高校动态多目标优化算法与其在pid控制中的应用分析

《高校动态多目标优化算法与其在pid控制中的应用分析》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、硕士学位论文AbstractInengineeringdesign，real-timecontrol，artificialintelligenceandotherengineeringfields，thereareawiderangeofoptimizationproblemsthattheobjectivesandconstraintswillchangeovertime．Andthiskindofproblemiscalleddynamicoptimizationproblem．Iftheoptimizationpro

2、blemhasmultipleconflictingoptimizationgoals，thenitisknownasadynamicmulti-objectiveoptimizationproblem．Thecomplexnatureofthedynamicmulti-objectiveoptimizationproblemitselfmakesthiskindofproblemverydifficulttosolve．Itrequiresthealgorithmtomakeadjustmentsaccordingtoc

3、hangesintheenvironmentcontinuously,andsearchthechangedParetooptimalsolutionsetasquicklyaspossible．Thiswillinevitablyrequiredynamicmulti—objectiveoptimizationalgorithmhashighsolutionefficiency．Especiallyinthefaceofincreasinglycomplexengineeringproblems，manyexisting

4、algorithmswillalsobecomepowerless．Tosolvethisproblem，ahighefficientdynamicmicromulti-objectivegeneticalgorithmbasedontheMicroMulti-objectivegeneticalgorithmissuggested．Thealgorithmusessmall-scaleevolutionarypopulationtoimprovethesolutionefficiency，andtrackschanges

5、oftheenvironmentthroughanintroductionofanenvironmentdetectionmechanismintheevolutionaryprocesstoensurethatthealgorithmsearchnewParetooptimalsolutionsetquickly．Whenthepopulationofthecurrentgenerationaccomplishesthegeneticmanipulation，thedetectionmechanismwillbestar

6、tedforenvironmenttesting．Calculatethevalueofalltheindividualobjectivefunctionsandtheconstraintfunctions，andcomparetheresultswiththecorrespondingfunctionvaluecalculatedinthepreviousgeneration．Ifthereisanydifference，weconsidertheproblemhaschangedovertime．Atthistime，

7、selecttheappropriatewaytoregenerateapopulation，andcombinedwiththesavedinformationsolutionsofpreviousgenerationandexternalpopulations，updatenon．dominatedsolutionssetintoanewevolutionaryprocess．Then，fourdifferenttypesofdynamicmulti-objectiveoptimizationtestproblemsa

8、reusedtotesttheperformanceoftheefficientdynamicmicromulti．objectivegeneticalgorithm．AndbycomparisonwithDNSGA-II，anexistingdynamicmulti。objectiveoptimiza