水晶球软件使用CrystalBall

ProjectManagementSession9CrystalBallOPIM5270精选课件Session9GoalsUnderstandwhyriskmustbeanalyzedKnowpros/consforthreewaystoanalyzeriskIdentifyrandomvariablesinmodelsKnowthefourstepsofasimulationprocessGeneraterandomnumberswithCrystalBallUsethefourstepsofasimulationprocessExplainhowCrystalBallsupportsProj.Mgmt.精选课件Mostreal-worldbusinesssituationstodayareprobabilistic,butthedecisionmodelsusedtodealwiththemaredeterministic.Howtodealwithrandomness?IgnoreitSimplifyproblemtomakeitanalyticallytractable,getsolution,thenignorereal-lifecomplicationsFindawaytoobtainanapproximatesolutiontoreal-worldproblemsDealingwithRandomness精选课件MonteCarlosimulationisamethodbywhichapproximatesolutionsareobtainedtorealistic(andthereforecomplicated)problemsThisisincontrasttoanalyticalmethods,whichobtainexactsolutionstohighlystylizedproblemsTradeoffbetweenrigorandrelevanceMonteCarloSimulation精选课件Whatisthis?Y=f(X1,X2,…,Xk)Often,thevaluesforoneormore"input"cellsareunknownoruncertainThiscreatesuncertaintyaboutthevalueofthe"output"cellSimulationcanbeusedtoanalyzethesetypesofmodelsIntroductiontoSimulation精选课件Arandomvariableisanyvariablewhosevaluecannotbepredictedorsetwithcertainty.Many“inputcells”inspreadsheetmodelsareactuallyrandomvariables.Forexample:thefuturecostofrawmaterialsfutureinterestratesfuturenumberofemployeesinafirmexpectedproductdemandDecisionsmadeusinguncertaininformationofteninvolverisk.Whatrisks?RandomVariables&Risk精选课件Usingexpectedvaluesforuncertaincellstellsusnothingaboutthevariabilityoftheperformancemeasure.Supposean$1,000investmentisexpectedtoreturn$2,000intwoyears.Wouldyouinvestif...theoutcomescouldrangefrom$1,060to$4,000?theoutcomescouldrangefrom$0to$2,100?Alternativeswiththesameexpectedvaluemayinvolveverydifferentlevelsofrisk.WhyAnalyzeRisk?精选课件Best-Case/Worst-CaseAnalysisWhat-ifAnalysisSimulationMethodsofRiskAnalysis精选课件Bestcase-pluginthemostoptimisticvaluesforeachoftheuncertaincells.Worstcase-pluginthemostpessimisticvaluesforeachoftheuncertaincells.Thisiseasytodoandboundstheoutcomes,buttellsusnothingaboutthedistributionofpossibleoutcomeswithinthebestandworst-caselimits.Otherproblemsorbenefits?Best-Case/Worst-CaseAnalysis精选课件PossiblePerformanceMeasureDistributionsWithinaRange精选课件Plugindifferentvaluesfortheuncertaincellsandseewhathappens.Benefits:ThisiseasytodowithspreadsheetsOther?Problems:Valuesmaybechoseninabiasedway.Hundredsorthousandsofscenariosmayberequiredtogeneratearepresentativedistribution.Doesnotsupplythetangibleevidence(factsandfigures)neededtojustifydecisionstomanagement.What-IfAnalysis精选课件Valuesforuncertaincellsareselectedrandomly(andinanunbiasedmanner).Thecomputergenerateshundreds(orthousands)ofscenarios.Weanalyzethescenarioresultstobetterunderstandthebehavioroftheperformancemeasure.Allowsdecisionsbasedonsolidempiricalevidence.Simulation精选课件Properriskassessmentrequiressimulation.Simulationisa4stepprocess:1)Identifytheuncertaincellsinthemodel.2)ImplementappropriateRandomNumberGenerators(RNGs)foreachuncertaincell.3)Replicate复制themodelntimes,andrecordthevalueofthebottom-lineperformancemeasure.4)Analyzethesamplevaluescollectedontheperformancemeasure.SimulationContinued精选课件ARandomNumberGeneratorisamathematicalfunctionthatrandomlygenerates(returns)avaluefromaparticularprobabilitydistribution.WecanimplementRandomNumberGeneratorsforuncertaincellstoallowustosamplefromthedistributionofvaluesexpectedfordifferentcells.RandomNumberGenerators精选课件TheRAND()functionreturnsuniformlydistributedrandomnumbersbetween0.0and0.9999999.Supposewewanttosimulatetheactoftossingafaircoin.Let1represent“heads”and2represent“tails”.ConsiderthefollowingRNG:=IF(RAND()<0.5,1,2)HowRandomNumberGeneratorsWork精选课件CrystalBallprovidestwodifferentwaysforcreatingRandomNumberGeneratorsinspreadsheetsCrystalBallfunctionsUsedinformulaslikeanyotherExcelfunctionRequireCBtobeinstalledonthemachinedisplayingthespreadsheet&donotsupportallCBfunctionalityTheDistributionGalleryDisplayanumber(notaformula)inacellbutgeneratesrandomnumbersforthatcellwhensimulatingthemodelDoesnotrequireCBtobeinstalledonthemachinetodisplaythespreadsheet&supportsallCBfunctionalityGeneratingRandomNumberswithCrystalBall精选课件ClickDefineAssumptioniconSelectdistributionSpecifyparametersUsingtheDistributionGallery精选课件DiscreteProbabilityDistributions精选课件ContinuousProbabilityDistributions精选课件DefineMenuRunMenuAnalyzeMenuCrystalBallToolbarCrystalBallinMSExcel精选课件Determinewhichmodelinputsareuncertainanddefineaprobabilitydistribution.Identifywhichforecastsyouwanttoanalyze/measure(e.g.,NPV,Sigmalevel,processefficiency)RunSimulationAnalyzeResultsGenerateReportHowdoesCrystalBallwork?精选课件ThefirststeptousingCrystalBallistodeterminewhichmodelinputsareuncertain.Whichvaluesareestimates?Whichareaverages?Onceyouhaveidentifiedthese,youuseyourknowledgeoftheuncertaintyaroundtheinputtocreateaprobabilitydistributionforthatcell(whatCrystalBallcallsanassumption).CrystalBallletsyoudefinethesedistributionsusingtheDistributionGalleryDefineYourDistributions精选课件EntervarietyofparameterstodefinedistributionsCanfitdistributionstorawdataCancellreferenceallfieldsCancorrelatepairsofassumptionsMarkerlinesAssumptionDialog精选课件Thenextstepistoidentifyaforecast.Aforecastisaformulacellthatyouwanttomeasureandanalyze.Inthismodel,youselecttheNetProfit(cellC23).IdentifyYourForecasts精选课件NumberofsimulationtrialsperformedDisplayrangeCertainty(probability)thattheforecastwillreach$2,812,558Partswithinthespeclimitsareshowninblue,partsoutsidespeclimitsareshownredNumberofdatapointsdisplayedinthechartCrystalBallusesMonteCarlosimulationtorandomlygeneratethousandsofwhat-ifscenariosEachscenarioisthencapturedandpresentedinafrequencychart–(ForecastChart)RunSimulation精选课件What’sresponsibleformostofthevariationintheforecast?Thesensitivitychartshowstheinfluenceeachassumptioncellhasontheforecast.AnalyzeResults精选课件ReportsSelectapre-definedreportorcreateyourowncustomreport.Reportsnowincludenewstatisticsandmorecontroloverdataandcharts.ExtractDataYoucanextractdatafrombothforecastsandassumptionsandextractmultipletypesofdata.GenerateReports精选课件GotoCrystalBallinExcelLet’sBuildOurOwnModel精选课件Withsimulationsoftware(likeCrystalBall)youcanaccountfor,andmanageagainsttheuncertaintyof:Time(howlongtheprojectmaytake)Money(howmuchtheprojectmaycost)Scope(variancefromspecification)RelevancetoProjectManagement精选课件Simulationisusedtodescribethebehaviorofabottom-lineperformancemeasurewhenvaluesofoneormoreinputvariablesareuncertain.Often,someinputvariablesareunderthedecisionmakerscontrol.Simulation(CrystalBall)canalsoassistinfindingtheoptimalvaluesofthecontrollableinputvariables.(SimulationOptimization)AdditionalUsesofSimulation精选课件感谢亲观看此幻灯片，此课件部分内容来源于网络，如有侵权请及时联系我们删除，谢谢配合！