Chapter 7_ Demand Forecasting IPE .ppt

IPE

1. Chapter 7Chapter 7
Demand ForecastingDemand Forecasting
Text Book:
William J. Stevenson, Operations
Management, 8th ed., McGraw-Hill. Garrison

2. What is Forecasting?What is Forecasting?
Process of predicting a
future event
Underlying basis of
all business decisions
Production
Material Requirement
Personnel
Facilities
Budgeting
Sales will
be $200
Million!

3. Steps in Forecasting ProcessSteps in Forecasting Process
Determine the purpose
Establish a time horizon
Select a forecasting technique
Gather and analyze the appropriate data
Prepare the forecast
Monitor the forecast

4. Short-range forecast
Usually less than 3 months
Job scheduling, worker assignments
Medium-range forecast
3 months to 3 years
Sales & production planning, budgeting
Long-range forecast
3
+
years
New product planning, facility location
Types of Forecasts by Time Types of Forecasts by Time
HorizonHorizon

5. Forecasting ApproachesForecasting Approaches
Used when situation is
‘stable’ & historical data
exist
Existing products
Current technology
Involves mathematical
techniques
e.g., Forecasting demand of
Rod of a steel plant
Quantitative Methods
Used when situation is
vague & little data exist
New products
New technology
Involves intuition,
experience
e.g., First arrival of android
phone.
Qualitative Methods

6. Qualitative MethodsQualitative Methods
Sales force composite
Jury of executive opinion
Delphi method
Consumer Market Survey

7. Sales Force Composite (or grass Sales Force Composite (or grass
roots forecasting)roots forecasting)
Person closest to customer ends
knows best
Combines each levels
Sales representatives know
customers’ wants
Retailer
Distributor
Manufacturer
SaleSale
ss
© 1995
Corel Corp.

8. Involves people from various positions in the
organization
Group estimates demand by working together
Relatively quick
Free exchange of ideas
Lower employee levels
may get intimidated
by higher management.
© 1995 Corel Corp.
Jury of Executive Opinion (or Jury of Executive Opinion (or
panel consensus)panel consensus)

9. Delphi MethodDelphi Method
Iterative group process. Conceals
identity of participants
Moderator makes a questionnaire.
Response are summed and new set of
questions are provided

10. Step by step procedureStep by step procedure
1.Choose the experts to participate.
2.Through a question set obtain forecasts from
the participants.
3.Summarize the results and redistribute them to
the participants along with new set of question.
4.Summarize again.
5.Repeat step 3 & 4 until consensus is reached

11. Consumer Market Survey (or Consumer Market Survey (or
Market Research)Market Research)
Ask customers about
purchasing plans
What consumers
say, and what they
actually do are often
different
Sometimes difficult to
answer
How many hours
will you use the
Internet next week?
© 1995 Corel
Corp.

12. Quantitative ApproachesQuantitative Approaches
Naïve approach
Moving averages
Exponential smoothing
Trend projection
Linear regression

13. Naive ApproachNaive Approach
Assumes demand in next period
is similar as demand in most
recent period without adjusting
them considering any factor.
e.g., If May sales were 48, then
June sales will be …48
Sometimes cost effective &
efficient
© 1995 Corel Corp.

14. MA is a series of arithmetic means
Used if little or no trend
Used often for smoothing
Provides overall impression of data over time
Equation
MAMA
nn
nn

Demand inDemand in PreviousPrevious PeriodsPeriods
Moving Average MethodMoving Average Method

15. You’re manager of a museum store that sells
historical replicas. You want to forecast sales
2000 for 2003 using a 3-period moving average.
1998 4
1999 6
2000 5
2001 3
2002 7
Moving Average ExampleMoving Average Example

16. Moving Average SolutionMoving Average Solution
Time Response
Yi
Moving
Total
(n=3)
Moving
Average
(n=3)
1998 4 NA NA
1999 6 NA NA
2000 5 NA NA
2001 3 4+6+5=15 15/3 = 5
2002 7
2003 NA

17. Moving Average SolutionMoving Average Solution
Time Response
Yi
Moving
Total
(n=3)
Moving
Average
(n=3)
1998 4 NA NA
1999 6 NA NA
2000 5 NA NA
2001 3 4+6+5=15 15/3=5.0
2002 7 6+5+3=14 14/3=4.7
2003 NA 5+3+7=15 15/3=5.0

18. Moving Average SolutionMoving Average Solution
A moving average forecast tends to smooth and lag
changes in the data

19. Moving Average SolutionMoving Average Solution
The more periods in a moving average, the greater the
forecast will lag changes in the data

20. Used when trend is present
Older data usually less important
Weights based on intuition
Often lay between 0 & 1, & sum to 1.0
Equation
Weighted Moving Average MethodWeighted Moving Average Method

21. Weighted Moving Average MethodWeighted Moving Average Method
What is the sales forecast for Month 5?
Ans: =0.4*95+ .3*105+ .2*90+ .1*100
= 97.5 units

22. Form of weighted moving average
Weights decline exponentially
Most recent data weighted most
Requires smoothing constant ()
Ranges from 0 to 1
Involves little record keeping of past data
Exponential Smoothing MethodExponential Smoothing Method

23. F
t = F
t-1 + (A
t-1 - F
t-1)
F
t = Forecast value
A
t = Actual value
 = Smoothing constant
A
t-1
- F
t-1=
Forecast error.
Exponential Smoothing EquationsExponential Smoothing Equations

24. During the past 8 quarters, the Port of Mongla has unloaded large quantities of grain.
( = .10). The first quarter forecast was 175..
QuarterActual
1 180
2 168
3 159
4 175
5 190
6 205
7 180
8 182
9 ?
Exponential Smoothing ExampleExponential Smoothing Example
Find the forecast
for the 9
th
quarter.

25. F
t = F
t-1 + (A
t-1 - F
t-1)

QuarterQuarterActualActual
Forecast, Ft
(αα= = .10.10))
11 180 175.00 (Given)
22 168168
33 159159
44 175175
55 190190
66 205205
175.00 +175.00 +
Exponential Smoothing SolutionExponential Smoothing Solution

26. QuarterQuarterActuaActual
Forecast, Ft
(αα= = .10.10))
11 180180 175.00 (Given)175.00 (Given)
22 168168175.00 + 175.00 + .10.10((
33 159159
44 175175
55 190190
66 205205
Exponential Smoothing SolutionExponential Smoothing Solution
F
t = F
t-1 + (A
t-1 - F
t-1)

27. QuarterQuarterActualActual
Forecast, Forecast, FFtt
((αα= = .10.10))
11 180180 175.00 (Given)175.00 (Given)
22 168168175.00 + 175.00 + .10.10(180(180 - -
33 159159
44 175175
55 190190
66 205205
Exponential Smoothing SolutionExponential Smoothing Solution
F
t = F
t-1 + (A
t-1 - F
t-1)

28. QuarterQuarterActualActual
Forecast, Ft
(αα= = .10.10))
11 180180 175.00 (Given)175.00 (Given)
22 168168175.00 + 175.00 + .10.10(180(180 - 175.00 - 175.00))
33 159159
44 175175
55 190190
66 205205
Exponential Smoothing SolutionExponential Smoothing Solution
F
t = F
t-1 + (A
t-1 - F
t-1)

29. QuarterQuarterActualActual
Forecast, Forecast, FFtt
((αα= = .10.10))
11 180180 175.00 (Given)175.00 (Given)
22 168168175.00 +175.00 + .10.10(180 (180 - 175.00- 175.00)) = 175.50 = 175.50
33 159159
44 175175
55 190190
66 205205
Exponential Smoothing SolutionExponential Smoothing Solution
F
t = F
t-1 + (A
t-1 - F
t-1)

30. F
t = F
t-1 + (A
t-1 - F
t-1)

QuarterQuarterActualActual
Forecast, Ft
(αα= = .10.10))
1 180 175.00 (Given)
22 168168175.00 + .10(180 - 175.00) = 175.50175.00 + .10(180 - 175.00) = 175.50
33 159159175.50175.50 ++ .10.10(168 -(168 - 175.50175.50)) = 174.75= 174.75
44 175175
55 190190
66 205205
Exponential Smoothing SolutionExponential Smoothing Solution

31. F
t = F
t-1 + (A
t-1 - F
t-1)
TimeActual
Forecast, Ft
(α= .10)
4 175174.75 + .10(159 - 174.75) = 173.18
5 190173.18 + .10(175 - 173.18) = 173.36
6 205173.36 + .10(190 - 173.36) = 175.02
Exponential Smoothing SolutionExponential Smoothing Solution
7 180
8
175.02 + .10(205 - 175.02) = 178.02
9 178.22 + .10(182 - 178.22) = 178.58
182178.02 + .10(180 - 178.02) = 178.22
?

32. Effect of Effect of αα in forecasting in forecasting

33. Linear Regression Linear Regression
Least Squares MethodLeast Squares Method
Deviation
Deviation
Deviation
Deviation
Deviation
Deviation
Deviation
Time
V
a
lu
e
s
o
f
D
e
p
e
n
d
e
n
t
V
a
r
ia
b
le
bxaYˆ
Actual
observation
Point on
regression
line

34. Used for forecasting linear trend line
Assumes relationship between response
variable, Y, and time, X, is a linear function
Estimated by least squares method
Minimizes sum of squared errors
iYabX
i

Linear Trend ProjectionLinear Trend Projection

35. Least Squares EquationsLeast Squares Equations
Equation:
ii bxaYˆ
Slope:
22
1
1
)(xnx
yxnyx
b
i
n
i
ii
n
i





Y-Intercept:
xbya

36. Using a Trend LineUsing a Trend Line
Year Demand (MW)
1997 74
1998 79
1999 80
2000 90
2001 105
2002 142
2003 122
The demand for
electrical power at N.
Y. Edison over the
years 1997-2003 is
given at the left. Find
the overall trend. Also
find the demand at
2004 and 2005.

37. Finding a Trend LineFinding a Trend Line
YearTime
Period
Power
Demand
x
2
xy
1997 1 74 1 74
1998 2 79 4 158
1999 3 80 9 240
2000 4 90 16 360
2001 5 105 25 525
2002 6 142 36 852
2003 7 122 49 854
x=28y=692x
2
=140xy=3,063

38. The Trend Line EquationThe Trend Line Equation
megawatts 151.56 10.54(9) 56.70 2005in Demand
megawatts 141.02 10.54(8) 56.70 2004in Demand
10.54X 56.70 Y Trend, Overall
56.70 10.54(4) - 98.86 xb - y a
10.54
28
295
(7)(4)140
86)(7)(4)(98.3,063
xnΣx
y xn -Σxy
b
98.86
7
692

n
Σy
y 4
7
28
n
Σx
x
222












39. Practice ProblemPractice Problem
Ans: Y= 441.6+ 359.6X

40. Non-Linear TrendNon-Linear Trend

41. Polynomial ModellingPolynomial Modelling
The general form of a quadratic polynomial regression
Y = a
0 + a
1t + a
2t
2
Where,
Y = Demand/ Forecasted Value
t = time period
a
0, a
1, a
2 = Model parameters need to determine

42. Polynomial ModellingPolynomial Modelling
YearDemand (MW)
1997 74
1998 79
1999 80
2000 90
2001 105
2002 142
2003 122
The demand for electrical
power at N. Y. Edison over
the years 1997-2003 is
given at the left. Find the
overall trend. Also find the
demand at 2004 and 2005.

43. Polynomial ModellingPolynomial Modelling
Calculation of the required values:

44. Polynomial ModellingPolynomial Modelling

45. Polynomial ModellingPolynomial Modelling
System of normal equations:
692 = 7a
0 + 28a
1 + 140a
2
3063 = 28a
0
+ 140a
1
+ 784a
2
16265 = 140a
0 + 784a
1 + 4676a
2
Solving the equations yields:
a
0 = 66
a
1
= 4.345
a
2 = 0.7738
The polynomial equation will be:
Y = 66 + 4.345t + 0.7738t
2