A Social Learning Based Particle Swarm Optimization Algorithm for Real-Parameter Single Objective Optimization Problems

Abstract

The Particle Swarm Optimization (PSO) algorithm is a simple and effective method that has been widely used to solve complex optimization problems. However, it can easily get trapped in a local optima due to the loss of population diversity. This paper presents a new variant of the PSO algorithm based on social learning (SL-PSO) that aims to improve performance of traditional PSO. This is encouraged by the ability shown by diverse animal species to learn from the behavior of more experienced individuals. Specifically, the historical information of the best particle is utilized to modify the position and direction of the stagnant particles, and improve the exploration capability of the swarm. Experiments conducted on unimodal and multimodal test functions demonstrate the effectiveness of the SL-PSO algorithm compared to other variants of the PSO algorithm.