OperationsManagementPlanning

Background information

Simple Moving Average (MA)

Thesimplemovingaverageconcept is based on the ideaofaveragingrandom fluctuations in atime series to identifytheunderlyingdirection inwhich thetime series is changing. Amovingaverage forecastis an averageofthemostrecent observations in a time series.

Step 1

Determine themodel's time frame(e.g., 3 periods, 5 periods etc.).In this example we will usea3 period movingaverage.

Step 2

Add the first 3 observations together and divideby3. The answeris the forecast forthefourthperiod.

Step 3

Toprogress further, add the observation numbers2, 3 and 4 together and divideby3. Theansweris the forecastforthe fifth period. Repeatsteps 1 through 3as necessary.

Assignment#2: Operations Management Planning

Background Information(Continued)

Forecast Errors

All forecasts aresubject to error.Understandingthe natureand sizeoferrors is important tomaking good decisions.

Forecasterroris the difference between the observed valueofthetime series and the forecast,or At-Ft.Takealook atthe table seen inthechartbelow.

Suppose that a forecastingmethodprovided the forecasts in column Eofthe sample chart. Theforecast errors arecomputed in column F.Becauseofthe inherent inabilityofanymodel toforecast accurately,weuse quantitative measuresof forecast accuracyto evaluatehowwell theforecastingmodel performs. Clearly, we want touse models thathavesmall forecast errors.

Generally, three types offorecast error metricsareused:Mean squareerror (MSE),Meanabsolute deviation (MAD) and Mean absolute percentage error(MAPE).

Assignment#2: Operations Management Planning

Background Information(Continued)

Mean SquareError(MSE)

Mean square error, orMSE, is calculated bysquaringthe individual forecast errors and thenaveragingtheresults over all T periods ofdatain thetime series.

The formulais:

In other words,youarecalculatingtheaveragevalueofthe24 "squared error" entries. Forthedatain thegraphic, this is computed in column H.Thesum ofthe squared errors is

87,910.6, and thereforeMSEis 87,910.6 / 24 =3,662.94. MSEis probablythemost commonlyused measureof forecast accuracy.

Mean Absolute Deviation (MAD)

Another common measureof forecastaccuracyis the mean absolutedeviation (MAD), computedas:

The measureis simplythe averageofthesum ofthe absolute deviations for all the forecasterrors. Usingtheinformation in columnJofthe tablein the graphic, wecompute MAD as: 1,197

/ 24 =49.88

Assignment #2: Operations Management Planning

Background Information(Continued)

Mean Absolute Percentage Error(MAPE)

A third measureof forecast erroris themean absolute percentage error. The formulais:

This is simplytheaverageofthe percentage errorfor each forecast value inthetime series.

Step 1:TaketheActualAmount subtract the forecast(assumethe valuetoalways be positive)=X

Step 2:TakeXand dividebythe Actual Amountandexpress this as a percentage(EG: Multiplyby100).

Step 3:RepeatSteps 1 and 2 for each timeperiod. Add the percentagesand then dividethis totalbythe total numberoftime periods.

Judgmental forecasting:

Judgmental forecasting relies on opinions andexpertiseof peoplein developingforecasts. Whenno historical dataareavailable, judgmental forecastingis theonlymethodpossible. But evenwhenhistorical dataareavailable and appropriate,theycannot bethesolebasis for prediction.

The demand forgoods and services is affected byavarietyoffactors, suchasglobal markets andcultures,interest rates, disposable income, inflation, and technology. Competitors'actions andgovernment regulations also have an impact.Thus,some element of judgmental forecastingisalwaysnecessary.

Oneexampleofthe roleof judgmentalforecastingoccurred duringthenational recession. Alleconomic indicators pointed toward afutureperiod oflow demand for manufacturers ofmachinetools. However, theforecasters ofonesuch companyrecognized that recentgovernmentregulationsforautomobilepollution control would requiretheauto industryto update its currenttechnologybypurchasingnewtools. As a result, this machinetool companywas preparedforthenew business.

Increasingor decreasing prices and promotingproducts aretools used bythe marketingdepartment to shapedemand. Theyneed to assist in developingjudgemental forecasting ratherthan simplyrelyingon historicalpatterns.

Assignment#2: Operations Management Planning

Assignmentrequirements:

Youareamanageratamediumsizedmanufacturingoperationthatsellsacustomizedproductdirectlytotheconsumersandthedatapatternislinearbutitcanhavetheoccasionalrandomness.Thecausesofthisrandomnessareneverapparentuntilaftertheevent.

Asmanager,youaretocompleteacorporateforecastforthisexpensiveandcustomizedproductthattrendsgenerallywiththecurrenteconomicsituation.

Requirement1

Completethecorporateforecastforthisproductusingthedatabelowtocompletea3and5yearsimplemovingaverageforecastforthenextyear(timeperiod8).Forthejudgmentalforecastingcomponentusethecurrenteconomicsituation.

TimePeriod	HistoricalDemand
1	473
2	425
3	708
4	493
5	480
6	569
7	453

Requirement2

Selectthebetterforecastingmodel(EG:3yearor5yearsimplemovingaverage)andsupportyouranswer.

Requirement3

TheplantisusingthechaseproductionoptionandtheAssembletoOrder(ATO)processdesign.TheBillofMaterial(BOM)isillustratedbelow.

ExplodetheaboveBOMforeachitemtodeterminetherequirementsofeachitembasedontheGrossRequirementof150unitsforItem'A';theparentitem.

Assignment#2: Operations Management Planning

Assignmentrequirements(Continued)

Requirement4

Thefactoryworksonaprocesslayoutbuttheorderlotsizingisopenfordetermination.Determinetheoptimalmethodforlotsizingusingyourbestjudgementofthelimiteddataavailableandsupportyouranswer.