Date post: | 06-Apr-2018 |
Category: |
Documents |
Upload: | poorti-garg |
View: | 223 times |
Download: | 0 times |
of 99
8/2/2019 Ch3. Demand Forecasting
1/99
1
1 2002 South-Western/Thomson Learning2002 South-Western/Thomson Learning TMTM
Slides preparedSlides prepared
by John Loucksby John Loucks
8/2/2019 Ch3. Demand Forecasting
2/99
2
Chapter 3
Demand Forecasting
8/2/2019 Ch3. Demand Forecasting
3/99
3
Overview
IntroductionQualitative Forecasting MethodsQuantitative Forecasting Models
How to Have a Successful Forecasting SystemComputer Software for ForecastingForecasting in Small Businesses and Start-UpVenturesWrap-Up: What World-Class Producers Do
8/2/2019 Ch3. Demand Forecasting
4/99
4
Demand Management
Independent demand items are the onlyitems demand for which needs to beforecastThese items include:
Finished goods andSpare parts
8/2/2019 Ch3. Demand Forecasting
5/99
5
Demand Management
A
Independent Demand(finished goods and spare parts)
B(4) C(2)
D(2) E(1) D(3) F(2)
Dependent Demand(components)
8/2/2019 Ch3. Demand Forecasting
6/99
6
Demand Management
The importance of forecasting in OM
8/2/2019 Ch3. Demand Forecasting
7/99
7
Introduction
Demand estimates for products and services are thestarting point for all the other planning in operationsmanagement.Management teams develop sales forecasts based inpart on demand estimates.The sales forecasts become inputs to both businessstrategy and production resource forecasts.
8/2/2019 Ch3. Demand Forecasting
8/99
8
Forecasting is an Integral Partof Business Planning
ForecastMethod(s)
DemandEstimates
SalesForecast
ManagementTeam
Inputs:Market,
Economic,Other
BusinessStrategy
Production ResourceForecasts
8/2/2019 Ch3. Demand Forecasting
9/99
9
Some Reasons WhyForecasting is Essential in OM
New Facility Planning It can take 5 years to designand build a new factory or design and implement anew production process.Production Planning Demand for products varyfrom month to month and it can take several monthsto change the capacities of production processes.Workforce Scheduling Demand for services (and
the necessary staffing) can vary from hour to hourand employees weekly work schedules must bedeveloped in advance.
8/2/2019 Ch3. Demand Forecasting
10/99
10
Examples of Production Resource Forecasts
ForecastHorizon Time Span Item Being Forecast Units of Measure
Long-Range Years
Product linesFactory capacitiesPlanning for new productsCapital expenditures
Facility location or expansionR&D
Dollars, tons, etc.
Medium-Range Months
Product groupsDepartment capacitiesSales planningProduction planning and budgeting
Dollars, tons, etc.
Short-Range Weeks
Specific product quantitiesMachine capacitiesPlanningPurchasingSchedulingWorkforce levelsProduction levels
Job assignments
Physical units of products
8/2/2019 Ch3. Demand Forecasting
11/99
11
Forecasting Methods
Qualitative ApproachesQuantitative Approaches
8/2/2019 Ch3. Demand Forecasting
12/99
12
Qualitative Approaches
Usually based on judgments about causal factors thatunderlie the demand of particular products or servicesDo not require a demand history for the product orservice, therefore are useful for new products/servicesApproaches vary in sophistication from scientificallyconducted surveys to intuitive hunches about futureevents
The approach/method that is appropriate depends on a products life cycle stage
8/2/2019 Ch3. Demand Forecasting
13/99
13
Qualitative Methods
Educated guess intuitive hunches Executive committee consensusDelphi method
Survey of sales forceSurvey of customersHistorical analogyMarket research scientifically conducted surveys
8/2/2019 Ch3. Demand Forecasting
14/99
14
Qualitative Forecasting ApplicationsSmall and Large Firms
Technique Low Sales(less than $100M)
High Sales(more than $500M)
Managers Opinion 40.7% 39.6%
ExecutivesOpinion 40.7% 41.6%
Sales ForceComposite 29.6% 35.4%
Number of Firms 27 48
Source: Nada Sanders and Karl Mandrodt (1994) Practitioners Continue to Rely on Judgmental ForecastingMethods Instead of Quantitative Methods, Interfaces , vol. 24, no. 2, pp. 92-100.Note: More than one response was permitted.
8/2/2019 Ch3. Demand Forecasting
15/99
15
Quantitative Forecasting Approaches
Based on the assumption that the forces thatgenerated the past demand will generate the futuredemand, i.e., history will tend to repeat itself Analysis of the past demand pattern provides a goodbasis for forecasting future demandMajority of quantitative approaches fall in thecategory of time series analysis
8/2/2019 Ch3. Demand Forecasting
16/99
16
Quantitative Forecasting ApplicationsSmall and Large Firms
Technique Low Sales(less than $100M)
High Sales(more than $500M)
Moving Average 29.6% 29.2
Simple Linear Regression 14.8% 14.6
Naive 18.5% 14.6
Single ExponentialSmoothing 14.8% 20.8
Multiple Regression 22.2% 27.1
Simulation 3.7% 10.4
Classical Decomposition 3.7% 8.3
Box-Jenkins 3.7% 6.3Number of Firms 27 48
Source: Nada Sanders and Karl Mandrodt (1994) Practitioners Continue to Rely on Judgmental ForecastingMethods Instead of Quantitative Methods, Interfaces , vol. 24, no. 2, pp. 92-100.Note: More than one response was permitted.
8/2/2019 Ch3. Demand Forecasting
17/99
17
A time series is a set of numbers where the order orsequence of the numbers is important, e.g., historicaldemandAnalysis of the time series identifies patternsOnce the patterns are identified, they can be used todevelop a forecast
Time Series Analysis
8/2/2019 Ch3. Demand Forecasting
18/99
18
Components of Time Series
Trends are noted by an upward or downward slopinglineSeasonality is a data pattern that repeats itself overthe period of one year or lessCycle is a data pattern that repeats itself... may takeyearsIrregular variations are jumps in the level of the series
due to extraordinary eventsRandom fluctuation from random variation orunexplained causes
8/2/2019 Ch3. Demand Forecasting
19/99
19
Seasonal Patterns
Length of Time Number of Before Pattern Length of Seasons
Is Repeated Season in Pattern
Year Quarter 4Year Month 12Year Week 52
Month Day 28-31Week Day 7
8/2/2019 Ch3. Demand Forecasting
20/99
20
Quantitative Forecasting Approaches
Linear RegressionSimple Moving AverageWeighted Moving Average
Exponential Smoothing (exponentially weightedmoving average)Exponential Smoothing with Trend (doubleexponential smoothing)
8/2/2019 Ch3. Demand Forecasting
21/99
21
Long-Range Forecasts
Time spans usually greater than one yearNecessary to support strategic decisions aboutplanning products, processes, and facilities
8/2/2019 Ch3. Demand Forecasting
22/99
22
Simple Linear Regression
Linear regression analysis establishes a relationshipbetween a dependent variable and one or moreindependent variables.In simple linear regression analysis there is only oneindependent variable.If the data is a time series, the independent variable isthe time period.
The dependent variable is whatever we wish toforecast.
8/2/2019 Ch3. Demand Forecasting
23/99
23
Simple Linear Regression
Regression EquationThis model is of the form:
Y = a + bX
Y = dependent variableX = independent variablea = y-axis interceptb = slope of regression line
8/2/2019 Ch3. Demand Forecasting
24/99
24
Simple Linear Regression
Constants a and bThe constants a and b are computed using thefollowing equations:
2
2 2x y- x xya =n x -( x)
2 2
xy- x yb = n x -( x)
n
8/2/2019 Ch3. Demand Forecasting
25/99
25
Simple Linear Regression
Once the a and b values are computed, a future valueof X can be entered into the regression equation and acorresponding value of Y (the forecast) can becalculated.
8/2/2019 Ch3. Demand Forecasting
26/99
26
Example: College Enrollment
Simple Linear RegressionAt a small regional college enrollments have grownsteadily over the past six years, as evidenced below.Use time series regression to forecast the studentenrollments for the next three years.
Students StudentsYear Enrolled (1000s) Year Enrolled (1000s)
1 2.5 4 3.22 2.8 5 3.33 2.9 6 3.4
8/2/2019 Ch3. Demand Forecasting
27/99
27
Example: College Enrollment
Simple Linear Regression x y x 2 xy1 2.5 1 2.5
2 2.8 4 5.63 2.9 9 8.74 3.2 16 12.85 3.3 25 16.5
6 3.4 36 20.4S x=21 S y=18.1 S x2=91 S xy=66.5
8/2/2019 Ch3. Demand Forecasting
28/99
28
Example: College Enrollment
Simple Linear Regression
Y = 2.387 + 0.180X
291(18.1) 21(66.5) 2.387
6(91) (21)a
6(66.5) 21(18.1) 0.180105
b
8/2/2019 Ch3. Demand Forecasting
29/99
29
Example: College Enrollment
Simple Linear RegressionY7 = 2.387 + 0.180(7) = 3.65 or 3,650 studentsY8 = 2.387 + 0.180(8) = 3.83 or 3,830 students
Y9 = 2.387 + 0.180(9) = 4.01 or 4,010 students
Note: Enrollment is expected to increase by 180students per year.
8/2/2019 Ch3. Demand Forecasting
30/99
30
Simple Linear Regression
Simple linear regression can also be used when theindependent variable X represents a variable otherthan time.In this case, linear regression is representative of aclass of forecasting models called causal forecastingmodels.
8/2/2019 Ch3. Demand Forecasting
31/99
31
Example: Railroad Products Co.
Simple Linear Regression Causal ModelThe manager of RPC wants to project the firms
sales for the next 3 years. He knows that RPCs long -range sales are tied very closely to national freight carloadings. On the next slide are 7 years of relevanthistorical data.
Develop a simple linear regression model
between RPC sales and national freight car loadings.Forecast RPC sales for the next 3 years, given that therail industry estimates car loadings of 250, 270, and300 million.
8/2/2019 Ch3. Demand Forecasting
32/99
32
Example: Railroad Products Co.
Simple Linear Regression Causal ModelRPC Sales Car Loadings
Year ($millions) (millions)
1 9.5 1202 11.0 1353 12.0 1304 12.5 150
5 14.0 1706 16.0 1907 18.0 220
8/2/2019 Ch3. Demand Forecasting
33/99
33
Example: Railroad Products Co.
Simple Linear Regression Causal Modelx y x 2 xy
120 9.5 14,400 1,140
135 11.0 18,225 1,485130 12.0 16,900 1,560150 12.5 22,500 1,875170 14.0 28,900 2,380190 16.0 36,100 3,040220 18.0 48,400 3,960
1,115 93.0 185,425 15,440
8/2/2019 Ch3. Demand Forecasting
34/99
34
Example: Railroad Products Co.
Simple Linear Regression Causal Model
Y = 0.528 + 0.0801X
2185,425(93) 1,115(15,440)a 0.528
7(185,425) (1,115)
27(15,440) 1,115(93)b 0.08017(185,425) (1,115)
8/2/2019 Ch3. Demand Forecasting
35/99
35
Example: Railroad Products Co.
Simple Linear Regression Causal ModelY8 = 0.528 + 0.0801(250) = $20.55 millionY9 = 0.528 + 0.0801(270) = $22.16 million
Y10 = 0.528 + 0.0801(300) = $24.56 million
Note: RPC sales are expected to increase by$80,100 for each additional million national freightcar loadings.
8/2/2019 Ch3. Demand Forecasting
36/99
36
Multiple Regression Analysis
Multiple regression analysis is used when there aretwo or more independent variables.An example of a multiple regression equation is:
Y = 50.0 + 0.05X 1 + 0.10X 2 0.03X 3 where: Y = firms annual sales ($millions)
X1 = industry sales ($millions)
X2 = regional per capita income ($thousands)X3 = regional per capita debt ($thousands)
8/2/2019 Ch3. Demand Forecasting
37/99
37
Coefficient of Correlation ( r )
The coefficient of correlation, r , explains the relativeimportance of the relationship between x and y.The sign of r shows the direction of the relationship.The absolute value of r shows the strength of therelationship.The sign of r is always the same as the sign of b.r can take on any value between 1 and +1.
8/2/2019 Ch3. Demand Forecasting
38/99
38
Coefficient of Correlation ( r )
Meanings of several values of r :-1 a perfect negative relationship (as x goes up, y
goes down by one unit, and vice versa)+1 a perfect positive relationship (as x goes up, y
goes up by one unit, and vice versa)0 no relationship exists between x and y
+0.3 a weak positive relationship
-0.8 a strong negative relationship
8/2/2019 Ch3. Demand Forecasting
39/99
39
Coefficient of Correlation ( r )
r is computed by:
2 2 2 2( ) ( )
n xy x yr
n x x n y y
8/2/2019 Ch3. Demand Forecasting
40/99
40
Coefficient of Determination ( r 2)
The coefficient of determination, r 2
, is the square of the coefficient of correlation.The modification of r to r 2 allows us to shift fromsubjective measures of relationship to a more specificmeasure.r 2 is determined by the ratio of explained variation tototal variation:
22
2( )( )Y yr y y
8/2/2019 Ch3. Demand Forecasting
41/99
41
Example: Railroad Products Co.
Coefficient of Correlationx y x 2 xy y 2
120 9.5 14,400 1,140 90.25
135 11.0 18,225 1,485 121.00130 12.0 16,900 1,560 144.00150 12.5 22,500 1,875 156.25170 14.0 28,900 2,380 196.00190 16.0 36,100 3,040 256.00220 18.0 48,400 3,960 324.00
1,115 93.0 185,425 15,440 1,287.50
8/2/2019 Ch3. Demand Forecasting
42/99
42
Example: Railroad Products Co.
Coefficient of Correlation
r = .9829
2 2
7(15,440) 1,115(93)
7(185,425) (1,115) 7(1,287.5) (93)r
8/2/2019 Ch3. Demand Forecasting
43/99
43
Example: Railroad Products Co.
Coefficient of Determinationr 2 = (.9829) 2 = .966
96.6% of the variation in RPC sales is explained by
national freight car loadings.
8/2/2019 Ch3. Demand Forecasting
44/99
44
Ranging Forecasts
Forecasts for future periods are only estimates and aresubject to error.One way to deal with uncertainty is to develop best-estimate forecasts and the ranges within which theactual data are likely to fall.The ranges of a forecast are defined by the upper andlower limits of a confidence interval.
8/2/2019 Ch3. Demand Forecasting
45/99
8/2/2019 Ch3. Demand Forecasting
46/99
46
Ranging Forecasts
The standard error (deviation) of the forecast iscomputed as:
2
yx
y - a y - b xys = n - 2
8/2/2019 Ch3. Demand Forecasting
47/99
47
Example: Railroad Products Co.
Ranging ForecastsRecall that linear regression analysis provided a
forecast of annual sales for RPC in year 8 equal to$20.55 million.
Set the limits (ranges) of the forecast so that thereis only a 5 percent probability of exceeding the limitsby chance.
8/2/2019 Ch3. Demand Forecasting
48/99
48
Example: Railroad Products Co.
Ranging ForecastsStep 1: Compute the standard error of the
forecasts, s yx.
Step 2: Determine the appropriate value for t.
n = 7, so degrees of freedom = n 2 = 5.Area in upper tail = .05/2 = .025Appendix B, Table 2 shows t = 2.571.
1287.5 .528(93) .0801(15,440) .57487 2yx
s
8/2/2019 Ch3. Demand Forecasting
49/99
49
Example: Railroad Products Co.
Ranging ForecastsStep 3: Compute upper and lower limits.
Upper limit = 20.55 + 2.571(.5748)
= 20.55 + 1.478= 22.028Lower limit = 20.55 - 2.571(.5748)
= 20.55 - 1.478= 19.072
We are 95% confident the actual sales for year 8will be between $19.072 and $22.028 million.
8/2/2019 Ch3. Demand Forecasting
50/99
50
Seasonalized Time Series Regression Analysis
Select a representative historical data set.Develop a seasonal index for each season.Use the seasonal indexes to deseasonalize the data.Perform linear regression analysis on thedeseasonalized data.Use the regression equation to compute the forecasts.Use the seasonal indexes to reapply the seasonalpatterns to the forecasts.
8/2/2019 Ch3. Demand Forecasting
51/99
51
Example: Computer Products Corp.
Seasonalized Times Series Regression AnalysisAn analyst at CPC wants to develop next years
quarterly forecasts of sales revenue for CPCs line of Epsilon Computers. She believes that the most recent8 quarters of sales (shown on the next slide) arerepresentative of next years sales.
8/2/2019 Ch3. Demand Forecasting
52/99
52
Example: Computer Products Corp.
Seasonalized Times Series Regression AnalysisRepresentative Historical Data Set
Year Qtr. ($mil.) Year Qtr. ($mil.)
1 1 7.4 2 1 8.31 2 6.5 2 2 7.41 3 4.9 2 3 5.4
1 4 16.1 2 4 18.0
8/2/2019 Ch3. Demand Forecasting
53/99
53
Example: Computer Products Corp.
Seasonalized Times Series Regression AnalysisCompute the Seasonal Indexes
Quarterly Sales
Year Q1 Q2 Q3 Q4 Total1 7.4 6.5 4.9 16.1 34.92 8.3 7.4 5.4 18.0 39.1
Totals 15.7 13.9 10.3 34.1 74.0Qtr. Avg. 7.85 6.95 5.15 17.05 9.25Seas.Ind. .849 .751 .557 1.843 4.000
8/2/2019 Ch3. Demand Forecasting
54/99
54
Example: Computer Products Corp.
Seasonalized Times Series Regression AnalysisDeseasonalize the Data
Quarterly Sales
Year Q1 Q2 Q3 Q41 8.72 8.66 8.80 8.742 9.78 9.85 9.69 9.77
8/2/2019 Ch3. Demand Forecasting
55/99
55
Example: Computer Products Corp.
Seasonalized Times Series Regression AnalysisPerform Regression on Deseasonalized Data
Yr. Qtr. x y x 2 xy
1 1 1 8.72 1 8.721 2 2 8.66 4 17.321 3 3 8.80 9 26.401 4 4 8.74 16 34.962 1 5 9.78 25 48.90
2 2 6 9.85 36 59.102 3 7 9.69 49 67.832 4 8 9.77 64 78.16
Totals 36 74.01 204 341.39
8/2/2019 Ch3. Demand Forecasting
56/99
56
Example: Computer Products Corp.
Seasonalized Times Series Regression AnalysisPerform Regression on Deseasonalized Data
Y = 8.357 + 0.199X
2204(74.01) 36(341.39)a 8.357
8(204) (36)
28(341.39) 36(74.01)b 0.199
8(204) (36)
8/2/2019 Ch3. Demand Forecasting
57/99
57
Example: Computer Products Corp.
Seasonalized Times Series Regression AnalysisCompute the Deseasonalized Forecasts
Y9 = 8.357 + 0.199(9) = 10.148
Y10 = 8.357 + 0.199(10) = 10.347Y11 = 8.357 + 0.199(11) = 10.546Y12 = 8.357 + 0.199(12) = 10.745
Note: Average sales are expected to increase by.199 million (about $200,000) per quarter.
l d
8/2/2019 Ch3. Demand Forecasting
58/99
58
Example: Computer Products Corp.
Seasonalized Times Series Regression AnalysisSeasonalize the Forecasts
Seas. Deseas. Seas.
Yr. Qtr. Index Forecast Forecast3 1 .849 10.148 8.623 2 .751 10.347 7.77
3 3 .557 10.546 5.873 4 1.843 10.745 19.80
Sh R F
8/2/2019 Ch3. Demand Forecasting
59/99
59
Short-Range Forecasts
Time spans ranging from a few days to a few weeksCycles, seasonality, and trend may have little effectRandom fluctuation is main data component
E l i F M d l P f
8/2/2019 Ch3. Demand Forecasting
60/99
60
Evaluating Forecast-Model Performance
Short-range forecasting models are evaluated on thebasis of three characteristics:Impulse responseNoise-dampening abilityAccuracy
E l i F M d l P f
8/2/2019 Ch3. Demand Forecasting
61/99
61
Evaluating Forecast-Model Performance
Impulse Response and Noise-Dampening AbilityIf forecasts have little period-to-period fluctuation,they are said to be noise dampening.Forecasts that respond quickly to changes in dataare said to have a high impulse response.A forecast system that responds quickly to datachanges necessarily picks up a great deal of
random fluctuation (noise).Hence, there is a trade-off between high impulseresponse and high noise dampening.
E l i F M d l P f
8/2/2019 Ch3. Demand Forecasting
62/99
62
Evaluating Forecast-Model Performance
AccuracyAccuracy is the typical criterion for judging theperformance of a forecasting approachAccuracy is how well the forecasted values matchthe actual values
M it i A
8/2/2019 Ch3. Demand Forecasting
63/99
63
Monitoring Accuracy
Accuracy of a forecasting approach needs to bemonitored to assess the confidence you can have in itsforecasts and changes in the market may requirereevaluation of the approach
Accuracy can be measured in several waysStandard error of the forecast (covered earlier)Mean absolute deviation (MAD)
Mean squared error (MSE)
M it i A
8/2/2019 Ch3. Demand Forecasting
64/99
64
Monitoring Accuracy
Mean Absolute Deviation (MAD)
n
periodsnfordeviationabsoluteof Sum=MAD
n
i ii=1
Actual demand -Forecast demandMAD =
n
M it i g A
8/2/2019 Ch3. Demand Forecasting
65/99
65
Mean Squared Error (MSE)MSE = (S yx)2
A small value for S yx means data points aretightly grouped around the line and error range issmall.
When the forecast errors are normally
distributed, the values of MAD and s yx are related:MSE = 1.25(MAD)
Monitoring Accuracy
Sh t R g F ti g M th d
8/2/2019 Ch3. Demand Forecasting
66/99
66
Short-Range Forecasting Methods
(Simple) Moving AverageWeighted Moving AverageExponential SmoothingExponential Smoothing with Trend
Simple Mo ing A erage
8/2/2019 Ch3. Demand Forecasting
67/99
67
Simple Moving Average
An averaging period (AP) is given or selectedThe forecast for the next period is the arithmeticaverage of the AP most recent actual demandsIt is called a simple average because each periodused to compute the average is equally weighted. . . more
Simple Moving Average
8/2/2019 Ch3. Demand Forecasting
68/99
68
Simple Moving Average
It is called moving because as new demand databecomes available, the oldest data is not usedBy increasing the AP, the forecast is less responsiveto fluctuations in demand (low impulse response and
high noise dampening)By decreasing the AP, the forecast is more responsiveto fluctuations in demand (high impulse response andlow noise dampening)
Weighted Moving Average
8/2/2019 Ch3. Demand Forecasting
69/99
69
Weighted Moving Average
This is a variation on the simple moving averagewhere the weights used to compute the average arenot equal.This allows more recent demand data to have a
greater effect on the moving average, therefore theforecast.. . . more
Weighted Moving Average
8/2/2019 Ch3. Demand Forecasting
70/99
70
Weighted Moving Average
The weights must add to 1.0 and generally decreasein value with the age of the data.The distribution of the weights determine the impulseresponse of the forecast.
Exponential Smoothing
8/2/2019 Ch3. Demand Forecasting
71/99
71
The weights used to compute the forecast (movingaverage) are exponentially distributed.The forecast is the sum of the old forecast and aportion ( a ) of the forecast error (A t-1 - Ft-1).
Ft = F t-1 + a (A t-1 - Ft-1)
. . . more
Exponential Smoothing
Exponential Smoothing
8/2/2019 Ch3. Demand Forecasting
72/99
72
Exponential Smoothing
The smoothing constant,a
, must be between 0.0 and1.0.A large a provides a high impulse response forecast.A small a provides a low impulse response forecast.
Example: Central Call Center
8/2/2019 Ch3. Demand Forecasting
73/99
73
Example: Central Call Center
Moving AverageCCC wishes to forecast the number of incomingcalls it receives in a day from the customers of one of its clients, BMI. CCC schedules the appropriate
number of telephone operators based on projected callvolumes.
CCC believes that the most recent 12 days of callvolumes (shown on the next slide) are representativeof the near future call volumes.
Example: Central Call Center
8/2/2019 Ch3. Demand Forecasting
74/99
74
Example: Central Call Center
Moving AverageRepresentative Historical Data
Day Calls Day Calls
1 159 7 2032 217 8 1953 186 9 1884 161 10 168
5 173 11 1986 157 12 159
Example: Central Call Center
8/2/2019 Ch3. Demand Forecasting
75/99
75
Example: Central Call Center
Moving AverageUse the moving average method with an AP = 3days to develop a forecast of the call volume in Day13.
F13 = (168 + 198 + 159)/3 = 175.0 calls
Example: Central Call Center
8/2/2019 Ch3. Demand Forecasting
76/99
76
Example: Central Call Center
Weighted Moving AverageUse the weighted moving average method with anAP = 3 days and weights of .1 (for oldest datum), .3,and .6 to develop a forecast of the call volume in Day
13.F13 = .1(168) + .3(198) + .6(159) = 171.6 calls
Note: The WMA forecast is lower than the MAforecast because Day 13s relatively low call volumecarries almost twice as much weight in the WMA(.60) as it does in the MA (.33).
Example: Central Call Center
8/2/2019 Ch3. Demand Forecasting
77/99
77
Example: Central Call Center
Exponential SmoothingIf a smoothing constant value of .25 is used andthe exponential smoothing forecast for Day 11 was180.76 calls, what is the exponential smoothing
forecast for Day 13?
F12 = 180.76 + .25(198 180.76) = 185.07F13 = 185.07 + .25(159 185.07) = 178.55
Example: Central Call Center
8/2/2019 Ch3. Demand Forecasting
78/99
78
Example: Central Call Center
Forecast Accuracy - MADWhich forecasting method (the AP = 3 movingaverage or the a = .25 exponential smoothing) ispreferred, based on the MAD over the most recent 9
days? (Assume that the exponential smoothingforecast for Day 3 is the same as the actual callvolume.)
Example: Central Call Center
8/2/2019 Ch3. Demand Forecasting
79/99
79
Example: Central Call Center
AP = 3 a = .25Day Calls Forec. |Error| Forec. |Error|
4 161 187.3 26.3 186.0 25.05 173 188.0 15.0 179.8 6.8
6 157 173.3 16.3 178.1 21.17 203 163.7 39.3 172.8 30.28 195 177.7 17.3 180.4 14.69 188 185.0 3.0 184.0 4.0
10 168 195.3 27.3 185.0 17.011 198 183.7 14.3 180.8 17.212 159 184.7 25.7 185.1 26.1
MAD 20.5 18.0
Exponential Smoothing with Trend
8/2/2019 Ch3. Demand Forecasting
80/99
80
Exponential Smoothing with Trend
As we move toward medium-range forecasts, trendbecomes more important.Incorporating a trend component into exponentiallysmoothed forecasts is called double exponential
smoothing.The estimate for the average and the estimate for thetrend are both smoothed.
Exponential Smoothing with Trend
8/2/2019 Ch3. Demand Forecasting
81/99
81
Exponential Smoothing with Trend
Model Form
FT t = S t-1 + T t-1where:
FT t = forecast with trend in period tSt-1 = smoothed forecast (average) in period t-1Tt-1 = smoothed trend estimate in period t-1
Exponential Smoothing with Trend
8/2/2019 Ch3. Demand Forecasting
82/99
82
Exponential Smoothing with Trend
Smoothing the Average
St = FT t + a (A t FT t)
Smoothing the Trend
Tt = T t-1 + b (FT t FT t-1 - T t-1)
where: a = smoothing constant for the average
b = smoothing constant for the trend
Criteria for Selecting
8/2/2019 Ch3. Demand Forecasting
83/99
83
a Forecasting Method
CostAccuracyData availableTime spanNature of products and servicesImpulse response and noise dampening
Criteria for Selecting
8/2/2019 Ch3. Demand Forecasting
84/99
84
a Forecasting Method
Cost and AccuracyThere is a trade-off between cost and accuracy;generally, more forecast accuracy can be obtainedat a cost.
High-accuracy approaches have disadvantages:Use more dataData are ordinarily more difficult to obtain
The models are more costly to design,implement, and operateTake longer to use
Criteria for Selecting
8/2/2019 Ch3. Demand Forecasting
85/99
85
a Forecasting Method
Cost and AccuracyLow/Moderate-Cost Approaches statisticalmodels, historical analogies, executive-committeeconsensus
High-Cost Approaches complex econometricmodels, Delphi, and market research
Criteria for Selectingh d
8/2/2019 Ch3. Demand Forecasting
86/99
86
a Forecasting Method
Data AvailableIs the necessary data available or can it beeconomically obtained?If the need is to forecast sales of a new product,then a customer survey may not be practical;instead, historical analogy or market research mayhave to be used.
Criteria for SelectingF i M h d
8/2/2019 Ch3. Demand Forecasting
87/99
87
a Forecasting Method
Time SpanWhat operations resource is being forecast and forwhat purpose?Short-term staffing needs might best be forecastwith moving average or exponential smoothingmodels.Long-term factory capacity needs might best be
predicted with regression or executive-committeeconsensus methods.
Criteria for SelectingF i M h d
8/2/2019 Ch3. Demand Forecasting
88/99
88
a Forecasting Method
Nature of Products and ServicesIs the product/service high cost or high volume?Where is the product/service in its life cycle?Does the product/service have seasonal demandfluctuations?
Criteria for SelectingF i M h d
8/2/2019 Ch3. Demand Forecasting
89/99
89
a Forecasting Method
Impulse Response and Noise DampeningAn appropriate balance must be achieved between:
How responsive we want the forecasting modelto be to changes in the actual demand dataOur desire to suppress undesirable chancevariation or noise in the demand data
Reasons for Ineffective Forecasting
8/2/2019 Ch3. Demand Forecasting
90/99
90
g
Not involving a broad cross section of peopleNot recognizing that forecasting is integral tobusiness planningNot recognizing that forecasts will always be wrongNot forecasting the right thingsNot selecting an appropriate forecasting methodNot tracking the accuracy of the forecasting models
Monitoring and ControllingF ti M d l
8/2/2019 Ch3. Demand Forecasting
91/99
91
a Forecasting Model
Tracking Signal (TS)The TS measures the cumulative forecast errorover n periods in terms of MAD
If the forecasting model is performing well, the TSshould be around zeroThe TS indicates the direction of the forecastingerror; if the TS is positive -- increase the forecasts,if the TS is negative -- decrease the forecasts.
n
i i1
(Actual demand - Forecast demand )TS =
MADi
Monitoring and ControllingF ti M d l
8/2/2019 Ch3. Demand Forecasting
92/99
92
a Forecasting Model
Tracking SignalThe value of the TS can be used to automaticallytrigger new parameter values of a model, therebycorrecting model performance.
If the limits are set too narrow, the parametervalues will be changed too often.If the limits are set too wide, the parameter values
will not be changed often enough and accuracywill suffer.
Tracking Signal: What do you notice?
8/2/2019 Ch3. Demand Forecasting
93/99
93
Tracking Signal: What do you notice?
20
25
30
35
40
0 1 2 3 4 5 6 7 8 9 10 11Period
S a
l e s
Computer Software for Forecasting
8/2/2019 Ch3. Demand Forecasting
94/99
94
p g
Examples of computer software with forecastingcapabilities
Forecast ProAutoboxSmartForecasts for WindowsSASSPSS
SAPPOM Software Library
Primarily for
forecasting
Have
Forecastingmodules
Forecasting in Small Businessesand Start Up Vent res
8/2/2019 Ch3. Demand Forecasting
95/99
95
and Start-Up Ventures
Forecasting for these businesses can be difficult forthe following reasons:
Not enough personnel with the time to forecastPersonnel lack the necessary skills to develop goodforecastsSuch businesses are not data-rich environmentsForecasting for new products/services is always
difficult, even for the experienced forecaster
Sources of Forecasting Data and Help
8/2/2019 Ch3. Demand Forecasting
96/99
96
Government agencies at the local, regional, state, andfederal levelsIndustry associationsConsulting companies
8/2/2019 Ch3. Demand Forecasting
97/99
Wrap-Up: World-Class Practice
8/2/2019 Ch3. Demand Forecasting
98/99
98
Predisposed to have effective methods of forecastingbecause they have exceptional long-range businessplanningFormal forecasting effort
Develop methods to monitor the performance of theirforecasting modelsDo not overlook the short run.... excellent short rangeforecasts as well
End of Chapter 3
8/2/2019 Ch3. Demand Forecasting
99/99