Forecasting

August 29, Wednesday

Course
Structure

Introduction
Operations Strategy & Competitiveness
Quality Management
Strategic Decisions (some)

Design of Products
and Services

Process Selection
and Design

Capacity and
Facility Decisions

Forecasting

Tactical & Operational Decisions

Forecasting
Predict the next number in the pattern:
a) 3.7,

3.7,

3.7,

3.7,

3.7,

b) 2.5,

4.5,

6.5,

8.5,

10.5,

c) 5.0, 7.5, 6.0, 4.5, 7.0, 9.5, 8.0, 6.5, ?

Forecasting
Predict the next number in the pattern:
a) 3.7,

3.7,

3.7,

3.7,

3.7, 3.7

b) 2.5,

4.5,

6.5,

8.5,

10.5, 12.5

c) 5.0, 7.5, 6.0, 4.5, 7.0, 9.5, 8.0, 6.5, 9.0

Outline
What is forecasting?

Types of forecasts
Time-Series forecasting
Nave

Moving Average
Exponential Smoothing
Regression

Good forecasts

What is Forecasting?
Process of predicting a future
event based on historical data
Educated Guessing
Underlying basis of
all business decisions

Production
Inventory
Personnel
Facilities

Why do we need to forecast?

In general, forecasts are almost always wrong. So,
Throughout the day we forecast very different
things such as weather, traffic, stock market, state
of our company from different perspectives.

Virtually every business attempt is based on

forecasting. Not all of them are derived from
sophisticated methods. However, Best" educated
guesses about future are more valuable for
purpose of Planning than no forecasts and hence
no planning.

Importance of Forecasting in OM
Departments throughout the organization depend on
forecasts to formulate and execute their plans.
Finance needs forecasts to project cash flows and
capital requirements.
Human resources need forecasts to anticipate hiring
needs.
Production needs forecasts to plan production
levels, workforce, material requirements,
inventories, etc.

Importance of Forecasting in OM
Demand is not the only variable of interest to
forecasters.
Manufacturers also forecast worker
absenteeism, machine availability, material
costs, transportation and production lead
times, etc.
Besides demand, service providers are also
interested in forecasts of population, of other
demographic variables, of weather, etc.

Types of Forecasts by Time Horizon

Quantitative
methods

Short-range forecast
Usually < 3 months
Job scheduling, worker assignments

Medium-range forecast

Detailed
use of
system

3 months to 2 years
Sales/production planning

Long-range forecast
> 2 years
New product planning

Design
of system
Qualitative
Methods

Forecasting During the Life Cycle

Introduction

Qualitative models
- Executive judgment
- Market research
-Survey of sales force
-Delphi method
Sales

Growth

Maturity

Quantitative models
- Time series analysis
- Regression analysis

Time

Decline

Qualitative Forecasting Methods

Qualitative
Forecasting

Executive
Judgement

Sales
Force
Composite

Market
Research/
Survey
Smoothing

Models
Delphi
Method

Qualitative Methods
Briefly, the qualitative methods are:
Executive Judgment: Opinion of a group of high level
experts or managers is pooled
Sales Force Composite: Each regional salesperson
provides his/her sales estimates. Those forecasts are then
reviewed to make sure they are realistic. All regional
forecasts are then pooled at the district and national levels
to obtain an overall forecast.
Market Research/Survey: Solicits input from customers
pertaining to their future purchasing plans. It involves the
use of questionnaires, consumer panels and tests of new
products and services.

Qualitative Methods
Delphi Method: As opposed to regular panels where the individuals
involved are in direct communication, this method eliminates the
effects of group potential dominance of the most vocal members. The
group involves individuals from inside as well as outside the
organization.
Typically, the procedure consists of the following steps:
Each expert in the group makes his/her own forecasts in form of
statements
The coordinator collects all group statements and
summarizes them
The coordinator provides this summary and gives another set
of questions to each
group member including feedback as to the input of other
experts.
The above steps are repeated until a consensus is reached.
.

Quantitative Forecasting Methods

Quantitative
Forecasting
Regression
Models

Time Series
Models

1. Naive

2. Moving
Average
a) simple
b) weighted

3. Exponential
Smoothing
a) level
b) trend
c) seasonality

Quantitative Forecasting Methods

Quantitative
Forecasting
Regression
Models

Time Series
Models

1. Naive

2. Moving
Average
a) simple
b) weighted

3. Exponential
Smoothing
a) level
b) trend
c) seasonality

Time Series Models

Try to predict the future based on past

data
Assume that factors influencing the past will
continue to influence the future

Time Series Models: Components

Random

Trend

Seasonal

Composite

Demand for product or service

Product Demand over Time

Year
1

Year
2

Year
3

Year
4

Product Demand over Time

Trend component
Demand for product or service

Seasonal peaks

Random
variation
Year
1

Year
2

Actual
demand line
Year
3

Year
4

Now lets look at some time series approaches to forecasting

Borrowed from Heizer/Render - Principles of Operations Management, 5e, and Operations Management, 7e

Quantitative Forecasting Methods

Quantitative
Time Series
Models
Models
1. Naive

2. Moving
Average
a) simple
b) weighted

3. Exponential
Smoothing
a) level
b) trend
c) seasonality

1. Naive Approach

Demand in next period is the same as

demand in most recent period
May sales = 48 June forecast = 48

Usually not good

2a. Simple Moving Average

Assumes an average is a good estimator of
future behavior

Used if little or no trend

Used for smoothing
A t + A t -1 + A t -2 + ...+ A t -n 1
Ft 1 =
n
Ft+1
n
At

= Forecast for the upcoming period, t+1

= Number of periods to be averaged
= Actual occurrence in period t

2a. Simple Moving Average

A t + A t -1 + A t -2 + ...+ A t -n 1
Ft 1 =
n

Youre manager in Amazons electronics

department. You want to forecast ipod sales for
months 4-6 using a 3-period moving average.

Month
1
2
3
4
5
6

Sales
(000)
4
6
5
?
?
?

2a. Simple Moving Average

A t + A t -1 + A t -2 + ...+ A t -n 1
Ft 1 =
n

Youre manager in Amazons electronics

department. You want to forecast ipod sales for
months 4-6 using a 3-period moving average.

Month
1
2
3
4
5
6

Sales
(000)
4
6
5
?
?
?

Moving Average
(n=3)
NA
NA
NA
(4+6+5)/3=5

2a. Simple Moving Average

What if ipod sales were actually 3 in month 4

Month
1
2
3
4
5
6

Sales
(000)
4
6
5
3?
?
?

Moving Average
(n=3)
NA
NA
NA
5

2a. Simple Moving Average

Forecast for Month 5?

Month
1
2
3
4
5
6

Sales
(000)
4
6
5
3
?
?

Moving Average
(n=3)
NA
NA
NA
5
(6+5+3)/3=4.667

2a. Simple Moving Average

Actual Demand for Month 5 = 7

Month
1
2
3
4
5
6

Sales
(000)
4
6
5
3
?7
?

Moving Average
(n=3)
NA
NA
NA
5
4.667

2a. Simple Moving Average

Forecast for Month 6?

Month
1
2
3
4
5
6

Sales
(000)
4
6
5
3
7
?

Moving Average
(n=3)
NA
NA
NA
5
4.667
(5+3+7)/3=5

2b. Weighted Moving Average

Gives more emphasis to recent data

Ft 1 = w 1A t + w 2 A t -1 + w 3A t -2 + ... + w n A t -n 1
Weights
decrease for older data
sum to 1.0

Simple moving
average models
weight all previous
periods equally

Ft 1 = w 1A t + w 2 A t -1 + w 3A t -2 + ... + w n A t -n 1

2b. Weighted Moving Average: 3/6, 2/6, 1/6

Month
1
2
3
4
5
6

Sales
(000)
4
6
5
?
?
?

Weighted
Moving
Average
NA
NA
NA
31/6 = 5.167

Ft 1 = w 1A t + w 2 A t -1 + w 3A t -2 + ... + w n A t -n 1

2b. Weighted Moving Average: 3/6, 2/6, 1/6

Month
1
2
3
4
5
6

Sales
(000)
4
6
5
3
7

Weighted
Moving
Average
NA
NA
NA
31/6 = 5.167
25/6 = 4.167
32/6 = 5.333

3a. Exponential Smoothing

Assumes the most recent observations
have the highest predictive value
gives more weight to recent time periods

Ft+1 = Ft + a(At - Ft)

et
Ft+1 = Forecast value for time t+1
At
= Actual value at time t
a
= Smoothing constant

Need initial
forecast Ft
to start.

3a. Exponential Smoothing Example 1

Ft+1 = Ft + a(At - Ft)

i

Week
1
2
3
4
5
6
7
8
9
10

Ai

Demand
820
775
680
655
750
802
798
689
775

Given the weekly demand

data what are the exponential
smoothing forecasts for
periods 2-10 using a=0.10?
Assume F1=D1

3a. Exponential Smoothing Example 1

Ft+1 = Ft + a(At - Ft)

i

Week
1
2
3
4
5
6
7
8
9
10

Ai

Fi

a = 0.1
Demand
0.6
820
820.00
820.00
775
820.00
820.00
= F1+ a(A793.00
680 F2815.50
1F1) =820+.1(820820)
655
801.95
725.20=820
750
787.26
683.08
802
783.53
723.23
798
785.38
770.49
689
786.64
787.00
775
776.88
728.20
776.69
756.28

3a. Exponential Smoothing Example 1

Ft+1 = Ft + a(At - Ft)

i

Week
1
2
3
4
5
6
7
8
9
10

Ai

Fi

a = 0.1
Demand
0.6
820
820.00
820.00
775
820.00
820.00
680
815.50
793.00
F3 = F2+ a(A2F2) =820+.1(775820)
655
801.95
725.20
750
787.26
683.08=815.5
802
783.53
723.23
798
785.38
770.49
689
786.64
787.00
775
776.88
728.20
776.69
756.28

3a. Exponential Smoothing Example 1

Ft+1 = Ft + a(At - Ft)

i

Week
1
2
3
4
5
6
7
8
9
10

Ai

Demand
820
775
680
655
750
802
798
689
775

Fi

a = 0.1

820.00
820.00
815.50
801.95
787.26
783.53
785.38
786.64
776.88
776.69

0.6
820.00
820.00
793.00
725.20
683.08
This process
723.23
continues
770.49
through week
787.00
10
728.20
756.28

3a. Exponential Smoothing Example 1

Ft+1 = Ft + a(At - Ft)

i

Week
1
2
3
4
5
6
7
8
9
10

Ai

Demand
820
775
680
655
750
802
798
689
775

Fi

a = 0.1

a = 0.6

820.00
820.00
815.50
801.95
787.26
783.53
785.38
786.64
776.88
776.69

820.00
820.00
793.00
725.20
683.08
723.23
770.49
787.00
728.20
756.28

What if the
a constant
equals 0.6

3a. Exponential Smoothing Example 2

Ft+1 = Ft + a(At - Ft)

i

Ai

Month Demand
January
120
February
90
March
101
April
91
May
115
June
83
July
August
September

Fi

a = 0.3

a = 0.6

100.00
106.00
101.20
101.14
98.10
103.17
97.12

100.00
112.00
98.80
100.12
94.65
106.86
92.54

What if the
a constant
equals 0.6

3a. Exponential Smoothing Example 3

Company A, a personal computer producer
purchases generic parts and assembles them to
final product. Even though most of the orders
require customization, they have many common
components. Thus, managers of Company A need
a good forecast of demand so that they can
purchase computer parts accordingly to minimize
inventory cost while meeting acceptable service
level. Demand data for its computers for the past 5
months is given in the following table.

3a. Exponential Smoothing Example 3

Ft+1 = Ft + a(At - Ft)

i

Month
January
February
March
April
May
June
July

Ai

Fi

Demand
80
84
82
85
89
??

a = 0.3

a = 0.5

84.00
82.80
83.16
82.81
83.47
85.13

84.00
82.00
83.00
82.50
83.75
86.38
??

What if the
a constant
equals 0.5

3a. Exponential Smoothing

How to choose
depends on the emphasis you want to place
on the most recent data

Increasing makes forecast more

sensitive to recent data

Forecast Effects of
Smoothing Constant a
or

Ft+1 = Ft + a (At - Ft)

Ft+1 = a At + a(1- a) At - 1 + a(1- a)2At - 2 + ...
w1

w2

w3

Weights

a=

a= 0.10
a= 0.90

Prior Period

2 periods ago 3 periods ago

a(1 - a)

a(1 - a)2

10%

9%

8.1%

90%

9%

0.9%

To Use a Forecasting Method

Collect historical data

Select a model
Moving average methods
Select n (number of periods)
For weighted moving average: select weights

Exponential smoothing
Select a

Selections should produce a good forecast

but what is a good forecast?

A Good Forecast
Has a small error
Error = Demand - Forecast

Measures of Forecast Error

et
n

a. MAD = Mean Absolute Deviation

- Ft

t=1

MAD =

n
n

b. MSE = Mean Squared Error

A
F
t t

MSE =

t =1

c. RMSE = Root Mean Squared Error RMSE = MSE

Ideal values =0 (i.e., no forecasting error)

MAD Example

MAD =

- Ft

t=1

What is the MAD value given the

forecast values in the table below?
At

Month
1
2
3
4
5

Ft

Sales Forecast
220
n/a
250
255
210
205
300
320
325
315

|At Ft|
5
5
20
10
= 40

= 40 =10
4

A
F
t t

MSE/RMSE Example

MSE =

t =1

= 550 =137.5
4

RMSE = 137.5

What is the MSE value?

=11.73
At

Month
1
2
3
4
5

Ft

Sales Forecast
220
n/a
250
255
210
205
300
320
325
315

|At Ft| (At Ft)2

5
5
20
10

25
25
400
100
= 550

Measures of Error
1. Mean Absolute Deviation
(MAD)

At

Ft

Jan

120

100

Feb

90

106

Mar

101

102

et
20

20

-16

16

-1

May
June

91
115
83

400
256
1

10

103

84
6

= 14

2a. Mean Squared Error

(MSE)

100

101
98

MAD

-10

April

|et|

et2

17

17

289

-20

20

400

-10

84

1,446

MSE

2
e

t
1

1,446
= 241
6

2b. Root Mean Squared Error

(RMSE)

An accurate forecasting system will have small MAD,

MSE and RMSE; ideally equal to zero. A large error may
indicate that either the forecasting method used or the
parameters such as used in the method are wrong.
Note: In the above, n is the number of periods, which is

RMSE MSE
= SQRT(241)
=15.52

Forecast Bias
How can we tell if a forecast has a positive or
negative bias?
TS = Tracking Signal
Good tracking signal has low values

(actual forecast )

t
t
RSFE
TS =
= t
MAD Mean absolute
deviation
MAD
30

Quantitative Forecasting Methods

Quantitative
Forecasting
Regression
Models

Time Series
Models

1. Naive

2. Moving
Average
a) simple
b) weighted

3. Exponential
Smoothing
a) level
b) trend
c) seasonality

Exponential Smoothing (continued)

We looked at using exponential
smoothing to forecast demand with
only random variations
Ft+1 = Ft + a (At - Ft)
Ft+1 = Ft + a At a Ft
Ft+1 = a At + (1-a) Ft

Exponential Smoothing (continued)

We looked at using exponential
smoothing to forecast demand with
only random variations
What if demand varies due to
randomness and trend?
What if we have trend and seasonality
in the data?

Regression Analysis as a Method for

Forecasting
Regression analysis takes advantage
of the relationship between two
variables. Demand is then
forecasted based on the
knowledge of this relationship and
for the given value of the related
variable.
Ex: Sale of Tires (Y), Sale of Autos (X)
are obviously related
If we analyze the past data of these
two variables and establish a
relationship between them, we may
use that relationship to forecast the
sales of tires given the sales of
automobiles.
The simplest form of the relationship
is, of course, linear, hence it is
referred to as a regression line.

Sales of Autos (100,000)

Formulas
y=a+bx
where,
x

xy n x y

b
x nx
2

x
y

a y bx

Regression Example

y = a+ b X

xy n x y

b
x nx
2

MonthAdvertising
January
3
February
4
March
2
April
5
May
4
June
2
July
TOTAL

20

Sales
1
2
1
3
2
1

10

a y bx

X 2 XY
9.00
3.00
16.00
8.00
4.00
2.00
25.00
15.00
16.00
8.00
4.00
2.00

74

38

General Guiding Principles for

Forecasting
1.
2.
3.
4.

Forecasts are more accurate for larger groups of items.

Forecasts are more accurate for shorter periods of time.
Every forecast should include an estimate of error.
Before applying any forecasting method, the total system
should be understood.
5. Before applying any forecasting method, the method
should be tested and evaluated.
6. Be aware of people; they can prove you wrong very easily
in forecasting

FOR JULY 2nd MONDAY

READ THE CHAPTERS ON

Forecasting
Product and service design