Title: Perfectly convergent particle swarm optimisation in multidimensional space

Authors: Devinder Kumar; N.K. Jain; Uma Nangia

Addresses: Department of Electrical Engineering, G B Pant institute of Technology, Okhla Phase III, New Delhi, India ' Department of Electrical Engineering, Delhi Technological University, Delhi, India ' Department of Electrical Engineering, Delhi Technological University, Delhi, India

Abstract: In this paper a novel evolutionary algorithm, perfectly convergent particle swarm optimisation (PCPSO) has been proposed. This is an intelligent algorithm which does not get trapped in local minima by using personal best value along with new parameters and new velocity update equation for better exploration in the search space. Velocity clamping is used to avoid the particles from escaping the search space by preventing the initial explosion of particle velocity. The velocity clamping effectively help to control the maximum velocity of the particles by limiting the particle step size and hence the particle will have to perform further steps in the search space to travel the same distance which determines its search precision for exploration or exploitation and finally align them towards the true global minimum. Experimental results show that by using perfect convergent particle swarm optimisation (PCPSO) approach computational efficiency is increased as compared to other variants of PSO and finds fast true global minimum.

Keywords: particle swarm optimisation; PSO; exploration; stagnation; premature convergence; velocity clamping; PCPSO.

DOI: 10.1504/IJBIC.2021.119997

International Journal of Bio-Inspired Computation, 2021 Vol.18 No.4, pp.221 - 228

Received: 10 Aug 2020
Accepted: 16 Apr 2021
Published online: 04 Jan 2022 *

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