Particle swarm optimizer with adaptive species radius for multimodal function optimization

摘要

Based on species-PSO, this paper proposes an adaptive species radius strategy for multimodal function optimization. In species-PSO, The radius of the species is a very important parameter, which determines the success and failure of the algorithm. However, the value of the radius extremely depends on the corresponding optimization problems. The task of choosing an appropriate radius is difficult for users. We define stability of the species to denote the search ability for every species. According to stability of species and the number of members in species, the radius of the species is adjusted adaptively. The experimental results show that adaptive species radius strategy makes species-PSO could handle a wide range of optimization problems without user intervention and achieve good performance for multimodal function optimization.