Parameter co-evolution mechanism of particle swarm optimisation algorithm. (7th May 2020)

• Record Type: Journal Article
• Authors: Zhao, Ming; Song, Xiaoyu; Gao, Yichen
• Is Part Of: International journal of simulation and process modelling. Volume 15:Number 3(2020)
• Page Start: 255
• Page End: 267
• Publication Date: 2020-05-07
• Subjects: particle swarm optimisation -- PSO -- parameter co-evolution -- inertia weight -- acceleration factor -- adjustment mechanism -- stochastic evolution speed -- population evolution -- numerical optimisation -- benchmark functions; Management -- Computer simulation -- Periodicals; Mathematical models -- Periodicals; Operations research -- Periodicals; Simulation methods -- Periodicals; 003.05
• Languages: English
• ISSNs: 1740-2123

Abstract

The running parameters are the important factors that influence the performance of PSO, and the optimisation of the selection and the adjustment strategy on them is one of the hot research directions. Based on the related research, this paper designs a co-evolution mechanism for the parameters of PSO including both the inertia weight and the acceleration factors, which defines stochastic evolution speed to reflect the current state of population evolution during the iterative process, and uses it as the feedback to set the inertia weight and the two acceleration factors. PSO with the parameter co-evolution mechanism can realise cooperative evolution of the running parameters with the population by dynamically adjusting parameter values according to population evolution state. Compared with five widely recognised parameter selection or adjustment strategies, on 20 numerical optimisation benchmark functions of different categories, the effectiveness and the efficiency of the proposed mechanism are verified.