基于Powell搜索法的混合微粒群算法

利用Powell搜索法求解精度高、收敛速度快和局部搜索能力强等优点,本文提出了一种与Powell搜索法相结合的改进微粒群算法实践.改进算法将微粒的搜索过程分为两阶段,第一阶段,将PSO算法的速度公式改进后进行搜索;第二阶段,将第一阶段的最后一代微粒作为Powell搜索法的初始点,让Powell搜索法与PSO算法交替进行.这样既克服了PSO算法易陷入局部最优的缺点,也大大提高了算法的求解精度和收敛速度,同时保持了微粒的多样性.仿真结果表明:同PSO算法相比,Powell-PSO算法具有较高的求解精度和较强的寻优能力,并且不论是对单峰函数还是多峰函数都能取得很好的优化效果.

A Hybrid Particle Swarm Optimization Algorithm Based on the Powell Search Method

In this paper we propose a new improved particle swarm algorithm in combination with Powell search method-Powell-PSO.Improved algorithm of the search process of particle is divided into two stages.First stage,the speed of the standard particle swarm algorithm formula is improved and carried out in accordance with the improved formula search.Second stage,the last generation particles of in the first phase is used as the initial point of Powell search method,and the Powell search method and PSO algorithm are used to alternates search.New algorithms can overcome the drawback of trapping in local optimum particle swarm algorithm,and greatly improves the precision of the algorithm,and improves the convergence speed and keep the diversity of the particles.The simulation results show that compared with the standard particle group algorithm,Powell-PSO has higher precision and stronger optimization ability,and whether to unimodal function or multimodal function can gain better optimization effect.