Social Behaviour Inspired Optimization Algorithm: An Approach for Solving Complex Optimization Problems

Author Name(s): Priya Chandel, Prashant Borkar
Author Email: priyachandel.263@gmail.com

Abstract

Inspired from the human behavior an advanced Social Behavior Inspired Optimization rule is given in this research work. This rule works in two phases inspired from human behavior in phase one candidate solution is improved considering the good behavior and by avoiding the bad behavior and in second phase solution is improved by knowledge transfer among the candidates. This algorithmic rule does not depend on any algorithm-specific controlling parameters and needs only common controlling parameters. The execution of Social behavior impressed optimization algorithm is tested on standard benchmark functions having different characteristics and compared with traditional optimization algorithms such as GA, DE, PSO, ABC, Jaya and also with TLBO algorithm.

 

Keywords: Social Behavior, Constrained Optimization, Unconstrained Optimization, Benchmark Function.

 

Introduction

An extensive variety of optimization rules are there in swarm and evolutionary computation like PSO, GA, ACO, HS, ABC, etc. are a few of them [16]. Swarm intelligence models the collective behavior of animals or swarms of insects and has attracted the attention of researchers.

Optimization is the act of getting the most effective result underneath given circumstances. In design, construction, and maintenance of engineering system, we’ve to require several technological and social control choices at many stages. The last goal of all such choices is either to reduce the efforts needed or to maximize the specified advantages. Since the efforts required or profit desired in any useful circumstance is frequently communicated as an objective function of some design variables. Optimization process can be viewed as the way toward discovering the conditions that can help in finding the maximum or minimum value of an objective function. Besides the evolutionary techniques and swarm insight based calculations, deal with the principles of various phenomenon such as HS algorithmic rule, GSA, BBO, GEM, and so on [4]. All the biological process and swarm knowledge basically based calculations require normal dominant parameters such as population size, a variety of generations, size of elite, etc. Besides the basic administration parameters, totally unique calculations require their own particular calculation particular administration parameters. For instance, GA, PSO, ABC, HS depends on different algorithmic-specific parameters and therefore need the variety of improvisations. Likewise, the contrary calculations like ES, DE, ACO, BFO, EP, AIA, SFL, and so on need the standardization of different calculation particular parameters. The right standardization of the algorithmic rule specific parameters could be a terribly difficult issue that affects the ability of most of the mentioned optimization techniques. The improper standardization of algorithm specific controlling parameters may increase processing effort or lead to the native best answer.

Assuming this reality, presented within the teaching learning primarily based optimization algorithmic rule that doesn’t need standardization of any algorithm specific controlling parameters. The TLBO algorithmic rule needs solely common dominant parameters for its operating. Keeping seeable of the success of the TLBO and Jaya algorithmic rule, another algorithmic rule specific parameter-less algorithm is planned. The planned algorithmic rule works in two phases and it needs solely the common control parameters for its operating. The social behavior inspired optimization technique is described in next section.

Conclusion

A Social Behavior Inspired Optimization algorithm which is algorithmic parameter less is presented in this research work having two phases. In first stage, the algorithm consistently tries to get closer to success (by considering the great behavior) and tries to stay away from disappointment (by disregarding the terrible behavior) and in second stage, candidates increment their answer by information exchange. The execution of the proposed Social Behavior Inspired Optimization algorithm is tested on 17 unconstrained benchmark capacities having diverse attributes. To assess the execution Social Behavior Inspired Optimization algorithm is contrasted with Differential Evolution (DE), Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Artificial Bee Colony Optimization (ABC), Jaya algorithm and Teaching Learning Based Optimization (TLBO). The results have affirmed the effectiveness of the Social Behavior Inspired Optimization algorithm when contrasted with different algorithms.

918 total views, no views today

Download PDF File

About the author: admin