Hindawi Publishing Corporation

e Scientific World Journal

Volume 2014, Article ID 237102, 11 pages
http://dx.doi.org/10.1155/2014/237102

Research Article
Cloud Model Bat Algorithm

Yongquan Zhou, Jian Xie, Liangliang Li, and Mingzhi Ma

College of Information Science and Engineering, Guangxi University for Nationalities, Nanning, Guangxi 530006, China

Correspondence should be addressed to Yongquan Zhou; [email protected]

Received 18 March 2014; Accepted 22 April 2014; Published 19 May 2014

Academic Editor: Xin-She Yang

Copyright © 2014 Yongquan Zhou et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Bat algorithm (BA) is a novel stochastic global optimization algorithm. Cloud model is an effective tool in transforming between
qualitative concepts and their quantitative representation. Based on the bat echolocation mechanism and excellent characteristics
of cloud model on uncertainty knowledge representation, a new cloud model bat algorithm (CBA) is proposed. This paper focuses
on remodeling echolocation model based on living and preying characteristics of bats, utilizing the transformation theory of cloud
model to depict the qualitative concept: “bats approach their prey.” Furthermore, Lévy flight mode and population information
communication mechanism of bats are introduced to balance the advantage between exploration and exploitation. The simulation
results show that the cloud model bat algorithm has good performance on functions optimization.

1. Introduction term of qualitative concept and its numerical representation

[10]. In recent years, the cloud model is applied in the field of
Metaheuristics is a new method for stochastic optimization; metaheuristics, such as cloud model based genetic algorithm
in recent years, more and more different metaheuristic (CGA) [11] and cloud model based evolutionary algorithm
algorithms have been proposed, such as particle swarm (CBEA) [12, 13]. In this paper, the bat algorithm was used
optimization (PSO) [1], differential evolution (DE) [2], and for reference, the echolocation mechanism based on cloud
bat algorithm (BA) [3], and some novel metaheuristic algo- model was remodeled, and two mechanisms were introduced:
rithms are proposed. The bat algorithm was proposed by population information communicating of each individual
Xin-She Yang in 2010, which is inspired by the echolocation and random Lévy flight; a cloud model bat algorithm (CBA)
behaviour of microbats. The bat algorithm controls the size was proposed, and the purpose is to improve the convergence
and orientation of bats moving speed through adjusting the rate and precision of bat algorithm. At the end of this paper,
frequency of each bat and then moves to a new location; the combination strategies and parameter settings of CBA are
intensive local search is controlled by the loudness and pulse discussed, several appropriate parameters are selected, and
emission rate. To some extent, PSO is a special case of suitably eight typical benchmark functions are tested, and the test
simplified BA. Due to the fact that BA combines with the results show that the proposed algorithm is feasible and
advantages of swarm intelligence, which utilizes a balanced effective.
combination of the advantages of the standard PSO and the
intensive local search controlled by the loudness and pulse
rate, BA is widely researched in different field applications. 2. Behaviors of Bats and Cloud Model
BA has some advantages over other algorithms, and the 2.1. Flight and Echolocation of Bats. Bats are the only volitant
number of adjustable parameters is fewer. Consequently, BA mammals in the world; after tens of millions of years of
has been used for solving engineering design optimization evolution, there are nearly 1,000 species of bats. Bats have
[4–6], classifications [7], fuzzy cluster [8], prediction [9], powered flight ability, which is much more complex than
neural networks, and other applications. glide; their flight can generate complex aerodynamic tracks,
The cloud model is proposed by Li et al. in 1995, which and the flight is accompanied with local self-similarity [14].
is a model of the uncertain transition between a linguistic Many microbats have amazing echolocation; these bats can
2 The Scientific World Journal

𝑇
Objective function 𝑓(𝑥), 𝑥 = [𝑥1 , 𝑥2 , . . . , 𝑥𝑑 ]
Initialize the bat population 𝑥𝑖 (𝑖 = 1, 2, . . . , 𝑛) and V𝑖
Define pulse frequency 𝑓𝑖 at 𝑥𝑖 Initialize pulse rates 𝑟𝑖 and the loudness 𝐴 𝑖
While (𝑡 < Max number of iterations)
Generate new solutions by adjusting frequency,
and updating velocities and locations/solutions [(1)]
if (rand > 𝑟𝑖 )
Select a solution among the best solutions
Generate a local solution around the selected best solution
end if
Generate a new solution by flying randomly
if (rand < 𝐴 𝑖 & 𝑓 (𝑥𝑖 ) < 𝑓 (𝑥∗ ))
Accept the new solutions
Increase 𝑟𝑖 and reduce 𝐴 𝑖
end if
Rank the bats and find the current best 𝑥∗
end while
Postprocess results and visualization.

Algorithm 1: Pseudocode of the bat algorithm (BA) [3].

emit a very loud and short sound pulse and receive the For the local search part, once a solution is selected
echo that reflects back from the surrounding objects by their among the current best solutions, a new solution for each bat
extraordinary big auricle. Then, they analyze this feedback is generated locally using random walk:
information of echo in their subtle brain. They not only can
𝑥new = 𝑥old + 𝜀𝐴 𝑡 , (2)
discriminate direction for their own flight pathway according
to the echo but also can distinguish different insects and where 𝜀 ∈ [−1, 1] is a random number, while 𝐴 𝑡 = ⟨𝐴𝑡𝑖 ⟩ is the
obstacles, to hunt prey and avoid collision effectively in the average loudness of all the bats at this time step.
day or night. Bats minimize the conspicuousness of their Furthermore, the loudness 𝐴 𝑖 and the rate 𝑟𝑖 of pulse
echolocation call to potential insect prey by reducing call emission have to be updated accordingly as the iterations
intensity and by changing the frequencies in the call [15]. proceed. These formulas are
Furthermore, the echolocation signal that one individual bat
uses to collect information can simultaneously serve as a 𝐴𝑡+1
𝑖 = 𝛼𝐴𝑡𝑖 ,
communication function, allowing, for example, group mem- (3)
bers to remain in contact with one another. Echolocation call 𝑟𝑖𝑡+1 = 𝑟𝑖0 [1 − exp (−𝛾𝑡)] ,
plays a crucial and hitherto underestimated role for social where 𝛼 and 𝛾 are constants.
communication in a highly mobile and gregarious nocturnal Based on these approximations and idealization, the basic
mammal and thus facilitates social communication in bats steps of the bat algorithm [3] can be summarized as the
population [16]. pseudocode shown in Algorithm 1.

2.3. Lévy Flight. Lévy flight is a random walk in which the

2.2. Bat Algorithm. In simulations, they use virtual bats
step-lengths have a probability distribution that is heavy-
naturally, to define the updated rules of their positions 𝑥𝑖
tailed. Lévy flight has several properties: “heavy tails,” sta-
and velocities V𝑖 in a 𝐷-dimensional search space. The new
tistical self-similarity, random fractal characteristics, and
solutions 𝑥𝑖𝑡 and velocities V𝑖𝑡 at time step 𝑡 are given by
infinite variance with an infinite mean value [17]. Lévy
distribution, Gaussian distribution, and Cauchy distribution,
which is a 𝛼 stable distribution; however, probability density
𝑓𝑖 = 𝑓min + (𝑓max − 𝑓min ) 𝛽, function (PDF) curves of Gaussian distribution and the
Cauchy distribution are symmetrical; Lévy distribution is
V𝑖𝑡 = V𝑖𝑡−1 + (𝑥𝑖𝑡 − 𝑥∗ ) 𝑓𝑖 , (1)
not symmetrical [18]. Probability density function of Lévy
distribution on 𝑥 > 𝜇 is
𝑥𝑖𝑡 = 𝑥𝑖𝑡−1 + V𝑖𝑡 ,
𝐶 𝑒−𝐶/2(𝑥−𝜇)
Lévy ∼ 𝑓 (𝑥) = √ , (4)
2𝜋 (𝑥 − 𝜇)3/2
where 𝛽 ∈ [0, 1] is a random vector drawn from a uniform
distribution. Here, 𝑥∗ is the current global best location where 𝜇 is the location parameter and 𝐶 is the scale param-
(solution) which is located after comparing all the solutions eter. PDF curve of the three distributions is presented in
among all the 𝑛 bats. Figure 1.
The Scientific World Journal 3

0.5 the bandwidth of the mathematical expected curve (MEC)

0.45 of the normal cloud showing how many elements in the
0.4 universe of discourse could be accepted to the linguistic
0.35
term, the greater 𝐸𝑛, and the broader coverage. It can be
0.3
Y 0.25 also considered as a measure of fuzziness of the concept,
0.2 representing the scope of the universe of discourse that can be
0.15 accepted by the concept. The hyper entropy 𝐻𝑒 is the entropy
0.1 of the entropy 𝐸𝑛. It is a measure of dispersion of the cloud
0.05 drops; it can be used a measure of thickness of the cloud,
0 which not only reflects the randomness of samples appearing
−5 −4 −3 −2 −1 0 1 2 3 4 5
X that represent qualitative concepts value but also reveals the
relatedness between fuzziness and randomness.
Cauchy Levy-0.75 Normal cloud model makes full use of the universality of
Gaussian Levy-1.5 the normal distribution and normal membership function,
Levy-0.5
which not only broaden the formation conditions of the nor-
Figure 1: The PDF curve of the three distributions. mal distribution but also make the normal membership func-
tion be the expectation of the random membership degree;
the randomness and fuzziness are represented uniformly by
Studies have shown that flight behaviour of many animals entropy and then the theoretical basis of universality of the
and insects has demonstrated the typical characteristics of normal cloud model is established [24]. Cloud model has the
Lévy lights. A recent study by Reynolds and Frye shows that 3𝜎 characteristics; there are 99.7% drops of cloud located in
fruit flies explore their landscape using a series of straight [𝐸𝑥 − 3𝐸𝑛, 𝐸𝑥 + 3𝐸𝑛]. These drops of cloud are generated by
flight paths punctuated by a sudden 90∘ turn, leading to a the normal cloud generator. Atomized feature of the cloud
Lévy flight-style intermittent scale free search pattern [19]. model: the drops of cloud spread around while the hyper
Studies on human behaviour such as the Ju/’hoansi hunter- entropy is increasing, but many drops still stand in the central
gatherer foraging patterns also show the typical feature of area of the cloud, which can be used to adjust the strategies
Lévy flights [20]. Subsequently, due to the remarkable proper- of the evolution and help to escaping from local optima [11].
ties of stable Lévy distribution, Lévy flight has been applied to The clouds with different digital characteristics are depicted
optimization and optimal search [21], and preliminary results in Figure 2.
show its promising capability.
3. Cloud Model Bat Algorithm
2.4. Cloud Model. Cloud model build a transformational
bridge between a linguistic term of qualitative concept Bats prey by emitting pulse with a certain frequency and
and quantitative representation, which reflects randomness, detection of the echo; they communicate with each other
fuzziness, and the relationship between randomness and using echolocation call. This paper assimilates its principle
fuzziness of uncertainty in knowledge representation [22, 23]. to idealize some of the echolocation characteristics of micro-
The cloud and cloud droplets are defined as follows. bats. Based on the excellent characteristics of cloud model on
Let 𝑈 be the set 𝑈 = {𝑥}, as the universe of discourse, and uncertainty knowledge representation, a bat algorithm based
let 𝐶 be a linguistic term associated with 𝑈. The membership on cloud model was proposed (cloud model bat algorithm,
degree of 𝑥 in 𝑈 to the linguistic term 𝐶, 𝜇(𝑥), is a random CBA).
number with a stable tendency. 𝜇(𝑥) takes the values in [0, 1].
A membership cloud, or compatibility cloud, is a mapping
from the universe of discourse 𝑈 to the unit interval [0, 1]. 3.1. Knowledge Representation of Bat Cloud. In order to depict
That is, the CBA, the habits of bats are used for reference, taking
advantage of the excellent properties of cloud model. Firstly,
𝜇 (𝑥) : 𝑈 󳨀→ [0, 1] , representation of relevant knowledge needs to be described;
(5) several concepts certain about CBA were given as follows.
∀𝑥 ∈ 𝑈, 𝑥 󳨀→ 𝜇 (𝑥) .
(1) Optimizing Generation. Optimizing generation indicates
The distribution of 𝑥 in universe of discourse 𝑈 is called cloud the number of iteration circles in the algorithm; each iteration
and each 𝑥 is called a drop of cloud [22]. circle may include several times replacement of population,
A normal cloud is defined with three digital characteris- simply, namely, 𝑡.
tics, expected value 𝐸𝑥, entropy 𝐸𝑛, and hyper entropy 𝐻𝑒
and a cloud, namely, 𝐶(𝐸𝑥, 𝐸𝑛, 𝐻𝑒). Expectation 𝐸𝑥 is the (2) Individual. In CBA, each bat is treated as an individual;
position at 𝑈 corresponding to the center of gravity of the when it is in flight, the position of each bat 𝑥𝑖𝑡 signifies a
cloud. In other words, the element 𝐸𝑥 in the universe of candidate solution of optimization problem, where 𝑖 is the
discourse is fully compatible with the linguistic term. The number of individuals and 𝑡 is the optimizing generation.
entropy 𝐸𝑛 is a measure of the coverage of the concept within For the high-dimensional optimization, 𝑥𝑖𝑡 represent a vector
the universe of discourse. In other words, 𝐸𝑛 is defined by under high-dimensional space, correspondingly, where each
4 The Scientific World Journal

0.8

Membership degree
0.6
Membership degree

1
0.8 0.4
0.6
0.4 He 0.2
0.2
3En
0 0
−8 −6 −4 −2 0 2 4 6 8 −4 −2 0 2 4
C(0, 2, 0.2) C(0, 1, 0.2)
(a) (b)
1

0.8
Membership degree

0.6

0.4

0.2

0
−10 −5 0 5 10
C(0, 1, 0.8)
(c)

Figure 2: The cloud with different digital characteristics. 𝐶(0, 2, 0.2) denote a cloud, 0 is expectation, 2 is entropy, and 0.2 is hyper entropy.

dimension denotes an attribute of the solution of optimiza- 𝑥𝑔best ; the population elite 𝑥𝑔best will be saved and be used in
tion problem. swarm information communication.

(3) Fitness Function. Fitness function denotes adaptation

degree of each individual aiming at their located environment 3.2. Cloud Model Bat Search Algorithm. In this paper, at the
in the community. It is used to evaluate the individual and basis of original BA and the habits of bats, based on the
decides which individual to retain or eliminate. Fitness func- cloud model and Lévy flights, cloud model bat algorithm is
tion usually is the expression of costs, profitability, variance, proposed under idealized simulation of echolocation of bats.
and so on. For simplicity, some idealized rulesareas follows.

(4) Population Bats Cloud. The bats live and prey together, and (1) Using the echolocation, bats not only can identify
many bats constitute a community. A cloud was generated, the direction, measure the distance, and determine
which is to represent the distribution characteristics of the the current status of their prey but also can avoid
same dimension of all individuals, called population bats collision, distinguish obstacles, and prey from back-
cloud, namely, 𝑃𝑏𝑐𝑗 (𝐸𝑥, 𝐸𝑛, 𝐻𝑒), where 𝑗 represents the 𝑗th ground clutter. This paper only simulates that bats
dimension of population and 𝐸𝑥, 𝐸𝑛, and 𝐻𝑒 are three digital search for a prey using echolocation mechanism in
characteristics of cloud model. a search space under ideal environment, where the
position of prey means an optimal solution of the
(5) Individual Experience. It denotes these individuals that are problem; each position of bats indicates a candidate
able to remember their own history during the process of solution of optimization problem. Bats may not prey
optimization. In the proposed algorithm, bats can memorize their target, but they gradually approach the target,
their own best location 𝑥𝑝best during moving. Its main close to the prey, approximately regard as successful
purpose is to guide the flight of bat and to promote the preying under a certain tolerance.
communication among the population.
(2) Each bat flies randomly with frequency 𝑓𝑖𝑡 ; the posi-
(6) Population Elite. In this proposed algorithm, population tion 𝑥𝑖𝑡 moves under the adjustment of frequency 𝑓𝑖𝑡 .
elite denotes the position of the optimal individual, namely, The frequency 𝑓𝑖𝑡 resembles an adjustment coefficient
The Scientific World Journal 5

of step length, and frequency 𝑓𝑖𝑡 of sound pulse is and update the individual experience 𝑥𝑝best for each bat and
changeable, where 𝑡 is the optimizing generation. population elite 𝑥𝑔best in each step.

(3) The adjustment of frequency 𝑓 caused the change (7) Pulse Emission Rate and Loudness Update. The rate of
of the wavelength 𝜆 (this is because of the fact that pulse emission 𝑅𝑖 and loudness 𝐿𝑑𝑖 for each bat need to
𝜆𝑓 = V is a constant and V is the speed of sound in update when the achieved optimal solution after steps (3)–(5)
air; typically, V = 340 m/s); such wavelengths 𝜆 are in is better than the optimal solution of last generation.
the same order of their prey sizes and help to locate
the target. Generally, the frequency 𝑓 is in a range (8) Termination Judgment. 𝑡 = 𝑡 + 1; execute steps (3)–(7)
[𝑓min , 𝑓max ], and each individual can communicate until 𝑡 reaches a predefined maximum number of optimizing
information with others by echolocation call in a generation.
population. In this algorithm framework, three problems need to be
solved: first of all, the formation of bat cloud model, second
(4) Each bat emits sonic pulse with emission rate 𝑅𝑖 ∈ information communication of bat population, and third the
[0, 1] and loudness 𝐿𝑑𝑖 . At the beginning of prey, updating of 𝑅, 𝐿𝑑.
bats have a smaller 𝑅𝑖 and larger 𝐿𝑑𝑖 . During the
process of locating prey, the pulse emission rate 3.2.1. Formation of Population Bats Cloud Model. This paper
increases and loudness reduces once the bat searches simulates the moving of bats when several bats pursue and
for target traces, which is figuratively indicated by “bat capture prey. Each bat expects to move toward the direction
is approaching the target.” of prey (the optimal solution). In the search space, the entire
population is trying to approximate the optimal solution;
(5) In exploration of bats for prey, their flight features
the position 𝑥𝑖𝑡 of each individual should move toward the
are accompanied by typical Lévy flight characteristic;
optimum position. Consequently, the same dimensions of
many insects and animals have it as well. Exploration
population have stable tendency. However, each individual
and traces its of to detect potential prey traces of a
has their own feature, and the implementations of position
random flight.
updating are random for each individual. Therefore, the
characteristics of approximated process that bats have to
On the basis of the above mentioned idealized rules, approximate prey are simulated by cloud model. Sequentially,
the properties of the cloud model are utilized which rep- they adapt bat cloud model to depict the qualitative concept:
resents the membership degree of qualitative concept and “bats approach their prey.”
reveals the relationship between randomness and fuzziness Normal cloud model of bat approach process utilizes
in uncertainty knowledge representation. According to nor- the characteristics of cloud model that are the uncertain
mal cloud model generator with expectations, entropy, and transition between qualitative and quantitative. The pop-
hyper entropy, many drops of the cloud with quantitative ulations bats cloud 𝑃𝑏𝑐𝑗 (𝐸𝑥, 𝐸𝑛, 𝐻𝑒) analogize to cloud
transformation value corresponding to qualitative concept 𝐶(𝐸𝑥, 𝐸𝑛, 𝐻𝑒), where the expected value 𝐸𝑥 is the 𝑗th dimen-
are produced. In this paper, updating the position of bats sion of population elite 𝑥𝑔best , the entropy 𝐸𝑛 is the average
by cloud model, swarm information communication in each loudness of all bats, and the hyper entropy 𝐻𝑒 is the average
individual and random Lévy flights are introduced, a cloud pulse emission rate of all bats. The population bats cloud
model bat algorithm is proposed, and the steps of CMBA 𝑃𝑏𝑐𝑗 (𝐸𝑥, 𝐸𝑛, 𝐻𝑒) is a 1-dimensional normal cloud. In order to
algorithm can be summarized as follows: update the position of each individual, each dimension of new
individual is generated by randomly selecting several drops
(1) Initialization. Randomly initialize the position of each bat of the cloud from cloud cluster, and then calculate the result
in the population and relevant parameters. which is mean of the membership degree of each selected
drop multiplied by expected value 𝐸𝑥. The membership
(2) Initial Evaluation. Evaluate these initial positions using
degree of each drop is the certainty degree of approximation
fitness function, and find out a population elite 𝑥𝑔best .
expectation 𝐸𝑥. The computational formula is described as
(3) Bats Cloud Updating. Generate bats cloud based on cloud follows:
model, and update the position of bats.
𝑥𝑖𝑖𝑡+1 = AVG (∑ 𝐸𝑥 × 𝑅𝑆 (𝑝𝑏𝑐𝑖 (𝐸𝑥, 𝐸𝑛, 𝐻𝑒))) , (6)
(4) Swarm Information Communication. Information com-
munication of bat population adopts a differential operator where 𝑖 denotes 𝑖th individual, 𝑗 denotes 𝑗th dimen-
that is similar to mutation strategy “DE/best/2” in differential sion, 𝐸𝑥 = 𝑥𝑔best,𝑗 , 𝐸𝑛 = AVG(∑ Rate𝑖 ), 𝐻𝑒 =
algorithm. AVG(∑ Loudness𝑖 ), 𝑅𝑆(⋅) denotes a function of the ran-
domly selected several records, and AVG(⋅) denotes averag-
(5) Bats Random Lévy Flight. Each bat randomly flights using ing function.
Lévy flight. The pulse emission rate 𝑅𝑖 increases and loudness 𝐿𝑑𝑖
decreases while the iteration is increasing, and the entropy
(6) Population Evaluation. For each population in steps (3)– 𝐸𝑛 and hyper entropy 𝐻𝑒 therewith update. Consequently,
(5), evaluate each individual by fitness function and find out different cloud clusters are generated, so those individuals
6 The Scientific World Journal

gradually approach the target. Position of the bat is updated be adjusted suitably when it moves to a better position than
by population bats cloud 𝑃𝑏𝑐𝑗 (𝐸𝑥, 𝐸𝑛, 𝐻𝑒), which quantifica- last generation 𝑡 − 1. In this paper, the updating formulas
tionally represents the qualitative concept that bat approaches adopt (11). It is worth noting that the loudness and emission
𝑡
target. Sequentially, which reflect the determine tendency rates will be updated only if the final population elite 𝑥𝑔best in
of swarm optimizing, meanwhile show the fuzziness and current generation are better than the final population elite
randomness of uncertainty knowledge representation. 𝑡−1
𝑥𝑔best in last generation:
The proposed algorithm introduces information commu-
nication in bat population under idealized conditions, which 𝐿𝑑𝑖𝑡+1 = 𝛼𝐿𝑑𝑖𝑡 ,
assure that the entire bat colony gets helpful information by
communicating experience among individuals in a bat popu- 1 (11)
lation. This mechanism guides these bats that are approaching 𝑅𝑖𝑡+1 = /2))+𝑅1 )
,
1 + 𝑒(−(10/𝑡max )×(𝑡−(𝑡max 𝑖
prey fast.
Bats can emit sound pulses with certain frequency range. where 𝛼 ∈ [0, 1] is a constant, 𝑡 is optimizing generation, 𝑡max
Generally, the frequency 𝑓 is in a range [𝑓min , 𝑓max ] and the denotes maximum optimizing generation, and 𝑅𝑖1 denotes
typical range is [0, 1] in implementation. This paper defines initial pulse emission rate of each bat.
the frequency 𝑓 updating formula as follows: Cloud model bat algorithm is inspired by the behavior of
bat; original BA and cloud model are used for reference, based
𝑡
𝑓1𝑖𝑡 = ((𝑓1,min − 𝑓1,max ) + 𝑓1,max ) 𝛽1 , (7) on the properties of cloud model and echolocation of bat
𝑛𝑡 in foraging behavior, to remodel the algorithm framework,
defining several idealized rules and constructing optimizing
𝑡
𝑓2𝑖𝑡 = ((𝑓2,max − 𝑓2,min ) + 𝑓2,min ) 𝛽2 , (8) mechanism; a cloud model bat algorithm is proposed. The
𝑛𝑡 cloud model bat algorithm is different from the original bat
algorithm, which uses echolocation predation mechanism
where 𝛽1 , 𝛽2 ∈ [0, 1] is a random vector drawn from a of bats as the starting point and uses the universality of
uniform distribution, 𝑓1,max = 𝑓2,max = 𝑓max , 𝑓1,min = the normal cloud model as the basis. Several predominant
𝑓2,min = 𝑓min , and 𝑛𝑡 is a constant. The frequency 𝑓 would mechanisms are integrated organically in the CBA.
be analogous to an adjustable parameter. The step length For the performance of the proposed algorithm, the pop-
of individual moving is adjusted by adjusting frequency ulations bats cloud 𝑃𝑏𝑐𝑗 (𝐸𝑥, 𝐸𝑛, 𝐻𝑒) utilizes the information
𝑓. Meanwhile, it can be interpreted as bats adjusting their provided by the current optimal solution to generate the
own position by adjusting their own frequency and com- drops of cloud. The loudness and pulse emission rate are
municating with other bats. Information communication of regarded as entropy 𝐸𝑛 and hyper entropy 𝐻𝑒, respectively,
bat population adopts a differential operator that is similar to control the measure of the coverage and randomness
to mutation strategy “DE/best/2” in differential algorithm, of optimal solution structure. Reduction of the loudness
which is described as follows: and increasing of the pulse rate emission show that bats
𝑥𝑖𝑡+1 = 𝑥𝑔best
𝑡
+ 𝑓1𝑖𝑡 (𝑥𝑟1
𝑡 𝑡
− 𝑥𝑟2 ) + 𝑓2𝑖𝑡 (𝑥𝑟3
𝑡 𝑡
− 𝑥𝑟4 ), (9) approach their target. To quantificationally represent this
qualitative concept by population, bats cloud model makes
𝑡 many individual clusters around the current optimal solution
where 𝑥𝑔best represents the current population elite after
𝑡
and forms a bat cloud, thus exploring much better solutions.
updating by bat cloud updating and 𝑥𝑟𝑖 is 𝑖th individual This proposed algorithm has strong stability with bat cloud
randomly selected in the population after bat cloud updating. updating, which can gradually approach the optimal solution.
In addition, the above mentioned mechanism accelerates The swarm information communication guides the whole
the convergence rate, while the Lévy flight behavior is population moving toward the optimal solution. The increase
introduced to greatly ensure the swarm diversity against the or decrease of frequency 𝑓 controls the scale of the indi-
premature convergence. Random Lévy flights are manipu- vidual moving forward or backward. Each individual can
𝑡 𝑡
lated at the basis of individual experience 𝑥𝑝best , where 𝑥𝑝best communicate information with others and ultimately move
represents the current individual experience after swarm toward the common goal or direction. On the one hand,
information communication. The random Lévy flights are (7) implements on the basis of population elite 𝑥𝑖𝑡 , which
used to improve the individual ability to escape from the can accelerate the convergence speed of proposed algorithm;
local optima; simultaneously, this mechanism also assures the however, it may lead to premature convergence. On the other
intensification. The detailed description is as follows: hand, the mechanism also reflects the importance of Lévy
flight.
𝑥𝑖𝑡+1 = 𝑥𝑝best
𝑡
́
+ 𝜇 × sign [rand − 0.5] ⊕ Levy, (10) From (7), we know that premature convergence may take
place; from (8), Lévy flight is implemented on the individual
where 𝜇 is a random parameter drawn from a uniform experience of population. This randomness of Lévy flight
distribution, sign ⊕ means entry-wise multiplications, rand ∈ can ensure the diversity of the population against premature
[0, 1], and random step length Lévy obeys Lévy distribution. convergence. Lévy flight has a certain role in escaping from
local optima; meanwhile, based on individual experience of
3.2.2. Method of Pulse Emission Rate and Loudness Updating. population it can accelerate the convergence rate to some
The pulse emission rate 𝑅𝑖 and loudness 𝐿𝑑𝑖 of each bat will extent.
The Scientific World Journal 7

105

10−5
100

log (fitness)
log (fitness)

10−5
10−10
10−10

10−15
10−15

10−20
10−20 0 20 40 60 80 100 120 140 160 180 200
1 2 3 4 5 6 7 8 9
Population size, ps Optimizing generation

ps = 10 ps = 35
ps = 15 ps = 40
ps = 20 ps = 45
ps = 25 ps = 50
ps = 30
(a) Box-and-whisker after 50 independent runs with different population (b) Iterative curve after one independent run with different population
size 𝑝𝑠, where 1–9 in abscissa axis correspond to 𝑝𝑠 = 10, 15, . . . , 50, size 𝑝𝑠, where 1–9 in abscissa axis correspond to 𝑝𝑠 = 10, 15, . . . , 50,
respectively respectively

Figure 3: Box-and-whisker diagram and iterative curve about the impact of different population size 𝑝𝑠.

The range of loudness 𝐿𝑑𝑖 and pulse emission rate 𝑅𝑖 value is 0 at (1, 1, . . . , 1); the global minimum is inside
may have influence on the performance of the proposed a long, narrow, and parabolic shaped flat valley. To find
algorithm; meanwhile, the number of the drops of cloud is the valley is trivial. To converge to the global minimum,
no exception. In [13], for the drops of cloud are generated however, is difficult. Statistical result of minimum fitness
by the normal cloud generator, there are 99.7% located in the after 50 independent runs was represented by box-and-
interval [𝐸𝑥 − 3𝐸𝑛, 𝐸𝑥 + 3𝐸𝑛]. Consequently, initial value of whisker diagram, and the iterative curve was depicted for
𝐿𝑑𝑖 initializes 0.5(𝑥max −𝑥min )/3, where 𝑥max and 𝑥min denote once independent run.
the upper and lower limit of the search space. The range of the To initialize the parameter frequency 𝑓 ∈ [0, 2], 𝑛𝑡 =
drops of cloud that are located around the expectation 𝐸𝑥 will 4000, select different population size 𝑝𝑠 for experiment. The
reduce while the loudness 𝐿𝑑𝑖 reduces gradually. In addition, purpose is to investigate the influence of the population size
the thickness of cloud cluster will increase while the pulse for the proposed algorithm. The descriptive statistics of the
emission rate increases gradually; even the cloud cluster is results are plotted in Figure 3(a), where 1–9 in abscissa axis
excessively discrete and represents atomized feature. In [25], correspond to 𝑝𝑠 = 10, 15, . . . , 50, respectively. As shown
the expectation 𝐸𝑥 can be approximated by reverse clouds as Figure 3(a), the precision of the optimum value gradually
generator and the error is less than 0.01 if only the number of increases while the population size 𝑝𝑠 increases, and the
the drops of cloud is more than 10. Similarly, to approximate increment of precision gradually decreases. The precision
entropy 𝐸𝑛 and assure relative error less than 0.01, the number of the optimum value is low and the extreme outliers
of the drops of cloud is more than 100. Consequently, this will appear when population only includes 10 individuals,
paper produces 100 drops of cloud in the implementation, which show that the population size is insufficient and the
as a cloud cluster, and randomly selects a larger sample with exploring ability is poor. The precision of the optimal value
50 drops of cloud from the cloud cluster to fit the structure increases properly and the outliers are mild when population
individual. size increases to 20. After the population size reaches 30,
performance of the proposed algorithm gradually stabilizes,
4. Simulations and Result Analysis and the incremental extent of the precision is inapparent.
Figure 3(b) is the iterative curves for one independent
In order to validate the validity of bat cloud model algo- run of CBA. As shown in the figure, the difference of the
rithm, several unconstrained high-dimensional benchmark optimum value is not outstanding after the population size
test functions are selected (simulation platform: Microsoft increases to 30. Considering the precision and calculation,
Windows XP Professional SP3, AMD Athlon (tm) II X4 640 𝑝𝑠 = 45 is a preferable balance, which has high precision and
3.00 GHz, 4.00 GB; programming tools: Matlab R2012a). less calculation; in addition, the convergence rate is relatively
fast.
4.1. Parameter Settings and Analysis. In this section, in order After setting the parameter 𝑝𝑠 = 45, 𝑛𝑡 = 4000 to test the
to test the sensibility of parameter settings, the 2-dimensional proposed algorithm with different upper limit of frequency
Rosenbrock function was selected. And its global minimum 𝑓 ∈ [0, 𝐹]. The purpose is to investigate the impact of the
8 The Scientific World Journal

105

100
10−5

log (fitness)
10−5
log (fitness)

10−10

10−10
10−15

10−20 10−15

10−20 0 20 40 60 80 100 120 140 160 180 200

1 2 3 4 5 6
Range of frequency [0, F] Optimizing generation

F = 0.5 F=3
F=1 F=4
F=2 F=5

(a) Box-and-whisker after 50 independent runs with different upper limit (b) Iterative curve after one independent run with different upper limit
of frequency 𝑓 ∈ [0, 𝐹], where 1–6 in abscissa axis correspond to 𝐹 = of frequency 𝑓 ∈ [0, 𝐹], where 1–6 in abscissa axis correspond to 𝐹 =
0.5, 1, 2, . . . 5, respectively 0.5, 1, 2, . . . 5, respectively

Figure 4: Box-and-whisker diagram and iterative curve about the impact of different upper limit of frequency 𝑓 ∈ [0, 𝐹].

105
10−10

10−12 100
log (fitness)
log (fitness)

10−14 10−5

10−16
10−10
−18
10
10−15
10−20

10−22 10−20
1 2 3 4 5 6 7 8 0 20 40 60 80 100 120 140 160 180 200
Parameter, nt

nt = 1000 nt = 5000
nt = 2000 nt = 6000
nt = 3000 nt = 7000
nt = 4000 nt = 8000
(a) Box-and-whisker after 50 independent runs with different parameter (b) Iterative curve after one independent run with different parameter
𝑛𝑡, where 1–8 in abscissa axis correspond to 𝑛𝑡 = 1000, 2000, . . . , 8000, 𝑛𝑡, where 1–8 in abscissa axis correspond to 𝑛𝑡 = 1000, 2000, . . . , 8000,
respectively respectively

Figure 5: Box-and-whisker diagram and iterative curve about the impact of different parameter 𝑛𝑡.

frequency range for the proposed algorithm. The descriptive curves for one independent run of CBA. Figure 4(b) shows
statistics of the results are plotted in Figure 4(a), where 1–6 in that the performance of algorithm and convergence speed are
abscissa axis correspond to 𝐹 = 0.5, 1, 2, . . . , 5, respectively. preferable to the other condition.
As shown in Figure 4(a), the algorithm is very sensitive to Confirm the parameter 𝑝𝑠 = 45, 𝑓 ∈ [0, 1], and then
the initial values, while 𝑓 ∈ [0, 0.5], the exploring ability investigate the impact of the parameter 𝑛𝑡 for the proposed
is weak and cannot avoid the premature convergence and algorithm. The descriptive statistics of the results are plotted
escape from local minima, and the stability is poor. The in Figure 5(a), where 1–8 in abscissa axis correspond to
performance of algorithm is relatively good when the range 𝑛𝑡 = 1000, 2000, . . . , 8000, respectively. Figure 5(b) is the
is [0, 1]; the performance of CBA gradually reduces while the iterative curves for one independent run of the proposed
upper limit of frequency increases. Figure 4(b) is the iterative algorithm with different parameter 𝑛𝑡. As shown in Figure 5,
The Scientific World Journal 9

Table 1: Benchmarking test functions.

Benchmarks Functions expression Exact value 𝑥∗ Search space

functions
𝑛

𝑓1 : Sphere 𝑓(𝑥) = ∑𝑥𝑖2 𝑓min = 0 (0, 0, . . . , 0) [−10, 10]

𝑖=1
𝑛 𝑛
󵄨 󵄨 󵄨 󵄨
𝑓2 : Schwefel 𝑓(𝑥) = ∑ 󵄨󵄨󵄨𝑥𝑖 󵄨󵄨󵄨 + ∏ 󵄨󵄨󵄨𝑥𝑖 󵄨󵄨󵄨 𝑓min = 0 (0, 0, . . . , 0) [−10, 10]
𝑖=1 𝑖=1
𝑛
2
𝑓3 : Rosenbrock 𝑓(𝑥) = ∑ [(𝑥𝑖 − 1)2 + 100(𝑥𝑖+1 − 𝑥𝑖 2 ) ] 𝑓min = 0 (1, 1, . . . , 1) [−2.408, 2.408]
𝑖=1
𝑛
1
𝑓 (𝑥) = 20 + 𝑒 − 20 exp [−0.2√( ) × ∑𝑥𝑖2 ]
𝑛 𝑖=1
𝑓4 : Ackley 𝑛
𝑓min = 0 (0, 0, . . . , 0) [−30, 30]
1
− exp [−0.2√ ( ) × ∑ cos (2𝜋𝑥𝑖 )]
𝑛 𝑖=1
𝑛 𝑛
1 2 𝑥
𝑓5 : Griewangk 𝑓(𝑥) = × ∑𝑥𝑖 − ∏ cos 𝑖 + 1 𝑓min = 0 (0, 0, . . . , 0) [−600, 600]
4000 𝑖=1 𝑖=1 √𝑖
𝑛

𝑓6 : Rastrigin 𝑓(𝑥) = 10𝑛 + ∑ [𝑥𝑖2 − 10 cos (2𝜋𝑥𝑖 )] 𝑓min = 0 (0, 0, . . . , 0) [−5.12, 5.12]
𝑖=1
5 5

𝑓7 : Shubert 𝑓 (𝑥, 𝑦) = [∑𝑖 cos (𝑖 + (𝑖 + 1) 𝑥)] ⋅ [∑𝑖 cos (𝑖 + (𝑖 + 1) 𝑦)] 𝑓min ≈ −186.7309 — [−10, 10]
𝑖=1 𝑖=1
2
𝑓8 : Easom 𝑓(𝑥, 𝑦) = − cos (𝑥) cos (𝑦) exp [− (𝑥 − 𝜋)2 + (𝑦 − 𝜋) ] 𝑓min = −1 (𝜋, 𝜋) [−10, 10]

LGC, CLG, CGL, GL, CL, and CG, respectively, where the
combination of letters represents the combination of different
strategies. As shown as Figure 6, LGC sometimes can find a
10−5 better solution, but it is unstable, and several extreme outliers
log (fitness)

appear sometimes, which represent that the algorithm may be

premature convergence. The CLG is the most unstable, which
10−10 is sensitive to the initial position. GL can repeatedly find
a better solution; however, several extreme outliers appear
likewise, which represent that the algorithm lacks stability.
10−15
CL optimizes difficultly and its performance is poor. With CG
several mild outliers will appear and the performance of CG
is somewhat less than CGL, and the reason is lack of diversity
1 2 3 4 5 6 without L. Nevertheless, the CGL loses trifling precision of
Strategies combination the optimum value, but the overall performance is the most
stable, and it can always find better solution.
Figure 6: Box-and-whisker after 20 runs independently about
different strategies combination.

4.3. Experimental Results and Analysis. In order to compare

the performance with other algorithms, eight test functions
the performance of the algorithm has improved to some are selected to test CBA convergence. In Table 1, the values
extent with gradually increasing parameter 𝑛𝑡. In general, listed in the search space column are used to specify the range
the algorithm parameter is not sensitive for parameter 𝑛𝑡. of the initial random particles’ position; the 𝑥∗ denotes the
Considering the precision of optimal value, convergence global optimum, and the 𝑓min is the corresponding fitness
speed, and stability, parameter 𝑛𝑡 around 5000 or 6000 is value.
preferable. In this paper, 𝑛𝑡 = 6000. In [11], a cloud model based genetic algorithm (CGA)
was proposed; CGA is based on both the idea of GA and the
4.2. Combination Strategies Analysis. In order to discuss the properties of randomness and stable tendency of a normal
impact of three strategies (bats cloud updating, swarm infor- cloud model. In [12], a cloud model based evolutionary
mation communication, and bats random Lévy flight) after algorithm (CBEA) was proposed by Zhang et al., which is
selecting proper parameter, firstly, set bats cloud updating as based on the outstanding characteristics of the cloud model
C, swarm information communication as G, and bats random on the process of transforming a qualitative concept to a
Lévy flight as L and then select Rosenbrock as test function. set of quantitative numerical values and integrates with the
The statistical results after 20 times run independently are basic principle of evolutionary computation. In [13], cloud
shown in Figure 6, where 1–6 in abscissa axis correspond to based evolutionary algorithm (CBEA) was proposed by Liu
10 The Scientific World Journal

Table 2: Experimental results comparison between CGA, CBA, and SAPSO.

Algorithms 𝑓2 𝑓3 𝑓5 𝑓6 𝑓7
S-S [−10, 10] [−2.048, 2.048] [−512, 512] [−10, 10] [−10, 10]
F-D 5 2 10 15 2
CGA 0.00013 2.5749𝑒 − 08 0.01170 1.8529𝑒 − 06 −186.6267
SAPSO — 3.5383𝑒 − 003 5.7773𝑒 − 008 1.9425𝑒 − 004 —
CBA 1.9850𝑒 − 93 1.2683𝑒 − 12 0 0 −186.7309

Table 3: Experimental results comparison between CBEA08, CBA, and SAPSO.

Algorithms 𝑓1 𝑓5 𝑓6 𝑓7 𝑓8
S-S [−100, 100] [−600, 600] [−5.12, 5.12] [−10, 10] [−100, 100]
F-D 10 10 10 2 2
CBEA08 0 0 0 −186.7309088310227 −1
SAPSO 0.04860173 5.7773𝑒 − 008 1.9425𝑒 − 004 — —
CBA 0 0 0 −186.7309088310230 −1

Table 4: Experimental results comparison between CBEA09, CBA, and SAPSO.

Algorithms 𝑓1 𝑓2 𝑓3 𝑓4 𝑓5
S-S [−5.12, 5.12] [−10, 10] [−30, 30] [−32.768, 32.768] [−32.768, 32.768]
CBEA09 1.1696𝑒 − 166 5.2927𝑒 − 97 26.178 0 0
SAPSO 0.04860173 — 3.5383𝑒 − 003 1.5684𝑒 − 003 5.7773𝑒 − 008
CBA 0 2.4707𝑒 − 223 28.879 0 0

et al., who discuss the atomized feature of cloud model; Table 4 shows the results of comparison between CBEA09
the selection pressure of evolution is adjusted by changing the and CBA, where results of CBEA09 in 50 runs are derived
hyper entropy that is the main factor in atomized feature. from [13], and results of SAPSO are derived from [6]. All
There are many ways to carry out the comparison of the selected functions are 30-dimensional function, and the
algorithm performance with different termination criteria; optimizing generation 𝑡 is set as 500 for all algorithms.
the preferable approaches is to compare their accuracies for a Tables 2, 3, and 4 not only show that the proposed
algorithm is feasible and effective but also demonstrate
fixed number of fitness function evaluations 𝐹𝐸𝑠. This paper
the superior approximation capabilities in high-dimensional
adopts fixed 𝐹𝐸𝑠 as the termination criterion (𝐹𝐸𝑠 = 𝑝𝑠 ×
space. The proposed algorithm can perform better perfor-
𝑡 × 𝑁, where 𝑝𝑠 is population size, 𝑡 is optimizing generation,
mance while the optimizing generation 𝑡 increases gradually.
and 𝑁 is the number of fitness function evaluations in each As shown as Table 2, CBA has higher precision than CGA
optimizing generation). In order to compare with different both unimodal and multimodal functions. From Table 3,
algorithms, search space (SS) and function dimension (FD) CBA has similar precision than CBEA08 both unimodal
of selected benchmark functions are consistent with corre- and multimodal function, and several theoretical values of
sponding algorithm. CBA is run 50 times independently for benchmark function can be achieved. Table 4 shows the
each function and the mean of the function values found in results that are tested under high-dimensional condition. The
50 runs was described in experimental result. CBA can perform with better precision, except for Rosen-
In [11], 𝑝𝑠 was set as 100, selected different test functions brock function which has no best performance. Rosenbrock
have different optimizing generation 𝑡, the minimum 𝐹𝐸𝑠 function has a narrow valley from the perceived local optima
is 106 , and the maximum 𝐹𝐸𝑠 is 2 × 107 . In this paper, to the global optimum.
optimizing generation 𝑡 is set as 200 in each run for corre-
sponding function. 𝐹𝐸𝑠 = 27000. The experimental results
are presented in Table 2, where results of CGA in 30 runs are
5. Conclusions
derived from [11] and results of SAPSO (simulated annealing In this paper, the bat algorithm is used for reference;
particle swarm optimization) are derived from [6]. two mechanisms were introduced, population information
Table 3 shows the results of comparison between CBEA08 communicating of each individual and random Lévy flights,
and CBA, where results of CBEA08 in 50 runs are derived to propose a cloud model bat algorithm based on normal
from [12], and results of SAPSO are derived from [6]. Here cloud model and echolocation mechanism. Cloud model
𝑝𝑠 = 1000, 𝑡 = 100, and 𝐹𝐸𝑠 = 105 . In this paper, for these builds a transformational bridge between a linguistic term
functions, 𝑡 is set as 500, corresponding to 𝐹𝐸𝑠 = 67500. of qualitative concept and quantitative representation, which
The Scientific World Journal 11

reflects randomness, fuzziness, and the relationship between [10] D. Y. Li, H. J. Meng, and X. M. Shi, “Membership clouds and
randomness and fuzziness of uncertainty in knowledge rep- membership cloud generations,” Journal of Computer Research
resentation. A population bats cloud model 𝑃𝑏𝑐𝑗 (𝐸𝑥, 𝐸𝑛, 𝐻𝑒) and Development, vol. 32, no. 6, pp. 15–20, 1995.
was built, which analogizes to cloud 𝐶(𝐸𝑥, 𝐸𝑛, 𝐻𝑒), which [11] C. Dai, Y. Zhu, W. Chen, and J. Lin, “Cloud model based genetic
depicts the qualitative concept: “bats approach their prey.” algorithm and its applications,” Acta Electronica Sinica, vol. 35,
CBA mainly considers the balance between the global no. 7, pp. 1419–1424, 2007.
random search and local search and the balance between [12] G. Zhang, R. He, Y. Liu, D. Li, and G. Chen, “Evolutionary
intensification and diversification. In addition, we discuss the algorithm based on cloud model,” Chinese Journal of Computers,
mechanism and parameter set of CBA; several appropriate vol. 31, no. 7, pp. 1082–1091, 2008.
parameters are set. The simulation results show that the [13] Y. Liu, D. Li, G. Zhang, and G. Chen, “Atomized feature in cloud
proposed algorithm is feasible, effective, and robust. based evolutionary algorithm,” Acta Electronica Sinica, vol. 37,
no. 8, pp. 1651–1658, 2009.
[14] A. Hedenström, L. C. Johansson, M. Wolf, R. von Busse, Y.
Conflict of Interests Winter, and G. R. Spedding, “Bat flight generates complex
aerodynamic tracks,” Science, vol. 316, no. 5826, pp. 894–897,
The authors declare that there is no conflict of interests 2011.
regarding the publication of this paper. [15] M. E. Bates, J. A. Simmons, and T. V. Zorikov, “Bats use
echo harmonic structure to distinguish their targets from
Acknowledgments background clutter,” Science, vol. 333, no. 6042, pp. 627–630,
2011.
This work is supported by the National Science Foundation [16] M. Knörnschild, K. Jung, M. Nagy, M. Metz, and E. Kalko, “Bat
of China under Grant no. 61165015, Key Project of Guangxi echolocation call facilitate social communication,” Proceedings
Science Foundation under Grant no. 2012GXNSFDA053028, of the Royal Society B, vol. 279, no. 1748, pp. 4827–4835.
and Key Project of Guangxi High School Science Foundation [17] A. V. Chechkin, R. Metzler, J. Klafter, and V. Y. Gonchar,
under Grant nos. 20121ZD008 and 201203YB072. “Introduction to the theory of Lévy flights,” in Anomalous
Transport: Foundations and Applications, pp. 129–162, Wiley-
VCH, 2008.
References [18] “Alpha-Stable distributions in MATLAB,” http://math.bu.edu/
people/mveillet/html/alphastablepub.html.
[1] J. Kennedy and R. Eberhart, “Particle swarm optimization,” in
Proceedings of the 1995 IEEE International Conference on Neural [19] A. M. Reynolds, M. A. Frye, and S. Rands, “Free-flight odor
Networks, vol. 4, pp. 1942–1948, December 1995. tracking in Drosophila is consistent with an optimal intermit-
tent scale-free search,” PLoS ONE, vol. 2, no. 4, article e354, 2007.
[2] R. Storn and K. Price, “Differential evolution—a simple and
efficient heuristic for global optimization over continuous [20] C. T. Brown, L. S. Liebovitch, and R. Glendon, “Lévy flights in
spaces,” Journal of Global Optimization, vol. 11, no. 4, pp. 341– dobe Ju/’hoansi foraging patterns,” Human Ecology, vol. 35, no.
359, 1997. 1, pp. 129–138, 2007.
[21] X. Yang and S. Deb, “Cuckoo search via Lévy flights,” in
[3] X. S. Yang, “A new metaheuristic Bat-inspired algorithm,” in
Proceedings of the World Congress on Nature and Biologically
Nature Inspired Cooperative Strategies for Optimization (NICSO
Inspired Computing (NABIC ’09), pp. 210–214, December 2009.
’10), vol. 284 of Studies in Computational Intelligence, pp. 65–74,
Springer, 2010. [22] D. Y. Li, C. Y. Liu, and W. Y. Gan, “Proof of the heavy-tailed
propety of normal cloud model,” Engineer and Science of China,
[4] X. S. Yang, “Bat algorithm for multi-objective optimisation,”
vol. 13, no. 4, pp. 20–23, 2011.
International Journal of Bio-Inspired Computation, vol. 3, pp.
267–274, 2011. [23] D. Y. Li and Y. Du, Artificial Intelligence with Uncertainty,
National Defense Industry Press, Beijing, China, 2005.
[5] A. H. Gandomi, X. S. Yang, S. Talatahari, A. H. Alavi, and S.
Talatahari, “Bat algorithm for constrained optimization,” Neural [24] D. Y. Li and C. Y. Liu, “Study on the universality of the normal
Computing and Applications, vol. 22, pp. 1239–1255, 2013. cloud model,” Engineer and Science of China, vol. 6, no. 8, pp.
28–33, 2004.
[6] X. S. He, W. J. Ding, and X. S. Yang, “Bat algorithm based
[25] H. Lu, Y. Wang, D. Li, and C. Liu, “Application of backward
on simulated annealing and gaussian perturbations,” Neural
cloud in qualitative evaluation,” Chinese Journal of Computers,
Computing and Applications, 2013.
vol. 26, no. 8, pp. 1009–1014, 2003.
[7] S. Mishra, K. Shaw, and D. Mishra, “A new meta-heuristic bat
inspired classification approach for microarray data,” Procedia
Technology, vol. 4, pp. 802–806, 2012.
[8] K. Khan, A. Nikov, and A. Sahai, “A fuzzy bat clustering method
for ergonomic screening of office workplaces,” in Proceedings
of the 3rd International Conference on Software, Services and
Semantic Technologies (S3T ’11), vol. 101 of Advances in Intelligent
and Soft Computing, pp. 59–66, Springer, 2011.
[9] K. Khan and A. Sahai, “A comparison of BA, GA, PSO, BP and
LM for training feed forward neural networks in e-learning
context,” International Journal of Intelligent Systems Technologies
and Applications, vol. 4, no. 7, pp. 23–29, 2012.
 Advances in Journal of
Industrial Engineering
Multimedia
Applied
Computational
Intelligence and Soft
Computing
The Scientific International Journal of
Distributed
Hindawi Publishing Corporation
World Journal
Hindawi Publishing Corporation
Sensor Networks
Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Advances in

Fuzzy
Systems
Modelling &
Simulation
in Engineering
Hindawi Publishing Corporation
Hindawi Publishing Corporation Volume 2014 http://www.hindawi.com Volume 2014
http://www.hindawi.com

Submit your manuscripts at

Journal of
http://www.hindawi.com
Computer Networks
and Communications Advances in
Artificial
Intelligence
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014

International Journal of Advances in

Biomedical Imaging Artificial
Neural Systems

International Journal of
Advances in Computer Games Advances in
Computer Engineering Technology Software Engineering
Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

International Journal of
Reconfigurable
Computing

Advances in Computational Journal of

Journal of Human-Computer Intelligence and Electrical and Computer
Robotics
Hindawi Publishing Corporation
Interaction
Hindawi Publishing Corporation
Neuroscience
Hindawi Publishing Corporation Hindawi Publishing Corporation
Engineering
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014