风险管理软件Crystal Ball使用指导.docx

编号：时间：2021年x月x日书山有路勤为径，学海无涯苦作舟页码：第页共NUMPAGES22页第PAGE\*MERGEFORMAT22页共NUMPAGES\*MERGEFORMAT22页Monte-CarloSimulationwithCrystalBall®TorunasimulationusingCrystalBall®:1.SetupSpreadsheetBuildaspreadsheetthatwillcalculatetheperformancemeasure(e.g.,profit)intermsoftheinputs(randomornot).Forrandominputs,justenteranynumber.2.DefineAssumptions—i.e.,randomvariablesDefinewhichcellsarerandom,andwhatdistributiontheyshouldfollow.3.DefineForecast—i.e.,outputorperformancemeasureDefinewhichcell(s)youareinterestedinforecasting(typicallytheperformancemeasure,e.g.,profit).4.ChooseNumberofTrialsSelectthenumberoftrials.IfyouwouldlaterliketogeneratetheSensitivityAnalysischart,choose“SensitivityAnalysis”underOptionsinRunPreferences.5.RunSimulationRunthesimulation.Ifyouwouldliketochangeparametersandre-runthesimulation,youshould“reset”thesimulation(clickonthe“ResetSimulation”buttononthetoolbarorintheRunmenu)first.6.ViewResultsTheforecastwindowshowingtheresultsofthesimulationappearsautomaticallyafter(orduring)thesimulation.Manydifferentresultsareavailable(frequencychart,cumulativechart,statistics,percentiles,sensitivityanalysis,andtrendchart).Theresultscanbecopiedintotheworksheet.CrystalBallToolbar:DefineDefineRunStartResetForecastTrendAssumptionsForecastPreferencesSimulationSimulationWindowChartWaltonBookstoreSimulationwithCrystalBall®RecalltheWaltonBookstoreexample:ItisAugust,andtheymustdecidehowmanyofnextyear’snaturecalendarstoorder.Eachcalendarcoststhebookstore$7.50andissoldfor$10.AfterFebruary,allunsoldcalendarsarereturnedtothepublisherforarefundof$2.50percalendar.SupposeWaltonpredictsdemandwillbesomewherebetween100and300(discreteuniform).Demand=d~Uniform[100,300]OrderQuantity=Q(decisionvariable)Revenue=$10*Min(Q,d)Cost=$7.50*QRefund=$2.50*Max(Q–d,0)Profit=Revenue–Cost+RefundStep#1(SetupSpreadsheet)WaltonBookstoreSimulationwithCrystalBall®Step#2(DefineAssumptions—i.e.,randomvariables)Selectthecellthatcontainstherandomvariable(B17)—colorcode(blue):andclickonthe“DefineAss