Title

The Progressive Analysis Of Particle Swarm Optimization

Keywords

Convergence; PSO; Stochastic component; Unimodal function

Abstract

Particle Swarm Optimization (PSO) has been proposed for more than ten years. Most of studies concentrate on the improvement and the application on PSO, but the theoretical analysis on PSO is very incomplete. The current approach of theoretical analyses on PSO is to simplify the system, which includes: Removal of stochastic factor, isolated single individuals, search stagnation. Based on current works, this paper will progressively analyze the convergence of PSO with extending conditions mentioned above that improved solution is found and the random factor is added. The following conclusions are drawn in this paper: 1) Without stochastic component, the trajectory of particle, x (t), must converges to the global optimal location yg on PSO being employed in unimodal function, while it must not in multimodal function. 2) With stochastic component, the trajectory of particle, x (t), must converges to the global optimal location yg whether PSO is employed in unimodal function or multimodal function. The experiments conducted in this paper demonstrate the conclusions. © 2008 Binary Information Press.

Publication Date

12-1-2008

Publication Title

Journal of Computational Information Systems

Volume

4

Issue

6

Number of Pages

2885-2891

Document Type

Article

Personal Identifier

scopus

Socpus ID

65649103452 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/65649103452

This document is currently not available here.

Share

COinS