0% found this document useful (0 votes)
5 views25 pages

Two-Strain SEIR Model Stability Analysis

This paper analyzes a two-strain SEIR epidemic model incorporating a quarantine strategy, represented mathematically by two parameters that reflect the efficiency of quarantine for each strain. The authors demonstrate the existence of four equilibrium points and derive two basic reproduction numbers, exploring the global stability of the model using Lyapunov's method. Numerical simulations are provided to validate the theoretical findings and illustrate the effectiveness of quarantine in reducing infection rates.

Uploaded by

El-Mehdi Farah
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
5 views25 pages

Two-Strain SEIR Model Stability Analysis

This paper analyzes a two-strain SEIR epidemic model incorporating a quarantine strategy, represented mathematically by two parameters that reflect the efficiency of quarantine for each strain. The authors demonstrate the existence of four equilibrium points and derive two basic reproduction numbers, exploring the global stability of the model using Lyapunov's method. Numerical simulations are provided to validate the theoretical findings and illustrate the effectiveness of quarantine in reducing infection rates.

Uploaded by

El-Mehdi Farah
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Chapter 1

Global stability analysis of two-strain SEIR


epidemic model with quarantine strategy

Zakaria Yaagoub ∗ , Jaouad Danane and Karam Allali

Abstract In this paper, we study a two-strain SEIR epidemic model with bilinear
and non-monotonic incidence functions in which the quarantine strategy is taken
into consideration, this strategy is presented mathematically by u1 and u2 , the first
represents the efficiency of the quarantine strategy concerning the first strain infection
rate and the second represents the efficiency of the quarantine strategy concerning
the second strain infection rate. By using the new generation matrix method, we
have shown that the model has four equilibrium points and two basic reproduction
numbers R01 and R02 which depend on u1 and u2 . Using the Lyapunov’s method we
have shown the global stability of different equilibrium point, this stability depends
on R01 and R02 . Finally, we have given two types of numerical simulations, the first to
confirm the theoretical and to illustrate the effectiveness of quarantine to minimize
the infection in the population.

Key words: two-strain; epidemic model; bilinear incidence; non-monotone inci-


dence; quarantine.

Zakaria Yaagoub
Laboratory of Mathematics, Computer Science and Applications, Faculty of Sciences and Tech-
nologies, University Hassan II of Casablanca, PO Box 146, Mohammedia 20650, Morocco, e-mail:
[Link]@[Link]
Jaouad Danane
Hassan First University, National School of Applied Sciences, Laboratory of Systems Modelization
and Analysis for Decision Support, Berrechid, Morocco, e-mail: jaouaddanane@[Link]
Karam Allali
Laboratory of Mathematics, Computer Science and Applications, Faculty of Sciences and Tech-
nologies, University Hassan II of Casablanca, PO Box 146, Mohammedia 20650, Morocco, e-mail:
allali@[Link]
∗ Corresponding author: [Link]@[Link]

1
2 Zakaria Yaagoub, Jaouad Danane and Karam Allali

1.1 Introduction

Infectious diseases are responsible for some disorders in living organisms in the
human body. They are usually caused by organisms such as: bacteria, viruses, fungi
or parasites. These are normally harmless, but under certain conditions they can
be dangerous causing fatal illnesses [1]. The mathematical literature contains many
works which already treated many kind of diseases by mathematical modeling. Most
of these models use specific notations for different categories of populations such as S
for susceptible individuals, I for infected individuals and R for recovered individuals.
The first mathematical model of the spread of infectious diseases is formulated and
analyzed by Daniel Bernoulli in 1760 [2] to assess the effectiveness of susceptible
variolation against smallpox. In 1911 Roland Ross, was preoccupied with studying
the dynamics of malaria, he came to develop a mathematical model for which model
the spread of malaria. Subsequently, inspired from the ideas of Roland Ross, in
1927 W.O. Kermack and A.G McKendrick [3], have studied infectious diseases with
special attention to the dynamics of transmission. When the infection takes a little
time to appear in individuals of a population, another category the population will
be added; this is called the population of exposed individuals noted by E which
represents those who are infected individuals but not yet infectious. However, they
will be infectious after a given time. To better understand how such an infectious
disease is transmitted in a population, some authors use the method of Lyapunov
functions to study the global stability of the equilibrium points of the model proposed
to describe for these infectious diseases (see [4–10]), and the choice of incidence
functions is very important to describe the transmission of such an infectious disease
noting that the incidence function represents the total number of newly infected per
unit of time. Some epidemiological models use a bilinear incidence function as: αSI
with α is the infection rate of the disease and N is total population, if the population
of infected is large compared to the population of susceptible then the amount of
contacts with these two populations will decrease because of the quarantine and
health precautions. Moreover, we know that viruses mutate, so we can have some
strains of a virus but most of the work in mathematical literature are only interested
in the case of two strains, therefore the infected population will be divided into two
sub-populations I1 and I2 with I1 the sub-population infected of strain 1 and I2 the
sub-population infected of strain 2. Additionally, these models have two incidence
rates, the primary is for the strain 1 and therefore the second is for the strain 2.
In order to describe the disease transmission rate, many epidemiological models
describe the transmission of an infectious disease by bilinear or non-monotonic
incidence rate (see[11–18]), recently Bentaleb and Amine [18] study a multi-strain
SEIR epidemic model with bilinear and non-monotone incidence functions and they
given the sufficient conditions of global stability of the determined equilibrium
points. More recently, the previous multi-strain SEIR epidemic have been improved
in [19] by taken to account two non-monotonic incidence rates, which have been
further studied in [20] by considering two general incidence rates. Many works
have shown the efficiency of adding different type of controls in order to reduce the
infection (see[20–36]). In this work, we consider the same model as [18] with the add
1 Global stability analysis of two-strain SEIR epidemic model with quarantine strategy 3

the effect of two quarantine controls. The first control will describe the quarantine for
the first strain sub-population, while the second control will represent the quarantine
for the second strain sub-population.
The present work is organized as follow. In Section 1.2, we formulate the two-
strain SEIR epidemic model with bilinear and non-monotonic incidence functions.
After we will prove the positivity and boundedness of solutions, In Section 1.3, we
will analyze the model starting with the determination of the equilibrium points of
our system and the basic reproduction numbers and finally we will give the sufficient
and necessary conditions for the stability of the determined equilibrium point. In
Section 1.5, two kinds of numerical simulation is given, the first one to show the
effect of different problem parameters on the infection spread and the second and to
illustrate the efficiency of controls, and finally the last section concludes the work.

1.2 Mathematical formulation

1.2.1 Mathematical model

The two-strain SEIR model is composed of six compartments. The compartment of


susceptible individuals S, strain-1 exposed E1 , strain-2 exposed E2 , strain-1 infected
I1 , strain-2 infected I2 and recovered individuals R. The model is given by the
following non-linear system of differential equations:

dS β(1 − u2 )SI2
= Λ − α(1 − u1 )SI1 − − δS,


1 + k I22




 dt

 dE1

= α(1 − u1 )SI1 − (γ1 + δ)E1,




 dt
β(1 − u2 )SI2

dE2


 dt = − (γ2 + δ)E2,


1 + k I22


(1.1)
 dI 1
= γ1 E1 − (µ1 + δ)I1,






 dt
dI2


= γ2 E2 − (µ2 + δ)I2,





 dt
dR


= µ1 I1 + µ2 I2 − δR.



 dt
This model can be illustrated by Fig. 1.1. The parameters description is given in
Table 1.1 with the fact that the initial conditions S(0), E1 (0), E2 (0), I1 (0), I2 (0) and
R(0) are positive. We will assume in our model that the individuals of the recovered
compartment do not become susceptible as shown in Fig. 1.1.
4 Zakaria Yaagoub, Jaouad Danane and Karam Allali

δE1 δI1

E1 (t) γ1 E1 I1 (t)

α(1 − u1 )SI1 µ1 I 1

Λ S(t) δS R(t) δR

β(1 − u2 )SI2
µ2 I2
1 + k I22

E2 (t) γ2 E2 I2 (t)

δE2 δI2

Fig. 1.1: The diagram of SEIR two-strain model.

Table 1.1: Description of parameters of the model (1.1)

Parameters Description
Λ Recruitment rate
1/δ Average life expectancy of the population
α Infection rate of the strain 1
β Infection rate of the strain 2
u1 The efficiency of the quarantine strategy concerning the first strain infection rate
u2 The efficiency of the quarantine strategy concerning the second strain infection rate
1/µ1 Average infection period of strain 1
1/µ2 Average infection period of strain 2
1/γ1 Average latency period of strain 1
1/γ2 Average latency period of strain 2
k Parameter that measures the psychological or inhibitory effect of strain 2

1.2.2 Posivity and boundedness of solutions

In this subsection, we will begin our study of our model by demonstrating the
well-posedness of the all solutions.
1 Global stability analysis of two-strain SEIR epidemic model with quarantine strategy 5

Proposition For any positive initial conditions S(0), E1 (0), E2 (0), I1 (0), I2 (0), R(0),
the variables of the model (1.1): S(t), E1 (t), E2 (t), I1 (t), I2 (t) and R(t) will remain
positive and bounded for all t > 0. 

Proof First, let T = sup{τ ≥ 0 | ∀t, 0 ≤ t ≤ τ such that S(t) ≥ 0, E1 (t) ≥


0, E2 (t) ≥ 0, I1 (t) ≥ 0, I2 (t) ≥ 0, R(t) ≥ 0}.
Let’s now prove that T = +∞.
Assume that 0 < T < +∞; by continuity of solutions, we have S(T) = 0 or E1 (T) = 0
or E2 (T) = 0 or I1 (T) = 0 or I2 (T) = 0 or R(T) = 0.
If S(T) = 0 before the other variables E1 E2, I1, I2, R, become zero. Therefore

dS(T) S(T) − S(t) −S(t)


= lim− = lim− 6 0.
dt t→T T −t t→T T − t

From the first equation of the system (1.1), we have

dS(T)
= Λ > 0.
dt
If E1 (T) = 0 before S, E2, I1, I2, R, become zero then

dE1 (T) E1 (T) − E1 (t) −E1 (t)


= lim− = lim− 6 0.
dt t→T T −t t→T T − t

From the second equation of the system (1.1) with the fact E1 (T) = 0, we will have

dE1 (T)
= α(1 − u1 )SI1 .
dt
Since α > 0 and 0 < u1 < 1, we have
dE1 (T)
> 0.
dt
If E2 (T) = 0 before S, E1, I1, I2, R, become zero then

dE2 (T) E2 (T) − E1 (t) −E2 (t)


= lim− = lim− 6 0.
dt t→T T −t t→T T − t

From the 3r d equation of the system (1.1) with the fact E2 (T) = 0, we will have

dE2 (T) β(1 − u2 )SI2


= .
dt 1 + k I22

Since β, k are positive and 0 < u2 < 1, we have:

dE2 (T)
> 0.
dt
If I1 (T) = 0 before S, E1, E2, I2, R, become zero then
6 Zakaria Yaagoub, Jaouad Danane and Karam Allali

dI1 (T) I1 (T) − I1 (t) −I1 (t)


= lim− = lim− 6 0.
dt t→T T −t t→T T − t

From the 4th equation of the system (1.1) with the fact I1 (T) = 0, we will have

dI1 (T)
= γ1 E1 .
dt
Since γ1 > 0, we have
dI1 (T)
= γ1 E1 > 0.
dt
If I2 (T) = 0 before S, E1, E2, I1, R, become zero then

dI2 (T) I2 (T) − I2 (t) −I2 (t)


= lim− = lim− 6 0.
dt t→T T −t t→T T − t

From the 5th equation of the system (1.1) with the fact I2 (T) = 0, we will have

dI2 (T)
= γ2 E2 .
dt
Since γ2 > 0, we have
dI2 (T)
= γ1 E2 > 0.
dt
If R(T) = 0 before S, E1, E2, I1, I2, become zero then

dR(T) R(T) − R(t) −R(t)


= lim− = lim− 6 0.
dt t→T T −t t→T T − t

From the 6th equation of the system (1.1) with the fact R(T) = 0, we will have

dR(T)
= µ1 I1 + µ2 I2 .
dt
Since µ1 > 0 and µ2 > 0 , we have

dR(T)
= µ1 I1 + µ2 I2 > 0.
dt
This is a contradiction, therefore T can note be finite. We shown that all the variables
of the model remain positive.
We will show after that they are bounded. Let the total population

N(t) = S(t) + E1 (t) + E2 (t) + I1 (t) + I2 (t) + R(t).

We add all the equations of the system (1.1), we will have


1 Global stability analysis of two-strain SEIR epidemic model with quarantine strategy 7

dN(t) dS(t) dE1 (t) dE2 (t) dI1 (t) dI2 (t) dR
= + + + + +
dt dt dt dt dt dt dt
= Λ − δN(t),

then,
Λ Λ
N(t) = + (N(0) − )e−δt ,
δ δ
so
Λ
lim N(t) =
,
t→+∞ δ
hence, ∀ > 0, ∃t0 > 0 such as ∀t ≥ t0 we have
Λ
N(t) < + .
δ
So N(t) and all variables are bounded for all t > 0 . 

1.3 Analysis of the model

In this section, we will show that the model has a disease-free equilibrium point
and three endemic equilibrium points and we will study the global stability of these
equilibrium points found using Lyapunouv’s function method. We know that the first
five equations of the system (1.1) are independent of R and also that the number of
the total population N achieves the Eq.(1.3), so we can reduce our system to another
(1.2) of five equations instead of six:

dS β(1 − u2 )SI2
= Λ − α(1 − u1 )SI1 − − δS,


1 + k I22




 dt

dE

1

= α(1 − u1 )SI1 − (γ1 + δ)E1,





 dt
β(1 − u2 )SI2
 dE2


= − (γ2 + δ)E2, (1.2)
 dt

 1 + k I22

 dI1
= γ1 E1 − (µ1 + δ)I1,






 dt
dI

 2

= γ2 E2 − (µ2 + δ)I2 .


 dt
With
R = N − E1 − E2 − I1 − I2 . (1.3)
8 Zakaria Yaagoub, Jaouad Danane and Karam Allali

1.3.1 The basic reproduction number calculation

We know that the biological definition of basic reproduction number is the number
of newly infected spawned by an infected individual in a population consisting
of susceptible individuals only. Mathematical the basic reproduction number R0
is the spectral radius of the next generation FV −1 . So R0 = ρ(FV −1 ), F is the no-
negative matrix of new infection cases, and V is the matrix of the transition infections
associated with model (1.2). Let

Λ
©0 0 α(1 − u1 ) 0
­ δ ª
®
Λ®
F = ­­0 0 β(1 − u2 ) ®® ,
­
0
­ δ®
­0 0 0 0 ®
«0 0 0 0 ¬
and
γ +δ 0 0 0
© 1
­ 0 γ2 + δ 0 0 ®
ª
V =­ ®.
­ −γ1 0 µ1 + δ 0 ®
« 0 −γ2 0 µ2 + δ ¬
So
α(1 − u1 )Λγ1 α(1 − u1 )Λ
0 0
­ δ(µ1 + δ)(γ1 + δ) δ(µ1 + δ)
© ª
®
β(1 − u2 )Λγ2 β(1 − u2 )Λ ®
­ ®
FV −1 =­ 0 0 ®.
­
­ δ(µ2 + δ)(γ2 + δ) δ(µ2 + δ) ®
0 0 0 0
­ ®
­ ®
« 0 0 0 0 ¬
Then
R0 = max{R01 ; R02 },
with
α(1 − u1 )Λγ1
 R0 = δ(µ1 + δ)(γ1 + δ) ,


 1


(1.4)
 R2 = β(1 − u2 )Λγ2 ,

 0 δ(µ + δ)(γ + δ)

 2 2
we note
a = γ1 + δ ; b = γ2 + δ ; c = µ1 + δ ; e = µ2 + δ.
Then,
α(1 − u1 )Λγ1
 R0 = ,

 1
δac


(1.5)
β(1 − u2 )Λγ2
 R02 = .


 δbe
1 Global stability analysis of two-strain SEIR epidemic model with quarantine strategy 9

1.3.2 Steady states

The model (1.2) has a free equilibrium point and three endemic equilibrium points
as follows:
Λ
1. The disease-free equilibrium E f = ( , 0, 0, 0, 0).
δ
2. The strain 1 endemic equilibrium E s1 = (Ss∗1 , E1,s
∗ , E ∗ , I ∗ , I ∗ ). Where
1 2,s1 1,s1 2,s1

ac 1 Λ
Ss∗1 = = 1 ,
γ1 α(1 − u1 ) R0 δ

c ∗

E1,s = I ,
1 γ1 1,s1
Λγ1 δ δ

I1,s = − = (R1 − 1),
1 ac α(1 − u1 ) α(1 − u1 ) 0

I2,s1
= 0 , E2,s

1
= 0.
∗ , E ∗ , I ∗ , I ∗ ). Where
3. The strain 2 endemic equilibrium E s2 = (Ss∗2 , E1,s2 2,s2 1,s2 2,s2

be(1 + k I2,s
∗2 )
1 Λ
Ss∗2 = 2
= (1 + k I2,s
∗2
) ,
γ2 β(1 − u2 ) 2
R02 δ
e ∗

E2,s = I ,
2 γ2 2,s2
q
−β(1 − u2 )be + (β(1 − u2 )be)2 − 4kdbe(1 − R02 )

I2,s = ,
2 2δbek

I1,s2
= 0, E1,s

2
= 0.
4. The both strains equilibrium E st = (St∗, E1,t
∗ , E ∗ , I ∗ , I ∗ ). Where
2,t 1,t 2,t

ac be(1 + k I2,t
∗2 )
St∗ = = ,
γ1 α(1 − u1 ) γ2 β(1 − u2 )
c ∗

E1,t = I ,
γ1 1,t
e ∗

E2,t = I ,
γ2 2,t
Λ δ β(1 − u2 )I2,s


I1,t = − − 2
,
α(1 − u1 )St∗ α(1 − u1 )δ α(1 − u1 )(1 + k I2,s ∗2 )
2
r
β(1 − u2 )St∗ γ2 1
I2,t =

− .
kbe k
10 Zakaria Yaagoub, Jaouad Danane and Karam Allali

Theorem 1.2 The equilibrium point E s1 (respectively, E s2 ) exists when R01 > 1
(respectively when R02 > 1). The equilibrium point E st exists when R01 > 1 and
R02 > 1.

Proof 1. For the equilibrium point E s1 to exist, it is necessary that


a. firstly
ac
Ss∗1 > 0 ⇔ > 0,
γ1 α(1 − u1 )
this is obvious because all the parameters of the model are positive,
b. secondly

I1,s1
> 0 ⇔ (R01 − 1) > 0
⇔ R01 > 1,

c. and thirdly

E1,s1
> 0 ⇔ I1,s

1
>0
⇔ R01 > 1.

So for the equilibrium E s1 point to exist it is necessary that R01 > 1.

2. For the equilibrium point E s2 to exist, it is necessary that


a. firstly
1 Λ
Ss∗2 > 0 ⇔ (1 + k I2,s
∗2
) > 0,
2
R02 δ
this is obvious because all the parameters of the model are positive,
∗ exists if:
b. secondly I2,s2

(β(1 − u2 )be)2 − 4kdbe(1 − R02 ) > 0 ⇔ R02 > 1

and
q
−β(1 − u2 )be + (β(1 − u2 )be)2 − 4kdbe(1 − R02 ) > 0
q
⇔ (β(1 − u2 )be)2 − 4kdbe(1 − R02 ) > β(1 − u2 )be ⇔ R02 > 1,

c. and thirdly

E2,s2
> 0 ⇔ I2,s

2
>0
⇔ R02 > 1.

So for the equilibrium E s2 point to exist it is necessary that R02 > 1.


1 Global stability analysis of two-strain SEIR epidemic model with quarantine strategy 11

3. For the equilibrium point E st to exist, it is necessary that


a.
ac
St∗ > 0 ⇔ > 0,
γ1 α(1 − u1 )
this is obvious because all the parameters of the model are positive,
b. secondly

Λ δ β(1 − u2 )I2,t ∗

I1,t >0⇔ >0 +
α(1 − u1 )St
∗ α(1 − u1 )δ α(1 − u1 )(1 + k I2,t
∗2 )

Λγ1 α(1 − u1 ) β(1 − u1 )I2,t



⇔ > 1+
δac α(1 − u1 )(1 + k I2,t
∗2 )

⇔ R01 > 1,

and
c ∗

E1,t = I > 0.
γ1 1,t
c. and thirdly
r
β(1 − u1 )St∗ γ2 1

I2,t >0⇔ − >0
kbe k
β(1 − u1 )St γ2

1
⇔ >
kbe k
β(1 − u2 )γ2 Λ
⇔ >1
δbe
⇔ R02 > 1,

and
e ∗

E2,t = I > 0.
γ2 2,t
So for the equilibrium point E st to exist it is necessary that R01 > 1 and R02 > 1. 

1.3.3 Global Stability

[Link] Global stability of the disease free equilibrium

Theorem 1.3 If R01 ≤ 1 and R01 ≤ 1, then the disease free equilibrium E f of the
system (1.2) is globally asymptotically stable.
Proof (Proof) First, we consider the following Lyapunov function in R5+
   
S S a b
L f (S, E1, E2, I1, I2 ) = S0 − ln − 1 + E1 + E2 + I1 + I2,
S0 S0 γ1 γ2
12 Zakaria Yaagoub, Jaouad Danane and Karam Allali

the time derivative is given by:


 
ÛL f (S, E1, E2, I1, I2 ) = 1 − S0 S(t)
Û + EÛ1 (t) + EÛ2 (t) + a IÛ1 (t) + b IÛ2 (t)
S γ1 γ2
!
β(1 − u2 )SI2
 
S0
= 1− Λ − δS − α(1 − u1 )SI1 −
S 1 + k I22
β(1 − u2 )SI2
+ α(1 − u1 )SI1 − aE1 +
1 + k I22
ac be
− bE2 + aE1 − I1 + bE2 − I2
γ1 γ2
   
S0 S ac
= δS0 2 − − + I1 α(1 − u1 )S0 −
S S0 γ1
!
β(1 − u2 )S0 be
+ I2 −
1 + k I22 γ2
 
S0 S ac be
≤ δS0 2 − − + I1 (R01 − 1) + I2 (R02 − 1).
S S0 γ1 γ2

Since the arithmetic mean is greater than or equal to the geometric mean, we have
S S0
2− − ≤ 0,
S0 S

therefore if R01 ≤ 1 and R02 ≤ 1, we will have LÛ f ≤ 0, then the disease-free


equilibrium point E f is globally asymptotically stable. 

[Link] Global stability of the strain 1 endemic equilibrium

Theorem 1.4 If R02 ≤ 1 < R01 . Then the strain 1 endemic equilibrium point of the
system (1.2) E s1 is globally asymptotically stable.

Proof First, we consider the following Lyapunov function in R5+


       
S S E1 E1
L1 (S, E1, E2, I1, I2 ) = Ss∗1 ∗ − ln ∗ − 1 + E1,s ∗
− ln − 1 + E2
Ss1 Ss1 1 E1,s1 E1,s1
! !
a ∗ I1 I1 b
+ I1,s1 ∗ − ln ∗ − 1 + I2,
γ1 I1,s1 I1,s1 γ2

the time derivative is given by


1 Global stability analysis of two-strain SEIR epidemic model with quarantine strategy 13
!
S∗
ÛL1 (S, E1, E2, I1, I2 ) = 1 − s1 Λ − δS − α(1 − u1 )SI1 − β(1 − u2 )SI2
 

S 1 + k I22

!
E1,s β(1 − u2 )SI2
+ 1− 1
(α(1 − u1 )SI1 − aE1 ) + − bE2
E1 1 + k I22

!
a I1,s b
+ 1− 1
(γ1 E1 − cI1 ) + (γ2 E2 − eI2 ) ,
γ1 I1 γ2

from the endemic state E s1 of the system (1.2), we have

 Λ = δSs1 + α(1 − u1 )Ss1 I1,s1 ,


 ∗ ∗ ∗


ac ∗
 α(1 − u1 )Ss1 I1,s1 = aE1,s1 = γ1 I1,s1 .
∗ ∗ ∗


By direct calculations, we have:
∗ ∗
!
S∗ S∗ SI1 E1,s E1 I1,s
 
ÛL1 = δSs∗ 2 − S − s1 + aE ∗ 3 − s1 − 1 1
1
Ss∗1 S 1,s1
S ∗ E − E∗ I
Ss∗1 I1,s1
1 1,s1 1
!
 
Ss1∗
ac be
+ α(1 − u1 )I1 Ss1 −

+ β(1 − u2 )I2 − .
α(1 − u1 )γ1 1 + k I22 β(1 − u2 )γ2

Since the arithmetic mean is greater than or equal to the geometric mean, we have
∗ ∗
Ss∗1 SI1 E1,s E1 I1,s
3− − ∗ ∗ 1 − ∗ 1 ≤0
S Ss1 I1,s1 E1 E1,s1 I1

and
S Ss∗
2− ∗ − 1 ≤ 0.
Ss1 S
β(1 − u2 )Λγ2
If R02 = ≤ 1, then
δbe
Λ be
≤ ,
δ β(1 − u2 )γ2
we know that the population size satisfies the equations:

N = S + E1 + E2 + I1 + I2

and
Λ
N(t) = + e−δt ,
δ
so
Ss∗1 Λ be
≤ Ss∗1 ≤ ≤ ,
1+ k I22 δ β(1 − u2 )γ2
14 Zakaria Yaagoub, Jaouad Danane and Karam Allali

therefore
Ss∗1 be
− ≤ 0.
1+ k I22 β(1 − u2 )γ2
Finally, after replacing the value of Ss∗1 we have
ac ac ac
Ss∗1 − = − = 0.
α(1 − u1 )γ1 α(1 − u1 )γ1 α(1 − u1 )γ1

Therefore when R02 ≤ 1 < R01 , we will have LÛ 1 ≤ 0, then the strain 1 endemic
equilibrium of the model (1.2) E s1 is globally asymptotically stable. 

To demonstrate the global stability of the endemic equilibrium points E s2 and E st


we will need the following number
Λ√
Rk = k,
δ
with Rk means the strain 2 inhibitory effect number.

[Link] Global stability of the strain 2 endemic equilibrium

Theorem 1.5 If max R01, Rk ≤ 1 < R02 . Then the strain 2 endemic equilibrium


point of the system (1.2) E s2 is globally asymptotically stable.

Proof First, we consider the following Lyapunov function in R5+


    ! !
S S E2 E2
L2 (S, E1, E2, I1, I2 ) = Ss∗2 ∗ − ln ∗ − 1 + E1 + E2,s ∗
∗ − ln ∗ −1
Ss2 Ss2 2 E2,s2
E2,s2
! !
a b ∗ I2 I2
+ I1 + I2,s2 ∗ − ln ∗ −1 ,
γ1 γ2 I2,s2 I2,s2

the time derivative is given by


!
Ss∗ β(1 − u2 )SI2
 
LÛ 2 (S, E1, E2, I1, I2 ) = 1 − 2 Λ − δS − α(1 − u1 )SI2 − + α(1 − u1 )SI2
S 1 + k I22

! !
E2,s β(1 − u 2 )SI2 a
− aE1 + 1 − 2
− bE2 + (γ1 E1 − cI1 )
E2 1 + k I2 2 γ 1

!
b I2,s
+ 1− 2
(γ2 E2 − eI2 ) ,
γ2 I2

from the endemic state E s2 of the system (1.2), we have


1 Global stability analysis of two-strain SEIR epidemic model with quarantine strategy 15

 β(1 − u2 )Ss∗2 I2,s



= δS ∗ + 2
,


 Λ s2
1 + k I2,s
∗2




2
β(1 − u2 )Ss∗2 I2,s

be ∗
= bE2,s = I ,

 2 ∗

+ ∗2 γ2 2,s2
 2


 1 k I2,s 2

By direct calculations, we have:


∗ ∗
!
S∗ S∗ SI2 E2,s E2 I2,s
 
ÛL2 = δSs∗ 2 − S − s2 + bE ∗ 3 − s2 − 2 2
1
Ss∗2 S 2,s2
S ∗ E − E∗ I
Ss∗2 I2,s2
2 2,s2 2
!
Ss∗2
 
ac be
+ α(1 − u1 )I1 Ss2 −

+ β(1 − u2 )I2 − .
α(1 − u1 )γ1 1 + k I2,s2 β(1 − u2 )γ2
2

Since the arithmetic mean is greater than or equal to the geometric mean, we have

S Ss∗2
2− − ≤0
Ss∗2 S

and ∗ ∗
Ss∗2 SI2 E2,s2
E2 I2,s2
3− − − ≤ 0.
S Ss∗2 I2,s
∗ E
2
2
∗ I
E2,s2
2

α(1 − u1 )Λγ1
If R01 = ≤ 1, then
δac
Λ ac
≤ ,
δ α(1 − u1 )γ1
we know that the population size satisfies the equations:

N = S + E1 + E2 + I1 + I2

and
Λ
N(t) = + e−δt ,
δ
so
Λ ac
Ss∗2 ≤ ≤ ,
δ α(1 − u1 )γ1
therefore
ac
Ss∗2 − ≤ 0.
α(1 − u1 )γ1
In order to prove that
Ss∗2 be
− ≤ 0,
1 + k I2,s
2 β(1 − u2 )γ2
2

we have
16 Zakaria Yaagoub, Jaouad Danane and Karam Allali

Ss∗2 be be
− ≤ Ss∗2 −
1 + k I2,s
2 β(1 − u2 )γ2 β(1 − u2 )γ2
2

and

be be(1 + k I2,s
∗2 )
be
Ss∗2 − = 2

β(1 − u2 )γ2 β(1 − u2 )γ2 β(1 − u2 )γ2
be
= ∗2
(1 − k I2,s ),
β(1 − u2 )γ2 2

we know that
N = S + E1 + E2 + I1 + I2
and
Λ −δt
N(t) = e ,
δ
so
Λ2
∗2
I2,s ≤ ,
2
δ2
Λ2 Λ√
∗2 ≤ k
then k I2,s ≤ 1 if Rk = k ≤ 1.
2 δ 2 δ
Therefore when max R0, Rk ≤ 1 < R02 , we will have LÛ 2 ≤ 0. Then the strain 2
 1

endemic equilibrium of the system (1.2) E s2 is globally asymptotically stable. 

[Link] Global stability of the total endemic equilibrium

Theorem 1.6 If R01 > 1, R02 > 1 and Rk ≤ 1. Then the endemic equilibrium point
E st is globally asymptotically stable.

Proof First, we consider the following Lyapunov function in R5+


    ! !
S S E1 E1
L3 (S, E1, E2, I1, I2 ) = St∗ − ln ∗ − 1 + E1,t ∗
∗ − ln E ∗ −1
St∗ St E1,t 1,t
! ! ! !
E2 E2 a ∗ I1 I1
+ E2,t

∗ − ln E ∗ − 1 + I1,t ∗ − ln ∗ − 1
E2,t 2,t γ1 I1,t I1,t
! !
b ∗ I2 I2
+ I2,t ∗ − ln I ∗ −1 ,
γ2 I2,t 2,t

the time derivative is given by


1 Global stability analysis of two-strain SEIR epidemic model with quarantine strategy 17
!
ÛL3 (S, E1, E2, I1, I2 ) = 1 − t Λ − δS − α(1 − u1 )SI1 − β(1 − u2 )SI2
S∗
 

S 1 + k I22

!
E1,t
+ 1− (α(1 − u1 )SI2 − aE1 )
E1

! !
E2,t β(1 − u2 )SI2
+ 1− − bE2
E2 1 + k I22
∗ ∗
! !
a I1,t b I2,t
+ 1− (γ1 E1 − cI1 ) + 1− (γ2 E2 − eI2 ) ,
γ1 I1 γ2 I2

from the endemic state E st of the system (1.2), we have

∗ + α(1 − u )S ∗ I ∗ + β(1 − u2 )St I2,t ,


∗ ∗
= δS

Λ

 t 1 t 1,t
1 + k I2,t
∗2




∗ = aE ∗ = ac I ∗ ,


α(1 − u1 )St∗ I1,t


 1,t γ1 1,t
β(1 − u2 )St I2,t
∗ ∗

∗ = be I ∗ ,


= bE2,t


+ γ2 2,t


 1 k I ∗2
 2,t

By direct calculations, we have:


∗ ∗
!
S∗ S∗ SI1 E1,t E1 I1,t
 
ÛL3 (S, E1, E2, I1, I2 ) = δSt∗ 2 − S − t + aE ∗ 3 − t −
St∗ S 1,t
S ∗ E − E∗ I
St∗ I1,t 1 1,t 1
∗ ∗
!
S∗ SI2 E2,t E2 I2,t
+ bE2,t

3− t − ∗ ∗ − ∗
S St I2,t E2 E2,t I2
 
ac
+ α(1 − u1 ) St∗ −
α(1 − u1 )γ1
!
St∗ be
+ β(1 − u2 )I2 −
1 + k I22 β(1 − u2 )γ2

Since the arithmetic mean is greater than or equal to the geometric mean, we have
∗ ∗
St∗ SI1 E1,t E1 I1,t
3− − ∗ ∗ − ∗ ≤0
S St I1,t E1 E1,t I1

and ∗ ∗
St∗ SI2 E2,t E2 I2,t
3− − ∗ ∗ − ∗ ≤ 0.
S St I2,t E2 E2,t I2
After replacing the value of St∗ we have
18 Zakaria Yaagoub, Jaouad Danane and Karam Allali
ac ac ac
St∗ − = − = 0.
α(1 − u1 )γ1 α(1 − u1 )γ1 α(1 − u1 )γ1
In order to prove that
St∗ be
− ≤ 0,
1 + k I2,t
2 β(1 − u2 )γ2
we have
St∗ be be
− ≤ St∗ −
1+ 2
k I2,t β(1 − u2 )γ2 β(1 − u2 )γ2
and

be be(1 + k I2,t
∗2 )
be
St∗ − = −
β(1 − u2 )γ2 β(1 − u2 )γ2 β(1 − u2 )γ2
be
= ∗2
(1 − k I2,t ),
β(1 − u2 )γ2
we know that
N = S + E1 + E2 + I1 + I2
and
Λ −δt
N(t) = e ,
δ
so
Λ2
∗2
I2,t ≤ ,
δ2
Λ2 Λ√
∗2 ≤ k
then k I2,t ≤ 1 if Rk = k ≤ 1.
δ2 δ
Therefore, if R0 > 1, R0 > 1 and Rk ≤ 1, we will have LÛ 3 ≤ 0. Then the total
1 2

endemic equilibrium of the system (1.2) E st is globally asymptotically stable. 

1.4 Numerical simulations and discussions

In this section we will give some numerical simulations to value and confirm the
results found in the theoretical part using the parameters given in the Table 1.2, two
kinds of numerical simulations are given, in the first simulation is concerning the
global stability of the equilibrium points where we will see that the theoretical results
coincide with the one found numerically results. The second numerical simulation
is to prove the effect of the used control for the two strains of the disease.
1 Global stability analysis of two-strain SEIR epidemic model with quarantine strategy 19

Table 1.2: The parameters values of the system (1.2)


Parameters Fig. 1.2 Fig. 1.3 Fig. 1.4 Fig. 1.5 Fig. 1.6
Λ 1 1 1 1 1
δ 0.2 0.2 0.2 0.2 0.2
α 0.2 0.5 0.2 0.6 0.6
β 0.15 0.15 0.5 0.5 0.5
u1 0.2 0.5 0.2 0.5 −
u2 0.15 0.15 0.5 0.5 −
γ1 0.4 0.4 0.4 0.5 0.5
γ2 0.3 0.3 0.3 0.5 0.5
µ1 0.65 0.4 0.4 0.15 0.15
µ2 0.75 0.4 0.17 0.15 0.15
k 0.01 0.01 0.01 0.01 0.01

12
S
E1
10 E2
I1
8 I2
Population

R
6

0
0 10 20 Time 30 40 50

Fig. 1.2: Stability of disease-free equilibrium E f of two-strain SEIR Model with R01 = 0.62 and
R02 = 0.40.
20 Zakaria Yaagoub, Jaouad Danane and Karam Allali

12
S
E1
10 E2
I1
8 I2

Population
R
6

0
0 10 20 Time 30 40 50

Fig. 1.3: Stability of the strain 1 endemic equilibrium E s1 of two-strain SEIR Model with R01 = 1.66
and R02 = 0.50.
12
S
E1
10 E2
I1
Population

8 I2
R
6

0
0 10 20 Time 30 40 50

Fig. 1.4: Stability of the strain 2 endemic equilibrium E s2 of two-strain SEIR Model with R01 =
0.80, R02 = 1.25 and Rk = 0.5.

12
S
E1
10 E2
I1
8
Population

I2
R
6

0
0 10 20 30 40 50
Time

Fig. 1.5: Stabilty of the endemic equilibrium E st two-strain SEIR Model with R01 = 3.06, R02 =
2.55 and Rk = 0.5
1 Global stability analysis of two-strain SEIR epidemic model with quarantine strategy 21

10 5
u1=u2=1 u1=u2=1
8 u1=u2=0.5 4 u1=u2=0.5
u1=u2=0 u1=u2=0
6 3
S E
1
4 2

2 1

0 0
0 10 20 30 Time 40 50 60 0 10 20 30 Time 40 50 60

5 4
u1=u2=1 u1=u2=1
4 u1=u2=0.5 3 u1=u2=0.5
u1=u2=0 I1 u1=u2=0
E23
2
2

1 1

0 0
0 10 20 30 Time 40 50 60 0 10 20 30 Time 40 50 60

4 4
u1=u2=1 u1=u2=1
3 u1=u2=0.5 3 u1=u2=0.5
I2 u1=u2=0 u1=u2=0
R
2 2

1 1

0 0
0 10 20 30 Time 40 50 60 0 10 20 30 Time 40 50 60

Fig. 1.6: The dynamics of the population for different values of quarantine.
22 Zakaria Yaagoub, Jaouad Danane and Karam Allali

Fig. 1.2 shows the evolution of all the SEIR model variables. We notice in this
figure all curves E1, E2, I1, I2 and R drop to zero, except the curve S representing
the susceptible individuals. So the results of numerical simulation coincides with the
results found in the analytical part concerning the stability of the free equilibrium:
the basic reproduction numbers found using the values of the parameters of the
model described in Table 1.2 for both strains are less than one: R01 = 0.31 < 1 and
R02 = 0.40 < 1. This corresponds to the Theorem 1.3, the disease free equilibrium
E f is globally asymptotically stable.
Fig. 1.3 we note that all elements of the first strain persist while the second strain dies
out. We notice also that the basic reproduction number of the first strain is greater
than one (R01 = 1.66 > 1) and the basic reproduction number of the second strain is
less than one (R02 = 0.50 < 1). This corresponds to the Theorem 1.4, then the strain
1 endemic equilibrium E s1 is globally asymptotically.
Fig. 1.4 we observe that all elements of the second strain persists while the first strain
dies out. We notice also that the basic reproduction number of the second strain is
greater than one (R02 = 1.25 > 1) and greater also than the basic reproduction
number of the first strain (R01 = 0.80 < 1), in addition the the strain 2 inhibitory
effect number is less than 1 (Rk = 0.5 < 1). This corresponds to the Theorem 1.5,
then the strain 1 endemic equilibrium E s2 is globally asymptotically.
Fig. 1.5 shows all elements of both strains persist and both basic reproduction
numbers of thus strains are greater than one: R01 = 3.06 > 1 and R02 = 2.55 > 1, in
addition the the strain 2 inhibitory effect number is less than 1 (Rk = 0.5 < 1). This
corresponds to the Theorem 1.6, then the both strains equilibrium E st is globally
asymptotically.
Fig. 1.6 illustrates the efficiency of the controls u1 and u2 , we observe that the number
of exposed E1, E2 and infected I1, I2 individuals decreases when we increase the
value of the controls u1 and u2 . Also, we observe that the susceptible population
increases when u1 and u2 are maximized.

1.5 Conclusion

Mathematical modeling in epidemiology is a very important tool in understanding


how an infectious disease is transmitted in a population. In this paper, we studied
a two-strain SEIR model with bilinear and non-monotonic incidence rates in which
the quarantine strategy is taken into consideration. The next generation matrix is
used to show that the model has two basic reproduction numbers R01 and R02 . We
established the necessary and sufficient conditions for the global stability of the
four equilibrium points given of our model using the Lyapunov function method
and we have supported these analytical results found by numerical simulations. We
have shown that if R01 and R02 are less than one, then the disease free equilibrium is
globally asymptotically stable. If R02 (respectively, R01 and Rk ) are less than one. Then
the strain 1 equilibrium point E s1 (respectively, the strain 2 equilibrium point E s2 )
is globally asymptotically stable, this means that the strain persists is that with the
1 Global stability analysis of two-strain SEIR epidemic model with quarantine strategy 23

higher basic reproduction number. And, we have shown if R01 and R02 are greater than
1 and if the strain 2 inhibitory effect number Rk is less than 1, then the total endemic
equilibrium point E st is globally asymptotically stable, this means that the strains
persists. A numerical simulation is given to value the theoretical results found. This
numerical simulation has two types, the first confirms the results of the theorem
given in the theoretical part using values of the parameters, the second is to show the
effectiveness of the controls used u1 and u2 to reduce the number of infected of the
two strains. We notice that our contribution which is to add two controls u1 and u2 .
The first represent the efficiency of the quarantine strategy concerning the first strain,
while the second stands for the efficiency of the quarantine strategy concerning the
second strain. Our theoretical results and numerical results have demonstrated that
these two controls can change considerably the infection dynamics. More precisely,
when the two controls are increased, the disease is considerably reduced. As future
directions of this present work, one can consider multi-strain infection problem and
also to compare the numerical results with COVID-19 clinical data.

References

[1] Irshad Ul Haq, Miaozhi Zhang, Pu Yang, Jan Dirk van Elsas, Chapter Five -
The Interactions of Bacteria with Fungi in Soil: Emerging Concepts, Editor(s):
Sima Sariaslani, Geoffrey Michael Gadd, Advances in Applied Microbiology,
Academic Press, Volume 89, 2014, Pages 185 − 215.
[2] Bernoulli, D. (1766) Essai une nouvelle analyse de la mortalite causee par la
petite verole. Mem. Math. Phys. Acad. Roy. Sci., Paris (1766).
[3] W.O. Kermack, A.G. McKendrick, A contribution to the mathematical the-
ory of epidemics. Proceedings of the Royal Society of London, Series A,
115(1927), 700-721.
[4] A. M. Elaiw, Global properties of a class of HIV models, Nonlinear Anal. Real
World Appl. 11 (2010) 2253 − 2263.
[5] A. Korobeinikov, Lyapunov functions and global stability for SIR and SIRS
epidemiological models with nonlinear transmission, Bull. Math. Biol. 30
(2006) 615 − 626.
[6] L. Liu, Global stability in a tuberculosis model incorporating two latent peri-
ods, Int. J. Biomath. 2 (2009) 357 − 362.
[7] G. P. Sahu and J. Dhar, Analysis of an SVEIS epidemic model with partial
temporary immunity and saturation incidence rate, Appl. Math. Model. 36
(2012) 908 − 923.
[8] J. P. Tian and J. Wang, Global stability for cholera epidemic models, Math.
Biosci. 232 (2011) 31 − 41.
[9] Tian, Canrong, Qunying Zhang, and Lai Zhang. "Global stability in a net-
worked SIR epidemic model." Applied Mathematics Letters 107 (2020):
106444.
24 Zakaria Yaagoub, Jaouad Danane and Karam Allali

[10] Elkaranshawy, Hesham A., Hossam M. Ezzat, and Nermeen N. Ibrahim. "Lya-
punov function and global asymptotic stability for a new multiscale viral
dynamics model incorporating the immune system response: Implemented
upon HCV." Plos one 16.10 (2021): e0257975.
[11] Shigui Ruan and Wendi Wang. Dynamical behavior of an epidemic model with
a nonlinear incidence rate. Journal of Differential Equations, 188(1):135−163,
2003.
[12] Herbert W Hethcote and P Van Den Driessche. Some epidemiological models
with nonlinear incidence. Journal of Mathematical Biology, 29(3):271 − 287,
1991.
[13] WR Derrick and P Van Den Driessche. A disease transmission model in a non
constant population. Journal of Mathematical Biology, 31(5):495 − 512, 1993.
[14] Herbert W Hethcote and P Van Den Driessche. Some epidemiological models
with nonlinear incidence. Journal of Mathematical Biology, 29(3):271 − 287,
1991.
[15] Wei-min Liu, Simon A Levin, and Yoh Iwasa. Influence of nonlinear incidence
rates upon the behavior of sirs epidemiological models. Journal of mathemat-
ical biology, 23(2):187 − 204, 1986.
[16] Abdullahi Baba, Evren Hincal, Global stability analysis of two-strain epidemic
model with bilinear and non-monotone incidence rates, Eur. Phys. J. Plus
(2017) 132: 208.
[17] Isa Abdullahi Baba, Evren Hincal and Sultan Hamed Khalifa Alsaadi, Global
stability analysis of a two-strain model with awareness, Advances in Differen-
tial Equations and Control Processes, 19(2)(2018), 83 − 100.
[18] Dounia Bentaleb, Saida Amine, Lyapunov function and global stability for a
two-strain SEIR model with bilinear and nonmonotone incidence, International
Journal of Biomathematics 2019; 12(2): 1950021.
[19] Adil Meskaf, Omar Khyar, Jaouad Danane and Karam Allali, Global stability
analysis of a two-strain epidemic model with non-monotone incidence rates,
Chaos, Solitons and Fractals 133 (2020) 109647.
[20] Khyar, Omar, and Karam Allali. "Global dynamics of a multi-strain SEIR
epidemic model with general incidence rates: application to COVID-19 pan-
demic." Nonlinear Dynamics 102.1 (2020): 489-509.
[21] S. Bowong, Optimal control of the transmission dynamics of tuberculosis,
Nonlinear Dyn. 61 (2010), 729 − 748.
[22] C. Castillo-Chavez, Z. Feng, To treat or not to treat: the case of tuberculosis,
J. Math. Biol. 35(1997), 629 − 656
[23] Kar, T. K., & Batabyal, A. (2011). Stability analysis and optimal control of an
SIR epidemic model with vaccination. Biosystems, 104(2-3), 127 − 135.
[24] Y. Emvudu, R. Demasse, D. Djeudeu, Optimal control of the lost to follow up
in a tuberculosis model, Comput. Math. Meth. Med. 2011 (2011), 398476.
[25] Yusuf, T. T., & Benyah, F. (2012). Optimal control of vaccination and treatment
for an SIR epidemiological model. World journal of modelling and simulation,
8(3), 194 − 204.
1 Global stability analysis of two-strain SEIR epidemic model with quarantine strategy 25

[26] K. Hattaf, M. Rachik, S. Saadi, Y. Tabit, N. Yousfi, Optimal control of tuber-


culosis with exogenous reinfection, Appl. Math. Sci. 3 (2009), 231 − 240.
[27] Laaroussi, A. E. A., Rachik, M., & Elhia, M. (2018). An optimal control
problem for a spatiotemporal SIR model. International Journal of Dynamics
and Control, 6(1), 384 − 397.
[28] Kandhway, K., & Kuri, J. (2014). How to run a campaign: Optimal control of
SIS and SIR information epidemics. Applied Mathematics and Computation,
231, 79 − 92.
[29] A. El Bhih, Y. Benfatah, S. Ben Rhila, M. Rachik, A. El Alami Laaroussi,
A Spatiotemporal Prey-Predator Discrete Model and Optimal Controls for
Environmental Sustainability in the Multifishing Areas of Morocco, Discr.
Dyn. Nat. Soc. 2020 (2020), 2780651.
[30] Zaman, G., Kang, Y. H., & Jung, I. H. (2008). Stability analysis and optimal
vaccination of an SIR epidemic model. BioSystems, 93(3), 240 − 249.
[31] E. Jung, S. Lenhart, Z. Feng, Optimal control of treatments in a two-strain
tuberculosis model, Discr. Contin. Dyn. Syst. Ser. B. 2 (2002), 473 − 482.
[32] Zaman, G., Kang, Y. H., & Jung, I. H. (2009). Optimal treatment of an SIR
epidemic model with time delay. BioSystems, 98(1), 43 − 50.
[33] Ledzewicz, U., & Schattler, H. (2011). On optimal singular controls for a
general SIR-model with vaccination and treatment. In Conference Publications
(Vol. 2011, No. Special, p. 981). American Institute of Mathematical Sciences.
[34] Djidjou-Demasse, Ramses, et al. "Optimal COVID-19 epidemic control until
vaccine deployment." MedRxiv (2020).
[35] Perkins, T. Alex, and Guido EspaÃśa. "Optimal control of the COVID-19
pandemic with non-pharmaceutical interventions." Bulletin of Mathematical
Biology 82.9 (2020): 1-24.
[36] Choi, Wongyeong, and Eunha Shim. "Optimal strategies for social distancing
and testing to control COVID-19." Journal of theoretical biology 512 (2021):
110568.

You might also like