients and several delays
are obtained. To do so, the theory of exponential dichotomy with the contraction mapping
principle (the classical Banach’s fixed-point principle) and inequality techniques are used. As
you will see, our results improve and generalize many previous known results in DGRNs.

Keywords: genetic regulatory networks, several delays, pseudo almost automorphic so-
lution.
2010 Mathematics Subject Classification: Primary 34C27, 34K14, 35B15.

Introduction
Dynamics and applications of differential equations (DE) and delay-differential equations (DDE)
have taken great interests of many authors in recent years. We refer the reader to neural net-
works [1, 2] and the genetic regulatory networks [3, 4]. These two models belong to the category
of delayed differential equations, which are widely used in many fields. Therefore, the study on
the dynamic behaviors of the genetic regulatory networks has important theoretical and practical
value. Qualitative analysis such as periodicity and automorphy of delayed differential equations
(DDEs) has been studied extensively by many authors. The authors in [4], have studied the fol-
lowing genetic regulatory networks with almost periodic coefficients and time-varying delays as
follows :
 n
dmi (t) P
= −c (t)m (t) + wij (t)hj (pj (t − ηj (t))) + Ii (t),


 dt

 i i

 j=1 (238)
 dp i (t)


dt = −bi (t)pi (t) + di (t)mi (t − τi (t))), i = 1, 2, . . . , n,

the initial conditions of (238) are associated to be of the form

mi (s) = φi (s), s ∈ [−τ, 0], τ = max sup{τi (t)}, and pi (s) = ψi (s), s ∈ [−η, 0], η = max sup{ηi (t)},
1≤i≤n t∈R 1≤i≤n t∈R

where φi (s) and ψi (s) are continuous functions on R. In model (238) :

Description 5.

250
 CNMA’2022
Second National Conference on
Mathematics and its Applications

• mi (t) and pi (t) are the concentrations of mRNA and protein of the ith node at a time t, respectively;

• ci (t) and bi (t) denote the degradation or dilution rates of mRNA and protein, respectively;

• di (t) represents the translation rate; and wij (t) is defined as follows :



 > 0, if transcription j is an activator of gene i,


wij (t)  = 0, if there is no link from node j to i,




 < 0, if transcription j is a repressor of gene i,


• hj (t) denotes the feedback regulation of the protein on the transcription, which is a monotonic function
in Hill form
pj (t)/βj
hj (pj (t)) = ,
1 + (pj (t)/βj )Hj
with Hj being the Hill coefficient and βj being a positive scalar.
Evidently, hj is a monotonically increasing and differentiable function with saturation, it satisfies

dhj (s)
0≤ ≤ Lj
ds
which is equivalent to
hj (s1 ) − hj (s2 )
0≤ ≤ Lj , f or all s1 , s2 ∈ R;
s1 − s2

• τi (t) and ηj (t) are the time-varying delays;

• Ii (t) is the base transcriptional rate of the repressor of gene i at a time t,

Motivated by the above discussions, in this paper, we consider a generalized class of genetic reg-
ulatory networks with pseudo almost automorphic coefficients and several delays more general
than the systems in [1], and (238) of [2], described by the following system of differential equa-
tions:
 n
dmi (t) P
= −ci (t)mi (t) + wij (t)hj (pj (b
ηj (t))) + Ii (t),


 dt

j=1 (239)



 dpi (t)


dt = −bi (t)pi (t) + di (t)mi (b
τi (t))), i = 1, 2, . . . , n, t > t0 ,

with the initial conditions associated are of the form


 mi (s) = φi (s), s ∈ [τi (t0 ), t0 ], t0 > 0,




 pi (s) = ψi (s), s ∈ [b
ηi (t0 ), t0 ], t0 > 0,



here,

• b
τi are delay factors and φi (.) ∈ C([b τi (t0 ), t0 ], R) denotes the initial value of mi on [b
τi (t0 ), t0 ].
Note that, τi (t) < t for all t > t0 . Then b τi (t) is referred to as delayed argument and the term
τi (t) is called delay.
t −b

• η
bi are delay factors and ϕi (.) ∈ C([b ηi (t0 ), t0 ], R) denotes the initial value of pi on [b
ηi (t0 ), t0 ].
Note that, ηi (t) < t for all t > t0 . Then ηi (t) is referred to as delayed argument and the term
b b
t−ηbi (t) is called delay.

251
CNMA’2022
Second National Conference on
Mathematics and its Applications

We distinguish two types of delay for b

τi or η
bi , in case of delays b
τi :

• if there exists M > 0 such that t − b

τi (t) < M for all t > t0 then we have the bounded delay,

• if t − b
τi (t) → ∞ as t → ∞ we talk about the unbounded delay.

A prototype of a bounded delay is the constant delay, i.e. τi (t) = t − τ ∗ , where τ ∗ ∈ R+ . On the
contrary, an example of an unbounded delay is the proportional one, i.e. τi (t) = qt, where q ∈ (0, 1)
is a real parameter.
Throughout this paper, BC(R, Rn ) denotes the set of bounded continued functions from R to Rn .
Note that (BC(R, Rn ), k . k∞ ) is a Banach space where the sup norm is given by :

k f k∞ := sup k f (t) k:= max sup | fi (t) | .

t∈R 1≤i≤n t∈R

Definition 46. [1, 2]. A continuous function f : R → Rn is said to be almost automorphic if for any
sequence of real numbers (sn0 ), there exists a subsequence (sn ) such that
g(t) = lim f (t + sn ) is well-defined for each t ∈ R and lim g(t − sn ) = f (t) for each t ∈ R.
n→∞ n→∞
Denote by AA(R, Rn ) the set of all such functions.

1 T
R
We denote by AA0 (R, Rn ) = {f ∈ BC(R, Rn ) : lim k f (t) k dt = 0}.
T −→∞ 2T −T

Definition 47. [1, 2]. A continuous function f ∈ BC(R, Rn ) is said to be pseudo almost automorphic if
it can be decomposed as f = f1 + f2 , where f1 ∈ AA(R, Rn ) and f2 ∈ AA0 (R, Rn ).
The collection of such functions will be denoted by P AA(R, Rn ).

Main results
In this section, we establish some results for the existence and the uniqueness of pseudo almost-
automorphic solutions of (239) in a suitable convex set.
For that, we assume that the following conditions hold.
(H1) For all 1 ≤ i ≤ n, the functions b bi : R+ → R+ , are continuously differentiables on R+ , and
τi , η
τi0 (.), η
b bi0 (.) > 0, are nondecreasing with
! !
τi (T )
b bi (T )
η
lim sup < ∞, lim sup <∞
T →∞ τi0 (t))
T inf+ (b T →∞ ηi0 (t))
T inf+ (b
t∈R t∈R

(H2) The functions t 7−→ ci (t), t 7−→ bi (t) are almost-automorphic with

inf(ci (t)) = ci∗ > 0, inf(bi (t)) = bi∗ > 0.

t∈R t∈R

(H3) The functions t 7−→ di (t), t 7−→ wij (t), t 7−→ Ii (t) are pseudo almost-automorphic with

di+ = sup | di (t) |, Ii+ = sup | Ii (t) |, wij

+
= sup | wij (t) | .
t∈R t∈R t∈R

In order to obtain the main result of this paper, we shall first state several lemmas which will be
useful in the proving the main result.

Lemma 44. [1, 2]. If ϕ, ψ ∈ P AA(R, R) and k ∈ R, then ϕ + ψ, ϕ × ψ, ϕ(. − k) ∈ P AA(R, R).

Lemma 45. Assume that both b

τi (.) satisfy (H1). If mi (.) ∈ P AA(R, R), then mi (b
τi (.)) ∈ P AA(R, R).

252
CNMA’2022
Second National Conference on
Mathematics and its Applications

Theorem 112. Under the conditions (H1)-(H3),and (H4) : assume that there exist nonnegative con-
stant r such that
n
P +
( ( Lj wij ) ( + ))
j=1 d
r = max max , max i < 1,
1≤i≤n ci∗ 1≤i≤n bi∗

then the delayed GRNs of (??) has a unique pseudo almost automorphic solution in the region

rβ
B = {(ϕ, ψ)/(ϕ, ψ) ∈ P AA(R, R2n ), k (ϕ, ψ) − (ϕ, ψ)0 k∞ ≤ },
(1 − r)

where
!T
t t Ii+
Z Rt Z Rt ( )
− c1 (u)du − cn (u)du
(ϕ, ψ)0 (t) = e s I1 (s)ds, . . . , e s In (s)ds, 0, . . . , 0 , and β = max .
−∞ −∞ 1≤i≤n ci∗

References
[1] M.S. M’hamdi, Pseudo almost automorphic solutions for multidirectional associative memory
neural network with mixed delays, Neural Processing Letters. 49.3 (2019), 1567-1592.

[2] C. Aouiti, M.S. M’hamdi, A. Touati, Pseudo almost automorphic solutions of recurrent neural
networks with time-varying coefficients and mixed delays, Neural Processing Letters. 45.1
(2017), 121–140.

[3] C. Aouiti, F. Dridi, Study of genetic regulatory networks with Stepanov-like pseudo-weighted
almost automorphic coefficients, Neural Computing and Applications. 33.16 (2021): 10175-
10187.

[4] L. Duan, F. Di, Z. Wang, Existence and global exponential stability of almost periodic so-
lutions of genetic regulatory networks with time-varying delays, Journal of Experimental &
Theoretical Artificial Intelligence. 32.3 (2020), 453–463.

253
CNMA’2022
Second National Conference on
Mathematics and its Applications

Caputo−Hadamard random fractional

differential equations in finite and infinite
dimensional Banach spaces

BEKADA Fouzia 1 bekadafouzia@[Link]

1 Department of mathematics , University Dr Moulay Tahar SIADA

Abstract: This article deals with some existence of random solutions for a class of Caputo-
Hadamard random fractional differential equations with two boundary conditions in finite
and infinite dimensional Banach spaces. Our results are based on some random fixed point
theorems and the measure of noncompactness. Some illustrative examples are presented in
the last section.

Keywords and phrases: Fractional differential equation; mixed Hadamard integral of

fractional order; Caputo-Hadamard fractional derivative; random solution; fixed point.
AMS (MOS) Subject Classifications: 26A33.

Introduction
The functional differential equations with random effects are differential equations with a stochas-
tic process, they play a very important fundamental role in the theory of random dynamic sys-
tems, in addition they are used in various branches of science and engineering.

In this paper we investigate the existence of random solutions for the following class of Caputo-
Hadamard fractional differential equation

(Hc D1r u)(t, w) = f (t, u(t, w), w); t ∈ I := [1, T ], w ∈ Ω, (240)

with the boundary conditions 

u(1, w) = u1 (w)


; w ∈ Ω, (241)
u 0 (T , w) = uT (w)



where r ∈ (1, 2], T > 1, f : I × R × Ω → R is a given function, u1 , uT : Ω → R, Hc D1r is the Caputo-

Hadamard fractional derivative of order r, and Ω is the sample space in a probability space (Ω, F).

Next, we consider the problem (240)-(241), where f : I × E × Ω → E is a given function, u1 , uT :

Ω → E, and E is a real (or complex) Banach space with a norm k · k.

Main results
In the sequel, we will use the following fixed point Theorems:

254
CNMA’2022
Second National Conference on
Mathematics and its Applications

Theorem 113. Let X be a nonempty, closed convex bounded subset of the separable Banach space E and
let N : Ω×X → X be a compact and continuous random operator. Then the random equation N (w)u = u
has a random solution.

Theorem 114. Let X be a separable closed convex subset of Banach space, f : Ω × X → X a condensing
random operator. Suppose that for any w ∈ Ω, f (w, X) is bounded. then there exists a random fixed
Point ξ : Ω → X of f .

Existence Results in the Scalar Case

Let C(I, E) be the Banach space of all continuous functions from I into E with the norm

kuk∞ = sup{ku(t)k : t ∈ I}.

In particular C := C(I, R) is assumed to be endowed with the standard norm

kuk∞ = sup{|u(t)| : t ∈ I}.

Theorem 115. Assume that the hypotheses (H1 ) and (H2 ) hold. If

(ln T )r (ln T )r
!
∗
p2 (w) +T < 1,
Γ (r + 1) Γ (r)

then the problem (240)-(241) has a random solution defined on I × Ω.

Existence Results in Banach Space

In this section we prove the existence of random solutions for our problem in the Banach space E
by using the measure of noncompactness.

Theorem 116. Assume (H10 )-(H30 ) hold. If

(ln T )r ∗ T (ln T )r ∗
" #
M := 4 l (w) + l (w) ≤ 1,
Γ (r + 1) 2 Γ (r) 2

then the problem (240)-(241) has at least one solution defined on I.

References
[1] S. Abbas, M. Benchohra and G. M. N’Guérékata, Topics in Fractional Differential Equations,
Springer, New York, 2012.

[2] S. Abbas, M. Benchohra and G. M. N’Guérékata, Advanced Fractional Differential and Inte-
gral Equations, Nova Science Publishers, New York, 2015.

[3] F. Bekada, S. Abbas, and M. Benchohra,Boundary value problem for Caputo–Fabrizio ran-
dom fractional differential equations, Moroccan J. Pure Appl. Anal.(MJPAA)6 (2) (2020),
218-230.

[4] F. Bekada, S. Abbas, M. Benchohra, and [Link], Dynamics and stability for Katugampola
random fractional differential equations, AIMS Mathematics (2021), 8654-8666.

255
CNMA’2022
Second National Conference on
Mathematics and its Applications

Asymptotic stability analysis of

Riemann-Liouville fractional stochastic
neutral differential equations

Chahra Kechar 1 chahra95kechar@[Link]

Abdelouaheb Ardjouni 2
1 Department of Mathematics and Informatics, University of SoukAhras, Algeria
2 Department of Mathematics, University of Annaba, Algeria

Abstract: The novelty of our paper is to establish results on asymptotic stability of mild
solutions in pth moment to Riemann-Liouville fractional stochastic neutral differential equa-
tions (for short Riemann-Liouville FSNDEs) of order a ∈ ( 12 , 1) using a Banach’s contraction
mapping principle. The core point of this paper is to derive the mild solution of FSNDEs
involving Riemann-Liouville fractional time-derivative by applying the stochastic version
of variation of constants formula. The results are obtained with the help of the theory of
fractional differential equations, some properties of Mittag-Leffler functions and asymptotic
analysis under the assumption that the corresponding fractional stochastic neutral dynamical
system is asymptotically stable.

Keywords: Riemann-Liouville fractional derivative, fractional stochastic neutral dynam-

ical systems, existence and uniqueness, asymptotic stability, continuity of mild solutions in
pth moment
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
Over the years, many results have been investigated on the theory and applications of stochas-
tic differential equations (SDEs) . The deterministic models often oscillate due to noise. Certainly,
the extension of these models is essential to consider stochastic models, in which the connected
parameters are considered as appropriate Brownian motion and stochastic processes. The mod-
eling of most problems in real-world problems is described by stochastic differential equations
rather than deterministic equations. Thus, it is of great importance to design stochastic effects
in the study of fractional-order dynamical systems. In particular, fractional stochastic differen-
tial equations (FSDEs) which are a generalization of differential equations by the use of fractional
and stochastic calculus are more popular due to their applications in modeling and mathematical
finance.
Recently, FSDEs are intensively applied to model mathematical problems in finance , dynamics
of complex systems in engineering and other areas . Most of the results on fractional stochastic
dynamical systems are limited to prove existence and uniqueness of mild solutions using fixed
point theorem .

256
CNMA’2022
Second National Conference on
Mathematics and its Applications

Results on the asymptotic behavior of solutions of fractional differential equations with Caputo
and Rieamnn-Liouville fractional time-derivative are relatively scarce in the literature. Mahmu-
dov < 17 > derived an explicit solution formula to linear inhomogeneous delayed Langevin equa-
tion involving two Riemann-Liouville fractional derivatives and studied existence and unique-
ness, and Ulam-Hyers stability of solutions. In < 9 >, Cong et al. investigate the asymptotic
behavior of solutions of the perturbed linear fractional differential system. Cong et al. < 10 >
proved the theorem of linearized asymptotic stability for fractional differential equations. More
precisely, they showed that an equilibrium of a nonlinear Caputo fractional differential equation
is asymptotically stable if its linearization at the equilibrium is asymptotically stable. There are
only a few papers related to asymptotic stability of solutions of fractional stochastic differential
equations which can be found in < 34 >. Sakthivel et al. < 28 > studied existence and asymptotic
stability in pth moment of a mild solution to a class of nonlinear fractional neutral stochastic
differential equations with infinite delays in Hilbert spaces. The same asymptotic stability in pth
moment of a mild solutions of nonlinear impulsive stochastic differential equations and impulsive
stochastic partial differential equations with infinite delays was discussed in < 29 > and < 30 >,
respectively.

Main results
To the best of our knowledge, the asymptotic stability of mild solutions for fractional stochas-
tic neutral differential equations with Riemann-Liouville fractional derivative are an untreated
topic in the present literature. Due to lack of asymptotic stability of mild solutions to Riemann-
Liouville FSNDEs, this motivates us to establish new results on the asymptotic analysis of frac-
tional stochastic differential equations with Riemann-Liouville fractional time-derivative involv-
ing matrix coefficients.
Therefore, the plan of this paper is systematized as below: Section 2 is a preparatory section
where we recall some basic notions and results from fractional calculus and fractional differen-
tial equations. Then we resort the setting for main results of the theory and we impose certain
assumptions, definitions of stability and asymptotic stability in pth moment of mild solutions to
Riemann-Liouville FSNDEs stochastic analysis. In Section 3, first we verify the continuity of op-
erator in pth moment on [0, ∞). Then we show global existence and uniqueness of mild solution
under various assumptions by a Banach’s contraction mapping principle. Section 4 is devoted to
proving asymptotic stability of mild solutions to Riemann-Liouville FSNDEs. With the help of
properties of Mittag-Leffler functions, we show that ψ is well-defined. Finally, we study asymp-
totic stability in pth of Riemann-Liouville FSNDEs of fractional-order a ∈ ( 12 , 1). Section 5 is for
the conclusion and future work by providing several open problems.
The main contribution of our results is that we have opened the possibility for a cooperative
investigation to solve several issues, for instance, combining the methods of this paper to study
the control theory, one may solve controllability of nonlinear case of our problem in finite and
infinite dimensional spaces and one can also discuss finite-time stability of semiliniear fractional
stochastic differential equations.

References
[1] A. Ahmadova, N. I. Mahmudov, Asymptotic stability analysis of Riemann-Liouville frac-
tional stochastic neutral differential equations (2021)

[2] A. Ahmadova, N. I. Mahmudov, Existence and uniqueness results for a class of frac-
tional stochastic neutral differential equations, Chaos Soliton Fract. 139 (2020)

257
CNMA’2022
Second National Conference on
Mathematics and its Applications

[Link] ,

[3] P. C. Jean, L. S. Gangaram, Stochastic fractional differential equations: Modeling,

method and analysis, Chaos, Solitons Fract. 45 (2012) 279-293. ,

[4] R. Sakthivel, P. Revathi, Y. Ren, Existence of solutions for nonlinear fractional

stochastic differential equations, Nonlinear Anal. Theory Methods Appl. 81 (2013) 70–
86. ,

[5] S. G. Samko, A. A. Kilbas, O. I. Marichev, Fractional Integrals and Derivatives: Theory

and Applications, Gordon and Breach, New York, 1993.,

258
CNMA’2022
Second National Conference on
Mathematics and its Applications

Crossing limit cycles for a class of

piecewise differential system separated by
a straight line

Meriem Barkate 1 [Link]@[Link]

Rebiha Benterki1 [Link]@[Link]
1 Department of mathematics, University Mohamed El Bachir El Ibrahimi of Bordj ´ Bou Arreridj 34000, El
Anasser, Algeria

Abstract: In this work we study the discontinuous planar piecewise differential systems
formed by two differential systems separated by straight line y = 0, when these differential
systems are polynomial potential systems or five families of quadratic and cubic isochronous
centers. We prove that a sharp upper bound for the number of crossing limit cycles are reached.

Keywords: polynomial potential systems, cubic isochronous centers, quadratic isochronous

centers.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
In this paper we deal with discontinuous piecewise differential systems of the form


− − − T
 F (x) = (F1 (x), F2 (x)) x ∈ Σ− ,



ẋ = F(x) = 
 F+ (x) = (F1+ (x), F2+ (x))T x ∈ Σ+ .



with x = (x, y). Where Σ− and Σ+ are two regions in the plane defined by

Σ− = {(x, y) : x ≤ 0}, and Σ+ = {(x, y) : x ≥ 0}.

The set Σ = {(x, y) : x = 0} is called the separation line of the plane.

In 1920 Andronov, Vitt and Khaikin started the study of the piecewise differential systems sep-
arated by straight line and nowadays such systems have deserved the attention by many re-
searchers. These differential systems are used extensively to model biological process as well
as some electronics and mechanical applications. A limit cycle is a periodic orbit of the differen-
tial system isolated in the set of all periodic orbits of the system. This concept was defined by
Poincaré [5] at the end of the 19th century. The study of the existence and the number of limit
cycles for the discontinuous piecewise differential systems is one of the main problem. Thus limit
cycles have played and still playing an important rol in physical phenomena, see for instance [2].
In [2] Han and Zhang conjectured that discontinuous piecewise linear differential systems in the
plane separated by a straight line have at most two crossing limit cycles, but in [3] Llibre and
Ponce provided a negative answer to this conjecture by presenting an example with exactly three

259
CNMA’2022
Second National Conference on
Mathematics and its Applications

limit cycles. Many papers are dedicated to study the existence of limit cycles for the piecewise
linear differential systems, when the curve of separation is either a straight line, or an algebraic
curve. In the literature we find many papers interesting in solving the second part of the sixteenth
Hilbert problem for linear discontinuous piecewise differential systems, but few papers devoted
to solve this problem for nonlinear piecewise differential systems. In this paper, we deal with
the following five classes of isochronous quadratic and cubic centers. The quadratic polynomial
differential systems with an isochronous center were classified into four classes by Loud [4]. we
list the four classes of quadratic isochronous centers and their first integrals as follows. (i) The
quadratic isochronous differential system

   
ẋ = x2 − y 2 , ẏ = x 2y + 1 ,

With the first integral

x2 + y 2
H1 (x, y) = .
2y + 1
(ii) The quadratic isochronous differential system

ẋ = x2 − y, ẏ = x 1 + y),

With the first integral

x2 + y 2
H2 (x, y) = .
(y + 1)2
(iii) The quadratic isochronous differential system

x2

ẋ = − y, ẏ = x 1 + y),
4
With the first integral
x2 + 4y + 8
H3 (x, y) = .
y +1
Now we give the two classes of isochronous cubic differential centers I) The first class is given by:
   
ẋ = y 2K1 x + 2K2 x2 − 1 , ẏ = K1 y 2 − x2 + 2K2 xy 2 ,

Which has the first integral

x2 + y 2 + 8
H4 (x, y) = .
1 − 2x(K1 + K2 x)
(II) The second class is:
8x 32y 2 4y 2

ẋ = y − − 1), ẏ = x − ,
3 9 3
and its first integral is
 2
H5 (x, y) = 3x − 4y 2 + 9y 2 .

The main result of this work is to study the upper bounds of crossing limit cycles for discontinuous
piecewise differential systems separatred by the straight line y = 0, and formed by polynomial
potential systems or one families of quadratic and cubic isochronous centers, after an arbitrary
affine change of variables.

260
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] R. Benterki and J. Llibre, . Crossing limit cycles of planar piecewise linear Hamiltonian
systems without Equilibrium Points, , Mathematics. 8 (2020) 755, pp. 14.

[2] M. Han M, W. Zhang, On hopf bifurcation in non-smooth planar systems, J. Differential

Equations.9 (2010), 2399–2416.

[3] S.M. Huan and X.S. Yang , On the number of limit cycles in general planar piecewise linear
systems, Disc. Cont. Dyn. Syst. 32 (2012), 2147–2164.

[4] [Link] W, Behaviour of the period of solutions of certain plane autonomous systems near cen-
ters. Contrib, Differ. Equations. 3 (1964), 21–36.

[5] M. Poincaré , Mémoire sur les courbes définies par une équations differentielle, IV. J. Math,
Pures Appl, Sér. 4 (1886), 155–217.

261
CNMA’2022
Second National Conference on
Mathematics and its Applications

Limit cycles of polynomial differential

systems with homogeneous nonlinearities
of degree 8 via the averaging method

Ahlam Belfar 1 [Link]@[Link]

Rebiha Benterki 1 [Link]@[Link]
1 Department of mathematics, University Mohamed El Bachir El Ibrahimi Bordj Bou Arreridj El Anasser, Algeria

Abstract: In this work, we give the seven global phase portraits in the Poincaré disc of the
Kukles differential systems given by

ẋ = −y, ẏ = x + ax8 + bx4 y 4 + cy 8 ,

where x, y ∈ R and a, b, c ∈ R with a2 + b2 + c2 , 0. Moreover, we perturb the system inside

all classes of polynomials of eight degrees, then we use the averaging theory until six order
to study the number of limit cycles which can bifurcate from the origin of coordinates of the
previous differential system.

Keywords: limit cycle, generalized Kukles differential system, averaging method, phase
portrait.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
We consider the so-called Kukles homogeneous differential system. Giné [3]

ẋ = −y, ẏ = x + Qn (x, y), (242)

who has a center at the origin, where Qn (x, y) denotes a homogeneous real polynomial of degree
n.
In 1999 Volokitin and Ivanov [4] prove that the systems (242) have a center at the origin definitely
if they are symmetric with respect to one of the coordinate axes. For n = 2 and n = 3, the authors
of the conjecture knew that it holds. Giné [3] in 2002 proved the conjecture for n = 4 and n = 5.
Giné et al. [3] proved the conjecture for all n under an additional assumption, that the authors
believe that it is redundant. The phase portraits for quadratic systems with center written in the
form (242), are known, see Vulpe [5]. The phase portraits of cubic differential systems symmetry
with respect to a straight line are also known and in particular those of system (242) with n = 3,
see Buzzi et al. [2]; Vulpe Sibirskii and Żoła̧dek [5]. The phase portraits of systems (242) with
n = 4 follows from Benterki and Llibre [?]. In Llibre and Silva classified the phase portraits of the
systems (242) for n = 5, 6. The phase portraits of systems (242) with n = 7 follows from Benterki
and Llibre [1].

262
CNMA’2022
Second National Conference on
Mathematics and its Applications

Figure 14: Global phase portraits of differential system (243).

In our work, we classify the global phase portraits of the polynomial differential systems

ẋ = −y,
(243)
ẏ = x + ax8 + bx4 y 4 + cy 8 .

The first main objective of this work is to study the phase portraits on the Poincaré disc of the
differential systems (243). The second objectif, is to study the number of limit cycles which can
bifurcate from the origin of coordinates of systems (243) when we perturb it inside all classes of
polynomial of eight degreewith, and we do this by using the averaging theory until six order.

Main results
Theorem 117. The set of all global phase portraits in the Poincaré disc of the differential system (243)
with a2 + b2 + c2 , 0 are topologically equivalent to the phase portraits given in Figure 14.

When we perturbed the polynomial differential system (243) with polynomials of degree eight,
we get

6
(s)
X X
ẋ = −y + εs αij xi y j ,
s=1 0≤i+j≤8
6 (244)
(s)
X X
s
ẏ = x + ax8 + bx4 y 4 + cy 8 + ε βij xi y j ,
s=1 0≤i+j≤8

where i, j ∈ N.
Our second result is given in the following Theorem.

Theorem 118. The number of limit cycles of the differential system (244) with ε , 0 is

(a) No limit cycles if we use the averaging theory of order 1 and 2,

(b) one limit cycle if we use the averaging theory of order 3 and 4,

(c) two limit cycles if we use the averaging theory of order 5 and 6.

263
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] R. Benterki and J. Llibre, The centers and their cyclicity for a class of polynomial differential
systems of degree 7 via averaging theory, J. Computational and Appl. Math. 368 (2020),
112456.

[2] C.A. Buzzi, J. Llibre, J.C. Medrado, Phase portraits of reversible linear differential systems
with cubic homogeneous polynomial nonlinearities having a non–degenerate center at the ori-
gin, Qual. Theory Dyn. Syst. 7 (2009), 369–403.

[3] J. Giné, Conditions for the existence of a center for the Kukles homogenenous systems, Comput.
Math. Appl. 43 (2002), 1261–1269.

[4] E.P. Volokitin and V.V. Ivanov, Isochronicity and Commutation of polynomial vector fields,
Siberian Mathematical Journal 40 (1999), 22–37.

[5] N.I. Vulpe and K.S. Sibirskii, Centro–affine invariant conditions for the existence of a center of
a differential system with cubic nonlinearities, (Russian) Dokl. Akad. Nauk SSSR 301 (1988),
1297–1301; translation in Soviet Math. Dokl. 38 (1989), 198–201.

264
CNMA’2022
Second National Conference on
Mathematics and its Applications

Synchronization between
Fractional-Order Lesser Date Moth Chaotic
System and Integer-Order Chaotic System
by Nonlinear Control

M. Labid 1 [Link]@[Link]
N. Hamri 2
1 Department of Mathematics, University Center of Mila
Mila 43000, Algeria
2 Laboratory of Mathematics and their interactions, Department of Science and Technology, University Center of
Mila, Mila 43000, Algeria

Abstract: This paper investigates the phenomenon of chaos synchronization between the
fractional-order lesser date moth and the integer-order chaotic system. Based on Lyapunov
stability theory and numerical differentiation, Nonlinear feedback control is the method used
to achieve the synchronization between fractional-order and integer-order chaotic systems.
Numerical examples are implemented to illustrate and validate the results.

Keywords: Chaos, Synchronization, Nonlinear control, Fractional-order chaotic system,

Integer-order chaotic system.
2010 Mathematics Subject Classification: 34H10, 37N35, 93C10, 93C15, 93C95.

Introduction
Chaos is a very interesting nonlinear phenomenon that has been intensively studied over the past
two decades. The chaos theory is found to be useful in many areas such as data encryption , fi-
nancial systems , biology and biomedical engineering [2], etc. Fractional-order chaotic dynamical
systems have begun to attract a lot of attention in recent years and can be seen as a generaliza-
tion of chaotic dynamic integer-order systems. The synchronization between the fractional-order
chaotic system and the integer-order chaotic system is thoroughly a new domain and began to at-
tract much attention in recent years because of its potential applications in secure communication
and cryptography . Obviously, the synchronization between fractional-order chaotic system and
integer-order chaotic system is more difficult than the synchronization between fractional-order
chaotic system or integer-order chaotic system for different order of their error dynamical sys-
tem. In this research work,we apply nonlinear control theory to synchronize two chaotic systems
when an fractional- order system is chosen as the drive system and a integer-order system serves
as the response system, we demonstrate the technique capability on the synchronization between
fractional-order lesser date moth chaotic system and integer-order chaotic system.

265
CNMA’2022
Second National Conference on
Mathematics and its Applications

Synchronisation of Fractional-Order Lesser Date Moth Chaotic System

and Integer-Order Chaotic System by nonlinear Control
In this section, to validate the nonlinear control method , we take the fractional- order lesser
date moth chaotic system as a drive system and the integer-order chaotic system as a response
system.

Thus, the drive and response systems are as follows:


x1 y1
D α x1 = x1 (1 − x1 ) − β+x



 1

γx1 y1

 α
 D y1 = −δy1 + β+x1 − y1 z1
 (245)


 D α z = −ηz + σ y z ,

 1 1 1 1

and 
x2 y2
ẋ2 = x2 (1 − x2 ) − β+x + u1 (t)



 2

γx2 y2

ẏ2 = −δy2 + β+x − y2 z2 + u2 (t) (246)



 2


 ż2 = −ηz2 + σ y2 z2 + u3 (t),


where u1 (t), u2 (t), u3 (t) are the nonlinear controls.

It is reported that the fractional-order lesser date moth system (245) with fractional order of
α = 0.95 can behave chaotically. The three-dimensional (3D) phase portraits of the lesser date
moth chaotic system with fractional-order and integer-order, respectively, are shown in Figure 15
and Figure 16 .

Figure 15: The 3D phase portrait of the fractional-order lesser date moth system

Figure 16: The 3D phase portrait of the integer-order lesser date moth system

Subtracting (246) from (245) gives the error system as below

266
CNMA’2022
Second National Conference on
Mathematics and its Applications


y x y2 x1 x1 y1

 ė1 = e1 − x12 + x22 − ( β+x2 )e1 + (β+x 1)(β+x e1 − β+x e2 + x1 − x12 − β+x − ẋ1 + u1 (t)
2)

 2 1 1 1

γy γx1 y2 x1 γx1 y1

ė2 = −δe2 + [ β+x2 − (β+x )(β+x ]e1 − (z2 − β+x )e2 − y1 e3 − δy1 + β+x − y1 z1 − ẏ1 + u2 (t) (247)

2)


 2 1 1 1


 ė3 = −ηe3 + σ (y1 e3 + z2 e2 ) − ηz1 + σ y1 z1 − ż1 + u3 (t),


where e1 = x2 − x1 , e2 = y2 − y1 , e3 = z2 − z1 .
We introduce a quadratic Lyapunov function
3
1X 2
V (e) = ei , (248)
2
i=1

and calculate the derivative of V (e) to obtain:

y2 x1 y 2 x1 x y
V̇ (e) = e1 [e1 − (x2 + x1 )e1 − ( )e1 + e1 − e2 + x1 − x12 − 1 1 − ẋ1 ]
β + x2 (β + x2 )(β + x1 ) β + x1 β + x1
γy2 γx1 y2 γx1 γx y
+ e2 [−δe2 + ( e1 − )e1 − (z2 e2 + )e2 − y1 e3 − δy1 + 1 1
β + x2 (β + x1 )(β + x2 ) β + x1 β + x1
− y1 z1 − ẏ1 ]
X3
+ e3 [−ηe3 + σ z2 e2 + σ y1 e3 − ηz1 + σ y1 z1 − ż1 ] + ui (t)ei (t). (249)
i=1

From the above equation, we deduce that if the active control functions ui (t) are chosen such that

y2 x1 y 2 x1 x y
u1 (t) = −[2e1 − (x1 + x2 )e1 − ( )e1 + e1 − e2 + x1 − x12 − 1 1 − ẋ1 ]
β + x2 (β + x1 )(β + x2 ) β + x1 β + x1
γy2 γx1 y2 γx1 γx1 y1
u2 (t) = −[( − )e − (z2 − )e − y e − δy1 + − y1 z1 − ẏ1 ]
β + x2 (β + x1 )(β + x2 ) 1 β + x1 2 1 3 β + x1
u3 (t) = −[σ z2 e2 + σ y1 e3 − ηz1 + σ y1 z1 − ż1 ],

equation (249) becomes

V̇ (e) = −(e12 + δe22 + ηe32 ) < 0. (250)

According to the inequality (248), the system (247) is asymptotically stable.

For the numerical simulations, we use some documented data for some parameters like γ = 3, δ =
η = 1, σ = 3, β = 1.15, h = 0.85, α = 0.95, then we have (x1 , y1 , z1 ) = (0.7, 0.3, 0.8) and (x2 , y2 , z2 ) =
(1.2, 0.12, 2.0). The simulation results are illustrated in Figure 17.

Figure 17: Synchronization between response system (246) and drive system (245)

267
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] A. E. Matouk, Chaos, feedback control and synchronization of a fractional-order modified au-
tonomous Van der Pol-Duffing circuit, Commun Nonlinear Sci. Numer. Simulat, 16 (2011)
975–986.

[2] B. Zsolt, Chaos theory and power spectrum analysis in computerized cardiotocography, Eur J
Obstet Gynecol Reprod Biol, 71(2) (1997) 163–168.

[3] D. Matignon , Stability result on fractional differential equations with applications to con-
trol processing, Computational Engineering in Systems and Application multi-conference,
IMACS, In: IEEE-SMC Proceedings, Lille, France, 2 (1996) 963–968.

268
CNMA’2022
Second National Conference on
Mathematics and its Applications

Chaos of a new one-dimensional

fractional map

Seyf El Islam Bouzeraa 1 [Link]@[Link]

Rabah Bououden 1
1 Abdelhafid Boussouf University Center, Mila, Algeria

Abstract: In this work, we propose a new one-dimensional fractional-order chaotic map based
rx(n)(1−x(n))
on the integer-order unified x(n + 1) = (1+x(n)) , x ∈ [0, 1]. The chaotic behavior of the new
fractional map is examinated using bifuraction diagrams and time series plots, all our resault
are validated by simulation in Matlab.

Keywords: One-dimensionl, bifurcation diagrams, fractional-order, chaotic map.

2010 Mathematics Subject Classification: Primary 34F10, 34H10, 39A70.

Introduction
The fractional systems have recently received considerable and growing attention from researchers,
because fractional calculus can accurately explain many more and a lot of real problems. Despite
the great interests and many results presented in the discrete fractional chaotic systems, identify-
ing chaos in the fractional chaotic map remains open topic.
In this work, we will propose and analyze a one-dimentional fractional chaotic map. In the next
section, we introduce the basic concepts of the discrete fractional calculus. In the last section,
we propose and analyze a one-dimentional fractional chaotic map, analyzed by bifurcations dia-
grams.

Preliminaries
Definition 48. [3] Let u : Na −→ R, and v > 0 by given. Then the fractional sum of v order is defined
by
t−v
1 X
∆−v
a u(t) = (t − σ (s))(v−1) u(s), t ∈ Na+v ,
Γ (v) s=a

where a is the starting point, σ (s) = s + 1, Na = {a, a + 1, a + 2....}, t (v) is the flling fractional function
defnied as
Γ (t + 1)
t (v) = .
Γ (t + 1 − v)
Definition 49. [1] For v > 0, v < N and u(t) defined on Na , the Caputo-like delta difference is defined

269
CNMA’2022
Second National Conference on
Mathematics and its Applications

by :
c −(m−v) m
∆va x(t) = ∆a ∆ x(t)
t−(m−v)
1 X
= (t − σ (s))(m−v−1) ∆m
s x(s),
Γ (m − v) s=a

where t ∈ Na+m−v , m = [v] + 1. , [v] means the integer order difference with starting point 0 and a,
respectively.

Theorem 119. [2] For the delta fractional difference equation.



 c ∆v u(t) = f (t + v − 1, u(t + v − 1))

 a

 ∆k u(a) = u , m = [v] + 1, k = 0, . . . , m − 1,


k

the equivalent discrete integral equation can be obtained as

t−v
1 X
u(t) = u0 (t) + (t − σ (s))(v−1) × f (s + v − 1, u(s + v − 1))
Γ (v) s=a+m−v
, t ∈ Na+m ,
where the initial iteration u0 (t) reads:
m−1
X (t − a)(k) k
u0 (t) = ∆ u(a)
k!
k=0

Chaos of new fractional map

We introduce a new one dimensional chaotic map:
rxn (1 − xn )
xn+1 = , (251)
(1 + xn )
Where r is known as control parameter, the map (251) is chaos behaviours for the values of r be-
tween 5.21 and 5.82 as in figure (18).

The fractional order of the map (251) is given as

C rx(t + v − 1)(1 − x(t + v − 1))

∆va x(t) = , (252)
(1 + x(t + v − 1))
From Theorem (119), we can obtain equivalent discrete integral of Eq (252)
t−v
v X Γ (t − s + v) rx(s + v − 1)(1 − x(s + v − 1))
x(t) = x(a) + ( − x(s + v − 1)).
Γ (v) Γ (t − s + 1) 1 + x(s + v − 1)
s=1−v

For a = 0 and j = s + v, the numerical solution of map (252) is given as

n
v X Γ (n − j + v) rx(j − 1)(1 − x(j − 1))
xn = x0 + ( − x(j − 1)). (253)
Γ (v) Γ (n − j + 1) 1 + x(j − 1)
j=1

270
CNMA’2022
Second National Conference on
Mathematics and its Applications

Figure 18: Bifurcation diagram of the map for v = 1.

Figure 19: Bifurcation diagrams of fractional map (251) for v = 0.9 and v = 0.2.

Using the numerical formula (253), assume x0 = 0.1, v = 0.9 and v = 0.2, we obtain the bifuraction
diagrams.

References
[1] T. Abdeljawad, On Riemann and Caputo fractional differences, Comput. Math. Appl. 62
(2011), 1602–1611.

[2] F.L. Chen, X.N. Luo and Y. Zhou, Existence results for nonlinear fractional difference equa-
tion, Adv. Differ. Equ. 70 (2011), 1–12.

[3] C. Guo and B. Dumitru, Discrete fractional logistic map and its chaos, Nonlinear Dyn. 75
(2014), 283–287.

[4] G.C. Wu and D. Baleanu, Discrete chaos in fractional delayed logistic maps, Nonlinear Dyn.
75 (2014), 1697–1703.

271
CNMA’2022
Second National Conference on
Mathematics and its Applications

Dynamics study of the modified

Beverton-Holt model

Chabane Bedjguelel 1 [Link]@[Link]

Hacene Gharout 1 [Link]@[Link]
1 Department of mathematics , University of Bejaia

Abstract: In this paper, we investigate a new four-parameter model that generalizes the
Beverton-Holt demographic model. The conditions for the existence and stability of fixed
points were determined. The Allee effect and essential extinction regions are noted. We use
numerical simulations to explain theoretical results.

Keywords: Stability , fixed point , allee effect, Beverton-Holt model..

2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
In recent years, increasing attention has been given to biological models. Many researchers study-
ing the field of dynamical systems have focused on population models. Population models in
dimension one, are the simplest models to understand since they are modeled by real-valued
functions, for example, the Ricker model, a model used to model the fish population was first
modeled by a monotonic and unimodal function, and also the Beverton-Holt model [1]. One of
the important phenomena that has attracted the attention of many researchers and that can be
observed in these models is the Allee phenomenon [3, 4, 5], a biological phenomenon character-
ized by a positive correlation between a population density and its per capita growth rate at low
densities, which can lead to critical population thresholds below which population extinction oc-
curs [2]. In general, a function expressing the Allee effect in the context of dynamical systems
or difference equations must contain three fixed points. The trivial fixed point that is asymp-
totically stable, a strictly positive unstable fixed point, called the threshold point, and a bigger
asymptotically stable fixed point. The model that will be the subject of our work is:
βxα
xn+1 = f (xn ), with f (x) = , (254)
1 + δxm
where β, α, δ, m are positive real parameters. We will limit our study in the space of the parame-
ters: n o
R0 = (β, α, δ, m) ∈ R4 : β, α, δ, m > 0 and m > α .
Main results
  1−α
m α−1
Proposition 33. Let α > 1, f the function defined by Equation??, and β ∗ = m−α+1 δ(m+1−α)
m
. Then,

1. If β = β ∗ , then f admits exactly two fixed points, the trivial fixed point x = 0 and x1 which is 0 is
stable with basin of attraction [0, x1 [,and it is verified that
lim f n (x) = 0, ∀x ∈ R+ \[x1 , x1∗ ], and lim f n (x) = x1 , ∀x ∈ [x1 , x1∗ ].
n→+∞ n→+∞

272
CNMA’2022
Second National Conference on
Mathematics and its Applications

2. If β > β ∗ , then admits three fixed point, the trivial fixed point and x2 are asymptotically stable,
and x1 is unstable. Moreover, the basin of attraction of x2 is ]x1 ; ∞)[, and the basin of attraction
of 0 is ]0; x1 [.

3. If β < β ∗ , then f admit x = 0 a single fixed point which represents a global attractor, i.e., lim f n (x) =
n→+∞
0, ∀x ∈ R+ .

Definition 50. [5] Let f : R+ −→ R+ the function defined by Equation 254 with Allee effect. We define
essential extinction region by:
 ! 1−α 
 m α − 1 m 
(β, α, δ, m) ∈ R0 : f 2 (c) < x1 , α > 1, β >
 
Ress =  ,
 
m − α + 1 δ(m + 1 − α)
 

 

where x1 is the threshold fixed point.

Proposition 34. Let f : R+ −→ R+ the function defined by Equation 254 with Allee effect and, consider
the values

a = min{f −1 (x1∗ )}, b = max{f −1 (x1∗ )}.

such that I =]a, b[⊂]x1 , x1∗ [. If (r, α, δ, m) ∈ Ress , then

lim f n (x) = 0, ∀x ∈ I,
x→+∞

where x1∗ = max{f −1 (x1 )}.

In this region, we have f admits at least three fixed points and this indicates that almost all initial
densities will lead to extinction see Figure 20.

Figure 20: The graph of f when f 2 (c) < x1 show that for any initial point ready in the interval
]a, b[ its orbit approaches the fixed point zero.

References
[1] M. Bohner, S. Streipert, Optimal harvesting policy for the Beverton–Holt model, Mathematical
Biosciences and Engineering, 13:4(2016), 673-695.

[2] F. Courchamp, L. Berec , J. Gascoigne, Allee Effects in Ecology and Conversation, Oxford Uni-
versity ,2008.

273
CNMA’2022
Second National Conference on
Mathematics and its Applications

[3] S. Elaydi, E. Kwessi, G. Livadiotis, Hierarchical competition models with the Allee effect
III:Multispecies, J. Biol. Dyn. 12 (2018), 271-287.

[4] [Link], A new flexible discrete-time model for stable populations, Discrete Contin. Dyn. Syst. 23
(2018), 2487-2498.

[5] J.L. Rocha, A-K Taha, Bifurcation analysis of the γ -Ricker populationmodel using the Lambert
W function, Int.J . Bifurc. Chaos, 30:7(2020), 2050108.

274
CNMA’2022
Second National Conference on
Mathematics and its Applications

Global dynamics of an age-structured

model

Fatima Zohra HATHOUT 1 [Link]@[Link]

Tarik Mohammed TOUAOULA 1 touaoula_tarik@[Link]
1 Université de Tlemcen

Abstract: In this research, we consider the influence of protection measures on the spread
of infectious diseases in an age-structured population. Protection strategy can take different
forms as isolation, treatment, or renewable vaccine; to mathematically represent it, we include
a new compartment p standing for protected individuals, in a classical age structured si model.
Global analysis of the proposed model is made by the introduction of total trajectory and a
suitable Lyapunov functional.
We give a particular importance to the protection strategy and many numerical simulations
are provided to illustrate our theoretical results.

275
CNMA’2022
Second National Conference on
Mathematics and its Applications

Stability analysis of a fractional-order

ecosystem with predator harvesting

Kerioui Nadjah1 [Link]@[Link]

1 Department of mathematics , University ,
2 Department of mathematics , Center University Mila

Abstract: The objective of the present work is to investigate the dynamics of a fractional-order
differential-algebraic predator-prey system with Holling type III functional response. This
model can be established as follow:

q1 x = x r 1 − x − axy
    
D ,


K  d + x2




bx2
!
y

 
 q
 D 2y = y s 1 − + − E ,
N d + x2





0 = E(py − c) − v,




The boundedness and positivity of solutions for this model are derived. Local stability of
the ecosystem near the coexistence equilibria have been thoroughly investigated when the
economic profit v varies in both commensurate and incommensurate fractional orders. The
influence of the commensurate fractional orders on the existence of the Hoph bifurcation for
the fractional-order ecosystem is explored. Finally, numerical illustrations are performed in
order to validate some of the important analytical findings.

Keywords: Differential-algebraic system, Fractional order, Stability, Hopf bifurcation,

Harvesting..
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
Fractional order differential equations as a generation of the classical integer order differential
equations have played a very important role in describing various phenomena in nature. Due to
the non-local property of the fractional derivative, fractional order differential equations are more
appropriate than integer order ones in biological, economic, and social systems where memory ef-
fects are important.
In this work, we will study the following differential-algebraic predator-prey model with fractional-
order:
x axy
    

 q
D x = x r 1−
1 − ,
K  d + x2



 2
!
y bx

 
 q
 D y = y s 1− N +
 2 −E ,


 d + x2

0 = E(py − c) − v,




such that q1 , q2 ∈ (0, 1) are the fractional order derivatives in the sense of Caputo. Hence,

276
CNMA’2022
Second National Conference on
Mathematics and its Applications

• 0 ≤ E(t) ≤ Emax the harvest effort exerted on the predators.

• r, s > 0 the intrinsic growth rates of prey and predator populations,

• K, N > 0 the carrying capacity of the two species,

• d > 0 half capturing saturation constant,

• a > 0 maximal efficiency of predation,

• b > 0 conversion coefficient.

• p > 0 unit price of the harvested population.

• c > 0 cost of the harvest effort.

• v > 0 economic profit.

We aim to find the coexistence equilibrium points and to investigate the behaviours of the ecosys-
tem around them by means of qualitative theory of dynamical systems (such as local stability and
Hopf bifurcation).

Main results

• The present work deals with a differential-algebraic biological economic system with fractional-
order. We have taken predator functional response to prey in a form that approaches to a
constant even when the prey population increases. We consider the dynamical behaviour of
the system when only the predator is subjected to harvesting. From the biological perspec-
tive, we only interest on the positive equilibrium points.

• The local stability of the interior equilibria is curried out by analysing their corresponding
characteristic equation and the proposed numerical example shows that the system has two
interior equilibria (see Figure 21) one them is unstable saddle point and the other one is
focus point that changes its stability property when varying the economic revenue v and
the fractional orders q1 and q2 . Moreover, one parameter bifurcation analysis is done with
respect to the commensurate fractional-order q (see Figure 22). It has been assumed that the
positive economic revenue v and commensurate fractional-order q are responsible for the
stability of the proposed model. The stability analysis has revealed that when the economic
profit v is less than a bifurcation value v ∗ both species converge to their steady states and
they will coexist over the time (see Figure 23). Moreover, it is shown that when the economic
profit is v ∗ < v < 1.4147, then the stability is depend on the fractional-order q ∈ (0, 1) (see
Figure 24).

• The proposed study allows us to point out that it is important for the government to adjust
revenue and draw up beneficial strategies to support to encourage or improve fishery or
mitigate emissions, so that the community can be driven to steady states that will lead to
the survival and sustainable growth of the prey-predator ecosystem.

277
CNMA’2022
Second National Conference on
Mathematics and its Applications

Figure 21: Number of interior equilibrium points of the system (??) with respect to the economic
profit v, for 0 < v ≤ 5.

Figure 22: Time evolution of x(t) and y(t) with v = 1.

References
[1] I. Podlubny, Fractional Differential Equations, Academic Press, New York, 1999.

[2] N. Kerioui and M. S. Abdelouahab, Stability and Hopf bifurcation of the coexistence equi-
librium for a differential-algebraic biological economic system with predator harvesting, Elec-
tronic Research Archive. 29(1) (2021), 1641–1660.

[3] H. A. A. El-Saka, L. Seyeon, J. Bongsoo, Dynamical analysis of fractional-order predator-

prey biological economic system with Holling type II functional response, Nonlinear Dynam-
ics. 1(96) (2019), 407–416.

[4] M. S. Abdelouahab, N. E. Hamri, J. Wang, Hopf bifurcation and chaos in fractional-order

modified hybrid optical system, Nonlinear Dynamics. (69)) (2012), 275–284.

278
CNMA’2022
Second National Conference on
Mathematics and its Applications

Figure 23: Time evolution of x(t) and y(t) with v = 0.941 and x(0) = 0.872, y(0) = 0.6777.

Figure 24: Time evolution of x(t), y(t) and E(t) with v = 1.251 and q = 0.85.

279
Operational research and
optimization

280
 CNMA’2022
Second National Conference on
Mathematics and its Applications

Interior Point Method for Quadratic

Programming
-Inverse Barrier Method-

Boutheina Fellahi 1 [Link]@[Link]

Bachir Merikhi 1
1 Department of mathematics , University Ferhat Abbas

Abstract: In this paper, we are interested to solve a quadratic optimization programming

problem using an inverse barrier method in which the inverse barrier function used is given
P  −1
by ni=1 xir where r ∈ R∗+ . The descent direction has been calculated using a descent Newton
method, however the step size has been calculated using a tangent technique. Our study is
supported by a numerical tests very encouraged.

Keywords: Quadratic programming, interior point methods, inverse function, tangent

technique.
2010 Mathematics Subject Classification: Primary 90C20, 90C25, 90C51.

Introduction
Lots of previous studies solve the mathematical programming using the interior point methods
give much attention to the logarithmic barrier methods [1] which proposed by Frish [3] and de-
veloped by Fiacco and McCormick [4].
In our study, we thought to use the inverse barrier function in [2] to solving a constrained convex
quadratic programming which had the same properties as the logarithmic barrier function. The
P  −1
non negativity constrains xi ≥ 0 are replaced by a penalty term r ni=1 xir , in which the barrier
parameter r tends to zero.
The quadratic problem to be studied in this work is given by:


min f (x)



(P)


x∈S




Under the following assumptions:

Description 6.

a- f is a quadratic, convex, twice continuously differentiable function on X.

In which: f (x) = 21 xt Qx + bt x + c,

280
 CNMA’2022
Second National Conference on
Mathematics and its Applications

And S = {x ∈ Rn : Ax = b, x ≥ 0} is the set of feasible solution of (P ),

S0 = {x ∈ Rn : Ax = b, x > 0} is the set of strictly feasible solution of (P ).

b- A ∈ Rm×n is a full rank matrix, b ∈ Rm , (m < n).

c- There exists x0 > 0 such that Ax0 = b.

d- The set of optimal solutions of (P ) is nonempty and bounded.

The penalization
In this section, we associate the following perturbed problem (Pr ) to the original problem (P ):


min fr (x)



(Pr )


x ∈ Rn




In which  P  −1
f (x) + r ni=1 xir , if Ax = b, x ≥ 0,





fr (x) = 




∞
 if not

The first and the second order derivatives of fr are given by

∇fr (x) = Qx + b − r 2 X −r−1 e

∇2 fr (x) = Q + r 2 (r + 1)X −r−2

In which X is the diagonal matrix with diagonal entries Xii = xi , ∀i = 1, n, and e = (1, 1, ..., 1)t ∈ Rn .
fr is a strictly convex, proper and lower semi continuous function since fr is differentiable on S0 .
Our idea is to develop a new approach, which consist to solve the perturbed unconstrained prob-
lem (Pr ) instead of solving the original problem (P ).
The resolution of (Pr ) is based on the classical Newton method to calculate the descent direction
and a tangent technique to calculate the optimal step size.

Solving the perturbed problem

In this section, we are interested in the numerical solution of the perturbed problem (Pr ).

The Newton descent direction

A descent direction d can be computed by different methods, in this work we choose the New-
ton’s method, therefore d is given by solving the following equation

∇2 fr (x)d = −∇fr (x)

The step size

281
CNMA’2022
Second National Conference on
Mathematics and its Applications

Generally, the most used methods in the line search are the classical iterative methods as Armijo-
Goldstein, Wolfe, Fibonnaci,..., but the computational cost in there becomes high when n is very
large. The line search function is given by:

θ(α) = fr (x + αd)

Pn 1
= f (x + αd) + r i=1 (xi +αdi )r

We have
0 Pn di
θ (α) = h∇f (x + αd), di − r 2 i=1 (xi +αdi )r+1

00 Pn di2
θ (α) = h∇2 f (x + αd), di + r 2 (r + 1) i=1 (xi +αdi )r+2

According to the complex form in which the function θ is defined, it is difficult to solve the equa-
0
tion θ (α) = 0 explicitly to obtain the optimal value α ∗ . We are interested to avoid this difficulty
using the tangent technique described bellow which is simple and more effective than the classical
methods.

Description of the tangent technique

This technique is based on two phases:

1. Finds an interval Ik =]αk , αk+1 [ such that its bounds satisfy the condition:

θ 0 (αk ).θ 0 (αk+1 ) < 0

Which means that the optimal step size α ∗ ∈ Ik .

2. Computes the optimal step size, in this phase we determine the tangents Tk and Tk+1 in
the bounds of the interval and we select the value corresponding to the intersection of the
tangents Tk and Tk+1 .

References
[1] [Link], [Link], A logarithm barrier method for semidefinite programming, RAIRO
Oper Res. 42 (2008), 123-139.

[2] [Link], [Link] and [Link] , Invrse barrier methods for linear programming, Oper Res.
28(2), 135-163, (1994).

[3] A. V. Fiacco, G. P. Mccormick, Nonlinear Programming : Sequential Unconstrained Mini-

mization Techniques, Wiley, New York, 1968.

[4] K. R. Frisch, The logarithmic potential method of convex programming, Technical report, Uni-
versity Institute of Economics, Oslo, Norway, 1955.

282
CNMA’2022
Second National Conference on
Mathematics and its Applications

PRIMAL-DUAL PATH-FOLLOWING
ALGORITHMS FOR SEMIDEFINITE
PROGRAMMING

Chafia DAILI 1 ch_dail@[Link]

Mohamed ACHACHE 1
1 Department of mathematics , University of Setif 1 El Baz

Abstract: In this paper, we are interested in the development of a realizable primal-dual

interior point algorithm with full Newton step, to solve semiedefinite programming problems.
To determine the directions of displacement, we use the Nesterov-Todd symmetrization.
Convergence and complexity analysis are given, followed by numerical results from the
application of this algorithm on some semidefinite problems of different sizes. We will end
this paper with a conclusion.

Keywords: semidefinite programming, interior-point methods, polynomial complexity,

pathfollowing methods, primal-dual algorithms
2010 Mathematics Subject Classification: Primary 65K05, 90C25, 90C30

Introduction
In recent years, semidefinite programming has become one of the most studied problems thanks
to their applications in different fields such as linear programming, convex and non-convex quadratic
programming, combinatorial optimization, approximation theory, system and control theory, me-
chanics, engineering...etc; and that these problems can be solved in polynomial time by interior-
point algorithms.
We define a semidefinite problem with linear constraints in primal form by the following opti-
mization problem :



 min C • X


(P ) Ai • X = bi , i = 1, ..., m





 X ∈ S n.

+

b ∈ Rm and the matrices C; Ai ∈ S n ; i = 1, ..., m.

the dual of (P ) is given by :



 max bT y


(D) C− m
P
i=1 yi Ai  0, i = 1, ..., m





 y ∈ Rm .


283
CNMA’2022
Second National Conference on
Mathematics and its Applications

where the unknown is (y, Z) ∈ Rm × S+n .

We have developed a realizable short-step primal-dual interior point algorithm for solving the
semidéfinite programming problems, at each iteration to calculate Newton’s direction we used
Nesterov and Todd symmetrization.

Main results
Through the examples we have used, we can notice that:
- decreasing the value of the parameter µ reduces the time and the number of iterations.
- the results obtained using the practical choice of θ are better than the theoretical choice.

References
[1] M. Achache, L. Guerra, A full Nesterov-Todd-step feasible primal-dual interior point algo-
rithm for convex quadratic semi-definite optimization, Appl. Math. Comput. 231 (2014),581–
590.

[2] E. De Kelerk, Aspects of semidefinite Programming, Applied Optimization, Vol.65, Kluwer

Acadimic, Dordrecht, 2002.

[3] M. Halicka, E. De Klerk and C. Roos, On the convergence of the central path in semidefinite
optimization, SIAM J. Optim. 12(4) (2002)1090–1099.

[4] J. Peng, C. Roos, T. Terlaky, Self-regularity functions and new search directions for linear
and semidefinite optimization, Math. Prog. 93 (2002), 129-171.

284
CNMA’2022
Second National Conference on
Mathematics and its Applications

Comparative Study on Multiple Chaotic

Maps Incorporated Global Optimization
Algorithms

Hamaizia Tayeb 1 h2tayeb@[Link]

Nacer Meriem 1
1 Department of mathematics , University Constantine 1

Abstract: In this paper we presented the performances of a new class of evolutionary algo-
rithms called chaotic optimization algorithm (COA). Proposed to solve nonlinear optimization
problems with bounded variables by Caponetto et al. Chaotic optimization is a new stochastic
optimization algorithm, which directly uses chaotic variables to find the optimal solution.
Different chaotic maps have been considered, combined with several working strategies.
We propose five different 2D chaotic maps in an optimization algorithm using a two-step
chaotic optimization method and compare them. This study reviews and compares chaotic
optimization algorithms from the literature. Moreover, the two-phase strategy is a commonly
used technique in a COA to refine the solution and help escape local optima. A performance
study is conducted to understand their impact on a chaotic optimization algorithm. .

Keywords: Chaotic optimization algorithm- Dynamic system- Chaos- Optimisation.

2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

References
[1] M. Bucolo, R. Caponetto, L. Fortuna, M. Frasca, A. Rizzo, Does chaos work better than
noise? IEEE Circuits and Systems , Magazine 2 (3) (2002) 4-19.

[2] T. Hamaizia, R. Lozi, An improved chaotic optimization algorithm using a new global locally
averaged strategy, Journal of Nonlinear Sys- tems and Applications 3 (2) (2012) 5863.

285
CNMA’2022
Second National Conference on
Mathematics and its Applications

An adaptive version of the conjugate

gradient parameter βkCD using the Newton
direction for unconstrained optimization

Naima HAMEL 1 hamelnaima24@[Link]

Noureddine BENRABIA 21 noureddinebenrabia@[Link]
Mourad GHIAT 1 mourad.ghi24@[Link]
Hamza GUEBBAI 1 guebaihamza@[Link]
1 Laboratoire de Mathématiques Appliquées et de Modélisation, Université 8 Mai 1945 Guelma, Algeria
2 Département de Mathématiques et Informatique, Université Mohamed-Chérif Messaadia, Algeria

Abstract: In this paper, we present a new nonlinear conjugate gradient method as a modifica-
tion of the Conjugate Descent (CD) method for solving unconstraind optimization. The search
direction of the proposed method accords with the Newton direction, the descent property
and global convergence are achieved under the strong Wolfe line search conditions. The
numerical tests show the efficiency of the proposed method, as it is better than the Conjugate
Descent (CD) method for some problems.

Keywords: unconstraind optimization, nonlinear conjugate gradient method, Newton

method, global convergence, numerical experiments.
2010 Mathematics Subject Classification: 90C26, 49M15, 90C30, 65K05

Introduction
Consider the following unconstrained optimization problem

minn f (x). (255)

x∈R

Where f : Rn 7→ R is a smooth nonlinear function.

There are many methods to solve the problem (255) which include the Steepest Descent method
[1]; the Newton method [2]; conjugate gradient methods [3] and Quasi-Newton methods [4]. all
these methods use the following iterative scheme

xk+1 = xk + αk dk ,

where, αk > 0 is a steplength usually computed by the strong Wolfe line search given as

f (xk + αk dk ) ≤ f (xk ) + δαk gkT dk

|g(xk + αk dk )T dk | ≤ −σ gkT dk

with: 0 < δ < σ < 12 .

And dk is a search direction.

286
CNMA’2022
Second National Conference on
Mathematics and its Applications

The nonlinear conjugate gradient methods are the most used ones for solving large-scale nonlinear
optimization due to the simplicity of their iterations, the search direction is defined as:

d0 = −g0 , dk+1 = −gk+1 + βk dk .

The different choices for the parameter βk correspond to different CG methods, for example the
Conjugate Descent (CD) method which given by:

k gk+1 k2
βkCD = .
−gkT dk

The Newton method is known by the quadratical convergence property, where the search direc-
tion is defined as
dk+1 = −∇2 f (xk+1 )−1 gk+1 ,
where, ∇2 f (xk+1 ) is the Hessian matrix.

Main results
In this paper, we propose a new conjugate gradient method by making some modifications to
the Conjugate Descent (CD) method so that the search direction accords with Newton’s direction:

k gk+1 k2
−skT ∇2 f (xk+1 )gk+1 − skT ∇2 f (xk+1 ) dk = −skT gk+1 ,
gkT dk

using a suitable condition and after some algebraic calculation, we get get an adaptative version
of βkCD without needing to compute or save the Hessian matrix.
The global convergence is achieved.
The Numerical test shows that our algorithm is more efficient and practical than the Conjugate
Descent (CD) algorithm.

References
[1] W. Hao, A gradient descent method for solving a system of nonlinear equations, Applied Math-
ematics Letters, 112, pp 106739 (2021).

[2] B.T. Polyak, Newton’s method and its use in optimization, European Journal of Operational
Research. 181 (3) (2007), 1086–1096.

[3] R. Dehghani, N. Bidabadi, H. Fahs and M.M. Hosseini, A Conjugate Gradient Method
Based on a Modified Secant Relation for Unconstrained Optimization, Numerical Functional
Analysis and Optimization . 4 (5), (2020), 621–634.

[4] B. A. Hassan, A new type of quasi-Newton updating formulas based on the new quasi-Newton
equation, Numerical Algebra, Control, Optimization. 10 (2), (2020), 227–235.

287
CNMA’2022
Second National Conference on
Mathematics and its Applications

An interior-point algorithm for linear

programming based on modified
displacement step

Linda Menniche 1 l_menniche@[Link]\

jamel Benterki2 dj_benterki@[Link]
1 LMPA, Department of mathematics , Mohammed seddik Ben Yahia University,Jijel
2 LMFN, Department of mathematics , Ferhat Abbas University, Setif-1

Abstract: In this paper, we present an interior-point method for solving a linear programming
problem. Newton’s method is used to compute the descent direction, and approximate func-
tion is used as an efficient alternative to line search methods to determine the displacement
step along the direction in order to reduce the computation cost. Finally, we present some
numerical simulations which show the effectiveness of the algorithm developed in this paper.

Keywords: Linear programming; interior-point methods, logarithmic barrier methods.

2010 Mathematics Subject Classification: 90C22, 90C51.

Introduction
In this paper, we present an algorithm for solving the optimization problem




 min bt y


 t
(D) 
 A y≥c


 y ∈ Rm ,


where A ∈ Rm×n is a matrix of full rank (rank(A) = m < n).

c and b are a vectors of Rn and Rm respectively.
The dual problem associated to (D) is defined as follows:



 max ct x


(P )  Ax = b




 x ∈ Rn x ≥ 0.


Solving (D) is equivalent to solving (P ): the optimal solutions of one problem being easily ob-
tained when the other problem is known. In this paper, the problem (D) is approximated by the
problem (Dr ), (r > 0).

 min fr (y)



(Dr ) 
 y ∈ Rm .



288
CNMA’2022
Second National Conference on
Mathematics and its Applications

Where r > 0 is the parameter barrier and fr is a barrier function defined by

n


 b t y + nr ln r − r P ln < e , At y − c > if At y − c > 0,

 i
fr (y) = 

i=1

 +∞ otherwise.



Where (e1 , e2 , ..., en ) is the canonical base in Rn . This problem is solved via a classical Newton
descent method. Firstly, we give the existence and uniqueness of the optimal solution of the
corresponding perturbed problem (Dr ) and verify its convergence to the optimal solution of the
original problem (D) when the barrier parameter approaches zero. Concerning the computation
of displacement step, especially while using line search methods is expensive and even more
delicate in particular in semidefinite programming problems [2]. Here, instead of minimizing fr
along the descent direction d at the current point y, we minimize the function θ(t) defined by
1
θ(t) = (fr (y + td) − fr (y)).
r
The minimization of the function θ(t) is difficult, we approximate this function by a new majorant
function θ(t)
e checks
θ(0) = θ(0),
e θ 0 (0) = θ
e0 (0) < 0.

This function θe needs to be appropriately chosen so that the optimal t is easily obtained and to
be close enough to θ in order to give a significant decrease of fr in the iteration step. We propose
in this paper a new function θe for which the optimal solution t is explicitly obtained and a good
quality approximation of θ by θe is ensured by the condition θ 00 (0) = θ
e00 (0).

Conclusion
In this study, we have presented an interior point method for solving the linear programming.
We have given the existence and uniqueness of the optimal solution of the corresponding per-
turbed problem and have verified its convergence to the optimal solution of the original problem
when the barrier parameter approaches zero. Newton’s method has been applied to find a new
iterative point by calculating a sufficient descent direction. Due to the high computational cost,
we have avoided using several methods, such as the line search methods, to calculate the displace-
ment step. Alternatively, a new approach based on approximate functions has been proposed to
accomplish this task to the optimal solution. The approximate functions technique is very reli-
able alternative that will be confirmed as the technique of choice for both (P L) and other classes
of optimization problems.

References
[1] J.P. Crouzeix, A. Seeger, New bounds for the extreme values of a finite sample of real numbers,
Journal of Mathematical Analysis and Applications, 197 (2008), 411–426.
[2] J.P. Crouzeix, B. Merikhi, A logarithm barrier method for semidefinite programming, RAIRO-
Operations Research. 42 (2008), 123–139.
[3] N. Karmarkar, , A new polynamial-time algorithm in linear programming, Combinatorica, 4
(1984), 373–395.
[4] A. Keraghel, D. Benterki, Sur les performances de l’algorithme de Karmarkar pour la program-
mation linéaire, Revue Roumaine des sciences techniques mécaniques appliquées, 4- (2001),
87–96.

289
CNMA’2022
Second National Conference on
Mathematics and its Applications

[5] M. Kojima, N. Megiddo, S. Mizuno, , A primal-dual infeasible interior point method for linear
programming, Math, 61 (1993), 263–280.

[6] L. Menniche, D. Benterki, A logarithmic barrier approach for linear programming, Journal of
computational and Applied Mathematics, 312 (2017), 267–275.

290
CNMA’2022
Second National Conference on
Mathematics and its Applications

Primal-dual feasible interior-point

algorithm for monotone linear
complementarity problems

Randa Chalekh 1 [Link]@[Link]

EL Amir Djeffal 1 [Link]@[Link]
1 Department of mathematics , University of Batna 2

Abstract: This paper is devoted to generalize the feasible interior-point algorithm for solving
monotone horizontal linear complementarity problems to solving monotone linear com-
plementarity problems. The algorithm has the advantage that no line-searches are needed,
since it uses only full-Newton
√ steps. Moreover, we prove that the complexity bound of the
proposed algorithm is O( n log nε ), which coincides with the well-known best iteration bound
for monotone LCP.

Keywords: Linear complementarity problems, interior-point methods, full-Newton step,

complexity.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
We consider the monotone linear complementarity problem (LCP) as follows :

s = Mx + q, xT s = 0, x ≥ 0, s ≥ 0. (256)

where M ∈ Rn×n is a real matrix supposed positive semidefinite, q ∈ Rn . The linear complemen-
tarity problems consists in finding the two vectors x and s in a real space Rn which satisfy (256).
There exist a differents types of LCP as the monotone horizontal linear complementarity problem
(HLCP) N s = Mx + q with N ∈ Rn×n . The HLCP reduced to the monotone LCP (256) if the matrix
N = I or N is nonsingular.
For solving monotone linear complementarity problems we use primal-dual path-following meth-
ods which is the most widely used method for solving optimization problems. This is because its
effectiveness has been proven among all interior-point methods introduced by Karmarker. The
most important property that
√ make this method the best because it has the best current polyno-
n
mial complexity namely, O( n log ε ).
Our goal is to present a full-Newton step feasible interior-point algorithm for monotone LCP.
The idea of this approach is to replace HLCP with LCP and we study a number of properties by
following the same schema introduced by Achache et al and Mansouri et al in [1, 3], respectively.
The algorithm is follow the centers of the perturbed LCP by using only full-Newton steps with
the advantage that no line search.

291
CNMA’2022
Second National Conference on
Mathematics and its Applications

This paper is organized as follows. First, we briefly recall the notion of the central path, which is
a basic concept of primal-dual IPMs. Then, in section 2.1 we deal with the new search directions.
Finally, in section 2.3 we state its polynomial complexity by investigate the feasibility of a full-
Newton step and we mainly prove that the iterate is locally quadratically convergent.

Main results
In this section, first we study the central-path of LCP and the search directions. Then, we will
show that the algorithm solves the LCP in polynomial complexity.
In this paper, we assume that the following assumptions hold :
1. Interior-point-condition (IPC). 2. The matrix M is a positive semidefinite matrix.
Hence, finding the solution of (256) is equivalent to solving the following system :

s = Mx + q, xs = 0, x ≥ 0, s ≥ 0. (257)

The basic idea of primal–dual IPMs is to replace the second equation in (257) by the parameterized
equation xs = µe with µ > 0, we find the following system :

s = Mx + q, xs = µe. (258)

Under the two assumptions, it is known that (258) has unique solution [2], denoted as (x(µ), s(µ))
for each µ > 0. We call (x(µ), s(µ)) the µ-centers of (258). The set of the µ-centers central path of
(256).

New search directions

To find a strictly feasible point taking a full-Newton step ((x+ , y+ ) = (x + ∆x, s + ∆s)) we need to
define search directions (∆x, ∆s). This end hold by applying Newton’s method for (258), we get
the following linear system :

∆s = M∆x, s∆x + x∆s = µe − xs. (259)

Under our assumptions,

 the system (259) has a unique solution. This is due to the fact that the
 −M I 
following matrix   is invertible (see [3]), where X = diag(x) and S = diag(s). By using
 
 S X 

r
xs v∆x v∆s
v= , dx = and ds =
µ x s

we can easily checks that the system (259), which defines the search directions, is rewritten as
follows
ds = Mdx , dx + ds = v −1 − v. (260)
where M = Mxs−1 . For the analysis of the algorithm, we define the following norm-based prox-
imity measure which uses to measure closeness of the iterates to the µ-center as follows
1 1
δ(v) := δ(x, s; µ) = kdx + ds k = kv −1 − vk. (261)
2 2

Complexity analysis of the algorithm

292
CNMA’2022
Second National Conference on
Mathematics and its Applications

Feasibility and quadratic convergence of the feasible full-Newton step

In this part, we guarantees feasibility of a full-Newton step then we prove that the full-Newton
steps are quadratically convergent.
First, using (259) and (260) we obtain
x+ s+
v+2 = = e + dx ds =⇔ x+ s+ = µ(e + dx ds ). (262)
µ

Lemma 46. (Lemma 2, [1]) The full-Newton step is positive if and only if e + dx dy > 0.

Proof The proof is similar to the proof of Lemma 2 in [1]. 

Lemma 47. (Lemma 1, [1]) Let δ > 0 and (dx , ds ) be a solution of system (260). Then, we have
√
0 ≤ dxT ds ≤ δ2 , kdx ds k∞ ≤ δ2 and kdx ds k ≤ 2δ2 . (263)

Proof The proof is similar to the proof of Lemma 1 in [1]. 

Lemma 48. If δ < 1, then the positive pair (x+ , s+ ) is strictly feasible for (256) and satisfied

δ2
δ+ := δ(x+ , s+ ; µ) ≤ p . (264)
2(1 − δ2 )

Proof The proof of the first part is similar to the proof of Lemma 3 in [1] and for the last statement
is similar to the proof of Lemma 4 in [1]. 

Corollary 14. If δ ≤ √2 , then δ+ ≤ δ2 . This imply the quadratic convergence of the full-Newton step
10
to the µ-center.

Updating the barrier parameter

In this part, we establish an upper bound for the duality gap after a full-Newton step and the
relation between the proximity before and after the update of the barrier parameter µ.

Lemma 49. If δ ≤ √2 , then the upper bound for the duality gap after a full-Newton step satisfies
10

3n
x+T s+ ≤ µ.
2

Proof The proof is similar to the proof of Lemma 5 in [1]. 

Lemma 50. Let δ ≤ √2 . From (262) and µ+ = (1 − θ)µ, then

10

2(1 − θ) 5nθ 2 + 2
δ2 (x+ , s+ ; µ+ ) ≤ + ,
15 20(1 − θ)

where 0 ≤ θ ≤ 1 and µ > 0.

293
CNMA’2022
Second National Conference on
Mathematics and its Applications

Proof The proof is similar to the first part of proof of Lemma 6 in [1]. 
Complexity analysis

In this last part, we give the iteration bound of our algorithm. Before this, we must show that
our algorithm is well defined and this imply the polynomial iteration complexity of the algo-
rithm which starts from a strictly feasible point (x, s) and a barrier parameter µ > 0 such that
δ ≤ τ = √2 . Then, after the update of barrier parameter to µ+ = (1 − θ)µ with θ = √1 , Lemma 50
10 3n
yields δ++ ≤ √2 . Hence,
10

2 ≤ 2(1 − θ) 5nθ 2 + 2 2(1 − θ) 11 8θ 2 − 16θ + 19

δ++ + ≤ + = .
15 20(1 − θ) 15 60(1 − θ) 60(1 − θ)

2
Assuming n ≥ 2 and f (θ) = 8θ60(1−θ)
−16θ+19
, we get that the function f is continuous, convex and mono-
     
tone increasing on 0, √1 . Then, f (θ) ≤ f √1 = 0.388 < 25 for all θ ∈ 0, √1 . Under Corollary 14,
6 6 6
we have δ2 (x + , s + ; µ+ ) ≤ √2
10

Theorem 120. If θ = √1 and µ0 = 23 . The number of iterations of our algorithm does not exceed
3n
√ n
10n log for getting the approximate solution.


References
[1] [Link], [Link], A full-Newton step feasible interior-point algorithm for mono-
tone horizontal linear complementarity problems, Optimization Letters, 1-19, 2018.

[2] [Link], [Link] and [Link], A primal-dual infeasible-interior-point algorithm for

linear programming, Math. Program. 61, 1993, 263–280.

[3] [Link], [Link], [Link], A full-Newton step O(n) infeasible-interior-point

algorithm for lineair complementary problems, Elsevier Ltd, 545-561, 2010.

294
CNMA’2022
Second National Conference on
Mathematics and its Applications

NEW ITERATIVE CONJUGATE GRADIENT

METHOD FOR NONLINEAR
UNCONSTRAINED OPTIMIZATION

Sabrina Ben Hanachi 1 [Link]@[Link]

Badreddine SellamiI 1
Mohammed Belloufi 1
1 Department of mathematics , Mohamed-Cherif Messaadia University Souk Ahras

Abstract: Conjugate gradient (CG) method is one of the most important iterative mathemat-
ical techniques used to solve unconstrained optimization problems because of its simplicity,
low memory requirements and global convergence properties. In this study, we consider a
new hybrid conjugate gradient method, which it is generated from a convex combination of
Conjugate Descent proposed by Fletcher ( abbreviated CD) and Al-Bayati and Al-Assady (
abbreviated BA) methods, our selected method produces the sufficient descent at each iteration
and global convergence property is established. Numerical results and their performances are
presented to show that our new hybrid conjugate gradient method usually gives more efficient
results than some of the known methods CD and BA.

Keywords: Unconstrained optimization; Hybrid conjugate gradient method; Sufficient

descent; Convex combination; Global convergence.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
A given an unconstrained optimization problem

min{f (x), x ∈ Rn }. (265)

Where f : Rn → R is continuously differentiable function, bounded from below and its gradient
is available. Conjugate gradient algorithms are one of efficient optimization algorithms that are
contribute greatly in obtaining the solution of problem (265), especially large scale problems.
Beginning with a starting guess x0 ∈ Rn , the CG method consists in creating a sequence {xk }∞
k=0 of
n vectors which is given as shown

xk+1 = xk + αk dk , k = 0, 1, 2, 3, . . . .

Where xk is the k-th iterative point, αk is a positive scalar and called the step length is obtained
by line search and dk is the search direction generated by the following forms

d0 = −g0 , dk+1 = −gk+1 + βk dk , k = 0, 1, 2, 3, . . . . (266)

295
CNMA’2022
Second National Conference on
Mathematics and its Applications

In the relation (266), gk is the gradient ∇f (xk ) of f (x) at the point xk and βk ∈ R is an important
scalar referred as CG coefficient. In computing the step size αk , we typically use line searches
algorithms include the strong Wolfe conditions which are given as shown [3], [4]

f (xk + αk dk ) − f (xk ) ≤ δαk gkT dk , (267)

T
|gk+1 dk | ≤ −σ gkT dk . (268)

Where dk is a descent direction and 0 < δ ≤ σ < 1.

The different values assigned to βk define variety of CG formulas. Over the years, many formulas
of this method have been developed. Moreover The most popular of these formulas are Fletcher-
Reeves method (FR). Dai-Yuan method (DY). Conjugate – Descent method (CD) [2]. Hestenes and
Stiefel method (HS). Polak-Ribiere– Polyak method (PR). Al-Bayati and Al-Assady method (BA)
[1]
||gk+1 ||2 ||gk+1 ||2 ||gk+1 ||2
βKFR = ||gk ||2
, βKDY = ykT sk
, βKCD = −gkT sk
,
(269)
T T 2
gk+1 yk gk+1 yk ||yk ||
βKHS = ykT sk
, βKP RP = ||gk ||2
, βKBA = dkT yk
.

Where ||.|| is the Euclidean norm and yk = gk+1 − gk .

In the case where f is a strictly convex quadratic function and if αk is obtained by exact line search,
the parameters βk are equal and the quadratic CG algorithm generates the same sequence {xk }∞ k=0 ,
but in the opposite case nonlinear function with inexact line searches we get different sequences
{xk }∞
k=0 which imply variety methods. One of the most useful CG methods is the Hybrid method
which combines various conjugate gradient methods to improve the behavior of these methods.
The objective of this study is to consider a convex combination of two methods, the first method
is the method CD and the second method is the method BA.

Main results
hyb
The parameter βk in the proposed method, denoted by βK , is defined by the following form
hyb
βk = (1 − θk )βkCD + θk βkBA , (270)

the direction dk+1 are given by

hyb hyb hyb
d0 = −g0 , dk+1 = −gk+1 + βk dk , k = 0, 1, 2, 3, . . . . (271)

hyb hyb
Note that, if θk = 0, then βk = βkCD , if θk = 1, then βk = βkBA . On the other side, if 0 < θk < 1,
hyb
then βk is a convex combination of βkCD and βkBA . The hybridization parameter θk in the relation
(270) is given by the following form

(g T y )(g T d )+||g ||2 (d T y )


0 if ||gk+1 k||2 (dkT yk )+||yk+1||2 (g T kd )k ≤ 0,





 k+1 k k k k k



 T
yk )(gkT dk )+||gk+1 ||2 (dkT yk ) T
yk )(gkT dk )+||gk+1 ||2 (dkT yk )

 (gk+1
 (gk+1
θk =  2 T 2 T , if 0 < ||gk+1 ||2 (dkT yk )+||yk ||2 (gkT dk )
< 1,
 ||gk+1 || (dk yk )+||yk || (gk dk )






(g T y )(g T d )+||g ||2 (d T y )

if ||gk+1 k||2 (dkT yk )+||yk+1||2 (g T kd )k ≥ 1.

1


k+1 k k k k k

296
CNMA’2022
Second National Conference on
Mathematics and its Applications

The following Theorem proved that the search direction dk generated by the selected method
satisfies the sufficient descent condition.

Theorem 121. Let strong Wolfe conditions (267) and (268) hold, if 0 < θk < 1 then dk+1 is given by
(271) satisfies the sufficient descent condition i.e

gkT dk ≤ −c||gk ||2 ,

where
c = (1 − σ )(1 − λ2 ) + 2(1 + µ)λ1 .

Now, we give the next theorem which establishes the global convergence of our proposed method
with strong Wolfe line searches.

Theorem 122. Consider the iterative method, let all conditions of theorem (121) hold, then either
gk = 0, for some k, or
lim inf ||gk || = 0.
k→∞

Numerical Results

The results are given in two Figures which are represented the performance profiles of our new
hybrid method versus CD and BA based on the CPU time and number of iterations , respectively.
They shown that the performance profiles of our method was higher than the other CG methods
on the testing problems.

References
[1] Al-Bayati, A.Y and Al-Assady, N.H, Conjugate gradient method, Technical Research,
school of computer studies, Leeds University . (1986)

[2] Fletcher, Roger, Practical Methods Of Optimization Vol. 1 Unconstrained Optimization,

John Wiley & Sons, (1987).

[3] Hager, William W and Zhang, Hongchao, A survey of nonlinear conjugate gradient meth-
ods, Pacific journal of Optimization, 2 (2006), pp. 35–58

[4] Yuan, Gonglin and Wei, Zengxin and Zhao, Qiumei, A modified Polak–Ribière–Polyak con-
jugate gradient algorithm for large-scale optimization problems , IIE Transactions , 46 (2014),
pp. 397–413 .

297
CNMA’2022
Second National Conference on
Mathematics and its Applications

Comparative Study on Multiple Chaotic

Maps Incorporated Global Optimization
Algorithms

Hamaizia Tayeb 1 h2tayeb@[Link]

Nacer Meriem1
1 Department of mathematics , University Constantine 1

Abstract: In this paper we presented the performances of a new class of evolutionary algo-
rithms called chaotic optimization algorithm (COA). Proposed to solve nonlinear optimization
problems with bounded variables by Caponetto et al. Chaotic optimization is a new stochastic
optimization algorithm, which directly uses chaotic variables to find the optimal solution.
Different chaotic maps have been considered, combined with several working strategies.
We propose five different 2D chaotic maps in an optimization algorithm using a two-step
chaotic optimization method and compare them. This study reviews and compares chaotic
optimization algorithms from the literature. Moreover, the two-phase strategy is a commonly
used technique in a COA to refine the solution and help escape local optima. A performance
study is conducted to understand their impact on a chaotic optimization algorithm.

Keywords: Chaotic optimization algorithm- Dynamic system- Chaos- Optimisation.

2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

References
[1] M. Bucolo, R. Caponetto, L. Fortuna, M. Frasca, A. Rizzo, Does chaos work better than
noise? IEEE Circuits and Systems , Magazine 2 (3) (2002) 4-19.

[2] T. Hamaizia, R. Lozi, An improved chaotic optimization algorithm using a new global locally
averaged strategy, Journal of Nonlinear Sys- tems and Applications 3 (2) (2012) 5863.

298
CNMA’2022
Second National Conference on
Mathematics and its Applications

Comparison of Some Scalarization

Methods in Multiobjective Optimization

Bouguern Wafa 1 [Link]@[Link]

1 Department of mathematics , Bordj Bou Arrérid University

Abstract: This paper is dedicated to study some scalarization methods commonly used in
multiobjective optimization, the properties of these methods are examined with respect to
basic features such as ordering cone, convexity and boundedness. We made numerical tests
and a comparison between these different methods.

Keywords: Weighted Sum method, ε-constraint method, Benson’s method, Elastic Con-
straint method, multiobjective optimization, pareto front.
2010 Mathematics Subject Classification: Primary 34C29, 34C25, 47H11.

Introduction
Several computational methods have been proposed for characterizing pareto optimal solutions
depending on the different scalarizations of the multiobjective optimization problem. Among the
many possible ways of scalarizing the multiobjective linear programming problems, the Weight-
ing method, ε-constraint method, Benson’s method, Elastic Constraint method.
For determining solutions of the multiobjective optimization problem (MOP)




 minx f (x) := (f1 (x), . . . , fp (x))


(MOP )



 subject to


 x ∈ X.


a widespread approach is the transformation of this problem to a scalar-valued parameter depen-

dent optimization problem. This is done for instance in the Weighted Sum method [5], where we
solve
Xp
(W SP (λ)) min λk fk (x)
x
k=1
p
The Weighted Sum problem (W SP (λ)) uses the vector of weights λ ∈ R≥ as a parameter. Another
scalarization is based on the minimization of only one of the p objectives while all the other
objectives are transformed into constraints by introducing upper bounds. This scalarization is
called ε-constraint method and is given by

299
CNMA’2022
Second National Conference on
Mathematics and its Applications


min fj (x)






x∈X



(ECj (ε))


s.c







 fk (x) ≤ εk k = {1, ...p} / {j} , k , j


Here the parameters are the upper bounds εk , k = {1, ...p} / {j} for a k ∈ {1, ...p} Surveys about differ-
ent scalarization approaches can be found in [5] [3].
For the ε-constraint method we have no results on properly efficient solutions. In addition,
the scalarized problem (ECj (ε)) may be hard to solve in practice due to the added constraints
fk (x) ≤ εk . In order to address this problem we can relax these constraints by allowing them to be
violated and penalizing any violation in the objective function. Ehrgott and Ryan (2002) used this
idea to develop the Elastic Constraint scalarization [5], the last method that we will present and
discuss in this paper is the Benson’s method.
Finally we made numerical tests and a comparison between these different methods.

Main results
It should be noted that the comparison of different methods is a delicate problem, because not
all methods may have the same comparable features.
The Weighted Sum scalarization method may guarantee to generate all proper efficient solutions
under the convexity assumption.
The Elastic Constraint method which is a modification of the ε-constraint method, gives condi-
tions on the characterization for efficient and properly efficient solutions.
The Weighted Sum method can be applied for any closed pointed convex cone serving as an or-
dering cone.
The boundedness below of objective space is not an essential condition when the Weighted Sum
method is applied. Nevertheless, for some weights λ, the problem (W SP (λ)) may have not a finite
optimal solution.
The convexity condition is essential. The ε-constraint method can be applied only in the case
p
when the ordering cone equals R+ .
p
The boundedness from below is not an essential condition for R+ method.
The ε-constraint method does not require convexity condition on the problem under considera-
tion.
The ε-constraint generates weakly efficient solutions and does not provide conditions for generat-
ing properly efficient solutions.
p
The Benson’s method can be applied only in the case when the ordering cone equals R+ .
The Benson’s method does not require convexity assumptions. The Bens’s method provides neces-
sary and sufficient conditions for efficient solutions, but does not provide conditions guaranteeing
the generation of properly efficient solutions.

300
CNMA’2022
Second National Conference on
Mathematics and its Applications

References
[1] A. Chinchuluun, P.M. Pardalos, A survey of recent developments in multiobjective optimiza-
tion,Annals of Operations Research, 154, issue 1, pp. 29-50, 2007.

[2] A. M. Rubinov and R. N. Gasimov, Scalarization and nonlinear scalar duality for vector
optimization with preferences that are not necessarily a pre-order relation, Journal of Global
Optimization, 29, pp. 455-477, 2004.

[3] I. Y. Kim and O. de Weck, Adaptive weighted sum method, for bi-objective optimization.
Structural and Multidisciplinary Optimization 29, 149–158, 2005.

[4] K. Miettinen and M.Mäkelä, On scalarizing functions in multiobjective optimization, OR

Spectrum, 24, pp. 193-213, 2002.

[5] M. Ehrgott, Multicriteria optimisation, volume 491 of Lect. Notes Econ. math. Syst.
Springer, Berlin, 2000.

[6] R. N. Gasimov, A. Sipahioglu, and T. Sarac, A multiobjective programming approach to

1.5-dimensional assortment problem, European J. Oper. Res., 179, pp. 64-79, 2007.

[7] R. S. Burachik, C. Y. Kaya and M. M. Rizvi, A new scalarization technique to approximate

Pareto fronts of problems with disconnected feasible sets, J. Optim. Theory Appl. 162, 428-
446, 2014.

[8] R. Kasimbeyli, Z.K. Ozturk, N. Kasimbeyli, G.D. Yalcin, B. Icmen, Conic Scalarization
Method in Multiobjective Optimization and Relations with Other Scalarization Methods, Pro-
ceedings of the 3rd International Conference on Modelling, Computation and Optimiza-
tion in Information Systems and Management Sciences-MCO 2015-Part I, In: L.T.H. An,
P.D. Tao, N. N. Thanh (Eds.), Volume of Series: Advances in Intelligent Systems and Com-
puting, Springer, Vol. 359, pp.319-329, 2015.

301
CNMA’2022
Second National Conference on
Mathematics and its Applications

A primal-dual interior-point method for

convex quadratic programming based on
a new kernel function with a hyperbolic
barrier term

Youssra Bouhenache 1 youssrabouhenache1997@[Link]

Wided Chikouche 1
1 Department of mathematics, University Mohamed Seddik Ben Yahia,

Abstract:In this paper, we describe a primal-dual interior-point algorithm for solving convex
quadratic programming based on a new parametric kernel function with a hyperbolic barrier
term. We study the complexity analysis of primal-dual interior-point methods based on the
proposed kernel function and derive the iteration bounds
√ that enjoy the currently best-known
iteration bounds for large-update methods, namely, O n log n log n .

Keywords: Convex quadratic programming, Kernel function, Hyperbolic barrier term,

Large-update methods.
2010 Mathematics Subject Classification: 90C20, 90C25, 90C51

Introduction
We present the standard Convex Quadratic programming (CQP) problem (P ) with its dual prob-
lem (D) as follows

1
min cT x + xT Qx,
2
(P ) Ax = b,
x ≥ 0,

where Q ∈ S+n , A ∈ Rm×n , rank (A) = m, b ∈ Rm , and x ∈ Rn with m ≤ n.

1
max bT y − xT Qx,
2
(D) AT y − Qx + s = c,
s ≥ 0,

where y ∈ Rm and s, c ∈ Rn .
Optimization is the field of applied mathematics where one wants to minimize or maximize an
objective function in several variables.

302
CNMA’2022
Second National Conference on
Mathematics and its Applications

Primal-dual interior-point methods (IPMs) are among the most efficient algorithms to solve lin-
ear programming (LP), complementarity programming (CP) and convex quadratic programming
(CQP). The new era of interior-point methods (IPMs) started in 1984 when Karmarkar [2] pro-
posed his LO algorithm, which enjoyed a polynomial complexity of O(nL) iterations bound.
After 10 years, Nesterov and Nemirovskii [3] extended IPMs from LO to more general
 convex op-
timization problems such as CQP and proved that their algorithms has O n log n iteration bound.
The purpose of this work is to describe a primal-dual interior-point method for CQP based on
a new kernel function with a hyperbolic barrier term. The obtained iteration bound for large-
√
update methods, namely, O n log n log n coincides with the currently best-known iteration bound.

Preliminaries

First, we assume that problems (P ) and (D) satisfy the interior point condition, i.e., there exist
(x0 , y 0 , s0 ) with (x0 , s0 ) > 0 so that

Ax0 = b, AT y 0 − Qx0 + s0 = c.

Then, finding an optimal solution for problems (P ) and (D) is equivalent to solve this system

Ax = b, x > 0, AT y − Qx + s = c, s ≥ 0, xs = 0.

The main idea behind primal-dual IPMs is to replace xs = 0 in (??) by xs = µe, where µ > 0.
Therefore, we get

Ax = b, x > 0, AT y − Qx + s = c, s ≥ 0, xs = µe.

This system has a unique solution (x(µ), y(µ), s(µ)) for any µ > 0, x(µ) and (y(µ), s(µ)) are called the
µ-center of (P ) and (D), respectively. The set of all µ-centers is named the central path of problems
(P ) and (D).
Now, we apply Newton’s method for the search direction (∆x, ∆y, ∆s)

A∆x = b, AT ∆y − Q∆x + ∆s = 0, s∆x + x∆s = µe − xs. (272)

As A is full rank, the system (272) has a unique solution (∆x, ∆y, ∆s), which is named Newton’s
direction. The new iterate is obtained as follows

x+ = x + α∆x, s+ = s + α∆s, y+ = y + α∆y,

where the step size α satisfies α ∈ (0, 1], with (x+ , s+ ) > 0.
Then, we define the scaled vector v and the scaled search directions dx and ds ;
r
xs v∆x v∆s
v= , dx = , ds = .
µ x s

Thus, the new system (272) can be represented as below

Ādx = 0, ĀT ∆y − Q̄dx + ds = 0, dx + ds = v −1 − v. (273)

Again, this system has a unique solution as (∆x, ∆y, ∆s). Therefore, the system (273) can be rewrit-
ten as below

Ādx = 0, Ā∆y − Q̄dx + ds = 0, dx + ds = −∇Φc (v),

303
CNMA’2022
Second National Conference on
Mathematics and its Applications

where Φc (v) is the barrier function of the classical logarithmic kernel function
n
X t2 − 1
Φc (v) = ψc (vi ), v ∈ Rn++ , where ψc (t) = − log t, t > 0.
2
i=1

The main idea in IPMs is to replace ψc (t) by any strictly convex function ψ(t) : Rn++ → Rn+ which is
minimal at t = 1 with ψ(1) = 0.

Main results

In this paper, we propose a new parametric kernel function with a hyperbolic barrier term. Then,
we provide some of its properties which are necessary in the complexity analysis of interior-point
methods based on this function. !
p+1 n
By simple tools, we show that the large-update IPM based on this kernel function has O pn log
2p

iterations complexity for large-update methods. By a special choice of the parameter p, this bound
yields the so far best known iteration bound for large-update methods in terms of hyperbolic ker-
nel functions ([4, 5]).

References
[1] S. Guerdouh, W. Chikouche and I. Touil, An efficient primal-dual interior point algorithm
for linear optimization problems based on a novel parameterized kernel function with a hyper-
bolic barrier term, 2021, halshs-03228790.

[2] N.K. Karmarka, A new polynomial-time algorithm for linear programming, In: Proceedings
of the 16th Annual ACM Symposium on Theory of Computing, vol. 4, pp. 373–395, 1984.

[3] Y.E. Nesterov and A.S. Nemirovski, Interior Point Polynomial Algorithms in Convex Pro-
gramming, SIAM Studies in Applied Mathematics, Vol. 13, SIAM, Philadelphia, 1994.

[4] I. Touil and W. Chikouche, Primal-dual interior point methods for semidefinite programming
based on a new type of kernel functions, Filomat. 34 (12), 3957-3969, 2020.

[5] I. Touil and W. Chikouche, Novel kernel function with a hyperbolic barrier term to primal-
dual interior point algorithm for SDP problems, Acta Mathematicae Applicatae Sinica.
Vol.38, No.1, 44–67, 2022.

304
CNMA’2022
Second National Conference on
Mathematics and its Applications

Interior-point algorithm for linear

programming based on a new descent
direction

Zaoui Billel 1 [Link]@[Link]

Benterki Djamel 1

Khelladi Samia 1
1 Department of Mathematics, Laboratory of Fundamental and Numerical Mathematics, Faculty of Sciences,
University of Ferhat Abbas Setif-1, Algeria

Abstract: This paper is devoted to introducing an interior-point algorithm for linear pro-
gramming with full-Newton step. The introduced method uses an algebraic equivalent
transformation on the centering equation of the system which defines the central path. This
technique leads to a new efficient search direction for the considered algorithm. Moreover, we
prove that the method finds the ε-optimal solution of the underlying problem in polynomial
time. The established numerical tests conclude that the newly proposed algorithm is not only
polynomial but requires a number of iterations clearly lower than that obtained theoretically.

Keywords: linear programming, interior-point methods, descent direction, primal-dual

algorithm.
2010 Mathematics Subject Classification: Primary 90C05, 90C51.

Introduction
Consider the standard linear optimization (LO) problem:

min{cT x : Ax = b, x ≥ 0} (LP)

and its dual problem:

max{bT y : AT y + s = c, s ≥ 0}, (LD)
where b = (b1 , b2 , . . . , bm )T ∈ Rm , c = (c1 , c2 , . . . , cn )T ∈ Rn and A is a m × n given matrix.
The LO problem is one of the most active research areas in mathematical programming. There
are many approaches for solving LO. Among them, interior-point methods which gain much more
attention than others.
Primal-dual interior-point method IPM is one of the most efficient numerical methods for solving
large classes of optimization problems and highly efficient in both theory and practice.
The determination of the search directions plays a key role in case of the IPMs. Therefore, the
search directions can be obtained by using the technique introduced by Darvay [1]. This last is
based on an algebraic equivalent transformation on the centering equations of the central path.
The new search directions are obtained by applying Newton’s method to the resulting system.

305
CNMA’2022
Second National Conference on
Mathematics and its Applications

In 2018, Z. Darvay [2] proposed another technique for obtaining a new descent direction for
solving linear problem (LP). The technique is based on applying the function ψ(t) = t 2 on both
sides of a specific algebraic equivalent transformation of the centering equations defined by Zhang
and Xu [3].
Motivated by the above work, we reconsider this technique with a new function ψ(t) = t 3 , which
 
  n + 24 1 − √1 
 √  3
4  
yields a new search direction with a polynomial complexity, namely O28 n log


iterations.

Main results
Our theoretical results prove the quadratic convergence neighborhood of the proximity measure
and the polynomial complexity of the obtained algorithm. Moreover, the numerical results are
acceptable and encouraging. Finally, the implementation with the update parameter θ reduces
significantly the number of iterations produced by this algorithm and leads this algorithm to
reach their real numerical performances. These numerical results consolidate and confirm our
theoretical purpose.

References
[1] Zs. Darvay, New interior point algorithms in linear programming, Adv. Model. Optim. 5(1)
(2003), 51–92.

[2] Zs. Darvay, P. R. Takàcs, New method for determining search directions for interior-point
algorithms in linear optimization, Optim. Lett. 12 (2018), 1099–1116 .

[3] L. Zhang, Y. Xu, A full-Newton step interior-point algorithm based on modified Newton direc-
tion, Oper. Res. Lett. 39 (2011), 318–322.

306
CNMA’2022
Second National Conference on
Mathematics and its Applications

A BFGS algorithm for global optimization

using random perturbation

Raouf Ziadi 1 [Link]@ [Link]

1 Laboratory of Fundamental and Numerical Mathematics (LMFN), Department of Mathematics, University Fer-
hat Abbas Setif, Algeria

Abstract: In this paper, we give a new representation to the limited memory BFGS methods,
and show how to use them efficiently for solving smooth global optimization problems,
by considering a random perturbation following a truncated Gauss’s law. Our approach is
suitable for solving large-scale bound-constrained global optimization problems. Theoretical
results ensure that the proposed method converges to a global minimizer almost surely.
Numerical experiments are achieved on some typical test problems and comparisons with
well-known methods are carried out to show the performance of our algorithm.

Keywords: Global optimization, Limited memory BFGS method, Stochastic perturba-

tion, Truncated Gauss’s law.
2010 Mathematics Subject Classification: 90C26, 90C90..

In this paper we consider the following bound-constrained global optimization problem

min f (x) (P)

x∈D

n
Y
where D is the hyper-rectangle D (i) , D (i) = [a(i) , b(i) ] and the objective function f (x) : Rn → R is
i=1
not necessarily convex but differentiable whose gradient ∇f is bounded. No additional smooth-
ness on ∇f is required.
The problem (P ) is of interest in many real-world applications involving objective functions
which are differentiable but non-convex. Many methods for solving differentiable global opti-
mization problems have been proposed, and are classified into deterministic and stochastic meth-
ods. As is well known, deterministic algorithms provide a theoretical guarantee of locating the
ε-global optimum. When dealing with an oscillating function in a large search space or in rel-
atively high dimensions, deterministic exploration methods (such as DIRECT methods, the ap-
proach based on the introduction of an auxiliary function or covering methods [2, 3, 5], etc.) are
not effective and can have unreasonable calculation times. Indeed, with these approaches, it is
hard to obtain useful information while exploring all the regions of the feasible domain.
Stochastic algorithms such as Simulated Annealing algorithm (SA), Classification and Regres-
sion Trees (CART), Random Walk, Tabu Search (TS), Variable Neighbourhood Search (VNS) etc.
[6], involve random sampling or a combination of random sampling and local search; they are the-
oretically well studied. They ensure the convergence to the global minimum only in probability.
Unfortunately, most of them are not well suited to efficiently solve high-dimensional problems,
particularly those containing more than 10 variables.

307
CNMA’2022
Second National Conference on
Mathematics and its Applications

In this paper, we suggest a method for solving large-scale problems. This method is a modifi-
cation of the limited memory BFGS method for bound-constrained problems (L-BFGS-B) and we
show how to use it efficiently to deal with global optimization problems by the adjunction of a
stochastic perturbation following a truncated Gauss’s law. The proposed method will be called
P-LBFGSB (Perturbed L-BFGS-B algorithm). Starting from a point X0 in D, our new sequence
{Xk }k≥0 is given by

n        o
Xk+1 ∈ arg min f G(Xk ) , f Pk1 , f Pk2 , . . . , f Pkm ,
with  
Pkl = Pl G(Xk ) for l = 1, 2, . . . r ,
where G(Xk ) is the
 last point obtained by a few iterations using the L-BFGS-B algorithm starting
from Xk and Pl G(Xk ) , for l = 1, 2, . . . , r, are the stochastic perturbations of the point G(Xk ), they
are renewed independently at each iteration k, having the following truncated density function

D 1D (x)
Γµ,σ (x) = Z .Nµ,σ (x)
Nµ,σ (x)dx
D

where 1D is the indicator function of D and Nµ,σ (x) is the multivariate Gauss density function with
σ a positive scale referred to as the diversity (standard deviation) and µ = (µ(1) , µ(2) , . . . , µ(n) ) is a
location vector. The perturbations Pk1 , Pk1 , . . . , Pkr are i.i.d (independent and identically distributed)
n-dimensional random vectors having the common law Γµ,σ D with the same diversity parameter

that decreases to zero slowly enough to prevent the sequence {Xk }k≥0 from converging to a local
minimum.

References
[1] Ali, M., Pant, M., Singh, V. P., Two modified differential evolution algorithms and their
applications to engineering design problems. World Journal of Modeling and Simulation,
6(1), 72-80, (2010).

[2] Ziadi, R., Bencherif-Madani, A., R., Ellaia. A deterministic method for continuous global
optimization using a dense curve. Mathematics and Computers in Simulation, 178, 62-91,
(2020).

[3] Ziadi,R, Bencherif-Madani, A., A covering method for continuous global optimization.
International Journal of Computing Science and Mathematics, 12, 672-684, (2022).

[4] Ziadi, R., Ellaia, R., Bencherif-Madani, A., Global optimization through a stochastic per-
turbation of the Polak-Ribière conjugate gradient method. Journal of Computational and
Applied Mathematics, 317, 672-684, (2017).

[5] Ziadi, R., Bencherif-Madani, A., Ellaia, R., Continuous global optimization through the
generation of parametric curves. Applied Mathematics and Computation, 282, 65-83,
(2016).

[6] Zhigljavsky, A., Zilinskas, A., Stochastic global optimization (Vol. 9). Springer Science &
Business Media, (2007).

308