Predicted Particle Swarm Optimization
Zhihua Cui, Jianchao Zeng, Guoji Sun · 2006
The standard particle swarm optimization (PSO) may prematurely converge on suboptimal solution partly because of the insufficiency information utilization of the velocity. The time cost by velocity is longer than position of each particle of the swarm, though the velocity, limited by the constant vmax, only provides the positional displacement. To avoid premature convergence, a new modified PSO, predicted PSO, is proposed owning two different swarms in which the velocity without limitation, considered as a predictor, is used to explore the search space besides providing the displacement while the position considered as a corrector. The algorithm gives some balance between global and local search capability. The optimization computing of some examples is made to show the new algorithm has better global search capacity and rapid convergence rate