Title: On the performance of particle swarm optimisation with(out) some control parameters for global optimisation
Authors: Aderemi Oluyinka Adewumi; Martins Akugbe Arasomwan
Addresses: School of Mathematics, Statistics and Computer Science, University of KwaZulu-Natal, Private Bag X54001, Durban 4000, South Africa ' School of Mathematics, Statistics and Computer Science, University of KwaZulu-Natal, Private Bag X54001, Durban 4000, South Africa
Abstract: This paper establishes that the basic particle swarm optimisation (BPSO) technique can perform efficiently without some (or any) of the control parameters in the particle velocity update formula and also presenting a modified BPSO (M-BPSO) that ameliorate the problem of premature convergence associated with PSO when optimising high dimensional multi-modal problems. Some modifications to BPSO are presented including dynamically decreasing the velocity limits of particles depending on the progressive minimum and maximum dimensional values of the entire swarm in order to control the exploration and exploitation activities of M-BPSO. Various experiments, based on benchmark problems, were conducted to achieve the above stated goal and compare the performances of BPSO with MPSO empirically. Variants of M-BPSO were also compared with other efficient techniques from literature. Experimental results demonstrate the superior performance of the proposed M-BPSO algorithm in terms of solution quality, convergence speed, global-local search ability and stability.
Keywords: particle swarm optimisation; control parameters; global optimisation; particle velocity updating; PSO performance; solution quality; convergence speed; global-local search ability; stability.
International Journal of Bio-Inspired Computation, 2016 Vol.8 No.1, pp.14 - 32
Available online: 09 Feb 2016 *Full-text access for editors Access for subscribers Purchase this article Comment on this article