Crystal Ball.docx

Crystal Ball.docx

《Crystal Ball.docx》由会员分享，可在线阅读，更多相关《Crystal Ball.docx（14页珍藏版）》请在冰点文库上搜索。

CrystalBall

Monte-CarloSimulationwithCrystalBall®

TorunasimulationusingCrystalBall®:

1.SetupSpreadsheet

Buildaspreadsheetthatwillcalculatetheperformancemeasure（e.g.,profit）intermsoftheinputs（randomornot）.Forrandominputs,justenteranynumber.

2.DefineAssumptions—i.e.,randomvariables

Definewhichcellsarerandom,andwhatdistributiontheyshouldfollow.

3.DefineForecast—i.e.,outputorperformancemeasure

Definewhichcell（s）youareinterestedinforecasting（typicallytheperformancemeasure,e.g.,profit）.

4.ChooseNumberofTrials

Selectthenumberoftrials.IfyouwouldlaterliketogeneratetheSensitivityAnalysischart,choose“SensitivityAnalysis”underOptionsinRunPreferences.

5.RunSimulation

Runthesimulation.Ifyouwouldliketochangeparametersandre-runthesimulation,youshould“reset”thesimulation（clickonthe“ResetSimulation”buttononthetoolbarorintheRunmenu）first.

6.ViewResults

Theforecastwindowshowingtheresultsofthesimulationappearsautomaticallyafter（orduring）thesimulation.Manydifferentresultsareavailable（frequencychart,cumulativechart,statistics,percentiles,sensitivityanalysis,andtrendchart）.Theresultscanbecopiedintotheworksheet.

CrystalBallToolbar:

DefineDefineRunStartResetForecastTrend

AssumptionsForecastPreferencesSimulationSimulationWindowChart

WaltonBookstoreSimulationwithCrystalBall®

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*Q

Refund=$2.50*Max（Q–d,0）

Profit=Revenue–Cost+Refund

Step#1（SetupSpreadsheet）

WaltonBookstoreSimulationwithCrystalBall®

Step#2（DefineAssumptions—i.e.,randomvariables）

Selectthecellthatcontainstherandomvariable（B17）—colorcode（blue）:

andclickonthe“DefineAssumptions”buttonintoolbar（orintheCellmenu）:

Selecttypeofdistribution:

Provideparametersofdistributions:

WaltonBookstoreSimulationwithCrystalBall®

Step#3（DefineForecast—i.e.,output）

Selectthecellthatcontainstheoutputvariabletoforecast（F17）:

clickonthe“DefineForecast”buttonintoolbar（orintheCellmenu）,

andfillintheDefineForecastdialoguebox.

Step#4（ChooseNumberofTrials）

Clickonthe“RunPreferences”buttonintoolbar（orintheRunmenu）:

andselectthenumberoftrialstorun.

WaltonBookstoreSimulationwithCrystalBall®

Step#5（RunSimulation）

Clickonthe“StartSimulation”buttonintoolbar（orRunintheRunmenu）:

Step#6（ViewResults）

Theresultsofthesimulationcanbeviewedinavarietyofdifferentways（frequencychart,cumulativechart,statistics,andpercentiles）.ChoosedifferentoptionsundertheViewmenuintheforecastwindow.

TheresultscanbecopiedintoaworksheetorWorddocument（chooseCopyundertheEditmenuinthesimulationoutputwindow.

UsingTrendChartstoFindtheImpactofOrderQuantityonPotentialProfit

Defineseveralforecastcells（G14:

G18）forseveralpossibleorderquantities（Q=100,150,200,250,300）.Usethesamerandomorderquantityforeachtocomparethemmoreequally（i.e.,oneassumptioncellfordemand—C14—withtherestsetequaltoC14）.

Afterrunningthesimulation,choose“OpenTrendChart”intheRunmenu.Thischartgives“certaintybands”fortheforecastcells.10%ofthetime,theprojectdurationwillfallwithintheinnerband（lightblue）,25%ofthetimewithinthe2ndband（red）,50%ofthetimewithinthethirdband（green）,and90%ofthetimewithintheoutsideband（darkblue）.

ProjectManagement—GlobalOil

GlobalOilisplanningtomovetheircreditcardoperationtoDesMoines,IowafromtheirhomeofficeinDallas.Themoveinvolvesmanydifferentdivisionswithinthecompany.Realestatemustselectoneofthreeavailableofficesites.PersonnelhastodeterminewhichemployeesfromDallaswillmove,howmanynewemployeestohire,andwhowilltrainthem.Thesystemsgroupandtreasurer’sofficemustorganizethenewoperatingprocedureandmakefinancialarrangements.Thearchitectswillhavetodesigntheinteriorspace,andoverseeneededstructuralimprovements.Eachsiteisanexistingbuildingwithsufficientopenspace,butofficepartitions,computerfacilities,furnishings,andsoon,mustallbeprovided.

Acomplicatingfactoristhatthereisaninterdependenceofactivities.Inotherwords,somepartsoftheprojectcannotbestarteduntilotherpartsarecompleted.Forexample,Globalcannotconstructtheinteriorofanofficebeforeithasbeendesigned.Neithercanithirenewemployeesuntilithasdetermineditspersonnelrequirements.

Thenecessaryactivitiesandtheirnecessarypredecessors（duetointerdependence）arelistedbelow.Threeestimatesaremadeforthecompletiontimeofeachactivity—theminimumtime,mostlikelytime,andmaximumtime.

Immediate

TimeEstimates（days）

Activity

Description

Predecessor

Minimum

MostLikely

Maximum

A

SelectOfficeSite

—

21

21

21

B

CreateOrg.&Fin.Plan

—

20

25

30

C

DeterminePersonnelReq.

B

15

20

30

D

DesignFacility

A,C

20

28

42

E

ConstructFacility

D

40

48

66

F

SelectPersonneltoMove

C

12

12

12

G

HireNewEmployees

F

20

25

32

H

MoveKeyEmployees

F

28

28

28

I

TrainNewPersonnel

E,G,H

10

15

24

GlobalOilSimulationwithCrystalBall®

Step#1（SetupSpreadsheet）

Step#2（DefineAssumptions—i.e.,randomvariables）

Eachoftherandomactivitytimes（B,C,D,E,G,andI）isassumedtofollowthetriangulardistribution.

GlobalOilSimulationwithCrystalBall®

Step#3（DefineForecast—i.e.,output）

CellJ15istheforecastcell:

Step#4（ChooseNumberofTrials）

500trialswererun.Inaddition,SensitivityAnalysiswasenabledintheOptionsoftheRunPreferencesdialoguebox.Thisallowsforthegenerationofsensitivityanalysisresultslater.

Step#5（RunSimulation）

Step#6（ViewResults）

AdditionalResultsAvailablewithCrystalBall®

Slidethetrianglesbelowthehistogramstodeterminetheprobabilitythattheoutput（projectduration）islessthanacertainvalue（e.g.,adeadline）,greaterthanacertainvalue,orbetweenanytwovalues（byslidingbothtriangles）.

Alternatively,youcantypeinvaluesforthelowerboundorupperboundtodeterminetheprobability.Youcanalsotypeinaprobability（in“Certainty”）,anditwilldeterminetherangethathasthatprobability.

Thereisa79%chancetheprojectwillbecompletedwithin150days.Thereisa2.4%chancethattheprojectwilltakemorethan160days.

SensitivityChart

Choose“OpenSensitivityChart”intheRunmenu.Notethatthischartisonlyavailableifyouselectedthe“SensitivityAnalysis”optionunderRunPreferences.Thischartgivesanindicationastowhichrandomvariables（activitytimes）havethegreatestimpactontheoutputcell（projectcompletiontime）.

VariabilityinactivityEhasthegreatestimpactonoverallprojectduration,followedbyactivityD,C,I,andB.VariabilityinactivityGhasalmostnoimpact.

FittingaDistribution

CrystalBallcanbeusedto“fit”adistributiontodata.

Thefollowingdatahasbeencollectedfortheprevious100phonecallstoamail-orderhouse:

（80rowshavebeenhidden）

FittingDatatoaDistribution

UsingCrystalBall®tofitdatatoadistribution

1.Selectaspreadsheetcell.

2.ChooseDefineAssumption.

3.ClicktheFitbutton,thenselectthesourceofthefitteddata.

4.ClicktheNextbutton,thenselectthedistributionstotrytofit.

5.ClickOK.

InterarrivalTime

ServiceTime