Enhanced Particle Swarm Optimization With Self-Adaptation Based On Fitness-Weighted Acceleration Coefficients. Issue 1 (2nd January 2016)

• Record Type: Journal Article
• Authors: Wang, Hei-Chia; Yang, Che-Tsung
• Is Part Of: Intelligent automation & soft computing. Volume 22:Issue 1(2016)
• Page Start: 97
• Page End: 110
• Publication Date: 2016-01-02
• Subjects: Particle swarm optimization -- Acceleration coefficient -- Adaptive -- Cognition learning -- Social learning; Artificial intelligence -- Periodicals; Intelligent control systems -- Periodicals; 003.5
• DOI: 10.1080/10798587.2015.1057956
• Languages: English
• ISSNs: 1079-8587

Abstract

Abstract: Acceleration coefficients are the control parameters used to tune the movements of cognition and social components in the particle swarm optimization (PSO) algorithm. Because most of the PSO algorithms treat individual particles equally despite the different positions of distinct particles, the inhomogeneous spread and scattering of the data samples during evolution are ignored. In this regard, the proposed PSO-FWAC algorithm aims to enhance the adaptability of individual particles by introducing the diverse acceleration coefficients according to their corresponding fitness values. The experimental results show that the PSO-FWAC outperforms the static and time-varying approaches.