1

Company selection
• 5 groups sent the company.
• Deadline: within week 2.

2
FORECASTING
 Learning Objectives
• Explain the role of demand planning in operations
management, in the firm and the supply chain
• Understand the necessity of forecasting
• Describe various qualitative and quantitative demand
forecast procedures
• Differentiate types of forecasting techniques
• Develop forecasts using various techniques
• Evaluate and select forecasting models using various
measures of accuracy
9
Uncertainty in supply and
demand

10
 Uncertainty in supply and
demand
• Demand may be

predictabl unpredictabl
e e

11
 Uncertainty in supply and
demand
• Demand may be

predictabl unpredictabl
e e
⇒ Both supply and demand uncertainty makes
planning and control more difficult.
12
Independent vs. Dependent Demand

13
Independent vs. Dependent Demand

14
 Independent vs. Dependent Demand
Independent
Demand
Finished Goods

15
 Independent vs. Dependent Demand
Independent
Demand
Finished Goods

Dependent
Demand
Raw Materials,
Component parts,
Sub-assemblies,
16
Independent vs. Dependent
Demand

17
Independent vs. Dependent
Demand

Raw Finished
materials goods 18
Forecasting

19
 Forecasting
• The process of projecting the values of one or
more variables into the future.

High Low
forecast forecast
20
Why do we forecast?

21
 Why do we forecast?

To minimise uncertainty

22
 Why do we forecast?

To minimise uncertainty To anticipate

change

23
 Why do we forecast?

To minimise uncertainty To anticipate To make better decisions

change

24
Types of Forecasts

27
Types of Forecasts
Qualitative Quantitative

28
 Types of Forecasts
Qualitative Quantitative
• Built on the
estimates and
opinions of experts.
• Panel method.

29
 Types of Forecasts
Qualitative Quantitative
• Built on the • Based only on a series
estimates and of past demands.
opinions of experts. • Assume that a demand
• Panel method. pattern of the past will
continue in the future.
30
 Forecasting Keys

Accurate
historical data

31
 Forecasting Keys

Accurate Understand
historical data Causes of change

32
 Forecasting Keys

Accurate Understand Examine

historical data causes of change forecast error

33
Quantitative Method:
Time Series Analysis

34
 Quantitative Method:
Time Series Analysis
• Many types of models available.
• Pick a model based on:
1. Fits previous data best.
2. Data availability.
3. Accuracy required.

35
Simple Moving Average

36
Simple Moving Average

37
 Simple Moving Average

• Dt = actual demand from period t

• Ft+1 = forecast of demand for period t+1 (next period that
has not occurred yet)
• Forecast for the next period t+1 = average from the last n
periods of actual demand.
38
Simple Moving Average Exercise 1

39
 Simple Moving Average Exercise 1
WEE DEMAND
K
• Let’s develop 3-week and 4- 1 821
2 770
week moving average forecasts 3 690
for demand for this item. 4 650
5 620
6 550
7
40
 Simple Moving Average Exercise 2
WEEK DEMAND
1 743
2 710
• Let’s develop 3-week 3 678
4 665
moving average forecasts 5 665
6 844

for demand for this item. 7 789

8 920
9 892
10
41
Weighted Moving Average

42
Weighted Moving Average

43
Weighted Moving Average

44
Weighted Moving Average

• Dt = actual demand from period t

• Ft+1 = forecast of demand for period t+1 (next period
that has not occurred yet)
• wt = the weight given to the demand value in period t
45
Weighted Moving Average
Exercise 1

46
 Weighted Moving Average
Exercise 1
• Determine the 3-period WEE DEMAND
K
weighted moving average 1 821
forecast with weight as 2 770
follow: 3 690
4 650
• t = 0.5
5 620
• t – 1 = 0.3 6 550
• t – 2 = 0.2 7
47
 Weighted Moving Average
Exercise 2
WEEK DEMAND
• Determine the 3-period 1 743
weighted moving average 2 710
3 678
forecast with weight as follow: 4 665
5 665
a. t = 0.5; t – 1 = 0.3; t – 2 = 0.2 6 844
7 789
b. t = 0.7; t – 1 = 0.2; t – 2 = 0.1 8 920
9 892
10
48
 CHECK
ATTENDANCE
49
BREAK-TIME
50
 Exponential Smoothing
(α is the smoothing parameter)

51
 Exponential Smoothing
(α is the smoothing parameter)

Ft+1 = α Dt + (1-
α)Ft

52
 Exponential Smoothing
(α is the smoothing parameter)

Ft+1 = α Dt + (1-
• α)Ft
Premise — we should determine how much weight to put on
recent information versus older information.
• 0<α<1
• High α (0.7) puts weight on recent demand.
• Low α (0.2) puts weight on previous periods.
53
Exponential Smoothing Exercise 1

54
 Exponential Smoothing Exercise 1
WEE DEMAND
K
• Determine exponential 1 821
smoothing forecasts by using α 2 770
= 0.7 3 690
4 650
• Let D1= F2 5 620
6 550
7
55
 Exponential Smoothing Exercise 2
WEEK DEMAND
• Determine exponential
1 743
smoothing forecasts by using 2 710
3 678
a. α = 0.6 4 665
5 665
b. α = 0.7 6 844
7 789
• Let D1= F2 8 920
9 892
10
56
Forecast Error

57
 Forecast Error
Measures of Forecast Et = Dt -
Error: Ft

58
 Forecast Error
Measures of Forecast Et = Dt -
Error: Ft
The error (E) which defines deviation between the
forecast and the actual demand.

59
 Forecast Error
Mean Forecast Error
(MFE)

60
 Forecast Error
Mean Absolute
Mean Forecast Error Deviation
(MFE) (MAD)

61
Forecast Errors Exercise

62
Forecast Errors Exercise 1
Determine the Forecast Error for the below periods

Forecas
Month
Sales t
1 220
2 225 255
3 215 225
4 350 320
5 300 315
63
Forecast Errors Exercise 2
Determine the Forecast Error for the below periods
Forecas
Month
Sales t
1 821
2 770 800
3 690 710
4 650 760
5 620 703
6 550 655
7 600 505 64
 Homework
• Historical demand for KMK Inc., is presented on the
Excel file.
a. Complete the Forecast using three methods.
b. Conduct the forecast error by using MFE and MAD.
c. Which method is the most appropriate to use?

65
THANK YOU!