粒子种群优化(PSO)算法的性能研究

近年来,一种新的基于种群优化的算法———粒子种群优化(PSO)算法,正受到人们的普遍关注。首先介绍了PSO原理及具体实现步骤,接着对各种常见PSO算法,例如原始算法、惯性权值算法、限制因子算法等进行了解释。在此基础上,对PSO算法典型模型的参数选择,如惯性权值、加权系数、最大速度等,进行了详细研究,并给出了实验结果,得出了相关结论,为今后参数的选择提供了参考。接着讨论了PSO在神经网络、模糊逻辑系统和进化计算等计算智能领域及其它工程领域的应用,最后给出了进一步的研究方向。

Study on the Performance of the Particle Swarm Optimization Algorithm

In recent years,a novel population based on optimization algorithm,particle swarm optimization(PSO) algorithm,has gained great attention.In this paper,the principle of and a variety of related methods of PSO are firstly introduced.Then,some commonly used PSO algorithms,such as original algorithm,inertia weight algorithm,constrain factor algorithm etc.,are explained.After that the parameter selection of a classical PSO model,such as inertia weight,weighting coefficient,maximum velocity etc.,are analyzed and compared in detail.Some simulation results are given and conclusions are drawn.It can act as references for the parameter selection in the future.Next,the applications of PSO in computational intelligence fields,such as Neural Network,Fuzzy Logic System,Evolutionary Computation etc.and other engineering fields are discussed.Finally,some research directions to be further studied are presented.