Advances in Engineering Software 112 (2017) 231–254

Contents lists available at ScienceDirect

Advances in Engineering Software

journal homepage: www.elsevier.com/locate/advengsoft

An astrophysics-inspired Grey wolf algorithm for numerical

optimization and its application to engineering design problems
Vijay Kumar a,∗, Dinesh Kumar b
a
Computer Science and Engineering Department, Thapar University, Patiala, Punjab, India
b
Computer Science and Engineering Department, GJUS&T, Hisar, Haryana, India

a r t i c l e i n f o a b s t r a c t

Article history: In this paper, modified schemes are proposed for preventing a grey wolf optimizer (GWO) from pre-
Received 26 February 2017 mature exploration and convergence on optimization problems. Three novel strategies are developed to
Revised 4 May 2017
improve the performance of existing GWO. The first strategy uses the concept of prey weight. The sec-
Accepted 14 May 2017
ond strategy uses the astrophysics concepts, which guide the grey wolves toward more promising areas
Available online 22 May 2017
of the search space. The beauty of this strategy is to let each grey wolf learn from not only movement
Keywords: of sun (symbolizes prey) in the search space but also the wolves are made to explore and exploit si-
Meta-heuristics multaneously. Third strategy combines the both, first and second strategies to take advantages of prey
Grey wolf optimizer weight and astrophysics strategies. The proposed improvements in GWO have been evaluated on thirteen
Astrophysics concept benchmark test functions. The performance of the proposed modifications has been compared with other
Function optimization five recently developed state-of-the-art techniques. The effects of scalability, noise, and control parame-
Constrained design problems
ter have also been investigated. The statistical tests have been performed to validate the significance of
modified variants. The proposed variants are also applied for seven well-known constrained engineering
design problems. The experimental results depict the supremacy of the proposed modified algorithm in
solving engineering design problems when compared with several existing techniques.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction Recently, Mirjalili et al. [15] developed a novel metaheuristic al-

Optimization is an active field of research, which explores the
best parameter values of a given problem under the specified con-
ditions. The main aim of optimization is to achieve the relevant
parameter values that allow an objective function to produce min-
ima or maxima. The formulation of an optimization problem be-
gins with the devise of an objective function. Many real-life prob-
lems, from various domains, are viewed as optimization of contin-
uous functions. The classical optimization algorithms fail to pro-
vide a satisfactory solution for many real-life problems due to their
size, dependency on algorithms, problem dimensions etc. There-
fore, metaheuristic algorithms took their place to tackle optimiza-
tion problems. To name a few of them are genetic algorithm (GA)
[7], particle swarm optimization (PSO) [11], ant colony optimiza-
tion (ACO) [2], gravitational search algorithm (GSA) [20], moth
flame optimization (MFO) [16], sine cosine algorithm (SCA) [18],
multiverse optimization (MVO) [17].
∗
Corresponding author.
E-mail addresses: vijaykumarchahar@gmail.com, vijay_kumar@thapar.edu (V. Ku-
mar), dinesh_chutani@yahoo.com (D. Kumar).
gorithm namely grey wolf optimizer (GWO), which mimics the so-
cial leadership and hunting behaviour of grey wolves. GWO is pre-
ferred over other metaheuristic techniques as it is easy to imple-
ment and has fewer control parameters. However, GWO and its ex-
isting variants are yet to prove their supremacy and show satisfac-
tory performance on some benchmarks. They also suffer from pre-
mature convergence, i.e., the most of the grey wolves lose their di-
versity before their convergence towards some solutions and hence
converge prematurely. This may be attributed to the fact that the
wolves lack information sharing among them. [12]. GWO needs a
better trade-off between exploitation and exploration. This neces-
sitated the introduction of new concepts/strategies incorporation
of which further improves GWO. These new concepts are inspired
from astrophysics. These strategies provide better balance between
exploration and exploitation and are able to find the optimal solu-
tion.
The proposed strategies utilize the basic concept of prey weight
and astrophysics-based learning [24]. The first strategy and hence
the algorithm called Modified Grey Wolf Algorithm-I (MGWO-I),
uses the concept of prey weight. It is responsible for exploration
of search space. The second strategy proposes an astrophysics-
based learning strategy to improve the performance of GWO for

http://dx.doi.org/10.1016/j.advengsoft.2017.05.008
0965-9978/© 2017 Elsevier Ltd. All rights reserved.
232 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254
solving optimization problems and the algorithm thus proposed is
called Modified Grey Wolf Algorithm-II (MGWO-II). This learning
strategy uses the elliptical orbit concept for wolves. The ellipti-
cal orbit is responsible for both exploration and exploitation. This
astrophysics-based learning strategy is incorporated in GWO to
speed up the search and increase the efficiency of search process.
The third strategy combines the both above-mentioned strategies
and hence the algorithm called Modified Grey Wolf Algorithm-
III (MGWO-III). The performance of the proposed modifications in
GWO has been evaluated on thirteen benchmark test functions.
These are compared with the original version of GWO and five
recently developed metaheuristic techniques. The effects of scal-
ability and noise have also been investigated over three variants
of GWO. The different values of control parameters have also been
studied. The proposed approach is also employed for engineering
structural design problems.
The remainder of this paper is organized as follows.
Section 2 describes the concepts of GWO and introduces an
overview of related works. The details of proposed modifications
are presented in Section 3. Section 4 presents the benchmark test
functions, experimental setups and results obtained. Section 5 in-
troduces the application of proposed algorithms on structural
design problems. Eventually, Section 6 draws the conclusions from
the experimental results.
2. Background
In this section, we first describe the basics of GWO followed by
brief description of recently developed modifications in GWO.
2.1. Classical Grey wolf optimizer
Grey wolf optimizer algorithm (GWO) was developed by Mir-
jalili et al. [15]. It is a recently developed metaheuristic algorithm
that mimics the leadership hierarchy and hunting behavior of grey
wolves in nature. The grey wolves are categorized into four groups
such as α , β , δ , and ω wolves [23]. These are involved in simu-
lating the leadership hierarchy. The α wolves are leading wolves.
They are responsible for taking decisions during the hunting pro-
cess. The second level of dominated wolves, in the group, is β
wolves. When α wolves die or become old, then they are upgraded
to α wolves. The δ wolves control the ω wolves and provide the
information to α and β wolves. The lowest level in wolves’ hierar-
chy is ω wolves. Besides social hierarchy, they have an interesting
hunting behaviour. The main steps of hunting process are: search-
ing for prey, encircling the prey, and attacking the prey. The fol-
lowing two equations model the encircling behaviour of the grey
wolves [15].
Dist = |C × YP (tcur ) − Y (tcur )|
Y (tcur + 1 ) = YP (tcur ) − A × Dist
where Dist represents the distance between the position of both
the prey (YP ) and a grey wolf (Y). tcur is the current iteration.
YP (tcur ) represents the position vector of prey. A and C are the co-
efficient vectors calculated as follows:
A = 2 × au × rand1 − au
C = 2 × rand2
The value of au is linearly decreased from 2 to 0 during the
iterations. The rand1 and rand2 are random variables and may have
any value in the range of [0, 1]. The exploration and exploitation
decisions are made based on the value of vector A.
It is assumed thatα , β , and δ have better knowledge of the po-
sition of the prey. Hence, these three solutions are saved for hunt-
ing mechanism. The position of ω is updated based on the posi-
tions of α , β , and δ . These are mathematically modelled as follows
[3].
 
Distα = |C1 × Yα − Y | Distβ = C2 × Yβ − Y  Distδ = |C3 × Yδ − Y |
(5)
Y1 = Yα − A1 × Distα Y2 = Yβ − A2 × Distβ Y3 = Yδ − A3 × Distδ
(6)
Y1 + Y2 + Y3
Y3 = (7)
3
Fig. 1(a) shows the flowchart of GWO.
2.2. Related works
In the past few years, it has been observed that grey wolf op-
timizer (GWO) has been adopted by many researchers to solve the
optimization problems. A number of improvements in GWO were
also proposed from time to time to increase its performance.
Kamboj [9] developed a novel approach combining both PSO
and GWO. This approach suffers from premature convergence
problem. Jitkongcheun [8] used an invasion-based migration oper-
ation, which was integrated with GWO. This migration was respon-
sible for information exchange within population and produced
new individuals. However, it fails to converge in problems that
possess high dimensionality. Saremi et al. [23] utilized the evolu-
tionary population dynamics (EPD) concept in GWO. EPD was used
to remove the poor search agents of GWO and reposition them to
enhance the exploitation. However, it requires random positioning
mechanism to avoid the local optima. Zhang and Zhou [28] pre-
sented an extended GWO algorithm based on Powell local opti-
mization method named as PGWO. The convergence property of
Powell’s method was utilized to improve the performance of GWO.
Zhu et al. [29] presented a hybridization of GWO and differential
evolution (DE). This approach is known as hybrid GWO (HGWO).
They have utilized the search capability of DE to remove the stag-
nation problem and to improve the convergence speed.
Kishor and Singh [12] provided an empirical study of classical
GWO. They proposed a modification in GWO, which improves the
information sharing mechanism of GWO and also improves its per-
formance. However, it suffers from premature convergence prob-
lem. Mittal et al. [19] developed a modified GWO (mGWO) to im-
prove the performance of GWO. They have introduced exponen-
tial decay function for fine tuning of parameters. The decay func-
tion utilizes the coefficient parameter, A. However, it suffers from
(1) premature convergence problem. Emary et al. [4] proposed a bi-
nary GWO (BGWO) for feature selection. They used two approaches
for converting the classical GWO into binary form. They developed
(2)
a classification-based fitness function to eliminate the redundant
data, which was optimized through BGWO.
Wen et al. [25] presented a hybrid GWO which combines
chaotic mapping and elite opposition based learning strategy to
improve the performance. The chaotic sequence was used to ini-
tiate position of individuals. The elite opposition based learning
(3) strategy is applied to elite individuals for better exploration. Wen
[26] designed a nonlinear adjustment strategy of control parame-
(4) ter au in GWO, known as improved GWO (IGWO). The opposition-
based learning strategy is introduced to initialize the population.
However, it fails to converge in high dimensionality problems.
Emary et al. [5] proposed a new variant of GWO (EGWO) that
uses reinforcement learning principles with neural networks, to
enhance the performance. The reinforcement learning is used to
 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254 233

Fig. 1. Flowchart of Grey wolf algorithms: (a) GWO (b) MGWO-I (c) MGWO-II (d) MGWO-III.
234 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254

Fig. 2. Convergence characteristics of GWO on benchmark functions (a) F1 (b) F13 .

set the parameter’s value based on individual agent. The neural Table 1
Unimodal benchmark test functions.
network is used to map a set of agents’ states to corresponding
actions that influence the exploration rate. The main drawback of Benchmark Test Function Search Range Dim.
this technique is that it employs complex methodology. 
D
F1 (Y ) = y2i [ − 100, 100] 30
Although these works maintain a balance between intensifi- i=1
cation and diversification, still these are far from yielding opti- D D
F2 (Y ) = |yi | + i=1 |yi | [ − 10, 10] 30
mal solution. Hence, there is a scope to think of some new con- i=1
2
D 
i
cepts and/or strategies that may help provide a better balance be- F3 (Y ) = ( yj) [ − 100, 100] 30
i=1 j=1
tween the two (intensification and diversification). This motivates
F4 (Y ) = max{|yi |, 1 ≤ i ≤ D} [ − 100, 100] 30
us to develop a modified algorithm based on new proposed con- D
i
−1
2
F5 (Y ) = [100(yi+1 − ) + ( yi − 1 ) ]
2
cepts/strategies. y2i [ − 30, 30] 30
i=1

D

3. The proposed Grey wolf algorithm variants F6 (Y ) = (|yi + 0.5| ) 2

[ − 100, 100] 30
i=1
D
F7 (Y ) = iy4i + random(0, 1) [ − 1.28, 1.28] 30
3.1. Motivation i=1

After the detailed study and investigation of GWO, it has been

observed that it has an obliging procedure to balance the explo-
ration and exploitation. The exploration capability of GWO mainly
depends upon the coefficient vector C. The other important coef-
ficient vector A, dictates whether the algorithm will exploit or ex-
plore. During the course of action, GWO may, however, get stuck
in local optima for some optimization problems. Moreover, it has
also been noticed that there is a scope to further increase the ex-
ploration capability of GWO. Hence, GWO needs modifications.
Based on the above facts, three novel modification approaches
have been proposed. In first approach, a modification in coefficient
vector C has been done to further improve the exploration capabil-
ity of GWO. The second approach uses the concept of astrophysics
to improve exploitation and exploration capabilities. The third ap-
proach combines both first and second approaches together to take
advantages of these approaches. The succeeding subsections ex-
plain the justification and concept behind these approaches.
computation [13] (Eq. 2). It emphasizes the effect of prey’s weight
provided C is greater than 1, otherwise it reduces the effect. Based
on this fact, the novel modification has been proposed in the con-
trol parameter C. This proposed approach improves the exploita-
tion of GWO. The value of Cchanges dynamically with the number
of iterations and the expression is given below:
C = 2 × rand2 − (au/2 ) (8)
3.2.1. Justification of proposed approach 1
Fig. 2(a) and 2 (b) show the evaluation of fitness function over
generations on two benchmark test functions taken from Tables
1 and 2. We have taken one unimodal function F1 and one mul-
timodal function F13 . The graph plots show that GWO with C > 1
provides better exploitation than GWO with C < 1. The graphs also
show that C > 1 provides better exploration of search space during
earlier generations as depicted.

3.2. Proposed approach 1: prey weight strategy 3.2.2. Modified Grey wolf Algorithm-I
The modified grey wolf algorithm-I (MGWO-I) follows the basic
The value of the important control parameter C, responsible for steps of GWO as mentioned in Section 2.1. The above-mentioned
exploration process, indicates the weight of the prey in distance proposed equation is introduced in Step 2. The Eq. (4) is replaced
 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254 235

Table 2
Multimodal benchmark test functions.

Benchmark Test Function Search Range Dim.


D 
F8 (Y ) = −yi sin( |yi | ) [ − 50 0, 50 0] 30
i=1
D
F9 (Y ) = [y2i − 10 cos(2π yi ) + 10] [ − 5.12, 5.12] 30
i=1     

D 
D
F10 (Y ) = −20 exp −0.2 1
D
y2i − exp 1
D
cos(2π yi ) + 20 + e [ − 32, 32] 30
i=1 i=1

D D

F11 (Y ) = 1
40 0 0
Yi2 − i=1 cos y
√i +1 [ − 60 0, 60 0] 30
i
i=1
D
−1
F12 (Y ) = πD {10 sin(π z1 ) + (zi − 1 )2 [1 + 10sin2 (π zi+1 )]}.....
i=1
D
..... + πD { (zD − 1 ) } + u(yi , 10, 100, 4)
2

i=1 [ − 50, 50] 30

k ( yi − a )m yi > a
zi = 1 + yi 4+1 , u(yi , a, k, m) = {0 − a < yi < a
k(−yi − a )m yi < −a

D
F13 (Y ) = 0.1{sin (3π y1 ) + (yi − 1 )2 [1 + sin2 (3π yi + 1)]}.....
2

i=1
[ − 50, 50] 30

D
..... + 0.1{ (yD − 1 ) [1 + sin (2π yD )]} + u(yi , 5, 100, 4)
2 2

i=1

with the proposed Eq. (8). The flowchart of MGWO-I is shown in Here, Ri, α indicates the Euclidean distance computed between
Fig. 1(b). The proposed approach of MGWO-I consists of following Yi and Yα . U( − 2, 2) represents uniformly distributed random num-
the steps. ber in the interval [ − 2, 2]. tcur is the current iteration. The reason
behind using U( − 2, 2) is that the proposed algorithm exploits if
Algorithm (MGWO-I).
U( − 2, 2) returns value near to 1, otherwise it explores.
Step 1. Initialize the algorithm parameters, such as number of gray The positions of K grey wolves (Yi , i = 1, ...., K) are updated only
wolves (or search agent), maximum number of iterations when their new positions are better than the original positions,
(MaxIteration), and the control coefficient (au). described as under:
Step 2. The positions of grey wolves are initialized randomly in the 
Yi, new (tcur + 1 ), i f F it (Yi, new ) < F it (Yi )
search space. Yi (tcur + 1 ) = (11)
Yi (tcur ), otherwise
Step 3. Compute the fitness value of each search agent which rep-
resents the distance between wolf and prey.
Step 4. Repeat the following steps until the maximum number of 3.3.1. Justification of proposed astrophysics strategy
iterations is reached: The rationale behind the proposed astrophysics strategy is that
(a) Based on the fitness values, the best (Yα ), second best planets move round the Sun in elliptical orbit. Due to this, the
(Yβ ), and third best (Yδ )solutions are identified. planets have different positions with respect to Sun at different
(b) Compute control coefficients parameters i.e., A and C us- times. Some planets are much closer to the Sun than others mean-
ing Eqs. (3) and (8) respectively. ing thereby that they (the closer ones) are best candidates for ex-
(c) Modify the position of grey wolves using Eq. (7). ploitation. Those, far from the Sun, are in better position to explore
(d) Update the fitness value using the modified position of whole search space. The elliptical orbital shape, in itself, is respon-
the grey wolves. sible for simultaneous exploration and exploitation. Here, the Sun
Step 5. The best grey wolf (Yα ) will yield the optimal solution at is analogous to the prey and planets are analogous to grey wolves.
the final iteration. Fig. 4 depicts how the elliptical orbit concept is responsible for
both exploration and exploitation in GWO.
3.3. Proposed approach 2: astrophysics strategy

3.3.2. Modified Grey wolf Algorithm-II

The proposed modifications are inspired from astrophysics con-
cept [24]. Here, we have used the concept of path followed by
planets around the Sun. The path, being elliptical, helps provide
better exploration and exploitation. This concept is incorporated
in GWO to speed up the search and increase the efficiency of
search process. The new learning strategy makes use of the ellipti-
cal orbital movement of wolves. Based on this concept, we assume
that grey wolves move around the prey in elliptical shape. The
different steps of proposed astrophysics strategy are depicted in
Fig. 3.
First of all, K grey wolves are randomly chosen from popula-
tion of n wolves. Thereafter, the positions of these K grey wolves
(Yi , i = 1, ...., K) are updated based on the position of best wolf (Yα ).
The new positions of K grey wolves (Yi, new , i = 1, ...., K) and that
of best grey wolf (Yα , new ) are computed as follows.
Yi, new (tcur + 1 ) = Yα (tcur ) + Ri, α × U (−2, 2 )
Yα , new (tcur + 1 ) = Yα (tcur ) × U (−2, 2 )
To improve the performance of GWO, the above-mentioned pro-
posed strategy is incorporated in GWO. This strategy also proves
quite useful, as illustrated in Section 4, in avoiding the solution
being stuck in local optima. The flowchart of proposed approach
namely Modified GWO-II (MGWO-II) is shown in Fig. 1(c). The pro-
posed approach of MGWO-II consists of following the steps.
Algorithm (MGWO-II).
Steps 1, 2 and 3 are same as mentioned in the MGWO-I algo-
rithm.
Step 4. Repeat the following steps until the maximum number
of iterations is reached:
(a) Based on the fitness values, identify the best (Yα ), second
best (Yβ ), and third best (Yδ ) solutions.
(9)
(b) Compute control coefficients i.e., A and C using Eqs.
(3) and (4) respectively.
(10) (c) Modify the positions of grey wolves using Eq. (7).
236 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254

Fig. 3. Proposed astrophysics strategy for GWO.

Fig. 4. Concept of astrophysics strategy (a) 2D position of solar system, (b) 2D position of grey wolves, (c) 3D position of solar system, (d) 3D position of grey wolves.

(d) Update the fitness values according to the modified posi-
tions of the grey wolves.
(e) Apply astrophysics strategy, as mentioned in Fig. (3), on
the positions of grey wolves.
Step 5. The best grey wolf (Yα )will yield the optimal solution at
the final iteration.
3.4. Proposed approach 3: combined strategy
The above-mentioned strategies, i.e., prey weight and astro-
physics, are combined to take the advantages of both strategies.
The combined strategy provides balance between exploration and
exploitation. This combined strategy is known as Modified GWO-III
(MGWO-III).
 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254
3.4.1. Justification of proposed approach 3
The proper balance between exploration and exploitation is
highly desirable for any metaheuristic technique. The exploration
helps avoid the solution being stuck in local optima. The exploita-
tion is responsible for search of better solutions in the identified
region. For MGWO-III, the proposed astrophysics strategy (i.e., el-
liptical path), mentioned in Section 3.2, is utilized to provide the
balance between both exploration and exploitation. Moreover, prey
weight strategy, mentioned in Section 3.3, further improves the ex-
ploitation process.
3.4.2. Modified Grey wolf Algorithm-III
To improve the performance of GWO, the above-mentioned pro-
posed strategy is incorporated in GWO. The flowchart of MGWO-III
is show in Fig. 1(d). The proposed approach of MGWO-III consists
of following the steps.
Algorithm (MGWO-III).
Steps 1, 2 and 3 are same as mentioned in the MGWO-I algo-
rithm.
Step 4. Repeat the following steps until the maximum number
of iterations is reached:
(a) Based on the fitness values, identify the best (Yα ), second
best (Yβ ), and third best (Yδ ) solutions.
(b) Compute control coefficients i.e., A and C using Eqs.
(3) and (8) respectively.
(c) Modify the positions of grey wolves using Eq. (7).
(d) Update the fitness values according to the modified posi-
tions of the grey wolves.
(e) Apply astrophysics strategy on the positions of grey
wolves.
Step 5. The best grey wolf (Yα ) will yield the optimal solution
at the final iteration.
3.5. Computational complexity
The complexity analysis of MGWO-III is presented in this sub-
section. Both time and space complexities of MGWO-III is de-
scribed below.
3.5.1. Time complexity
1. Initialization of MGWO-III requires O(n × d) time where n indi-
cates the number of wolves and d is the dimension of the test
function.
2. Control parameter calculation and position update steps of
MGWO-III requires O(n × d) each.
3. Fitness evaluation requires O(n × d) time.
4. Computation of astrophysics-based strategy requires O(K × d)
time where K indicates the number of selected wolves.
Therefore, summing up the complexities of all the above
steps and considering that K  n, the total time complexity
becomes O(n × d) per generation. The total time complexity of
MGWO-III for maximum number of iterations is O(n × d × MaxItera-
tion). Here MaxIterationindicates the maximum number of genera-
tions/iterations.
3.5.2. Space complexity
The space requirement of MGWO-III is due to its initialization
process. Thus, the total space complexity of MGWO-III is O(n × d).
4. Experimental results and discussion
This section compares the performance of MGWO’s variants
with six other competitive algorithms and validates over thirteen
well-known benchmark test functions. These benchmark test func-
tions are as follows:
237
Table 3
Parameter settings for algorithms used.
Algorithm Parameter Values
GSA Gravitational Constant 100
Alpha Coefficient 20
PSO Inertia Weight 0.72
Cognitive Coefficient 2.0
Social Coefficient 2.0
MVO Wormhole Existence Probability [0.2, 1]
Travelling Distance Rate [0.6, 1]
Exploitation Accuracy 6
MFO Convergence Constant [−1, −2]
Logarithmic Spiral Constant 0.75
SCA Number of Elites 2
GWO Control Parameter (au) [2, 0]
4.1. Benchmark test functions
The effectiveness of the MGWO’s variants is demonstrated on
thirteen benchmark test functions. Table 1 and 2 describe uni-
modal and multimodal benchmark test functions with their prop-
erties [27].
4.2. Experimental setup
In order to investigate the performance of MGWO’s variants,
these are compared with original version of GWO and five recently
developed techniques such as GSA, PSO, MVO, MFO and SCA. The
maximum number of iterations (or generations) and number of
search agents (or population size) for these algorithms is set as
10 0 0 and 30 respectively. The other parameters of GSA [20], PSO
[11], MVO [17], MFO [16], SCA [18], and GWO [15], are set as they
are recommended in their original papers. The parameter settings
of the above-mentioned algorithms are reported in Table 3. The
above-mentioned algorithms are run 30 times for each benchmark
function.
4.3. Experiment 1: performance comparison
To illustrate the performance of MGWO’s variants, its results are
compared with six above-mentioned techniques on benchmark test
functions stated in Section 4.1. The dimension of search space is
set to 30 for this comparison, as mentioned in literature.
Tables 4 and 5 represent the performance comparisons of the
proposed MGWO’s variants with above-mentioned existing tech-
niques for unimodal and multimodal benchmark test functions re-
spectively. The results have been compared in terms of ’mean’ and
’standard deviation’ over 30 independent simulation runs for each
case. The results reveal that the proposed MGWO-III is able to de-
termine the optimal solution for all unimodal test functions in al-
most every run exceptF6 . For F1 − F4 functions, it has been found
that MGWO-II and MGWO-III outperform all the above-mentioned
techniques. For F5 function, MGWO-III provides better results than
the other competitive techniques. GSA provides much better results
than the other competitive techniques for F6 function. MGWO-III is
the third best algorithm after PSO. For F7 function, MGWO-III per-
forms better than GSA, PSO, MVO, MFO, SCA, GWO, MGWO-I and
MGWO-II.
It has been observed from results presented in Table 5 that
MGWO-III is able to find the optimal solution for almost all mul-
timodal test function in every run. For F8 function, MGWO-III
outperforms the other techniques. For F9 , F10 , and F11 functions,
MGWO-III and MGWO-II perform better than the other competi-
tive techniques. For F12 function, MGWO-III provides better results
than the other techniques. GWO and MGWO-I provide comparable
results. For F13 function, GSA is superior to other algorithms such
as PSO, MVO, MFO, SCA, GWO and MGWO’s variants.
238 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254

Table 4
Results on unimodal benchmark test functions.

GSA PSO MVO MFO SCA GWO MGWO-I MGWO-II MGWO-III

F1 1.05E−16 3.92E−10 2.78E−01 1.02E−04 1.33E−03 8.01E−59 1.54E−59 0.0 0E+ 0 0 0.0 0E+ 0 0
(6.10E−17) (1.40E−08) (1.11E−01) (5.99E−04) (1.06E−01) (1.85E−58) (4.77E−58) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F2 5.31E−08 1.03E−04 3.44E−01 4.00E+ 01 3.53E−06 1.23E−34 8.58E−35 0.0 0E+ 0 0 0.0 0E+ 0 0
(9.29E−01) (1.84E−03) (1.41E−01) (2.16E+ 01) (8.57E−05) (2.80E−34) (6.69E−35) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F3 3.96E+ 02 1.17E+ 01 4.26E+ 01 3.07E+ 03 4.07E+ 03 9.22E−15 8.06E−15 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.56E+ 02) (7.13E+ 00) (8.97E+ 00) (3.71E+ 03) (8.21E+ 00) (3.29E−14) (1.92E−14) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F4 8.85E−01 5.79E−01 9.35E−01 6.76E+ 01 1.71E+ 01 1.65E−14 1.51E−14 0.0 0E+ 0 0 0.0 0E+ 0 0
(9.89E−01) (1.72E−01) (2.50E−01) (1.06E+ 01) (8.21E+ 00) (2.02E−14) (2.89E−14) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F5 2.75E+ 01 2.80E+ 01 3.26E+ 01 6.28E+ 02 5.32E+ 01 2.70E+ 01 2.70E+ 01 1.79E+ 01 1.63E+ 01
(3.47E+ 01) (3.89E+ 01) (1.43E+ 02) (3.98E+ 03) (1.98E+ 03) (7.84E−01) (6.63E−01) (1.20E+ 01) (1.25E+ 01)
F6 9.60E−17 1.27E−09 2.98E−02 1.02E−04 4.62E+ 00 5.97E−01 5.65E−01 2.09E−05 2.04E−05
(4.00E−17) (1.78E−08) (9.98E−02) (2.01E−04) (9.75E−01) (3.67E−01) (3.81E−01) (9.16E−06) (6.76E−06)
F7 5.23E−02 6.40E−02 1.92E−02 1.89E−01 2.04E−02 8.38E−04 7.30E−04 3.16E−05 3.10E−05
(2.77E+ 00) (2.87E−02) (7.43E−03) (2.93E−01) (5.79E−02) (5.87E−04) (4.02E−04) (2.72E−05) (2.76E-05)

Table 5
Results on multimodal benchmark test functions.

GSA PSO MVO MFO SCA GWO MGWO-I MGWO-II MGWO-III

F8 −6.58E+ 03 −6.22E+ 03 −6.86E+ 03 −8.18E+ 03 −3.82E+ 03 −5.90E+ 03 −6.04E+ 03 −6.59E+ 03 −6.70E+ 03

(5.72E+ 02) (1.30E+ 03) (9.19E+ 02) (8.80E+ 02) (2.53E+ 02) (9.99E+ 02) (7.47E+ 02) (1.47E+ 03) (1.18E+ 03)
F9 2.74E+ 01 4.78E+ 01 1.01E+ 02 1.59E+ 02 9.73E+ 00 1.89E−14 1.14E−14 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.19E+ 01) (1.03E+ 01) (1.89E+ 01) (3.74E+ 01) (3.25E+ 01) (2.73E−14) (2.75E−14) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F10 8.13E−09 3.14E−05 1.10E+ 00 1.88E+ 01 2.01E+ 01 1.59E−14 1.51E−14 8.88E−16 8.88E−16
(1.90E−09) (2.11E−01) (7.87E−01) (6.18E+ 00) (8.11E+ 00) (2.90E−15) (1.87E−15) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F11 7.96E+ 00 3.36E−09 5.64E−01 1.25E−02 2.68E−01 2.83E−04 1.24E−04 0.0 0E+ 0 0 0.0 0E+ 0 0
(3.70E+ 00) (7.39E−03) (1.12E−01) (1.74E−01) (2.89E−01) (1.55E−03) (3.47E−04) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F12 1.10E−01 6.44E−02 8.04E−01 6.92E−01 1.14E+ 00 4.20E−02 4.03E−02 2.99E−02 2.65E−02
(3.14E−01) (2.06E−01) (1.02E+ 00) (1.83E+ 00) (2.47E+ 02) (2.03E−02) (2.37E−02) (2.63E−02) (5.46E−02)
F13 3.75E−32 1.75E−09 5.66E−02 1.11E−01 3.88E+ 00 4.79E−01 4.01E−01 3.53E−01 2.02E−01
(8.95E−32) (5.39E−03) (4.33E−02) (1.48E+ 00) (8.63E−04) (2.11E−01) (1.56E−01) (6.79E−01) (5.54E−01)

4.4. Experiment 2: statistical analysis
Besides basic statistical analysis (i.e., mean and standard devia-
tion), ANOVA test has been conducted for comparison of GSA, PSO,
MVO, MFO, SCA, GWO and MGWO’s variants. It is used to deter-
mine whether the proposed MGWO’s variants are statistical signifi-
cant or not. A p-value determines the significance level of MGWO’s
variants. An algorithm is statistically significant if and only if the
p-value is less than 0.05. The result analysis of the ANOVA test for
the benchmark test functions is tabulated in Table 6.
The proposed variants are statistical different from the above-
mentioned algorithms for F1 function. The p-value obtained from
MGWO’s variants is less than 0.05. For F2 function, MGWO’s vari-
ants are statistically significant as compared to other algorithms.
For F3 − F6 functions, the MGWO’s variants are statistically signifi-
cant than the existing algorithms. For F7 function, the p-value ob-
tained from MGWO’s variants is smaller than 0.05. MGWO-II and
MGWO-III are statistically different from the existing algorithms.
MGWO-II and MGWO-III are statistically similar for F1 − F7 func-
tions.
For F8 function, MGWO-III is significantly different from the
existing algorithms (i.e., GSA, PSO, MVO, MFO, SCA, GWO, and
MGWO-I). MGWO-II is not statistically different from MGWO-I and
MGWO-III. For F9 and F10 functions, the p-value obtained from
MGWO-II and MGWO-III is 0.00. Hence, MGWO-II and MGWO-
III are statistically significant from other existing algorithms. For
F11 function, MGWO’s variants are statistically different from
the above-mentioned algorithms except MFO. For F12 function,
MGWO’s variants are statistically significant from the all above-
mentioned algorithms except SCA. MGWO’s variants are statisti-
cally similar for F13 function.
The results demonstrate that the p-value, obtained using
MGWO’s variants (i.e., MGWO-I, MGWO-II and MGWO-III) with re-
spect to the existing algorithms, is much smaller than 0.05 for al-
most all benchmark test functions. Therefore, MGWO’s variants are
statistically significant from the existing algorithms for both uni-
modal and multimodal benchmark test functions.
4.5. Experiment 3: scalability study
The next experiment is carried out to perceive the effect of scal-
ability on above-mentioned algorithms for all benchmark test func-
tions. The dimensionality of benchmark test functions is made to
vary as 30, 50, 70, 90 and 100. Tables 7 and 8 depict the perfor-
mance of algorithms on scalable unimodal and multimodal bench-
mark test functions. For F1 − F4 functions, MGWO-II and MGWO-III
provide same results over all dimensions. This may be attributed
to the fact that they have better exploration and exploitation ca-
pability owing to the proposed approaches. For F5 − F7 functions,
the performance of MGWO’s variants decreases with increase in
dimension. The results also show that the degradation in the per-
formance of GSA, PSO, MVO, MFO, SCA, and GWO techniques with
increase in dimensionality of search space for all unimodal bench-
mark test functions.
The results mentioned in Table 8 reveal that all the above-
mentioned algorithms namely GSA, PSO, MVO, MFO, SCA, GWO,
and MGWO’s variants improve their performance with increase
in dimensionality forF8 . However, MGWO-III provides better re-
sults than the other competitive algorithms. The performance of
MGWO-II and MGWO-III is not influenced by scalable search spaces
ofF9 , F10 and F11 . It has also been noticed that the performance
of GSA, PSO, MVO, MFO, SCA, and GWO degrades with increase
in dimension. For these functions, MGWO-I and GWO are the sec-
ond and third best algorithms among all used for comparison. For
F12 and F13 functions, the performance of GSA, PSO, MVO, MFO,
SCA, GWO and MGWO’s variants reduces with increase in dimen-
sionality of search space.
The results illustrated that the performance of MGWO’s vari-
ants is not much affected with increase in dimensionality of search
Table 6
Anova test results on benchmark test functions.
p-value Algorithms for comparison
GSA PSO MVO MFO SCA GWO MGWO-I MGWO-II MGWO-III
F1 0.0 0E+ 0 0 MVO, MFO, SCA, MVO, MFO, SCA, GSA, PSO, MFO, SCA, GSA, PSO, MVO, SCA, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO,
GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, MFO, GWO, MGWO-I, MFO, SCA, MGWO-II, MFO, SCA, MGWO-II, MFO, SCA,GWO, MFO, SCA,GWO,
MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-III MGWO-III MGWO-I MGWO-I
F2 0.0 0E+ 0 0 MFO MVO, MFO, SCA, PSO, MFO, SCA, GSA, PSO, MVO, SCA, PSO, MVO, MFO, PSO, MVO, MFO, PSO, MVO, MFO, PSO, MVO, MFO, PSO, MVO, MFO,

V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254

GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, SCA, MGWO-II, SCA, MGWO-II, SCA, GWO, MGWO-I SCA, GWO, MGWO-I
MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-III MGWO-III
F3 0.0 0E+ 0 0 PSO, MVO, MFO, MVO, MFO, SCA, GSA, PSO, MFO, SCA, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO,
SCA, GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, MFO, SCA MFO, SCA, MGWO-II, MFO, SCA, MGWO-I MFO, SCA, MGWO-I
MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-III
F4 0.0 0E+ 0 0 PSO, MFO, SCA, GSA, MVO, MFO, PSO, MFO, SCA, GSA, PSO, MVO, SCA, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO,
GWO, MGWO-I, SCA, GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, MFO, GWO, MGWO-I, MFO, SCA, MGWO-II, MFO, SCA, MGWO-II, MFO, SCA, GWO, MFO, SCA, GWO,
MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-III MGWO-III MGWO-I MGWO-I
F5 0.0 0E+ 0 0 MVO, MFO, MVO, MFO, GWO, GSA, PSO, MFO, GSA, PSO, MVO, SCA, MFO PSO, MVO, MFO, PSO, MVO, MFO, GSA, PSO, MVO, GSA, PSO, MVO,
MGWO-II, MGWO-III MGWO-I, MGWO-II, GWO, MGWO-I, GWO, MGWO-I, MGWO-I, MGWO-II, GWO, MGWO-II, MFO, GWO, MGWO-I MFO, GWO, MGWO-I
MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-III MGWO-III
F6 0.0 0E+ 0 0 PSO, MVO, MFO, GSA, MVO, MFO, GSA, PSO, MFO, SCA, GSA, PSO, MVO, SCA, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO,
SCA, GWO, MGWO-I, SCA, GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, MFO, GWO, MGWO-I, MFO, SCA, MGWO-II, MFO, SCA, MGWO-II, MFO, SCA, GWO, MFO, SCA, GWO,
MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-III MGWO-III MGWO-I MGWO-I
F7 1.97E-03 MGWO-II, MGWO-III MVO, MFO, SCA, PSO, GWO, MGWO-I, PSO, MVO, SCA, PSO, MFO, GWO, PSO, MVO, MFO, PSO, MFO, MVO, GSA, PSO, MVO, GSA, PSO, MVO,
GWO, MGWO-I, MGWO-II, MGWO-III GWO, MGWO-I, MGWO-I, MGWO-II, SCA, MGWO-II, SCA, MGWO-II, MFO, SCA, GWO, MFO, SCA, GWO,
MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-III MGWO-III MGWO-III MGWO-I MGWO-I
F8 0.0 0E+ 0 0 PSO, MVO, MFO, GSA, MVO, MFO, GSA, PSO, MFO, SCA, GSA, PSO, MVO, SCA, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO,
SCA, GWO, MGWO-I, SCA, GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, MFO, GWO, MGWO-I, MFO, SCA, MGWO-III MFO, SCA, MGWO-III MFO, SCA MFO, SCA, GWO,
MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-I
F9 0.0 0E+ 0 0 PSO, MVO, MFO, GSA, MVO, MFO, GSA, PSO, MFO, SCA, GSA, PSO, MVO, SCA, PSO, MVO, MFO, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO,
GWO, MGWO-I, SCA, GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, MFO, SCA MFO, SCA, MGWO-II, MFO, SCA, MGWO-I MFO, SCA, MGWO-I
MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-III
F10 0.0 0E+ 0 0 MVO, MFO, GWO, MVO, MFO, SCA GSA, PSO, MFO, SCA, GSA, PSO, MVO, GSA, PSO, MVO, GSA, MVO, MFO, GSA, MVO, MFO, GSA, MVO, MFO, GSA, MVO, MFO,
MGWO-I, MGWO-II, GWO, MGWO-I, GWO, MGWO-I, MFO, GWO, MGWO-I, SCA, MGWO-II, SCA, MGWO-II, SCA, GWO, MGWO-I SCA, GWO, MGWO-I
MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-III MGWO-III
F11 0.0 0E+ 0 0 PSO, MVO,MFO, SCA, GSA, MVO, SCA, GSA, PSO, MFO, SCA, GSA, MVO, SCA GSA, PSO, MVO, GSA, PSO, MVO, SCA, GSA, PSO, MVO, SCA, GSA, PSO, MVO, SCA, GSA, PSO, MVO, SCA,
GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, MFO, GWO, MGWO-I, MGWO-I, MGWO-II, GWO GWO GWO
MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-III
F12 2.33E-02 PSO, MVO,MFO, GSA, MVO, MFO, GSA, PSO, SCA, GWO, GSA, PSO, GWO, MVO GSA, PSO, MVO, MFO GSA, PSO, MVO, MFO GSA, PSO, MVO, MFO GSA, PSO, MVO, MFO
GWO, MGWO-I, GWO, MGWO-I, MGWO-I, MGWO-II, MGWO-I, MGWO-II,
MGWO-II, MGWO-III MGWO-II, MGWO-III MGWO-III MGWO-III
F13 1.19E-03 PSO, MVO,MFO, GSA, MVO, MFO, GSA, PSO, MFO, GSA, PSO, MVO NA GSA, PSO, MVO, GSA, PSO, MVO, GSA, PSO, MVO GWO, MGWO-I
GWO, MGWO-I, GWO, MGWO-I, GWO, MGWO-I, MGWO-III MGWO-III
MGWO-II MGWO-II MGWO-II

239
240 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254

Table 7
Performance of algorithms on high dimensional unimodal benchmark functions.

Dim GSA PSO MVO MFO SCA GWO MGWO-I MGWO-II MGWO-III

F1 30 1.05E-16 3.92E-10 9.21E−01 1.02E−04 1.33E−03 8.01E−59 1.54E−59 0.0 0E+ 0 0 0.0 0E+ 0 0
(6.10E-17) (1.40E−08) (1.27E−01) (5.99E−04) (1.06E−01) (1.85E−58) (4.77E−58) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
50 1.06E−04 9.12E−04 6.81E+ 00 9.33E+ 03 8.04E+ 01 1.05E−43 1.18E−58 0.0 0E+ 0 0 0.0 0E+ 0 0
(5.86E−05) (9.39E−04) (1.26E+ 00) (1.11E+ 04) (1.36E+ 02) (1.46E−43) (2.26E−58) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
70 3.32E−04 1.17E−01 2.79E+ 01 1.35E+ 04 1.19E+ 03 6.55E−36 6.45E−49 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.34E−04) (8.98E−02) (5.22E+ 00) (1.07E+ 04) (1.14E+ 03) (1.31E−35) (1.46E−48) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
90 1.00E+ 01 1.21E+ 00 6.83E+ 01 2.16E+ 04 4.05 + 03 3.88E−31 7.18E−43 0.0 0E+ 0 0 0.0 0E+ 0 0
(4.52E+ 01) (7.61E−01) (1.24E+ 01) (1.04E+ 04) (3.56E+ 03) (4.77E−31) (1.03E−42) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
100 5.76E+ 02 2.97E+ 00 1.19E+ 02 3.51E+ 04 6.10E+ 03 2.09E−29 1.06E−40 0.0 0E+ 0 0 0.0 0E+ 0 0
(3.05E+ 02) (1.69E+ 00) (2.28E+ 01) (1.22E+ 04) (4.20E+ 03) (2.79E−29) (1.64E−40) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F2 30 5.31E−08 1.03E−04 9.95E−01 4.00E+ 01 3.53E−06 1.23E−34 8.58E−35 0.0 0E+ 0 0 0.0 0E+ 0 0
(9.29E−01) (1.84E−03) (1.27E+ 00) (2.16E+ 01) (8.57E−05) (2.80E−34) (6.69E−35) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
50 6.23E−02 1.50E−01 4.26E+ 02 7.22E+ 01 1.27E−02 5.30E−26 2.18E−34 0.0 0E+ 0 0 0.0 0E+ 0 0
(8.66E−03) (1.17E−01) (1.39E+ 03) (3.29E+ 01) (1.63E−02) (4.99E−26) (2.46E−34) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
70 1.38E−01 1.66E+ 00 4.25E+ 09 1.16E+ 02 3.21E−01 7.53E−22 4.50E−29 0.0 0E+ 0 0 0.0 0E+ 0 0
(3.11E−02) (8.52E−01) (1.70E+ 10) (3.88E+ 01) (5.19E−01) (4.48E−22) (5.45E−29) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
90 2.45E−01 6.29E+ 00 6.02E+ 17 1.52E+ 02 9.55E−01 5.81E−19 8.71E−26 0.0 0E+ 0 0 0.0 0E+ 0 0
(4.67E−02) (2.41E+ 00) (3.28E+ 18) (5.34E+ 01) (9.71E−01) (3.68E−19) (7.29E−26) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
100 9.90E+ 00 1.18E+ 01 1.47E+ 23 1.67E+ 02 1.46E+ 00 5.29E−18 1.50E−24 0.0 0E+ 0 0 0.0 0E+ 0 0
(4.76E+ 00) (3.61E+ 00) (5.62E+ 23) (4.77E+ 01) (1.84E+ 00) (3.15E−18) (1.16E−24) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F3 30 3.96E+ 02 1.17E+ 01 1.38E+ 02 3.07E+ 03 4.07E+ 03 9.22E−15 8.06E−15 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.56E+ 02) (7.13E+ 00) (4.85E+ 01) (3.71E+ 03) (8.21E+ 00) (3.29E−14) (1.92E−14) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
50 1.19E+ 02 6.26E+ 02 4.67E+ 03 4.80E+ 04 3.27E+ 04 3.73E−06 2.88E−09 0.0 0E+ 0 0 0.0 0E+ 0 0
(5.42E+ 01) (1.69E+ 02) (1.29E+ 03) (1.59E+ 04) (1.19E+ 04) (1.32E−05) (1.52E−08) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
70 7.16E+ 02 2.89E+ 03 1.92E+ 04 9.97E+ 04 8.79E+ 04 9.73E−03 2.63E−04 0.0 0E+ 0 0 0.0 0E+ 0 0
(2.35E+ 02) (7.80E+ 02) (3.33E+ 03) (3.99E+ 04) (2.27E+ 04) (2.26E−02) (1.13E−03) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
90 1.66E+ 03 7.76E+ 03 4.64E+ 04 1.61E+ 05 1.63E+ 05 2.58E+ 00 3.61E−03 0.0 0E+ 0 0 0.0 0E+ 0 0
(4.36E+ 02) (1.54E+ 03) (5.90E+ 03) (5.00E+ 04) (2.95E+ 04) (6.37E+ 00) (1.15E−02) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
100 8.77E+ 03 1.12E+ 04 6.14E+ 04 1.76E+ 05 2.04E+ 05 6.92E+ 00 3.97E−01 0.0 0E+ 0 0 0.0 0E+ 0 0
(2.00E+ 03) (3.16E+ 03) (8.42E+ 03) (4.92E+ 04) (4.79E+ 04) (1.21E+ 01) (1.73E+ 00) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F4 30 8.85E−01 5.79E−01 1.59E+ 00 6.76E+ 01 1.71E+ 01 1.65E−14 1.51E−14 0.0 0E+ 0 0 0.0 0E+ 0 0
(9.89E−01) (1.72E−01) (4.56E−01) (1.06E+ 01) (8.21E+ 00) (2.02E−14) (2.89E−14) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
50 7.21E+ 00 2.26E+ 00 1.52E+ 01 8.38E+ 01 5.73E+ 01 2.21E−09 3.11E−15 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.51E+ 00) (3.31E−01) (5.77E+ 00) (4.11E+ 00) (8.0 0E+ 0 0) (3.57E−09) (1.08E−14) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
70 1.19E+ 01 4.52E+ 00 3.70E+ 01 9.02E+ 01 7.52E+ 01 1.91E−06 1.35E−11 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.77E+ 00) (7.45E−01) (7.67E+ 00) (2.81E+ 00) (6.19E+ 00) (2.68E−06) (2.74E−11) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
90 1.42E+ 01 8.29E+ 00 5.05E+ 01 9.23E+ 01 8.43E+ 01 2.82E−03 9.05E−05 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.25E+ 00) (1.41E+ 00) (6.13E+ 00) (2.36E+ 00) (3.76E+ 00) (1.27E−02) (5.18E−04) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
100 1.55E+ 01 9.06E+ 00 5.71E+ 01 9.44E+ 01 8.60E+ 01 2.15E−03 6.34E−06 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.55E+ 00) (1.61E+ 00) (6.86E+ 00) (1.80E+ 00) (2.71E+ 00) (3.32E−03) (2.71E−05) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F5 30 2.75E+ 01 2.80E+ 01 2.91E+ 02 6.28E+ 02 5.32E+ 01 2.70E+ 01 2.70E+ 01 1.79E+ 01 1.63E+ 01
(3.47E+ 01) (3.89E+ 01) (6.02E+ 02) (3.98E+ 03) (1.98E+ 03) (7.84E−01) (6.63E−01) (1.20E+ 01) (1.25E+ 01)
50 4.73E+ 01 1.62E+ 02 5.58E+ 02 5.36E+ 06 1.31E+ 06 4.73E+ 01 4.76E+ 01 3.00E+ 01 3.50E+ 01
(1.01E+ 01) (7.20E+ 01) (4.75E+ 02) (2.03E+ 07) (1.70E+ 06) (8.84E−01) (7.13E−01) (2.16E+ 01) (1.92E+ 01)
70 8.47E+ 01 5.87E+ 02 2.00E+ 03 2.25E+ 07 8.69E+ 06 6.74E+ 01 6.79E+ 01 5.17E+ 01 5.14E+ 01
(3.35E+ 01) (5.58E+ 02) (1.51E+ 03) (4.13E+ 07) (7.57E+ 06) (8.93E−01) (6.45E−01) (2.65E+ 01) (2.57E+ 01)
90 1.39E+ 02 1.46E+ 03 5.58E+ 03 2.42E+ 07 3.58E+ 07 8.76E+ 01 8.81E+ 01 6.13E+ 01 6.73E+ 01
(4.33E+ 01) (9.93E+ 02) (5.40E+ 03) (5.25E+ 07) (2.58E+ 07) (7.88E−01) (5.41E−01) (3.66E+ 01) (3.13E+ 01)
100 1.12E+ 04 1.93E+ 03 6.00E+ 03 2.74E+ 07 6.65E+ 07 9.75E+ 01 9.80E+ 01 8.21E+ 01 6.61E+ 01
(7.41E+ 03) (8.16E+ 02) (4.40E+ 03) (5.46E+ 07) (3.48E+ 07) (7.11E−01) (5.19E+ 01) (3.07E+ 01) (4.18E+ 01)
F6 30 9.60E−17 1.27E−09 9.30E−01 1.02E−04 4.62E+ 00 5.97E−01 5.65E−01 2.09E−05 2.04E−05
(4.00E−17) (1.78E−08) (2.31E−01) (2.01E−04) (9.75E−01) (3.67E−01) (3.81E−01) (9.16E−06) (6.76E−06)
50 1.15E−04 9.40E−04 6.57E+ 00 9.38E+ 03 7.73E+ 01 2.24E+ 00 4.50E+ 00 2.15E+ 00 2.65E+ 00
(4.75E−05) (1.30E−03) (1.43E+ 00) (9.78E+ 03) (9.80E+ 01) (6.85E−01) (6.69E−01) (2.11E+ 00) (2.91E+ 00)
70 3.66E−04 1.11E−01 2.78E+ 01 1.41E+ 04 1.45E+ 03 5.04E+ 00 7.81E+ 00 3.39E+ 00 3.27E+ 00
(1.70E−04) (7.50E−02) (4.97E+ 00) (9.83E+ 03) (1.95E+ 03) (8.55E−01) (6.69E−01) (3.39E+ 00) (4.21E+ 00)
90 5.09E−01 1.09E+ 00 7.08E+ 01 2.47E+ 04 3.40E+ 03 7.68E+ 00 1.14E+ 01 3.54E+ 00 7.22E+ 00
(1.68E+ 00) (6.13E−01) (1.14E+ 01) (1.54E+ 04) (2.21E+ 03) (8.37E−01) (6.76E−01) (4.88E+ 00) (6.74E+ 00)
100 6.03E+ 02 2.73E+ 00 1.12E+ 02 3.45E+ 04 3.99E+ 03 9.13E+ 00 1.34E+ 01 7.42E+ 00 6.91E+ 00
(2.83E+ 02) (1.62E+ 00) (1.75E+ 01) (1.34E+ 04) (3.27E+ 03) (9.94E−01) (1.03E+ 00) (6.28E+ 00) (7.81E+ 00)
F7 30 5.23E−02 6.40E−02 2.88E−02 1.89E−01 2.04E−02 8.38E−04 7.30E−04 3.16E−05 3.10E−05
(2.77E+ 00) (2.87E−02) (1.26E−02) (2.93E−01) (5.79E−02) (5.87E−04) (4.02E−04) (2.72E−05) (2.76E−05)
50 1.85E−01 4.16E−01 1.08E−01 2.24E+ 01 7.18E−01 1.45E−03 4.76E−04 4.13E−05 3.88E−05
(4.78E−02) (1.39E−01) (4.24E−02) (2.77E+ 01) (8.97E−01) (5.46E−04) (2.64E−04) (4.20E−05) (3.12E−05)
70 3.42E−01 6.65E+ 01 2.10E−01 5.11E+ 01 6.24E+ 00 2.04E−03 7.33E−04 3.37E−05 3.35E−05
(7.90E−02) (1.23E+ 02) (5.81E−02) (4.49E+ 01) (6.34E+ 00) (9.12E−04) (3.24E−04) (2.85E−05) (2.58E−05)
90 6.33E−01 1.08E+ 03 4.28E−01 1.15E+ 02 3.76E+ 01 2.23E−03 8.98E−04 3.72E−05 3.54E−05
(2.04E−01) (2.51E+ 02) (1.17E−01) (1.08E+ 02) (3.50E+ 01) (1.18E−03) (6.06E−04) (3.74E−05) (2.78E−05)
100 1.47E+ 01 1.37E+ 03 5.54E−01 1.30E+ 02 5.57E+ 01 2.38E−03 9.83E−04 3.85E−05 4.06E−05
(2.04E+ 01) (3.60E+ 02) (1.11E−01) (9.78E+ 01) (5.24E+ 01) (1.07E−03) (4.15E−04) (3.75E−05) (3.44E−05)
 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254 241

Table 8
Performance of algorithms on high dimensional multimodal benchmark functions.

Dim GSA PSO MVO MFO SCA GWO MGWO-I MGWO-II MGWO-III

F8 30 −6.58E+ 03 −6.22E+ 03 −7.69E+ 03 −8.18E+ 03 −3.82E+ 03 −5.90E+ 03 −6.04E+ 03 −6.59E+ 03 −6.70E+ 03

(5.72E+ 02) (1.30E+ 03) (5.74E+ 02) (8.80E+ 02) (2.53E+ 02) (9.99E+ 02) (7.47E+ 02) (1.47E+ 03) (1.18E+ 03)
50 −3.15E+ 03 −9.82E+ 03 −1.24E+ 04 −1.34E+ 04 −5.24E+ 03 −8.81E+ 03 −7.31E+ 03 −9.39E+ 03 −8.76E+ 03
(6.43E+ 02) (2.61E+ 03) (8.53E+ 02) (1.25E+ 03) (4.91E+ 02) (1.35E+ 03) (1.92E+ 03) (2.14E+ 03) (2.02E+ 03)
70 −4.14E+ 03 −1.37E+ 04 −1.67E+ 04 −1.77E+ 04 −6.01E+ 03 −1.19E+ 04 −9.21E+ 03 −1.45E+ 04 −1.17E+ 04
(7.55E+ 02) (3.49E+ 03) (1.41E+ 03) (2.21E+ 03) (4.64E+ 02) (1.80E+ 03) (2.54E+ 03) (2.78E+ 03) (2.97E+ 03)
90 −4.42E+ 03 −1.74E+ 04 −2.12E+ 04 −2.15E+ 04 −7.00E+ 03 −1.51E+ 04 −1.08E+ 04 −1.71E+ 04 −1.54E+ 04
(6.08E+ 02) (5.01E+ 03) (1.27E+ 03) (2.09E+ 03) (4.64E+ 02) (2.06E+ 03) (3.63E+ 03) (3.72E+ 03) (3.38E+ 03)
100 −4.70E+ 03 −2.11E+ 04 −2.36E+ 04 −2.36E+ 04 −7.23E+ 03 −1.62E+ 04 −1.36E+ 04 −2.00E+ 04 −1.71E+ 04
(9.41E+ 02) (4.21E+ 03) (1.11E+ 03) (2.38E+ 03) (5.57E+ 02) (2.32E+ 03) (2.54E+ 03) (3.99E+ 03) (3.94E+ 03)
F9 30 2.74E+ 01 4.78E+ 01 1.20E+ 02 1.59E+ 02 9.73E+ 00 1.89E−14 1.14E−14 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.19E+ 01) (1.03E+ 01) (3.29E+ 01) (3.74E+ 01) (3.25E+ 01) (2.73E−14) (2.75E−14) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
50 4.51E+ 01 1.23E+ 02 2.56E+ 02 2.99E+ 02 6.14E+ 01 1.54E+ 00 0.0 0E+ 0 0 0.0 0E+ 0 0 0.0 0E+ 0 0
(8.80E+ 00) (1.92E+ 01) (4.24E+ 01) (4.90E+ 01) (4.64E+ 01) (4.10E+ 00) (0.0 0E+ 0 0) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
70 6.97E+ 01 2.37E+ 02 4.04E+ 02 4.79E+ 02 1.32E+ 02 2.92E−01 7.58E−15 0.0 0E+ 0 0 0.0 0E+ 0 0
(9.64E+ 00) (4.12E+ 01) (7.31E+ 01) (6.73E+ 01) (6.86E+ 01) (1.13E+ 00) (3.66E−15) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
90 8.29E+ 01 3.81E+ 02 5.66E+ 02 6.48E+ 02 1.85E+ 02 7.23E−01 0.0 0E+ 0 0 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.10E+ 01) (6.30E+ 01) (7.32E+ 01) (7.47E+ 01) (8.50E+ 01) (1.75E+ 00) (0.0 0E+ 0 0) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
100 2.14E+ 02 4.46E+ 02 6.68E+ 02 7.39E+ 02 1.97E+ 02 1.02E+ 00 3.79E−15 0.0 0E+ 0 0 0.0 0E+ 0 0
(2.86E+ 01) (7.72E+ 01) (8.65E+ 01) (7.91E+ 01) (1.46E+ 02) (2.25E+ 00) (2.08E−14) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F10 30 8.13E−09 3.14E−05 1.54E+ 00 1.88E+ 01 2.01E+ 01 1.59E−14 1.51E−14 8.88E−16 8.88E−16
(1.90E−09) (2.11E−01) (4.87E−01) (6.18E+ 00) (8.11E+ 00) (2.90E−15) (1.87E−15) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
50 6.33E−03 6.75E−01 2.93E+ 00 1.93E+ 01 1.74E+ 01 3.30E−14 1.39E−14 8.88E−16 8.88E−16
(1.29E−03) (6.59E−01) (5.79E−01) (1.08E+ 00) (6.91E+ 00) (3.90E−15) (1.94E−15) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
70 4.12E−02 1.60E+ 00 3.53E+ 00 1.95E+ 01 1.82E+ 01 5.43E−14 1.79E−14 8.88E−16 8.88E−16
(1.75E−01) (5.01E−01) (4.72E−01) (5.73E−01) (5.29E+ 00) (6.76E−15) (3.66E−15) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
90 2.36E+ 00 2.41E+ 00 6.45E+ 00 1.97E+ 01 1.86E+ 01 8.80E−14 2.91E−14 8.88E−16 8.88E−16
(8.95E−01) (2.99E−01) (5.45E+ 00) (3.26E−01) (5.12E+ 00) (8.38E−15) (2.27E−15) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
100 2.91E+ 00 2.61E+ 00 9.33E+ 00 1.99E+ 01 2.06E+ 01 1.09E−13 2.91E−14 8.88E−16 8.88E−16
(4.73E−01) (2.85E−01) (7.08E+ 00) (1.78E−01) (3.86E+ 00) (7.73E−15) (2.27E−15) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F11 30 7.96E+ 00 3.36E−09 7.87E−01 1.25E−02 2.68E−01 2.83E−04 1.24E−04 0.0 0E+ 0 0 0.0 0E+ 0 0
(3.70E+ 00) (7.39E−03) (8.16E−02) (1.74E−01) (2.89E−01) (1.55E−03) (3.47E−04) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
50 2.39E+ 01 5.62E−03 1.06E+ 00 1.06E+ 02 3.01E+ 00 2.04E−03 0.0 0E+ 0 0 0.0 0E+ 0 0 0.0 0E+ 0 0
(5.33E+ 00) (8.32E−03) (1.75E−02) (9.76E+ 01) (5.81E+ 00) (4.83E−03) (0.0 0E+ 0 0) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
70 5.53E+ 01 5.80E−03 1.25E+ 00 1.47E+ 02 8.34E+ 00 3.65E−04 0.0 0E+ 0 0 0.0 0E+ 0 0 0.0 0E+ 0 0
(8.37E+ 00) (7.61E−03) (5.77E−02) (8.36E+ 01) (9.07E+ 00) (2.00E−03) (0.0 0E+ 0 0) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
90 1.02E+ 02 2.83E−02 1.64E+ 00 2.40E+ 02 3.76E+ 01 1.67E−03 0.0 0E+ 0 0 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.44E+ 01) (1.86E−02) (1.02E−01) (1.15E+ 02) (3.86E+ 01) (5.12E−03) (0.0 0E+ 0 0) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
100 9.29E+ 01 3.84E−02 2.01E+ 00 3.22E+ 02 3.45E+ 01 5.46E−04 0.0 0E+ 0 0 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.42E+ 01) (2.81E−02) (1.12E−01) (1.18E+ 02) (4.04E+ 01) (2.99E−03) (0.0 0E+ 0 0) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F12 30 1.10E−01 6.44E−02 1.70E+ 00 6.92E−01 1.14E+ 00 4.20E−02 4.03E−02 2.99E−02 2.65E−02
(3.14E−01) (2.06E−01) (1.40E+ 00) (1.83E+ 00) (2.47E+ 02) (2.03E−02) (2.37E−02) (2.63E−02) (5.46E−02)
50 6.10E−01 1.25E−02 5.51E+ 00 2.56E+ 07 1.54E+ 06 9.66E−02 2.24E−01 1.76E−01 1.49E−01
(9.53E−01) (3.43E−02) (2.45E+ 00) (7.81E+ 07) (3.67E+ 06) (6.14E−02) (7.48E−02) (3.59E−01) (2.80E−01)
70 1.0 0E+ 0 0 1.65E−01 9.54E+ 00 2.73E+ 07 3.69E+ 07 1.51E−01 3.12E−01 1.87E−01 1.52E−01
(9.97E−01) (2.56E−01) (2.88E+ 00) (7.76E+ 07) (5.97E+ 07) (3.20E−02) (7.72E−02) (3.49E−01) (2.01E−01)
90 1.89E+ 00 1.45E+ 00 1.59E+ 01 3.38E+ 07 1.23E+ 08 2.24E−01 3.72E−01 1.98E−01 1.77E−01
(8.44E−01) (1.17E+ 00) (5.17E+ 00) (7.03E+ 07) (6.79E+ 07) (6.40E−02) (7.40E−02) (2.62E−01) (3.06E−01)
100 4.86 + 00 1.61 E+ 00 1.77E+ 01 4.35E+ 07 1.64E+ 08 2.38E−01 4.06E−01 4.27E−01 2.34E−01
(1.44E+ 00) (9.47E−01) (5.17E+ 00) (1.77E+ 08) (9.09E+ 07) (3.91E−02) (4.51E−02) (4.82E−01) (3.11E−01)
F13 30 3.75E−32 1.75E−09 5.66E−02 1.11E−01 3.88E+ 00 4.79E−01 4.01E−01 3.53E−01 2.02E−01
(8.95E−32) (5.39E−03) (4.33E−02) (1.48E+ 00) (8.63E−04) (2.11E−01) (1.56E−01) (6.79E−01) (5.54E−01)
50 2.46E−01 9.76E−03 4.45E+ 00 1.37E+ 07 1.98E+ 06 1.82E+ 00 2.63E+ 00 2.24E+ 00 1.89E+ 00
(8.54E−01) (1.38E−02) (1.09E+ 01) (7.49E+ 07) (4.10E+ 06) (3.79E−01) (3.10E−01) (2.07E+ 00) (2.31E+ 00)
70 6.25E−01 4.25E−01 5.36E+ 01 1.11E+ 08 5.37E+ 07 3.59E+ 00 4.47E+ 00 3.26E+ 00 3.25E+ 00
(1.69E+ 00) (3.21E−01) (2.37E+ 01) (1.93E+ 08) (6.91E+ 07) (4.63E−01) (2.90E−01) (3.05E+ 00) (3.22E+ 00)
90 6.57E−01 4.89E+ 00 1.21E+ 02 1.67E+ 08 2.04E+ 08 5.57E+ 00 6.48E+ 00 4.66E+ 00 4.30E+ 00
(1.20E+ 00) (3.05E+ 00) (2.63E+ 01) (1.77E+ 08) (1.48E+ 08) (3.10E−01) (2.96E−01) (4.06E+ 00) (4.22E+ 00)
100 7.34E+ 00 9.50E+ 00 1.62E+ 02 4.19E+ 08 2.32E+ 08 6.25E+ 00 7.47E+ 00 5.30E+ 00 4.95E+ 00
(9.68E+ 00) (4.48E+ 00) (2.66E+ 01) (2.30E+ 08) (1.45E+ 08) (4.07E−01) (3.24E−01) (4.71E+ 00) (4.90E+ 00)

space for unimodal as well as multimodal benchmark functions. fined as [14].

The reason behind this is that these variants have better search
capability and superior balance between exploration and exploita-
tion.
4.6. Experiment 4: noise study
The next experiment is carried out to see the effect of noisy
benchmark functions on the performance of the above-mentioned
algorithms. The noisy benchmark function is mathematically de-
FNoisy (Y ) = F (Y ) + Nd (0, 1 ) (12)
where Nd (0,1) represents normal distribution with zero mean and
standard deviation of one. Tables 9 and 10 show the results ob-
tained from noisy unimodal and multimodal benchmark test func-
tions respectively.
For F1 function, MGWO’s variants provide better fitness value
than the other competitive algorithms. MGWO-III provides better
results than GWO, MGWO-I and MGWO-II for all unimodal bench-
mark functions. The noise significantly degrades the performance
242 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254

Table 9
Results on noisy and high−dimensional unimodal benchmark test functions.

GSA PSO MVO MFO SCA GWO MGWO−I MGWO−II MGWO−III

F1 5.65E+ 02 2.68E+ 00 4.57E+ 01 3.37E+ 04 3.56E+ 03 −3.72E+ 00 −3.84E+ 00 −4.15E+ 00 −4.20E+ 00

(2.79E+ 02) (1.35E+ 00) (7.13E+ 00) (1.49E+ 04) (5.31E+ 03) (2.95E−01) (2.72E−01) (2.92E−01) (1.81E−01)
F2 4.12E+ 00 8.92E+ 00 5.76E+ 07 2.08E+ 02 4.60E+ 00 −2.99E+ 00 −3.45E+ 00 −4.11E+ 00 −4.17E+ 00
(6.17E+ 00) (2.77E+ 00) (8.47E+ 15) (4.46E+ 01) (3.27E+ 00) (5.03E−01) (4.09E−01) (2.68E−01) (3.92E−01)
F3 7.72E+ 03 1.05E+ 04 4.18E+ 04 1.71E+ 05 1.92E+ 05 −3.22E+ 00 −3.49E+ 00 −4.06E+ 00 −4.10E+ 00
(2.04E+ 03) (3.08E+ 03) (4.71E+ 03) (5.10E+ 04) (4.52E+ 04) (4.72E−01) (3.41E−01) (2.31E−01) (2.60E−01)
F4 1.28E+ 01 8.92E+ 00 6.74E+ 01 9.09E+ 01 8.80E+ 01 −2.43E+ 00 −3.18 + 00 −4.15E+ 00 −4.20E+ 00
(1.81E+ 00) (8.85E−01) (1.46E+ 01) (2.12E+ 00) (3.10E+ 00) (6.56E−01) (4.17E−01) (2.21E−01) (3.62E−01)
F5 1.13E+ 04 1.58E+ 03 1.49E+ 03 7.43E+ 07 6.47E+ 07 2.50E+ 01 2.51E+ 01 2.41E+ 01 2.40E+ 01
(2.84E+ 04) (1.03E+ 03) (8.87E+ 02) (5.19E+ 07) (4.56E+ 07) (4.29E−01) (3.30E−01) (4.25E+ 00) (4.31E+ 00)
F6 5.13E+ 02 1.81E+ 00 4.91E+ 01 3.20E+ 04 4.30E+ 03 −5.31E−01 5.52E−01 1.67E+ 00 1.56E+ 00
(2.28E+ 02) (1.46E+ 00) (7.50E+ 00) (1.70E+ 04) (4.59E+ 03) (5.64E−01) (6.23E−01) (2.42E+ 00) (2.70E+ 00)
F7 1.89E+ 01 1.46E+ 03 2.02E+ 00 1.91E+ 02 4.94E+ 01 −3.30E+ 00 −3.35E+ 00 −3.74E+ 00 −3.77E+ 00
(2.47E+ 01) (2.16E+ 02) (1.27E+ 00) (1.01E+ 02) (3.40E+ 01) (4.07E−01) (3.02E−01) (2.81E−01) (3.87E−01)

Table 10
Results on noisy and high−dimensional multimodal benchmark test functions.

GSA PSO MVO MFO SCA GWO MGWO−I MGWO−II MGWO−III

F8 −4.91E+ 03 −2.05E+ 04 −2.24E+ 04 −2.23E+ 04 −6.96E+ 03 −1.63E+ 04 −1.18E+ 04 −2.36E+ 04 −1.45E+ 04

(8.03E+ 02) (4.34E+ 03) (1.47E+ 03) (3.11E+ 03) (5.99E+ 02) (3.00E+ 03) (4.24E+ 03) (1.45E+ 03) (5.41E+ 03)
F9 2.01E+ 02 4.54E+ 02 6.45E+ 02 7.84E+ 02 2.26E+ 02 −2.50E+ 00 −3.67E+ 00 −4.15E+ 00 −4.19E+ 00
(2.64E+ 01) (5.16E+ 01) (7.95E+ 01) (8.41E+ 01) (9.25E+ 01) (1.45E+ 00) (4.47E−01) (2.71E−01) (2.75E−01)
F10 1.97E+ 01 2.69E+ 00 1.71E+ 01 1.71E+ 01 1.69E+ 01 1.22E+ 01 −1.09E+ 00 −4.06E+ 00 −4.22E+ 00
(4.56E−01) (2.83E−01) (4.32E−02) (3.08E−01) (2.90E−01) (6.29E+ 00) (3.52E+ 00) (2.67E−01) (3.30E−01)
F11 8.69E+ 01 4.00E+ 02 1.12E+ 01 3.15E+ 02 5.11E+ 01 −2.08E+ 00 −2.51E+ 00 −4.11E+ 00 −4.23E+ 00
(1.21E+ 01) (1.65E+ 02) (2.89E+ 00) (1.70E+ 02) (3.69E+ 01) (5.48E−01) (3.47E−01) (3.22E−01) (3.35E−01)
F12 1.49E+ 00 1.37E+ 00 2.00E+ 01 1.84E+ 07 1.24E+ 08 −2.22E+ 00 −2.48E+ 00 −2.88E+ 00 −3.30E+ 00
(1.21E+ 00) (9.23E−01) (4.97E+ 00) (1.52E+ 08) (8.64E+ 07) (3.41E−01) (3.33E−01) (2.97E−01) (3.61E−01)
F13 9.65E+ 00 1.06E+ 01 1.31E+ 02 1.47E+ 08 3.03E+ 08 6.50E+ 00 6.36E+ 00 3.14E+ 00 3.06E+ 00
(7.62E+ 00) (6.41E+ 00) (2.84E+ 01) (2.38E+ 08) (1.72E+ 08) (6.69E−01) (6.12E−01) (3.40E+ 00) (3.70E+ 00)

of all the above-mentioned algorithms. The results obtained from
MGWO-II are far better than other four algorithms for F8 function.
For F9 − F13 functions, MGWO-III shows best results followed by
MGWO-II, MGWO-I and GWO. The performance of GSA, PSO, MVO,
MFO, and SCA is affected to a greater extent with noisy benchmark
test functions.
It has also been noticed that performance of MGWO-II and
MGWO-III is least affected as compared to other techniques for
noisy benchmark functions. This is attributed to the fact that the
proposed approach has better exploration capability due to in-
corporation of the astrophysics concept. Due to this feature, it is
able to determine the optimal value of fitness notwithstanding the
noisy benchmark functions.
4.7. Experiment 5: performance comparison with existing variants of
GWO
In order to prove the supremacy of MGWO’s variants, these are
compared with the five well-known existing algorithms. These are
HGWO [29], PGWO [28], IGWO [26], mGWO [19], and EGWO [5].
Tables 11 and 12 show the performance comparison of MGWO’s
variants with above-mentioned algorithms.
For F1 − F4 functions, MGWO-II and MGWO-III provide better
solution than the other competitive algorithms. EGWO performs
better than MGWO-I and other existing variants of GWO. The re-
sults obtained from MGWO-II and MGWO-III are far better than
other variants of GWO. For F8 function, EGWO provides best fit-
ness value as compared to other techniques. For F9 − F11 functions,
MGWO-II and MGWO-III perform much better than other tech-
niques. EGWO is the third best algorithm that provides better so-
lution than the other competitive algorithms. MGWO-III is the best
algorithm for solving F12 and F13 functions. MGWO-II is the second
best algorithm for these functions.
4.8. Experiment 6: sensitivity analysis
The effect of au on the performance of MGWO’s variants has
also been investigated. The values of auused in experimentation
are set to six different intervals such as [2, 0], [2.5, 0], [3, 0], [3.5,
0], [4, 0], and [4.5, 0] while keeping the dimensionality of search
space as 30. Fig. 5 shows the performance of MGWO’s variants over
different values of au.
The result reveal that the MGWO-II and MGWO-III does
not affected from value of au for F1 , F2 , F3 , F4 , F9 , F10 and F11
functions. Therefore, a straight line is observed for the above-
mentioned functions. For F1 , F2 , F3 , F4 , F10 , and F11 functions,
performance of MGWO-I’s performance is greatly reduce with in-
crease in the value of au. The optimum value of these functions
for MGWO-I is obtained at au = 2. For F9 function, the value of fit-
ness produced from MGWO-I is first decrease, then increase with
increase in the value of au. The optimum value of F9 is obtained at
au = 2.5.
For F5 , F6 , F7 , F8 , and F13 functions, the performance of
MGWO’s variants greatly reduce with increase in the value of au.
The optimum value of these functions is obtained at au = 2. For
F13 function, the value of this function produced from MGWO-I
increases with increase in the value of au. The optimum value of
this function is obtained at au = 2. For MGWO-II and MGWO-III,
the value of this function is first decreases until the value of au
reaches at 2.5. Thereafter, its value increases with increase in value
of au. The optimum value of this function is obtained at au = 2.5.
The results obtained from Fig. 5 reveal that the performance of
MGWO-II and MGWO-III is least affected as compared to MGWO-I
for different values of au. The reason behind is that the MGWO-II
and MGWO-III have better exploration and exploitation capability
due to incorporate the astrophysics concept. Due to this feature, it
V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254 243

Fig. 5. Effect of au on the performance of MGWO’s variants.

244 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254

Fig. 5. Continued
V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254 245

Fig. 5. Continued
246 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254

Fig. 5. Continued
 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254 247

Fig. 5. Continued

Table 11
Performance comparison with existing variants of GWO on unimodal functions.

IGWO HGWO mGWO PGWO EGWO MGWO−I MGWO−II MGWO−III

F1 2.92E−57 2.21E−58 2.42E−59 2.54E−58 1.12E−59 1.54E−59 0.0 0E+ 0 0 0.0 0E+ 0 0
(2.32E−57) (1.87E−58) (0.0 0E+ 0 0) (7.98E−53) (3.58E−59) (4.77E−58) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F2 9.33E−34 7.83E−35 8.84E−35 5.93E−34 7.41E−35 8.58E−35 0.0 0E+ 0 0 0.0 0E+ 0 0
(5.03E−33) (6.92E−32) (4.96E−35) (1.56E−32) (1.05E−34) (6.69E−35) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F3 5.68E−15 3.98E−15 3.74E−15 4.44E−15 2.93E−15 8.06E−15 0.0 0E+ 0 0 0.0 0E+ 0 0
(8.71E−14) (1.32E−13) (1.91E−15) (0.0 0E+ 0 0) (1.10E−16) (1.92E−14) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F4 2.27E−14 1.99E−14 1.69E−14 1.16E−14 1.00E−14 1.51E−14 0.0 0E+ 0 0 0.0 0E+ 0 0
(4.76E−14) (3.98E−14) (4.52E−14) (2.84E−14) (1.15E−14) (2.89E−14) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F5 2.98E+ 01 2.77E+ 01 2.69E+ 01 2.62E+ 01 1.98E+ 01 2.70E+ 01 1.79E+ 01 1.63E+ 01
(7.16E−01) (8.68E−01) (8.52E−01) (5.59E−01) (4.27E−01) (6.63E−01) (1.20E+ 01) (1.25E+ 01)
F6 8.69E−01 7.27E−01 7.86E−01 5.61E−01 2.21E−03 5.65E−01 2.09E−05 2.04E−05
(2.21E−01) (2.10E−01) (2.44E−01) (1.59E−01) (1.37E−02) (3.81E−01) (9.16E−06) (6.76E−06)
F7 6.79E−03 6.64E−03 6.09E−04 6.83E−04 4.36E−04 7.30E−04 3.16E−05 3.10E−05
(2.27E−02) (3.12E−02) (1.76E−04) (3.42E−04) (3.12E−04) (4.02E−04) (2.72E−05) (2.76E−05)

Table 12
Performance comparison with existing variants of GWO on unimodal functions.

IGWO HGWO mGWO PGWO EGWO MGWO−I MGWO−II MGWO−III

F8 −6.11E+ 03 −6.03E+ 03 −6.21E+ 03 −5.19E+ 03 −8.90E+ 03 −6.04E+ 03 −6.59E+ 03 −6.70E+ 03

(2.42E+ 03) (7.11E+ 02) (4.17E+ 02) (3.64E+ 02) (7.23E+ 02) (7.47E+ 02) (1.47E+ 03) (1.18E+ 03)
F9 1.74E−14 1.69E−14 1.61E−14 1.73E−14 1.19E−14 1.14E−14 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.46E−14) (1.70E−14) (1.41E−13) (2.25E−14) (1.34E−14) (2.75E−14) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F10 8.98E−14 8.90E−14 8.89E−14 8.23E−14 9.30E−15 1.51E−14 8.88E−16 8.88E−16
(2.09E−14) (2.87E−14) (2.80E−14) (2.17E−14) (1.46E−15) (1.87E−15) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F11 1.73E−03 1.69E−03 1.54E−03 1.26E−03 1.19E−04 1.24E−04 0.0 0E+ 0 0 0.0 0E+ 0 0
(1.39E−03) (1.26E−03) (1.42E−03) (1.93E−03) (2.51E−04) (3.47E−04) (0.0 0E+ 0 0) (0.0 0E+ 0 0)
F12 4.19E−02 4.15E−01 4.10E−01 4.04E−01 4.00E−02 4.03E−02 2.99E−02 2.65E−02
(1.61E−01) (1.28E−01) (1.37E−01) (2.74E−02) (1.65E−02) (2.37E−02) (2.63E−02) (5.46E−02)
F13 4.71E−01 4.65E−01 4.52E−01 4.38E−01 4.09E−01 4.01E−01 3.53E−01 2.02E−01
(2.93E−01) (2.33E−01) (2.10E−01) (2.63E−01) (1.81E−01) (1.56E−01) (6.79E−01) (5.54E−01)

is capable to determine the optimal value of fitness functions, even
the value of au is significantly varied. The optimum value for most
of the benchmark test functions is obtained at au = 2.
4.9. Result and discussion
The experimental results reveal that MGWO-III has better fit-
ness function value that makes the approach to converge to better
optima. The improved performance is attributed to the fact that
MGWO-III can exploit and explore well because of the search path
is being oval shaped. The prey weight control helps the approach
jump to more promising regions quickly, especially in complex
search space. The results mentioned in Tables 4 and 5 demonstrate
that MGWO-III has almost zero standard deviation, which ensures
the robustness of the approach regardless of random factor used.
The performance on noisy and high dimensional test data also de-
picts that MGWO-III is a better performer than other algorithms.
The MGWO-III, due to incorporation of the astrophysics concept,
improvises superior balance between exploration and exploitation.
248 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254

Table 14
Statistical results for different algorithms for three-bar truss design problem.

Algorithm Best Mean Worst Std. Dev. Median

MGWO-III 263.89586 263.89613 263.89662 0.0 0 0217 263.89621

MGWO-II 263.89588 263.89618 263.89664 0.0 0 0267 263.89601
MGWO-I 263.89589 263.89633 263.89755 0.0 0 0429 263.89628
GWO 263.89592 263.89663 263.89800 0.0 0 0437 263.89634
SCA 263.89838 263.93564 263.99511 0.028844 263.93062
MVO 263.89594 263.89614 263.89709 0.0 0 0249 263.89608
PSO 264.21827 265.95531 267.45900 1.386148 265.82749
GSA 264.82996 271.03477 279.79251 4.128589 271.07328

Fig. 6. Three-bar truss design.

Table 13
Comparison of best solution obtained from different algorithms
for three−bar truss.

Algorithm Optimal values for variables Optimal Cost

DW DC

MGWO−III 0.788693 0.408199 263.89586

MGWO−II 0.788861 0.407724 263.89588
MGWO−I 0.788561 0.408572 263.89589
GWO 0.788409 0.409003 263.89592
SCA 0.789068 0.407162 263.89838
MVO 0.788993 0.407351 263.89594
PSO 0.781224 0.432548 264.21827 Fig. 7. Welded beam design.
GSA 0.777622 0.448853 264.82996

terms of statistical results including worst, mean and standard de-

5. Constrained engineering design applications
In order to validate the performance of proposed approaches,
they have been applied to real-life engineering problems. The pro-
posed MGWO’s variants are applied to seven well-known con-
strained engineering design problems such as three-bar truss de-
sign, welded beam design, pressure vessel design, tension spring
design, speed reducer design, rolling element bearing. These prob-
lems were used to show the validity and effectiveness of proposed
approaches in real-life problems. The algorithm’s parameter set-
tings are same as used in Section 4.2.
5.1. Constrained problem I: three-bar truss design
The three-bar truss problem (see Appendix A.1), is one of the
most popular constrained engineering design problem. The objec-
tive of this problem is to minimize the weight of structure while
supporting a total load P acting vertically downward. The geom-
etry of the problem is shown in Fig. 6. The design variables are
cross-sectional areas of structural members. Since the system is
symmetric, only cross sections with A1 ( = y1 ) and A2 ( = y2 )are op-
timized [21].
Table 13 shows the comparison of the best solution of the re-
ported methods. From this table, MGWO-III outperforms all other
algorithms in terms of best solution. The comparison of obtained
statistical results for MGWO’s variants with recently developed
techniques including GSA, PSO, MVO, MFO, SCA and GWO is pre-
sented in Table 14. As it can be seen in Table 14, the proposed
MGWO-III shows superiority to other optimization techniques in
viation. MGWO-II performs slightly better than MGWO-I in terms
of best solution.
5.2. Constrained problem II: welded beam design
The another popular engineering design problem is welded
beam design (see Appendix A.2). The objective of this problem is
to find the minimum fabrication cost of the welded beam subject
to constraints on shear stress (τ ), bending stress (σ ), buckling load
(Pc ), end defection (δ ), and side constraint [22]. There are four de-
sign variables associated with this problem such as thickness of
the weld Twh ( = y1 ), length of the welded joint Lwj ( = y2 ), width of
the beam Wb ( = y3 ), and thickness of the beam Tb ( = y4 ), as shown
in Fig. 7.
The comparison of best solutions among several algorithms
is provided in Table 15. The statistical results so obtained are
given in Table 16. Among these algorithms, the proposed MGWO’s
variants provide optimal solution at y∗ = (0.205667, 3.471899,
9.036679, 0.205733) with corresponding optimal cost equal to
F(y∗ ) = 1.724984. It is worth mentioning that MGWO-III provides
better statistical results in terms of worst, mean, median, and stan-
dard deviation.
5.3. Constrained problem III: pressure vessel design
In this constrained problem (see Appendix A.3), the aim is to
minimize the total cost, including cost of material, welding, and
forming [10]. One side of the vessel has flat and other side has
hemispherical shape. The objective is to find the minimum fabri-
cation cost of the vessel subject to constraints on thickness of the
 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254 249

Table 15 Table 18
Comparison of best solution obtained from different algorithms for welded beam. Statistical results for different algorithms for pressure vessel design problem.

Algorithm Optimal values for variables Optimal Cost Algorithm Best Mean Worst Std. Dev. Median

Twh Lwj Wb Tb MGWO-III 5884.0616 5884.7820 5886.3961 0 0 0.977 5884.3491

MGWO-II 5884.0810 5885.5175 5888.5104 001.011 5885.4237
MGWO-III 0.205667 3.471899 9.036679 0.205733 1.724984
MGWO-I 5884.2166 5885.6175 5887.5313 001.259 5885.0638
MGWO-II 0.205667 3.471899 9.036679 0.205733 1.724984
GWO 5884.3689 5886.4441 5892.3207 001.893 5886.2282
MGWO-I 0.205667 3.471899 9.036679 0.205733 1.724984
SCA 6137.3724 6326.7606 6512.3541 126.609 6318.3179
GWO 0.205678 3.471403 9.036964 0.205729 1.724995
MVO 6011.5148 6477.3050 7250.9170 327.007 6397.4805
SCA 0.204695 3.536291 9.004290 0.210025 1.759173
PSO 5891.3879 6531.5032 7394.5879 534.119 6416.1138
MVO 0.205611 3.472103 9.040931 0.205709 1.725472
GSA 11,550.2976 23,342.2909 33,226.2526 5790.625 24,010.0415
PSO 0.197411 3.315061 10.0 0 0 0 0 0.201395 1.820395
GSA 0.147098 5.490744 10.0 0 0 0 0 0.217725 2.172858

Table 16
Statistical results for different algorithms for welded beam problem.

Algorithm Best Mean Worst Std. Dev. Median

MGWO-III 1.724984 1.725156 1.725420 0.0 0 0137 1.725116

MGWO-II 1.724984 1.725159 1.725599 0.0 0 0164 1.725117
MGWO-I 1.724984 1.725156 1.725568 0.0 0 0196 1.725058
GWO 1.724995 1.725228 1.725664 0.0 0 0187 1.725187
SCA 1.759173 1.817657 1.873408 0.027543 1.820128
MVO 1.725472 1.729680 1.741651 0.004866 1.727420
PSO 1.820395 2.230310 3.048231 0.324525 2.244663
GSA 2.172858 2.544239 3.003657 0.255859 2.495114

Fig. 9. Tension/Compression spring design.

Table 19
Comparison of best solution obtained from different algorithms for
spring problem.

Algorithm Optimal values for variables Optimal Cost

DW DC NAC

MGWO-III 0.051658 0.355975 11.33351 0.012666186

MGWO-II 0.051402 0.349839 11.70569 0.012668724
MGWO-I 0.051334 0.348197 11.80809 0.012669594
GWO 0.051178 0.344541 12.04249 0.012672321
Fig. 8. Pressure vessel design. SCA 0.050780 0.334779 12.72269 0.012709667
MVO 0.050 0 0 0.315956 14.22623 0.012816930
Table 17 PSO 0.050 0 0 0.310414 15.0 0 0 0 0.013192580
Comparison of best solution obtained from different algorithms for pressure vessel GSA 0.050 0 0 0.317312 14.22867 0.012873881
design problem.

Algorithm Optimal values for variables Optimal Cost

TSh THd IR L observed from Table 18 that MGWO-III surpassed other algorithms
for offering the best solution in terms of standard deviation, mean,
MGWO-III 0.778197 0.384675 40.315493 199.95932 5884.0616
worst and median. MGWO-II and MGWO-I are the second and third
MGWO-II 0.778403 0.384823 40.331283 199.74989 5884.0810
MGWO-I 0.778181 0.384821 40.318142 199.93194 5884.2166 best algorithm respectively.
GWO 0.778535 0.384960 40.337793 199.65029 5884.3689
SCA 0.817577 0.417932 41.74939 183.57270 6137.3724
MVO 0.845719 0.418564 43.816270 156.38164 6011.5148 5.4. Constrained problem IV: tension/compression spring design
PSO 0.778961 0.384683 40.320913 20 0.0 0 0 0 0 5891.3879
GSA 1.085800 0.949614 49.345231 169.48741 11,550.2976 The tension/compression spring design problem (see
Appendix A.4), as shown in Fig. 9, is a minimization constrained
problem. The objective is to design a tension/compression spring
having minimum weight and satisfying constraints such as deflec-
shell (TSh ), thickness of the head (THd ), inner radius (IR), and length tion, shear stress, surge frequency, and limits on outer diameter
of cylindrical section of the vessel (L). There are four design vari- and design variables [1]. There are three design variables such
ables such as TSh ( = y1 ), THd ( = y2 ), IR( = y3 ), and L( = y4 ), as shown as the wire diameter Dw ( = y1 ), mean coil diameter DC ( = y2 ), and
in Fig. 8. number of active coils NAC ( = y3 ).
Table 17 represents the comparisons of different algorithms Table 19 shows the comparison of the best solutions for the six
for the best obtained solutions. As seen from Table 17, it is ob- optimization algorithms and proposed MGWO’s variants in terms
served that the proposed MGWO-III provides optimal solution at of design variables and function values. The statistical optimiza-
y∗ = (0.778197, 0.384675, 40.315493, 199.95932) with correspond- tion results for the reported algorithms are given in Table 20. From
ing fitness value equal to F(y∗ ) = 5884.0616. The simulation results Table 20, it is evident that the proposed MGWO-III produced the
reveal that MGWO-III provides better optimization than the other best solution with considerable improvement with respect to other
existing techniques. The statistical results thus obtained for pres- algorithms. The statistical results of MGWO-III in terms of mean,
sure vessel problem were compared and shown in Table 18. It is worst, median, and standard deviation were also best among all.
250 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254

Table 20
Statistical results for different algorithms for spring design problem.

Algorithm Best Mean Worst Std. Dev. Median

MGWO-III 0.012666186 0.012679158 0.012720783 0.0 0 0 014 0.012669749

MGWO-II 0.012668724 0.012679599 0.012719337 0.0 0 0 012 0.012672997
MGWO-I 0.012669594 0.012679953 0.012705662 0.0 0 0 011 0.012675766
GWO 0.012672321 0.012697116 0.012720757 0.0 0 0 021 0.012692686
SCA 0.012709667 0.012839637 0.012998448 0.0 0 0 078 0.012844664
MVO 0.012816930 0.014464372 0.017839737 0.001622 0.014021237
PSO 0.013192580 0.014817181 0.017862507 0.002272 0.013192580
GSA 0.012873881 0.013438871 0.014211731 0.0 0 0287 0.013367888

Fig. 10. Speed reducer design.

Table 21
Comparison of best solution obtained from different algorithms for Speed Reducer.

Algorithm Optimal values for variables Optimal Cost

W M P L1 L2 D1 D2

MGWO-III 3.50 010 0 0.7 17 7.300568 7.806687 3.35068 5.286903 2996.798

MGWO-II 3.500569 0.7 17 7.319115 7.80414 3.350477 5.286695 2996.905
MGWO-I 3.500485 0.7 17 7.302095 7.8 3.351258 5.286973 2997.007
GWO 3.500690 0.7 17 7.310933 7.814726 3.351047 5.286741 2997.288
SCA 3.508755 0.7 17 7.3 7.8 3.461020 5.289213 3030.563
MVO 3.508502 0.7 17 7.392843 7.816034 3.358073 5.286777 3002.928
PSO 3.50 0 019 0.7 17 8.3 7.8 3.352412 5.286715 3005.763
GSA 3.60 0 0 0 0 0.7 17 8.3 7.8 3.369658 5.289224 3051.120

5.5. Constrained problem V: speed reducer design Table 22

Statistical results for different algorithms for Speed Reducer problem.

The speed reducer (see Appendix A.4) is a gear box between Algorithm Best Mean Worst Std. Dev. Median
engine and propeller of airplane to permit each to rotate at ef- MGWO-III 2996.798 2997.978 2999.825 0.97880 2997.021
ficient speed, as demonstrated in Fig. 10. The aim of this design MGWO-II 2996.905 2998.517 30 0 0.739 1.10299 2997.223
problem is to minimize the weight of speed reducer subject to MGWO-I 2997.007 2998.927 3002.242 1.81536 2997.980
constraints on bending stress of gear teeth, transverse deflections GWO 2997.288 2999.640 3003.889 1.93193 2999.187
SCA 3030.563 3065.917 3104.779 18.0742 3065.609
of the shafts, surface stress, and stresses in the shafts [6]. There
MVO 3002.928 3028.841 3060.958 13.0186 3027.031
are seven design variables such as face width W( = y1 ), module of PSO 3005.763 3105.252 3211.174 79.6381 3105.252
teeth M( = y2 ), number of teeth on pinion P( = y3 ), length of the GSA 3051.120 3170.334 3363.873 92.5726 3156.752
first shaft between bearings L1 ( = y4 ), length of second shaft be-
tween bearings L2 ( = y5 ), diameter of first shaft D1 ( = y6 ) and di-
ameter of second shaft D2 ( = y7 ). The complexity of this problem is
sults, MGWO-III surpassed other reported algorithms for offering
high as compared to already mentioned problems. This is due to
the best solution.
the fact that it as higher number of constraints.
The optimization algorithms such as GSA, PSO, MVO, MFO, SCA,
GWO, and MGWO variants were applied to this problem. The com- 5.6. Constrained problem VI: rolling element bearing design
parison for the best solution given by such algorithms is presented
in Table 21. The best solution was obtained by MGWO-III at de- The aim of this design problem (See Appendix A.6) is to maxi-
sign variables y∗ = (3.50010, 0.7, 17, 7.300568, 7.806687, 3.35068, mize the dynamic load carrying capacity of a rolling element bear-
5.286903) with an objective function value of F(y∗ ) = 2996.798. The ing as demonstrated in Fig. 11. There are 10 decision variables
obtained statistical results for Speed Reducer design problem were such as pitch diameter Dm ( = y1 ), ball diameter Db ( = y2 ), number
compared and tabulated in Table 22. In terms of statistical re- of ballsZ( = y3 ), inner raceway curvature coefficient fi ( = y4 ), outer
raceway curvature coefficients fo ( = y5 ), KDmin ( = y6 ), KDmax ( = y7 ),
 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254 251

Table 23
Comparison of best solution obtained from different algorithms for Rolling Element Bearing.

Algorithm Optimal values for variables Optimal Cost

Dm Db Z fi fo KDmin KDmax ε e ζ
MGWO-III 125.7051 21.42294 10.99905 0.515 0.515 0.4872 0.63834 0.3 0.07605 0.67328 85,525.305
MGWO-II 125.6730 21.42199 10.99759 0.515 0.515 0.4641 0.67631 0.300282 0.034802 0.66995 85,510.338
MGWO-I 125.6959 21.42125 10.99907 0.515 0.515 0.4289 0.64169 0.30 0 022 0.03578 0.63704 85,513.323
GWO 125.6999 21.42129 10.99781 0.515 0.515 0.4687 0.68807 0.300151 0.04254 0.61701 85,507.111
SCA 125 21.14834 10.96928 0.515 0.515 0.5 0.7 0.3 0.02778 0.62912 83,431.117
MVO 125.6402 21.42250 10.99338 0.515 0.5150 0 0 0.4998 0.68782 0.301348 0.04617 0.60061 85,491.266
PSO 125 20.75388 11.17342 0.515 0.5150 0 0 0.5 0.61503 0.30 0 0 0 0 0.05161 0.60 0 0 0 81,691.202
GSA 125 20.85417 11.14989 0.515 0.517746 0.5 0.61827 0.304068 0.020 0 0 0.624638 82,276.941

Table 24
Statistical results for different algorithms for Rolling Element Bearing design prob-
lem.

Algorithm Best Mean Worst Std. Dev. Median

MGWO-III 85,525.305 85,446.308 85,286.688 0067.332 85,466.592

MGWO-II 85,510.338 85,456.099 85,281.326 0061.278 85,477.609
MGWO-I 85,513.323 85,443.149 85,236.152 0074.050 85,469.449
GWO 85,507.111 85,424.858 85,253.876 0086.684 85,460.241
SCA 83,431.117 81,005.232 77,992.482 1710.777 81,035.109
MVO 85,491.266 85,353.685 85,200.834 0092.431 85,352.601
PSO 81,691.202 50,435.017 32,761.546 13,962.150 42,287.581
GSA 82,276.941 78,002.107 71,043.110 3119.904 78,398.853

Fig. 12. Multiple disk clutch brake design.

parison of the best solutions for above-mentioned six techniques

Fig. 11. Rolling element bearing design.
ε( = y8 ), e( = y9 ) and ζ ( = y10 ). The latter five variables come into
view of constraints and indirectly affect the internal geometry [6].
Table 23 shows the comparison of the best solutions for com-
pared algorithms and proposed MGWO’s variants in terms of func-
tion values and design variables. The statistical results for opti-
mization of rolling element bearing problem using different opti-
mizers are presented in Table 24. Among the compared algorithms,
MGWO-III yielded the best solution thus far. In terms of statistical
results including worst, mean and standard deviation, the proposed
MGWO-III shows superiority to other optimization algorithms.
5.7. Constrained problem VII: multiple disk clutch brake design
The objective of this constrained design problem (Appendix A.7)
is to minimize the mass of the multiple disc clutch brake using
five discrete design variables [22]. These are inner radius Ri ( = y1 ),
outer radius Ro ( = y2 ), thickness of the disc T( = y3 ), actuating force
F( = y4 ), and number of friction surfaces Z( = y5 ), as shown in
Fig. 12.
The problem of multiple disk clutch brake was optimized using
GSA, PSO, MVO, MFO, SCA, and GWO. Table 25 shows the com-
and the proposed approaches in terms of design variables and
function values. The comparison of obtained statistical results for
MGWO’s variants with above-mentioned optimization techniques
is presented in Table 26. From Table 26, it can be seen that the
proposed MGWO-III detected the best solution with considerable
improvement over other algorithms. The optimization statistical re-
sults also indicate that MGWO-III outperforms the other consider-
able algorithms in terms of worst, mean and standard deviation.
MGWO-II is the second best performing algorithm.
These results reveal that the proposed MGWO-III has capability
in handling various constrained engineering design problems and
can provide the optimum solutions. Therefore, it can be concluded
that the MGWO-III is an attractive and effective optimization algo-
rithm for solving both constrained and unconstrained optimization
problems.
6. Conclusions
In this paper, the variants of grey wolf algorithm, based on prey
weight and astrophysics concepts, have been proposed. These have
been implemented and tested on thirteen benchmark test func-
tions. On comparing the results of proposed technique with the
others, it has been observed that MGWO-III outperformed GSA,
PSO, MVO, MFO, SCA, GWO, and other MGWO’s variants. This pa-
per also investigated the effect of scalability and noise on the per-
formance of GSA, PSO, MVO, MFO, SCA, GWO, and MGWO’s vari-
ants. The results reveal that the performance of MGWO’s variants
is less susceptible to scalability and noise as compared to other.
Simulation results also demonstrated that MGWO-III provided the
252 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254

Table 25
Comparison of best solution obtained from different algorithms for Multiple Disk Clutch Brake
Design.

Algorithm Optimal values for variables Optimal Cost

Ri Ro T F Z

MGWO-III 69.99997 90.0 0 0 0 0 1.0 0 0 0 0 10 0 0.0 0 0 0 0 2.31281 0.259771

MGWO-II 69.99989 90.0 0 0 0 0 1.0 0 0 0 0 10 0 0.0 0 0 0 0 2.31285 0.259775
MGWO-I 69.99937 90.0 0 0 0 0 1.0 0 0 0 0 999.925778 2.31329 0.259778
GWO 69.99984 90.0 0 0 0 0 1.0 0 0 0 0 10 0 0.0 0 0 0 0 2.31299 0.259787
SCA 69.86292 90.0 0 0 0 0 1.0 0 0 0 0 10 0 0.0 0 0 0 0 2.32496 0.262286
MVO 69.99844 90.0 0 0 0 0 1.0 0 0 0 0 10 0 0.0 0 0 0 0 2.31283 0.259790
PSO 69.02503 90.0 0 0 0 0 1.0 0 0 0 0 10 0 0.0 0 0 0 0 2.33737 0.272782
GSA 76.48821 97.02592 1.0 0 0 041 797.74530 2.67498 0.320926

Table 26
Statistical results for different algorithms for Multiple Disk Clutch Brake design.
Sub ject to g1 (y ) = τ (y ) − τmax ≤ 0,
Algorithm Best Mean Worst Std. Dev. Median g2 (y ) = σ (y ) − σmax ≤ 0,
MGWO-III 0.259771 0.259840 0.259915 0.0 0 0 031 0.259835 g3 ( y ) = y1 − y4 ≤ 0 ,
MGWO-II 0.259775 0.259837 0.259913 0.0 0 0 031 0.259833 g4 (y ) = δ (y ) − δmax ≤ 0
MGWO-I 0.259778 0.259817 0.259879 0.0 0 0 031 0.259778 g5 (y ) = P − Pc (y ) ≤ 0
GWO 0.259787 0.259838 0.259891 0.0 0 0 033 0.259835
SCA 0.262286 0.267783 0.272872 0.002719 0.268412
g6 (y ) = 0.125 − y1 ≤ 0
MVO 0.259790 0.260426 0.267507 0.001712 0.259887 g7 (y ) = 1.10471y21 + 0.04811y3 y4 (14.0 + y2 )
PSO 0.272782 0.324099 0.371779 0.029831 0.327459 −5.0 ≤ 0
GSA 0.320926 0.343254 0.393482 0.020212 0.337767 Range 0.1 ≤ y1 ≤ 2, 0.1 ≤ y2 ≤ 10, 0.1 ≤ y3 ≤ 10,
0.1 ≤ y4 ≤ 2

where
best results for high-dimensional and noisy benchmark problems.  y

The sensitivity analysis of parameters has also been investigated on P MR

MGWO’s variants, where it was noticed that MGWO-II and MGWO- τ (y ) = τ12 + 2τ1 τ2 2
+ τ22 ; τ1 = √ ; τ2 =
2R 2y 1 y 2 J
III are less prone as compared to MGWO-I. The proposed approach
is applied on seven well-known constrained engineering design

  y + y
2 
problems. The results reveal that MGWO-III provided best optimal y2 √ y2 1 3
results amongst all competitive algorithms. M = P L+ ; J (y ) = 2 2y 1 y 2 2 + ;
2 4 2

y22
y + y
2
1 3
R= +
4 2
Appendix A   
y23 y64 
4.013E 36 y3 E
A.1. Three-bar truss design problem Pc (y ) = 1− ;
L2 2L 4G

6P L 6P L3
σ (y ) = ; δ (y ) =
 √ y4 y23 Ey33 y4
Minimize F ( y ) = 2 2y 1 + y 2
P = 60 0 0 lb, L = 15 in, G = 12 × 106 psi, E = 30 × 106 psi
δmax = 0.25 in, τmax = 13600 psi, σmax = 30 0 0 0 psi
√
2y 1 + y 2
Sub ject to g1 (y ) = √ P − σ ≤ 0, A.3. Pressure vessel design problem
2y21 + 2y1 y2
y2
g2 ( y ) = √ P − σ ≤ 0,
2y21 + 2y1 y2
Minimize F (y ) = 0.6224y1 y2 y3 y4 + 1.7781y2 y23
1
g3 ( y ) = √ P − σ ≤ 0, +3.1661y21 y4 + 19.84y31 y3
2y 2 + y 1
Range 0 ≤ y1 ≤ 1, 0 ≤ y2 ≤ 1 Sub ject to g1 (y ) = −y1 + 0.0193y3 ≤ 0,
L = 100 cm, P = 2 kN/cm , 2
σ = 2kN/cm 2 g2 (y ) = −y2 + 0.00954y3 ≤ 0,
g3 (y ) = −π y23 y4 − 43 π y33 + 12960 0 0 ≤ 0,
g4 (y ) = y4 − 240 ≤ 0

A.2. Welded beam design problem

A.4. Tension/compression spring design problem

Minimize F (y ) = 1.10471y21 y2 + 0.0481y3 y4 (14.0 + y2 ) Minimize F (y ) = (y3 + 2 )y2 y21

 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254 253

where
y32 y3 ⎡  ⎤−0.3
Sub ject to g1 ( y ) = 1 − ≤ 0,  1.72  0.41 10/3
7178y41 1−γ f i (2 f 0 − 1 )
fc = 37.91⎣1 + 1.04 ⎦
4y22 − y1 y2 1 1+γ f 0 (2 f i − 1 )
g2 ( y ) =  + − 1 ≤ 0,
12566 y2 y31 − y41 5108y21   0.41
γ 0.3 (1 − γ )1.39 2 fi
140.45y1 ×
g3 ( y ) = 1 − ≤ 0, (1 + γ )1/3 2 fi − 1
y22 y3
y2 + y1 φ0 = 2π − 2cos−1
g4 ( y ) = −1≤0  
1.5
{(D − d )/2 − 3(T /4)}2 + {D/2 − T /4 − Db}2 − {d/2 + T /4}2
2{(D − d )/2 − 3(T /4 )}{D/2 − T /4 − Db}

Db ri r0
A.5. Speed reducer design problem γ= , fi = , f0 = , T = D − d − 2 Db
Dm Db Db
D = 160, d = 90, Bw = 30, ri = r0 = 11.033
 Range 0.4 ≤ KDmin ≤ 0.5, 0.6 ≤ KDmax ≤ 0.7, 0.3 ≤ e ≤ 0.4,
Minimize F (Y ) = 0.785y1 y22 3.3333y23 + 14.9334y3 − 43.0934
  0.02 ≤ e ≤ 0.1, 0.6 ≤ ζ ≤ 0.85
− 1.508y1 y26 + y27 + · · · 7.4777 y36 + y37 0.5 ( D + d ) ≤ Dm ≤ 0.6 ( D + d ),

+ 0.78054 y4 y26 + y5 y27 0.15(D − d ) ≤ Db ≤ 0.45(D − d ),
4 ≤ Z ≤ 50, 0.515 ≤ fi , f0 ≤ 0.6

A.7. Multiple disk clutch brake design problem

27 397.5
Sub ject to g1 ( y ) = − 1 ≤ 0, g2 ( y ) = − 1 ≤ 0,
y1 y22 y3 y1 y22 y23
1.93y34 1.93y35 

g3 ( y ) = − 1 ≤ 0, g4 ( y ) = − 1 ≤ 0, Minimize F (y ) = π R2o − R2i T (Z + 1 )ρ
y2 y3 y46 y2 y3 y47

2
1.0 750.0y5 Sub ject to g 1 ( y ) = R o − R i − r ≥ 0 ,
g5 ( y ) = + 16.9 × 106 − 1 ≤ 0,
110y36 y2 y3 g2 (y ) = Lmax − (Z + 1 )(T + δ ) ≥ 0,

2 g3 (y ) = Pmax − Prz ≥ 0,
1.0 750.0y5 g4 (y ) = Pmax vsr max − Prz vsr ≥ 0
g6 ( y ) = + 157.5 × 106 − 1 ≤ 0,
85y37 y2 y3 g5 (y ) = vsr max − vsr ≥ 0
y2 y3 5y2 g6 (y ) = Tmax − t ≥ 0
g7 ( y ) = − 1 ≤ 0 , g8 ( y ) = − 1 ≤ 0, g7 ( y ) = Mh − sMs ≥ 0
40 y1
g8 ( y ) = t ≥ 0
y1 1.5y6 + 1.9
g9 ( y ) = − 1 ≤ 0, g10 (y ) = − 1 ≤ 0,
12y2 y4 where
1.1y7 + 1.9 2 R3o − R3i F
g11 (y ) = −1≤0 Mh = μF K , Prz = 
,
y5 3 R2o − R2i π R2o − R2i


2π n R3o − R3i Iz π n
vrz =  2
, Z= 

90 Ro − Ri 2 30 Mh + M f
A.6. Rolling element bearing design problem
r = 20 mm, IZ = 55kgmm2 , Pmax = 1 MPa,
Fmax = 10 0 0 N, Zmax = 15 s, μ = 0.5,
s = 1.5, Ms = 40 Nm, M f = 3N, n = 250 r pm,
Maximize Cd = fc Z 2/3 D1b.8 i f D ≤ 25.4 mm
vsrmax = 10 m/s, Lmax = 30 mm,
Cd = 3.647 fc Z 2/3 D1b.4 i f D > 25.4 mm
Ri min = 60, Rimax = 80, Ro min = 90, Romax = 110,
Fmin = 600, Fmax = 1000, Hmin = 2,
Hmax = 9, Tmin = 1.5, Tmax = 3.
φ0
Sub ject to g1 ( y ) = − Z + 1 ≥ 0,
2sin−1 ( Db /Dm ) References
g2 (y ) = 2Db − KDmin (D − d ) ≥ 0,
g3 (y ) = KDmax (D − d ) − 2Db ≥ 0, [1] Arora JS. Introduction to optimum design. New York: McGraw-Hill; 1989.
g4 ( y ) = ζ B w − Db ≤ 0 , [2] Dorigo M, Birattari M, Stutzle T. Ant colony optimization. IEEE Comput Intell
Mag 2006;1(4):28–39.
g5 ( y ) = Dm − 0 .5 ( D + d ) ≥ 0 , [3] El-Fergany AA, Hasanien HN. Single and multi-objective optimal power flow
g6 ( y ) = ( 0 .5 + e ) ( D + d ) − Dm ≥ 0 , using grey wolf optimizer and differential evolution algorithms. Electr Power
g7 ( y ) = 0 .5 ( D − Dm − Db ) − ε Db ≥ 0 , Compon Syst 2015;43(13):1548–59.
[4] Emary E, Zawbaa HM, Hassanien AE. Binary Grey wolf optimization approaches
g8 (y ) = fi ≥ 0.515, for feature selection. Neurocomputing 2016;172:371–81.
g9 (y ) = f0 ≥ 0.515, [5] Emary E, Zawbaa HM, Grosan C. Experienced grey wolf optimization through
reinforcement learning and neural networks. IEEE Trans Neural Netw Learn
Syst 2017:1–14.
254 V. Kumar, D. Kumar / Advances in Engineering Software 112 (2017) 231–254
[6] Eskandar H, Sadollah A, Bahreininejad A, Hamdi M. Water cycle algorithm: a
novel metaheuristic method for solving constrained engineering optimization
problems. Comput Struct 2012;110:151–66.
[7] Goldberg D. Genetic algorithms in Search, optimization and machine learning.
MA: Addison-Wesley Professional; 1989.
[8] Jitkongchuen, D., 2015, A Hybrid differential evolution with grey wolf opti-
mizer for continuous global optimization, International Conference on Infor-
mation Technology and Electrical Engineering, pp. 51–54, Chiang Mai.
[9] Kamboj VK. A novel hybrid PSO-GWO approach for unit commitment problem.
Neural Comput Appl 2015:1–13 2015.
[10] Kannan BK, Kramer SN. An augmented lagrange multiplier based method for
mixed integer discrete continuous optimization and its applications to me-
chanical design. J Mech Des. 1994;116:405–11.
[11] Kennedy J, Eberhart R. Particle swarm optimization. In: IEEE International Con-
ference on Neural Networks, 4; 1995. p. 1942–8.
[12] Kishor A, Singh PK. Empirical study of grey wolf optimizer. In: International
Conference on Soft Computing for Problem Solving; 2016. p. 1037–49.
[13] Korayem L, Khorsid M, Kassem SS. Using grey wolf algorithm to solve the ca-
pacitated vehicle routing problem. Mater Sci Eng 2015;83:1–10.
[14] Kumar V, Chhabra JK, Kumar D. Variance-based harmony search algorithm for
unimodal and multimodal optimization problems with application to cluster-
ing. Cybern Syst 2014;45:486–511.
[15] Mirjalili S, Mirjalili SM, Lewis A. Grey wolf optimizer. Adv Eng Softw
2014;69:46–61.
[16] Mirjalili S. Moth-flame optimization algorithm: a novel nature-inspired heuris-
tic paradigm. Knowledge-based Syst 2015;89:228–49.
[17] Mirjalili S, Mirjalili SM, Hatamlou A. Multi-Verse optimizer: a nature-inspired
algorithm for global optimization. Neural Comput Appl 2016;27:495–513.
[18] Mirjalili S. SCA: a sine cosine algorithm for solving optimization problems.
Knowledge-based Syst 2016;96:120–33.
[19] Mittal N, Singh U, Sohi BS. Modified grey wolf optimizer for global engineering
optimization. Appl Comput Intell Soft Comput 2016:1–16.
[20] Rashedi E, Nezamabadi-pour H, Saryazdi S. GSA: a gravitational search algo-
rithm. Inf Sci 2009;179(13):2232–48.
[21] Ray T, Liew KM. Society and civilization: an optimization algorithm based on
the simulation of social behavior. IEEE Trans Evol Comput 2003;7:386–96.
[22] Sadollah A, Bahreininejad A, Eskandar H, Hamdi M. Mine blast algorithm: a
new population based algorithm for solving constrained engineering optimiza-
tion problems. Appl Soft Comput 2013;13:2592–612.
[23] Saremi S, Mirjalili SZ, Mirjalili SM. Evolutionary population dynamics and grey
wolf optimizer. Neural Comput Appl 2015;26:1257–63.
[24] Sarafrazi S, Nezamaadi-pour H, Seydnejad SR. A novel hyrid algorithm of
GSA with kepler algorithm for numerical optimization. Comput Inf Sci
2015;27:288–96.
[25] Wen L, Shao-Hong C, Jian-Jun J, Wen-Zhuan Z, Ming-Zhu T. Hybrid grey
wolf optimization algorithm for high dimensional optimization. Control Decis
2016;31(11):1991–7.
[26] Wen L. Grey wolf optimizer based on nonlinear adjustment control parameter.
In: International Conference on Sensors, Mechatronics and Automation; 2016.
p. 643–8.
[27] Yao X, Liu Y, Lin G. Evolutionary programming made faster. IEEE Trans Evol
Comput 1999;3:82–102.
[28] Zhang S, Zhou YQ. Grey wolf optimizer based on powell local optimization
method for clustering analysis. Discrete Dyn Nature Soc 2015:1–17.
[29] Zhu A, Xu C, Li Z, Wu J, Liu Z. Hybridizing grey wolf optimization with dif-
ferential evolution for global optimization and test scheduling for 3D stacked
SoC. J Syst Eng Electron 2015;26:317–28.