Incorporating the Regional Dominance Strategy into Multi-Swarm PSO

In this paper, we improve the exploitation capability of Multi-Swarm Particle Swarm Optimization (MSPSO) by integrating a dominance policy implemented between competing sub-swarms in some randomly generated regions defined by hyperspheres. Furthermore, we modify a stage of the classical MSPSO using Mesh Adaptive Direct Search in the place of a line search optimization method to enhance the exploration ability. The proposed method has been experimented on some well-known benchmark functions for the continuous optimization problems.