求解多峰函数的改进粒子群算法的研究

针对标准粒子群算法进行多峰函数优化时存在的易陷入局部极值和搜寻效率低的问题，提出了子种群划分和自适应惯性权重改进方法来求解多峰函数．根据群体微粒的相似度将粒子群分成子群体，各子群体围绕一个有最佳适应值的群体中心进行建立，并通过几个经典函数进行求解．实验表明：改进的粒子群算法能快速有效地找到多峰函数的全局最佳值．

Study on an Improved Particle Swarm Optimization Algorithm for Solving Multi-peak Function

In application of particle swarm optimization algorithm, the solution is usually misled to local minima resulting in compromised searching efficiency. To cope this drawback, the division of subpopulation and self-adaptive inertia weight are incorporated to PSO. The particle swarm is divided by subpopulation into subgroups close to the center of the swarm with the best fitness value judged by the similarity of particle swarm. Test of the improved algorithm using six benchmark functions indicates that the algorithm is capable of locating the global optimal value with less time-consumption and higher efficiency.