Applied Soft Computing 46 (2016) 230–245

Contents lists available at ScienceDirect

Applied Soft Computing

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

An improved krill herd algorithm with global exploration capability

for solving numerical function optimization problems and its
application to data clustering
R. Jensi ∗ , G. Wiselin Jiji
Dr. Sivanthi Aditanar College of Engineering, Tiruchendur, Tamilnadu, India

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

Article history: Krill herd algorithm is a stochastic nature-inspired algorithm for solving optimization problems. The
Received 28 December 2015 performance of krill herd algorithm is degraded by poor exploitation capability. In this study, we propose
Received in revised form 7 April 2016 an improved krill herd algorithm (IKH) by making the krill the global search capability. The enhancement
Accepted 20 April 2016
comprises of adding global search operator for exploration around the defined search region and thus the
Available online 10 May 2016
krill individuals move towards the global best solution. The elitism strategy is also applied to maintain
the best krill during the krill update steps. The proposed method is tested on a set of twenty six well-
Keywords:
known benchmark functions and is compared with thirteen popular optimization algorithms, including
Global optimization
High convergence
original KH algorithm. The experimental results show that the proposed method produced very accurate
Global exploration results than KH and other compared algorithms and is more robust. In addition, the proposed method
Metaheuristic has high convergence rate. The high performance of the proposed algorithm is then employed for data
Data clustering clustering problems and is tested using six real datasets available from UCI machine learning laboratory.
The experimental results thus show that the proposed algorithm is well suited for solving even data
clustering problems.
© 2016 Elsevier B.V. All rights reserved.

1. Introduction
Over the last few decades, many nature-inspired algorithms
have been proposed for solving numerical optimization problems.
Nature-inspired algorithms [3–5] play a vital role in solving many
engineering optimization problems [6–10,23–25,32] owing to the
global exploration and exploitation ability. These algorithms imi-
tate the behavior of living things in nature such as animals, birds,
fishes, etc. Several heuristic algorithms have been developed in the
literature. Genetic algorithm (GA) [17] was proposed by Goldberg
in 1998 simulating the survival of fittest among individuals in the
population over many generations. Evolutionary strategy (ES) [16]
and differential evolution (DE) [13,14] algorithms belong to the
sub class of evolutionary algorithms which use selection, muta-
tion and recombination operators. Population-based Incremental
Learning (PBIL) [18] developed by Shumeet Baluja in 1994, is an
optimization algorithm which combines the genetic algorithm with
simple competitive learning. Particle Swarm Optimization (PSO)
[19] was proposed by Eberhart and Kennedy in 1995, imitates the
∗ Corresponding author.
E-mail address: r jensi@yahoo.co.in (R. Jensi).
food foraging behavior of bird flocks or fish school. Ant Colony
Optimization (ACO) [12] was initially developed by Marco Dorigo,
simulates the behavior of ants searching for a path between food
source and colony. Artificial bee colony (ABC) [11] is another opti-
mization algorithm developed by Karaboga in 2005 based on the
foraging behavior of honey bee swarm. Biogeography-based opti-
mization (BBO) [15] is developed by Simon in 2008, is imitating
the migrating behavior of species between islands. StudGA (SGA)
[20], a variant of GA, is an optimization algorithm in which fittest
individual is selected rather than stochastic selection and shares
this information among others using genetic algorithm operators.
Cuckoo search (CS) [21] which is a recently proposed optimization
algorithm by Yang et al. in 2009, inspired by the obligate brood
parasitic behavior of cuckoo species. Firefly algorithm (FFA) [22]
is developed by Yang in 2010 based on the flashing behaviour of
fireflies. In Refs. [2,35] authors proposed a new PSO algorithm com-
bined with levy flight for solving optimization problems. Krill herd
(KH) [1] is introduced by Gandomi and Alavi in recent times based
on the imitation of krill individuals behavior. Even though there
exists several optimization algorithms, research is still going on
in the development of optimization algorithm which will provide
high convergence rate and global optimum solution.

http://dx.doi.org/10.1016/j.asoc.2016.04.026
1568-4946/© 2016 Elsevier B.V. All rights reserved.
 R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245 231

The remaining section of the paper is organized as follows.
Section 2 provides the variants of krill herd algorithm in solving
function optimization problems. Section 3 briefly explains the orig-
inal krill herd algorithm, and Section 4 presents the proposed IKH
approach. Section 5 provides the experimental results, Section 6
provides the experimental results of data clustering and Section 7
is concluded with future discussion.
where Ni is the movement induced by other krill individuals, Fi is
the foraging action and Di is the random physical diffusion of the
ith krill individuals.
The direction of motion induced, ∝i , depends on the three com-
ponents, namely local swarm density, a target swarm density and
a repulsive swarm density. The movement of a krill individual Ni is
defined as:

Ninew = N max ∝i + ωn Niold (2)

2. Variants of krill herd algorithm
where
In Ref. [1], the authors tested four different krill herd algo- target
∝i = ∝local
i
+ ∝i (3)
rithms such as KH without any genetic operators, KH with crossover
operator, KH with mutation operator and KH with crossover and and Nmax is the maximum induced speed, ωn is the inertia weight,
mutation operators. They concluded that KH with crossover oper- Niold is the motion induced previously, ∝local
i
is the local effect
ator performed well for solving the optimization problems. Hence, offered by neighbours and ∝i
target
is the best krill individual’s target
hereafter KH refers to KH with crossover operator. Krill herd effect.
would fail to reach the global optima due to the poor exploita- The second movement of KH approach foraging action Fi
tion capability. In order to overcome the problem of KH, the depends on two parameters, namely current food location and
researchers developed novel variants of KH method. Krill herd (KH) information about previous food location. The ith krill individual’s
is hybridized with other methods by utilizing the advantages of all motion is described as:
the methods employed.
Wang et al. [28] employed an updated genetic reproduction Fi = Vf ˇi + ωf Fiold (4)
mechanism such as stud selection and crossover operator into
where
KH instead of stochastic selection while updating krill individ-
ual position. The algorithm is named as Stud krill herd because ˇi = ˇifood + ˇibest (5)
the operators are derived from Stud GA. G.G. Wang et al. [29]
introduced an effective biogeography-based krill herd (BBKH) in and Vf is the foraging speed, ωf is the inertia weight of the foraging
which migration operator from BBO is combined with KH. In addi- action, Fiold is the previous foraging motion, ˇifood is the food attrac-
tion, Wang et al. [30] proposed a hybrid differential evolution KH tive and ˇibest is the best fitness found by the ith krill so far. The
method in which differential evolution operator is hybridized with value for ωn , ωf is equal to 0.9 at the first iteration and decreases
KH during krill individual position updating process. Wang et al. gradually to 0.1 at the end of the iteration.
[31] improved KH performance by including levy flight method The third movement of KH approach is random physical diffu-
into KH resulting Levy-flight krill herd algorithm. Wang et al. [37] sion. The physical diffusion of the ith krill individual depends on
developed a hybrid metaheuritic algorithm called HS/KH in which two components, namely maximum diffusion speed and a random
Harmony search is combined with KH in order to perform explo- directional vector and it is defined as:
ration and exploitation effectively. Wang et al. [38] introduced a  I

Levy flight distribution and elitism strategy for updating KH motion Di = Dmax 1 − ı (6)
Imax
calculation. Wang et al. [39] proposed a hybridized algorithm FKH
where operators used in firefly algorithm is combined into KH. Li where Dmax is the maximum diffusion speed, ı is the random vec-
et al. [43] proposed an improved KH algorithm with linear decreas- tor in the range [–1, 1], I is the current generation and Imax is the
ing step. In Ref. [44] authors introduced opposition based learning maximum generation.
strategy and free search operator into KH algorithm in order to Based on the three movements defined above, the position of
avoid the diversity problem. Wang et al. [45] developed a hybrid ith krill individual during the time interval is:
metaheuristic cuckoo search and krill herd (CSKH) algorithm where dXi
operators from cuckoo search are combined with KH to enhance its Xi (t + t) = Xi (t) + t (7)
dt
effectiveness and reliability. Wang et al. [46] proposed a hybrid
It is clearly seen that t is an important parameter and its value
simulated annealing-based krill herd algorithm for solving global
determines the convergence speed. For more details refer [1].
optimization problems.

4. The proposed IKH algorithm

3. Krill herd (KH) algorithm
Sometimes KH gets stuck in local optima due to the poor explo-
Krill herd (KH) [1] is a new metaheuristic population based ration and exploitation. To improve the performance of KH, in
global optimization algorithm. The inspiration of KH algorithm this study we incorporate global search towards the best solution
is the herding behaviour of krill swarm when looking for food into KH so that the proposed algorithm reaches promising region
and communication with each other. The implementation of KH within the defined boundary. In the proposed method, original KH
method is based on three movements such as: algorithm is performed to move the krill into new position. After
exploration stage, global search is carried out with a small step to
(i) Movement influenced by other krill individual. achieve global best solution. Here the step size is determined by
(ii) Foraging action. using the dimension of the problem. Step size is calculated using
(iii) Physical diffusion. Eq. (8):
⎧ exp (−nd)
KH approach follows Lagrangian model for effective search and ⎪
⎨ exp (−2) , if nd > 10
it is described as:
stepsize = (8)
dXi ⎪
⎩ 1
= Ni + Fi + Di (1) , otherwise
dt exp (−0.5)
232 R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245
where nd is the dimension of the problem.
Krill individual position is fine-tuned towards the best solution
with a small step to explore the search space effectively. The fine-
tuning of the krill position is done based on the probability. If the
probability is less than 0.5 then krill position is changed according
to Eq. (9), otherwise Eq. (10) is used to generate new krill position.
 
j j
Xmax − Xmin
⊕ random [−1, 1]
Xinew = Xit+1 + 0.001 ⊕
2 tion f24. While comparing the mean results of the algorithms, IKH
Xinew = Xit+1 ⊕ step
where
 f6, LFPSO on f17, f19, GA on f21, f24 functions. Although ACO, FFA,
 j j
Xmax − Xmin
step = Xit+1 − Xbest ⊕ random [−1, 1] + stepsize ⊕
2
random [−1, 1]
random [−1, 1] is the normally distributed pseudorandom num-
j j
bers in the range [–1,1], stepsize is from Eq. (8), Xmax , Xmin are
maximum and minimum bound values of the jth dimension, respec-
tively and ⊕ represents element-by- element multiplication.
After finding new position of the krill, individual fitness is eval-
uated. The greedy selection technique is employed for choosing the
better krill position (i.e.) if the fitness of the newly calculated krill is
better than current fitness of the krill, then old position is replaced
by the newly found krill position, otherwise the old position is
retained. The above procedure is repeated for all krill individuals.
At the end of current generation, elitism approach is employed to
keep the best krill in the population to the next generation. By this
method, krill individuals are exploring the search space with ran-
dom jump towards the best solution and thus diversity problem is
avoided and converges quickly. The psuedocode and flowchart of
the proposed algorithm is shown in Figs. 1 and 2, respectively.
5. Experimental results
In this section, the experimental results of the proposed work
are given. In order to test the performance of the proposed algo-
rithm, 26 benchmark functions [36,37,40–42] are selected and its
properties are given in Table 1. We intend to improve the perfor-
mance of KH algorithm for solving both unimodal functions and
multimodal functions, and hence the efficiency of the proposed
work is compared with thirteen methods, such as ABC [11], CS [21],
FFA [22], ACO [12], BBO [15], DE [13,14], ES [16], GA [17], PBIL [18],
PSO [19], [3], SGA [20], LFPSO [2] and KH [1]. The code that was used
to generate the results of ACO, BBO, DE, ES, GA, PBIL, PSO, SGA is
available at http://academic.csuohio.edu/simond/bbo and KH code
at [34]. For ACO, BBO, DE, ES, GA, HS, PBIL, and PSO, the parameters
are set as in Refs. [15,37]. The control parameter values for LFPSO,
ABC, CS, and FFA are set as per the suggestions of the authors in
Refs. [2,11,21,22] respectively. For KH and proposed IKH, we set
the same parameter as [1]: Vf = 0.02, Dmax = 0.005, Nmax = 0.01.
For easy comparison, population size for all algorithms is set
to 50 and maximum generation is 50. All the algorithms are exe-
cuted in Matlab 8.2 with Windows OS environment using Intel Core
i3, 3.30 GHz, 3.41 GB RAM. Each algorithm is executed 100 times
independently for each benchmark functions. The mean and best
optimization results of all algorithms over 100 runs for each test
function are given in Tables 2 and 3 respectively. The minimum
value obtained by the algorithms for each test function is marked
in bold.
5.1. Comparison of optimization results
As seen from Table 2, IKH achieves optimal results for 21
functions out of 26 functions. IKH outperforms other thirteen algo-
rithms for functions f1–f7, f9–f16, f18, f20, f22–23, f25–f26 except 5
functions namely f8, f17, f19, f21 and f24. BBO algorithm obtains
optimal mean results of 9.12E + 02 for f8 function (Rosenbrock)
only. LFSPO gets mean optimal results for f17, f19 and f21 functions
(9) and CS algorithm finds mean result of 6.17E − 02 for the func-
indeed outperforms well and reaches optimal values for most of
(10) the unimodal and multimodal functions.
While examining the best optimal value obtained by the algo-
rithms as given in Table 3, IKH algorithm obtains best result on 21
functions of the 26 functions. ACO obtains best value on f5, FFA on
⊕ GA get best value on some functions but the mean optimal results
for those functions are not satisfactory as seen from Table 2, which
(11) indicates that these algorithms fail to obtain the optimal results in
all runs. LFPSO obtains best and mean optimal results on f17 and
f19 functions. IKH outperforms well for 21 functions.
IKH algorithm utilizes exploration and exploitation capability of
KH algorithm in addition it uses the global exploration operator to
update the krill position towards the best krill. IKH enables the krill
to move randomly around the search space thus making the krill to
have more diverse position while updating its position in each iter-
ation. IKH also ensures that it has high convergence rate compared
to KH algorithm. As KH fails to perform global exploration, it does
not reach the global optimum solution for any one of the tested
benchmark functions. Based on the analysis of the experimental
results, we conclude that IKH generates optimal or near optimal
solutions and is robust when compared to other methods.
Wilcoxon’s rank sum test is done to determine statistically sig-
nificant between IKH and the other thirteen methods. The test is
performed on optimum results obtained by 100 independent runs
of the corresponding methods with significance level at ˛ = 0.05.
The test result values for 26 benchmark functions are given in
Tables 4–6 . From the values in Tables 4–6, it can be seen that IKH’s
result is statistically significant when compared to other methods
except for functions f08, f17, f19, f21 and f24.
In order to compare the convergence characteristics of proposed
and compared algorithm for solving the function optimization
problems, convergence graphs are drawn. The convergence graphs
for Rosenbrock, Schwefel2.26, Step, Penalized1, Alpine and Wavy
functions are shown in Fig. 3(a)–(f). The values taken for draw-
ing the graph are true mean optimal solution obtained through
100 independent runs. On seeing the convergence behavior of all
the methods, IKH quickly converges and achieves optimal result
for Rosenbrock, Step, Penalized1, Alpine and Wavy functions and
FFA, LFPSO, KH algorithms perform well next to IKH. For Schwe-
fel2.26 function, BBO and SGA converge slowly and reach the better
solution compared to IKH method. IKH method ranks third for
Schwefel2.26 function among thirteen approaches.
6. Application of IKH to data clustering
Data clustering [47] is the method in which a group of data
objects are divided into groups or clusters in such a way that the
objects within the clusters are having high similarity while the data
objects in different clusters are dissimilar. Data clustering is an
unsupervised technique due to the unknown class labels. The sim-
ilarity between data objects is measured by some distance metric.
There are several distance measurements [47] such as Euclidean
distance, Minkowski metric, Manhatten distance, Cosine similarity,
Jaccard coefficient, Pearson correlation coefficient, and so on.
Table 1
Benchmark functions. D: dimension, C: characteristic, U: unimodal, M: multimodal.

ID Name Function C D Range Global minimum

n

f01 Sphere f1 (x) = xi 2 U 20 −100 ≤ xi ≤ 100 0

i=1

n n

f02 Schwefel 2.22 f2 (x) = |xi | + |xi | U 20 −10 ≤ xi ≤ 10 0

i=1 i=1

i
2
n
f03 Schwefel 1.2 f3 (x) = xj U 20 −100 ≤ xi ≤ 100 0
i=1
j=1
 
f04 Schwefel 2.21 f4 (x) = max |xi |, 1 ≤ i ≤ n U 20 −100 ≤ xi ≤ 100 0
i

R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245

n−1   2

f05 Rosenbrock f5 (x) = 100 xi+1 − xi 2 + (xi − 1)2 U 20 −30 ≤ xi ≤ 30 0
i=1

n
2
f06 Step f6 (x) = (xi + 0.5) U 20 −100 ≤ xi ≤ 100 0
i=1

f07 Quartic with noise f7 (x) = ixi 4 + random (0, 1) U 20 −1.28 ≤ xi ≤ 1.28 0
i=1

n   
f08 Schwefel 2.26 f8 (x) = 418.9828 × n − xi sin |xi | M 20 −512 ≤ xi ≤ 512 0
i=1

n
 
f09 Rastrigin f9 (x) = 10n + xi 2 − 10cos (2xi ) M 20 −5.12 ≤ xi ≤ 5.12 0
i=1
  n
  n

1 1
f10 Ackley f10 (x) = −20 exp −0.2 n
xi 2 − exp n
cos(2xi ) + (20 + e) M 20 −32 ≤ xi ≤ 32 0
i=1 i=1

n n  
x
f11 Griewank f11 (x) = 1
4000
xi 2 − cos √i +1 M 20 −600 ≤ xi ≤ 600 0
i
i=1 i=1
 2 n−1 2
 2
  n
f12 Penalized 1 f12 (x) = 
n
(10sin (y1 ) + (yi − 1) 1 + 10sin (yi+1 ) + (yn − 1)2 + u (xi , 10, 100, 4) M 20 −50 ≤ xi ≤ 50 0
i=1 i=1
1
yi = 1 + 4 (xi + 1)
 2 n−1 2
 2
  2
 n
f13 Penalized 2 f13 (x) = 1
10
sin (3x1 ) + (xi − 1) 1 + sin (3xi+1 ) + (xn − 1)2 1 + sin (2xn ) + u (xi , 5, 100, 4) M 20 −50 ≤ xi ≤ 50 0
i=1 i=1

f14 Alpine f14 (x) = |xi sin (xi ) + 0.1xi | M 20 −10 ≤ xi ≤ 10 0

i=1

233
 234
Table 1 (Continued)

ID Name Function C D Range Global minimum


n−1
(x2 +1)
 x2 +1
f15 Brown f15 (x) = (xi2 ) i+1 + xi+1
2 i
U 20 −1 ≤ xi ≤ 4 0
i=1

n
 2
f16 Dixon & Price f16 (x) = (x1 − 1)2 + i 2xi2 − xi−1 U 20 −10 ≤ xi ≤ 10 0
i=2

n n
2
f17 Trid10 f17 (x) = (xi − 1) − xi xi−1 M 10 −100 ≤ xi ≤ 100 −200
i=1 i=2

R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245

f18 Holzman 2 function f18 (x) = ixi 4 U 20 −10 ≤ xi ≤ 10 0
i=1

n−1
2
 2
 2
 2

f19 Levy f19 (x) = (xi − 1) 1 + sin (3xi+1 ) + sin (3x1 ) + |xn − 1| 1 + sin (3xn ) M 20 −10 ≤ xi ≤ 10 0
i=1

n−1  2 2 
sin2 100x +x −0.5
f20 Pathological function f20 (x) = 0.5 +  i i+1
2 M 20 −100 ≤ xi ≤ 100 0
1+0.001 x2 −2xi xi+1 +x2
i=1 i i+1

 2
n n
k  xi k
f21 Perm f21 (x) = i +ˇ i
−1 M 4 −4 ≤ xi ≤ 4 0
k=1 i=1

n/4
2 2 4 4
f22 Powell f22 (x) = (x4i−3 + 10x4i−2 ) + 5(x4i−1 − x4i ) + (x4i−2 − x4i−1 ) + 10(x4i−3 − x4i ) U 20 −4 ≤ xi ≤ 5 0
i=1

f23 Sum square f23 (x) = ixi 2 U 20 −10 ≤ xi ≤ 10 0

i=1

n
 n
2
f24 Power sum f24 (x) = xik − bk M 4 0 ≤ xi ≤ 4 0
k=1 i=1
where b = [8,18,44,114]

n n
2 n
4
1 1
f25 Zakharov f25 (x) = xi2 + 2
ixi + 2
ixi M 20 −5 ≤ xi ≤ 10 0
i=1 i=1 i=1

n
−x2
i
1
f26 Wavy1 f26 (x) = 1 − n
cos (kxi ) e 2 M 20 − ≤ xi ≤  0
i=1
Table 2
Mean real optimization results in twenty six benchmark functions.

ID ABC CS FFA ACO BBO DE ES GA PBIL PSO SGA LFPSO KH IKH

f01 1.24E + 03 8.00E + 03 1.75E + 01 2.04E + 03 1.01E + 03 2.43E + 03 1.60E + 04 4.76E + 03 2.51E + 04 8.86E + 03 7.63E + 02 1.64E + 01 1.86E + 02 2.61E − 35
f02 5.49E − 01 4.64E + 01 3.27E + 00 5.06E + 01 7.13E + 00 2.10E + 01 7.92E + 01 3.66E + 01 6.10E + 01 4.93E + 01 1.06E + 01 1.13E + 00 1.23E + 06 2.70E − 25

R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245

f03 1.91E + 04 1.59E + 04 1.03E + 03 1.62E + 04 1.17E + 04 2.52E + 04 2.82E + 04 2.02E + 04 2.89E + 04 1.94E + 04 1.70E + 04 9.42E + 02 1.39E + 04 1.14E − 07
f04 6.54E + 01 4.96E + 01 5.26E + 00 4.05E + 01 4.60E + 01 5.08E + 01 6.37E + 01 5.09E + 01 6.68E + 01 5.34E + 01 3.64E + 01 6.16E + 00 7.59E + 00 3.31E − 17
f05 8.86E + 04 6.47E + 06 1.02E + 03 4.84E + 07 2.94E + 05 1.24E + 06 3.55E + 07 2.96E + 06 5.46E + 07 1.02E + 07 7.62E + 04 7.29E + 02 5.22E + 04 1.87E + 01
f06 1.57E + 03 8.13E + 03 2.04E + 01 2.13E + 03 8.84E + 02 2.32E + 03 1.56E + 04 4.41E + 03 2.54E + 04 9.08E + 03 7.82E + 02 2.27E + 01 2.37E + 02 0.00E + 00
f07 8.58E − 01 2.73E + 00 6.21E − 02 1.01E + 01 6.83E + 00 8.81E + 00 2.29E + 01 1.03E + 01 2.59E + 01 1.10E + 01 8.08E + 00 4.42E − 02 8.63E − 02 3.96E − 04
f08 2.64E + 03 3.69E + 03 6.32E + 03 1.78E + 03 9.12E + 02 3.44E + 03 4.21E + 03 1.47E + 03 5.37E + 03 5.11E + 03 9.91E + 02 3.71E + 03 5.72E + 03 1.03E + 03
f09 7.90E + 01 1.59E + 02 9.20E + 01 1.59E + 02 3.53E + 01 1.38E + 02 2.18E + 02 1.49E + 02 2.23E + 02 1.65E + 02 4.78E + 01 4.85E + 01 1.09E + 02 0.00E + 00
f10 1.52E + 01 1.79E + 01 2.68E + 00 1.59E + 01 8.60E + 00 1.30E + 01 1.89E + 01 1.71E + 01 1.96E + 01 1.65E + 01 9.12E + 00 2.74E + 00 5.51E + 00 8.88E − 16
f11 1.33E + 01 7.23E + 01 1.13E + 00 1.29E + 01 9.12E + 00 2.23E + 01 1.04E + 02 4.41E + 01 2.30E + 02 8.23E + 01 7.94E + 00 1.16E + 00 2.47E + 00 2.79E − 02
f12 3.97E − 01 4.30E + 06 1.44E + 00 8.76E + 07 5.05E + 04 2.90E + 05 4.28E + 07 7.00E + 05 8.82E + 07 6.24E + 06 2.50E + 01 7.89E − 01 1.84E + 02 7.65E − 03
f13 7.07E + 04 1.86E + 07 1.94E − 04 2.02E + 08 3.47E + 05 2.52E + 06 1.14E + 08 4.41E + 06 2.25E + 08 2.78E + 07 2.12E + 04 6.73E − 04 1.39E + 03 3.27E − 05
f14 6.09E + 00 1.63E + 01 2.07E + 00 1.28E + 01 2.54E + 00 1.49E + 01 2.68E + 01 1.18E + 01 2.95E + 01 1.91E + 01 3.18E + 00 4.97E − 01 3.85E + 00 7.53E − 26
f15 8.43E − 03 1.73E + 01 2.06E + 01 4.58E + 20 5.94E + 00 7.05E + 00 4.61E + 03 2.53E + 01 5.76E + 02 1.32E + 05 2.44E + 00 4.27E + 01 1.52E + 02 4.16E − 56
f16 2.07E + 02 3.50E + 04 1.70E + 01 4.05E + 04 2.18E + 03 7.86E + 03 2.44E + 05 1.88E + 04 2.91E + 05 3.54E + 04 1.04E + 03 7.28E + 00 3.79E + 02 6.74E − 01
f17 −1.42E + 02 6.77E + 02 −6.90E + 01 2.46E + 02 1.39E + 02 5.44E + 02 2.62E + 03 5.65E + 02 4.07E + 03 8.38E + 02 2.68E + 02 -1.91E + 02 −4.75E + 01 −1.79E + 02
f18 2.12E + 01 8.68E + 03 2.01E + 00 1.07E + 03 3.70E + 02 1.75E + 03 3.22E + 04 2.51E + 03 7.05E + 04 8.26E + 03 1.39E + 02 2.57E − 01 8.27E + 01 5.88E − 97
f19 2.55E + 01 1.44E + 02 2.03E + 00 1.32E + 02 1.46E + 01 3.86E + 01 3.22E + 02 8.76E + 01 3.51E + 02 1.15E + 02 1.18E + 01 8.47E − 01 1.72E + 01 1.27E + 00
f20 6.77E + 00 4.52E + 00 7.27E + 00 6.70E + 00 5.24E + 00 4.96E + 00 3.59E + 00 5.25E + 00 7.49E + 00 6.36E + 00 4.97E + 00 6.48E + 00 4.54E − 01 4.16E − 01
f21 1.29E + 00 2.08E + 00 7.43E + 01 4.56E + 00 7.27E + 00 2.53E + 00 1.86E + 01 1.13E + 01 6.06E + 00 4.21E + 00 1.31E + 01 9.32E − 01 1.18E + 02 1.70E + 00
f22 4.83E + 01 6.84E + 02 7.30E + 00 3.23E + 03 1.96E + 02 1.36E + 03 4.96E + 03 6.31E + 02 4.04E + 03 8.84E + 02 8.74E + 01 5.50E + 00 3.00E + 02 2.47E − 50
f23 1.19E + 02 7.08E + 02 8.79E + 00 9.16E + 02 9.20E + 01 2.09E + 02 2.07E + 03 5.26E + 02 2.39E + 03 6.92E + 02 7.23E + 01 1.77E + 00 1.67E + 00 2.28E − 48
f24 1.09E − 01 6.17E − 02 5.25E − 01 2.25E − 01 9.50E − 01 4.09E − 01 8.07E + 00 5.16E + 00 6.31E + 00 2.59E + 00 5.21E + 00 7.38E + 00 4.72E − 01 1.22E − 01
f25 2.12E + 02 2.10E + 02 5.35E + 01 5.51E + 06 7.86E + 04 5.20E + 04 8.68E + 06 2.53E + 05 5.39E + 06 1.44E + 07 1.41E + 04 1.36E + 02 1.04E + 05 2.10E − 48
f26 4.38E − 01 5.81E − 01 4.88E − 01 7.26E − 01 1.61E − 01 5.96E − 01 7.35E − 01 6.02E − 01 7.02E − 01 6.60E − 01 2.78E − 01 3.26E − 01 5.91E − 01 0.00E + 00
Total 0 1 0 0 1 0 0 0 0 0 0 3 0 21

*The best value obtained by the algorithms for each benchmark function is shown in boldface.

235
 236
Table 3
Best real optimization results in twenty six benchmark functions.

ABC CS FFA ACO BBO DE ES GA PBIL PSO SGA LFPSO KH IKH

F01 1.88E + 00 4.69E + 03 5.71E − 01 8.91E + 02 3.05E + 02 1.12E + 03 1.04E + 04 8.13E + 02 1.46E + 04 5.56E + 03 1.35E + 02 2.17E + 00 8.65E + 00 3.82E − 48
F02 2.92E − 01 3.15E + 01 1.25E + 00 3.26E + 01 3.10E + 00 1.45E + 01 4.97E + 01 1.68E + 01 4.57E + 01 2.72E + 01 4.30E + 00 5.35E − 01 5.19E + 01 2.42E − 27

R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245

F03 1.04E + 04 9.35E + 03 3.39E + 02 6.47E + 03 4.91E + 03 1.68E + 04 1.60E + 04 6.38E + 03 1.55E + 04 1.09E + 04 7.23E + 03 2.44E + 02 6.13E + 03 6.69E − 44
F04 4.26E + 01 3.75E + 01 1.66E + 00 2.06E + 01 2.70E + 01 3.55E + 01 4.97E + 01 1.94E + 01 4.79E + 01 3.01E + 01 1.83E + 01 3.15E + 00 3.91E + 00 1.13E − 22
F05 1.54E + 02 8.23E + 05 6.30E + 01 2.42E − 26 1.66E + 04 3.27E + 05 1.68E + 07 1.41E + 05 2.82E + 07 2.20E + 06 6.47E + 03 6.28E + 01 2.11E + 03 1.84E + 01
F06 4.00E + 00 3.87E + 03 0.00E + 00 5.42E + 02 3.21E + 02 1.13E + 03 9.16E + 03 1.19E + 03 1.50E + 04 4.75E + 03 1.66E + 02 6.00E + 00 2.50E + 01 0.00E + 00
F07 1.09E − 01 7.36E − 01 1.27E − 02 7.23E + 00 5.41E + 00 6.62E + 00 1.34E + 01 6.71E + 00 1.38E + 01 6.61E + 00 6.44E + 00 6.78E − 03 1.37E − 02 5.19E − 06
F08 1.76E + 03 2.79E + 03 5.08E + 03 1.01E + 03 4.13E + 02 2.47E + 03 3.29E + 03 3.70E + 02 4.08E + 03 3.53E + 03 2.72E + 02 1.73E + 03 4.72E + 03 1.19E + 02
F09 4.85E + 01 1.16E + 02 3.91E + 01 1.08E + 02 1.82E + 01 9.82E + 01 1.59E + 02 9.14E + 01 1.80E + 02 1.33E + 02 2.20E + 01 1.71E + 01 3.99E + 01 0.00E + 00
F10 9.49E + 00 1.52E + 01 1.35E + 00 1.19E + 01 6.20E + 00 1.11E + 01 1.69E + 01 1.35E + 01 1.84E + 01 1.42E + 01 6.28E + 00 1.04E + 00 3.27E + 00 8.88E − 16
F11 1.04E + 00 3.20E + 01 3.27E − 01 4.38E + 00 3.33E + 00 9.78E + 00 6.49E + 01 1.13E + 01 1.32E + 02 4.39E + 01 2.28E + 00 9.95E − 01 1.10E + 00 0.00E + 00
F12 3.90E − 03 1.45E + 05 9.70E − 02 3.17E − 02 5.60E + 00 9.61E + 02 1.00E + 07 7.62E + 00 7.63E + 06 3.64E + 05 2.25E + 00 7.25E − 02 9.53E + 00 2.09E − 03
F13 6.14E − 07 1.90E + 06 2.64E − 09 1.00E + 00 2.60E + 01 3.07E + 05 1.69E + 07 4.05E + 03 8.38E + 07 4.02E + 06 8.63E + 00 2.83E − 10 4.48E − 08 7.50E − 12
F14 2.24E + 00 1.09E + 01 3.74E − 01 7.64E + 00 1.31E + 00 1.10E + 01 2.02E + 01 6.93E + 00 1.78E + 01 1.33E + 01 1.31E + 00 6.64E − 02 1.82E − 01 2.44E − 28
F15 1.07E − 03 5.74E + 00 8.66E + 00 6.02E − 02 1.92E + 00 2.93E + 00 3.45E + 01 4.38E + 00 3.58E + 01 2.94E + 01 5.71E − 01 4.51E + 00 2.39E + 01 4.30E − 71
F16 2.12E + 00 6.35E + 03 2.22E + 00 1.05E + 04 5.04E + 02 2.49E + 03 7.61E + 04 1.25E + 03 8.23E + 04 1.05E + 04 3.33E + 02 1.18E + 00 2.25E + 01 6.67E − 01
F17 −2.08E + 02 3.89E + 01 −1.75E + 02 −1.10E + 02 −1.84E + 02 −7.10E + 01 6.57E + 02 −1.79E + 02 1.39E + 03 2.20E + 02 −1.48E + 02 -2.10E + 02 −2.09E + 02 −2.07E + 02
F18 3.24E − 05 1.82E + 03 6.57E − 02 8.16E + 01 2.97E + 01 2.87E + 02 2.59E + 03 2.31E + 02 3.26E + 04 1.79E + 03 7.98E + 00 5.66E − 03 5.83E + 00 1.83E − 106
F19 2.52E − 01 6.73E + 01 4.23E − 01 6.17E + 01 4.99E + 00 1.70E + 01 2.28E + 02 2.65E + 01 2.05E + 02 5.99E + 01 4.69E + 00 8.76E − 02 3.84E + 00 2.33E − 01
F20 5.55E + 00 2.70E + 00 5.99E + 00 6.09E + 00 4.08E + 00 3.50E + 00 2.30E + 00 3.88E + 00 6.47E + 00 2.30E + 00 3.50E + 00 4.62E + 00 2.23E − 01 0.00E + 00
F21 4.22E − 02 7.25E − 02 1.55E − 01 2.96E − 01 1.33E − 01 3.80E − 02 2.79E − 01 0.00E + 00 7.46E − 02 5.85E − 02 4.68E − 01 6.25E − 04 7.05E − 01 2.54E − 03
F22 1.84E + 00 1.36E + 02 1.28E + 00 9.05E + 02 4.18E + 01 4.85E + 02 1.70E + 03 9.09E + 01 1.89E + 03 3.30E + 02 1.84E + 01 6.54E − 01 9.89E + 01 3.55E − 55
F23 8.94E − 02 3.84E + 02 1.70E + 00 2.83E + 02 2.09E + 01 9.35E + 01 1.16E + 03 9.66E + 01 1.34E + 03 3.85E + 02 2.09E + 01 3.41E − 01 4.28E − 02 3.32E − 52
F24 3.77E − 03 1.46E − 03 1.57E − 03 0.00E + 00 0.00E + 00 6.71E − 03 1.52E − 01 0.00E + 00 2.15E − 01 1.34E − 01 4.87E − 02 1.89E − 02 8.99E − 03 3.02E − 04
F25 1.13E + 02 9.39E + 01 1.94E + 01 1.02E + 02 1.94E + 03 1.32E + 04 1.57E + 06 1.15E + 04 1.40E + 06 3.97E + 05 3.12E + 02 2.55E + 01 1.58E + 02 3.09E − 64
F26 3.09E − 01 5.10E − 01 2.65E − 01 6.36E − 01 5.57E − 02 4.93E − 01 6.32E − 01 5.08E − 01 6.01E − 01 5.65E − 01 1.77E − 01 1.93E − 01 4.55E − 01 0.00E + 00
Total 0 0 1 1 0 0 0 2 0 0 0 2 0 21

*The best value obtained by the algorithms for each benchmark function is shown in boldface.
Table 4
Statistical results of IKH for the benchmark functions f01–f11 using Wilcoxon rank sum test (number of runs = 100).

Wilcoxon rank sum test IKH vs.

ABC CS FFA ACO BBO DE ES GA PBIL PSO SGA LFPSO KH

f01 p-value 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157

f02 p-value 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −12.2158 −12.2158 −12.2157 −12.2157 −12.2158 −12.2158 −12.2157 −12.2158 −12.2157 −12.2157

2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34

R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245

f03 p-value
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2158 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157

f04 p-value 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −12.2158 −12.2158 −12.2157 −12.2157 −12.2158 −12.2158 −12.2157 −12.2158 −12.2157 −12.2157

f05 p-value 2.56E − 34 2.56E − 34 2.56E − 34 9.26E − 32 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −11.727 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157

f06 p-value 5.64E − 39 5.64E − 39 2.1E − 38 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.5E − 39 5.63E − 39
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −13.0591 −13.0591 −12.9585 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.061 −13.0592

f07 p-value 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157

f08 p-value 2.56E − 34 2.56E − 34 2.56E − 34 9.33E − 24 0.004937 2.56E − 34 2.56E − 34 6.49E − 10 2.56E − 34 2.56E − 34 0.208708 2.72E − 34 2.56E − 34
h-value 1 1 1 1 1 1 1 1 1 1 0 1 1
Zval −12.2157 −12.2157 −12.2157 −10.0484 2.811119 −12.2157 −12.2157 −6.17811 −12.2157 −12.2157 1.257127 −12.2108 −12.2157

f09 p-value 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591

f10 p-value 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591

f11 p-value 2.89E − 34 2.56E − 34 1.54E − 33 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 7.76E − 34 2.64E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.206 −12.2157 −12.0691 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2158 −12.1253 −12.2133

Total 11 11 11 11 11 11 11 11 11 11 10 11 11

237
 238
Table 5
Statistical results of IKH for the benchmark functions F12–F22 using Wilcoxon rank sum test (number of runs = 100).

Wilcoxon rank sum test IKH vs.

ABC CS FFA ACO BBO DE ES GA PBIL PSO SGA LFPSO KH

f12 p-value 1.21E − 29 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −11.3068 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157

f13 p-value 5.41E − 29 2.56E − 34 0.000132 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 0.001549 4.15E − 26
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −11.1748 −12.2157 −3.82268 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −3.16541 −10.5689

2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34

R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245

f14 p-value
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157

f15 p-value 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157

f16 p-value 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157

f17 p-value 0.000763 2.56E − 34 1.32E − 30 5.42E − 34 2.91E − 32 2.64E − 34 2.56E − 34 1.95E − 33 2.56E − 34 2.56E − 34 9E − 34 1.71E − 05 5.72E − 30
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −3.36577 −12.2157 −11.4998 −12.1546 −11.8248 −12.2133 −12.2157 −12.0496 −12.2157 −12.2157 −12.1131 4.299143 −11.3728

f18 p-value 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157

f19 p-value 1.03E − 22 2.56E − 34 1.87E − 08 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 1.08E − 09 2.64E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −9.80898 −12.2157 −5.62346 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 6.097477 −12.2133

f20 p-value 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.22E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 0.000131
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2272 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −3.82513

f21 p-value 0.944483 0.005206 9.53E − 26 1.5E − 12 7.34E − 15 3.32E − 05 1.41E − 28 1.91E − 15 4.78E − 19 9.53E − 12 1.03E − 18 5.75E − 05 6.97E − 27
h-value 0 1 1 1 1 1 1 1 1 1 1 1 1
Zval 0.069637 −2.79402 −10.4907 −7.07484 −7.77855 −4.1501 −11.0893 −7.94717 −8.91715 −6.81339 −8.83167 4.02304 −10.735

f22 p-value 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157

Total 10 11 11 11 11 11 11 11 11 11 11 11 11
Table 6

R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245

Statistical results of IKH for the benchmark functions F23–F26 using Wilcoxon rank sum test (number of runs = 100).

Wilcoxon rank sum test IKH vs.

ABC CS FFA ACO BBO DE ES GA PBIL PSO SGA LFPSO KH

f23 p-value 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157

f24 p-value 0.022124 0.746128 1.54E − 07 7.36E − 12 1.79E − 20 1.19E − 16 4.21E − 33 1.85E − 23 1.25E − 33 2.51E − 32 9.29E − 27 6.17E − 33 1.36E − 15
h-value 1 0 1 1 1 1 1 1 1 1 1 1 1
Zval −2.28823 0.323749 −5.24718 −6.85056 −9.27405 −8.28431 −11.986 −9.98075 −12.0863 −11.837 −10.7085 −11.9543 −7.98866

f25 p-value 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34 2.56E − 34
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157 −12.2157

f26 p-value 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.63E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39 5.64E − 39
h-value 1 1 1 1 1 1 1 1 1 1 1 1 1
Zval −13.0591 −13.0591 −13.0591 −13.0591 −13.0592 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591 −13.0591

Total 4 3 4 4 4 4 4 4 4 4 4 4 4

239
240 R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245

1: Initialize algorithm parameters Vf, Dmax , Nmax , NP, MI , and minimum and maximum bounds
2: Randomly generate krill individuals (solutions)
3: Evaluate the objective function value f and find the worst and best fitness values
4: iteration=1
5: while (stopping condition is not met) do
6: Store the pre-specified number of best krill
7: for each krill
8: Calculate Movement influenced by other krill individual
9: Calculate Foraging action
10: Calculate Physical diffusion
11: Implement crossover operator.
12: Update krill position using Eq. (7).
13: if rand < 0.5 then
14: Fine-tune the krill position using Eq. (9)
15: else
16: Fine-tune the krill position using Eq. (10)
17: end if
18: Evaluate the objective function value f and update the krill individual if necessary
19: end for
20: Replace the worst krill with the best krill stored before
21: end while
22: Output the global best solution.

Fig. 1. Pseudocode of the proposed algorithm.

Clustering is widely used in many fields of science and engi- where, cj is cluster center for a cluster j and is calculated as follows:
neering and it must often be solved as part of complicated tasks in
pattern recognition, data mining, information retrieval and image 1
cj = di (14)
analysis. The clustering algorithms are mainly classified into two nj
di ∈ cj
[47]: hierarchical and partitional. The most well known partitional
clustering algorithm is K-means which is the center-based clus- where nj is the total number of objects in cluster j.
tering algorithm. The advantage of K-means algorithm is simple
and efficient. But K-means suffers from initial cluster seed selection 6.2. K-means algorithm
since it is easily trapped in local minima. In this section, we intend
to apply the efficiency of proposed IKH algorithm to solve the data K-means [47] is the simplest partitional clustering algorithm
clustering problems. Several clustering approaches are introduced and it is widely used due its simplicity and efficiency. Given a set of
in the literature [26,27,48–57]. N data objects and the number of clusters k, the k-means algorithm
proceeds as follows:
Step 1: Randomly select ‘k’ cluster centers.
6.1. The problem statement
Step 2: Calculate the Euclidean distance between each data point
and cluster centers.
Clustering is the process of partitioning the set of N data objects
Step 3: Assign the data point to the cluster center whose distance
into K clusters or groups based on some distance (or similarity)
from the cluster center is minimum of all the cluster centers.
metric. Let D = {d1 , d2 ,. . .,dN } be a set of N data objects to be
Step 4: Update cluster center using Eq. (14).
partitioned and each data object di , i = 1,2,. . .,N is represented as
Step 5: If no data point was reassigned then stop, otherwise
di = {di1 ,di2 ,. . ..,dim } where dim represents mth dimension value of
repeat from step 2.
data object i.
The aim of clustering algorithm is to find a set of K partitions
C = {C1 ,C2 , . . ..,Ck |∀k: Ck =
/ ∅ and ∀l =/ k: Ck ∩Cl = ∅} in such a way 6.3. IKH for data clustering
that objects within the clusters are more similar and dissimilar
to objects in different clusters. These similarities are measured by To start with IKH for solving data clustering problems, two mod-
some optimization criterions, especially squared error function and ifications have to be done in the IKH algorithm. The first one is
it has been calculated as follows: solution representation and the other is fitness calculation. Given
the data set D and the number of clusters k as input to the algorithm,
k k   the solution is represented as a row vector of size k × m, where k
f = min E di , cj (12) is the number of clusters and m is the number of features for the
j=1 i=1
clustering problem and it is shown in Fig. 4.
where cj represents a jth cluster center; E is a distance measure Then the population of solutions is represented as given below:
between a data object di and a cluster center cj . This optimization ⎡ ⎤
S1
criterion is used as the objective function value in our study. There
⎢ S2 ⎥
are many distance metric used in literature. In our study Euclidean ⎢ ⎥
⎢ ⎥
distance is used as distance metric which is defined as follows: P = ⎢ S3 ⎥ (15)
⎢ . ⎥
 ⎣ .. ⎦

  M

E di , cj = 
2 SN
dim − cjm (13)
m=1 Si = [C1, C2. . .., Ck] (16)
 R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245 241

Initialize the parameters

Randomly initialize the swarm

Calculate objective function values

Sort the krill and store the best krill

j=1

Select jth krill

Calculate induced motion, foraging motion, physical

diffusion by using (2), (4), and (6) respectively

Apply crossover operator

Update krill position

No Yes
rand<0.5

Fine-tune position in stepwise Fine-tune position using

using Eq. (9) Eq. (11)

Calculate objective function value

Update krill position if improved

j=j+1
Yes

j<NP
No
Replace the worst krill with the best krill and
find the current best krill in the swarm

Yes
iteration<
maxiter

No
Output Best Solution

Fig. 2. Flowchart of the proposed algorithm.

 
Cj = [cj1, cj2, . . .cjm] , ∀j ∈ 1, 2, . . ., k
where k is the number of clusters, m is the set of features, N is the
number krill individuals.
The population solutions are randomly initialized with the data
objects as cluster centroids. The second modification made to the
algorithm is fitness calculation. Fitness is calculated by using Eq.
(17) (12). The remaining steps of KH and IKH algorithms are executed
as described in Sections 3 and 4 respectively.
6.4. Experimental results of IKH for data clustering
To evaluate the performance of proposed algorithm for data
clustering, six datasets have been used. The datasets, namely,
242 R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245

Fig. 3. Performance Comparison for (a) Rosenbrock (b) Schwefel2.26 (c) Step (d) Penalized1 (e) Alpine (f) Wavy functions.

c12 c13 … …
c11 c1m c21 c22 c23 c2m ck1 ck2 ck3 … ckm

C1 C2 …. Ck

Fig. 4. Representation of a candidate solution for k clusters and m features.

Table 7
Test dataset descriptions.

S.no. Dataset name # of features # of classes # of instances (size of each class)

1 Iris 4 3 150 (50,50,50)

2 Wine 13 3 178 (59,71,48)
Glass 9 6 214 (70,17,76,13,9,29)
3 Cancer 9 2 683 (444,239)
4 CMC 10 3 1473 (629,333,511)
5 Vowel 3 6 871 (72, 89, 172, 151, 207, 180)
6 Liver disorders (LD) 6 2 345 (145,200)

Iris, Wine, Glass, Wisconsin Breast Cancer (WBC), Contraceptive
Method Choice (CMC), Vowel and Liver Disorder (LD), are collected
from Machine Learning Laboratory [33]. The datasets used in this
study are described in Table 7.
In order to evaluate the performance and accuracy of the cluster-
ing result, total intra-cluster distances as defined in Eq. (12) criteria
is used. The low value of the sum is, the higher the quality of the
clustering is.
With the aim of compare the performance of our proposed algo-
rithm, several heuristic methods in the literature are used such as
K-means, K-means++ [55], GA [49], SA [48], TS [52], ACO [51], HBMO
[54], PSO [53], whose results are directly taken from Refs. [56,57]
and is given in Table 8. Table 8 lists the best, worst, average and
standard deviation of solutions and ranks the algorithms based on
the mean value for all datasets in Table 7. As compared algorithm
results are directly taken from Ref. [56], KH and IKH algorithms
are executed 100 times independently with same parameters as
described in section 5 except that the number of krill individuals
are set to 25 and maximum number of generations is 400 (Max.
NFE = 10,000).
The experimental results given in Table 8 show that proposed
algorithm obtains near optimal solutions in compare to those
of other methods. Our proposed algorithm achieves much better
results for almost all datasets with small standard deviation. For
iris dataset, KH and IKH algorithms converge to 96.6555 for each
run.
For wine dataset, IKH obtains better solution for worst, mean
and standard deviation than K-means, K-means++, GA, SA, TS, ACO,
HBMO, PSO, and KH. KH obtains best value of 16292.19, but it fails
to reach that value in all runs and thus the KH algorithm gets sixth
rank among all others. As for glass data set, KH and IKH achieve
best optimum value of 210.242 and 210.252, respectively which
needs 10,000 fitness function evaluations. The worst value of IKH
is 215.9355, while the worst values over 100 runs for glass dataset
for the algorithms K-means, K-means++, GA, SA, TS, ACO, HBMO,
PSO, KH are 227.35, 223.71, 286.77, 287.18, 286.47, 280.08, 249.54,
283.52, 251.2749 respectively. Thus IKH algorithm reaches near
best value in all runs.
The best solution for Cancer dataset obtained by IKH and KH
algorithm is 2964.387. In some cases KH gets stuck in local min-
imum and thus it does not obtain optimum mean value when
compared to IKH. For CMC dataset, MKH-H achieves best, worst
and mean solutions of 5693.72, 5693.779 and 5693.735 with a
standard deviation of 0.007975, while K-means, K-means++, GA,
Table 8
Comparison of objective function values for different datasets with other methods.

Dataset K-means K-means++ GA SA TS ACO HBMO PSO KH IKH

Iris Best 97.3259 97.3259 113.98650 97.4573 97.36597 97.10077 96.752047 96.8942 96.6555 96.6555
Worst 123.9695 122.2789 139.778272 102.0100 98.569485 97.808466 97.757625 97.8973 96.6555 96.6555
Mean 106.5766 98.5817 125.197025 99.9570 97.868008 97.171546 96.95316 97.2328 96.6555 96.6555
Std 12.938 5.578 14.563 2.01 0.53 0.367 0.531 0.347168 1.9E − 06 9.8E − 06
Rank 9 7 10 8 6 4 3 5 1 1

R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245

Wine Best 16,555.68 16,555.68 16,530.5338 16,473.4 16,666.23 16,530.533 16,357.2844 16,345.97 16,292.19 16,292.21
Worst 18,294.85 18,294.85 16,530.5338 18,083.25 16,837.54 16,530.534 16,357.2844 16,562.32 18,293.6 16,294.305
Mean 17,251.35 16,816.55 16,530.5338 17,521.09 16,785.45 16,530.53 16,357.2843 16,417.47 16,579.66 16,292.847
Std 874.148 637.140 0 753.084 52.073 0 0 85.4974 424.9147 0.7067427
Rank 9 8 5 10 7 4 2 3 6 1

Glass Best 215.73 15.36 282.32 275.16 283.79 273.46 247.71 270.57 210.2421 210.252
Worst 227.35 223.71 286.77 287.18 286.47 280.08 249.54 283.52 251.2749 222.8008
Mean 218.70 217.56 278.37 282.19 279.87 269.72 245.73 275.71 215.7225 215.9355
Std 2.456 2.455 4.138712 4.238 4.192734 3.584829 2.438120 4.557134 5.44876 2.737919
Rank 4 3 8 10 9 6 5 7 1 2

Cancer Best 2988.43 2986.96 3249.46 2993.45 3251.37 3046.06 3112.42 2973.50 2964.387 2964.387
Worst 2999.19 2988.43 3427.43 3421.95 3434.16 3242.01 3210.78 3318.88 3580.312 2964.393
Mean 2988.99 2987.99 2999.32 3239.17 2982.84 2970.49 2989.94 3050.04 2971.977 2964.389
Std 2.469 0.689 229.734 230.192 232.217 90.50028 103.471 110.8013 62.26148 0.001258
Rank 6 5 8 10 4 2 7 9 3 1

CMC Best 5703.20 5703.20 5756.5984 5849.03 5993.594 5819.1347 5713.9800 5700.985 5693.72 5693.72
Worst 5704.57 5705.37 5812.6480 5966.94 5999.805 5912.4300 5725.3500 5923.249 6755.956 5693.779
Mean 5705.37 5704.19 5705,6301 5893.48 5885.062 5701.9230 5699.2670 5820.965 5737.234 5693.735
Std 1.033 0.955 50.3694 50.867 40.84568 45.63470 12.690000 46.95969 178.0245 0.007975
Rank 5 4 10 9 8 3 2 7 6 1

Vowel Best 149,398.66 149,394.56 159,153.498 149,370.47 162,108.5381 159,458.1438 161,431.0431 148,976.0152 148,967.246 148,967.247
Worst 162,455.69 161,845.54 165,991.6520 165,986.42 165,996.4280 165,939.8260 165,804.671 149,121.1834 158,503.045 158,600.525
Mean 151,987.98 151,445.29 149,513.735 161,566.28 149,468.268 149,395.602 149,201.632 148,999.8251 150,035.986 150,172.425
Std 3425.250 3119.751 3105.5445 2847.085 2846.23516 3485.3816 2746.0416 28.8134692 1707.84248 1732.45161
Rank 9 8 5 10 4 3 2 1 6 7

Mean rank 7 5.83 7.67 9.5 6.33 3. 67 3.5 5.33 3.83 2.167
Final rank 8 6 9 10 7 3 2 5 4 1

243
244 R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245
SA, TS, ACO, HBMO, PSO, fail to obtain the best solutions. Similar
to Cancer dataset, KH algorithm achieves best solution of 5693.72
for CMC dataset, but fails to reach the value in all runs. For vowel
dataset, PSO performs well than all other algorithms including IKH.
Nevertheless, KH and IKH achieve best solution of 148967.24. As
a conclusion, our proposed scheme reaches better optimal solu-
tions with a small standard deviation in limited number iterations.
Based on the ranking analysis, IKH gets first rank among all others.
The time consumptions of the algorithms are nearly the same for
solving each data clustering problems.
7. Conclusion
Krill herd (KH) is a new optimization method for solving many
complex global optimization problems. In this paper, we pre-
sented an improved krill algorithm to solve function optimization
problems. The original krill herd saturated quickly and hence
trapped in local minimum. To alleviate the shortcomings of krill
herd, improved krill herd was proposed by introducing global
exploration operator. Using these modifications, IKH algorithm
converges to optimal solutions quickly. The simulation results show
that our proposed method is fast and efficient for solving function
optimization problems. The proposed algorithm IKH is then applied
to data clustering problem. The experimental results are compared
with other methods in the literature and indeed it reveals that the
proposed algorithm is suitable for solving clustering problems.
In the future study, we aim to do the following
(1) The proposed method is applied to solve other practical engi-
neering problems, such as scheduling, path planning, text
documents clustering and constrained optimization.
(2) The performance of KH can be improved by hybridizing KH with
other optimization strategies.
References
[1] A.H. Gandomi, A.H. Alavi, Krill herd: a new bio-inspired optimization
algorithm, Commun. Nonlinear Sci. Numer. Simul. 17 (12) (2012) 4831–4845,
http://dx.doi.org/10.1016/j.cnsns.2012.05.010.
[2] Hüseyin Haklı, Harun Uguz, A novel particle swarm optimization algorithm
with Levy flight, Appl. Soft Comput. 23 (2014) 333–345.
[3] R.C. Eberhart, Y. Shi, J. Kennedy, Swarm Intelligence, Morgan Kaufmann, 2001.
[4] X.-S. Yang, Engineering Optimization an Introduction with Metaheuristic
Applications, first ed., John Wiley & Sons, New Jersey, 2010.
[5] X.S. Yang, Nature-Inspired Metaheuristic Algorithms, second ed., Luniver
Press, 2010.
[6] D.M. Van, A.P. Engelbrecht, Data clustering using particle swarm
optimization, in: Proceedings IEEE Congress on Evolutionary Computation,
Canbella, Australia, 2003, pp. 215–220.
[7] X. Cui, T. Potok, P. Palathingal, Document clustering using particle swarm
optimization, Proc. of IEEE Swarm Intelligence Symposium, IEEE Press (2005).
[8] A. Sarangi, R.K. Mahapatra, S.P. Panigrahi, DEPSO and PSO-QI in digital filter
design, Expert Syst. Appl. 38 (2011) 10966–10973.
[9] Y.-P. Chang, C.-N. Ko, A PSO method with nonlinear time-varying evolution
based on neural network for design of optimal harmonic filters, Expert Syst.
Appl. 36 (2009) 6809–6816.
[10] H.-C. Yang, S.-B. Zhang, K.-Z. Deng, P.-J. Du, Research into a feature selection
method for hyper spectral imagery using PSO and SVM, J. China Univ. Min.
Technol. 17 (2007) 473–478.
[11] D. Karaboga, B. Basturk, A powerful and efficient algorithm for numerical
function optimization: artificial bee colony (ABC) algorithm, J. Glob. Optim. 39
(3) (2007) 459–471, http://dx.doi.org/10.1007/s10898-007-9149-x.
[12] M. Dorigo, T. Stutzle, Ant Colony Optimization, MIT Press, Cambridge, 2004.
[13] R. Storn, K. Price, Differential evolution—a simple and efficient heuristic for
global optimization over continuous spaces, J. Glob. Optim. 11 (4) (1997)
341–359.
[14] X. Li, M. Yin, Application of differential evolution algorithm on self-potential
data, PLoS One 7 (12) (2012) http://dx.doi.org/10.1371/journal.pone.0051199.
[15] D. Simon, Biogeography-based optimization, IEEE Trans. Evol. Comput. 12 (6)
(2008) 702–713.
[16] H. Beyer, The Theory of Evolution Strategies, Springer, New York, 2001.
[17] D.E. Goldberg, Genetic Algorithms in Search. Optimization and Machine
Learning, Addison-Wesley, New York, 1998.
[18] B. Shumeet, Population-Based Incremental Learning: A Method for
Integrating Genetic Search Based Function Optimization and Competitive
Learning, Carnegie Mellon University, Pittsburgh, 1994.
[19] J. Kennedy, R. Eberhart, Particle swarm optimization, in: Proceeding of the
IEEE International Conference on Neural Networks, Perth, Australia, 1995, pp.
1942–1948.
[20] W. Khatib, P. Fleming, The stud GA: a mini revolution? in: A. Eiben, T. Back, M.
Schoenauer, H. Schwefel (Eds.), Proceeding of the 5th International
Conference on Parallel Problem Solving from Nature, Springer-Verlag,
London, 1998, pp. 683–691.
[21] X.S. Yang, S. Deb, Cuckoo search via Levy flights, in: World Congress on Nature
& Biologically Inspired Computing, IEEE Publication, USA, 2009, pp. 210–214.
[22] X.S. Yang, Firefly algorithm, Levy flights and global optimization, in: M.
Bramer, R. Ellis, M. Petridis (Eds.), Research and Development in Intelligent
Systems, XXVI, Springer, London, 2010, pp. 209–218.
[23] L.-Y. Chuang, H.-W. Chang, C.-J. Tu, C.-H. Yang, Improved binary PSO for
feature selection using gene expression data Comput. Biol. Chem. 32 (2008)
29–38.
[24] Y. Zhang, D. Huang, M. Ji, F. Xie, Image segmentation using PSO and PCM with
Mahalanobis distance, Expert Syst. Appl. 38 (2011) 9036–9040.
[25] J.-P. Yang, C.-K. Kung, F.-T. Liu, Y.-J. Chen, C.-Y. Chang, R.-C. Hwang, Logic
circuit design by neural network and PSO algorithm, in: 2010 First
International Conference on Pervasive Computing Signal Processing and
Applications (PCSPA), Harbin, China, 2010, pp. 456–459.
[26] R. Jensi, G. Wiselin Jiji, Hybrid data clustering approach using k-means and
flower pollination algorithm, Adv. Comput. Intell. 2 (2) (2015).
[27] R. Jensi, G. Wiselin Jiji, MBA-LF: a new data clustering method using modified
bat algorithm and Levy flight, ICTACT J. Soft Comput. 6 (1) (2015) 1093–1101.
[28] Gai-Ge Wang, Amir H. Gandomi, Amir H. Alavi, Stud krill herd algorithm,
Neurocomputing 128 (2014) 363–370.
[29] Gai-Ge Wang, Amir H. Gandomi, Amir H. Alavi, An effective krill herd
algorithm with migration operator in biogeography-based optimization, Appl.
Math. Model. 38 (2014) 2454–2462.
[30] Gai-Ge Wang, Amir H. Gandomi, Amir H. Alavi, Guo-Sheng Hao, Hybrid krill
herd algorithm with differential evolution for global numerical optimization,
Neural Comput. Appl. 25 (2014) 297–308.
[31] Gaige Wang, Lihong Guo, Amir Hossein Gandomi, Lihua Cao, Amir Hossein
Alavi, Hong Duan, Jiang Li, Lévy-flight krill herd algorithm, Math. Prob. Eng.
2013 (2013), Article ID 682073, http://dx.doi.org/10.1155/2013/682073, 14
pages.
[32] R. Jensi, G. Wiselin Jiji, A survey on optimization approaches to text document
clustering, Int. J. Comput. Sci. Appl. 3 (6) (2013) 31–44.
[33] C.L. Blake, C.J. Merz, University of California at Irvine Repository of Machine
Learning Databases, 1998 http://www.ics.uci.edu/mlearn/MLRepository.html.
[34] http://www.mathworks.com/matlabcentral/fileexchange/55486-krill-herd-
algorithm.
[35] R. Jensi, G. Wiselin Jiji, An enhanced particle swarm optimization with levy
flight for global optimization, Appl. Soft Comput. 43 (c) (2016), http://dx.doi.
org/10.1016/j.asoc.2016.02.018.
[36] X. Yao, Y. Liu, G. Lin, Evolutionary programming made faster, IEEE Trans.
Evolut. Comput. 3 (1999) 82–102.
[37] G. Wang, L. Guo, H. Wang, H. Duan, L. Liu, J. Li, Incorporating mutation scheme
into krill herd algorithm for global numerical optimization, Neural Comput.
Appl. 24 (3) (2014) 853–871, http://dx.doi.org/10.1007/s00521-012-1304-8.
[38] Lihong Guo, Gai-Ge Wang, Amir H. Gandomi, Amir H. Alavi, Hong Duan, A
new improved krill herd algorithm for global numerical optimization,
Neurocomputing 138 (2014) 392–402.
[39] Gai-Ge Wang, Amir H. Gandomi, Amir H. Alavi, Yong-Quan Dong, A hybrid
meta-heuristic method based on firefly algorithm and krill herd, in: A Hybrid
Meta-Heuristic Method Based on Firefly Algorithm and Krill Herd, IGI Global,
2015, pp. 521–540, http://dx.doi.org/10.4018/978-1-4666-9479-8.ch019.
[40] Dervis Karaboga, Bahriye Akay, A comparative study of artificial bee colony
algorithm, Appl. Math. Comput. 214 (2009) 108–132.
[41] Momin Jamil, Xin-She Yang, A literature survey of benchmark functions for
global optimization problems, Int. J. Math. Model. Numer. Optim. 4 (2) (2013)
150–194.
[42] Ali R. Alroomi, The Farm of Unconstrained Benchmark Functions, University
of Bahrain, Electrical and Electronics Department, Bahrain (October) 2013,
[Online] Available: http://www.al-roomi.org/cv/publications.
[43] Junpeng Li, Yinggan Tang, Changchun Hua, Xinping Guan, An improved krill
herd algorithm: krill herd with linear decreasing step, Appl. Math. Comput.
234 (2014) 356–367.
[44] Liangliang Li, Yongquan Zhou, Jian Xie, A free search krill herd algorithm for
functions optimization, Math. Prob. Eng. 2014 (2014), http://dx.doi.org/10.
1155/2014/936374.
[45] G.-G. Wang, A.H. Gandomi, X.-S. Yang, A.H. Alavi, A new hybrid method based
on krill herd and cuckoo search for global optimisation tasks, Int. J. Bio
Inspired Comput. (in press).
[46] Gai-Ge Wang, Lihong Guo, Amir Hossein Gandomi, Amir Hossein Alavi, Hong
Duan, Simulated Annealing-Based Krill Herd Algorithm for Global
Optimization, 2013 (2013), http://dx.doi.org/10.1155/2013/213853.
[47] Jiawei han, Michelin Kamber, Data Mining Concepts and Techniques, Elsevier,
2010.
[48] S.Z. Selim, K.S. Al-Sultan, A simulated annealing algorithm for the clustering
problem, Pattern Recognit. 24 (10) (1991) 1003–1008.
 R. Jensi, G.W. Jiji / Applied Soft Computing 46 (2016) 230–245
[49] Ujjwal Maulik, Sanghamitra Bandyopadhyay, Genetic algorithm-based
clustering technique, Pattern Recognit. 33 (2000) 1455–1465.
[50] C. Sung, H. Jin, A tabu-search-based heuristic for clustering, Pattern Recognit.
33 (2000) 849–858.
[51] P.S. Shelokar, V.K. Jayaraman, B.D. Kulkarni, An ant colony approach for
clustering, Anal. Chim. Acta 509 (2) (2004) 187–195.
[52] Y. Liu, Z. Yi, H. Wu, M. Ye, K. Chen, A tabu search approach for the minimum
sum-of-squares clustering problem, Inf. Sci. 178 (2008) 2680–2704.
[53] Yi-Tung Kao, Erwie Zahara, I-Wei Kao, A hybridized approach to data
clustering, Expert Syst. Appl. 34 (3) (2008) 1754–1762.
[54] M. Fathian, B. Amiri, A honey-bee mating approach on clustering, Int. J. Adv.
Manuf. Technol. 38 (2008) 809–821.
245
[55] D. Arthur, S. Vassilvitskii, K-means++: the advantages of careful seeding, in:
Proceedings of the eighteenth annual ACM-SIAM symposium on discrete
algorithms SODA ’07, Philadelphia, PA, USA: Society for Industrial and Applied
Mathematics, 2007, pp. 1027–1035.
[56] Taher Niknam, Babak Amiri, An efficient hybrid approach based on PSO, ACO
and k-means for cluster analysis, Appl. Soft Comput. 10 (2010) 183–197.
[57] Ganesh Krishnasamy, Anand J. Kulkarni, Raveendran Paramesran, A hybrid
approach for data clustering based on modified cohort intelligence and
K-means, Expert Syst. Appl. 41 (2014) 6009–6016.