一种基于和声搜索的动态交叉粒子群算法 - 浙江工业大学ivsn group.pdf

小型微型计算机系统 2013年9月第9期 ofChinese V01．34No．92013 Journal ComputerSystems 一种基于和声搜索的动态交叉粒子群算法 陈晋音，杨东勇 (浙江工业大学信息工程学院，杭州310023) E-mail：chenjinyin@163．tom 摘要：针对粒子群优化算法(PSO)在优化多维问题时容易陷入局部最优的问题，提高其全局搜索能力和拓展能力，提出了一 种基于和声搜索的动态交叉粒子群算法．引入动态交叉操作，使得粒子在更新速度时实现共享有效信息，保证粒子进化过程中 且算法时间复杂度相比传统PSO增加不明显． 关键词：自适应和声搜索；动态交叉率；交叉粒子群；种群多样性 中图分类号：TPl8 文献标识码：A 文章编号：1000·1220(2013)09-2152-06 CrossoverParticleSwarm BasedOil Search Dynamic OptimizationHarmony CHEN Jin-yin，YANGDong-yong ofInformation (College EngineeringZhejiangUniversityofTechnology．Hangzhou310023，China) at local swarm is to its search the trap of improveglobal Abstract：Aiming problemparticle optimization(PSO)-thispapersupposed and carriedouta crossoverPSObasedon crossover is to exploitability．We dynamic HarmonySearch．Dynamicoperationadopted make shareusefulinformationeachotherwhichwould botll and search particles among guaranteepopulationdiversityglobal ability． HS—DCPSOis combinedwithrandomsearch of search．ParameterHtch up adjustingrate(Par)andarbitrary brought abilityharmony distancebandwidthbwwouldbe to of aretestifiedonvariousbenchmark self-adaptiveimproveexploitabilityparticles．HS—DCPSO has functionsand with andvarious PSOs．Thesimulationresults thatHS-DCPSObetter HS，PSO per- compared improved proved formancesof searchand withoutobvious incensement