社會系統(tǒng)仿真課件

I.IntroductionM.PeterJurkatCS452/Mgt532SimulationforManagerialDecisionsTheRobertO.AndersonSchoolsofManagementUniversityofNewMexico

2/5/20231MPJ/UNMCS452/Mgt532I.IntroductionDefinitionsSimulation:processofexperimentingwithamodelofadynamicsystems(e.g.,process)tostudyortestthebehaviorofthesystemimprove,problemsolvedesignand/orselectnewsystems,and/ortrainoperatorsonamodelofanexistingsystemsSystem:purposeful,interrelatedcomponentswithinterdependenciesandcomplexityBehavior:purposeful,interrelatedsequencesofactivitiesDynamic:timevarying(staticsystemsaredull!)2/5/20232MPJ/UNMCS452/Mgt532I.IntroductionExamplesServiceSystems:TrafficonNetworks:messagesto/fromcomputers,carsonroads/rails,airplanesto/fromairports/gates,shipsto/fromharbors/piers,elevatorsRetail/Service:storessellinggoods,service/repairshops,logistics/inventory/distribution/MRPManufacturingSystems:Materials,Chemicals,BiologicalsAppliances,Automobiles/Trucks,Toys,ClothingElectronics,WeaponsSystemsComputationsusingmodelsfromotherdisciplinesMacroeconomic:taxation/interestratecost/benefitsPollution:environmentalinterventioncost/benefitsProjectManagement:completiontimevsresources2/5/20233MPJ/UNMCS452/Mgt532I.IntroductionWhySimulate?ToovercomehumanlimitationsinPhysicalcapability:avoidinjuryanddeath;beabletocontrolsystemswhosedynamicsarenotyetknown,…Mentalcapability:attention,memory,processing,…Analysis:allowsustostudysystemstoocomplexforanalyticdescriptionand/ortoodangerousforhumansafety–gainknowledgeDesign:attemptchangesinIVstodriveoneormoreDVstowardan“optimal”valueorcombinationofvaluesfordesign,improvement,and/orproblemsolving2/5/20234MPJ/UNMCS452/Mgt532I.IntroductionWhennottoSimulate!WhentheorycandeterminesufficientresultsWhenitwillcostmoretosimulatethanthereturnontheknowledgegainedWhenthereisincompleteinformationaboutthesystem(canhandleimprecisebutnotmissingpieces)Needatleastinputsandrelatedoutputsfor“blackboxes”Canassumemissinginformationandcheckagainstknownresults–ifagreement,supportforassumptionsWhenitisnotpossibletodeveloparepresentative,tractablesimplificationofthesystem2/5/20235MPJ/UNMCS452/Mgt532I.IntroductionDefinitions(cont.)Model:representationofasystem–threephases:Verbal–alwaysincludedinanyrepresentationGraphical–seepages22,39,50,54,367,and536Algorithmand/orcomputerprogramExperimentation:purposeful,structured,andcontrolledchangeoftheinputsfactors(independentvariables–IVs,exogenous,…)ofaproductand/orprocesstoobserveresultingchangesinoutputs(dependentvariables-DVs,responses,results,outcomes,…)BothIVs,DVsalsocalledmeasuresormetricsInsimulationliteraturearunisoneexecutionofthesimulationprogramatonecombinationofinputvariablevalues–alsocalledareplication2/5/20236MPJ/UNMCS452/Mgt532I.IntroductionGraphicalRepresentation:

LogicalSymbolsBCNN4thEd.,Figure2.1,page22:SingleServerQueuingSystem2/5/20237MPJ/UNMCS452/Mgt532I.IntroductionGraphicalRepresentation:

StateVariableTrackingBCNN4thEd.,Example2.2,Figure2.11,page392/5/20238MPJ/UNMCS452/Mgt532I.IntroductionGraphicalRepresentation:

PhysicalLayoutBCNN4thEd.,Example2.6,Figure2.15,page502/5/20239MPJ/UNMCS452/Mgt532I.IntroductionGraphicalRepresentation:

NetworkModelBCNN4thEd.,Example2.8,Figure2.18,page542/5/202310MPJ/UNMCS452/Mgt532I.IntroductionGraphicalRepresentation:

“BlackBox”BCNN4thEd.,Figure10.5,page3672/5/202311MPJ/UNMCS452/Mgt532I.IntroductionGraphicalRepresentation:

ComponentRelationshipBCNN4thEd.,Example14.4,Figure14.10,page536:Websiteconfiguration2/5/202312MPJ/UNMCS452/Mgt532I.IntroductionSimulationStudyRepresentation

(afterBanksetal,Figure1.3,Page15)ProblemFormulationSetObjectivesandProjectPlan(Re)ConceptualizeModel&CollectDataTranslateModelCanModelbeVerified?NoCanModelbeValidated?YesNoDOE-DesignExperimentsRunsandReplicationsAnalysisResultsClearandAbletobeDescribed?NoDocument,ReportandRecommendYesYes2/5/202313MPJ/UNMCS452/Mgt532I.IntroductionSimulationStudyIdentifyproblem(s),improvement(s),and/orplannewcapabilitiesSpecifythesystem–selectboundaries,identifyinputs,entities,attributes,events,activities,processes,andstatevariables-specifyoutput(s)andtheirdesiredvaluesBuildaconceptualandoperationalmodelofthesystem–buildarepresentationofinputs,entities,…2/5/202314MPJ/UNMCS452/Mgt532I.IntroductionSimulationStudy(cont.)VerifyandValidate(asbestyoucan)theoperationalmodelagainstexistingsystem–onlypartialmodelverification/validationmaybepossiblefornewsystemsPerformscreeningexperiment(s)toidentifyIVswithsignificanteffectondesiredoutput(s)–proceedwithonlytheseIVsSelectrangesofIVswhichreducevariabilitytoacceptablelevels,ifnecessary(CriticalStep!!!)Experimentwithmodeltoidentifyvaluesofinputswhich“optimize”outputor“achieve”goalBuildsystemorprototypetotestresultsofstudy2/5/202315MPJ/UNMCS452/Mgt532I.IntroductionSystemDescription,Problem,Objectives,ProjectPlanVerbaldescription/linguisticanalysisIdentifyproblemsand/or(re)designobjectivesIdentifyingrelevantEntitiesAttributesEventsActivities/processes,andstatevariables toaddressproblem(s)and/orobjectivesDevelopprojectplan–mayfollowSTEPSFOREXPERIMENTALDESIGNinSchmidtandLaunsbyonpagesI-26andI-272/5/202316MPJ/UNMCS452/Mgt532I.IntroductionSimulationModelComponentsEntities:namedphysical/conceptualobjects(impropernounsusedforUMLclasses,propernounsforUMLobjects)Attributes:namedcharacteristicorproperty(adjectives)Methods:namedactivitiesoroperationstheentitycanperform(predicates=verb+direct/indirectobject(s))States:namedsetofconditions,standings,circumstances,andpositionsdescribinganentityataparticulartime(adjectives,verbalnouns=gerunds)Processes:namedgroupsofactivitiesEvents:namednoteworthyoccurrences,oftenatthebeginningorcompletionofoneormoreactivitiesand/orprocesses2/5/202317MPJ/UNMCS452/Mgt532I.IntroductionIdentifyVariablesOutput(dependent)variableswhosevalueswillbetheproblemsolution/designimprovementOperationaldefinitionsRangeofvaluesAppropriateoutputanalysisTransientvs.steadystateStatisticaltools(confidenceintervals,t-tests,ANOVA,regression/modelbuilding)2/5/202318MPJ/UNMCS452/Mgt532I.IntroductionIdentifyVariables(cont.)FactorsamongwhosecombinationofvalueswillprovidetheproblemsolutionofoptimumdesignThesewillbevariedbytheinvestigatoraccordingtosomeexperimentaldesign(DOE)Operationaldefinitions,rangeofvalues,levelvalues,potentialinteractions(foreventualassignmenttoDOEcolumns)Factormodel:relatesfactorstooutputvariables–developedinmodelingexperiments2/5/202319MPJ/UNMCS452/Mgt532I.IntroductionIdentifyVariables(cont.)StatevariableswhosechangeofvaluesdeterminetheeventsOthervariablesnecessaryforacompletemodelIdentifystochasticvariablesandcollectdatatospecifytheirdistributionsIfclosetoknownmathematicaldistributionsthenidentifytheirparametersElseuseasempiricaldistributionsCollectdataforconstants–thesemayhavetobefittedfromthedata2/5/202320MPJ/UNMCS452/Mgt532I.Introduction(Re)ConceptualizeSimulationModelandCollectDataSimulationmodel:relatesallvariablestooutputvariablesRepresentationtools:naturalordomainspecificlanguage/jargonmathematicalnotationcode(e.g.,Java,GPSS)andpseudo-code(primitiveaction,choice,iteration)flowchartsUMLPERT/CPMdiagramspictorialimagesstoryboards/moviesBuildSimulationModelandtheSimulationitself2/5/202321MPJ/UNMCS452/Mgt532I.IntroductionVerifyandValidateVerifythatcalculationsinimplementationarecorrectValidatetheresultsagainstoutputknowntobeanaccuratereflectionofrealityMayonlybepossibleforpartsofthemodelorhighlyrestrictedsituationsIfnotmake“reasonablenesschecks”2/5/202322MPJ/UNMCS452/Mgt532I.IntroductionDesignandConductExperimentalRunsDoexperimentsScreen:experimentalruns(2-level?)tofindthesignificantfewfactorsModel:furtherornewsetofexperimentalruns(3or5levels)todevelopfactormodelequationsfitequationsbyregressionOptimize:solveequationsforoptimumormakeexperimentalrunstodrilldowntobestcombinationsoffactorsCheck:localoptimum(simulateallneighbors)2/5/202323MPJ/UNMCS452/Mgt532I.IntroductionSolutions/DesignIdentificationandReportFromsimulationrunsidentifythesolutiontotheproblemand/ortheoptimumdesignWriteReportAbstract(mayonlybeneededforresearchorarchivereports)ExecutiveSummary:non-technicalproblemstatement,solution/design,justification(notusuallyinresearchreports)TechnicalReport:completedetailssothatentireprojectcouldberepeatedbyothers–includingequations,code,distributions,runresultsTechnicalAppendix2/5/202324MPJ/UNMCS452/Mgt532I.IntroductionSimulationReport

SeeSimulationStudy&ReportOutline.docfordetailsofeachsectionAbstractExecutiveSummaryFullTechnicalReportSituation,Problems,Opportunities,Goals,andObjectivesBackgroundSystemSpecificationPerformanceMeasuresInputFactorsSystemRepresentation/ModelProjectActivitiesInputSpecificationandModelImplementationVerificationandValidationExperimentsandResultsoftheSimulationRunsAnalysisandResultsConclusionandRecommendationsTechnicalAppendix2/5/202325MPJ/UNMCS452/Mgt532I.IntroductionAssignmentsChooseoneapplicationfromBanks1.1oryourselectionforaDESSproject:writesections3.a)-c)ofthereport(specifytheentities&makeasymbolicrepresentationusingflowcharts,UML,or…).Thiscanbeagroupexercise.Individualexercises,Banks1.6Prepareabriefwrittenreport(includecopyofpapersifpossible)andPrepareanevenbriefersetofslidesforpresentationtotheclass(unlessthesubjectofyourpaperisparticularlyinterestingyoumaynotbeaskedto