A Survey on Nature Inspired Meta-Heuristic

Algorithms with its Domain Specifications

Rajakumar. R Dhavachelvan. P Vengattaraman. T
Research Scholar Professor Assistant Professor
Department of CSE Department of CSE Department of CSE
Pondicherry University Pondicherry University Pondicherry University
ramukshare@gmail.com dhavachelvan@gmail.com vengattaraman.t@gmail.com

Abstract—Nature inspired metaheuristics algorithms are well solution that is “good enough” in a computing time that is
known economical approaches for solving several hard “small enough”. Meta-heuristic optimization algorithms are
optimization problems. This paper furnishes a survey of some of aids to solve wide range of real-time problems due to its (i)
the main metaheuristics. It provides the components and simplicity and easy to implement; (ii) does not need slope
concepts that are employed in these algorithms so as to research
information; (iii) avoid local optima; (iv) can be exploited in
their similarities and variations. The classification adopted in this
paper differentiates by the behaviors obtained to develop the an ample range of problems wrapping different disciplines [2].
wide variety of nature inspired algorithms. The literature survey Unique feature of Meta-heuristics algorithm is different
is guided by the presentation of control parameters, methods of search process. These algorithms most uniformly
intensification, and diversification used in these algorithms and contribute in two phases: intensification and diversification
its domain specifications. [3]. Intensification phase search process begins to find the
Keywords— metaheuristics; control parameters; intensification; local best solution within its adjacent location however this
diversification; domain specifications. process otherwise called as local search. Diversification phase
start the search process globally in the provided search space
I. INTRODUCTION which intend to attain the global solution however this process
In recent years, optimization is a booming research area also called as global search. Most challenging task in the
for providing an optimal solution to complex real-time development of any meta-heuristic algorithm is to find a
problems. Multi-dimensionality, multi-modality, multi- suitable balance the intensification and diversification.
objective, differentiability and different combinatorial The rest of the paper is structured as follows. Section 2
characteristics are coped with these problems. The demand for describes the detail survey on some popular meta-heuristic
solving real-time problems has attracted many researchers to algorithms in this study. Nature inspired meta-heuristics
develop fast, accurate and computationally powerful algorithm with their application are presented and discussed in
optimization algorithms. Researchers from various domains Sections 3 and 4, respectively, for concludes this study and
have introduced many numerical optimization techniques to suggests directions for future research.
attain better solution for these problems. Historical problem-
solving techniques are classified into two techniques: Extract II. CLASSIFICATION OF NATURE INSPIRED META-
and Heuristics methods. Logical and mathematical HEURISTICS ALGORITHMS
programming are involved in Extract methods to solve NP-
Nature inspired meta-heuristic algorithm are classified
complete problems whereas heuristics method seems to be
into four major divisions as shown in Fig.1:
superior in solving NP-hard and complex optimization
x Evolution- based method
problems, specifically where the traditional methods fail [1].
Any advance to problem-solving, learning or discovery which x Physics-based method
spotlight on immediate near optimality rather than exact x Swarm-based method and
results, using realistic methods can be described as a x Human-based method
heuristic.The present trend to use heuristic techniques over A. Evolution Based Meta-Heuristics Algorithms
precise ones is thanks to proven fact that several time period Evolution-based methods [2, 4] are inspired by the laws
issues are shown to stay forever wild to exact algorithms, of natural evolution. Initially the set of population are
regardless of the ever increasing computational power, merely generated stochastically then it begins search process over
thanks to unrealistically massive running times. Nature subsequent generations. The best individuals are collected in
inspired meta-heuristic algorithms mentions to high-level every generation and it taken to the next generation process
heuristics that mimics the biological or physical phenomena. likewise the process goes on until it reaches the termination
Meta-heuristics are refined scientifically to find anoptimal criteria or the optimal solution is obtained. Genetic algorithm
(GA) is most popular evolution-inspired technique that Eberhart developed PSO to solve real-time problems by
imitated by the principles of Charles Darwin Theory of pertaining the best solution identification in a given search
survival of the fittest [5]. This method involves base process space. In PSO, Individual is considered as particles which
of selection, crossover and mutation to replace the worst
search around the search space to find the best solution.
solution in every generation.GA begins by initializing a
population of solution (chromosome). Then each individual Cognitive and social parameters are aids to explore the
evaluates the fitness using an appropriate objective function exploitation (local search) and exploration (global search) in a
for the problem. The best individual is selected into the search space. Ant colony optimization (ACO) [10] is
mating pool, where they undergo cross over and mutation to considered as another well-known Swarm based algorithm.
produce a new set of solution (offspring). Memetic algorithm Vast variety of optimization algorithms are introduced in
(MA) [6] is another evolution based algorithm that mimics the Swarm Intelligence. Still swarm based approach attracts many
behavior of GA and this algorithm improve the worst solution
researchers to develop effective algorithms for different
in each generation based on its probability ratio. Evolution
Strategy (ES), Genetic programming, Biogeography-Based engineering applications. Bacterial foraging behavior (BFO),
Optimizer (BBO) are the other popular algorithms. Virulence Cuckoo Search (CS) algorithm, Krill Herd (KH) algorithm,
optimization algorithm (VAO) [7] are newly developed Dolphin Optimization Algorithm (DOA), Shuffled Frog
evolutionary algorithm. Leaping Algorithm (SFLA), Artificial Bee Colony (ABC)
algorithm, Dragon Flies (DF) algorithm, Bat Algorithm (BA)
B.Physics Based Meta-Heuristics Algorithms are other swarm based algorithms. Whale optimization
Physics-based methodsmimic the physical rules in the algorithm (WOA), Ageist Spider Monkey Optimization
universe [2]. Simulated Annealing (SA) [8]models the (ASMO) [11], Lions Algorithm (LA) [12] are newly
physical process of warming a material and then gradually introduced algorithms.
decreasing the temperature to decrease defects, thus reducing
the system energy.Simulated annealing introduces just the D. Human Based Meta-Heuristics Algorithms
right amount of unpredictability into things to getaway local Human based methods inspired by the advancement in
maxima early in the process without getting off course late in level of searching strategy [2]. Rao et al [13] proposed an
the game, when a solution is in close proximity. This algorithm named as Teaching-Learning-Based Optimization
algorithm better suits in identifying optimal solution, and (TLBO), which the behavior of traditional teaching-learning
never consider the strategy for starting point.Big-Bang Big- phenomenon of a class room. Some of the other most popular
Crunch (BBBC), Gravitational Local Search (GLSA), algorithms are as Tabu (Taboo) Search (TS) , Harmony
Gravitational Search Algorithm (GSA), Central Force Search (HS), Group Search Optimizer (GSO), Imperialist
optimization (CFO), Black Hole (BH) algorithm, Artificial Competitive Algorithm (ICA) [14], Firework Algorithm,
Chemical Reaction Optimization Algorithm (ACROA), League Championship Algorithm (LCA), Interior Search
Galaxy-Based Search Algorithm (GbSA), Ray Optimization Algorithm (ISA), Colliding Bodies Optimization (CBO), Mine
(RO) algorithm, Charged System Search (CSS), Small-World Blast Algorithm (MBA), Soccer League Competition (SLC)
Optimization Algorithm (SWOA),and Curved Space algorithm, Exchange Market Algorithm (EMA) [15] , Seeker
Optimization (CSO) are the most popular algorithms. Optimization Algorithm (SOA), Social-Based Algorithm
(SBA), and Group Counseling Optimization (GCO) algorithm.
C. Swarm Based Meta-Heuristics Algorithms In such a way, metaheuristics algorithms are divided in
Swarm-based method mimics the social behavior of such which try to find the best optimum values of objective
groups of animals [2]. One of the most popular algorithms is functions for problem specific. The following section
Particle Swarm Optimization (PSO) [9], which mimics the describes these metaheuristics algorithms based on its use of
behavior of fish schooling and birds flocking. Kennedy and domain applications.
 Fig.1. Classification on Nature Inspired Metaheuristics Algorithms

III. META-HEURISTICS ALGORITHMS WITH behavior of physical rules in the universe. Swarm
ITS DOMAIN SPECIFICATION based methods mimics the behavior of group of
animals whereas human based methods inspired by
In the last decade, nature inspired metaheuristics the self-learning. Generally, almost of all the methods
algorithms are emerging as viable tools and performs with heuristics population-based search
alternatives to more traditional real-time applications methodologies that integrate stochastic diversity and
[4]. Among the many metaheuristics algorithm, some selection. It has been endorsed that the growth of
of the main algorithms are tabulated with their meta-heuristics and applications in the past years is
developers, control parameters, domain very extreme and has been practiced to various
specifications, intensification and diversification. problems. Further this survey can extend to freshly
Researchers are developed their metaheuristics introduce meta-heuristic hybrid algorithms with their
algorithm with two different motives such as problem modifications (in improving diversity and
specific algorithms and generic algorithms with intensifications) and its potency in real world
improvising intensification (local search) and applications.
diversification (global search) in search space as
shown in table 1. The intensification and REFERENCES
diversification factors is measured based on the [1] Binitha S, S Siva Sathya, “A Survey of Bio inspired
Optimization Algorithms”, International Journal of Soft
control parameter modifications in the entire
Computing and Engineering (IJSCE), Vol. 2, Issue. 2, pp.
discussed algorithm. 137-151, May, 2012.
[2] Seyedali Mirjalili, Andrew Lewis, “Whale Optimization
IV. CONCLUSION AND FUTURE WORK Algorithm”, Journal on Advances in Engineering Software,
This work reviewed several important nature vol. 25, pp. 51-67, February, 2016.
inspired meta-heuristics algorithms closely young [3] Ilhem Boussaïd, Julien Lepagnot, Patrick Siarry, “A survey
on optimization metaheuristics”, Journal on Information
and emerging with new ideas and applications. We
Sciences, vol, 237, pp. 82-117, March, 2013.
classified this emerging research area into four [4] Arpan Kumar Kar, “Bio Inspired Computing – A Review of
divisions based on its behaviors. Evolution based Algorithms and Scope of Applications”, Expert Systems with
methods purely mimics the behavior of biological Applications, vo. 59, pp. 20-32, April, 2016.
[5] Holland JH, “Genetic algorithms”, Sci. Am., vol. 267, pp.
evolution however physics based methods mimics the
66–72, 1992.
[6] Dawkins R. “The selfish gene”, Oxford: Oxford University [12] Rajakumar B, “The Lion's Algorithm: A New Nature-
Press; 1976. Inspired Search Algorithm”, Procedia Technology, vol. 6, pp.
[7] Morteza Jaderyan, Hassan Khotanlou, “Virulence 126-135, 2012.
Optimization Algorithm”, Applied Soft Computing, Vol. 43, [13] Rao RV, Savsani VJ, Vakharia DP, “Teaching–learning-
pp. 596-618, 2016. based optimization: an optimization method for continuous
[8] Kirkpatrick S, Gelatt CD, Vecchi MP, “Optimization by non-linear large scale problems”, Inf. Sci., vol. 183, pp. 1–15,
simmulated annealing”, Science, vol. 220, pp. 671–80, 1983. 2012.
[9] Kennedy J, Eberhart R, “Particle swarm optimization”, In: [14] Atashpaz-Gargari E, Lucas C, “Imperialist competitive
Proceedings of the 1995 IEEE international conference on algorithm: an algorithm for optimization inspired by
neural networks, pp. 1942–8, 1995. imperialistic competition” In: Proceedings of the 2007 IEEE
[10] Dorigo M, Birattari M, Stutzle T, “Ant colony optimization”, congress on evolutionary computation, CEC, pp. 4661–7,
IEEE Comput. Intell., vo.1, pp.28–39, 2006. 2007.
[11] Avinash Sharma, Akshay Sharma, B.K. Panigrahi, Deep [15] Ghorbani N, Babaei E, “Exchange market algorithm”, Appl.
Kiran, Rajesh Kumar, “Ageist Spider Monkey Optimization Soft. Comput., vol. 19, pp. 177–87, 2014.
algorithm”, Swarm and Evolutionary Computation, vol. 28,
pp. 58-77, February, 2016.
 Name of Control Intensificat Diversificatio
S.No Classification Publication Domain Specifications
Algorithms Parameters ion n
Population Size
(Chromosome), Engineering, Computer Science,
Genetic Holland J
Selection, Mathematics, Physics and Astronomy, High High
Algorithm (1975)
Crossover, Materials Science
Mutation, Elitism
Population Size
(Chromosome), Mathematics, Engineering, Computer
Memetic Dawkins R
Selection, Science, Materials Science, Physics and Very High Medium
Algorithm (1976)
Crossover, Astronomy
Mutation, Elitism
Population Size
Rechenberg (chromosome), Computer Science, Engineering,
Evolution
I Selection and Mathematics, Physics and Astronomy, Medium High
Strategy (ES)
(1977) Recombination, Materials Science
Mutation
Population Size
Evolution (chromosome),
1
Based Crossover, computer networks, control systems,
Genetic
Koza J R mutation, bioinformatics, data mining, game theory,
programming High Medium
(1992) permutation, biometrics, power systems, image
(GP)
Editing, processing, industry and engineering
Encapsulation,
Decimation
Population Size
Biogeography- (Habitat),
Engineering, Embedded Systems,
Based Simon D migration
Computer Science, Mathematics, and Very High High
Optimizer (2008) (emigration and
Physics
(BBO) immigration) ,
mutation
Scheduling heuristics, Data mining and
Virulence Population Size
Morteza rule extraction, Assignment and
optimization (Virus and host
Jaderyan et management, Bioinformatics, Control High Very High
algorithm cells), Drifting
al (2016) systems and robotics, Pattern recognition
(VAO) and shifting
and image processing
Population Size
Kennedy, Computer science, Engineering,
(particles), particle
PSO Eberhart Mathematics, Energy, Physics and High High
velocity, inertia
(1995) Astronomy
weight
Population Size
Dorigo et. (ants), pheromone Computer science, Engineering,
ACO Al evaporation rate, Mathematics, Decision Sciences, Social High Very High
(1996) pheromone reward Sciences
factor
Population Size
(frogs), number of
memeplexes,
Eusuff MM Engineering, Computer Science,
SFLA maximum step High High
et al (2003) Mathematics, Energy, Social Science
size, number of
shuffling
2 Swarm Based
iterations
Population size
Karaboga (Bees), Engineering, Computer Science, Neural
Artificial Bee
and Basturk Reproduction, Network, Mathematics, Physics and Medium High
Colony (ABC)
(2007) replacement of Astronomy
bee, selection
Seyedali Population Size
Mirjalili, (Whales), Mechanical, Engineering, Computer
Whale
Andrew Encircling prey, Science, Neural Network and Wireless High High
Optimization
Lewis Bubble-net, Sensor Network
(2016) Search for prey
Population Size
Ageist Spider Avinash Computer Science, Engineering,
(Spider Monkey),
Monkey Sharma et Mathematics, Decision Sciences, Physics Very High High
Local Leader
Algorithm al (2016) and Astronomy
Selection,
TABL
E I.
META
HEUR
ISTIC
ALGO Kirkpatrick Population Size
Simulated Mathematics, Computer Science,
RITH Annealing (SA)
S et al., (state) , Selection,
Engineering, Physics and Astronomy
Medium Medium
(1983) Permutation,
MS
Population Size
WITH Gravitational
Rashedi E (agents), Total Physics and Astronomy, Engineering,
Search
THEI et al Force, Mathematics, Computer Science, Neural High Medium
Algorithm
(2009) acceleration and Network, Power systems
R (GSA)
velocity
CHAR Ray
A. Kaveh, , Population Size
M. (particles), Mathematics, Computer Science,
ACTE Optimization
Khayatazad velocity, cockshy, Engineering design problems, Physics
Medium High
(RO)
RISTI (2012) motion refinement
CS, Population Size
3 Physics Based
(charged particle), Physics and Astronomy, Mathematics,
AND Charged Kaveh A,
acceleration Computer Science, Mechanical,
System Search Talatahari High Medium
ITS coefficient, Engineering Design problems, Power
(CSS) (2010)
Velocity, Systems
DOM Probability
AIN Small-World Population size
Computer networks, Bio informatics,
Optimization Haifeng Du (node), bit
SPECI Algorithm et al (2006) oriented adding,
Engineering design problems, High High
Mathematics, Physics
FICAT (SWOA) minus, multiply
IONS Galaxy-Based Population Size,
Shah- Computer science, Mathematics,
search spiral movement
Hosseini, Engineering, Physics and Astronomy High Medium
Algorithm vector, spiral
H. (2011) (Continuous optimization problems)
(GbSA) galaxy size
B. Rao RV
Teaching- Population Size
, Savsani Mathematics, Engineering Design
Learning-Based (Marks), teacher
VJ , problems, Mechanical, Computer Medium High
Optimization phase, learner
Vakharia Science, Networks, Image processing
(TLBO) phase, difference
DP (2011)
Esmaeil
Imperialist Population size
Atashpaz-
Competitive (colonies, Engineering, Computer Science,
Gargari, Medium Medium
Algorithm Imperialists), Mathematics, Decision Sciences
Caro Lucas
(ICA) Probability
(2007)
League Population
Computer science, Engineering,
Championship Kashan AH (teams),
Mathematics, Engineering Design Medium Medium
Algorithm (2009) formation/Playing,
problems, Bio informatics
4 Human Based (LCA) number of seasons
Mine Blast Ali Population Size,
Engineering Design Problems, Computer
Algorithm Sadollah et exploration factor, Medium High
Science
(MBA) al (2013) shrapnel pieces
Naser Population Size
Exchange
Ghorbani, (stock),
Market Engineering, computer science, Bio
Ebrahim Shareholder costs, Medium High
Algorithm informatics, Mechanical, Physics
Babaei oscillation market,
(EMA)
(2014) balance market
Population
Seeker
Chaohua (Seeeker), search
Optimization Computer science, Artificial Intelligent,
Dai et al direction, search Medium Medium
Algorithm Bio informatics, Engineering
(2007) radius, trust
(SOA)
degree