Title: Species-conserving particle swarm optimisation for multimodal functions

Authors: Jian-Ping Li, Alastair S. Wood

Addresses: School of Engineering, Design and Technology, University of Bradford, Chesham Building, Richmond Road, Bradford, West Yorkshire, BD7 1DP, UK. ' School of Engineering, Design and Technology, University of Bradford, Chesham Building, Richmond Road, Bradford, West Yorkshire, BD7 1DP, UK

Abstract: A species-conserving particle swarm optimisation (SCPSO) is proposed to seek multiple solutions of multimodal functions. The SCPSO is implemented by modifying the velocity function to allow a PSO to locate multiple solutions, using species conservation to maintain the diversity of particles. In order to maximise the diversity of particles, the personal best position of a particle is updated only when the particle and its personal best belong to the same species. Species seeds are identified from both particles and their personal best positions. Particle velocity is influenced by its nearest species and by the nearest global solution. Numerical examples illustrate that the proposed SCPSO can comfortably outperform other related existing algorithms in terms of function evaluations on a set of benchmark multimodal functions.

Keywords: multimodal functions; particle swarm optimisation; PSO; species conservation; speciation.

DOI: 10.1504/IJMIC.2009.030075

International Journal of Modelling, Identification and Control, 2009 Vol.8 No.4, pp.290 - 300

Published online: 09 Dec 2009 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article