Chapter 8 - Forecasting

Chapter 8 - Forecasting
Operations Management
6th Edition
R. Dan Reid & Nada R. Sanders
Copyright © 2016 John Wiley & Sons, Inc.

Learning Objectives
2
Identify principles of forecasting.

Explain the steps involved in the
forecasting process.
Identify types of forecasting methods
and their characteristics.
Describe time series models.

Learning Objectives - cont'd
3
Describe causal modeling using linear

regression.
Compute forecast accuracy.
Explain the factors that should be
considered when selecting a forecasting
model.
Explain the nine-step process of CPFR.

Forecasting - Defined
4
Predicting future events

One of the most important business functions as
decisions are based on a forecast of the future
Goal: Generate good forecasts on the average over
time and keep errors low
Forecasting is an ongoing process

Principles of Forecasting
5
Many types of forecasting models differ in complexity and

amount of data & way they generate forecasts.
Common features include:
1. Forecasts are rarely perfect
2. Forecasts are more accurate for grouped data
than for individual items
3. Forecast are more accurate for shorter than
longer time periods

Steps in the Forecasting Process
6
Decide what needs to be forecast

 Level of detail, units of analysis & time horizon required
Evaluate and analyze appropriate data
 Identify needed data & whether it’s available
Select and test the forecasting model
 Cost, ease of use & accuracy
Generate the forecast
Monitor forecast accuracy over time

Types of Forecasting Methods – cont’d
7
Classified into two groups:

Types of Forecasting Models
8
Qualitative methods – judgmental methods

Forecasts generated subjectively by the forecaster
 Educated guesses
Quantitative methods – based on mathematical

modeling:
 Forecasts generated through mathematical modeling

Qualitative Methods
9

Quantitative Methods
10
Time Series Models: Causal Models or

 Assumes information needed Associative Models
to generate a forecast is  Explores cause-and-effect
contained in a time series of relationships
data  Uses leading indicators to
 Assumes the future will predict the future
follow same patterns as the  Housing starts and appliance
past
sales

Time Series Models
11
 Forecaster looks for data patterns as

 Data = historic pattern + random variation
 Historic pattern to be forecasted:
 Level (long-term average) – data fluctuates around a constant mean
 Trend – data exhibits an increasing or decreasing pattern
 Seasonality – any pattern that regularly repeats itself and is of a
constant length
 Cycle – patterns created by economic fluctuations
 Random Variation cannot be predicted

Time Series Patterns
12

Time Series Models
13
Naive: Ft 1  At
 The forecast is equal to the actual value observed during the last
period – good for level patterns
Simple Mean: Ft 1   A t / n
 The average of all available data - good for level patterns
Simple Moving Average: Ft 1   A t / n
 The average value over a set time period
(e.g.: the last four weeks)
 Each new forecast drops the oldest data point & adds a new
observation
 More responsive to a trend but still lags behind actual data - good
for level patterns; trend + level = bad forecast
Time Series Models cont'd
14
Weighted Moving Average: Ft 1   C t A t

 Method in which “n” of the most recent observations are
averaged and past observations may be weighted differently
 All weights must add to 100% or 1.00
e.g. Ct .5, Ct-1 .3, Ct-2 .2 (weights add to 1.0)

 Allows emphasizing one period over others; above indicates
more weight on recent data (Ct=.5)
 Differs from the simple moving average that weighs all
periods equally - more responsive to trends

Time Series Models cont'd
15
Exponential Smoothing: Ft 1  αA t   1  α  Ft
 Most frequently used time series method because of ease of
use and minimal amount of data needed
 Need just three pieces of data to start:
 Last period’s forecast (Ft)
 Last periods actual value (At)
 Select value of smoothing coefficient, ,between 0 and 1.0
 If no last period forecast is available, average the last few
periods or use naive method
 
Higher values (e.g. .7 or .8) place a lot of weight on current
periods actual demand and influenced by random variation
Time Series Problem
16
 Determine forecast for periods 7 & 8 Period Actual

 2-period moving average 1 300
 4-period moving average 2 315
 2-period weighted moving average 3 290
with t-1 weighted 0.6 and t-2 weighted
4 345
0.4
 Exponential smoothing with alpha=0.2
5 320
and the period 6 forecast being 375 6 360
7 375
8
Time Series Problem Solution
17
Exponential
Period Actual 2-Period 4-Period 2-Per.Wgted. Smoothing
1 300
2 315
3 290
4 345
5 320
6 360
7 375 340.0 328.8 344.0 372.0
8 367.5 350.0 369.0 372.6

Forecasting Trend
18
 Basic forecasting models for trends compensate for the lagging that
would otherwise occur
 One model, trend-adjusted exponential smoothing uses a three step
process
 Step 1 - Smoothing the level of the series
S t  αA t  (1  α)(S t 1  Tt 1 )
 Step 2 – Smoothing the trend
Tt  β(S t  S t 1 )  (1  β)Tt 1
 Step 3 - Forecast including the trend
FITt 1  S t  Tt
Forecasting trend problem: a company uses exponential smoothing with trend to forecast usage of its lawn care
products. At the end of July the company wishes to forecast sales for August. July demand was 62. The trend
through June has been 15 additional gallons of product sold per month. Average sales have been 57 gallons per
month. The company uses alpha+0.2 and beta +0.10. Forecast for August.
19
 Smooth the level of the series:

S July  αA t  (1  α)(S t 1  Tt 1 )   0.2 62   0.8 57  15  70
 Smooth the trend:

TJuly  β(St  St 1 )  (1  β)Tt 1   0.1 70  57   0.915  14.8
 Forecast including trend:

FITAugust  S t  Tt  70  14.8  84.8 gallons
Linear Trend Line
20
A time series technique that computes a forecast with

trend by drawing a straight line through a set of data
using this formula:
Y = a + bx
where
Y = forecast for period X
X = the number of time periods from X = 0
A = value of y at X = 0 (Y intercept)
B = slope of the line

Forecasting Seasonality
21
Remember it is a regularly repeating pattern

Examples:
 University enrollment varies between quarters or semesters; higher in
the fall than in the summer
Seasonal Index:
 Percentage amount by which data for each season are above
or below the mean.

Forecasting Seasonality Steps
22
1. Calculate the average demand per season

 E.g.: average quarterly demand
2. Calculate a seasonal index for each season of

each year:
 Divide the actual demand of each season by the average
demand per season for that year
3. Average the indexes by season
 E.g.: take the average of all Spring indexes, then of all
Summer indexes, ...

Forecasting Seasonality Steps - cont'd
23
4. Forecast demand for the next year & divide by

the number of seasons
 Use regular forecasting method & divide by four for
average quarterly demand
5. Multiply next year’s average seasonal demand by
each average seasonal index
 Result is a forecast of demand for each season of next
year

Seasonality problem: a university must develop forecasts for the next year’s quarterly enrollments. It has
collected quarterly enrollments for the past two years. It has also forecast total enrollment for next year to
be 90,000 students.
What is the forecast for each quarter of next year?
24
Quarter Year Seasonal Year Seasonal Avg. Year3

1 Index 2 Index Index
Fall 24000 1.2 26000 1.238 1.22 27450
Winter 23000 22000
Spring 19000 19000
Summer 14000 17000
Total 80000 84000 90000
Average 20000 21000 22500

Causal Models
25
 Often, leading indicators can help to predict changes in

future demand e.g. housing starts
 Causal models establish a cause-and-effect relationship
between independent and dependent variables
 A common tool of causal modeling is linear regression:
Y  a  bx
 Additional related variables may require multiple
regression modeling

Linear Regression
26
 Identify dependent (y) and

independent (x) variables
 Solve for the slope of the line
b
 XY   X  Y 
 X 2    X  X  b
 XY  n XY

2
X  nX
2
 Solve for the y intercept
a  Y  bX
 Develop your equation for the
trend line
Y=a + bX
Linear Regression Problem: A maker of golf shirts has been tracking the relationship
between sales and advertising dollars. Use linear regression to find out what sales might be
if the company invested $53,000 in advertising next year.
27
Sales $
(Y)
Adv.$
(X)
XY X^2 Y^2
b
 XY  n XY
 X  nX 2 2
1 130 32 4160 2304 16,900
2 151 52 7852 2704 22,801 b
28202  4 47.25 147.25
 1.15
9253  4 47.25
2
3 150 50 7500 2500 22,500 a  Y  b X  147.25  1.15 47.25

a  92.9
4 158 55 8690 3025 24964 Y  a  bX  92.9  1.15X
Y  92.9  1.15 53  153.85
5 153.85 53
Tot 589 189 28202 9253 87165
Avg 147.25 47.2
5

Correlation Coefficient- How Good is the Fit?
28
 Correlation coefficient (r) measures the direction and strength of the linear
relationship between two variables. The closer the r value is to 1.0 the better the
r
       
regression line fits the data points.
n XY  X Y
n  X     X  * n  Y    Y 
2 2
2 2
4 28,202   189 589 

r  .992
4(9253) - (189) * 4 87,165   589 
2 2
r 2   .982   .964
2
 Coefficient of determination ( r 2) measures the amount of variation in the dependent

variable about its mean that is explained by the regression line. Values of (r 2 ) close to
1.0 are desirable.
Multiple Regression
29
An extension of linear regression but:

 Multiple regression develops a relationship between a
dependent variable and multiple independent variables. The
general formula is:

Measuring Forecast Accuracy
30
Forecasts are never perfect

Need to measure over time
Need to know how much we should rely on our
chosen forecasting method
Measuring forecast error:
E t  A t  Ft
Note that over-forecasts = negative errors and
under-forecasts = positive errors

Measuring Forecasting Accuracy
31
 Mean Absolute Deviation (MAD)

 measures the total error in a forecast
without regard to sign MAD 
 actual  forecast
n
 Cumulative Forecast Error (CFE)
 Measures any bias in the forecast
 Mean Square Error (MSE) CFE    actual  forecast 

 Penalizes larger errors
 Tracking Signal
  actual - forecast  2
 Measures if your model is working; MSE 

quality n
CFE
TS 
MAD

Accuracy & Tracking Signal Problem: A company is comparing the accuracy of two forecasting methods. Forecasts
using both methods are shown below along with the actual values for January through May. The company also uses a
tracking signal with ±4 limits to decide when a forecast should be reviewed. Which forecasting method is best?
32
Method A Method B
Month Actual F’cast Error Cum. Tracking F’cast Error Cum. Tracking
sales Error Signal Error Signal
Jan. 30 28 2 2 2 27 2 2 1
Feb. 26 25 1 3 3 25 1 3 1.5
March 32 32 0 3 3 29 3 6 3
April 29 30 -1 2 2 27 2 8 4
May 31 30 1 3 3 29 2 10 5
MAD 1 2
MSE 1.4 4.4
Selecting the Right Forecasting Model
33
1. The amount & type of available data

 Some methods require more data than others
2. Degree of accuracy required
 Increasing accuracy means more data
3. Length of forecast horizon
 Different models for 3 month vs. 10 years
4. Presence of data patterns
 Lagging will occur when a forecasting model meant for a
level pattern is applied with a trend

Forecasting Software
34
Spreadsheets
Microsoft Excel, Quattro Pro, Lotus 1-2-3
 Limited statistical analysis of forecast data
Statistical packages
SPSS, SAS, NCSS, Minitab
 Forecasting plus statistical and graphics
Specialty forecasting packages

 Forecast Master, Forecast Pro, Autobox, SCA
 Extensive range of forecasting capability

Guidelines for Selecting Software
35
 Does the package have the features you want?

 What platform is the package available for?
 How easy is the package to learn and use?
 Is it possible to implement new methods?
 Do you require interactive or repetitive forecasting?
 Do you have any large data sets?
 Is there local support and training available?
 Does the package give the right answers?
 What is the cost of the package?
 Is it compatible with our existing software

Collaborative Planning Forecasting &
Replenishment(CPFR)
36
 Establish collaborative relationships between buyers and sellers

 Create a joint business plan
 Create a sales forecast
 Identify exceptions for sales forecast
 Resolve/collaborate on exception items
 Create order forecast
 Identify exceptions for order forecast
 Resolve/collaborate on exception items
CPFR is an
 Generate order
iterative
process.

Forecasting Within OM: How It All Fits
Together
37
 Forecasts impact not only other business functions but all other
operations decisions. Operations managers make many forecasts,
such as the expected demand for a company’s products.
 These forecasts are then used to determine:
Product designs that are expected to sell (Ch 2)
The quantity of product to produce (Chs 5 and 6)
The amount of needed supplies and materials (Ch 12)
Future space requirements (Ch 10)
Capacity and location needs (Ch 9)
The amount of labor needed (Ch 11)

Forecasting within OM - cont'd
38
Forecasts drive strategic operations decisions, such

as:
Choice of competitive priorities, changes in processes, and
large technology purchases (Ch 3)
Forecast decisions serve as the basis for tactical planning;
developing worker schedules (Ch 11)
Virtually all operations management decisions are

based on a forecast of the future.

Forecasting Across the Organization
39
Forecasting is critical to management of all

organizational functional areas
 Marketing relies on forecasting to predict demand and
future sales
 Finance forecasts stock prices, financial performance,
capital investment needs..
 Information systems provides ability to share databases and
information
 Human resources forecasts future hiring requirements

Chapter 8 Highlights
40
 Three basic principles of forecasting are: forecasts are rarely

perfect, are more accurate for groups than individual items, and are
more accurate in the shorter term than longer time horizons.
 The forecasting process involves five steps: decide what to
forecast, evaluate and analyze appropriate data, select and test
model, generate forecast, and monitor accuracy.
 Forecasting methods can be classified into two groups: Qualitative
methods are based on subjective opinion of forecaster and
quantitative methods are based on mathematical modeling.

Chapter 8 Highlights - cont'd
41
 Time series models are based on the assumption that all

information needed is contained in the time series of data.
Causal models assume that the variable being forecast is related
to other variables in the environment.
 There are four basic patterns of data: level or horizontal, trend,
seasonality, and cycles. In addition, data usually contain random
variation. Some forecast models used to forecast the level of a
time series are: naïve, simple mean, simple moving average,
weighted moving average, and exponential smoothing. Separate
models are used to forecast trends and seasonality.

Chapter 8 Highlights - cont'd
42
 A simple causal model is linear regression in which a straight-line

relationship is modeled between the variable we are forecasting
and another variable in the environment. The correlation measures
the strength of the linear relationship between these two variables.
 Three useful measures of forecast error are mean absolute
deviation (MAD), mean square error (MSE) and tracking
signal.
 There are four factors when selecting a model: amount and type
of data available, degree of accuracy required, length of forecast
horizon, and patterns present in the data.

Chapter 8 - Forecasting

Uploaded by

Copyright:

Available Formats

Chapter 8 - Forecasting

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 8 - Forecasting

Uploaded by

Copyright:

Available Formats

Chapter 8 - Forecasting

Copyright © 2016 John Wiley & Sons, Inc.

Identify principles of forecasting.

Copyright © 2016 John Wiley & Sons, Inc.

Describe causal modeling using linear

Copyright © 2016 John Wiley & Sons, Inc.

Predicting future events

Copyright © 2016 John Wiley & Sons, Inc.

Many types of forecasting models differ in complexity and

Copyright © 2016 John Wiley & Sons, Inc.

Decide what needs to be forecast

Copyright © 2016 John Wiley & Sons, Inc.

Classified into two groups:

Copyright © 2016 John Wiley & Sons, Inc.

Qualitative methods – judgmental methods

Quantitative methods – based on mathematical

Copyright © 2016 John Wiley & Sons, Inc.

Copyright © 2016 John Wiley & Sons, Inc.

Time Series Models: Causal Models or

Copyright © 2016 John Wiley & Sons, Inc.

 Forecaster looks for data patterns as

Copyright © 2016 John Wiley & Sons, Inc.

Copyright © 2016 John Wiley & Sons, Inc.

Weighted Moving Average: Ft 1   C t A t

e.g. Ct .5, Ct-1 .3, Ct-2 .2 (weights add to 1.0)

Copyright © 2016 John Wiley & Sons, Inc.

 Determine forecast for periods 7 & 8 Period Actual

7 375 340.0 328.8 344.0 372.0

8 367.5 350.0 369.0 372.6

 Smooth the level of the series:

 Smooth the trend:

 Forecast including trend:

A time series technique that computes a forecast with

Copyright © 2016 John Wiley & Sons, Inc.

Remember it is a regularly repeating pattern

Copyright © 2016 John Wiley & Sons, Inc.

1. Calculate the average demand per season

2. Calculate a seasonal index for each season of

Copyright © 2016 John Wiley & Sons, Inc.

4. Forecast demand for the next year & divide by

Copyright © 2016 John Wiley & Sons, Inc.

Quarter Year Seasonal Year Seasonal Avg. Year3

Winter 23000 22000

Spring 19000 19000

Summer 14000 17000

Total 80000 84000 90000

Average 20000 21000 22500

 Often, leading indicators can help to predict changes in

Copyright © 2016 John Wiley & Sons, Inc.

 Identify dependent (y) and

3 150 50 7500 2500 22,500 a  Y  b X  147.25  1.15 47.25

Copyright © 2016 John Wiley & Sons, Inc.

4 28,202   189 589 

 Coefficient of determination ( r 2) measures the amount of variation in the dependent

An extension of linear regression but:

Copyright © 2016 John Wiley & Sons, Inc.

Forecasts are never perfect

Copyright © 2016 John Wiley & Sons, Inc.

 Mean Absolute Deviation (MAD)