Study on Hybrid Algorithm Based on Particle Swarm Optimization and Simulated Annealing Algorithm

Zhong Wang
2011-01-01
Abstract:An improved method is derived from the standard particle swarm optimization.Firstly,the particles are initialized by chaos technique which is ergodic,stochastic and regular.To enhance the diversity and to hold the better particles,all the particles are initialized selectively again for a period of time.Finally,all the individual best positions except the index of global best position are mutated.The results show that convergence speed and accuracy of algorithm are enhanced by the improved particle swarm optimization.But it is impossible to find the best positions of some functions.Then test functions are calculated again by a hybrid algorithm based on improved particle swarm optimization and simulated annealing algorithm,and the results show that the global optimal solution is achieved by the hybrid algorithm.
What problem does this paper attempt to address?