Multimodal Function Optimization Based on Multigrouped Mutation Particle Swarm Optimization

An improved hybrid particle swarm algorithm, named the multigrouped mutation particle swarm optimization(MMPSO), is provided in this paper. It keeps the basic concepts of the PSO, at the same time embeds the mutation operator of genetic algorithm, thus, it shows a more straightforward convergence ratio and the global searching ability compared to conventional PSO. Moreover, the MMPSO has a unique advantage in that on can search many superior peaks of a multimodal function when the number of the groups is N. Two multimodal functions were tested by the MMPSO algorithm, and results show the MMPSO can obtain the best optima and the rest optimum of those multimodal function.