Outline

▶ What is Forecasting?

CHAPTER 2
Forecasting
2 ▶ The Strategic Importance of Forecasting

▶ Seven Steps in the Forecasting System

▶ Forecasting Approaches

▶ Time-Series Forecasting

POM for Engineer Course: Chapter 2: Forecasting Techniques 4-1 POM for Engineer Course: Chapter 2: Forecasting Techniques 4-2

Learning Objectives
When you complete this chapter you
What is Forecasting?
should be able to : + Process of predicting a
+ Understand the three time horizons and which models future event
apply for each + Underlying basis
of all business decisions
??
+ Explain when to use each of the four qualitative models
+ Production
+ Apply the naive, moving-average, exponential smoothing,
+ Inventory
and trend methods
+ Personnel
+ Compute three measures of forecast accuracy + Facilities

POM for Engineer Course: Chapter 2: Forecasting Techniques 4-3 POM for Engineer Course: Chapter 2: Forecasting Techniques 4-4
 Forecasting Time Horizons Distinguishing Differences
1. Short-range forecast
► Up to 1 year, generally less than 3 months 1. Medium/long range forecasts deal with more
► Purchasing, job scheduling, workforce levels, job comprehensive issues and support management
assignments, production levels decisions regarding planning and products, plants
and processes
2. Medium-range forecast
► 3 months to 3 years 2. Short-term forecasting usually employs different
methodologies than longer-term forecasting
► Sales and production planning, budgeting
3. Short-term forecasts tend to be more accurate
3. Long-range forecast than longer-term forecasts
► 3+ years
► New product planning, facility location, capital
expenditures, research and development
POM for Engineer Course: Chapter 2: Forecasting Techniques 4-5 POM for Engineer Course: Chapter 2: Forecasting Techniques 4-6

Types of Forecasts Strategic Importance of Forecasting

1. Economic forecasts ► Supply Chain Management – Good supplier
► Address business cycle – inflation rate, money relations, advantages in product innovation, cost
supply, housing starts, etc. and speed to market
2. Technological forecasts ► Human Resources – Hiring, training, laying off
► Predict rate of technological progress
workers
► Impacts development of new products ► Capacity – Capacity shortages can result in
undependable delivery, loss of customers, loss of
3. Demand forecasts market share
► Predict sales of existing products and services

POM for Engineer Course: Chapter 2: Forecasting Techniques 4-7 POM for Engineer Course: Chapter 2: Forecasting Techniques 4-8
 Seven Steps in Forecasting
The Realities!
1. Determine the use of the forecast
► Forecasts are seldom perfect,
2. Select the items to be forecasted unpredictable outside factors may impact
the forecast
3. Determine the time horizon of the forecast
► Most techniques assume an underlying
4. Select the forecasting model(s) stability in the system
5. Gather the data needed to make the forecast ► Product family and aggregated forecasts
are more accurate than individual product
6. Make the forecast forecasts
7. Validate and implement the results
POM for Engineer Course: Chapter 2: Forecasting Techniques 4-9 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 10

Forecasting Approaches Forecasting Approaches

Qualitative Methods Quantitative Methods
► Used when situation is vague and little ► Used when situation is ‘stable’ and
data exist historical data exist
► New products ► Existing products
► New technology ► Current technology
► Involves intuition, experience ► Involves mathematical techniques
► e.g., forecasting sales on Internet ► e.g., forecasting sales of color televisions

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 11 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 12
 Overview of Qualitative Methods Overview of Qualitative Methods
3. Sales force composite
1. Jury of executive opinion ► Estimates from individual salespersons are
Pool opinions of high-level experts, sometimes reviewed for reasonableness, then
augmented by statistical models aggregated

2. Delphi method 4. Market Survey

Panel of experts, queried iteratively ► Ask the customer

 Replaced by using POS data (point of sales)

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 13 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 14

Jury of Executive Opinion Delphi Method

► Iterative group
► Involves small group of high-level experts and process, continues Decision Makers
managers (Evaluate responses
until consensus is and make decisions)
► Group estimates demand by working together reached
► Combines managerial experience with statistical
models
► Three types of Staff
(Administering
participants survey)
► Relatively quick
► Decision makers
► ‘Group-think’
► Staff
disadvantage
► Respondents Respondents
(People who can make
valuable judgments)
POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 15 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 16
 Sales Force Composite Market Survey
► Ask customers about purchasing plans
► Each salesperson projects his or her sales
► Useful for demand and product design and
► Combined at district and national levels planning
► Sales reps know customers’ wants ► What consumers say and what they
► May be overly optimistic actually do may be different
► May be overly optimistic

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 17 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 18

Overview of Quantitative Time-Series Forecasting

Approaches
► Set of evenly spaced numerical data
1. Naive approach
► Obtained by observing response variable at
2. Moving averages regular time periods
3. Exponential Time-series ► Forecast based only on past values, no
smoothing models other variables important
4. Trend projection ► Assumes that factors influencing past and
5. Linear regression Associative present will continue influence in future
model

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 19 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 20
 Time-Series Components Components of Demand
Trend
component

Trend Cyclical

Demand for product or service

Seasonal peaks

Actual demand
line

Average demand
over 4 years

Seasonal Random Random variation

| | | |
1 2 3 4
Time (years)
Figure 4.1

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 21 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 22

Trend Component Seasonal Component

► Regular pattern of up and down fluctuations
► Persistent, overall upward or downward
► Due to weather, customs, etc.
pattern
► Occurs within a single year
► Changes due to population, technology,
age, culture, etc. PERIOD LENGTH “SEASON”
” LENGTH NUMBER OF “SEASONS”
” IN PATTERN
Week Day 7
► Typically several years duration Month Week 4 – 4.5
Month Day 28 – 31
Year Quarter 4
Year Month 12
Year Week 52

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 23 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 24
 Cyclical Component Random Component
► Repeating up and down movements ► Erratic, unsystematic, ‘residual’ fluctuations
► Affected by business cycle, political, and ► Due to random variation or unforeseen
economic factors events
► Multiple years duration ► Short duration
and nonrepeating
► Often causal or
associative
relationships

0 5 10 15 20 M T W T
POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 25 POM for Engineer Course: Chapter 2: Forecasting Techniques F 4 - 26

Naive Approach Moving Averages

► Assumes demand in next ► MA is a series of arithmetic means

period is the same as
demand in most recent period ► Used if little or no trend
► e.g., If January sales were 68, then ► Used often for smoothing
February sales will be 68 ► Provides overall impression of data over time
► Sometimes cost effective and efficient
► Can be good starting point ådemand in previous n periods
Moving average =
n

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 27 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 28
 Moving Average Example
Weighted Moving Average
n=3
MONTH ACTUAL SHED SALES 3-MONTH MOVING AVERAGE ► Used when some trend might be present
January 10
February 12 ► Older data usually less important
March 13
April 16 (10 + 12 + 13)/3 = 11 2/3 ► Weights based on experience and intuition
(12 + 13 + 16)/3 = 13 2/
May 19 3

June 23 (13 + 16 + 19)/3 = 16

July 26 (16 + 19 + 23)/3 = 19 1/3
August
September
30
28
(19 + 23 + 26)/3 = 22 2/3
(23 + 26 + 30)/3 = 26 1/3 moving =
((
Weighted å Weight for period n Demand in period n )( ))
October 18 (26 + 30 + 28)/3 = 28 average å Weights
November 16 (30 + 28 + 18)/3 = 25 1/3
December 14 (28 + 18 + 16)/3 = 20 2/3

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 29 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 30

Weighted Moving Average

Weighted Moving Average
n = 3, weights are 3, 2, 1
MONTH ACTUAL SHED SALES 3-MONTH WEIGHTED MOVING AVERAGE MONTH ACTUAL SHED SALES 3-MONTH WEIGHTED MOVING AVERAGE
January 10 January 10
February 12 February 12
March 13 March 13
April 16 [(3 x 13) + (2 x 12) + (10)]/6 = 12 1/6 April 16 [(3 x 13) + (2 x 12) + (10)]/6 = 12 1/6
May 19 May 19 [(3 x 16) + (2 x 13) + (12)]/6 = 14 1/3
June WEIGHTS
23 APPLIED PERIOD June 23 [(3 x 19) + (2 x 16) + (13)]/6 = 17
July 26 3 Last month July 26 [(3 x 23) + (2 x 19) + (16)]/6 = 20 1/2
August 30 2 Two months ago August 30 [(3 x 26) + (2 x 23) + (19)]/6 = 23 5/6
September 28 1 Three months ago September 28 [(3 x 30) + (2 x 26) + (23)]/6 = 27 1/2
October 18 6 Sum of the weights October 18 [(3 x 28) + (2 x 30) + (26)]/6 = 28 1/3
November Forecast for
16this month = November 16 [(3 x 18) + (2 x 28) + (30)]/6 = 23 1/3
December 3 x14
Sales last mo. + 2 x Sales 2 mos. ago + 1 x Sales 3 mos. ago December 14 [(3 x 16) + (2 x 18) + (28)]/6 = 18 2/3
Sum of the weights

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 31 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 32
 Potential Problems With Graph of Moving Averages
Moving Average Weighted moving average (from Example 2)

30 –

1. Increasing n smooths the forecast but 25 –

makes it less sensitive to changes

Sales demand
20 –
2. Does not forecast trends well
15 – Actual sales
3. Requires extensive historical data
10 – Moving average
(from Example 1)
5–
| | | | | | | | | | | |

J F M A M J J A S O N D
Figure 4.2 Month

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 33 Copyright © 2020 Pearson Education Ltd. All Rights Reserved. 4 - 34

Exponential Smoothing Exponential Smoothing

► Form of weighted moving average New forecast = Last period’s forecast
+ α (Last period’s actual demand
► Weights decline exponentially
– Last period’s forecast)
► Most recent data weighted most
► Requires smoothing constant (α) Ft = Ft – 1 + α(At – 1 – Ft – 1)
► Ranges from 0 to 1
where Ft = new forecast
► Subjectively chosen
Ft – 1 = previous period’s forecast
► Involves little record keeping of past data α = smoothing (or weighting) constant (0 ≤ α ≤ 1)
At – 1 = previous period’s actual demand

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 35 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 36
 Exponential Smoothing Exponential Smoothing
Example Example
Predicted demand = 142 Ford Mustangs Predicted demand = 142 Ford Mustangs
Actual demand = 153 Actual demand = 153
Smoothing constant α = .20 Smoothing constant α = .20

New forecast = 142 + .2(153 – 142)

Copyright © 2020 Pearson Education Ltd. All Rights Reserved. 4 - 37 Copyright © 2020 Pearson Education Ltd. All Rights Reserved. 4 - 38

Exponential Smoothing Exponential smoothing

Example ACTUAL
TONNAGE FORECAST WITH
QUARTER UNLOADED FORECAST WITH α = .10 α = .50
1 180 175 175
Predicted demand = 142 Ford Mustangs 2 168 175.50 = 175.00 + .10(180 – 175) 177.50
Actual demand = 153 3 159 174.75 = 175.50 + .10(168 – 175.50) 172.75
Smoothing constant α = .20 4 175 173.18 = 174.75 + .10(159 – 174.75) 165.88

5 190 173.36 = 173.18 + .10(175 – 173.18) 170.44

6 205 175.02 = 173.36 + .10(190 – 173.36) 180.22

New forecast = 142 + .2(153 – 142) 7 180 178.02 = 175.02 + .10(205 – 175.02) 192.61

= 142 + 2.2 8 182 178.22 = 178.02 + .10(180 – 178.02) 186.30

9 ? 178.59 = 178.22 + .10(182 – 178.22) 184.15

= 144.2 ≈ 144 cars

Copyright © 2020 Pearson Education Ltd. All Rights Reserved. 4 - 39 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 40
 Effect of
Smoothing Constants Impact of Different α

▶ Smoothing constant generally .05 ≤ α ≤ .50 225 –

▶ As α increases, older values become less Actual α = .5

significant 200 –
demand

Demand
WEIGHT ASSIGNED TO
MOST 2ND MOST 3RD MOST 4th MOST 5th MOST 175 –
RECENT RECENT RECENT RECENT RECENT
SMOOTHING PERIOD PERIOD PERIOD PERIOD PERIOD
CONSTANT (α ) α (1 – α ) α (1 – α )2 α (1 – α )3 α (1 – α )4 α = .1
α = .1 .1 .09 .081 .073 .066 150 – | | | | | | | | |
α = .5 .5 .25 .125 .063 .031 1 2 3 4 5 6 7 8 9
Quarter
POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 41 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 42

Impact of Different α Selecting the Smoothing Constant

225 –
The objective is to obtain the most accurate
forecast no matter the technique
Actual α = .5
► 200 –Choose high values
demand of α
We generally do this by selecting the model
when underlying average
Demand

that gives us the lowest forecast error

is likely to change according to one of three preferred measures:
175 –
► Choose low values of α
when underlying average ► Mean Absolute Deviation (MAD)
α = .1
is–stable ► Mean Squared Error (MSE)
150 | | | | | | | | |
► Mean Absolute Percent Error (MAPE)
1 2 3 4 5 6 7 8 9
Quarter
POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 43 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 44
 Common Measures of Error Determining the MAD
ACTUAL FORECAST ABSOLUTE FORECAST ABSOLUTE
TONNAGE WITH DEVIATION WITH DEVIATION
QUARTER UNLOADED α = .10 FOR α = .10 α = .50 FOR α = .50
Mean Absolute Deviation (MAD) 1 180 175 5.00 175 5.00
2 168 175.50 7.50 177.50 9.50
3 159 174.75 15.75 172.75 13.75

MAD =
å Actual - Forecast 4 175 173.18 1.82 165.88 9.12
5 190 173.36 16.64 170.44 19.56
n
6 205 175.02 29.98 180.22 24.78
7 180 178.02 1.98 192.61 12.61
8 182 178.22 3.78 186.30 4.30
Sum of absolute deviations: 82.45 98.62
Σ|Deviations|
MAD = 10.31 12.33
n

FORECAST WITH α = .10 is better than α = .50

because MAD (α = .10) is less than MAD (α = .50)
POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 45 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 46

Common Measures of Error Determining the MSE

ACTUAL
TONNAGE FORECAST FOR
QUARTER UNLOADED α = .10 (ERROR)2

Mean Squared Error (MSE) 1

2
180
168
175
175.50
52 = 25
(–7.5)2 = 56.25
3 159 174.75 (–15.75)2 = 248.06
2

MSE =
å(Forecast errors) 4
5
175
190
173.18
173.36
(1.82)2 = 3.31
(16.64)2 = 276.89
n 6 205 175.02 (29.98)2 = 898.80
7 180 178.02 (1.98)2 = 3.92
8 182 178.22 (3.78)2 = 14.29
Sum of errors squared = 1,526.52

MSE =
å(Forecast errors) = 1,526.52 / 8 = 190.8
n
POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 47 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 48
 Common Measures of Error Determining the MAPE
ACTUAL
TONNAGE FORECAST FOR ABSOLUTE PERCENT ERROR
QUARTER UNLOADED α = .10 100(|ERROR|/ACTUAL)
1 180 175.00 100(5/180) = 2.78%

Mean Absolute Percent Error (MAPE) 2

3
168
159
175.50
174.75
100(7.5/168) = 4.46%
100(15.75/159) = 9.90%

n 4 175 173.18 100(1.82/175) = 1.05%

å100 Actual −Forecast i i

/ Actuali 5
6
190
205
173.36
175.02
100(16.64/190) = 8.76%
100(29.98/205) = 14.62%
MAPE = i=1
7 180 178.02 100(1.98/180) = 1.10%
n 8 182 178.22 100(3.78/182) = 2.08%
Sum of % errors = 44.75%

MAPE =
åabsolute percent error = 44.75% = 5.59%
n 8
POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 49 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 50

Comparison of Measures Comparison of Forecast Error

Rounded Absolute Rounded Absolute
TABLE 4.1 Comparison of Measures of Forecast Error
Actual Forecast Deviation Forecast Deviation
Tonnage with for with for
MEASURE MEANING APPLICATION TO CHAPTER EXAMPLE Quarter Unloaded α = .10 α = .10 α = .50 α = .50
Mean absolute How much the forecast For α = .10 in Example 4, the forecast for grain 1 180 175 5.00 175 5.00
deviation (MAD) missed the target unloaded was off by an average of 10.31 tons.
2 168 175.5 7.50 177.50 9.50
Mean squared The square of how much For α = .10 in Example 5, the square of the forecast
error (MSE) the forecast missed the error was 190.8. This number does not have a 3 159 174.75 15.75 172.75 13.75
target physical meaning, but is useful when compared to the 4 175 173.18 1.82 165.88 9.12
MSE of another forecast.
5 190 173.36 16.64 170.44 19.56
Mean absolute The average percent For α = .10 in Example 6, the forecast is off by 5.59%
percent error error on average. As in Examples 4 and 5, some forecasts 6 205 175.02 29.98 180.22 24.78
(MAPE) were too high, and some were low. 7 180 178.02 1.98 192.61 12.61
8 182 178.22 3.78 186.30 4.30
82.45 98.62

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 51 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 52
 Comparison of Forecast Error Comparison of Forecast Error
Rounded Absolute Rounded Absolute Rounded Absolute
2 Rounded Absolute
Actual ∑ |deviations|
Forecast Deviation Forecast Deviation ∑ (forecast
Actual Forecasterrors)
Deviation Forecast Deviation
MAD =
Tonnage with for with for MSE =Tonnage with for with for
Quarter Unloaded α =n.10 α = .10 α = .50 α = .50 Quarter Unloaded α =n.10 α = .10 α = .50 α = .50
1 180 175 5.00 175 5.00 1 180 175 5.00 175 5.00
2 For = .10
α 168 175.5 7.50 177.50 9.50 2 For = .10
α 168 175.5 7.50 177.50 9.50
3 159 = 82.45/8
174.75 = 15.75
10.31 172.75 13.75 3 159 174.75
= 1,526.52/8 = 15.75
190.8 172.75 13.75
4 175 173.18 1.82 165.88 9.12 4 175 173.18 1.82 165.88 9.12
5 For α = .50 173.36
190 16.64 170.44 19.56 5 For α = .50 173.36
190 16.64 170.44 19.56
6 205 175.02 29.98 180.22 24.78 6 205 175.02 29.98 180.22 24.78
7 180 = 98.62/8
178.02 = 12.33
1.98 192.61 12.61 7 = 1,561.91/8
180 178.02 = 195.24
1.98 192.61 12.61
8 182 178.22 3.78 186.30 4.30 8 182 178.22 3.78 186.30 4.30
82.45 98.62 82.45 98.62
MAD 10.31 12.33

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 53 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 54

Comparison of Forecast Error Comparison of Forecast Error

n
Rounded
∑ Absolute
Actual 100|deviation
Forecast
Rounded
i|/actualiForecast
Deviation
Absolute
Deviation Actual
Rounded
Forecast
Absolute
Deviation
Rounded
Forecast
Absolute
Deviation
MAPEUnloaded
Quarter
= i = 1 α with
Tonnage
= .10
for
α = .10
with
α = .50
for
α = .50 Quarter
Tonnage
Unloaded
with
α = .10
for
α = .10
with
α = .50
for
α = .50
n
1 180 175 5.00 175 5.00 1 180 175 5.00 175 5.00
2 168= .10 175.5
For α 7.50 177.50 9.50 2 168 175.5 7.50 177.50 9.50
3 159 174.75 15.75
= 44.75%/8 = 5.59% 172.75 13.75 3 159 174.75 15.75 172.75 13.75
4 175 173.18 1.82 165.88 9.12 4 175 173.18 1.82 165.88 9.12
5 For 190=
α .50 173.36 16.64 170.44 19.56 5 190 173.36 16.64 170.44 19.56
6 205 175.02 29.98 180.22 24.78 6 205 175.02 29.98 180.22 24.78
7 180 = 54.00%/8
178.02 =1.98
6.75% 192.61 12.61 7 180 178.02 1.98 192.61 12.61
8 182 178.22 3.78 186.30 4.30 8 182 178.22 3.78 186.30 4.30
82.45 98.62 82.45 98.62
MAD 10.31 12.33 MAD 10.31 12.33
MSE 190.82 195.24 MSE 190.82 195.24
MAPE 5.59% 6.75%
POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 55 POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 56
 Exercise
The monthly sales were as follows:

a) Compute the forecasts for each month using a 2-month moving average
b) Compute the forecasts for each month using a 3-month moving average,
with the weights are 0.6, 0.3 and 0.1 respectively.
c) Compute the forecasts for each month using exponential smoothing, with an
initial forecast for January of 22. Use α = 0.3.
d) Which above model is the best using the mean absolute deviation (MAD)?

POM for Engineer Course: Chapter 2: Forecasting Techniques 4 - 57