资源描述:
《基于遗传算法的转炉炉修优化研究》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、ShanghaiJiaoTongUniversityMasterDegreeofEngineeringAbstractStudyonOptimizationofConverterRepairProjectBasedonGeneticAlgorithmAbstractThenetworkplanningtechnologyapplicationinengineeringprojectmanagementhasbeenwidelyused.Ittendtooptimizesingleortwogoals,B
2、utitismoredifficulttoachieveoptimizationwhendealwithmultipletargetssuchastime,cost,qualityetc.Intheselectionofnetworkplanningoptimization,thetargetsmustbediscrete.Itmustbeone-to-onecorrespondenceinmultipleobjectiveoptimization,somoreworkmustbedonewhensta
3、rttosolve.Forthelimitationsinstructureofthemodel,itcannotcompletelycovertheoptimalsolutionscheme.Multi-objectivegeneticalgorithmisdevelopedinrecentyears,itisaneffectivemethodinthesolutionofmulti-objectiveoptimizationandnon-inferior.Becauseofitshigheffici
4、ency,practicalfeatures,sotherearemoreandmoreacademicattentions.ThispaperisbasedonthepracticeinShanghaiBaosteelsteelmakingconverterproject.Itputforwardtheoptimizingstructurewhichisbasedonamulti-objectivegeneticalgorithm,auxiliarynetworkplanandcurvefitting
5、methodsetc.tobuildthemulti-objectiveoptimizationmodelinconverterrepairingprojectmanagement.Consideringgeneticsearchcanbedoneincontinuousinterval,whendesignthemodelcode,onlyeverycomponent’stimelimitsarejudgedascodingfactors----todealwiththeproblemthatyouh
6、avetobuildone-to-onecorrespondencediscretemodelwhenyouusemultiplecodingfactortocode.Itcanreducethesubjectivefactors’interferencetooptimalsolution.Intheknownarrangementsforthedurationofeachsub-project,youcansolvethecorrespondingcriticalpathanditscorrespon
7、dingtotaldurationaccordingtothenetworkplanningtechnique.Afteranalyzedthekeyproblemoffurnacerepairprojectmanagement,youcangetthemutualinfluenceamongthetimelimit,costandquality,andcorrespondingequationcanalsobebuiltthroughthecurvefitting,withtheadditionofe
8、verycomponent’sduration,youcangettheprojectcostandqualityobjectives.Takingtimelimit,costandqualityasfitnessfunction,transformingthefurnacerepairprojectmanagementintoamulti-objectiveoptimizationproblem,usingtheNSGA-IIinmult