资源描述：
《copula-based multivariate input models for stochastic simulation》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。
1、OPERATIONSRESEARCHinforms®Vol.57,No.4,July–August2009,pp.878–892issn0030-364Xeissn1526-54630957040878doi10.1287/opre.1080.0669©2009INFORMSCopula-BasedMultivariateInputModelsforStochasticSimulationBaharBillerTepperSchoolofBusiness,CarnegieMellonUnivers
2、ity,Pittsburgh,Pennsylvania15213,billerb@andrew.cmu.eduAslarge-scalediscrete-eventstochasticsimulationbecomesatoolthatisusedroutinelyforthedesignandanalysisofstochasticsystems,theneedforinput-modelingsupportwiththeabilitytorepresentcomplexinteractionsandi
3、nterde-pendenciesamongthecomponentsofmultivariatetime-seriesinputprocessesismorecriticalthanever.Motivatedbythefailureofindependentandidenticallydistributedrandomvariablestorepresentsuchinputprocesses,acomprehen-siveframeworkcalledVector-Autoregressive-To
4、-Anything(VARTA)hasbeenintroducedformultivariatetime-seriesinputmodeling.Despiteitsflexibilityincapturingawidevarietyofdistributionalshapes,weshowthatVARTAfallsshortinrepresentingdependencestructuresthatariseinsituationswhereextremecomponentrealizationsocc
5、urtogether.WedemonstratethatitispossibletoextendVARTAtoworkforsuchdependencestructuresviatheuseofthecopulatheory,whichhasbeenusedprimarilyforrandomvectorsinthesimulationinput-modelingliterature,formultivariatetime-seriesinputmodeling.Weshowthatourcopula-b
6、asedmultivariatetime-seriesinputmodel,whichincludesVARTAasaspecialcase,allowsthedevelopmentofstatisticallyvalidfittingandfastsamplingalgorithmswellsuitedfordrivinglarge-scalestochasticsimulations.Subjectclassifications:correlation;estimation;sampling;timese
7、ries.Areaofreview:Simulation.History:ReceivedJuly2006;revisionsreceivedJune2007,January2008;acceptedFebruary2008.PublishedonlineinArticlesinAdvanceJune3,2009.1.IntroductionAcloselookattheexistingsimulationinput-modelingliteraturerevealsthatmuchoftheprevio
8、usworkontime-Animportantstepinthedesignofstochasticsimulationisseriesinputprocessesisbasedonlinear,univariatetime-inputmodeling,i.e.,modelingtheuncertaintyintheinputseriesmodelssuchastheautoregressivemovingaverageen
