Abstract: Particle swarm optimization algorithm is of simple in form, flexible in parameter setting, easy to operate, and capable of fast convergence， so it has attracted many attentions. However, the traditional particle swarm algorithm also has its drawbacks: the slow convergence rate and the vulnerability to local optimization. In order to solve this problem, this paper uses the niche method to initialize the population, in the initial stages of evolution, dividing the initial population into subpopulations, and using different mutation strategies for different subpopulations; in the process of evolution, different inertia weighting factors are set for different subpopulations, in order to enhance the global search ability and the local search ability. The results of test functions show that the algorithm has faster convergence than the traditional particle swarm algorithm, the global optimal solution is closer to the real solution set, and the convergence accuracy is also higher.

Key words: particle swarm optimization, the niche method, local PSO, variation strategy