A Hybrid Particle Swarm Optimization Strategy for Multimodal Function Optimization

Particle swarm optimization is easy to fall into local minima, defects and poor precision. In order to solve the above problem, a hybrid particle swarm optimization named HPSO has been proposed in this paper. The new method focuses on the change of the position of particle, which is updated by means of a radial symmetric function of the center in the iterative process. And to avoid premature convergence, simulated annealing algorithm is employed to dynamically adjust the inertia weight and social cognitive parameters for avoiding falling into local optimal optimum in the searching process. Finally, experiments are carried out on six multimodal functions for testing the hybrid efficiency and scalability, and the results of the simulation and comparison show that the hybrid particle swarm optimization is verified to be effective and scalable.