具有随机惯性权重的PSO算法  

A Particle Swarm Optimization Model with Stochastic Inertia Weight

作　　者：胡建秀[1] 曾建潮[1] 

机构地区：[1]太原科技大学系统仿真与计算机应用研究所,山西太原030022

出　　处：《计算机仿真》2006年第8期164-167,共4页Computer Simulation

基　　金：教育部重点科研项目(204018)

摘　　要：微粒群算法(PSO算法)是模拟鸟类、鱼群等的群体智能行为的一种优化算法,当前,在相关领域内,倍受国内外学者关注。该文在分析基本PSO算法的速度进化方程的基础上,提出一种能更好描述微粒进化过程的速度方程,由其引出一种具有随机惯性权重的PSO算法;通过五个典型测试函数的仿真实验,验证了其可行性,同时也表明具有随机惯性权重的PSO算法较具有线性递减惯性权重的PSO算法在收敛速度和全局收敛性方面有明显提高。The particle swarm optimization is an optimization method through simulating social behavior of bird flocking or fish schooling, which attracts much attention of many scholars in this region. Based on the analysis of evolutionary equations of the standard PSO, the paper proposes a velocity evolutionary equation which could describe more precisely particle＇s evolving process, and further gives a PSO model with stochastic inertia weight. The results of five benchmark functions prove the model to be feasible, at the same time show that the performance of the PSO with stochastic inertia weight is improved obviously than that of the PSO with linearly decreasing inertia weight.

关 键 词：微粒群算法 惯性权重 随机惯性权重 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]