The Use of Global Best Position in Rerun of Particle Swarm Optimization

Abstract

This paper studies the use of particle best position (GBEST) in rerun when particle swarm optimization (PSO) traps in local optima. Reinitialize particles positions are often used to restart PSO to get better results when trapping in local optima. This paper proposed the use of GBEST to further force particle movement out of previous local optima instead of only reset GBEST. The proposed method is tested on 26 benchmark test functions with satisfactory results.