A switching particle swarm optimization for multimodal optimization problem

Particle swarm optimization (PSO) has a variety of applications on optimization problems, and it has been proved better convergence performance than former evolutionary algorithms (as GA), but the standard PSO algorithm is sensitive to fall into local optima, especially for multimodal optimization problem. To deal with this case, this paper proposes an switching PSO algorithm with a novel velocity update mechanism and switching mode based on entropy of swarm and the global optima, according to which the proposed PSO changes velocity and particle update formula. Some benchmark tests were performed, and numerical results show advantages in comparison with performance of standard PSO.