欢迎来到天天文库
浏览记录
ID:27281951
大小:3.53 MB
页数:113页
时间:2018-12-02
《基于动态子群策略的改进pso算法与其应用分析》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、-------西安理工大学硕士学位论文andshowsitssuperiorityinsolvingpracticalproblems.Onthebasisofintroducingthepowerfulteachingandlearningglobalsearchoperator,westudyanddesignanimprovedparticleswarmoptimizationalgorithmwithanewsubgroupsplittingmethodanddynamicsubgroupstrategy.We
2、adoptarandomgroupingstrategyinthenewsubgroupsplittingmethod.WenamethenewalgorithmasDynamicSubgroupStrategyParticleSwarmOptimizationalgorithm(DSSPSO).Toexaminetheeffectivenessoftheproposedalgorithm,itistestedon10benchmarkfunctions.ExperimentalresultsshowthatDSSPS
3、Oisaneffectivemethodtooptimizecomplexfunctions,andperformssignificantlybetterthantheoriginalPSOalgorithmandotherimprovedPSOalgorithmi.e.PSO-wFIPSandUPSO.(1)Inmanypracticaloptimizationproblems,therearesomeconstraintconditions.Swarmintelligencealgorithmshavebeensu
4、ccessfullyappliedforconstrainedoptimizationproblems.However,thedifferencebetweeneachconstraintproblemsdifferinthousandsofways,andweusethepenaltyfunctionmethodtosolvetheconstrainedoptimizationproblemisverydifficulttosettheparametersoftheappropriatefactors.Hencewe
5、introduceanewmechanismtoimprovetheparametersconfiguretermeddynamicpenaltymechanismwhichisconducivetosearchforthemobilefeasiblesolutionregiontoacceleratetheprocessofevolution.Basedonthenewproposedmechanismofdynamicpenalty,weapplytheimprovedDSSPSOalgorithmtosolvec
6、rudeoilblendingproblemwithmulti-constraintconditions.ThenwegivethemainflowofcrudeoilblendingoptimizationproblemusingdynamicpenaltymechanismandDSSPSO.Finally,weusetwotestdatatotestifythevalidityofthemethod,andthetestresultsshowtheeffectivenessofthedynamicpenaltym
7、echanismandDSSPSOalgorithm.Keywords:DSSPSO;DynamicSubgroupStrategy;ConstrainOptimization;DynamicPenalty-----------ii万方数据-----------目录目录1绪论............................................................................................................................
8、..........................................11.1课题研究背景..................................................................................................................
此文档下载收益归作者所有