欢迎来到天天文库
浏览记录
ID:31977598
大小:524.21 KB
页数:56页
时间:2019-01-29
《基于多智能体遗传算法的约束优化方法-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、AbstractIIIAbstractEvolutionaryAlgorithmprovidesanewwaytosolvecomplexoptimizationproblems.Becauseofitsintelligence,universality,robustnessandglobalsearchability,EAshavebeenwidelyusedinthisfieldandhaveagreatsuccessinrecentlyseveraldecades.Itiscommontofaceanumberofoptimization
2、problemsinmanyareasoftherealworld,especiallyinthescienceandengineeringfields.However,theseproblemsareoftenconstrained.Becauseofthedifferentfeaturesoftheseproblems,thetraditionalmethodsarehardtosolvetheseproblemseffectively.Asroustpopulation-basedglobalsearchmethods,Evolution
3、aryAlgorithms(EAs)areverypromisingtosolvetheconstrainedoptimizationproblems.TheaimofthisdissertationistoexplorethetheoriesandmechanismsofEAs,andtodothecorrespondingtheoreticandexperimentalanalyses.Themainresearchworkinthisdissertationconsistofthefollowingaspects.(1)Weextendt
4、hemultiagentgeneticalgorithm(MAGA)tosolveconstrainedoptimizationproblems(COPs)(MAGA_COPs)bycombiningtheneighborhoodcompetitionoperatorwithanefficientconstrainthandlingtechnique.Thismethodcanmakegooduseoftheinformationofinfeasiblesolutionswhichisaimatguidingthesearchtowardthe
5、globaloptimaofCOPs.Thisalgorithmistestedon12benchmarkfunctions,andtheresultshowsthat12benchmarkfunctionscanfindglobaloptima.(2)AnapproximationstrategyforfeasibleregionalsousedinMAGA_COPs,thismethodmakethesolutionoffunctionstoapproachtheglobaloptimalsolutionandeffectiveinprev
6、entingthealgorithmtrappingintolocaloptima.Thealgorithmistestedon12benchmarkfunctions,andtheresultshowsthatthealgorithmisoutperformsotherscomparedwithsomeotherstate-of-the-artalgorithms.(3)WeimproveMAGA_COPsbycombiningMAGA_COPswithtraditionalmethods.Hybridapproachbasedonmulti
7、agentgeneticalgorithmisgiven,inordertoovercometheslowerconvergenceofthestandardgeneticalgorithm,andinwhichthelocalsearchisweak.Thealgorithmistestedon12benchmarkfunctions,andtheresultshowsthatthealgorithmisanefficientandconvergenthybridgeneticalgorithm.(4)Weimprovethemultiage
8、ntgeneticalgorithm(MAGA)tosolvelayoutoptimizationbycombiningtheneighborhood
此文档下载收益归作者所有