A Hybrid Attractive and Repulsive Particle Swarm Optimization Based on Gradient Search

摘要

As an evolutionary computing technique, particle swarm optimization (PSO) has good global search ability, but its search performance is restricted because of stochastic search and premature convergence. In this paper, attractive and repulsive PSO (ARPSO) accompanied by gradient search is proposed to perform hybrid search. On one hand, ARPSO keeps the reasonable search space by controlling the swarm not to lose its diversity. On the other hand, gradient search makes the swarm converge to local minima quickly. In a proper solution space, gradient search certainly finds the optimal solution. In theory, The hybrid PSO converges to the global minima with higher probability than some stochastic PSO such as ARPSO. Finally, the experiment results show that the proposed hybrid algorithm has better convergence performance with better diversity than some classical PSOs.