欢迎来到天天文库
浏览记录
ID:37370945
大小:2.65 MB
页数:58页
时间:2019-05-22
《改进的并行遗传算法在知识库中的应用研究》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、大连交通大学硕士学位论文改进的并行遗传算法在知识库中的应用研究姓名:赵戈申请学位级别:硕士专业:计算机应用技术指导教师:梁旭20081025大连交通人学1:学硕十学佗论文AbstractSchedulingtheorybelongstocombinationoptimizationproblem.Introducingtheoptimizationtheorytothefieldofjobschedulinginworkshopcallimprovetheperformanceofalgorithm,makethealgorithmapplytob
2、roaderfieldsandcompletethewholesystemofthealgorithm,whichisasubjectincludingboththeoreticalmeaningandpracticalvalues.Theparallelgeneticalgorithmhasalwaysbeenthehotspotoftheresearchongeneticalgorithm.Thisarticleproposesallimprovedparallelgeneticalgorithmbaseduponthetraditionalc
3、oarsegraingeneticalgorithm.Theproposedalgorithmimprovedthetraditionalonemainlyintwoaspects:ononehand,itemploysamechanismwhichCandynamicallyadjusteachisland’Smigrationrateaccordingtoitsaveragefitness.Therefore,ifoneisland’Saveragefitnessishigh,itsdiversi黟willbeincreasedbyincrea
4、singotherislands’migrationratetothisislandinordertoavoidprematurity;theimprovedalgorithmalsocalculatesthesurvivingperiodforeachchromosomeinvolvedinmigration,notonlyenablethechromosomewithhighfitnesstosurvivelongerbutalsoavoidtheexplosionofthepopulationofsomeisland.Theseimprove
5、dmeasuresareofgreatsignificanceonavoidingprematurityofthealgorithm,aswellasincreasingglobalsearchingabilityofthealgorithm.Theimprovedalgorithmistestedthroughusingexamplesinthestandardtestsetandtheactualprobleminscheduling.Thesimulationresultsofthetestshowthatthisalgorithmcanco
6、mpletethetaskofthefastsearchinthegivenrangeandglobaloptimization.Inthemeantime,accordingtothepracticalsituationofthemanufacturingbusiness,thisarticledesignsanewmethodtoinitializethepopulmionwhichCanapplytopracticaljobshopproblem.Ifcomparedwiththeconventionalpopulationinitializ
7、ingalgorithmwhichonlyrandomlydividestheinitialpopulmiontomanysmallergroups,thenewalgorithmismuchmoreeffectiveintheaspectofincreasingthediversityofthepopulation.Thisarticledesignedandimplementedanintelligentschedulingknowledgebasesystemplatformforworkshop,whichistargetingatsome
8、factory’Spracticalproblem.Furthermore,theimprovedalgorithmhas
此文档下载收益归作者所有