一种基于路径信息衰减的粒子群改进算法  

An Improved Particle Swarm Optimization Algorithm based on The Attenuation of Path Information

作　　者：杨蕊[1] 陈华[2] 张艺丹 

机构地区：[1]东营职业学院,山东东营257000 [2]中国石油大学,山东青岛266580

出　　处：《电脑知识与技术（过刊）》2015年第3X期190-191 196,196,共3页Computer Knowledge and Technology

摘　　要：针对标准微粒群易于收敛于局部最优解的问题,提出了一种通过添加粒子群搜索历史信息的改进微粒群算法。标准粒子群算法没有用到粒子群以往的迭代信息,再加上粒子群算法的快速收敛性,造成了算法易于收敛到局部最优解。改进的算法在运行时可以增加算法搜索空间,使算法更加平稳的收敛于全局最优解。通过用典型的Benchmarks函数进行模拟试验,实验结果证实了所提出的算法更加平稳,收敛速度更快。The standard particle swarm optimization algorithm is easily converging to the local optimal solution. The algorithmabove seldom contains the historical information of particle swarm search path, adding the fast convergence of this algorithm, theymake this optimization converging to the local optimal solution too fast. So an improved particle swarm optimization by adding thehistorical information of particle swarm search is presented, which adds the algorithm search space to make the algorithm converg-ing to global optimal solution steadily during the running time. Several classic Benchmarks functions are tested and the resultsshow that the rate of convergence is faster and the algorithm is more stable.

关 键 词：微粒群算法 全局最优解 Benchmarks函数 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]