Improvement of original particle swarm optimization algorithm based on simulated annealing algorithm

Particle swarm optimization (PSO) algorithm is an optimization algorithm in the filed of Evolutionary Computation, which has been applied widely in function optimization, artificial neural networks´ training, pattern recognition, fuzzy control and some other fields. Original PSO algorithm could be trapped in the local minima easily, so in this paper we improved the original PSO algorithm using the idea of simulated annealing algorithm, which makes the PSO algorithm jump out of local minima. In this paper, two improved strategies was proposed, and after testing and comparing the two improved algorithms with the original PSO algorithm again and again, we conclude at last that efficiency of searching global about the two improved strategies is better than the original PSO.