A New Strategy Based on GSABAT to Solve Single Objective Optimization Problem

A New Strategy Based on GSABAT to Solve Single Objective Optimization Problem

H.A. Sattar (Universiti Pendidikan Sultan Idris, Tanjung Malim, Malaysia), Alaa Cheetar (Mustansiriyah University, Baghdad, Iraq) and Iraq Tareq (University of Baghdad, Baghdad, Iraq & University Putra Malaysia, Seri Kembangan, Malaysia)
Copyright: © 2019 |Pages: 22
DOI: 10.4018/IJSIR.2019070101

Abstract

This article proposes a new strategy based on a hybrid method that combines the gravitational search algorithm (GSA) with the bat algorithm (BAT) to solve a single-objective optimization problem. It first runs GSA, followed by BAT as the second step. The proposed approach relies on a parameter between 0 and 1 to address the problem of falling into local research because the lack of a local search mechanism increases intensity search, whereas diversity remains high and easily falls into the local optimum. The improvement is equivalent to the speed of the original BAT. Access speed is increased for the best solution. All solutions in the population are updated before the end of the operation of the proposed algorithm. The diversification feature of BAT has solved the problem of weakness in diversity observed in the algorithm by applying the parameters used in BAT. Moreover, balance is achieved through the intensification properties of the algorithms.
Article Preview
Top

1. Introduction

Classical optimization algorithms do not provide a suitable solution for optimization problems with a high-dimensional search space because of the exponentially increasing search space concerning problem size. Therefore, solving such problem using exact techniques, such as comprehensive research, is impractical (Alatas, 2010).

Naturalistic phenomena inspire the increasing interest in algorithms over the last decade (Dorigo et al., 1996; Farmer et al., 1986; Kennedy & Eberhart, 1995; Kim et al., 2007; Kirkpatrick et al., 1983; Tang et al., 1996). Many researchers have proven that these algorithms are well suited for solving complex computational problems, such as the optimization of objective functions (Du & Li, 2008; Yao et al., 1999), pattern recognition (Tan & Bhanu, 2006; Liu et al., 2008), control objectives (Baojiang & Shiyong, 2007; Karakuzu, 2009; Kim et al., 2008), image processing (Cordón et al., 2006; Nezamabadi et al., 2006), and filter modeling (Kalinli & Karaboga, 2005; Lin et al., 2008). Various heuristic approaches have been adopted in research, such as simulated annealing (Kirkpatrick et al., 1983), genetic algorithm (GA) (Tang et al., 1996), ant colony search algorithm (Dorigo et al., 1996), tabu search (Glover 1989 & 1990), evolutionary algorithms (Van et al., 2006), gravitational search algorithm (GSA) (Rashedi et al., 2009), particle swarm optimization (PSO) (Kennedy & Eberhart, 1995), and bat algorithms (BAT) (Yang, 2001). Algorithms are gradually powered in different areas (Wolpert et al., 1997; Lozano et al., 2008; Tripathi et al., 2007; Glover, 1989; Glover, 1990; Rashedi et al., 2009) to solve various optimization problems. However, no specific algorithm is used to find the best solutions for all the problems in finite iterations, and certain algorithms exhibit better performance for particular problems compared with others. Thus, searching for new heuristic optimization algorithms is an open problem (Tripathi et al., 2007). For example, GSA is based on the movement of particles that are affected by the gravitational force. Moreover, GSA can be used to improve the convergence rates of BAT during iterations and enhance BAT behaviour for high-dimensional problems.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 11: 4 Issues (2020): 2 Released, 2 Forthcoming
Volume 10: 4 Issues (2019)
Volume 9: 4 Issues (2018)
Volume 8: 4 Issues (2017)
Volume 7: 4 Issues (2016)
Volume 6: 4 Issues (2015)
Volume 5: 4 Issues (2014)
Volume 4: 4 Issues (2013)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing