Self-adaptive large neighborhood search algorithm for parallel machine scheduling problems

在线阅读 下载PDF 导出详情
摘要 Aself-adaptivelargeneighborhoodsearchmethodforschedulingnjobsonmnon-identicalparallelmachineswithmultipletimewindowsispresented.Theproblems'anotherfeatureliesinoversubscription,namelynotalljobscanbescheduledwithinspecifiedschedulinghorizonsduetothelimitedmachinecapacity.Theobjectiveisthustomaximizetheoverallprofitsofprocessedjobswhilerespectingmachineconstraints.Afirst-infirst-outheuristicisappliedtofindaninitialsolution,andthenalargeneighborhoodsearchprocedureisemployedtorelaxandreoptimizecumbersomesolutions.Amachinelearningmechanismisalsointroducedtoconvergeonthemostefficientneighborhoodsfortheproblem.Extensivecomputationalresultsarepresentedbasedondatafromanapplicationinvolvingthedailyobservationschedulingofafleetofearthobservingsatellites.Themethodrapidlysolvesmostprobleminstancestooptimalornearoptimalandshowsarobustperformanceinsensitiveanalysis.
机构地区 不详
出版日期 2012年02月12日(中国期刊网平台首次上网日期,不代表论文的发表时间)