You are on page 1of 22

Artif Intell Rev

DOI 10.1007/s10462-011-9206-1

A review of clonal selection algorithm and its applications

Berna Haktanirlar Ulutas · Sadan Kulturel-Konak

© Springer Science+Business Media B.V. 2011

Abstract Recently, clonal selection theory in the immune system has received the attention
of researchers and given them inspiration to create algorithms that evolve candidate solu-
tions by means of selection, cloning, and mutation procedures. Moreover, diversity in the
population is enabled by means of the receptor editing process. The Clonal Selection Algo-
rithm (CSA) in its canonical form and its various versions are used to solve different types
of problems and are reported to perform better compared with other heuristics (i.e., genetic
algorithms, neural networks, etc.) in some cases, such as function optimization and pattern
recognition. Although the studies related with CSA are increasingly popular, according to
our best knowledge, there is no study summarizing the basic features of these algorithms,
hybrid algorithms, and the application areas of these algorithms all in one paper. Therefore,
this study aims to summarize the powerful characteristics and general review of CSA. In
addition, CSA based hybrid algorithms are reviewed, and open research areas are discussed
for further research.

Keywords Artificial immune systems · Clonal selection algorithms · Review ·


Hybrid algorithms

1 Introduction

Artificial Immune System (AIS) algorithms imitate the principles of immune systems and
are inspired by the features such as the distributed memory, robustness, self-organizing, and
decentralized control mechanism. Among the Evolutionary Algorithms (EAs), AIS is a bio-

B. Haktanirlar Ulutas (B) · S. Kulturel-Konak


Department of Industrial Engineering, Eskisehir Osmangazi University, Eskisehir 26480, Turkey
e-mail: bhaktan@ogu.edu.tr
S. Kulturel-Konak
Management of Information Systems, Penn State Berks, Tulpehocken Road, P.O. Box 7009, Reading,
PA 19610-6009, USA
e-mail: sadan@psu.edu

123
B. Haktanirlar Ulutas, S. Kulturel-Konak

logically-inspired computation paradigm that emerged in recent years. Research in AIS has
become popular in the late 1990s. Dasgupta (1998) published one of the early studies on
AIS with theories, models, simulations, and applications. Other pioneering studies include
De Castro and Von Zuben (1999, 2000b), and De Castro and Timmis (2002). Zheng et al.
(2010) review the general applications of the AIS and address open issues to make the AIS
a real world problem solving technique.
Dasgupta and Nino (2008) compare the basic principles of EAs for optimization, Analyt-
ical Neural Networks (ANN) for classification, and AIS. The related information is provided
using chromosome strings for EAs, connection strings for ANN, and component concentra-
tion/network for AIS.
De Castro and Timmis (2002) give details of the main models in AIS such as clonal selec-
tion, immune networks, and negative selection theories. Timmis et al. (2008) present a simple
example for each of the main types of AIS algorithms and then give the details of the theo-
retical analysis. The immune system based models and algorithms are used to solve different
types of problems. Hart and Timmis (2008) state that the recent applications of AIS include
the areas, such as computer security, numerical function optimization, combinatorial optimi-
zation (i.e., scheduling), learning, bio-informatics, image processing, robotics (i.e., control
and navigation), adaptive control systems, virus detection, and web mining. Since the aim
of this paper is not outlining all the developments within AIS, the interested readers might
refer to recent review papers and books, such as Garrett (2005), Dasgupta (2006), Timmis
(2007), Dasgupta et al. (2007), Dasgupta and Nino (2008), and Wong and Lau (2009), for
further detailed information.
Wong and Lau (2009) note the future directions of AIS algorithms as follows: exploration
of novel immune theories, analyzing distinct advantages of AIS compared to other EAs, inte-
gration of AIS with other techniques, and applications of AIS algorithms to solve industrial
problems. Timmis (2007) states that a large part of AIS works have been based on the clonal
selection theory. Therefore, this paper aims to provide up-to-date information, specifically
for CSA, as well as future research directions.
Hofmeyr and Forrest (2000) discuss that clonal selection can be considered as genetic
algorithm (GA) without crossover. However, De Castro and Von Zuben (2000a) state that
GA cannot cover all features of clonal selection such as affinity proportional reproduction and
hypermutation. De Castro and Von Zuben (2000b) review CSA with engineering applications
and provide the details of binary character recognition, multi-modal optimization, and travel-
ling salesperson problems (TSP) applications. Brownlee (2007) provides a detailed summary
of CSA and proposes the algorithm taxonomy, a standardized nomenclature and a general
model of such algorithms. In addition, he discusses the similarities and differences between
CSA and Evolutionary Computation (EC) as well as general research trends for CSA.
Besides general theoretical or application review studies of CSA, based on the best knowl-
edge, there is no review paper considering CSA based algorithms in two groups, such as
canonical CSA and hybrid CSA. By keeping these two groups in mind, this paper also aims
to note on comparison of CSA results with other heuristics and specific problem applications.
This paper is organized as follows. In the following section, general principles of canonical
CSA and its versions are summarized. Then, the novel algorithms, successful comparisons
with other heuristics, the problem types solved using CSA, and the trends in the appli-
cation areas are identified. The next section provides the hybrid CSA algorithms and the
last section forms general guidelines for future studies that would be developed based on
CSA.

123
A review of clonal selection algorithm and its applications

2 CSA methods and applications

The clonal selection theory in an immune system is used to explain the basic response of the
adaptive immune system to an antigenic stimulus. The theory depends on the idea that only
cells that are capable of recognizing an antigen will proliferate (Burnet 1959).
This study focuses on the clonal selection theory inspired CSA that is generally formed by
the operators, such as initial population generation, selection, cloning, hypermutation, and
receptor editing. The common use is to randomly generate the antibodies in the population.
It is known that choosing a population size that is too small can lead the algorithm towards
a fast convergence, and if it is too large, the algorithm may need extensive computational
resources. Therefore, determining the right population size might be challenging. De Castro
and Timmis (2002) state that the evolutionary selection mechanisms, such as, roulette wheel,
elitist, rank-based, bi-classist, and tournament selections, can be used for the selection and
cloning in immune based algorithms. The hypermutation operator in CSA is inversely pro-
portional to the affinity of the antibody. Finally, to attain diversity in the population and
explore new search regions, the receptor editing procedure is applied.
The general steps of CLONALG (CLONal selection ALGorithm) by De Castro and Von
Zuben (2002) that was initially proposed to solve pattern recognition problems are defined
as follows:

Step 1: Initialization: Randomly initialize a population of individuals (N ),


Step 2: Evaluation: Given a set of patterns to be recognized (P), for each pattern, determine
its affinity (match) with each element of the population,
Step 3: Selection and cloning: Select a number (n) of the best highest affinity elements of
N and generate copies of these individuals proportionally to their affinity with the
antigen,
Step 4: Hypermutation: Mutate all the copies with a rate proportional to their affinity with
the input pattern,
Step 5: Receptor editing: Add the mutated individuals to the population and reselect a
number (d) of the maturated (optimized) individuals as memory,
Step 6: Repeat steps 2–5 until a termination criterion is met.

In order to identify the general trend of CSA, the keywords such as AIS, immune (immuno)
engineering, CSA, and hybrid CSA are searched in the databases that mainly cover computer
science related topics (i.e., INSPECT, EBSCO, etc.). Figure 1 illustrates the number of stud-
ies with publication years. It is observed that the number of studies concerning both CSA
and hybrid CSA has an increasing trend since 2002.

Fig. 1 The number of CSA and hybrid CSA papers according to corresponding publication years

123
B. Haktanirlar Ulutas, S. Kulturel-Konak

2.1 Versions of CSA

This section provides a brief summary and differences of the CSA available in the acces-
sible literature compared with the CLONALG features. The CLONALG is relatively low
in complexity and has a small number of user parameters when compared to other AIS
systems such as Artificial Immune Recognition System (AIRS) proposed by Watkins et al.
(2004). CLONALG has seven user defined parameters which include; antibody population
size, memory pool size, selection pool size, remainder replacement size, clonal factor, num-
ber of generations, and the random number generator seed (required for all techniques with a
stochastic element). Antibody population size (N) specifies the total number of antibodies to
be maintained by the system. This includes both the memory antibody pool and the remainder
antibody pool. Selection pool size (n) refers to the total number of antibodies with greatest
affinity to draw from the antibody population for cloning. Smaller values can decrease the
diversity of the population by ensuring only the best affinity antibodies are cloned. Remainder
replacement size (d) is the total number of lowest affinity antibodies to replace with random
antibodies on each antigen and provides a mechanism to introduce additional diversity (more
than the local diversity of mutation), into the antibody population. When the rank based
selection is applied, clonal factor (β) is defined as a scaling factor for the number of clones
created for selected antibodies. The lower the value of β, the more search in the local area
(in relation to current antibodies) is performed by the algorithm. Number of generations (G)
specifies the total number of algorithm iterations to perform, where a single iteration sees
the system exposed to all known antigens. The parameter G controls the amount of learning
that the system will perform on the problem domain. High generation values may result the
problem being stuck on a locally optimal solution.
Parallel CLONALG depends on the idea that the immune system is inherently a parallel
system both in terms of distributed nature of its control, response, maturation process, and
the management of the cell population. Watkins et al. (2003) augment CLONALG to be more
efficient by operating in parallel. The parameters were defined as N , n, and d. The parallel
version of CLONALG was tested on the binary character recognition problem previous used
in DeCastro and Von Zuben (2000a, 2002). The Clonal Selection Classification Algorithm
(CSCA) is an immune inspired classifier and an augmented version of CLONALG. White
and Garrett (2003) prove the technique to be robust, self-tuning, insensitive to user param-
eters and competitive as a classification system on the binary pattern recognition dataset.
Adaptive Clonal Selection (ACS) is defined as the parameter free version of CLONALG
(Garrett 2004). The technique is applied to static function optimization problem domains.
The parameter values for CLONALG such as population size N , d, number of function eval-
uations and initial values for β and n are stated as a worthy area for investigation. It is also
shown that real-valued antibodies and the use of evolution strategy like mutation schemes
on antibodies might be a useful contribution for the technique. Yang et al. (2008a, 2008b)
and Gong et al. (2010a) substitute the mutation in CSA with the local search technique in
Lamarckian Clonal Selection Algorithm (LCSA) and adopt real-valued representation and
tournament selection method. On the other hand, Gong et al. (2010b) introduce Baldwin-
ian learning that can flatten out the fitness landscape around the optimal regions and used
real-coded representation.
The psychoclonal algorithm introduced by Tiwari et al. (2005) is inspired from the basic
features from Maslow’s Need Hierarchy, human psychology and clonal selection theory. The
psychoclonal algorithm studied by Dashora et al. (2008) selects a number of highest affinity
antibodies during the selection step and uses an inverse mutation followed by a pairwise
mutation during the maturation step. Liu et al. (2009) adopt Randomized ClonalExpansion

123
A review of clonal selection algorithm and its applications

Strategy (RCSA), where each antibody reproduces clones randomly. Khilwani et al. (2008)
introduce Gaussian mutation and chaos mutation. The chaos mutation, which is used to per-
form mutation in low affinity antibodies, helps in escaping from local optima and favours less
computation time in exploiting local neighbourhood; whereas, the Gaussian mutation due to
its fine tuning capability and small step size supports the algorithm to reach global optima.
Ma et al. (2009) and Shi and Jing (2009) apply a vaccination operator to modify the genes
in accordance to prior knowledge to gain higher fitness and present a super mutation operator
to avoid fast local convergence. Li and Chen (2009) introduce Reconfigurable Space Clone
Selection Algorithm (RSCSA) that investigates suitable adjustment of antibody population
size and antibody search space. Engin and Doyen (2004) use a special version of the roulette
wheel method during cloning process and a two phased mutation procedure where the gen-
erated clones undergo an inverse mutation procedure at first, then if there is no improvement,
apply pairwise interchange mutation method.
Besides these outstanding CSA applications, there are also a number of studies which
present different kinds of encoding schemes, mutation types, and selection procedures. (Cruz-
Cortes et al. 2005, b) investigate CLONALG with binary and gray encoding schemes as well
as a real-valued encoding scheme with a mutation scheme based on Gaussian and Cauchy
random numbers. Wang (2005) proposes a CSA based on CLONALG with a static clone
size. Campelo et al. (2005) propose a Real-coded Clonal Selection Algorithm (RCSA) with
Gaussian-based mutation and rank based selection. Cutello et al. (2005) define the cloning
potential as a truncated exponential function and an aging (i.e., pure aging and elitist aging)
operator during the elimination of old B-cells. Acan (2004) propose Artificial Immune Sys-
tem with Mutation Multiplicity (AISMM) that uses multiple concurrent mutation operators.
Ramaswamy et al. (2007) define a goal directed mutation process. Panigrahi et al. (2007)
utilize binary flip mutation with the probability of mutation varying from 0.010 to 0.035
and use tournament selection. Lu and Yang (2009) introduce the Cauchy mutation for the
Improved CSA (IMCSA).
It is observed that, traditionally, a deterministic selection rule is adopted to select better
antibodies for proliferation in CSA. However, deterministic selection rule selects only the
best antibodies for proliferation, and may lead to the premature convergence of the algorithm.
To overcome this difficulty, a roulette wheel selection based rule is adopted in a number of
studies for the selection of antibodies for proliferation. The variation in antibodies is per-
formed by a hypermutation and receptor editing mechanism in CSA. The hypermutation
operator works on a similar fashion to the mutation operator of GAs (Tiwari et al. 2005).
However, in CSA, the inferior antibodies are hyper-mutated at a higher rate on high affinity
antibodies. On the other hand, the use of two mutation types in parallel or consecutively aug-
ment the algorithm’s exploiting the local neighbourhood and fine tuning capability (i.e., to
reach global optima). Hypermutation guides the system toward local optima, while receptor
editing helps to avoid local optima.

2.2 Problem types solved with CSA

In this study, the problem types that are mainly studied by means of CSA are grouped as
follows: function optimization (i.e., multi modal optimization and continuous function opti-
mization), pattern recognition (PR) (i.e., binary character and face detection), design (i.e.,
continuous design, electromagnetic design, and hardware/software design), scheduling (i.e.,
job shop scheduling and project scheduling), industrial engineering (IE) related problems
(i.e., facility location, layout, assembly planning, and material handling systems), TSP, and

123
B. Haktanirlar Ulutas, S. Kulturel-Konak

Fig. 2 Problem types solved using CSA

others (i.e., time series prediction, classification, fault diagnosis, machine learning, and virus
detection, etc.).
The most common type of problem solved by CSA is function optimization (Acan 2004;
Cutello et al. 2005; De Castro and Von Zuben 2002; Dilettoso and Salerno 2006; Garrett 2004;
Du et al. 2002; Khilwani et al. 2008; Kelsey and Timmis 2003; Liu et al. 2009; Ramaswamy
et al. 2007; Yang et al. 2008a, 2008b). Then, the pattern recognition type problems have
a notable number of studies (Garain et al. 2006; Li et al. 2008; Ma et al. 2006; Watkins
et al. 2003; White and Garrett 2003; and Zhang and Hou 2003) followed by design prob-
lems (Akdagli et al. 2007; Campelo et al. 2004, 2005; Dong et al. 2007; Garain et al. 2006;
Moghaddam and Kardan 2009; Wang 2005; Wu et al. 2009; and Zuo and Fan 2003). Darmoul
et al. (2006) provide a review of scheduling problems solved with AIS. Recent studies, such
as Coello et al. (2003); Engin and Doyen (2004); Lu and Yang (2009); Ong et al. (2005),
and Pan et al. (2006) solve scheduling problems by modified versions of CSA. The problems
that fall in IE related problem classification in this study (Pang and Feng 2006; Cao and Xiao
2007; Mak and Lau 2007; Dashora et al. 2008; Ulutas and Islier 2009; Kumar et al. 2009, and
Ulutas and Islier (2010)) have a potential for further research. The TSP is another promising
area for CSA (Zhu et al. 2007; Zhao et al. 2008; and Dai et al. 2009). Besides these studies,
researchers, such as Cutello et al. (2005), Ding and Li (2009); Li and Chen (2009); Ma et al.
(2009), Yu and Hou (2004), Kim and Bentley (2002), Brownlee (2005), Cutello and Nicosia
(2006), Litvinenko et al. (2005), and Panigrahi et al. (2007) assess the performance of CSA
versions on various problem types. Figure 2 provides the summary of problem types that
are commonly solved using canonical CSA. When the publication years of the studies are
considered, there is no significant evidence of an apparent trend for any type of problem.

2.3 CSA comparison with other heuristics

De Castro and Von Zuben (1999) propose CLONALG with the various properties, such as
maintenance of a specific memory set, selection and cloning of most stimulated antibod-
ies, death of non-stimulated antibodies, affinity maturation, and maintenance of diversity.
Many researchers have been developing algorithms to improve CLONALG for various prob-
lem types, such as pattern matching and multi-modal function optimization. Cutello et al.
(2005) propose two modified versions called CLONALG1 and CLONALG2 with varying

123
A review of clonal selection algorithm and its applications

elitist strategies which were raced against the optimum-Immune Algoritm (opt-IA). Dilet-
toso and Salerno (2006) use CLONALG as a niching technique and compared the results
with traditional EC niching approaches. Ma et al. (2006) compare their CSA and template
matching based algorithm results with immune genetic, standard GAs, and evolutionary
GAs. Ramaswamy et al. (2007) compare their algorithm with the State Dependent Algebraic
Riccati Equation (SDARE) method. Cao and Xiao (2007) compare the immune optimiza-
tion approach with GAs. Mak and Lau (2007) compare age AIS with different immune
based algorithms, nearest neighbour heuristics, GA, and ant colony systems. Ulutas and
Islier (2009) assess the performance of CSA against thirty eight solutions approaches such
as, dynamic programming, GA, SA, and hybrid ant colony. The results of CSA are always
reported to perform better compared with the aforementioned heuristics (i.e., GAs and ant
colony optimization (ACO)).
In the scope of this study, the CSA based studies in the literature, the majority of which
are conference proceedings papers, are investigated. Table 1 summarizes the recent CSA
studies where related reference is given in the first column, and the algorithm name, main
characteristics and problem type are given in the following columns.

3 Hybrid CSA methods and applications

Besides the novel CSA theories, it is observed from Fig. 1 that the number of hybrid CSA
studies is also increasing. Although CSA has reported to provide better solution results com-
pared with a number of heuristics for specific problem types as noted in the previous section,
researchers have identified some weaknesses of CSA, often associated with the convergence
behaviour (Cutello et al. 2007; Hong 2008). Therefore, hybrid algorithms were proposed
to improve the performance of the canonical CSA. In addition, there are also a number of
studies where CSA is used to improve the performance of heuristics, such as PSO, GA, etc.
Detailed summary of immune, swarm and EAs are available at De Castro and Von Zuben
(2000a,b) for interested readers.
The problem types that are solved using hybrid CSA methods are grouped as function
optimization, classification (i.e., clustering and data mining), scheduling, TSP, PR, and others
(i.e., test data generation problem, routing, etc.). The most common use of hybrid CSA meth-
ods are for function optimization problems (Wang et al. 2006a,b,c; Tsai et al. 2007; Mitra
and Venayagamoorthy 2008; Chen and Guo 2008; Gao et al. 2009; Hong 2009; and Gong
et al. 2009). The solutions for classification type problems are obtained by Fang et al. (2008),
Karakasis and Stafylopatis (2006, 2008), Li and Chen (2009), Rabbani AND Panahi (2008),
and Wang et al. (2009). TSP (Dai et al. 2009; Gao et al. 2008a,b; Yong and Sunjun 2009) and
scheduling (Chang et al. 2009; Jia et al. 2008; Naderi et al. 2009; Yang et al. 2008a,b), are the
other common problem types for CSA hybrid applications. Figure 3 provides a summary for
the common problem types that are solved using hybrid CSA. There also exist some novel
problem types that attract researchers’ interest. To name a few, Huang et al. (2009) use CSA
for license plate pattern recognition problems in a parking unit and Liaskos and Roper (2008)
study automatic test data generation problem.
CSA is a random search heuristic that may not sufficiently utilize system feedback infor-
mation. ACO has distributed parallel overall search ability and can converge on an optimal
path by the accumulation and update of an information pheromone. However, there is a lack
of early stage pheromone, and the solving speed is slow. Therefore, Gao et al. (2008b) intro-
duce the hybrid algorithm to eliminate the weak sides of CSA and ACO. EAs cannot perform
well in the case of Multimodal Dynamic Optimization (MDO) because they tend to converge

123
Table 1 A summary of recent CSA characteristics and applications

Reference Algorithm name Main characteristics Problem type

123
De Castro and Von Zuben CLONALG Explicitly takes into account the affinity maturation of Multimodal and
(2002) the immune response combinatorial
optimization
Khilwani et al. (2008) Fast clonal algorithm Uses an advanced mutation strategy and elitism based
(FCA) immune memory
Du et al. (2002) Antibody clone algorithm Is able to avoid premature evolution
Dilettoso and Salerno CLONALG Modifies CLONALG as a niching technique, is
(2006) compared with traditional EC niching approaches
Kelsey and B-Cell algorithm (BCA) Uses elitist replacement, contiguous somatic Continuous function
Timmis (2003) hypermutation optimization
Garrett (2004) Adaptive clonal selection Adaptively determines the mutation parameter, the
(ACS) number of antibodies selected for cloning, and the
number of clones produced for each antibody
Acan (2004) AIS with mutation Uses multiple concurrent mutation operators
multiplicity (AISMM)
Cutello et al. (2005) CLONALG1, Use varied elitist strategies, is compared with the opt-IA Function optimization
CLONALG2
Liu et al. (2009) Randomized clonal expansion strategy (RCSA) Focuses on the ratio of clonal expansion, is compared
with conventional CSA
Ramaswamy et al. CSA Operates a new mutation operator on real values and Non-linear functions
(2007) aids in fast convergence, is compared with State
Dependent Algebraic Riccati Equation (SDARE)
method
Yang et al. (2008a) Lamarckian CSA (LCSA) Incorporates a recombination and tournament selection Numerical optimization
operator, is compared with CSA
Lamarckian CSA (LCSA) Utilizes the information of each individual to reinforce
Gong et al. (2010a) the exploitation with the help of Lamarckian local
search, compared with CSA
Gong et al. (2010b) Baldwinian Clonal Improves population by clonal proliferation, Baldwinian
Selection Algorithm learning, hypermutation, and clonal selection
(BCSA)
B. Haktanirlar Ulutas, S. Kulturel-Konak
Table 1 continued

Reference Algorithm name Main characteristics Problem type

Li et al. (2008) Reconfigurable space clone selection algorithm Reduces the search space to accelerate the convergent Function optimization
(RSCSA) speed and image matching
Watkins et al. Parallel CLONALG Collects all prepared memory pools based on a root Binary character
(2003) process, combines or merges them into a master recognition
memory pool
Zhang and Hou Niching clonal selection Combines negative Pattern matching problem
(2003) algorithm (NCSA) selection and refinement of anomaly detection
using CSA
White and Garrett Clonal classification Generalizes the pattern recognition version of Binary pattern
(2003) (CLONCLAS) CLONALG classification
Ma et al. (2006) CSA and template Is compared with immune GA, standard GA, and Face detection
matching evolutional GA
Garain et al. CSA Formulates a Handwritten numeral
(2006) retraining scheme recognition
for CSA
Wang (2005) Clonal selection Uses static clone Power filter design
A review of clonal selection algorithm and its applications

Campelo et al. Multi objective CSA Provides multi Electromagnetic


(2004) (MOCSA) objective version of design
RCSA
Campelo et al. Real-coded CSA Uses a
(2005) (RCSA) Gaussian-based
mutation
Zuo and Fan (2003) Chaotic search Integrates elements of Tuning Radial-Basis
immune algorithm the CLONALG Functions (RBF) in
(CSIA) real-time controller
design
Akdagli et al. CLONALG Takes discrete phase Design a reconfigurable dual
(2007) shifters into account beam linear antenna array
during synthesis
Dong et al. (2007) CSA The immune memory Stack filter design
performed better
than GA

123
Table 1 continued

Reference Algorithm name Main characteristics Problem type

123
Wu et al. (2009) Directional immune Adopts directional RF design
CSA (DICSA) operator during
optimization
process
Moghaddam and Clonal selection for Uses CSA for optimization of partitioning Hardware/software
Kardan (2009) partitioning (CSPA) codesign
Coello et al. CLONALG Uses hypermutation, library of antibodies, is compared Job shop scheduling
(2003) with greed search
Lu and Yang Improved CSA (IMCSA) Combines the global search ability of the CSA and the
(2009) diversity ability of the immune network theory
Engin and Doyen CSA Presents a generic systematic procedure to determine Hybrid flow shop
(2004) the optimum system parameters
Ong et al. (2005) ClonaFlex Is a clonal-selection Flexible job-shop
based algorithm
Pan et al. (2006) Dynamic CSA Includes crossover, mutation, and death Project scheduling
Pang and Feng (2006) Improved CSA Introduces a special string as the shape-space, a version Facility location
of the roulette wheel
Ulutas and Islier CSA Assess the performance of CSA for dynamic facility Facility layout
(2009) layout problems against thirty eight solution
approaches
Kumar et al. (2009) CSA Defines clone size as an increasing function of the Loop layout
affinity measure, carries inverse and pairwise mutation
in the hypermutation phase
Ulutas and Islier CSA Methodology to optimize the front page layout of an Layout optimization
(2010) internet newspaper
Cao and Xiao (2007) Immune optimization Uses immune selection, clonal selection, inoculation Assembly planning
approach (IOA) and immune metabolism, is compared with GA
B. Haktanirlar Ulutas, S. Kulturel-Konak
Table 1 continued

Reference Algorithm name Main characteristics Problem type

Mak and Lau (2007) Age artificial immune Is based on the clonal selection principle and the aging Optimal order pickings in
system (AAIS) concept, is compared with versions of AIS, nearest an Automated Storage
neighbor heuristics, GA and ant colony systems and Retrieval System
(AS/RS)
Dashora et al. Psycho-clonal Inherits Maslow’s theory related to psychology, is Process sequencing
(2008) algorithm compared with TS, SA
Zhu et al. (2007) Improved clonal Combines receptor editing with clonal selection, TSP
algorithm implements a self-crossover operator, is compared
with classical clonal selection
Zhao et al. (2008) Novel clonal selection Improves some
algorithm (NCSA) fundamental
problems for the
CSA
Dai et al. (2009) Immune CSA (ICSA) Introduces chaotic
initialization
mechanism and
A review of clonal selection algorithm and its applications

quantum interference
mechanism
Shi and Jing (2009) Immune clonal Introduce clonal proliferation, super mutation operators Capacitated vehicle
algorithm and clonal selection operators, improves the global routing problem
convergence speed, and can avoid prematurity (CVRP)
Ma et al. (2009) Improved immune Improves the global convergence speed, and introduces Vehicle routing
CSA (ICSA) clonal proliferation, super mutation and clonal
selection operators
Ding and Li (2009) CSA- support vector Selects a proper subset of features and optimal Classification
machines (SVM) parameters of SVMs classifier
Li and Chen (2009) Immune CSA (ICSA-SVM) Optimizes the parameter of SVM classifier by applying Fault diagnosis
the immunity CSA

123
Table 1 continued

Reference Algorithm name Main characteristics Problem type

123
Yu and Hou (2004) Clonal Selection Algorithm (CsAL) Is based on negative selection approach Virus detection
Kim and Bentley Dynamic clonal Is based on T-cell inspired negative selection Detection generation
(2002) selection
(DynamiCS)
Brownlee (2005) Clonal selection classification algorithm (CSCA) Develops supervised Machine learning
learning
Cutello and Nicosia Cloning, information gain, aging (CLIGA) Uses generational aging, information gain stopping Graph coloring
(2006) criteria
Litvinenko et al. Enhanced clonal Developed using CSA and idea GEP Time series prediction
(2005) algorithm (ECA)
Panigrahi et al. (2007) CSA Is compared with elitist GAs method, two-phase neural Economic load
network and particle swarm optimization (PSO) dispatch
B. Haktanirlar Ulutas, S. Kulturel-Konak
A review of clonal selection algorithm and its applications

Fig. 3 Problem types solved by hybrid CSA

through a single optimum. They may lose diversity as the optimum solution proliferates and
spreads over the population. By enabling diversity in the population by means of CSA, this
drawback is attempted to be solved (Nanas and De Roeck 2007).
CSA is used to improve local search and to avoid premature convergence in PSO by
Wang et al. (2006a; 2006c), Mitra and Venayagamoorthy (2008), and Hong (2009). Unlike
GAs, CSA performs a hyper mutation on the selected solutions with a rate that is inversely
proportionate to their fitness, and does not use crossover operator. CSA introduces the notion
of memory (i.e., high quality solutions are stored for future use, leading to a faster immune
response). Dai et al. (2009) take advantage of the clonal selection mechanism and the learning
capability of the CLONALG. By introducing the multi-parent crossover and neighbourhood
mutation operator, crossover-based CSA achieves a dynamic balance between exploration
and exploitation. An empirical study has shown that the Harmony Search (HS) method usu-
ally suffers from a slow search speed. To overcome this drawback, Wang et al. (2009) present
a hybrid optimization method based on the fusion of the CSA and HS technique. Jie et al.
(2006) and Wang et al. (2008) use Hopfield neural network (NN), whereas Gao et al. (2009)
utilize differential evolution (DE) to improve affinity of CSA. Gao et al. (2008a) use SA
that helps to explore CSA solution space more efficiently. Huang et al. (2009) use CSA to
dynamically select the best training data set for support vector machines throughout training.
The Artificial Fish Swarm Algorithm (AFSA) is used for parameter optimization. CSA helps
to reduce the size of training dataset and training time.
CSA can be considered as a local search algorithm and when combined CSA with swarm
optimization algorithms, such as PSO, ACO, and AFSA, which are global search algorithms,
enables the hybrid algorithm to efficiently search the solution space. The well-known meta-
heuristics, such as GAs, gene expression programming (GEP), simulated annealing (SA),
and NN, are also used with CSA. Besides these, there are also novel CSA hybrid algorithms,
that are considered in others group in this study (i.e., bacterial mutation, leap frog algorithm,
and HS, etc.). Table 2 provides a summary of recent hybrid CSA applications.
The performances of various hybrid CSA approaches are assessed by comparing with
other heuristics. To name a few: Tsai et al. (2007) compare their results with EAs; Gong et al.
(2009) with intelligent evolutionary algorithm (IEA) and CSA; Rabbani AND Panahi (2008)
with ACO, GAs, SA, and tabu search (TS); and Su et al. (2008) with Frequency Sampling

123
Table 2 A summary of recent hybrid CSA characteristics and applications

Reference Algorithm name Main characteristics Problem type

123
Wang et al. (2006a) CLONALG_PSO Combines Function optimization
CLONALG with
PSO
Wang et al. (2006c) PSO with clonal selection strategy and simplified CSA improves the local search ability, and avoids to
quadratic interpolation (SQI) escape from the local optima
Mitra and CSA and small Requires less memory, CSA increases the exploration
Venayagamoor- population based capability and reduces the chances of convergence to
thy PSO (SPPSO) local minima
(2008)
Chen and Guo (2008) CSA-PSO Introduces self-adaptive inertia weight factor in PSO
Gong et al. (2009) Hybrid immune Avoids the decrease of gene diversity, is compared with
algorithm with IEA and CSA
recombination
(HIAR)
Wang et al. (2006b) PSO based on CSA, and mind evolutionary embeds clone function from CSA and MEC in the Nonlinear function optimization
computation (MEC) original PSO
Gao et al. (2009) CSA and DE DE employs to improve the affinities of the clones of the
antibodies (Abs) in the CSA
Tsai et al. (2007) Taguchi-immune Integrates the Taguchi method for the local search to Global numerical
algorithm (MTIA) improve the capabilities of exploration and optimization,
exploitation, is compared with EAs job-shop scheduling
Hong (2009) CSA-based PSO Is based on CSA, avoids premature convergence and multi-modal function
guarantees diversity optimization
Karakasis and CSA gene expression CSA regulates the immune response and GP eliminates Data mining
Stafylopatis (2006) programming (GEP) the genotype-phenotype coincidence
Karakasis and Enhanced clonal Combines
Stafylopatis (2008) algorithm (ECA) CLONALG with
GEP
Fang et al. (2008) Hybrid immune Integrates clonal selection, bone marrow model, ant Feature selection
algorithm (HAIA) colony optimization, and K nearest neighbor classifier
B. Haktanirlar Ulutas, S. Kulturel-Konak
Table 2 continued

Reference Algorithm name Main characteristics Problem type

Nabil et al. (2009) Immuno-Genetic Mutation rate is introduced by simple degrading Binary character
Hybrid Algorithm function and the crossover is merged into the CSA to recognition
increase the exploration and combined with fuzzy
logic
Neggaz and Hybrid models based CSA improves the learning of neural networks, is Speech recognition
Benyettou (2009) on biological compared with Differential Evolution (DE), Multiple
approaches Populations Genetic Algorithms (MPGA)
Rabbani AND Panahi Hybrid immune Uses immune algorithm and bacterial optimization Clustering
(2008) algorithm (HIA) principles, is compared with ACO, GAs, SA, and TS
Li et al. (2009) Clonal selection principle and Fuzzy C-Means Fuzzy clustering is overcomes the disadvantages Classification
clustering (FCM) algorithm CLONALG for classification
Wang et al. (2009) CSA and HS CSA improves the harmony memory members in the HS
Gao et al. (2008a) SA polyclonal Exploits the solution space and explores more TSP
selection algorithm efficiency., the probability of local minimum is
(SAPCSA) reduced by using SA
A review of clonal selection algorithm and its applications

Gao et al. (2008b) Hybrid clonal selection and ant colony optimization Propose a pheromone-linker to combine these CSA and
(CSA-ACO) ACO
Yong and Sunjun (2009) CSA-ant colony Generates pheromone distribution, and ant colony
algorithm for optimal solving
Dai et al. (2009) Crossover-based CSA Points the lack of information exchange, and introduces
an improved quantum crossover operator
Yang et al. (2008b) CSA based memetic Uses CSA for exploration, and SA for local optimization Job shop scheduling
algorithm
Jia et al. (2008) CSA and Receding Infeasibility degree (IFD) handles constraints very well Aircraft landing
Horizon Control and guides the optimization process effectively, and scheduling (ALS)
(RHC) excellent gene segment spread (EGSS) improves the
convergence speed
Naderi et al. (2009) AIA is hybridized Incorporates advanced Scheduling
with SA features of AIA

123
Table 2 continued

Reference Algorithm name Main characteristics Problem type

123
Chang et al. (2009) Hybrid Cooperates with each other to search optimal solution Flow shop scheduling
genetic-immune by searching different objective functions
algorithm (HGIA)
Huang et al. (2009) CSA and AFSA CSA dynamically selects the best training data set for License plate
SVMs throughout training, AFSA optimizes character
parameters recognition (LPCR)
Liaskos and Roper (2008) CSA and GA Improves the Automatic test data
performance of generation
GAs, is compared
with GA
Bhaduri and Bhaduri (2009) Clonal selection based Is compared with Image segmentation
shuffled frog SFLA and K-means
leaping algorithm algorithm
(CSSFLA)
CSA and GA Helps the population Constrained
Bernardino et al. move into the optimization
(2008) feasible region
Vijayalakshmi and Artificial immune CSA handles the Routing in large scale
Radhakrishnan based GA constraints, and networks
(2008) introduces handling
dynamism into GA
Goncalves et al. (2007) Cultural immune Guides the clonal selection principle by sources stored Economic load
system in the belief space of a cultural algorithm dispatch
Jie et al. (2006) CSA and Hopfield Introduces the HNN into the CSA as an “immune Multiuser detection
neural network operator” to further improve the affinity of the
(HNN) antibodies at each generation.
Wang et al. (2008) Hybrid immune CSA Introduces embedded HNN to accelerate the search
(ICSA) convergence and improve local search capability
Su et al. (2008) CSA and frequency Uses the sampled frequencies of transition to optimize Design of 2-D finite
sampling method band of digital filters, is compared with methods such impulsive response
(CSA-FSM) as FSM, GAs, and Chebyshev (FIR) digital filter
B. Haktanirlar Ulutas, S. Kulturel-Konak
A review of clonal selection algorithm and its applications

Method (FSM), GAs and Chebyshev Sampling Method. The results state that the various
versions of CSA perform better than comparing with the aforementioned meta-heuristics to
solve optimization and pattern recognition problems.

4 Conclusions and emerging trends

The early studies on AIS were basically on the theoretical issues. Hart and Timmis (2008)
discuss that AIS algorithms are still worth studying. Recently, researchers have applied AIS
in a wide spectrum of challenging problems. Furthermore, the number of studies where CSA
is used with other meta-heuristics in a hybrid form is increasing. This study gives an overview
of CSA in literature and their application areas. However, due to the extensive number of
studies in literature, this study does not claim to cover all the studies in this area although
it is tried to be comprehensive. Moreover, only the databases that concern computer science
studies were investigated for the scope of this study. Although there may exist a notable num-
ber of CSA based studies written in other languages than English, they are not considered in
this study.
This study considers CSA based algorithms as canonical and hybrid. It is promising that
the studies have an increasing trend since 2002. It is observed that function optimization and
pattern recognition problems are leading problem types for canonical CSA. In this study,
scheduling problems are considered in a separate group. Further, a new problem type group
is introduced named as IE related problems which include NP-hard problems such as man-
ufacturing, facility layout and location that usually deal with the complex, dynamic real life
applications. It is expected that the number of studies that propose the versions of CSA to solve
problems that fall in this group will increase in the near future. For the CSA hybrid algorithms,
the most commonly studied problem type is function optimization followed by classification,
TSP, scheduling, and PR problems. The hybrid studies mainly focus on to improve the per-
formance of the algorithm and consequently obtain better results for the problem in concern.
Current studies focus on optimization, mainly function optimization, problems to present the
performance of the hybrid algorithms. It is also observed that especially swarm optimization
algorithms have successful hybrid applications with CSA. However, the hybrid approaches
that aim to enhance the global convergence performance of CSA have open issues that need
to be studied.

Acknowledgments This research was supported by The Scientific and Technological Research Council of
Turkey (TUBITAK) 2219-International Postdoctoral Research Scholarship Program.

References

Acan A (2004) Clonal selection algorithm with operator multiplicity. In: Congress on Evolutionary Compu-
tation (CEC2004), pp 1909–1915
Akdagli A, Guney K, Babayigit B (2007) Clonal selection algorithm for design of reconfigurable antenna
array with discrete phase shifters. J Electromagnet Waves Appl 21(2):215–227
Bernardino HS, Barbosa HJC, Lemonge ACC, Fonseca LG (2008) A new hybrid AIS-GA for constrained
optimization problems in mechanical engineering. In: IEEE Congress on Evolutionary Computation
(CEC2008), pp 1455–1462
Bhaduri A, Bhaduri A (2009) Color image segmentation using clonal selection-based shuffled frog leaping
algorithm. In: Proceedings of the international conference on advances in recent technologies in com-
munication and computing (ARTCom 2009), pp 517–520

123
B. Haktanirlar Ulutas, S. Kulturel-Konak

Brownlee J (2005) Clonal selection theory and CLONALG the clonal selection classification algorithm
(CSCA). Technical Report 202, Swinburne University of Technology
Brownlee J (2007) Clonal selection algorithms. Technical Report 070209A, Swinburne University of
Technology
Burnet FM (1959) The clonal selection theory of acquired immunity. Cambridge University Press, Cambridge
Campelo F, Guimares FG, Saldanha RR, Igarashi H, Noguchi S, Lowther DA, Ramirez JA (2004) A novel
multiobjective immune algorithm using nondominated sorting. In: 11th International IGTE symposium
on numerical field calculation in electrical engineering, Seggauberg, Austria, pp 308–313
Campelo F, Guimaraes FG, Igarashi H, Ramirez JA (2005) A clonal selection algorithm for optimization in
electromagnetic. IEEE Trans Magnet 41(5):1736–1739
Cao PB, Xiao RB (2007) Assembly planning using a novel immune approach. Int J Adv Manufact Technol
31(7/8):770–782
Chang PC, Huang WH, Ting CJ, Wu LC, Lai CM (2009) A hybrid genetic-immune algorithm with improved
offspring and elitist antigen for flow-shop scheduling problems. In: 11th IEEE international conference
on high performance computing and communications (HPCC2009), pp 591–596
Chen AL, Guo Q (2008) An effective hybrid optimization algorithm based on self-adaptive particle swarm
optimization algorithm and artificial immune clone algorithm. In: 4th International conference on natural
computation (ICNC2008), pp 129–132
Coello CCA, Rivera DC, Cortes NC (2003) Use of an artificial immune based system for job shop scheduling.
In: Timmis J et al (ed) ICARIS 2003. LNCS, vol 2787. Springer, Heidelberg, pp 1–10
Cruz-Cortes N, Trejo-Perez D, CoelloCoello AC (2005) Handling constraints in global optimization using
an artificial immune system. In: Proceedings of the 4th international conference on artificial immune
systems (ICARIS 2005), Banff, pp 234–247
Cutello V, Nicosia G (2006) A clonal selection algorithm for coloring, hitting set and satisfiability problems. In:
Workshop on natural and artificial immune systems, Revised Selected Papers. LNCS, vol 3931. Springer,
Heidelberg, pp 324–337
Cutello V, Narzisi G, Nicosia G, Pavone M (2005) Clonal selection algorithms: a comparative case study using
effective mutation potentials. In: Proceedings artificial immune systems: 4th international conference,
ICARIS 2005, Banff, pp 13–28
Cutello V, Nicosia G, Romeo M, Oliveto PS (2007) On the convergence of immune algorithms. In: Proceedings
of the 1st IEEE symposium on foundations of computational intelligence (FOCI 2007)
Dai H, Li C (2009) Chaotic immune clonal selection algorithm and its application. In: Proceedings of the 2nd
international symposium on knowledge acquisition and modeling (KAM 2009), pp 65–68
Dai H, Yang Y, Li C, Shi J, Gao S, Tang Z (2009) Quantum interference crossover-based clonal selection
algorithm and its application to traveling salesman problem. IEICE Trans Inf Syst 1:78–85
Darmoul S, Pierreval H, Gabouj SH (2006) Scheduling using artificial immune system metaphors: a review. In:
2006 International conference on service systems and service management (IEEE Cat. No. 06EX1406)
Dasgupta D (1998) An overview of artificial immune systems and their applications. Springer, Berlin
Dasgupta D (2006) Advances in artificial immune systems. IEEE Comput Intell Mag 40–49
Dasgupta D, Nino LP (2008) Immunological Computation: Theory and Applications. CRC Press,
Dasgupta D, Azeem R, Balachandran S, Yu S, Majumdar N, Nino F (2007) Artificial immune systems: a
bibliograph. Computer Science Department the University of Memphis, USA CS Technical Report No.
CS-07-004
Dashora Y, Tiwari MK, Karunakaran KP (2008) A psycho-clonal-algorithm-based approach to solve operation
sequencing problem in a CAPP environment. Int J Comput Integrat Manufact 21(5):510–525
De Castro LN (2002a) Immune, swarm, and evolutionary algorithms. Part I: basic models. In: Proceedings of
the 9th international conference on neural information (ICONIP 2002), vol 3, pp 1464–1468
De Castro LN (2002b) Immune, swarm, and evolutionary algorithms part II: philosophical comparisons.
In: Proceedings of the international conference on neural information processing (ICONIP), vol 3, pp
1469–1473
De Castro LN, Von Zuben FJ (1999) Artificial immune systems: part I-basic theory and applications. Technical
Report, TR-DCA 01/99
De Castro LN, Von Zuben FJ (2000a) Artificial immune systems: part II-a survey of applications. Technical
Report, DCA-RT 02/00
De Castro LN, Von Zuben FJ (2000b) The clonal selection algorithm with engineering applications. In:
Proceedings of genetic and evolutionary computation conference (GECCO), Las Vegas
De Castro LN, Timmis J (2002) Artificial immune systems: a new computational intelligence approach.
Springer, Heidelberg
De Castro LN, Von Zuben FJ (2002) Learning and optimization using the clonal selection principle. IEEE
Trans Evolut Comput 6(3):239–251

123
A review of clonal selection algorithm and its applications

Dilettoso E, Salerno N (2006) A self-adaptive niching genetic algorithm for multimodal optimization of elec-
tromagnetic devices. IEEE Trans Magnet 42:1203–1206
Ding S, Li SX (2009) Clonal selection algorithm for feature selection and parameters optimization of support
vector machines. In: Proceedings of the 2nd international symposium on knowledge acquisition and
modeling (KAM 2009), pp 17–20
Dong W, Shi G, Zhang L (2007) Immune memory clonal selection algorithms for designing stack filters.
Connection Neurocomput 70:777–784
Du HF, Jiao LC, Wang SA (2002) Clonal operator and antibody clone algorithms. In: Proceedings of 2002
international conference on machine learning and cybernetics (Cat.No.02EX583), vol 1, pp 506–510
Engin O, Doyen A (2004) A new approach to solve hybrid flow shop scheduling problems by artificial immune
system. Future Generat Comput Syst 20:1083–1095
Fang T, Fu D, Zhao Y (2008) A hybrid artificial immune algorithm for feature selection of ovarian cancer data.
In: 2008 International workshop on education technology and training & 2008 international workshop
on geoscience and remote sensing (ETT and GRS), pp 681–685
Gao S, Tang Z, Dai H, Zhang J (2008a) A hybrid clonal selection algorithm. Int J Innovat Comput Inf Control
4(4):995–1008
Gao S, Wang W, Dai H, Li F, Tang Z (2008b) Improved clonal selection algorithm combined with ant colony
optimization. IEICE Trans Inf Syst E91-D(6):1813–1823
Gao XZ, Wang X, Ovaska SJ (2009) Fusion of clonal selection algorithm and differential evolution method
in training cascade-correlation neural network. Neurocomputing 72(10–12):2483–2490
Garain U, Chakraborty MP, Dasgupta D (2006) Recognition of handwritten indict script using clonal selec-
tion algorithm. In: 5th International conference, ICARIS 2006. LNCS, vol 4163. Springer, Heidelberg,
pp 256–266
Garrett SM (2004) Parameter-free, adaptive clonal selection. In: Proceedings of congress on evolutionary
computing (CEC 2004), Portland Oregon, pp 1052–1058
Garrett SM (2005) How do we evaluate artificial immune systems?. Evolut Comput 13(2):145–177
Goncalves RA, de Almeida CP, Delgado MR, Goldbarg EF, Goldbarg MC (2007) A cultural immune system
for economic load dispatch with non-smooth cost functions. In: Proceedings 6th international conference
(ICARIS 2007). LNCS, vol 4628. Springer, Heidelberg, pp 382–394
Gong M, Jiao L, Ma W, Shang R (2009) Hybrid immune algorithm with intelligent recombination. In: IEEE
congress on evolutionary computation, pp 1807–1814
Gong M, Jiao L, Yang J, Liu F (2010a) Lamarckian learning in clonal selection algorithm for numerical
optimization. Int J Artif Intell Tools 19(1):19–37
Gong M, Jiao L, Zhang L (2010b) Baldwinian learning in clonal selection algorithm for optimization. Inf Sci
180:1218–1236
Hart E, Timmis J (2008) Application areas of AIS: the past, the present and the future. Appl Soft Comput
8(1):191–201
Hofmeyr SA, Forrest S (2000) Architecture for an artificial immune system. Evolut Comput 8(4):443–473
Hong L (2008) On the convergence rates of clonal selection algorithm. In: 2008 International symposium on
information science and engineering
Hong L (2009) A novel particle swarm optimization method using clonal selection algorithm. In: 2009 Inter-
national conference on measuring technology and mechatronics automation (ICMTMA), pp 471–474
Huang R, Tawfik H, Nagar AK (2009) Licence plate character recognition based on support vector machines
with clonal selection and fish swarm algorithms. In: Proceedings of 11th international conference on
computer modeling and simulation, pp 101–106
Jia X, Cao X, Guo Y, Oiao H, Zhang J (2008) Scheduling aircraft landing based on clonal selection algorithm
and receding horizon control. In: 2008 11th International IEEE conference on intelligent transportation
systems, pp 357–362
Jie M, Gao HY, Ming D (2006) Multiuser detection using the clonal selection algorithm and hopfield neural
network. In: 2006 International conference on communications, circuits and systems (IEEE Cat. No.
06EX1237)
Karakasis VK, Stafylopatis A (2006) Data mining based on gene expression programming and clonal selection.
In: IEEE congress on evolutionary computation (CEC 2006), pp 514–521
Karakasis VK, Stafylopatis A (2008) Efficient evolution of accurate classification rules using a combination of
gene expression programming and clonal selection. In: IEEE transactions on evolutionary computation,
vol 12(6), pp 662–678
Kelsey J, Timmis J (2003) Immune inspired somatic contiguous hypermutation for function optimisation.
In: Proceedings, part I genetic and evolutionary computation conference (GECCO 2003), Chicago,
pp 207–218

123
B. Haktanirlar Ulutas, S. Kulturel-Konak

Khilwani N, Prakash A, Shankar R, Tiwari MK (2008) Fast clonal algorithm. Eng Appl Artif Intell 21(1):
106–128
Kim J, Bentley PJ (2002) Towards an artificial immune system for network intrusion detection: an investi-
gation of dynamic clonal selection. In: Congress on evolutionary computation (CEC-2002), Honolulu,
pp 1015–1020
Kumar RMS, Asokan P, Kumanan S (2009) Artificial immune system-based algorithm for the unidirectional
loop layout problem in a flexible manufacturing system. Int J Adv Manufact Technol 40:553–565
Li D, Chen Z (2009) SVM optimized by immune clonal selection algorithm for fault diagnostics.
In: Proceedings Pacific-Asia conference on circuits, communications and systems (PACCS 2009),
pp 702–705
Li G, Zhuang J, Hou H, Yu D (2009) An improved clonal selection classifier incorparating fuzzy clustering.
In: International conference on measuring technology and mechatronics automation (ICMTMA 2009),
pp 179–182
Li JH, Gao HW, Wang S (2008) A novel clone selection algorithm with reconfigurable search space ability
and its application. In: 2008 4th International conference on natural computation, pp 612–616
Liaskos K, Roper M (2008) Hybridizing evolutionary testing with artificial immune systems and local
search. In: IEEE international conference on software testing verification and validation workshop
(ICSTW2008), pp 211–220
Litvinenko VI, Bidyuk PI, Bardachov JN, Sherstjuk VG, Fefelov AA (2005) Combining clonal selection algo-
rithm and gene expression programming for time series prediction. In: Proceedings of the 3rd workshop
2005 IEEE intelligent data acquisition and advanced computing systems: technology and applications,
pp 133–138
Liu X, Shi L, Chen R, Chen H (2009) A novel clonal selection algorithm for global optimization problems.
In: International conference on information engineering and computer science (ICIECS 2009)
Lu H, Yang J (2009) An improved clonal selection algorithm for job shop scheduling. In: International sym-
posium on intelligent ubiquitous computing and education (IUCE2009), pp 34–37
Ma J, Gao L, Shi G (2009) An improved immune clonal selection algorithm and its applications for VRP. In:
IEEE international conference on automation and logistics (ICAL2009), pp 2097–2100
Ma W, Shang R, Jiao L (2006) A novel clonal selection algorithm for face detection. In: 19th Conference
on artificial intelligence, advances in artificial intelligence (AI 2006). Proceedings, Lecture Notes in
Artificial Intelligence, vol 4304, pp 799–807
Mak KL, Lau PSK (2007) An age artificial immune system for order pickings in an AS/RS with Multiple I/O
Stations. In: World congress on engineering 2007, pp 1050–1055
Mitra P, Venayagamoorthy GK (2008) Empirical study of a hybrid algorithm based on clonal selection and
small population based PSO. In: 2008 IEEE Swarm Intelligence Symposium
Moghaddam MZ, Kardan A (2009) Clonal selection algorithm for partitioning and scheduling of code-
sign systems. In: 5th International colloquium on signal processing & its applications (CSPA 2009),
pp 267–272
Nabil E, Badr A, Farag I (2009) An immuno-genetic hybrid algorithm. Int J Comput Commun Control
4(4):374–385
Naderi B, Khalili M, Tavakkoli-Moghaddam M (2009) A hybrid artificial immune algorithm for a realistic
variant of job shops to minimize the total completion time. Comput Ind Eng 56:1494–1501
Nanas N, De Roeck A (2007) Multimodal dynamic optimization: from evolutionary algorithms to artificial
immune systems. In: De Castro LN, Von Zuben FJ, Knidel H (eds) ICARIS 2007. LNCS, vol 4628.
Springer, Heidelberg, pp 13–24
Neggaz N, Benyettou A (2009) Hybrid models based on biological approaches for speech recognition. Artif
Intell Rev 32:45–57
Ong ZX, Tay JC, Kwoh CK (2005) Applying the clonal selection principle to find flexible job-shop schedules.
In: International conference of artificial immune systems (ICARIS 2005). LNCS, vol 3627. Springer,
Heidelberg, pp 442–455
Pang XL, Feng YQ (2006) Solving competitive facilities location problem with the clonal selection algorithm.
In: 2006 International conference on management science and engineering (IEEE Cat. No. 06EX1529),
vol 1, pp 413–417
Pan X, Liu F, Jiao L (2006) A dynamic clonal selection algorithm for project optimization scheduling. In: 6th
International conference (SEAL 2006). LNCS, vol 4247, pp 821–828
Panigrahi BK, Yadav SR, Agrawal S, Tiwari MK (2007) A clonal algorithm to solve economic load dispatch.
Electr Power Syst Res 77:1381–1389
Rabbani M, Panahi H (2008) An efficient hybrid artificial immune algorithm for clustering. In: 8th International
conference on hybrid intelligent systems (HIS 2008), pp 374–379

123
A review of clonal selection algorithm and its applications

Ramaswamy SAP, Venayagamoorthy GK, Balakrishnan SN (2007) Optimal control of class of non-linear
plants using artificial immune systems. In: 22nd IEEE international symposium on intelligent control
(ISIC 2007), pp 249–254
Shi G, Jing Y (2009) Research of improved immune clonal algorithms and its applications. In: International
conference on computational intelligence for measurement systems and applications (CIMSA) Hong
Kong, China
Su Tj, Kuo CH, Tsai WP, Hou CC (2008) A hybrid of clonal selection algorithm and frequency sampling
method for designing A 2D FIR Filter. In: 4th IEEE international symposium on electronic design, test
and applications (DELTA 2008), pp 274–278
Timmis J (2007) Artificial immune systems-today and tomorrow. Nat Comput 6:1–18
Timmis J, Hone A, Stibor T, Clark E (2008) Theoretical advances in artificial immune systems. Theor Comput
Sci 403:111–132
Tiwari MK, Prakash Kumar A, Mileham AR (2005) Determination of an optimal assembly sequence using
the PsychoCA. J Eng Manufact 219:137–149
Tsai JT, Ho WH, Liu TK, Chou JH (2007) Improved immune algorithm for global numerical optimization
and job-shop scheduling problems. Appl Math Comput 194(2):406–424
Ulutas BH, Islier AA (2009) A clonal selection algorithm for dynamic facility layout problems. J Manufact
Syst 28:123–131
Ulutas BH, Islier AA (2010) A novel attribute-based dynamic content area layout for internet newspapers.
Internet Res 20(2):135–153
Vijayalakshmi K, Radhakrishnan S (2008) Artificial immune based hybrid GA for Qos based multicast routing
in large scale networks (AISMR). Comput Commun 31(17):3984–3994
Wang X (2005) Clonal selection algorithm in power filter optimization. In: Proceedings of the 2005 IEEE
mid-summer workshop on soft computing in industrial applications, pp 122–127
Wang J, Zhang X, Jiao L (2006a) Integrated the simplified interpolation and clonal selection into the parti-
cle swarm optimization for optimization problems. In: Proceedings of 6th international conference on
simulated evolution and learning (SEAL 2006). LNCS, vol 4247. Springer, Heidelberg, pp 433–440
Wang O, Wang C, Gao XZ (2006b) A hybrid optimization algorithm based on clonal selection principle and
particle swarm intelligence. In: 6th International conference on intelligent systems design and applica-
tions (ISDA 2006), pp 975–979
Wang X, Gao XZ, Ovaska SJ (2006c) A hybrid particle swarm optimization method. In: 2006 IEEE conference
on systems, man, and cybernetics, vol 4151(7), pp 8–11
Wang A, Xu B, Zhou R, He Z (2008) Near-optimal MIMO multiuser detection using hybrid immune clonal
selection algorithm. In: 3rd International conference on communications and networking in China (CHI-
NACOM2008), pp 983–989
Wang X, Gao XZ, Ovaska SJ (2009) Fusion of clonal selection algorithm and harmony search method in
optimization of fuzzy classification systems. Int J Bio-Inspired Comput 1(1–2):80–88
Watkins A, Bi X, Phadke A (2003) Parallelizing an immune-inspired algorithm for efficient pattern recogni-
tion, intelligent engineering systems through artificial neural networks: smart engineering system design.
Neural Networks, pp 225–230
Watkins A, Timmis J, Boggess L (2004) Artificial immune recognition system (AIRS): an immune-inspired
supervised learning algorithm. Gen Program Evol Mach 5:291–317
White J, Garrett SM (2003) Improved pattern recognition with artificial clonal selection. In: Proceedings of
ICARIS2003, Edinburgh, pp 181–193
Wong EYC, Lau HKY (2009) Advancement in the twentieth century in artificial immune systems for optimi-
zation: review and future outlook. In: Proceedings of the 2009 IEEE international conference on systems,
man, and cybernetics, pp 4195–4202
Wu Q, Jiao L, Li Y (2009) Design RF diplexer by directional immune clonal selection algorithm. In: Pro-
ceedings of the 2009 IEEE international conference on systems, man and cybernetics. SMC 2009, pp
4214–4217
Yang J, Gong M, Jiao L, Zhang L (2008a) Improved clonal selection algorithm based on Lamarckian local
search technique. In: IEEE congress on evolutionary computation (CEC 2008), pp 535–541
Yang J, Sun L, Lee HP, Qian Y, Liang Y (2008b) Clonal selection based memetic algorithm for job shop
scheduling problems. J Bion Eng 5:111–119
Yong L, Sunjun L (2009) A hybrid model for solving TSP based on artificial immune and ant colony. In:
International conference on information engineering and computer science (ICIECS 2009)
Yu Y, Hou CZ (2004) A clonal selection algorithm by using learning operator. In: Proceedings of international
conference on machine learning and cybernetics, pp 2924–2929
Zhang Y, Hou C (2003) A clone selection algorithm with niching strategy inspiring by biological immune
principles for change detection. In: IEEE international symposium on intelligent control, pp 1000–1005

123
B. Haktanirlar Ulutas, S. Kulturel-Konak

Zhao MY, Tang K, Lu G, Zhou MT, Fu C, Yang F, Zhang CG (2008) A novel clonal selection algorithm and its
application. In: International conference on apperceiving computing and intelligence analysis (ICACIA
2008), pp 385–388
Zheng J, Chen Y, Zhang W (2010) A survey of artificial immune applications. Artif Intell Rev, doi:10.1007/
s10462-010-9159-9
Zhu Y, Gao S, Dai H, Li F, Tang Z (2007) Improved clonal algorithm and its application to traveling salesman
problem. Int J Comput Sci Network Secur 7(8):109–113
Zuo XO, Fan YS (2003) The Chaos artificial immune algorithm and its application to rbf neuro-fuzzy controller
design. In: IEEE international conference on systems, man and cybernetics, pp 2809–2814

123

You might also like