粒子群优化算法的改进方法-研究

《粒子群优化算法的改进方法-研究》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、AbstractParticleSwarmOptimization(PSO)iSanewmetllodusedinthesolutionofOptimizationProblems．Sinceitwasproposedin1995，theresearcharticlesaboutithaveincreasedquickly．TheconceptofPSOissimpleandeasytoimplement．Itrandomlyinitializesacertainscaleofparticleswarm，inwhicheachparticle、ⅣitIlcertai

2、nintelligenceisusedtorepresentthecandidatesolutioninthespecificoptimizationproblem，andthenusestheinformationofthegroupandindividualparticlestofindanoptimalsolutionquicklythroughtheiterativeevolution．Tmsalgorithm．basedonthetwotheories—swarmintelligenceandevolutionarycomputation，isf．aVor

3、edbroadlybecauseofitsmanyadvantagessuchasitsfewparameters，simplesetupprocedureandthefastconverge．NowadaystherehavebeenalargenumberofpapersabouttheimprovedalgorithminordertomakethePSOalgorithmperformbetter．Theyhavebeensuccessfullyappliedtotheengineeringoptimizationproblems．Astheresearch

4、ismoreandmoredeeply,itsapplicationfieldsarealsoexpanding，anditsperformanceisalsogreatlyimproved．Tl：lispaperfh'stlydoesanin—depthstudyofthetheoreticalbasis，basicprincipleandrealizationprocessofthePSOalgorithm．Itanalyzestheinfluenceofrelatedparametersonthearithmeticperformance，theefficie

5、ncyofthealgorithmandtheimplementationsonthebasisofsimulationexperiments．ConsideringthedefectsofthePSO，thispaperanalyzeswhatandwhichaspectsshouldbeimprovedonthebasisoftheparticlemovementcharacteristics，andthepaperalsoexplainstheimprovementandappliedscopeofthealgorithm．Byanalyzingthebasi

6、ctheoriesandtheimprovedmethods，thispaperproposesamulti—agentparticleswarmoptimizationalgorithmbasedonthetheoryofmulti—agents，andgivesspecificproceduresofthisalgorithm．n圮authorprovestheeffectivenessofitsimprovementthroughtheMATLABsimulationexperiment，analyzesconcretelytheimprovedresults

7、，andmeanwhilepointsouttheweakpoints．Finally,thispapersummarizestheauthor’SstudyonthePSOalgorithm，andproposesafurtherresearchplan．ItisprovedthattheimprovedMAPSOismoreeffectivethanthebasicPSOwhensolvingthemorecomplexmultimodalfunctionoptimizationproblems．111eMAPSOtrulyrealizestheglobal