Improving the Particle Swarm Optimization algorithm using the Simplex Method at late stage

Abstract

This article proposes a hybrid Particle Swarm Optimization (PSO) based on the Nonlinear Simplex Method (NSM). At late stage of PSO running, when the promising regions of solutions have been located, the algorithm isolates particles which are very close to the extrema and applies the NSM to them to enhance the local exploitation. Experimental results on several benchmark functions demonstrate that this approach is very effective and efficient, especially for multimodal function optimizations. It yields better solution qualities and success rates compared to other methods taken from the literature. © 2005 by International Federation for Information Processing.