A Random Opposition-Based Learning Grey Wolf Optimizer

Wen Long,Shaohong Cai,Jianjun Jiao,Ximing Liang,Ming Xu

doi:10.1109/access.2019.2934994

Abstract

Grey wolf optimizer (GWO) algorithm is a swarm intelligence optimization technique that is recently developed to mimic the hunting behavior and leadership hierarchy of grey wolves in nature. It has been successfully applied to many real world applications. In the GWO algorithm, “C”is an important parameter which favoring exploration. At present, the researchers are few study the parameter “C”in GWO algorithm. In addition, during the evolution process, the other individuals in the population move towards to the α, β, and δ wolves which are to accelerate convergence. However, GWO is easy to trap in the local optima. This paper presents a modified parameter “C”strategy to balance between exploration and exploitation of GWO. Simultaneously, a new random opposition-based learning strategy is proposed to help the population jump out of the local optima. The experiments on 23 widely used benchmark test functions with various features, 30 benchmark problems from IEEE CEC 2014 Special Session, and three engineering design optimization problems. The results reveal that the proposed algorithm shows better or at least competitive performance against other compared algorithms on not only global optimization but also engineering design optimization problems.

Highlights

A wide variety of nature-inspired optimization algorithms are there in swarm intelligence and evolutionary computation literatures
In 2018, inspired by particle swarm optimization (PSO), Long et al [31] presented a novel variant of grey wolf optimizer (GWO) based on the nonlinear control parameter and the modified position-updating equation to balance between exploration and exploitation of the conventional
We investigate the performance of random oppositionbased learning (ROL)-GWO using 23 widely used benchmark test functions and 30 benchmark test functions taken from IEEE CEC2014

Summary

INTRODUCTION

A wide variety of nature-inspired optimization algorithms are there in swarm intelligence and evolutionary computation literatures. In 2018, inspired by PSO, Long et al [31] presented a novel variant of GWO based on the nonlinear control parameter and the modified position-updating equation to balance between exploration and exploitation of the conventional. In 2019, Long et al [35] proposed a novel GWO variant based on refraction learning for solving global optimization problem. In 2018, Gupta and Deep [39] proposed a modified version of GWO based on random walk strategy to improve the global search ability of the basic GWO algorithm. MODIFIED PARAMETER ‘‘C’’ STRATEGY All population-based optimization techniques aim to achieve a balance in both the exploration and exploitation to obtain the promising regions of the search space and eventually. Based on the above facts, the Eq (9) can effectively balance between global exploration and local exploitation of the GWO algorithm

RANDOM OPPOSITON LEARNING STRATEGY

EXPERIMENTS AND DISCUSSION

CONCLUSION