Dynamic Multi-objective Optimisation Using Multi-guide Particle Swarm Optimisation

Abstract

This study conducts a sensitivity analysis of the recently proposed multi-guide particle swarm optimisation (MG-PSO) algorithm for dynamic multi-objective optimisation problems (DMOPs). The MGPSO is a multi-swarm approach where each subswarm optimises one of the objectives. This paper further adapts the MGPSO algorithm to solve DMOPs by proposing alternative balance coefficient update strategies to allow efficient tracking of the changing Pareto-optimal front (POF). A total of twenty-nine benchmark functions and six performance measures were implemented to help with this task. The experiments were run against five different environment types to determine whether the MGPSO can solve problems with various spatial and temporal severities. The best control parameter update strategy was then compared with other state-of-the-art dynamic multi-objective optimisation algorithms (DMOAs). An extensive empirical analysis shows that MGPSO with the balance coefficient parameter re-initialized after the environment change achieves very competitive and oftentimes better performance when compared with the competing algorithms.