Abstract:

In view of both fixed inertia weight and premature convergence  obvious flaws of particle swarm optimization (PSO) algorithm, a dynamic adaptive adjustment strategy for inertia weight was proposed on the basis of a detailed analysis of the relationship among the inertia weight, population size, particle fitness and search space dimension,  which effectively enhances the global and local optimization abilities of the algorithm. For the problem of premature, the chaotic mapping method was used to increase the diversity of the population, while the group extreme was adjusted in the direction of negative gradient, which greatly reduces the probability of fall into the local extreme. The correctness and effectiveness of the proposed  PSO algorithm were verified to improve by some common used test functions compared with those by other algorithms.

Key words: particle swarm optimization (PSO) algorithm, chaotic, inertia weight, adaptive