粒子群优化算法-研究与其在动态数据校正中的应用

《粒子群优化算法-研究与其在动态数据校正中的应用》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、江苏大学硕士学位论文ABSTRACTInrecentyears，particleswarmoptimization(PSO)algorithmasintelligentoptimizationalgorithmshasbeenwidelyattentionandresearched，whichderivedfromthesimulationofflockforagingbehavior．PSOalgorithmhastheadvantagewithconciseandeasy,feweradjustedparameter,fastconvergence

2、rateandstrongcommonality,however，PSOalgorithmalsohasshortcomingsofprematureconvergence，easytogetintolocaloptimal．Therefore，thispapertheoreticallyanalysistheprematureconvergenceofPSOalgorithm，thenproposedtwomodifiedalgorithmwhichavoidedprematureconvergenceandenhancetheoverallconv

3、ergenceability．Basedonthecontinuousstirredtankreaction(CSTR)systemastheresearchobject，introducePSOalgorithmintothedynamicdatarectificationfield，andproposedthenewmethodsfordynamicdatarectificationtechniqueresearch．Finally,thesimulationresultsverifythetwokindsofPSOalgorithmCanavoi

4、dprematureconvergenceandimprovedglobalconvergence，anddynamicdatarectificationcanbeincreasedsignificantlyinoptimizationefficiencyandperformanceinCSTR．Themaincontentsareoutlinedasfollows：1．Uncertainknowledgebasedontheparticleswarnloptimizationalgorithm(PSO—UK)fordynamicdatarectifi

5、cationmethodisproposed，introduceuncertainknowledgeintothestandardPSOalgorithm，makeparticleateachstepcanachievecompleteknowledge，restraindeteriorationoftheparticle，andovercomeprematurephenomenon．SimulationswithtestingfunctionsconfirmedthePSO-UKalgorithmisbetterthanotherimprovedPS

6、Oalgorithms．2．Theperiodicevolutionstrategybasedonparticleswarmoptimizationalgorithm(PSO—PES)isproposed．Thealgorithmbuiltthedisturbancestrategiesofoptimalvaluebasedongroupenergy,makingtheparticleswarmtocyclicalevolution．Thismethodcaneffectivelyavoidthe”premature”problemofparticle

7、．ComparingtheperformancewithPSO—UKthoughtestfunctions，indicatedthePSO-PESalgorithmhasbetteroptimizationcapabilityandglobalconvergenceperformance．III江苏大学硕士学位论文3．ApplicationsindynamicdatarectificationbasedonimprovedPSOalgorithm．TheproposednewdatarectificationmethodbasedonPSO··UKan

8、dPSO··PESalgorithmcaneffectivelyidentifygrosser