Particle swarm optimizer based on dynamic neighborhood topology and mutation operator

A variant of particle swarm optimizer(PSO) based on dynamic neighborhood topology and mutation factor(DNMPSO) is proposed in this paper. In DNMPSO,the neighborhood of a particle are not fixed but dynamically changed,and the learning mechanism of a particle includes two parts,the historical best experience of the particle itself,and the experiences of its all neighbors. To effectively solve multimodal problems,the parallel hybrid mutation is used to work for local search,which improves the ability of escaping from local optima. The results demonstrate good performance of the DNMPSO algorithm in solving complicated multimodal problems when compared with other PSO algorithms.