Multiplicate Particle Swarm Optimization Algorithm

Using Particle Swarm Optimization to handle complex functions with high-dimension it has the problems of low convergence speed and sensitivity to local convergence. The convergence of particle swarm algorithm is studied, and the condition for the convergence of particle swarm algorithm is given. Results of numerical tests show the efficiency of the results. Base on the idea of specialization and cooperation of particle swarm optimization algorithm, a multiplicate particle swarm optimization algorithm is proposed. In the new algorithm, particles use five different hybrid flight rules in accordance with section probability. This algorithm can draw on each other ' s merits and raise the level together The method uses not only local information but also global information and combines the local search with the global search to improve its convergence. The efficiency of the new algorithm is verified by the simulation results of five classical test functions and the comparison with other algorithms. The optimal section probability can get through sufficient experiments, which are done on the different section probability in the algorithms.