Accepted Manuscript

A Review on Optimization of Composite Structures Part II: Functionally Graded

Materials

S. Nikbakht, S. Kamarian, M. Shakeri

PII: S0263-8223(18)33147-7
DOI: [Link]
Reference: COST 10622

To appear in: Composite Structures

Received Date: 28 August 2018

Revised Date: 20 December 2018
Accepted Date: 29 January 2019

Please cite this article as: Nikbakht, S., Kamarian, S., Shakeri, M., A Review on Optimization of Composite
Structures Part II: Functionally Graded Materials, Composite Structures (2019), doi: [Link]
[Link].2019.01.105

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Review

A Review on Optimization of Composite Structures Part II:

Functionally Graded Materials

S. Nikbakhtc, S. Kamariana,b,c*, M. Shakeric

a. Department of Mechanical Engineering, Ilam Branch, Islamic Azad University, Ilam, Iran

b. Young Researchers Club, Ilam Branch, Islamic Azad University, Ilam, Iran

c. Department of Mechanical Engineering, Amirkabir University of Technology, Tehran, Iran

Abstract
Functionally Graded (FG) structures are a novel design through which the material properties vary
smoothly and this feature leads these structures to have better mechanical or thermal performances.
They are mostly constituted from two or more materials with gradually varying volume fraction
distribution. Most of the publications on the optimization of laminated composite structures were
investigated in the first part of the authors’ review paper [1]. In this research which acts as the
second part, the majority of publications on optimization of FG structures are reviewed. In addition
to FG beams, plates and shells, various structures such as tubes, implants, rotating disks, sport
instruments, etc. are investigated. Furthermore, the key outputs of each publication are represented
to make this article an asset source for mechanical engineers since there has not been any
comprehensive review article on optimal designs of FG structures in the literature.
Keywords: Optimization; Composite structures; Functionally graded materials

Contents
1. Introduction ………………………………………………………………….. 2
1.1. Volume fraction distribution functions 2
………………………………………
1.2. Material properties variation functions 5
………………………………………
1.3. Motivation 6
…………………………………………………………………...
2. Functionally graded beams …………………………………………………... 9
2.1. Transversely functionally graded beams 9
…………………………………….
2.2. Axially functionally graded beams 12
…………………………………………..
2.3. Functionally graded sandwich beams 13
………………………………………..
2.4. Functionally graded porous beams 14
…………………………………………..
3. Functionally graded plates …………………………………………………… 14
3.1. Transversely functionally graded plates 15
……………………………………..
3.2. 2D functionally graded plates 18
………………………………………………..
3.3. Functionally graded sandwich plates 19
………………………………………...
3.4. Step-formed functionally graded plates 20
……………………………………...
3.5. Multi-layered plates containing functionally graded layers 21
………………….
3.6. Functionally graded carbon nanotube plates 23
…………………………………

1
 3.7. Functionally graded foam-filled plates 23
………………………………………
3.8. Other functionally graded plates 24
……………………………………………..
4. Functionally graded shells …………………………………………………… 24
4.1. Functionally graded cylindrical and conical shells 25
…………………………..
4.2. Functionally graded vessels and pipes 27
……………………………………….
4.3. Functionally graded spheres 28
…………………………………………………
5. Functionally graded tubes ……………………………………………………. 28
5.1. Functionally graded foam-filled tubes 29
……………………………………….
5.2. Functionally graded thickness tubes 31
…………………………………………
5.3. Functionally graded thickness inversion tubes 32
………………………………
6. Other functionally graded structures ………………………………………… 33
7. Conclusion …………………………………………………………………… 39
References …………………………………………………………………… 40
1. Introduction
In 1980s, a group of materials scientists in Japan introduced a new class of materials called
Functionally Graded Materials (FGMs) resistant to high temperatures and suitable for many
engineering applications, such as thermal barrier coatings, engine components or rocket nozzles.
FGMs are advanced composite materials whose equivalent properties vary gradually along one (or
more) direction(s), usually in the thickness direction, to obtain modified response to external
loadings. FG structures are used in a variety of engineering applications including aircraft,
construction and transportation where strong, stiff and light structures are required. The high
specific stiffness and strength of these structures along with their low weight can be mentioned as
the advantages of these materials over their homogeneous counterparts. Furthermore, in contrast
to conventional composites whose steep shifts in material properties cause the problem of
interfacial stresses, the strategy of changing the geometry or the volume fraction distribution
gradually through the FG structure avoids this problem. In the simplest form of FGMs, two
different material components change gradually from one surface to the other according to a certain
function. The most familiar FGM is compositionally graded from a refractory ceramic to a metal
(Fig 1).

2
 Fig.1. A FGM slab with volume fraction distribution through the thickness direction

The variation in material properties of FGM is examined via two types of functions. The first set
of functions are related to the volume fraction distribution of these materials. The second group of
functions are associated with the relationship between material gradation and the material
properties variation through FGMs.

1.1. Volume fraction distribution functions

The majority of FG structures consist of two phases whose volume fractions vary gradually.
Therefore, the functions employed for determining the volume fraction distribution of material
constituents of these structures are of significant importance. The most applicable formulas
regarding material gradation through FGMs are given below.

𝑝
Power-law model:
𝑉1 = +( )
1 𝑧
2 ℎ Eq.1
Where 𝑝 is the power-law index which defines the level of material inhomogeneity through the
FG structure.

𝑝
Trigonometric model: 2
𝑉1 = 𝑆𝑖𝑛 [( ) ]
1 𝑧
+
2 ℎ Eq.2

Viola-Tornabene three-parameter model: 𝑐 𝑝

(
1 𝑧 1 𝑧
𝑉1 = ‒ + 𝑏 +
2 ℎ 2 ℎ ( )) Eq.3

Viola-Tornabene four-parameter model: 𝑐 𝑝

( ( ) ( ))
1 𝑧 1 𝑧
𝑉1 = 1 ‒ 𝑎 ‒ + 𝑏 +
2 ℎ 2 ℎ
Eq.4

3
Where 𝑎, 𝑏 and 𝑐 are constant parameters which dictate the material propagation in the FG
structure.

Four-parameter model: 𝛽 𝛾

[ () ()]
𝑉1 = 𝐶 1 ‒
𝑧
ℎ
+𝛼
𝑧
ℎ
Eq.5

𝛾
Five-parameter trigonometric model:
[
1 𝛼
𝑉1 = 𝐶 ‒ sin
2 2
𝜂𝜋𝑧
ℎ
+𝜑 ( )] Eq.6.a

𝛾
Five-parameter trigonometric model
for 2D FGM: [ | (
𝑉1 = 1 ‒ sin
𝜂𝑥𝜋𝑥
𝐿𝑥 ) (
+ 𝜑𝑥 sin
𝜂𝑦𝜋𝑦
𝐿𝑦
+ 𝜑𝑦 )|] Eq.6.b

Where 𝐶, 𝛼, 𝛽, 𝛾, 𝜂 and 𝜑 are controlling parameters which determine the material variation pattern
through the FG structure. Two other functions for volume fraction distribution of 2D FG plates
have been recently introduced by Lieu et al. [2], [3] as follows.
𝑘𝑥 𝑘𝑦
𝑉1 = ()()
𝑥
𝑎
𝑦
𝑏 Eq.7

𝑉1 = (𝑉1,𝑥)(𝑉1,𝑦) Eq.8.a

{
𝑘𝑥

𝑉1,𝑥 =
( 2𝑥
𝑎 ) ,0 ≤ 𝑥 ≤
𝑎
2
𝑘𝑥

( 𝑎 ) ,2 ≤ 𝑥 ≤ 𝑎
2 ‒
2𝑥 𝑎 Eq.8.b

{( )
𝑘𝑦

𝑉1,𝑦 =
( )2𝑦
𝑏
,0 ≤ 𝑦 ≤
𝑏
2
𝑘𝑦
2𝑦 𝑏 Eq.8.c
2‒ , ≤𝑦≤𝑏
𝑏 2

Where 𝑎 and 𝑏 are the dimensions of FG plates in 𝑥 and 𝑦 directions, respectively. 𝑘𝑥 and 𝑘𝑦 are
the constants which define the level of material inhomogeneity in the in 𝑥 and 𝑦 directions, respectively.

4
 {(
𝑝
Sigmoid model: 1

𝑉1 =
1‒ 1‒
2 (2𝑧
ℎ
;0 ≤ 𝑧 ≤ )
ℎ
2
Eq.9
𝑝
1 2𝑧
2
1+
ℎ
ℎ
)
;‒ ≤𝑧≤0
2

There are a number of volume fraction approximating functions defined for FG structures
considering a certain number of control points through them. In other words, according to these
models, FGMs are assumed as multi-layered structures and the volume fraction distribution is
determined in each layer individually according to the volume fraction and/or the slope of volume
fraction gradation at the top and bottom of each layer. These functions are given as follows.
𝑛
B-Spline Basis Function:
𝑉1(𝜉) = ∑𝐵 𝑖,𝑝(𝜉)𝑉1,𝑖(𝑧) Eq.10
𝑖=1

Where 𝑉1,𝑖(𝑧), p and n are the volume fraction of the first phase at the ith control point, the number
of basis function and the polynomial order, respectively. 𝐵𝑖,𝑝(𝜉) are the B-Spline basis functions
defined via different formula and can be found in the literature.
Piece-wise cubic
interpolation 𝑉1(𝑧) = 𝑉𝑖𝐻1(𝑧) + 𝑆𝑖𝐻2(𝑧) + 𝑉𝑖 + 1𝐻3(𝑧) + 𝑆𝑖 + 1𝐻4(𝑧);𝑧𝑖 ≤ 𝑧 ≤ 𝑧𝑖 + 1 Eq.11
function:

Where 𝑉𝑖 and 𝑆𝑖 are the volume fraction and slope of volume fraction at the ith control point,
respectively. These values along with Hermite basis functions 𝐻𝑘(𝑧) can be found in a study
represented by Vel and Pelletier [4].
In all of the above mentioned functions, 𝑉1 is the volume fraction of the first phase and the volume
fraction of the second phase can be obtained as
𝑉2(𝑧) = 1 ‒ 𝑉1(𝑧) Eq.12

It is to be noted that all the above equations are defined in Cartesian coordinate system and it is
assumed that the volume fraction of each phase varies through the thickness direction (Fig. 1).
These formulations can be calculated for axially FG beams in which the volume fraction of
materials vary through longitudinal direction by applying a simple mathematical technique. For
FG hollow cylinders, these equations should be defined in cylindrical coordinate system. In order
to avoid any redundancy, the authors decided to only mention the functions related to plates in
which the material constituents vary through their thickness.

1.2. Material properties variation functions

Regarding the aforementioned volume fraction distribution functions, equivalent material
properties of FGMs could be determined by various functions. In this sub-section, a number of the

5
most important formulas in terms of material properties variation in FGMs are introduced. In order
to avoid any repetitive information, only the formulas related to the elastic modulus gradation are
mentioned here.
Rule of Mixture (ROM): 𝐸 = 𝐸1𝑉1 + 𝐸2𝑉2 Eq.13

Mori-Tanaka Scheme [5]: 𝐾 ‒ 𝐾1 𝑉2

= Eq.14.a
𝐾2 ‒ 𝐾1 3𝑉1(𝐾2 ‒ 𝐾1)
1+
(3𝐾1 + 4𝐺1)
𝐺 ‒ 𝐺1 𝑉2
= Eq.14.b
𝐺2 ‒ 𝐺1 𝑉1(𝐺2 ‒ 𝐺1)
1+
(𝐺1 + 𝑓1)
𝐺1 9𝐾1 ‒ 8𝐺1)
(
𝑓1 = Eq.14.c
6(𝐾1 + 2𝐺1)
9𝐾𝐺
𝐸= Eq.14.d
3𝐾 + 𝐺
3𝐾 ‒ 2𝐺
𝜈= Eq.14.e
2(3𝐾 + 𝐺)

Where 𝐾, 𝐺 and 𝜈 are the local bulk modulus, local shear modulus and Poisson’s ratio, respectively.
Exponential function [6]: 𝐸(𝑧) = 𝐴𝑒𝑥𝑝(𝐵𝑧) Eq.15

TTO model: The name of this model which was firstly represented by Tamura et al. [7] indicates
the initials of the names of the researchers who introduced it. The basic assumption of this model
is that the magnitude of stresses and strains in each area of a two-phased material (𝜎, 𝜀) depends
on the stresses and strains each phase undergoes (𝜎1, 𝜎2, 𝜀1, 𝜀2) and their volume fractions (𝑉1, 𝑉2).
𝜎 = 𝜎1𝑉1 + 𝜎2𝑉2 Eq.16.a
𝜀 = 𝜀1𝑉1 + 𝜀2𝑉2 Eq.16.b

𝑉1𝐸1( 𝑞 + 𝐸2
𝑞 + 𝐸1) + 𝑉2𝐸2
Eq.16.c
𝐸=
𝑉1(𝑞 + 𝐸1 )
𝑞 + 𝐸2
+ 𝑉2

𝜎1 ‒ 𝜎2
𝑞= ;0 ≤ 𝑞 ≤ ∞ Eq.16.d
|𝜀1 ‒ 𝜀2|

Where 𝑞 is the stress to strain ratio. This parameter was proved to depend on the type of materials
and the microstructural interactions between the two phases. TTO model is also capable of

6
determining the yielding strength of FGMs as well as other material properties such as Young’s
modulus and thermal expansion coefficient.

[ ( )( )
𝜎𝑌 = 𝜎𝑌1 𝑉1 +
𝑞 + 𝐸1 𝐸2
𝑞 + 𝐸2 𝐸1
(1 ‒ 𝑉1) ] Eq.17

Where 𝜎𝑌1 is the yielding strength of the ductile phase. Fig. 2 clearly demonstrates the yielding
strength variation of a FGM constituted of metal (ductile phase) and ceramic (brittle phase).

Fig. 2. The stress-strain curves for an arbitrary two phased composite [8]

Bhattacharyya et al. [9] conducted an experimental study and claimed that TTO model
outperformed ROM in determining material properties of FGMs. Nikbakht et al. [10] and
Komarsofla et al. [11] exploited TTO scheme to obtain the yielding initiation of FG plates and
shells, respectively. More detailed information about this model could be found in the mentioned
references.

1.3. Motivation
Technological advances in engineering designs motivated many scientists to investigate the
mechanical behavior of novel materials such as FGMs under static, dynamic and impact loadings.
The huge number of studies in these fields convinced researchers to compile review articles
regarding FGMs with different perspectives. Liew et al. [12] presented a review paper in the area
of static, dynamic, free vibration and buckling analyses of FG structures with meshless methods.
Their main concentration was on Element Free Galerkin (EFG) and reproducing kernel particle
methods. The cons and pros of each study were represented and possible future works were
suggested in their article as well. A critical review on FG plates was carried out by Jha et al. [13]
in which the most publications in the field of vibration and thermo-elastic analyses of these
structures were reviewed. As a conclusion, the researchers compared 2D and 3D methods in terms
of their accuracy and computational costs. A bulk of publications associated with FG plates and

7
shells subjected to thermal and mechanical loads were gathered in a comprehensive review article
represented by Thai and Kim [14]. The main focus of their paper was on the implemented theories.
Classical, firs order shear deformation theories, higher order shear deformation theories and 3D
elasticity theory employed for analyzing FG structures were also detailed thoroughly in their
research. Compiling a review article on FG plates and shells, Gupta and Talha [15] discussed
different fabrication processes of FGMs, different adopted theories for static and dynamic analyses
of FG structures and the recent application of these novel materials. They also represented a
number of suggestions for future studies regarding FGMs. Swaminathan and Sangeetha [16]
attempted to gather most of the published researches on thermal analysis of FG structures under
various types of thermal loadings. After reviewing the studies regarding static, vibrational and
buckling analyses of FG structures, these authors compared different solution methods and
mentioned their positive and negative points.
Albeit numerous review studies on mechanical behavior of FG structures, lack of a comprehensive
review article on optimization of these materials can be seen in the literature. To this end, in the
present work, researches on the optimum design of various types of FG structures are reviewed
and the most prominent results of each publication are highlighted. According to the structural
classification in this research, firstly the publications related to FG beams are reviewed. Following
this, two sections are devoted to publications on optimization of FG plates and shells.
Subsequently, FG tubes which are advanced crash structures with vast application in industries are
investigated in the next section. Finally, other structures such as implants, disks, sport instruments,
etc. made of FGMs are detailed in the last section. For the convenience of the readers, all the
abbreviations exploited in this review article are gathered in Table 1. in alphabetical order.

Table 1. The alphabetically ordered abbreviations

Abbreviation Stands for

AEDEA Adaptive Elicit Differential Evolution Algorithm
AHEFA Adaptive Hybrid Evolutionary Firefly Algorithm
ANFIS Adaptive Neuro-Fuzzi Interference System
ANN Artificial Neural Network
APDL ANSYS Parametric Design Language
ASA Adaptive Simulated Annealing
BESO Bi-directional Evolutionary Structural Optimization
CAMD Continuous Approximation of Material Distribution
CG Conjugate Gradient
CNT Carbon Nanotube
CPSO Co-evolutionary Particle Swarm Optimization
DAF Dynamic Amplification Factor
DE Differential Evolution
DMS Direct Multi-Search
DMO Discrete Material Optimization
DNN Deep Neural Network
DOE Design of Experiments
DQM Differential Quadrature Method
EFG Element Free Galerkin
FA Firefly Algorithm
FAIPA Feasible Arc Interior Point Algorithm

8
FEM Finite Element Method
FG Functionally Graded
FGCNT Functionally Graded Carbon Nanotube
FGF Functionally Graded Foam
FGFCS Functionally Graded Foam-filled Cellular Structure
FGFGT Functionally Graded Foam-filled Graded Thickness
FGFTT Functionally Graded Foam-filled Tapered Tubes
FGH Functionally Graded Honeycomb
FGM Functionally Graded Materials
FGPT Functionally Graded Piezoelectric Transducers
FGS Functionally Graded Sandwich
FGT Functionally Graded Thickness
FGTIT Functionally Graded Thickness Inversion Tube
FLGFT Functionally Lateral Graded Foam-filled Tubes
GA Genetic Algorithm
GDO Goal Driven Optimization
GLODS Global and Local Optimization using Direct Search
GS Golden Section
HAP Hydroxyapatite
HFT Heterogeneous Feature Tree
HSDT Higher order Shear Deformation Theory
ICA Imperialistic Competitive Algorithm
MBB Messerschmitt-Bolkow-Blohm
MC Monte-Carlo
MIGA Multi-Island Genetic Algorithm
MOGA Multi-Objective Genetic Algorithm
MOPSO Multi-Objective Particle Swarm Optimization
mSOS modified Symbiotic Organisms Search
MWCNT Multi-Walled Carbon Nanotube
NN Neural Network
NSGA Non-dominated Sorting Genetic Algorithm
NURBS Non-Uniform Rational Basis Spline
PCF Peak Crushing Force
PRS Polynomial Response Surface
PSACO Particle Swarm with Ant Colony Optimization
PSO Particle Swarm Optimization
PSOPC Particle Swarm Optimization with Passive Congregation
PSPCACO Particle Swarm with Passive Congregation and Ant Colony Optimization
PSOSM Particle Swarm Optimization Sliding Mode
PVDF Polyvinylidene Fluoride
PZT Lead Zirconate Titanate
RBDO Reliability Based Design Optimization
RBF Radial Basis Function
RCGA Real Coded Genetic Algorithm
ROM Rule of Mixture
RSM Response Surface Methodology
SA Simulated Annealing
SCM Simple Cell Mapping
SEA Specific Energy Absorption
SGO Social Group Optimization
SLP Sequential Linear Programming
SMOSA Suppapitnarm Multi-Objective Simulated Annealing
SOS Symbiotic Organisms Search
SQP Sequential Quadratic Programming
SVR Super Vector Regression
TDFGFT Two Dimensional Functionally Graded Foam-filled Tube

9
 TSM Traditional Sliding Mode
UF Uniform Foam-filled
UFCS Uniform Foam-filled Cellular Structure
UHF Uniform Honeycomb Filled

2. Functionally graded beams

Beams are one of the main structures in mechanical engineering that primarily resist loads applied
perpendicular to their longitudinal axis. They are traditionally descriptions of civil engineering
structural elements. However, many primary and secondary structural elements such as helicopter
rotor blades, turbine blades, robot arms and space erectable booms can be assumed as beams in
accordance with their geometry. As beams are vastly implemented in industries, a wide range of
materials, from homogeneous to composite and FGM, are utilized to construct them. FG beams
are divided into two main categories with regard to the direction of material variation. These two
categories are referred to as transversely and axially FG beams which can be seen in Fig. 3. Many
researchers have investigated optimization of FG beams for different objectives which are
mentioned in this section.

Fig.3. Schematic of FG beams. (a) Transversely FG beam. (b) Axially FG beam.

2.1. Transversely functionally graded beams

Goupee and Vel [17] represented an optimization approach combining EFG and Real Coded
Genetic Algorithm (RCGA) to reveal the results associated with the fundamental frequency
maximization of three different models of FG beams and to show the outperformance of these
materials over their monotonic counterparts. They proved that significantly larger amounts of
fundamental frequencies could be achieved for beams made up of copper/tungsten FGM compared
to those homogeneous materials made of copper or tungsten. Moreover, adding geometrical
parameters to the design variables of their study resulted in more light-weight structures. Almeida
et al. [18] implemented Continuous Approximation of Material Distribution (CAMD) method for
the topology optimization of FG beams subjected to thermal and mechanical loads. Their aim was
to find the best material distribution for such materials which resulted in the minimum compliance
10
of the beams taking into consideration symmetry and pattern repetition constraints. In order to
consider the safety of mechanical structures, scientists input safety factor in the formulation of
their optimization approaches. These methods however were found to be too conservative or too
unsafe. To tackle this problem, Reliability Based Design Optimization (RBDO) method is
suggested which restrains the design to a certain reliability level. This method in the framework of
Sequential Quadratic Programming (SQP) was utilized by Noh et al. [19] to minimize the thermal
stresses of FG beams subjected to thermal loads. These researchers managed to decrease the
thermal stress of such structures significantly regardless of the reliability. Subsequently, adding
RBDO method to their calculations led to a slightly higher amount of thermal stresses while the
reliability of the beams increased noticeably. Different structures with various shapes constituted
from FGM were optimized by Taheri and Hassani [20] in an attempt to modify the structural
eigenfrequency. The mass of an FG beam with clamped-clamped boundary conditions was
minimized with eigenfrequency as the constraint. Next, the natural frequency of FG cylinders with
different cross-sections was maximized. Subsequently, the difference between the first and second
eigenfrequencies of a square plate with different types of holes in the middle was maximized.
Finally, a particular eigenfrequency of a complex geometry made of FGM was increased
considering the mass as the constraint. In all cases above, both material distribution and shape of
the structures were selected as the design variables, and the researchers revealed that inserting the
shape as an additional design variable could lead to better optimized results. Combining
Imperialistic Competitive Algorithm (ICA) with Artificial Neural Network (ANN), Yas et al. [21]
improved the speed of an optimization approach in which the material variation profile of FG
beams was optimized. They demonstrated the efficiency of this hybrid method comparing the
optimal results with those of GA. Subsequently, Kamarian et al. [22] proved the outperformance
of the integration of Firefly Algorithm (FA) and Adaptive Neuro-Fuzzi Interference System
(ANFIS) over hybrid algorithms such as ICA integrated with ANN or Genetic Algorithm (GA)
combined with ANN in optimizing the material distribution of FG beams resting on a two-
parameter elastic foundation. The objective function of both studies was to maximize the
fundamental frequency of the beams and a three-parameter model was employed to examine the
material distribution through the thickness of beams. Roque and Martins [23] implemented
Differential Evolution (DE) optimization method to obtain the optimum material distribution of
FG beams which led to the maximum first fundamental frequency of these structures choosing the
volume fraction of one of the material constituents as the constraint. For this purpose, they
investigated different materials and the influence of the ratio of material constituents’ Young’s
moduli on the optimum design. Using a three-parameter material model referred to as Viola-
Tornabene formulation helped them to include a wider variety of material distribution in their
study. In another work carried out by Roque et al. [24], the same material model and optimization
methodology were employed along with a modified couple stress theory to minimize the free
vibration frequency of the Timoshenko FG nano beams. The researchers concluded that the
optimal parameters of the applied material model did not depend on the scale parameter and
boundary conditions. Trying to increase accuracy and decrease the computational time of an
optimization approach regarding the fundamental frequency of FG nano beams, Vosoughi and
Darabi [25] combined GA and Conjugate Gradient (CG) method. These researchers took the
volume fraction distribution along with the small scale parameter as the design variables of their

11
study. They also revealed results associated with FG nano beams with different boundary
conditions and proved the efficiency and accuracy of their methodology. Pham et al. [26]
employed the novel Social Group Optimization (SGO) algorithm to maximize the first natural
frequency of FG beams. Three different material models with 3, 4 and 5 variables were utilized in
their study and the last case was demonstrated to be the most efficient in representing the best
optimal material distribution leading to maximum fundamental frequency. Taati and Sina [27]
employed NSGA-II to optimize the thickness and the level of inhomogeneity in FG-micro beams
resting on elastic foundation. The objective functions of their study were to minimize the
maximum stress and deflection and to maximize the fundamental frequency and critical buckling
load. The coefficient of elastic foundation and internal length scale parameter were found to have
noticeable influence on the optimal design. Selecting lateral and flexural-torsional buckling of I-
shaped FG beams as the objective function, Nguyen and Lee [28] conducted a GA-based
optimization approach in which the material distribution and geometrical parameters were the
design variables. A piece-wise cubic interpolation function was utilized in their study to examine
the material distribution through the thickness of the beams. The researchers claimed that although
both geometrical and material parameters are influential on buckling of FG beams, the latter was
more dominant for shorter beams. Composite box beams with variable volume fraction of fibers
and ply thickness were optimized by Maalavi [29] using MATLAB optimization toolbox. The
volume fraction of materials was assumed to vary according to a power-law function. Maximizing
the natural frequency and harnessing them to certain target values were the key objective functions,
and the fiber orientation angle and volume fraction along with the ply thickness distribution formed
the design variables.
With the aim of minimizing the compliance of FG beams and bridge structures, Xia and Wang
[30] developed a level set based optimization approach and found the optimal topology and volume
fraction distribution of these structures. The material was assumed to vary through these structures
from a soft material to a stiff material smoothly. According to their results, although the FG designs
had more compliance than their homogeneous counterparts made of the stiff material, FG
structures contained less volume fraction of stiff material and this could lead to a design with lower
cost. Maleki Jebeli and Shariat Panahi [31] integrated their Bi-directional Evolutionary Structural
Optimization (BESO) method with GA to prove the outperformance of their methodology over
that of Xia and Wang [30] when the design variables, objective function and the type of structures
in the two aforementioned researches were the same. Less computational time, less minimized
compliance and design manufacturability after the optimization process were of virtues of their
approach over Xia and Wang’s.
Multi-material L-shaped brackets and Messerschmitt-Bolkow-Blohm (MBB) beams constituted
from three material phases were optimized by Taheri and Suresh [32] utilizing perimeter
penalization technique. As the volume fraction of material phases varied through these structures
smoothly, they were considered as FG structures. However, there was no regular material
gradation pattern in the optimized results. The total mass and volume fraction of particular phases
were chosen as the design variables in this topology optimization research whose aim was to
minimize the compliance of the mentioned beams and brackets. The authors noticed that this
technique led to more efficient and manufacturable designs than the other conventional methods

12
in the literature. A comprehensive multi-resolution topology optimization approach based on
isogeometric analysis on different types of structures namely MBB beams, cantilever beams,
bridge structures, curved beams and L-shaped beams was conducted by Lieu and Lee [33]. All of
these structures were multi-material and the objective function was to minimize their compliance.
The researchers changed each multi-phase optimization problem containing several volume
fraction constraints into a number of 2-phase sub-problems with one volume fraction constraint.
Every sub-problem was subsequently solved by optimality criteria method. The authors finally
demonstrated the outperformance of their methodology over conventional topology approaches in
terms of computational costs. The weight of FG lattice beams and L-shaped brackets were
minimized in a topology optimization approach conducted by Cheng et al. [34] with the stress as
the constraint. The researchers implemented a modified Hill’s yield criterion and asymptotic
homogenization method to obtain the yielding strength and effective elastic properties of these
structures, respectively. After conducting experimental tests, they proved the accuracy of the
utilized scheme. According to their results the optimized FG structures showed significantly better
elastic properties than their uniform counterparts.

2.2. Axially functionally graded beams

Alshabatat and Naghshineh [35] utilized Finite Element Method (FEM), lumped parameter model
and GA to represent the optimal design of clamped and cantilever axially FG beams with the aim
of minimizing the sound power radiation and maximizing the fundamental frequency of these
structures. Four and five-parameter trigonometric models were represented to determine the
material distribution of the beams. These models were proved to be more flexible than
conventional power-law material models. Axially FG cantilever polymer composite beams
reinforced by Multi-Walled Carbon Nanotube (MWCNT) were optimized by Rokni et al. [36] to
maximize the fundamental frequency. The proportion of Carbon Nanotube (CNT) through axial
direction was approximated by an exponential function, and different indices were investigated to
obtain the best CNT distribution regarding this objective function. According to the final results,
the optimized FG distribution of these structures could lead to nearly 11% higher values of
fundamental frequency than that of the uniformly distributed design while both contained 1 wt.%
of MWCNT. The combination of Particle Swarm Optimization (PSO) method and Simple Cell
Mapping (SCM) helped He and Sun [37] to optimize the material distribution of axially FG beams.
The objective functions were to achieve a design with minimum total mass and maximum radiated
sound power with upper and lower bounds of the desired frequencies as the constraints. The
authors also observed that the hybrid SCM integrated with an evolutionary algorithm outperformed
an individual evolutionary algorithm. The combination of isogeometric analysis based on 2D Non-
Uniform Rational Basis Spline (NURBS) functions with DE method helped Kim et al. [38] to
optimize the material distribution of 2D FG beams. The purpose of their study was to maximize
the first three fundamental frequencies. Investigating the influence of different parameters on the
optimal design of 2D FG beams, they reached two important conclusions. Firstly, the most
significant influence of the optimization process can be seen in clamped-free boundary conditions
while the increase in the fundamental frequency of simply supported beams were less than those

13
with other boundary conditions such as clamped-clamped and clamped-simply supported.
Secondly, the slenderness has negligible influence on the maximization of fundamental frequency
of the beams.

2.3. Functionally graded sandwich beams

Sandwich structures are implemented in various engineering applications where high strength and
lightweight structures are required. The main disadvantage of these structures is the mismatch of
stiffness properties between their core and face sheets. This causes a certain type of damage
referred to as debonding particularly when the sandwich beam or plate is subjected to impact
loadings. In order to decline the vulnerability of sandwich structures to this type of failure,
scientists have combined the concept of FGMs with sandwich structures and this has brought them
a new type of structures named Functionally Graded Sandwich (FGS) structures. There are two
main types of FGS structures. The first one is a FG core sandwiched between two homogeneous
face sheets. In the second type, a homogeneous core is sandwiched between two FG face sheets.
Both of these structures are depicted in Fig.4.

Fig.4. Schematic of FGS structure. (a) Type-I. (b) Type-II.

In recent years, many researchers have been motivated to optimize the FGS structures. In order to
illustrate the applicability of PSO method in optimizing FGM and exploiting re-initialization
strategy in this very algorithm, Loja [39] attempted to optimize FGS beams using this algorithm.
The objective function of this research was to maximize the bending stiffness of these beams. To
this end, two case studies with different sets of design variables were chosen. In the first case
which was a symmetric sandwich beam, the thickness of ceramic core and the material distribution
of FG face sheets were the design variables. The second case was an asymmetric FGS beam and
the thickness and material distribution of face sheets were taken as the design variables. The author
finally noted the importance of re-initialization in the process of PSO method. The material
distribution of FGS beams and plates was optimized by Shi and Shimoda [40] to minimize their
compliance using traction method. In these structures, face sheets and cores were constituted of
metal and ceramic, and the shape of interfaces between the core and face sheets were assumed to

14
vary according to a shape gradient function, considering the total volume of the structure as the
constraint. This function was the design variable of their study. The authors realized that the larger
the differences between the Young’s moduli of the ceramic and metal phases, the more
optimization procedure could lead to lower values of compliance. In a comprehensive optimization
study provided by Correia et al. [41] FGS plates with FG cores were optimized considering several
conflicting objective functions such as weight and cost minimization, natural frequency
maximization and stress reduction. The thickness of core and face sheets along with the material
inhomogeneity were chosen as the design variables while manufacturing limitations were also
taken into account. GLODS and DMS were the implemented optimization methods in their study.

2.4. Functionally graded porous beams

Porous materials which bring engineers lightweight designs with advanced performances are
widely utilized in different sections of industries these days. Engineers are also inspired by the
primary natural porous bones and have succeeded in modifying the characteristics of porous
materials even more by introducing the FG porosity idea and optimizing the porosity pattern in
these materials. Implementing Non-dominated Sorting Genetic Algorithm (NSGA-II), Beluch and
Hatlas [42] managed to optimize the porosity distribution of FG porous cantilever beams subjected
to vertical loads. The objective functions of this multi-objective optimization procedure were
minimizing the maximum deflection and maximizing the total porosity of the beams. A polynomial
function was used for determining the porosity variation in the mentioned structure. The optimal
results proved that the minimized maximum deflection at the end of the beam was achieved when
the porosity was minimized. A multi-objective optimization procedure founded on NSGA-II was
adopted in Jamshidi and Arghavani’s work [43] to optimize the porosity distribution of 2-D FG
porous beams. The objective functions were to maximize the buckling load and minimize the
weight. The researchers presented several examples with different symmetrical clamped and
hinged boundary conditions. They claimed that the concentration of porosity was on the middle of
the beams rather than outer corners in most of the optimal designs. Moreover, boundary conditions
were found to be influential on the final results and optimal porosity distribution.

3. Functionally Graded Plates

Various shapes of plates such as circular, square, rectangular, trapezoidal and skew plates have
been investigated by the scientists since these structures are remarkably applicable in engineering
designs. The splendid thermo-mechanical behavior of FGMs have inspired many researchers, over
the last twenty years, to probe the optimal design of FG plates with regards to different objective
functions. Fig. 5 depicts a schematic of an arbitrary rectangular FG plate and its coordinate system.

15
 Fig.5. Schematic of transversely FG plate and the corresponding coordinate system

3.1. Transversely functionally graded plates

Fares et al. [44] attempted to optimize the structure and closed loop control function of FG plates.
The objective function associated with the control problem was defined as minimizing the dynamic
responses of these plates while the structural objective was to maximize the critical buckling load.
These optimization problems were imposed to constraints such as control energy and total
thickness. It was proved that ignoring the normal strain effects could lead to errors up to 22% when
the power-law exponent by which the material distribution of the plates was determined was higher
than 1. A vibrational optimization study was presented by Alshabatat et al. [5] which was
associated with FG plates with various boundary conditions. The distinctive in-plane material
distribution was taken into consideration by using a five-parameter trigonometric function
(Eq.6.b). The coefficients of this function were chosen as the design variables in the two key
problems of their work. The first one dealt with maximizing the fundamental frequency while the
second one involved minimizing the kinetic energy of the mentioned structure vibrating under a
particular excitation frequency. The effective material properties were calculated via Mori-Tanaka
scheme and the researchers reported that this type of material variation model had great influence
on the vibrational characteristics of these plates. Correia et al. [45] implemented Direct Multi-
Search (DMS) method and GLODS method to optimize the thickness and material distribution of
FG plates in three different examples. The objective functions were to maximize critical buckling
load and natural frequency and to minimize the mass and cost of FG plates. Several optimum
design domains were achieved and the best possible design for these structures was proposed
depending on the priority of the four mentioned objective functions.
Ootao et al. [46] implemented ANN to optimize the material distribution of FG plates under
uniform heating on one surface with the aim of minimizing the thermal stresses. They assumed the
materials to be temperature-dependent and finally compared their results with those of an ordinary
non-linear mathematical programming. The results were in line with each other. Goupee and Vel
[6] introduced two optimum solutions for simply supported FG plates subjected to thermo-
mechanical loads using RCGA. In the first model, they managed to decrease the peak effective
stress of FG plates made of Ni/Al2O3 nearly 55% by optimizing the material distribution through
the thickness of these plates. Using the same design variable in the second model, they managed

16
to decrease the total volume fraction of ceramic from 0.5 to 0.44 and maximum temperature that
the metal experienced from 500°C to 275°C while the amount of effective stress was considered
as the constraint. Mori-Tanaka and self-consistent homogenization models were utilized in their
study to obtain the material properties variation. The maximum Von Mises stress, shear stress and
deflection of rectangular FG plates were minimized in a research performed by Helal and Shi [47]
employing ANSYS Parametric Design Language (APDL). The volume fraction distribution
through the thickness of these plates was chosen as the design variable. The researchers used
different types of material models namely power-law, exponential and sigmoid models to
investigate the influence of these models on the optimal design and finally managed to decrease
the deflection, Von Mises stress and shear stress by 24%, 22% and 11%, respectively.
Additionally, the best optimal results were related to those following sigmoid and power-law
models. Ding and Wu [48] adopted a GA-based optimization approach to minimize the peak
thermal induced stress of FG plates whose surfaces were imposed to particular temperatures and
heat convention. The plates were artificially assumed as multi-layered plates and four different
material distribution models namely power-law, sigmoid, linear layer-wise and step layer-wise
models were taken into account. Regarding this specific objective function, the results proved that
the best material model was the linear layer-wise scheme and the authors suggested that a higher
order layer-wise model would lead to even better optimum results.
Na and Kim [49] provided an optimization research related to thermo-mechanical buckling of FG
plates. They calculated the best volume fraction distribution through the thickness of these plates
which resulted in minimum temperature of the metal phase and the minimum total stress both
phases experienced. Improving the thermo-mechanical loads was another objective of their study.
Utilizing GA and PSO method, Ashjari and Khoshravan [50] optimized the material distribution
of simply supported FG plates under sinusoidal and uniformly distributed transverse loads in order
to minimize the total mass. Deflection and stress were considered as the constraints individually
and simultaneously through three case studies. The best results were achieved when both
constraints were imposed on the process of optimization. The researchers demonstrated that the
best performance was associated with symmetrical sandwich plates with homogeneous cores and
FG face sheets. They also claimed that PSO method was slightly more efficient than GA. Hossein
and Mulani [51] chose to analyze the shear and uniaxial buckling of simply supported and clamped
FG plates. These plates were actually made up of aluminum reinforced with silicon carbide nano-
particles. The authors subsequently decided to minimize the required amount of reinforcing
material which led to the minimum cost of these structures. They let material distribution be the
design variable which was determined by an in-plane polynomial function. The researchers
claimed that this function outperformed other schemes such as power-law since it was capable of
controlling the material distribution more efficiently. Finally, they managed to minimize the
amount of ceramic powder approximately 55% after the optimization process. A similar approach
was taken by the same authors [52] to decrease the cost and total volume fraction of silicone
carbide in these types of simply supported FG plates under uniformly distributed loading, non-
uniform loading, central point loading and flutter speed constraint. After the optimization process,
this value decreased by 16.3%, 22%, 82% and 18.6% for these four load cases, respectively. These
authors [53] also tried another optimization approach towards minimization of the reinforcement
material (silicon carbide) in three types of FG plates namely unstiffened, stiffened and plates with
17
cutouts with critical buckling load as the constraint. The impact of several parameters such as
boundary conditions, the type of structure, the direction of material variation and the type of
imposed loads on the optimal design were demonstrated. It was proved when looking at the final
optimal results that the most decrease in the amount of reinforcement material was related to
clamped rather than simply supported boundary conditions, vertical (perpendicular to the direction
of stiffeners) rather than horizontal material variation direction and shear rather uniaxial loadings.
The largest amount of silicon carbide saving was found to be up to 200% compared to that of
homogeneous panels.
In order to obtain a certain desired temperature distribution through FG plates, Golbahar Haghighi
et al. [54] assumed the six surfaces of these structures to be imposed to heat fluxes. Accordingly,
these authors employed CG method combined with Differential Quadrature Method (DQM) to
minimize the discrepancy between the desired and actual temperatures in the presence of heat
fluxes. They also provided different examples dealing with various shapes, material distributions
and desired temperature domain distributions to show the efficiency of their method. Chiba and
Sugano [55] utilized GA to optimize the material variation of FG plates under temperature loads
with regard to the microstructural morphology and microstresses into consideration. The objective
function was to maximize the safety index defined as the relationship between microstress domains
and phase-specific failure criteria. The researchers claimed that exploiting the knowledge of
microstresses in the optimization process had considerable influence on the final optimal volume
fraction distribution of the material constituents. Chen et al. [56] performed an elastic analysis of
FG plates via a meshless collocation method. This method was founded on generalized multi-
quadrics Radial Basis Functions (RBF) which was defined by a function with two parameters
referred to as the shape parameter and exponent. These authors then employed GA to minimize
the discrepancies between the results of their study and those obtained by Zenkour’s research [57].
Both of these studies were based on Higher order Shear Deformation Theory (HSDT). The optimal
design variables in their study were the exponent and shape parameter which were found to be
0.73 and 1.3, respectively. Lieu and Lee [58] approximated the volume fraction of ceramic of a
FG plate at specific grid points along the thickness using Greville abscissae and material
distribution through the thickness via a B-spline basis function. The constituent materials were
assumed to be temperature dependent. The plates were subjected to thermo-mechanical loads, and
the optimum material distribution which led to the minimum compliance was determined using
Adaptive Hybrid Evolutionary Firefly Algorithm (AHEFA). According to the optimal results, the
middle plane was metal-rich and the surfaces were ceramic-rich. Additionally, the mentioned
algorithm resulted in more accurate results with less computational costs.

3.2. 2D functionally graded plates

2D-FG plates are those in which the volume fraction of each constituent material varies through
the plane in both directions (Fig. 6). The application of this kind of FGMs can be seen mostly in
designs which contain cutouts and consequently the variation of materials in two dimensions
results in better stress distribution or effective high temperature-resistance.

18
 Fig.6. In-plane volume fraction distribution in a 2D-FG plate. [59]

Gradual variation of materials in two or more directions can be seen in many different natural
phenomena. This can be seen nowhere clearer than the microstructure of bones. Although bones
contain a lot of holes, they could carry huge loads while maintaining their low weight. These
significant features are attributed to the multi-directional material distribution around these holes.
Inspired by the structure of bones, Huang and Rapoff [60] analyzed plates with a circular hole in
the middle when the material around the hole was assumed to vary according to a power-law
function. They demonstrated that this design led to a higher ratio of strength to weight of the plates
compared to the homogeneous plates. Another optimization study on FG plates with a hole in the
middle was performed by Venkataraman and Sankar [61] employing GA. The hole was assumed
to be encircled by a number of rings when the material type and properties of these rings varied
gradually. Thus, this design was considered as a 2D-FG plate with a hole. The objective function
in this study was to minimize the failure index by finding the best material distribution around the
hole. The authors compared their design with perforated homogeneous plates and proved the
significant influence of gradually varying materials around the hole on the strength of the plate. A
2D-FG plate, constituted from tungsten and copper, with a hole near to its southern edge was
optimized by Vel and Goupee [62] implementing NSGA-II. The northern edge was the tungsten
rich edge which was assumed to have 500°C while the cutout edge’s temperature was 0°C. The
researchers’ goal was to maximize the dynamic and static safety factors of these plates under the
mentioned temperature gradient. To this end, they tried to find various optimal results and
represented a Pareto front regarding the objective functions. According to the final optimal design,
as the volume fraction of tungsten increased, the static safety factor increased while the dynamic
safety factor showed an opposite downward trend. Nemat-allah [63] conducted an optimization
approach to minimize the thermal and residual stresses of 2D-FG plates composed of ZrO2/6061-
T6/Ti-6Al-4V subjected to several thermal loading cycles. The volume fraction of material
constituents were assumed to vary via power-law functions in both directions. Thus, the power-
law exponents were considered as the design variables and the optimum values for these
parameters which led to minimized maximum temperature and minimized maximum equivalent
stress were found. A plate with semicircular cutouts on one edge, made up of 2D-FGM, was
optimized by kou et el. [59] to minimize the mass and maximum equivalent Von Mises stress. This
plate was subjected to a linear temperature gradient. The main objective of this study was to
illustrate the superiority of PSO method over mathematical programming methods such as

19
attractive set and trust region algorithms. The material distribution of the mentioned FG plates was
determined by a Heterogeneous Feature Tree (HFT) model which allowed the researchers to
approximate a versatile material propagation via a number of explicit functions. The authors finally
managed to decrease the mentioned objectives by nearly 4% and 24%, respectively. In order to
demonstrate the outperformance of modified Symbiotic Organisms Search (mSOS) method over
DE and Symbiotic Organisms Search (SOS) methods Do et al. [64] conducted two optimization
studies on 2D FG plates. The fundamental frequency and buckling load were maximized by
obtaining the optimum material gradation through these structures. These authors also proved that
combining mSOS with Deep Neural Network (DNN) method decreased the computational time
significantly and outperformed the mSOS in the framework of isogeometric analysis. Lieu and Lee
[65] exploited AHEFA in the framework of isogeometric multi-mesh design to maximize the
fundamental frequency of multi-directional FG plates. Different plates with constant and variable
thickness were analyzed in their study and the material inhomogeneity as well as thickness
variation profile were taken as the design variables. The aforementioned method enabled the
researchers to generate two independent NURBS surfaces which led the optimization process to
have remarkably lower computational costs.

3.3. Functionally graded sandwich plates

Utilizing DE method, Silva and Loja [66] could optimize FGS panels constituted from aluminum
core and two FG face sheets aiming at minimizing the thermal residual stress distribution. The
design variables were defined as the thicknesses of the core and face sheets and the volume fraction
variation pattern in the face sheets which was calculated by a power-law function. The researchers
restrained the range of power-law index between 0 and 10. The larger this parameter was, the less
the thermal residual stresses were. Therefore, the optimal exponent was calculated to be 10. Ashjari
and Khoshravan [67] represented an optimization solution based on NSGA-II to minimize the
mass and deflection of FGS plates subjected to mechanical loads with the stress field as the
constraint. The material volume fraction was defined in some control points and its distribution
was calculated via a piecewise cubic interpolation function. These control points, the volume
fraction distribution and the thickness of the face sheets were the design variables of this study. As
transverse shear stress played an important role in the maximum equivalent Von Mises stress at
the vertical edges, the researchers reported that the maximum shear stress occurred in designs with
stiffer face sheets and steeper material variation. Utilizing PSO method, Torabi and Afshari [68]
optimized trapezoidal FGS symmetric plates as a model of aircraft wings in order to maximize the
critical aerodynamic pressure. These plates were made of two FG face sheets and a homogeneous
core. Material variation through the thickness as well as several geometrical parameters such as
the thickness of core and plate, aspect ratio and plate angles were taken as the design variables.
The best optimal design regarding the mass and angles of each corner as the constraints was
achieved for these symmetric FGS plates. However, the authors recommended that asymmetric
designs might be more capable of increasing the mentioned objective function. Lieu et al. [69]
developed AHEFA to optimize the material variation of FGS plates with regards to their
vibrational behavior. These plates are considered as multi-layer plates. Therefore, a multi-patch

20
B-spline basis function was employed to determine the thickness of each layer and the volume
fraction of ceramic at each control point. The set of control points was examined by Greville
abscissae. Two different problems with different objective functions and constraints were
explored. The objective function of the first problem was to maximize the natural frequency having
total volume as the constraint. The second problem dealt with minimizing the mass imposing
constraints on fundamental frequencies. The authors employed DE method and FA as well as
AHEFA and demonstrated the outperformance of the last case in terms of accuracy and
convergency speed. Moreover, they revealed that the B-Spline functions led to a totally different
optimal design compared to their power-law counterparts. It is to be noted that utilizing B-spline
functions in determining the material propagation of FG plates resulted in more feasible and
convenient manufacturing process.

3.4. Step-formed functionally graded plates

Step-formed FG plates are those which are made of a number of homogeneous layers, and the
volume fraction of materials varies gradually through the thickness from a layer to another.
Detailed schemes of these types of FG plates and an arbitrary step-formed functionally graded
volume fraction through the thickness of these plates can be observed in Fig. 7.

Fig.7. (a) Geometry of a step-formed FG plate. (b) An arbitrary volume fraction distribution of ceramic in the step-
formed FG plate. [70]

Ootao et al. [71] applied GA to optimize the material composition of ten-layered step-formed FG
plates whose ceramic rich surface was subjected to uniform heating. The objective function was
to decrease the maximum thermal stress in the transient state. Three different cases were
considered in this study. In the first one, eight design variables associated with the volume fraction
of the eight middle layers were chosen while the volume fractions of metal in the first and tenth
layers were 0 and 1 as the constraints. These constraints were omitted in the second case. In the
third case, the material distribution of layers through the thickness was assumed to follow a power-

21
law function and the number of design variables declined to only the power-law index. The
researchers noted that the best stress reduction was related to the second design. The material
distribution of step-formed FG plates was optimized by Na and Kim [70] to reduce the stress and
to enhance their thermo-mechanical buckling. These types of structures were composed of several
layers made up of homogeneous materials and the material properties of the layers were assumed
to be temperature-dependent and vary gradually from layer to layer by controlling the volume
fraction of their material constituents. A similar optimization approach was performed by these
authors [72] on FGS plates consisting of an FG core and homogeneous face sheets considering
core to face sheet thickness ratio and material variation pattern as the design variables. Comparing
these plates with a two-layered composite plate made up of the same ceramic and metal, the best
results were related to those of FGS plates. However, the optimum results of step-formed FG plates
were found to be close to those of the aforementioned FGS plates.

3.5. Multi-layered plates containing functionally graded layers

Shape control optimization of FG beams and plates was conducted by Liew et al. [73] utilizing
GA. Both of the mentioned structures were sandwiched between two layers containing piezo-
electric patches. The top and bottom layers were playing the role of actuator and sensor,
respectively. The researchers finally obtained the optimum voltage distribution through these
layers which led to the desired shape of the beams and plates. They also investigated the influence
of material distribution through the thickness of these FG structures restrained to clamped and
simply supported boundary conditions on their optimal designs. FG plates sandwiched between
two piezo-electric layers made of Lead Zirconate Titanate (PZT) and/or Polyvinylidene Fluoride
(PVDF) were analyzed by Nourmohammadi and Behjat [74] in terms of their bending behavior.
The material properties were assumed to vary according to a power-law function and the exponent
was considered as the design variable by which the deflection of these plates was intended to be
minimized. These plates were subjected to electrical, mechanical and thermal loads. The
researchers proved that each type of loading led to a distinct optimum value of power-law index.
Three different layer arrangements namely PVDF-PVDF, PVDF-PZT and PZT-PZT were
investigated and the last case proved to result in the lowest minimized deflection.
Multilayered plates made up of ZrO2 and Ti6Al4V alloy whose thermal properties significantly
depend on temperature were optimized by Parashkevova et al. [75] when undergoing temperature
gradient. The layers were made up of ceramic and metal decussately and the thicknesses of layers
were assumed to vary through the thickness on the basis of an exponential function. Optimizing
the distribution of thicknesses through the thickness, the researchers intended to minimize the
stress jumps. Tsai-Wu, Mohr, maximal stress and deformation criteria were utilized in this research
to determine the stress through the thickness of these plates. Moreover, various boundary
conditions were imposed on these plates amongst which the simply supported boundary conditions
were observed to lead to the highest stress jumps. Gunes et al. [76] made an attempt to optimize
the adhesively bonded FG single lap joints to maximize their fundamental frequency and to
minimize their modal strain energy. The material distribution of FG plates and the geometry of the
joint formed the design variables of this research. The effects of several parameters such as the

22
thickness of the plate, overlap length, material composition of plate and the thickness of the
adhesive layer on the mentioned objectives were investigated, and the last parameter had negligible
influence in contrast with the first two variables. According to the optimum design, when the
overlap length increased, the thickness of the plates had to be decreased and when the overlap
length decreased, the thickness of plates had to be increased. Additionally, increasing the volume
fraction of ceramic in plates led to higher natural frequencies. In order to demonstrate the
outperformance of plates constituted from two homogeneous face sheets and a FG core over a two-
layered composite made up of the same metal and ceramic in terms of thermo-mechanical buckling
behavior, Na and Kim [77] compared these two types of plates and carried out an optimization
study on the former case to obtain the best volume fraction distribution of ceramic through the
thickness. The constituent materials were assumed to be temperature dependent varying through
the thickness via a power-law function. Protecting the battle vehicle bodies against radar
surveillance and blast waves motivated Zhu et al. [78] to design an elaborate type of sandwich
plates constituted from a glass composite face sheet at the top, a carbon composite face sheet at
the bottom and an Functionally Graded Foam (FGF) core. Additionally, these authors conducted
an optimization approach based on Response Surface Methodology (RSM) to find the best material
variation in the core which led to the minimum vulnerability of this intrinsic structure against
electro-magnetic or shock wave mitigation. The material distribution of the core was assumed to
follow a power-law index and the optimum value of this design variable was found to be between
8 and 10. Moreover, the ascending gradient model (increasing the stiffness of the FGF from the
bottom to the top of the core) led to the best performance of these plates. Utilizing Particle Swarm
Optimization Sliding Mode (PSOSM) control system, Moghaddam and Bagheri [79] optimized a
three-layer composite made up of an FG panel sandwiched between two piezo-electric layers at
the top and bottom as the actuator and sensor, respectively. The aim of their study was to suppress
the vibration of these structures which were imposed to three free and one clamped boundary
conditions. The researchers also compared the results with those of Traditional Sliding Mode
(TSM) control system and demonstrated the efficiency and outperformance of PSOSM over TSM
in terms of convergency speed. Utilizing the combination of GA and ANN, Daynes et al. [80]
conducted a topology optimization on flat sandwich panels containing an FG lattice core. They
calculated the best orientation and geometrical properties of the lattice cells which led to the
highest ratio of strength to weight. According to the final results of this research, the stiffness and
strength of FG lattice panels could be increased significantly through the optimization process.
The stiffness and strength of the optimum FG lattice panels were found to be 101% and 172%
more than those of their uniform lattice counterparts, respectively, while their weights were equal.

3.6. Functionally graded carbon nanotube plates

Setoodeh and Shojaei [81] conducted an optimization study based on GA to maximize the buckling
load of Functionally Graded Carbon Nanotube (FGCNT) reinforced composite quadrilateral
plates. Three different volume fraction distributions of CNTs through the thickness of these plates
were considered. The fiber orientation angles were chosen as the design variables in this study and
the influence of other parameters such as volume fraction of CNTs, geometry and boundary

23
conditions on the buckling of these plates were also demonstrated. The best results were related to
the plates in which the volume fraction of the CNTs at the center and surfaces of the plate was
maximum and minimum, respectively. Vo-Duy et al. [82] proposed an optimization method based
on Adaptive Elicit Differential Evolution Algorithm (AEDEA) to optimize the fiber orientation
angles of CNTs in FGCNT reinforced composite quadrilateral plates considering maximization of
fundamental frequency as the objective function. The influences of other parameters such as
volume fraction distribution of CNTs through the thickness, number of layers and boundary
conditions on the optimal design were illustrated the last of which was found to be the most
influential. Hussein and Mulani [83] employed RBDO method and a deterministic optimization
approach based on SQP to determine the best possible material gradation of in-plane FGCNT-
reinforced polymer composite plates aiming at improving the stiffness of these structures.
Accordingly, these researchers calculated the optimum volume fraction distribution through plates
which led to the minimum amount of required stiffening CNT considering deflection as the
constraint. A comparison between the optimized FGCNT plates with homogeneous CNT
composites made evident that the former case required 45% less amount of CNT for the same
values of stiffness. Additionally, the optimal volume fraction of CNT in the FGCNT plates
examined by RBDO method was 48% more than that of the deterministic approach.

3.7. Functionally graded foam-filled plates

Thermal protection Systems containing FG metal foam insulations were optimized in a study
performed by Zhu et al. [84], implementing SQP. As these systems should keep the maximum
temperature below a certain limit, the objective function was to minimize the maximum
temperature by optimizing the solidity profile of the FGF. The total thickness and mass of the
panel which was supposed to be insulated by FGF were chosen as the constraints. According to
the outcomes related to the optimal design of these systems, the solidity of the inner layer (which
was cooler) was high while the solidity of the outer layer (which was hotter) was low. Furthermore,
the optimal results of transient heat conduction showed a huge discrepancy with those associated
with steady state analysis. FGF insulations were optimized by Venkataraman et al. [85] in order to
minimize the heat transmission. These foams were compared to their uniform counterparts and
were proved to outperform them regarding the mentioned objective function. Closed and open
cells were considered and the heat transmission of the former did not change significantly during
the optimization process. In contrast, the open cell FGF’s heat transmission was proved to be 10%
higher than that of the uniform foam. Moreover, a FGF could transmit the same amount of heat as
the uniform foam while it is 12% thinner than its uniform density counterpart. RSM helped Zhang
and Zhang [86] to optimize the density variation of FGF blocks so as to enhance the
crashworthiness behavior of these structures. Accordingly, these blocks were assumed to be under
middle and high velocity ball impact loads. Various density distributions for these foam blocks
were studied and the linear decreasing density model was proved to outperform other designs in
terms of crush force efficiency. This design was also the most optimal when the foam was
sandwiched between two layers.

24
 3.8. Other functionally graded plates
A commercial optimization program named DOT helped Aboudi et al. [87] to optimize FG
composite plates in which the fiber volume fraction varied through the thickness. These plates
were assumed to be under temperature gradient. The optimization problem was solved with and
without constraint on fiber spacing and the objective function in both problems was to minimize
the moment resultants attributed to thermal loading. Moreover, the researchers entered generalized
plane strain and plane strain boundary conditions into the optimization formulations and compared
the results. According to the final results, the boundary conditions and the mentioned constraint
had great influence on the reduction of the objective function and the pattern by which the volume
fraction of fibers varied in the optimal designs. Increasing the buckling load of rib stiffened FG
plates, Birman and Byrd [88] decided to add some stringers to these plates. After investigating the
positive influence of these stringers on the critical buckling load of FG plates, the researchers
minimized the mass of the stringers letting critical buckling load be the constraint. Considering
neither the local buckling nor the buckling of stringers is the flaw which can be seen in this
research. Putting these two objectives aside, the best optimized results were achieved when the
minimum number of stringers were widely situated. Hedia et al. [89] utilized the ANSYS package
to optimize the material distribution of a perforated porous FG plate aiming at minimizing stress
concentration near the hole when undergoing thermal and pressure loads individually and
simultaneously. A non-liner function was employed to determine the effective material properties
of these plates. This non-linearity is attributed to the porosity distribution through these materials.
After obtaining the optimized results, the authors claimed that while one of the loads (thermal or
pressure) was imposed, a ceramic-rich design was optimum. On the other hand, a metal-rich plate
was found to be the best optimized design when both loads were applied on the plates.

4. Functionally Graded Shells

Shell structures are vastly employed in different engineering designs such as fluid containers or
conveyers, aerospace and naval construction industries. Their curvature helps them to carry out
huge loadings effectively. Therefore, they are often used as load bearing structures for aircrafts,
rockets, submarines and missile bodies. As FGMs have extraordinary features, many researchers
have been encouraged to optimize FG shells with different purposes recently. There are several
types of shells namely cylindrical, conical and spherical shells. In this section, these types of shells
along with cylinders, vessels and pipes made of FGM are investigated.

25
 Fig.8. Geometry of an FG cylindrical shell and the corresponding coordinate system

4.1. Functionally graded cylindrical and conical shells

The material distribution of FG cylindrical shells was optimized by Ootao et al. [90] using
mathematical programming and Neural Network (NN) method. These structures were assumed to
be under various thermal conditions and the aim was to minimize the thermal stresses. The volume
fraction of material constituents through the thickness of these shells was considered to vary
according to the power-law function and the exponent was taken as the design variable. The
optimum outcomes of the aforementioned methods were in line with each other. The researchers
also noted that although the first training of the neural network approach required a long time of
CPU processing, once completed, learned and unlearned thermal conditions would take a very
short amount of time for CPU processing. This feature was highlighted as the advantage of NN
over the mathematical programing method. In order to show how thermo-mechanical behavior of
FGM depends on the volume fraction distribution of materials through the thickness of these
structures, Cho and Ha [91] were motivated to adopt an optimization approach to this type of
materials. They incroporated a combined interior penalty function and GS method to optimize their
material propagation aiming at minimizing the steady-state thermal stresses. The material
properties were assumed to vary according to TTO model. The researchers proved that varying the
material distribution gradually has a significant effect on decreasing the thermal stresses by
presenting examples of a FG beam and cylinder. A four-layered cylinder made up of a steel layer,
a metal layer, a FG layer and a ceramic layer subjected to temperature gradient and inner and outer
pressure was optimized by Huang et al. [92] to reduce the tensile stresses at the outer ceramic rich
surface. Material distribution through the thickness of FG layer was estimated by a power-law
function and the exponent as well as the geometry of the cylinder were chosen as the design
variables of this optimization study which was based on a quadratic programming method. The
optimal outcomes demonstrated that increasing the thickness of the FG layer resulted in lower
tensile stress of the ceramic outer surface. Axially FG cylinders made up of nickel and alumina
were modeled by Choi and Cho [93] as a ten-layer structure. With the purpose of minimizing the
maximum hoop stress at the metal layer, they found the optimum material distribution of the multi-
layered cylinders and subsequently with the aid of an interpolation function, they managed to
obtain the best volume fraction distribution through the longitudinal direction of these FG
cylinders. The peak stress which the metal rich surface experienced was declined from 430Mpa to

26
231Mpa after the optimization process. Batra [94] analyzed the FG incompressible linear elastic
spherical and cylindrical shells subjected to inner and outer hydrostatic pressure. The aim of this
research was to find the best shear modulus variation through the thickness of these shells which
led to optimal circumferential and hoop stresses. Linear distribution of shear modulus through the
thickness was finally found to be the best design according to the aforementioned objectives for
both types of shells. Vel and Pelletier [2] applied an elastic NSGA to optimize the material
distribution of plates and shells made of FGM introducing two models. In the first model, the
objective function was to minimize the peak hoop stress and mass simultaneously under thermal
loads considering the maximum temperature of the metal as the constraint. The researchers
reported that the optimized design led to a slight increase in mass but a remarkable decrease in
hoop stress. In the second model, they took minimizing the mass and maximizing the safety factor
of shell under heat flux as the objective functions. Asgari [95] implemented Multi-Objective
Genetic Algorithm (MOGA) to optimize the volume fraction distribution of 2D-FG hollow
cylinders. The objective functions were considered to be minimizing the mass, stress wave
propagation and the displacement of these structures under impulsive internal pressure. Cubic
spline interpolation functions were chosen for determining the material distribution variation in
both radial and axial directions. Moreover, the outperformance of these functions over regular
power-law functions was noticed. In order to reduce the vibration and suppress the undesirable
vibrations of shear deformable FG cylindrical shells, Biglar and Mirdamadi [96] tried GA to
optimize the orientation angle and location of piezo-electric sensors and actuators installed on the
surface of these shells. After analyzing three different examples, the researchers claimed that
attaching sensors and actuators at the optimal angle and position could lead to higher transmission
of mechanical energy to the shells and this resulted in a very quick elimination of vibrations. The
influences of the material distribution of FG shells and control gain on the suppression of
vibrations were determined as well.
In a GA-based optimization approach, Nabian and Ahmadian [97] maximized the natural
frequency and minimized the weight of simply supported FG hollow cylinders by finding the best
material gradation pattern through the thickness of these structures. Piece-wise cubic interpolation
function was implemented in this research to determine the volume fraction of material
constituents at specific control points and to examine the volume fraction distribution through the
thickness of cylinders. The authors finally extracted a Pareto front for these two contradictory
objectives as a benchmark for other engineers. Tornabene and Ceruti [98] employed PSO method,
GA and Monte-Carlo (MC) algorithm to optimize material variation in FG panels and doubly
curved shells to minimize the deflection and maximize the fundamental frequency of these
structures. The analyzed shells were assumed to consist of a number of laminas and the material
distribution of each lamina was determined by a four-parameter power-law function. A comparison
between the mentioned algorithms depicted that PSO method and GA represented more exact
results than MC did; particularly when both of the objective functions were taken into
consideration. Looking from the lens of a designer, the main disadvantage of this approach was
attributed to the lack of any relationship between the four parameters of the power-law function
and the dynamic or static performance of the mentioned shells. Maximizing the sound transmission
loss of FG cylindrical shells which stems from the optimum first resonant frequenct encouraged
Nouri and Astaraki [99] to optimize the volume fraction distribution of materials and geometry of
27
these shells via GA. The thickness, frequency range and weight of these structures acted as the
constraints. Different sets of materials were analyzed in this research and finally, the combination
of nickel-aluminum and steel-aluminum were found to present the maximum sound transmission
loss and minimum weight. In order to spend the minimum amount of control energy for
suppressing the vibration of FG truncated conical shells, Fares et al. [100] attempted to optimize
the control forces and deflection of these shells when they were imposed to simply supported and
clamped boundary conditions. The integration of total energy of the shell and control energy was
taken as the objective function which was intended to be minimized using Lyapunov-Bellman
theory. The most important quality of this investigation was related to incorporating the stretching
effects in optimization formulations which was claimed by the authors to be significantly
influential on the optimal design; especially in case of moderately thick and short conical shells.
Feasible Arc Interior Point Algorithm (FAIPA) helped Moita et al. [101] to optimize FG plates
and cylindrical panels choosing fundamental frequency, mass, central deflection and critical load
as the objective functions or the constraints. Material variation through the thickness was assumed
to follow a power-law function and the power-law index plus the thickness were taken as the design
variables in all case studies. Finally, after representing five different examples, the researchers
introduced the utilized algorithm as an efficient and accurate method of optimization in case of
FGM.

4.2. Functionally graded vessels and pipes

Pressure vessels and pipes which are mechanical elements for containing and conveying fluids
used to be made of steel. During the past decades, engineers have been attracted to manufacturing
these structures from composite materials. As the safety of these elements are of significant
importance, researchers tried to design FG pressure vessels and pipes and optimize their material
propagation with the aim of achieving designs with the best thermal and mechanical performances.
Maalavi and El-Sayed [102] exploited a mathematical programming technique via MATLAB
toolbox to optimize the distribution of materials in the longitudinal direction of FG pipes. The aim
of their research was to maximize the ratio of stiffness to weight and the stability of the pipes
without increasing the mass even slightly when the fluid current reached the divergence velocity.
They considered three different boundary conditions namely hinged-hinged, clamped-hinged and
clamped-clamped for these pipes. In a comparison between the results of the FG pipes with the
existing baseline designs, the former showed greater divergence velocity and higher bending
stability. It is to be noted that the most important parameter in this optimization approach was the
length of pipe elements. Utilizing APDL, Wang et al. [103] optimized the material composition of
FG vessels with two closed ends subjected to thermal loads and internal pressure. They aimed to
find the best stress propagation through these structures. The material properties such as Young’s
modulus and coefficient of thermal expansion were determined by power-law functions with
different indices. Finally, these optimal indices which led to the minimum equivalent Von Mises
stresses average variance were obtained and presented by the researchers. In order to minimize the
failure probability of step-formed FG pressure vessels, Shabana et al. [104] optimized the number
of layers and material distribution through the thickness of these shells using PSO method. To this

28
end, the researchers decided to choose the maximum hoop stress jump at the interfaces and the
maximum hoop stress through the thickness as the objective functions. Their results documented
that changing the priority of the objective functions led to different structural stiffness. Moreover,
they managed to decrease these stresses by nearly 18% and 5%, respectively.

4.3. Functionally graded spheres

A gradient based method was implemented by Boussaa [105] on temperature-dependent FG thick
hollow spheres under thermal gradient to minimize the maximum Von Mises equivalent stress.
Material distribution through the thickness was chosen as the design variable. The final optimized
design was compared to homogeneous spheres with the same material constituents, geometry and
loading and the optimized FG spheres was found to be capable of bearing nearly 100% higher
magnitude of thermal loads than their homogeneous counterparts. The applicability of various
types of PSO namely Particle Swarm Optimization with Passive Congregation (PSOPC), Particle
Swarm with Ant Colony Optimization (PSACO) and Particle Swarm with Passive Congregation
and Ant Colony Optimization (PSPCACO) methods was explained by Fereidoon et al. [106] in an
optimization approach with two main examples. The first one dealt with an FG plate cooled down
from 300K to 100K while minimizing its peak residual stress. The second example was related to
temperature-dependent FG spheres subjected to gradient temperature which were intended to have
the maximum magnitude of safety factor against yielding. The material variation profile of these
structures was the design variable in both cases. The safety factor of the optimal design of the
spheres was found to be 28% higher than that of their non-optimized counterparts.

5. Functionally Graded Tubes

In order to reduce injuries in accidents, crash structures are vastly utilized in automobiles. The
quality of making passengers safe during any accident is referred to as crashworthiness
characteristics. This factor is defined by two parameters namely Specific Energy Absorption
(SEA) and Peak Crushing Force (PCF). An optimum design of any structure with high
crashworthiness quality has the maximum SEA and minimum PCF. Empty tubes and tubes filled
with foam are of the most suitable crash structures which have attracted many scientists’ attention
to optimize them so as to improve their crash behavior. To this end, during the past decade, many
scientists have attempted to design FG tubes which are divided into two types. The first case are
FGF tubes which contain foam with variable density. The other type is called Functionally Graded
Thickness (FGT) tubes which are those with gradually varying thickness along the longitudinal
direction. These two kinds of FG tubes are shown in Fig. 9 and Fig. 10.

29
 Fig.9. Schematic of an FGF tube with square cross section under lateral impact loading.

Fig.10. Schematic of an FGT tube with circular cross section under axial impact loading.

5.1. Functionally graded foam-filled tubes

Two types of Functionally Lateral Graded Foam-filled Tubes (FLGFT) were analyzed by Yin et
al. [107] to boost the crashworthiness characteristics of these structures when undergoing axial
impact loads. After comparing these tubes with their Uniform Foam-filled (UF) counterparts, the
researchers realized the outperformance of FLGFTs over UF tubes. Subsequently, Multi-Objective
Particle Swarm Optimization (MOPSO) method helped these authors to optimize the material
gradation of foam evaluated by a power-law function to reduce the PCF and to increase the SEA
even more. The Pareto front related to these two conflicting objectives was provided as a
benchmark for engineers. A comprehensive comparative optimization study was performed by Yin
et al. [108] in terms of crashworthiness behavior of FG tubes. These researchers optimized the
density distribution pattern through Functionally Graded Foam-filled Tapered Tubes (FGFTT)
utilizing the combination of dynamic ensemble metamodeling and MOPSO methods. Afterwards,
they compared the optimized FGFTT with the conventional UF tubes and demonstrated that the
ascending pattern of foam in FGFTTs had worse performance than UF tubes. However, the Pareto
front extracted for the descending pattern of FGFTT proved to have better crashworthiness
characteristics than UF tubes. Another comparison of this research was related to the efficiency of
different methods. A number of traditional optimization methods such as RBF, Design of
30
Experiment (DOE), Super Vector Regression (SVR) and Polynomial Response Surface (PRS)
were compared with the employed hybrid technique. The MOPSO method combined with the
metamodeling method was observed to be the most efficient approach. Fang et al. [109]
investigated the impact behavior of foam-filled thin walled tubes subjected to lateral impact
loadings. These tubes were assumed to be filled with FGF and UF and the crash behavior of these
two cases were compared with each other. This comparison demonstrated that FGF tubes had
higher SEA but lower PCF compared to their UF counterparts. Subsequently, these researchers
utilized PSO method and Kriging model to optimize both types of these tubes aiming to improving
their crashworthiness behavior. In FGF cases, density gradation was considered as the design
variable with the wall thickness, range of foam density and yielding strength as the constraints
while in the UF case, the foam density was taken as the design parameter. According to the optimal
designs, FGF showed better Pareto fronts. Moreover, axially graded FGF tubes outperformed their
transversely distributed foam-filled counterparts in terms of SEA. In another study, Li et al. [110]
applied NSGA-II to optimize the material gradation of FGF tubes subjected to oblique impact
loads which led to enhanced crash behavior. Four impact angles of 0°, 10°, 20° and 30° were
considered. Each impact load led to a particular set of optimal designs, PCF and SEA. Therefore,
the authors decided to take all these load angles into consideration by adding weighting factors
attributed to each angle in the optimization formulation. They concluded that incorporating all the
impact angles resulted in a different optimal material variation through the longitudinal direction
of the tubes. A comparison between conical tubes filled with FGFs and UFs subjected to axial
impact loads was also represented by Mohammadiha and Beheshti [111]. Similar to cylindrical
tubes, FGF tubes outperformed UF tubes regarding SEA and PCF. After this comparison, the
researchers employed RSM to optimize the material distribution of foam determined by a power-
law function in the tubes. The thickness of the wall was the other design variable in their study.
The results showed that the optimal design, maximized SEA and minimized PCF were
significantly influenced by whether the variation of materials from the top to the bottom of conical
tubes were ascending or descending. The idea of filling crash boxes with honeycomb instead of
foam inspired Mohammadiha et al. [112] to design and analyze these structures under oblique
impact loads. These authors firstly compared Uniform Honeycomb Filled (UHF) crash boxes with
their Functionally Graded Honeycomb (FGH) counterparts and noticed that the PCF of FGH tubes
was always higher than UHF tubes. Moreover, they attempted to optimize FGH tubes considering
the material density propagation through the box as the design variable. With this intention, four
different material models namely ascending, descending, double ascending and double descending
models were studied. In double patterns, density variation was calculated by two different power-
law functions for each half of the crash box and these models were found to outperform simple
ascending and descending models. Another factor significantly influencing the SEA was the
number of honeycomb grading layers. As this parameter increased, the SEA saw a remarkable rise.
Mohammadiha and Ghariblu [113] attempted to develop optimum results related to crash behavior
of FG tubes and conducted a RSM-based optimization study on multi-tubes filled with FGF. In
their study, sinusoidal density patterns with various numbers of extremum points were considered
as well as ascending and descending patterns determined by a power-law function. According to
the optimal results, ascending pattern led to higher SEA compared to its descending counterpart.
Moreover, considering 11 extremum points resulted in better crashworthiness characteristics than

31
5 extremum pointed sinusoidal pattern. Two Dimensional Functionally Graded Foam-filled Tubes
(TDFGFT) in which the foam density was assumed to vary along the longitudinal and transversal
directions were optimized by Ebrahimi et al. [114] utilizing PSO method. Two power-law
functions with different exponents namely n and m were used in the two aforementioned directions,
respectively. Several models with respect to the range of these exponents were introduced as the
optimal designs. Among all these optimized tubes, the ascending pattern with 0.2 ≤ m,n ≤ 1 and
the descending pattern with 2 ≤ m,n ≤ 10 were proved to have the best outcomes in SEA and PCF,
respectively. The researchers also pointed the noticeable superiority of TDFGFTs over UF tubes.
Uniform Foam-filled Cellular Structures (UFCS) and Functionally Graded Foam-filled Cellular
Structures (FGFCS) were optimized by Yin et al. [115] to enhance their crashworthiness
characteristics under lateral impact loadings. NSGA-II helped these scientists to optimize the
density variation of foam in the FGFCSs which led to maximum SEA and minimum PCF. Nine
different cross sections which were defined by the number of cells (1, 2, 3, 4, 6 and 9) and their
configurations were studied. Among all these cross sections, that with 9 cells represented the
highest SEA regardless of PCF as the constraint. Taking PCF into account, several Pareto fronts
were presented according to the final optimal designs of this research which are an excellent
benchmark for engineers in terms of manufacturing car bodies. Ying et al. [116] utilized RBF
combined with NSGA-II to minimize PCF and maximize the SEA of thin walled tubes made up
of steels with FG strength in which the strength was assumed to vary via a power-law function
along the tube. The gradient exponent was considered as the design variable. Critical impact load
angle was another objective which rose as the gradient exponent approached 1. The results were
validated by experimental and theoretical studies. The researchers also compared FG strength
tubes with their uniform strength counterparts and showed the outperformance of the former over
later regarding their crashworthiness characteristics.

5.2. Functionally graded thickness tubes

FGT tubes were analyzed by Sun et al. [117] in terms of their crashworthiness characteristics. The
thickness tended to vary following a power-law function. The authors compared the SEA and PCF
of these tubes with those of uniform thickness tubes and observed that the former outperformed
the latter. Following this, they implemented RSM and NSGA-II to optimize the power-law index
indicating the variation pattern of thickness aiming to maximize the SEA and minimize the PCF.
They adopted several response surface models to obtain the optimal exponent and demonstrated
that the higher order response surface models did not necessarily represent better results. To
illustrate, in some cases, a 2nd order polynomial function led to better results than its third order
counterparts. The crash behavior of FGT tubes was investigated by Sun et al. [118] when
undergoing lateral impact loads. After a comparison between these tubes and the uniform thickness
tubes, it was claimed that SEA and PCF of FG tubes were higher than those of uniform thickness
tubes. Subsequently, these authors employed PSO method to improve the crashworthiness
behavior of FG tubes under lateral impact by finding the best thickness gradation pattern. Although
the results of some FGT cases could not outperform the uniform tubes in terms of PCF, the
researchers managed to improve the overall crash behavior of FGT tubes by increasing SEA and

32
decreasing PCF to a great extent. RBF models combined with NSGA-II helped Yin et al. [119] to
optimize the thickness distribution pattern of FGT multi-cell tubes whose thicknesses were
assumed to vary based on power-law functions. Improving their crashworthiness behavior defined
by maximizing SEA and minimizing PCF acted as the objective function. Two types of RBF
namely adaptive RBF and conventional static RBF were employed in this research and the former
proved to be more accurate and efficient. FGT tubes made up of aluminum were optimized by
Baykasoglu and Baykasoglu [120] regarding their crash behavior under impact loads. The cross
section of these tubes was circular and the thickness distribution through the longitudinal direction
was examined by a power-law function. The aspect ratio (length to diameter ratio) and the power-
law index formed the design variables of this research. The researchers utilized GA combined with
ANN acting as the objective function evaluator to carry out their optimization approach. Finally,
the Pareto fronts with respect to PCF and SEA as the objective functions were represented.
Depending on the priority of the mentioned objectives, the optimal design could be obtained from
these Pareto fronts. Implementing an integrated NSGA-II and RSM, the same optimization
problem was solved by Yao et al. [121], and a number of interesting results were extracted.
Increasing the aspect ratio and decreasing the power-law exponent led to higher SEA while
increasing the two mentioned design variables resulted in lower PCF. Oblique impact loads were
also investigated in this study by considering 10°, 20°, 30° and 40° of load angles. Increasing the
angle loads caused decreasing trend in the PCF and SEA of the mentioned structures. In order to
make the design of crash tubes more flexible and improve their crashworthiness characteristics,
Yin et al. [122] modeled a new type of these structures in which the foam density varied along the
transverse direction and the thickness of the tube varied along the longitudinal direction. These
structures were referred to as Functionally Graded Foam-filled Graded Thickness (FGFGT) tubes.
According to the final results of FGFGT tubes, the best optimal design was associated with tubes
whose middle was the thickest part. In this case, when the weight of the optimized FGFGT tubes
were equal to that of their UF tubes, the former showed better crashworthiness behavior. On the
other hand, the design with minimum thickness the middle of FGFGT tubes showed lower SEA in
comparison to their UF counterparts with the same weight. The grading pattern of the thickness
and foam density were the design variables of this study.

5.3. Functionally graded thickness inversion tubes

An Inversion tube is another crash structure which consists of a die and a tube. When this structure
undergoes impact loads, the solid die fills the tube and collapses it. Therefore, the energy of impact
load transforms to distortion of the tube. Functionally Graded Thickness Inversion Tubes (FGTIT)
are an innovative design of crash structures in which the thickness of the tube varies gradually
along the longitudinal direction. These structures can be seen in Fig. 11.

33
 Fig.11. (a) An FGTIT specimen under impact loading. (b) A schematic of the die and FGT in FGTIT [123]

Mohammadiha and ghariblou [123] performed a numerical optimization study on the

crashworthiness characteristics of FGTIT which contained a FGT tube and a circular die. Dynamic
Amplification Factor (DAF) with which dynamic deformation performance of the tubes are
assessed was considered as extraordinary objective function along with SEA and PCF. Die radius,
coefficient of friction between die and tube and the gradient index by which the thickness varied
were taken as the design variables in this research. The results demonstrated that concave function
along the tube could significantly increase SEA with no remarkable increase in PCF. The optimum
friction coefficient values were found to be under 0.04 which noticeably enhanced the
crashworthiness characteristics of FGTITs. In another similar study regarding FGTITs subjected
to 5° and 10° oblique impact loads, these authors [124] considered the same objective functions
and design variables. The best friction coefficient was calculated to be under 0.033, and the
optimum die diameter was 3.5mm. Moreover, the researchers proved that the non-linear thickness
variation of the tubes resulted in the best crashworthiness behavior of FGTITs under oblique loads.
The results of this numerical study was verified by experimental tests and in both cases of oblique
and axial impact loads, the authors countered that FGTITs were superior to uniform thickness
tubes.

6. Other Functionally Graded Structures

Numerous advantages of FGMs have attracted many scientists to implement these materials in
different areas such as orthopedic and dental implants, rotating disks, turbine disks, sports

34
equipment, structures composed of piezoelectric materials, etc. These types of structures which
cannot be classified in three main mechanical structures namely beams, plates and shells as well
as tubes, are compiled in this section. Moreover, there are a few studies describing the behavior of
FGMs regardless of the structural shape which are reviewed and presented at the end of this
section.
There are two key parameters in designing a dental implant namely Osseo-integration and bone
remodeling. The mechanical properties of implants’ synthetic materials mismatch the native
biomaterials and this fact deteriorates the two aforementioned essential parameters of dental
implants. Manufacturing these implants from FGM varying from biomaterials to synthetic
materials would be a wise idea to overcome this problem. Accordingly, Lin et al. [125] designed
and constructed a model containing a FG dental implant and four other teeth. They subsequently
calculated the bone remodeling stemming from implementing different FG implants over a four-
year period. The results helped these scientists to optimize the material gradation of FG implants
and enhance the process of bone remodeling which was determined by three specialized functions.
The implant material distribution was assumed to follow a power-law function and the results
showed that reducing the power-law exponent led to higher performance of the implant. However,
reduction in the power-law index resulted in lower stiffness which made the whole design
vulnerable over the healing period. Therefore, Pareto fronts were represented by the researchers
as a benchmark to tailor these types of structures in the most appropriate manner. Following this
research, Sadollah and Bahreininejad [126] utilized Suppapitnarm Multi-Objective Simulated
Annealing (SMOSA) and MOGA to improve the results of RSM taking the same objective
function and design variables into consideration. Their results were improved up to 33% compared
to those of RSM. They also reported that the two meta-heuristic algorithms showed close optimal
results. They finally represented a wider range of optimal power-law index varying from 0.1 to
0.65. A similar study conducted by Sadollah et al. [127] employing MOPSO method sought the
same objectives and led to quite the same results as the previous research. Utilizing the
optimization toolbox of Solidworks software founded on the DOE Rechtschafner design, Ichim et
al. [128] represented a new design of dental implants enveloped by two coatings made of porous
zirconia to minimize the stresses in their surrounding bones. These implants were analyzed when
undergoing loads with angles of 0°, 15°, 30° and 45°. The elastic moduli of the coatings were the
design parameters of this study and the researchers demonstrated that covering the stiff core of the
implant (Ecore = 110Mpa) with softer porous coatings ( Ecoating = 45Mpa ) was sufficient. In other
words, a grading pattern of changing the young modulus from the stiff core to the soft coatings
had negligible impact on stress minimization. According to the final optimum results, these
resilient coatings led the average stresses in peri-implant bone to decrease by 15%. The equivalent
value for compressive and tensile stresses was 50%.
After hip implant surgeries, in order to reduce all the adverse consequences which are attributed
to two main factors namely bone resorption and bone-implant interface instability, Khonaki and
Pasini [129] made an attempt to design a novel graded cellular hip implant. Subsequently, they
determined the bone-implant interface instability and bone resorption by two specific functions
and carried out an NSGA-II-based optimization experiment to minimize both parameters. Material
gradation was chosen as the design variable of their study. In order to illustrate how efficient and

35
superior this novel design was, the researchers compared their optimal design with hip implants
made up of titanium and those constituted from foam. The interface stresses between the implant
and bone as well as bone resorption of the optimized design were found to be 50% and 70% less
than those of Titanium design, respectively. These values were 65% and 53% while comparing the
foam design with optimized graded hip implant in favor of the latter case. Hedia and Fouda [130]
attempted to design artificial hip stems which outperformed the former designs in terms of strength
and life of the joint implant. The rivals of this novel FG design consisting of collagen and
Hydroxyapatite (HAP) were titanium stems with and without HAP coatings. Choosing the
elimination of stress shielding at the medial proximal region of the femur as the objective function,
the authors conducted an optimization approach in which the material distribution and thickness
of their novel design played the role of the design variables. The volume fraction distribution of
material constituents was calculated by a power-law function, and the optimized power-law index
was found to be 0.1. This design demonstrated noticeable declines in the maximum lateral and
medial shear stress while the Von Mises equivalent stress at the medial proximal region increased
significantly. Bahraminasab et al. [131] employed FEM combined with RSM to optimize the
femoral component of knee implants made up of FGM. This optimization study was carried out to
eliminate the loosening problem of knee implants. Therefore, the authors defined maximizing
stresses in the distal femur at the interface of the bone/implant, minimizing micro-motions at the
interface between the bone and the implant and minimizing wear index of polyethylene insert as
the objective functions. The volume fraction of material constituents and their porosity variation
pattern were taken as the design variables in this multi-objective study. The researchers compared
their optimal result with a benchmark alloy made of cobalt and chromium and achieved averages
of 3.8%, 13.6% and 0.6% enhancement in the aforementioned objectives, respectively. Mehboob
and Chang [132] took advantage of ABAQUS software to simulate fractured tibia and the bone
fragments which were assembled by a FG biodegradable bone plate and a number of screws.
Subsequently, mechano-regulation theory was adopted along with deviatoric strain and FEM
analyses to determine the healing performance of the fractured bone. Finally, Taguchi method and
DOE were exploited to maximize the healing performance of the mentioned structures. The
average elasticity modulus, material distribution and thickness of these FG structures were the
design variables of this research and the optimum values of these parameters were obtained after
the optimization process.
Improving the performance of safety helmets motivated Rueda et al. [133] to optimize the foam
liner part without increasing the size or weight of these helmets. Increasing the weight of these
structures leads to higher inertia of the helmets. Therefore, the possibility of neck injuries are
increased. The performance of helmets depends on reduction in peak acceleration which is related
to three key features namely contact areas, distribution of material stresses and dissipated plastic
energy density. The researchers claimed that varying the foam material through the thickness and
increasing the contact area had great influence on the energy absorption of these helmets. In
another study, the same authors [134] decided to use FGF liners to avoid crack propagation and
delamination of liners. After modelling these new designed helmets, they attempted to find the
optimum foam configuration through the thickness of liners which led to the best energy absorbing
behavior. They finally compared FGF liner helmets with UF liner helmets and demonstrated the

36
outperformance of the former when both types were assumed to be subjected to various impact
velocities and positions.
Corrugated rods which contain a number of curvatures with specific steps and amplitude are
implemented as reinforcement in many engineering structures. FG corrugated rods were
introduced and optimized by Andrianov et al. [135] to increase their stiffness. The study succeeded
in optimizing the variation of amplitude and steps of corrugation in these structures and concluded
that these parameters are mostly effective on longitudinal stiffness and compression-extension
stiffness, respectively. In order to demonstrate the outperformance of FG strength structures
consisting a gradual range of low strength to high strength materials over their uniform strength
counterparts in terms of crashworthiness characteristics, Ying et al. [136] analyzed these two types
of materials using nonlinear finite element program LS_DYNA. For this purpose, columns made
of these materials were considered to be under crashing and oblique impact loads. Subsequently,
an NSGA-II-based multi-objective optimization study was conducted to find the best strength
propagation through FG strength columns which led to minimum PCF and maximum SEA.
According to the results, the best crashworthiness behavior was related to columns with the lowest
strength gradation index and designs with the highest values of top strength.
In order to increase the safety of roller compacted concrete gravity dams in terms of seepage and
cracking, Zhang et al. [137] designed dams in which the material properties along the water flow
varied functionally. These dams are called FG partition structures . Implementing optimization
approaches based on GA, Multi-Island Genetic Algorithm (MIGA) and Adaptive Simulated
Annealing (ASA), they managed to minimize the cost of these structure considering their hydraulic
fracture as the constraint. ASA was the fastest algorithm among the mentioned utilized methods
according to the final results.
As disks are vastly implemented in different industrial applications and undergo large amounts of
thermal and centrifugal loads, there is a desire to boost their mechanical characteristics. An
optimization approach founded on Sequential Linear Programming (SLP) was conducted by Chen
and Tong [138] to minimize the total weight of FGM subjected to heat flux and temperature loads.
A long slab, a circular disk and a turbine disk made up of these materials were investigated as
examples. In the last example, the researchers took deflection and maximum equivalent stress as
the constraints and managed to minimize the weight of FG turbine disk by 10%. Utilizing CAMD
method, Stump et al. [139] attempted to minimize the volumetric density of FG rotating and turbine
disks. The material variation profile was the design parameter of their study with the Von Mises
yield criterion as the constraint. The same turbine disks subjected to centrifugal force was
optimized by these authors [140] with the aim of minimizing the volume fraction of one of the
material constituents considering Von Mises equivalent stress as the constraint. SLP was the
optimization method they utilized to optimize the material distribution of FG turbine disks.
Khorsand and Tang [141] employed a combined Co-evolutionary Particle Swarm Optimization
(CPSO) method and DQM to optimize the weight of FGT rotating hollow disks and to minimize
the deflections and centrifugal and Von Mises equivalent stresses of these structures. These disks
were assumed to be under thermo-mechanical loads. The weight reduction results associated with
this approach were found to be greater than those of other methods. The influences of angular

37
velocity along with the thickness variation pattern on the mentioned stresses, deflection and weight
optimization were also evaluated in this research. The material propagation and geometry of FG
rotating hollow disks were optimized by Tharun et al. [142] in order to minimize the mass,
maximum deflection and stresses of these structure when they undergo thermo-mechanical loads.
Mechanical loading was attributed to rotating with certain angular speed and thermal load stem
from temperature gradient along the radial direction. The researchers also considered constant and
variable thickness disks in their study. Finally, they demonstrated that the most significant factors
on the optimality of these structures were the ratio of inner to outer radii and angular speed while
the temperature had the minimum influence on the optimal design of these disks.
The tribological application of coatings encouraged many surface engineers to optimize their
characteristics. The reason lies in the significant influence of coatings on enhancing the
environment sustainability and declining the life-cycle costs of many structures made of sensitive
materials. Nadeau and Ferrari [143] implemented Downhill simplex method to optimize the
material properties distribution of FG composite coatings made up of silicon/carbide and
aluminum. The properties and propagation pattern of the reinforcing fibers embedded in the matrix
of these composites acted as the design variables of this research. The curvature and strain energy
density of these layers were taken as the objective functions. The authors compared the results of
FG coatings with the equivalent optimized bi-layered coatings and claimed that the former could
represent nearly 86% better performance taking both objective functions into account. Xu et al.
[144] employed a modified PSO method to minimize the thermal residual stresses of step formed
five-layered FG carbon/silicon carbide coatings as face sheets of sandwich structures with
carbon/carbon composite core after cooling down. They considered the thicknesses and material
distributions as the design variables. The volume fraction distribution through the thickness of
these coatings followed a power-law function with the exponent varying from 0.02 to 2. With
respect to the optimized results, the best optimized power-law index was nearly 0.08 which led to
the minimum amount of residual stresses. In addition, two completely different optimum designs
were achieved depending on whether 𝑝 ≥ 0.08 or 𝑝 < 0.08.
The application of gears in industries is remarkable, and the most prominent objective of these
structures is to transmit energy. Minimizing the weight of gears has always been important to
engineers since this leads to less energy consumption and higher efficiency of these structures.
Jing et al. [145] utilized Goal Driven Optimization (GDO) method to optimize gear teeth
constituted of FGM. The aim of their research was defined as minimizing the maximum deflection
of the teeth and the total weight of these gears taking the maximum allowable stress as the
constraint. After optimizing the teeth material, the researchers compared the gears with FG teeth
with their totally homogeneous counterparts and reported that the former design had 32% less
weight and 6% less maximum deflection in the teeth. They also recommended an optimization
analysis on gears whose all parts are made of FGM. Considering the printability constraint was
another suggestion the authors mentioned which could result in a more practical design.
Injection molding is the process of producing specific parts in large scale. As the mechanical
stability of injection molds deters during molding cycles, it is important to improve their design
by increasing their load bearing capacity and their resistance against thermal loads. Using FGMs

38
for manufacturing injection molds and optimizing them are of viable options to reach these
objectives. FG lattice injection molds were optimized by Wu et al. [146] to reduce their weight
with their thermal conduction and stiffness being the constraints. Accordingly, a number of lattice
cells were distributed through the core of the injection mold so that the mass of these structures
reduced significantly with a negligible penalty on the thermal and mechanical constraints. The
final optimum design demonstrated nearly 9% and 3.5% reduction in the stiffness and heat
conduction, respectively while the structural weight reduced up to 30%. Although huge amount of
weight increase was saw in this optimization process, enhancements in stiffness and heat
conduction of the optimized injection molds were perceived as compensation.
Piston rings which are located between the cylinder and piston of engine have four principal
functions namely compression gas sealing, lubricating oil film control, heat transfer and supporting
piston in the cylinder. Due to the significant role these structures play in the engine, engineers are
driven to improve their performance. Carvalho et al. [147] optimized the material gradation of
AlSi-CNT FG composites to make a compromise between the wear, fatigue, tensile strength,
hardness and cost of these materials which were intended to be implemented in producing piston
rings. The best results related to fatigue and wear characteristics of these materials was obtained
when the composite was reinforced with 6% wt. CNT. In this case however, the other two
mechanical properties degraded. 2% wt. CNT designs were found to have the best behavior
regarding all the mentioned characteristics while the cost increased significantly. Therefore, the
researchers finally introduced a linearly graded pattern for CNT as the optimal design so that the
surfaces contained 2% wt. CNT and the middle of these structures contained 0% wt. CNT.
There is a long history with regards to materials of golf club heads from wood to steel, titanium
and recently FG porous materials. The main reason for producing different designs of golf club
heads was related to their beauty and functionality. Over the last decade, the dominance of the
latter is getting more and more obvious. In order to improve the performance of golf clubs, Ray et
al. [148] optimized the porous propagation of club heads made up of porous titanium. Three
different objective functions namely reducing the weight, reducing the coefficient of restitution
and increasing the contact time were taken into account, and all of these parameters were found to
be influential on the shot performance of clubs to a great extent. It is noteworthy that the numerical
results of the impact behavior of club heads against the steel balls were extracted from LS DYNA
commercial software.
Piezoelectric materials are novel designed materials which are capable of converting electrical
energy to mechanical energy and vice versa. Since these materials are vastly implemented in
various applications such as computer, military, medical and automotive industries, researchers
are trying to improve their characteristics. Recently, employing the concept of FGM in
piezoelectric materials has enabled scientists to promote the performance of these novel materials.
Energy harvesting of FG piezocomposites which is mostly attributed to the coefficient of
electromechanical coupling and microscopic stresses was maximized in an approach based on
Discrete Material Optimization (DMO) method presented by Vatanabe et al. [149]. Accordingly,
the polarization direction of piezoelectric material and material gradation were selected as the two
design parameters. The researchers compared the FG piezocomposites with their non-FG

39
counterparts. Although the latter case showed lower electromechanical coupling coefficient, the
microscopic stresses in the FG piezocomposites were claimed to be significantly lower than non-
FG designs. Therefore, the FG pattern in piezocomposites resulted in harvesting more electrical
energy. Material gradation in Functionally Graded Piezoelectric Transducers (FGPT) was
optimized through a topology optimization approach represented by Rubio et al. [150]. They
intended to find the maximum eigenfrequency of a selective eigenmode. The optimization process
was conducted through the framework of SLP and CAMD. Looking at the optimal results, the
gradation in the thickness direction was significantly more effective on the eigenfrequencies than
the gradation through the longitudinal direction.
Schaller and Yang [151] implemented swarm search algorithm and GA in an optimization study
with the goal of stress minimization in FGMs. They incorporated a three-layer FGS structure
whose core was constituted from FGM and the face sheets were homogeneous metal and ceramic.
The objective was to minimize the stress of the ceramic surface when this structure was assumed
to be under thermal loads. They finally managed to decrease maximum stress at the ceramic surface
from 45Mpa to 20Mpa by optimizing material gradation and changing the pattern from a linear to
an optimal function. Goupee and Vel [152] implemented NSGA combined with EFG method to
optimize the material distribution of FGM subjected to thermal loadings in two problems. The
objective functions in the first problem were defined as simultaneously minimizing mass and peak
effective stress under temperature gradient. The second problem involved minimizing the mass
and maximizing the safety factor of the structure simultaneously under intense heat flux. In this
case, the maximum temperature experienced by the metal phase as well as the factor of safety
played the role of constraints. The researchers finally prepared Pareto fronts offering a wide range
of minimized mass and maximum equivalent Von Mises stresses as a benchmark to designers. In
order to enhance the heat treatability and minimize the thermal stress distribution of FG slabs in
thermal environments, Molaei Najafabadi et al. [153] investigated the influences of Fourier
number in the thermo-elastic solution and the impact of material distribution of these structures on
their deflection, stresses and temperature. This optimization problem which was solved via quasi-
Newton method was restrained by the critical temperature. The researchers realized that exploiting
materials (whether the ceramic or metal phase) with higher temperature diffusivity led to better
heat treatability. Furthermore, it was claimed that for all values of Fourier number, as the critical
temperature increased, the optimized power-law exponent was seen to be higher.

7. Conclusion

The majority of publications on optimization of FG structures were reviewed and compiled in this
paper. These researches were categorized in accordance with the type of structures. Beams, plates,
shells, tubes and other structures constituted from FGMs were sequentially investigated in this
research. As the key feature of FGMs is their gradual and smooth variation of material constituents,
the most common design variable in the reviewed researches was the material distribution pattern.
Stress distribution, critical buckling load, fundamental frequency and the weight of FG structures

40
were the most popular objectives researchers sought to improve through optimization studies.
Methodologies based on GA and PSO were the most frequently utilized algorithms in the discussed
field. There were a number of studies carried out with the aim of reducing the computational time
and increasing the accuracy. According to these researches, combining stimulating techniques such
as ANN or ANFIS with a metaheuristic algorithm could lead researchers towards these objectives.
To conclude, the following optimization approaches are suggested by the authors to conduct in the
future.
 There were a few optimization studies in which the manufacturability of the FG structure
was taken into consideration. Adding manufacturability to the optimization studies is
recommended since it leads to more practical designs with prospects of being produced in
large scales.
 There were a few studies on FG structures with piezo electric patches or layers. As piezo-
electric materials demonstrated to be influential on the thermo-mechanical behavior of
various designs, it is suggested to investigate the optimum FG structures embedded
between two piezo-electric layers or FG structures with attached piezo patches to their
surfaces.
 The cost of a design is one of the most important issues which should be considered by
designers. However, there were a few studies selecting this prominent feature as the
objective function. Therefore, adding the cost of optimal design of a FG structure to the
optimization procedure is highly recommended.
 The impact performance of FG beams, plates or shells has attracted little attention in the
literature. Therefore, the authors suggest more contribution to the development of optimal
designs for these structures when undergoing impact loads.

References

[1] Nikbakht S, Kamarian S, Shakeri M. A Review on Optimization of Composite Structures Part I:

Laminated Composites. Composite Structures 2018; 195: 158-185.
[2] Lieu QX, Lee D, Kang J, Lee J. NURBS-based modeling and analysis for free vibration and
buckling problems of in-plane bi-directional functionally graded plates. Mechanics of Advanced
Materials and Structures 2018: 1-17.
[3] Lieu QX, Lee S, Kang J, Lee J. Bending and free vibration analyses of in-plane bi-directional
functionally graded plates with variable thickness using isogeometric analysis. Composite
Structures 2018; 192: 434-451.
[4] Vel SS, Pelletier JL. Multi-objective optimization of functionally graded thick shells for thermal
loading. Composite structures 2007; 81(3): 386-400.
[5] Alshabatat NT, Myers K, Naghshineh K. Design of in-plane functionally graded material plates
for optimal vibration performance. Noise Control Engineering Journal 2016; 64(2): 268-278.
[6] Goupee AJ, Vel SS. Two-dimensional optimization of material composition of functionally
graded materials using meshless analyses and a genetic algorithm. Computer methods in applied
mechanics and engineering 2006; 195(44-47): 5926-5948.

41
[7] Tamura I, Tomota Y, Ozawa M. Strength and ductility of Fe-Ni-C alloys composed of austenite
and martensite with various strengths. In Proc. Conf. on Microstructure and Design of Alloys,
Institute of Metals and Iron and Steel Institute, London. 1973; 1(129): 611-615.
[8] Jin ZH, Paulino GH, Dodds Jr RH. Cohesive fracture modeling of elastic–plastic crack growth in
functionally graded materials. Engineering Fracture Mechanics 2003; 70(14): 1885-1912.
[9] Bhattacharyya M, Kapuria S, Kumar AN. On the stress to strain transfer ratio and elastic
deflection behavior for Al/SiC functionally graded material. Mechanics of Advanced Materials
and Structures 2007; 14(4): 295-302.
[10] Nikbakht S, Jedari Salami S, Shakeri M. Three dimensional analysis of functionally graded plates
up to yielding, using full layer-wise finite element method. Composite Structures 2017; 182: 99-
115.
[11] Komarsofla MK, Jedari Salami J., Shakeri M. Thermo elastic-up to yielding behavior of three
dimensional functionally graded cylindrical panel based on a full layer-wise theory. Composite
Structures 2018; 208: 261-275.
[12] Liew KM, Zhao X, Ferreira AJ. A review of meshless methods for laminated and functionally
graded plates and shells. Composite Structures 2011; 93(8): 2031-2041.
[13] Jha DK, Kant T, Singh RK. A critical review of recent research on functionally graded plates.
Composite Structures 2013; 96: 833-849.
[14] Thai HT, Kim SE. A review of theories for the modeling and analysis of functionally graded
plates and shells. Composite Structures 2015; 128: 70-86.
[15] Gupta A, Talha M. Recent development in modeling and analysis of functionally graded materials
and structures. Progress in Aerospace Sciences 2015; 79: 1-14.
[16] Swaminathan K, Sangeetha DM. Thermal analysis of FGM plates–A critical review of various
modeling techniques and solution methods. Composite Structures 2017; 160: 43-60.
[17] Goupee AJ, Vel SS. Optimization of natural frequencies of bidirectional functionally graded
beams. Structural and Multidisciplinary Optimization 2006; 32(6): 473-484.
[18] Almeida SR, Paulino GH, Silva EC. Layout and material gradation in topology optimization of
functionally graded structures: a global–local approach. Structural and Multidisciplinary
Optimization 2010; 42(6): 855-868.
[19] Noh YJ, Kang YJ, Youn SJ, Cho JR, Lim OK. Reliability-based design optimization of volume
fraction distribution in functionally graded composites. Computational Materials Science, 69,
435-442.
[20] Taheri AH, Hassani B. Simultaneous isogeometrical shape and material design of functionally
graded structures for optimal eigenfrequencies. Computer Methods in Applied Mechanics and
Engineering 2014; 277: 46-80.
[21] Yas MH, Kamarian S, Pourasghar A. Application of imperialist competitive algorithm and neural
networks to optimise the volume fraction of three-parameter functionally graded beams. Journal
of Experimental & Theoretical Artificial Intelligence 2014; 26(1): 1-12.
[22] Kamarian S, Yas MH, Pourasghar A, Daghagh M. Application of firefly algorithm and ANFIS
for optimisation of functionally graded beams. Journal of Experimental & Theoretical Artificial
Intelligence 2014; 26(2): 197-209.
[23] Roque CMC, Martins PALS. Differential evolution for optimization of functionally graded
beams. Composite Structures 2015; 133: 1191-1197.
[24] Roque CMC, Martins PALS, Ferreira AJM, Jorge RMN. Differential evolution for free vibration
optimization of functionally graded nano beams. Composite Structures 2016; 156: 29-34.
[25] Vosoughi AR, Darabi A. A new hybrid CG-GAs approach for high sensitive optimization
problems: With application for parameters estimation of FG nanobeams. Applied Soft Computing
2017; 52: 220-230.

42
[26] Pham AH, Vu TV, Tran TM. (2017, August). Optimal Volume Fraction of Functionally Graded
Beams with Various Shear Deformation Theories Using Social Group Optimization. In
International Conference on Advances in Computational Mechanics (pp. 395-408). Springer,
Singapore.
[27] Taati E, Sina N. Multi-objective optimization of functionally graded materials, thickness and
aspect ratio in micro-beams embedded in an elastic medium. Structural and Multidisciplinary
Optimization 2018: 1-21.
[28] Nguyen TT, Lee J. Optimal design of thin-walled functionally graded beams for buckling
problems. Composite Structures 2017; 179: 459-467.
[29] Maalawi KY. Dynamic optimization of functionally graded thin-walled box beams. International
Journal of Structural Stability and Dynamics 2017; 17(09): 1750109.
[30] Xia Q, Wang MY. Simultaneous optimization of the material properties and the topology of
functionally graded structures. Computer-Aided Design 2008; 40(6): 660-675.
[31] Maleki Jebeli S, Shariat Panahi M. An evolutionary approach for simultaneous optimization of
material property distribution and topology of FG structures. Engineering Computations 2015;
32(2): 234-257.
[32] Taheri AH, Suresh K. An isogeometric approach to topology optimization of multi‐material and
functionally graded structures. International Journal for Numerical Methods in Engineering 2017;
109(5): 668-696.
[33] Lieu QX, Lee J. A multi-resolution approach for multi-material topology optimization based on
isogeometric analysis. Computer Methods in Applied Mechanics and Engineering 2017; 323:
272-302.
[34] Cheng L, Bai J, To AC. Functionally graded lattice structure topology optimization for the design
of additive manufactured components with stress constraints. Computer Methods in Applied
Mechanics and Engineering 2019; 344: 334-359.
[35] Alshabatat NT, Naghshineh K. Optimization of natural frequencies and sound power of beams
using functionally graded material. Advances in Acoustics and Vibration 2014: 752631.
[36] Rokni H, Milani AS, Seethaler RJ. Size-dependent vibration behavior of functionally graded
CNT-reinforced polymer microcantilevers: modeling and optimization. European Journal of
Mechanics-A/Solids 2015; 49: 26-34.
[37] He MX, Sun JQ. Multi-objective structural-acoustic optimization of beams made of functionally
graded materials. Composite Structures 2018; 185: 221-228.
[38] Kim NI, Huynh TA, Lieu QX, Lee J. NURBS-based optimization of natural frequencies for
bidirectional functionally graded beams. Archives of Mechanics 2018; 70(4).
[39] Loja MA. On the use of particle swarm optimization to maximize bending stiffness of functionally
graded structures. Journal of Symbolic Computation 2014; 61: 12-30.
[40] Shi JX, Shimoda M. Interface shape optimization of designing functionally graded sandwich
structures. Composite Structures 2015; 125: 88-95.
[41] Correia VMF, Madeira JA, Araújo AL, Soares CMM. Multiobjective optimization of functionally
graded material plates with thermo-mechanical loading. Composite Structures 2019; 207: 845-
857.
[42] Beluch W, Hatłas M. (2016, September). Multiscale Evolutionary Optimization of Functionally
Graded Porous Materials. In International Congress on Technical Diagnostic (pp. 429-438).
Springer, Cham.
[43] Jamshidi M, Arghavani J. Optimal design of two-dimensional porosity distribution in shear
deformable functionally graded porous beams for stability analysis. Thin-Walled Structures 2017;
120: 81-90.
[44] Fares ME, Elmarghany MK, Atta D. The influence of the normal strain effect on the control and
design optimization of functionally graded plates. Composites Part B: Engineering 2015; 77: 440-
453.

43
[45] Correia VMF, Madeira JA, Araújo AL, Soares CMM. Multiobjective optimization of ceramic-
metal functionally graded plates using a higher order model. Composite Structures 2018; 183:
146-160.
[46] Ootao Y, Kawamura R, Tanigawa Y, Nakamura T. Neural network optimization of material
composition of a functionally graded material plate at arbitrary temperature range and temperature
rise. Archive of Applied Mechanics 1998; 68(10): 662-676.
[47] Helal WM, Shi D. Optimum material gradient for functionally graded rectangular plate with the
finite element method. Indian Journal of Materials Science 2014.
[48] Ding S, Wu CP. Optimization of material composition to minimize the thermal stresses induced
in FGM plates with temperature-dependent material properties. International Journal of
Mechanics and Materials in Design 2017: 1-23.
[49] Na KS, Kim JH. Optimization of volume fractions for functionally graded panels considering
stress and critical temperature. Composite structures 2009; 89(4): 509-516.
[50] Ashjari M, Khoshravan MR. Mass optimization of functionally graded plate for mechanical
loading in the presence of deflection and stress constraints. Composite Structures 2014; 110: 118-
132.
[51] Hussien O, Mulani SB. Two-Dimensional Optimization of Functionally Graded Material Plates
Subjected to Buckling Constraints. In 58th AIAA/ASCE/AHS/ASC Structures, Structural
Dynamics, and Materials Conference 2017 (p. 1546).
[52] Hussein OS, Mulani SB. Multi-dimensional optimization of functionally graded material
composition using polynomial expansion of the volume fraction. Structural and Multidisciplinary
Optimization 2017; 56(2): 271-284.
[53] Hussein OS, Mulani SB. Optimization of in-plane functionally graded panels for buckling
strength: Unstiffened, stiffened panels, and panels with cutouts. Thin-Walled Structures 2018;
122: 173-181.
[54] Golbahar Haghighi MR, Malekzadeh P, Rahideh H. Three-dimensional transient optimal
boundary heating of functionally graded plates. Numerical Heat Transfer, Part B: Fundamentals
2011; 59(1): 76-95.
[55] Chiba R, Sugano Y. Optimisation of material composition of functionally graded materials based
on multiscale thermoelastic analysis. Acta Mechanica 2012; 223(5): 891-909.
[56] Chen YT, Xiang S, Zhao WP. Generalized Multiquadrics with Optimal Shape Parameter and
Exponent for Deflection and Stress of Functionally Graded Plates. In Applied Mechanics and
Materials 2015; 709: 121-124. Trans Tech Publications.
[57] Zenkour AM. Generalized shear deformation theory for bending analysis of functionally graded
plates. Applied Mathematical Modelling 2006; 30(1): 67-84.
[58] Lieu QX, Lee J. Modeling and optimization of functionally graded plates under thermo-
mechanical load using isogeometric analysis and adaptive hybrid evolutionary firefly algorithm.
Composite Structures 2017; 179: 89-106.
[59] Kou XY, Parks GT, Tan ST. Optimal design of functionally graded materials using a procedural
model and particle swarm optimization. Computer-Aided Design 2012; 44(4): 300-310.
[60] Huang J, Rapoff AJ. Optimization design of plates with holes by mimicking bones through
nonaxisymmetric functionally graded material. Proceedings of the Institution of Mechanical
Engineers, Part L: Journal of Materials: Design and Applications 2003; 217(1): 23-27.
[61] Venkataraman S, Sankar B. Elasticity analysis and optimization of a functionally graded plate
with hole. In 44th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and
Materials Conference 2003 (p. 1466).
[62] Vel SS, Goupee AJ. (2008, February). Multi‐objective Optimization of Geometric Dimensions
and Material Composition of Functionally Graded Components. In AIP Conference Proceedings
(Vol. 973, No. 1, pp. 610-615). AIP.

44
[63] Nemat-Alla M. Reduction of thermal stresses by composition optimization of two-dimensional
functionally graded materials. Acta mechanica 2009; 208(3-4): 147-161.
[64] Do DT, Lee D, Lee J. Material optimization of functionally graded plates using deep neural
network and modified symbiotic organisms search for eigenvalue problems. Composites Part B:
Engineering 2019; 159: 300-326.
[65] Lieu QX, Lee J. An isogeometric multimesh design approach for size and shape optimization of
multidirectional functionally graded plates. Computer Methods in Applied Mechanics and
Engineering 2019; 343: 407-437.
[66] Silva TAN, Loja MAR. Differential evolution on the minimization of thermal residual stresses in
functionally graded structures. In Computational Intelligence and Decision Making 2013: 289-
299. Springer, Dordrecht.
[67] Ashjari M, Khoshravan MR. Multi-objective optimization of a functionally graded sandwich
panel under mechanical loading in the presence of stress constraint. Journal of the Mechanical
Behavior of Materials 2017; 26(3-4): 79-93.
[68] Torabi K, Afshari H. Optimization of flutter boundaries of cantilevered trapezoidal functionally
graded sandwich plates. Journal of Sandwich Structures & Materials 2017: 1-29.
[69] Lieu QX, Lee J, Lee D, Lee S, Kim D, Lee J. Shape and size optimization of functionally graded
sandwich plates using isogeometric analysis and adaptive hybrid evolutionary firefly algorithm.
Thin-Walled Structures 2018; 124: 588-604.
[70] Na KS, Kim JH. Volume fraction optimization for step-formed functionally graded plates
considering stress and critical temperature. Composite Structures 2010; 92(6): 1283-1290.
[71] Ootao Y, Tanigawa Y, Ishimaru O. Optimization of material composition of functionally graded
plate for thermal stress relaxation using a genetic algorithm. Journal of Thermal Stresses 2000;
23(3): 257-271.
[72] Na KS, Kim JH. Volume fraction optimization of functionally graded composite panels for stress
reduction and critical temperature. Finite Elements in Analysis and Design 2009; 45(11): 845-
851.
[73] Liew KM, He XQ, Meguid SA. Optimal shape control of functionally graded smart plates using
genetic algorithms. Computational Mechanics 2004; 33(4): 245-253.
[74] Nourmohammadi H, Behjat B. Optimum response of functionally graded piezoelectric plates in
thermal environments. Materials Science-Poland 2017; 35(3): 606-617.
[75] Parashkevova L, Ivanova J, Bontcheva N. Optimal design of functionally graded plates with
thermo-elastic plastic behaviour. Comptes Rendus Mecanique 2004; 332(7): 493-498.
[76] Gunes R, Apalak MK, Yildirim M. The free vibration analysis and optimal design of an
adhesively bonded functionally graded single lap joint. International Journal of Mechanical
Sciences 2007; 49(4): 479-499.
[77] Na KS, Kim JH. (2008, February). Volume Fraction Optimization of Functionally Graded
Composite Plates for Stress Reduction and Thermo‐Mechanical Buckling. In AIP Conference
Proceedings (Vol. 973, No. 1, pp. 706-711). AIP.
[78] Zhu F, Chou CC, Yang KH. On the design and optimisation of a multi-functional lightweight
vehicular armour plate with functionally graded foams (FGF). International Journal of Vehicle
Safety 2013; 6(4): 320-332.
[79] Moghaddam JJ, Bagheri A. Suppressing vibration in a functionally graded material plate using
genetic algorithm particle swarm optimization and sliding mode control system. Proceedings of
the Institution of Mechanical Engineers, Part E: Journal of Process Mechanical Engineering 2016;
230(2): 120-133.
[80] Daynes S, Feih S, Lu WF, Wei J. Optimisation of functionally graded lattice structures using
isostatic lines. Materials & Design 2017; 127: 215-223.

45
[81] Setoodeh AR, Shojaee M. Critical buckling load optimization of functionally graded carbon
nanotube‐reinforced laminated composite quadrilateral plates. Polymer Composites 2018; 39(S2):
E853-E868.
[82] Vo-Duy T, Truong-Thi T, Ho-Huu V, Nguyen-Thoi T. Frequency optimization of laminated
functionally graded carbon nanotube reinforced composite quadrilateral plates using smoothed
FEM and evolution algorithm. Journal of Composite Materials 2018; 52(14): 1971-1986.
[83] Hussein OS, Mulani SB. Reliability analysis and optimization of in-plane functionally graded
CNT-reinforced composite plates. Structural and Multidisciplinary Optimization 2018: 1-12.
[84] Zhu H, Sankar BV, Haftka RT, Venkataraman S, Blosser M. Optimization of functionally graded
metallic foam insulation under transient heat transfer conditions. Structural and multidisciplinary
optimization 2004; 28(5): 349-355.
[85] Venkataraman S, Haftka RT, Sankar B, Zhu H, Blosser ML. Optimal functionally graded metallic
foam thermal insulation. AIAA journal 2004; 42(11): 2355-2363.
[86] Zhang X, Zhang H. Optimal design of functionally graded foam material under impact loading.
International Journal of Mechanical Sciences 2013; 68: 199-211.
[87] Aboudi J, Pindera MJ, Arnold SM. Microstructural optimization of functionally graded
composites subjected to a thermal gradient via the coupled higher-order theory. Composites Part
B: Engineering 1997; 28(1-2): 93-108.
[88] Birman V, Byrd LW. (2008, February). Methodology for selection of optimum light stringers in
functionally graded panels designed for prescribed fundamental frequency or buckling Load. In
AIP Conference Proceedings (Vol. 973, No. 1, pp. 371-376). AIP.
[89] Hedia H, Aldousari S, EI Midany TT, Kamel M. Optimization of material composition of
functionally graded materials plate containing a hole using FEM. Metall 2009; 63(5).
[90] Ootao Y, Tanigawa Y, Nakamura T. Optimization of material composition of FGM hollow
circular cylinder under thermal loading: a neural network approach. Composites Part B:
Engineering 1999; 30(4): 415-422.
[91] Cho JR, Ha DY. Volume fraction optimization for minimizing thermal stress in Ni–Al2O3
functionally graded materials. Materials Science and Engineering: A 2002; 334(1-2): 147-155.
[92] Huang J, Lu YB, Shen C. Thermal elastic-plastic limit analysis and optimal design for composite
cylinders of ceramic/metal functionally graded materials. In Materials Science Forum 2003; 423:
681-686. Trans Tech Publications.
[93] Choi JH, Cho JR. Optimum Material Design of Metal-Ceramic Hybrid Functionally Graded
Composite. In Materials Science Forum 2008; 569: 121-124. Trans Tech Publications.
[94] Batra RC. Optimal design of functionally graded incompressible linear elastic cylinders and
spheres. AIAA journal 2008; 46(8): 2050-2057.
[95] Asgari M. Material optimization of functionally graded heterogeneous cylinder for wave
propagation. Journal of Composite Materials 2016; 50(25): 3525-3528.
[96] Biglar M, Mirdamadi HR. Configuration optimization of piezoelectric patches attached to
functionally graded shear-deformable cylindrical shells considering spillover effects. Journal of
Intelligent Material Systems and Structures 2016; 27(3): 295-313.
[97] Nabian M, Ahmadian MT. (2011, January). Multi-Objective Optimization of Functionally Graded
Hollow Cylinders. In ASME 2011 International Mechanical Engineering Congress and
Exposition (pp. 583-590). American Society of Mechanical Engineers.
[98] Tornabene F, Ceruti A. Mixed static and dynamic optimization of four-parameter functionally
graded completely doubly curved and degenerate shells and panels using GDQ method.
Mathematical Problems in Engineering 2013: 1-33.
[99] Nouri A, Astaraki S. Optimization of sound transmission loss through a thin functionally graded
material cylindrical shell. Shock and Vibration 2014.

46
[100] Fares ME, Elmarghany MK, Atta D. Suppressing vibrational response of functionally graded
truncated conical shells by active control and design optimization. Thin-Walled Structures 2018;
122: 480-490.
[101] Moita JS, Araújo AL, Correia VF, Soares CMM, Herskovits J. Material distribution and sizing
optimization of functionally graded plate-shell structures. Composites Part B: Engineering 2018;
142: 263-272.
[102] Maalawi K, El-Sayed H. (2011, July). Stability Optimization of Functionally Graded Pipes
Conveying Fluid. In Proceedings of the International Conference of Mechanical Engineering,
World Academy of Science, Engineering and Technology, Paris, France (pp. 178-183).
[103] Wang ZW, Zhang Q, Xia LZ, Wu JT, Liu PQ. Stress analysis and parameter optimization of an
FGM pressure vessel subjected to thermo-mechanical loadings. Procedia Engineering 2015; 130:
374-389.
[104] Shabana YM, Elsawaf A, Khalaf H, Khalil Y. Stresses minimization in functionally graded
cylinders using particle swarm optimization technique. International Journal of Pressure Vessels
and Piping 2017; 154: 1-10.
[105] Boussaa D. Optimization of temperature-dependent functionally graded material bodies.
Computer Methods in Applied Mechanics and Engineering 2009; 198(37-40): 2827-2838.
[106] Fereidoon A, Sadri F, Hemmatian H. Functionally graded materials optimization using particle
swarm-based algorithms. Journal of Thermal Stresses 2012; 35(4): 377-392.
[107] Yin H, Wen G, Hou S, Qing Q. Multiobjective crashworthiness optimization of functionally
lateral graded foam-filled tubes. Materials & Design 2013; 44: 414-428.
[108] Yin H, Wen G, Fang H, Qing Q, Kong X, Xiao J, Liu Z. Multiobjective crashworthiness
optimization design of functionally graded foam-filled tapered tube based on dynamic ensemble
metamodel. Materials & Design 2014; 55: 747-757.
[109] Fang J, Gao Y, Sun G, Zhang Y, Li Q. Parametric analysis and multiobjective optimization for
functionally graded foam-filled thin-wall tube under lateral impact. Computational Materials
Science 2014; 90: 265-275.
[110] Li G, Zhang Z, Sun G, Xu F, Huang X. Crushing analysis and multiobjective optimization for
functionally graded foam-filled tubes under multiple load cases. International Journal of
Mechanical Sciences 2014; 89: 439-452.
[111] Mohammadiha O, Beheshti H. Optimization of functionally graded foam-filled conical tubes
under axial impact loading. Journal of Mechanical Science and Technology 2014; 28(5): 1741-
1752.
[112] Mohammadiha O, Beheshti H, Aboutalebi FH. Multi-objective optimisation of functionally
graded honeycomb filled crash boxes under oblique impact loading. International journal of
crashworthiness 2015; 20(1): 44-59.
[113] Mohammadiha O, Ghariblu H. Crush behavior optimization of multi-tubes filled by functionally
graded foam. Thin-Walled Structures 2016; 98: 627-639.
[114] Ebrahimi S, Vahdatazad N, Liaghat G. Crashworthiness efficiency optimisation for two-
directional functionally graded foam-filled tubes under axial crushing impacts. International
Journal of Crashworthiness 2017; 22(3): 307-321.
[115] Yin H, Chen C, Hu T, Wen G. Optimisation for bending crashworthiness of functionally graded
foam-filled cellular structure. International Journal of Crashworthiness 2017: 1-15.
[116] Ying L, Dai M, Zhang S, Ma H, Hu P. Multiobjective crashworthiness optimization of thin-walled
structures with functionally graded strength under oblique impact loading. Thin-Walled
Structures 2017; 117: 165-177.
[117] Sun G, Xu F, Li G, LiQ. Crashing analysis and multiobjective optimization for thin-walled
structures with functionally graded thickness. International Journal of Impact Engineering 2014;
64: 62-74.

47
[118] Sun G, Tian X, Fang J, Xu F, Li G, Huang X. Dynamical bending analysis and optimization design
for functionally graded thickness (FGT) tube. International Journal of Impact Engineering 2015;
78: 128-137.
[119] Yin H, Fang H, Wen G, Wang Q, Xiao Y. An adaptive RBF-based multi-objective optimization
method for crashworthiness design of functionally graded multi-cell tube. Structural and
Multidisciplinary Optimization 2016; 53(1): 129-144.
[120] Baykasoğlu A, Baykasoğlu C. Multiple objective crashworthiness optimization of circular tubes
with functionally graded thickness via artificial neural networks and genetic algorithms.
Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical
Engineering Science 2017; 231(11): 2005-2016.
[121] Yao S, Xing Y, Zhao K. Crashworthiness analysis and multiobjective optimization for circular
tubes with functionally graded thickness under multiple loading angles. Advances in Mechanical
Engineering 2017; 9(4): 1687814017696660.
[122] Yin H, Dai J, Wen G, Tian W, Wu Q. Multi-Objective Optimization Design of Functionally
Graded Foam-Filled Graded-Thickness Tube Under Lateral Impact. International Journal of
Computational Methods 2017, 1850088.
[123] Mohammadiha O, Ghariblu H. Multi-objective optimization of functionally graded thickness
tubes under external inversion over circular dies. International Journal of Mechanical and
Materials Engineering 2016; 11(1): 8.
[124] Mohammadiha O, Ghariblu H. Optimal shape design of functionally graded thickness inversion
tubes subjected to oblique loading. Structural and Multidisciplinary Optimization 2017; 56(3):
587-601.
[125] Lin D, Li Q, Li W, Zhou S, Swain MV. Design optimization of functionally graded dental implant
for bone remodeling. Composites Part B: Engineering 2009; 40(7): 668-675.
[126] Sadollah A, Bahreininejad A. Optimum gradient material for a functionally graded dental implant
using metaheuristic algorithms. Journal of the mechanical behavior of biomedical materials 2011;
4(7): 1384-1395.
[127] Sadollah A, Bahreininejad A, Eskandar H, Hamdi M. Optimum material gradient for functionally
graded dental implant using particle swarm optimization. In Advanced Materials Research 2013;
647: 30-36. Trans Tech Publications.
[128] Ichim PI, Hu X, Bazen JJ, Yi W. Design optimization of a radial functionally graded dental
implant. Journal of Biomedical Materials Research Part B: Applied Biomaterials 2016; 104(1):
58-66.
[129] Khanoki SA, Pasini D. Multiscale design and multiobjective optimization of orthopedic hip
implants with functionally graded cellular material. Journal of biomechanical engineering 2012;
134(3): 031004.
[130] Hedia HS, Fouda N. Design optimization of cementless hip prosthesis coating through
functionally graded material. Computational Materials Science 2014; 87: 83-87.
[131] Bahraminasab M, Sahari BB, Edwards KL, Farahmand F, Hong TS, Arumugam M, Jahan A.
Multi-objective design optimization of functionally graded material for the femoral component of
a total knee replacement. Materials & Design 2014; 53: 159-173.
[132] Mehboob H, Chang SH. Optimal design of a functionally graded biodegradable composite bone
plate by using the Taguchi method and finite element analysis. Composite Structures 2015; 119:
166-173.
[133] Rueda MF, Cui L, Gilchrist MD. Optimisation of energy absorbing liner for equestrian helmets.
Part I: Layered foam liner. Materials & Design 2009; 30(9): 3405-3413.
[134] Cui L, Rueda MF, Gilchrist MD. Optimisation of energy absorbing liner for equestrian helmets.
Part II: Functionally graded foam liner. Materials & Design 2009; 30(9): 3414-3419.

48
[135] Andrianov IV, Awrejcewicz J, Diskovsky AA. Optimal design of a functionally graded
corrugated rods subjected to longitudinal deformation. Archive of Applied Mechanics 2015;
85(2): 303-314.
[136] Ying L, Dai M, Zhang S, Ma H, Hu P. Multiobjective crashworthiness optimization of thin-walled
structures with functionally graded strength under oblique impact loading. Thin-Walled
Structures 2017; 117: 165-177.
[137] Zhang M, LiM, Shen Y, Zhang J. Isogeometric shape optimization of high RCC gravity dams
with functionally graded partition structure considering hydraulic fracturing. Engineering
Structures 2019; 179: 341-352.
[138] Chen B, Tong L. Thermomechanically coupled sensitivity analysis and design optimization of
functionally graded materials. Computer methods in applied mechanics and engineering 2005;
194(18-20): 1891-1911.
[139] Stump FV, Silva EC, Paulino GH. Optimization of material distribution in functionally graded
structures with stress constraints. Communications in numerical methods in engineering 2007;
23(6): 535-551.
[140] Stump FV, Silva EC, Paulino GH. (2008, February). Topology optimization with stress
constraints: reduction of stress concentration in functionally graded structures. In AIP Conference
Proceedings (Vol. 973, No. 1, pp. 303-308). AIP.
[141] Khorsand M, Tang Y. Design functionally graded rotating disks under thermoelastic loads:
Weight optimization. International Journal of Pressure Vessels and Piping 2018; 161: 33-40.
[142] Tharun P, Siddarth MD, Prakash D, Babu K. Analysis and Optimization on Functionally Graded
Rotating Disk Using Grey Relational Method. In Advances in Manufacturing Processes 2019 (pp.
297-308). Springer, Singapore.
[143] Nadeau JC, Ferrari M. Microstructural optimization of a functionally graded transversely
isotropic layer. Mechanics of Materials 1999; 31(10): 637-651.
[144] Xu Y, Zhang W, Chamoret D, Domaszewski M. Minimizing thermal residual stresses in C/SiC
functionally graded material coating of C/C composites by using particle swarm optimization
algorithm. Computational Materials Science 2012; 61: 99-105.
[145] Jing S, Zhang H, Zhou J, Song G. Optimum weight design of functionally graded material gears.
Chinese Journal of Mechanical Engineering 2015; 28(6): 1186-1193.
[146] Wu T, Liu K, Tovar A. (2016, August). Multiphase thermomechanical topology optimization of
functionally graded lattice injection molds. In ASME 2016 International Design Engineering
Technical Conferences and Computers and Information in Engineering Conference (pp.
V02AT03A036-V02AT03A036). American Society of Mechanical Engineers.
[147] Carvalho O, Buciumeanu M, Madeira S, Soares D, Silva FS, Miranda G. Optimization of AlSi–
CNTs functionally graded material composites for engine piston rings. Materials & Design 2015;
80: 163-173.
[148] Ray P, Chahine G, Smith P, Kovacevic R. (2011, August). Optimal Design of a Golf Club using
Functionally Graded Porosity. In The 22 International SFF Symposium–An Additive
Manufacturing Conference.
[149] Vatanabe SL, Paulino GH, Silva ECN. Design of functionally graded piezocomposites using
topology optimization and homogenization–Toward effective energy harvesting materials.
Computer Methods in Applied Mechanics and Engineering 2013; 266: 205-218.
[150] Rubio WM, Silva ECN. Design of Functionally Graded Piezoelectric Transducers Maximizing
Selective Eigenfrequency by Using Topology Optimization Method. International Conference on
Engineering Optimization 2008.
[151] Schaller W, Yang YY. Stress optimization in applications with functionally graded materials. In
Materials science forum 1999; 308: 942-947. Trans Tech Publications.
[152] Goupee AJ, Vel SS. Multi-objective optimization of functionally graded materials with
temperature-dependent material properties. Materials & design 2007; 28(6): 1861-1879.

49
[153] Molaei Najafabadi M, Taati E, Basirat Tabrizi H. Optimization of functionally graded materials
in the slab symmetrically surface heated using transient analytical solution. Journal of Thermal
Stresses 2014; 37(2): 137-159.

50