Sunteți pe pagina 1din 48

PRODUCTION AND OPERATIONS MANAGEMENT

Ch. 5: Forecasting

POM - J. Galvn

Learning Objectives


Understand techniques to foresee the future

POM - J. Galvn

What is Forecasting?
Process of predicting a future event Underlying basis of all business decisions
Production Inventory Personnel Facilities

POM - J. Galvn 3

Sales will be $200 Million!

Types of Forecasts by Time Horizon




ShortShort-range forecast
Up

to 1 year; usually < 3 months Job scheduling, worker assignments


3

MediumMedium-range forecast
months to 3 years Sales & production planning, budgeting

LongLong-range forecast
3+

years New product planning, facility location


POM - J. Galvn 4

Short-term vs. Longer-term ShortLongerForecasting




Medium/long range forecasts deal with more comprehensive issues and support management decisions regarding planning and products, plants and processes. ShortShort-term forecasting usually employs different methodologies than longerlongerterm forecasting ShortShort-term forecasts tend to be more accurate than longer-term forecasts. longerPOM - J. Galvn 5

Influence of Product Life Cycle




Stages of introduction & growth require longer forecasts than maturity and decline Forecasts useful in projecting
staffing levels, inventory levels, and factory capacity

as product passes through stages


POM - J. Galvn 6

Types of Forecasts


Economic forecasts
Address

business cycle e.g., inflation rate, money supply etc.




Technological forecasts
Predict

technological change Predict new product sales




Demand forecasts
Predict

existing product sales


POM - J. Galvn 7

Seven Steps in Forecasting


  

   

Determine the use of the forecast Select the items to be forecast Determine the time horizon of the forecast Select the forecasting model(s) Gather the data Make the forecast Validate and implement results
POM - J. Galvn 8

Realities of Forecasting
 

Forecasts are seldom perfect Most forecasting methods assume that there is some underlying stability in the system Both product family and aggregated product forecasts are more accurate than individual product forecasts
POM - J. Galvn 9

Forecasting Approaches
Qualitative Methods Quantitative Methods
Used when situation is Used when situation vague & little data is stable & historical exist data exist New products Existing products New technology Current technology Involves intuition, Involves mathematical experience techniques e.g., forecasting sales e.g., forecasting sales on Internet of color televisions
POM - J. Galvn 10

Overview of Qualitative Methods




Jury of executive opinion Pool opinions of high-level executives, sometimes highaugment by statistical models Sales force composite estimates from individual salespersons are reviewed for reasonableness, then aggregated Delphi method Panel of experts, queried iteratively Consumer Market Survey Ask the customer

POM - J. Galvn

11

Jury of Executive Opinion


Involves small group of high-level managers

Group estimates demand by working together

Combines managerial experience with statistical models Relatively quick Group-think disadvantage
POM - J. Galvn 12

1995 Corel Corp.

Sales Force Composite


Each salesperson projects their sales Combined at district & national levels Sales reps know customers wants Tends to be overly optimistic
POM - J. Galvn

Sales

1995 Corel Corp.

13

Delphi Method


Iterative group process 3 types of Staff people (What


Decision Staff Respondents

Decision Makers
(Sales?) (Sales will be 50!)

makers will sales


be? survey)

Reduces groupgroupthink
POM - J. Galvn

Respondents
(Sales will be 45, 50, 55)
14

Consumer Market Survey


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.

POM - J. Galvn

15

Overview of Quantitative Approaches


  

Nave approach Moving averages Exponential smoothing Trend projection Linear regression
POM - J. Galvn

Time-series Models

Causal models
16

5-22

Quantitative Forecasting Methods


(Non(Non-Naive)
Quantitative Forecasting Time Series Models Causal Models

Moving Average

Exponential Smoothing

Trend Projection

Linear Regression

POM - J. Galvn

17

What is a Time Series?




Set of evenly spaced numerical data Obtained by observing response variable at regular time periods Forecast based only on past values Assumes that factors influencing past, present, & future will continue Example Year: 1993 1994 1995 1996 1997 Sales: 78.7 63.5 89.7 93.2 92.1

POM - J. Galvn

18

Time Series Components

Trend

Cyclical

Seasonal
POM - J. Galvn

Random
19

Trend Component


 

Persistent, overall upward or downward pattern Due to population, technology etc. Several years duration
Response

Mo., Qtr., Yr.


POM - J. Galvn

1984-1994 T/Maker Co. 20

Cyclical Component
 

Repeating up & down movements Due to interactions of factors influencing economy Usually 2-10 years duration 2Cycle Response

Mo., Qtr., Yr.


POM - J. Galvn 21

Seasonal Component


 

Regular pattern of up & down fluctuations Due to weather, customs etc. Occurs within 1 year
Summer Response
1984-1994 T/Maker Co.

POM - J. Galvn Mo., Qtr.

22

Random Component


Erratic, unsystematic, residual fluctuations Due to random variation or unforeseen events


strike Tornado
Union

Short duration & nonrepeating


POM - J. Galvn 23

General Time Series Models




Any observed value in a time series is the product (or sum) of time series components Multiplicative model Yi = Ti Si Ci Ri (if quarterly or mo. data) Additive model Yi = Ti + Si + Ci + Ri (if quarterly or mo. data)

POM - J. Galvn

24

Naive Approach
Assumes demand in next period is the same as demand in most recent period e.g., If May sales were 48, then June sales will be 48 Sometimes cost effective & efficient
1995 Corel Corp.
POM - J. Galvn 25

Moving Average Method


  

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 Demand in Previous n Periods MA ! n


POM - J. Galvn 26

Moving Average Graph


Sales 8 6 4 2 0 93
Actual

Forecast

94

95 96 Year
POM - J. Galvn

97

98
27

Disadvantages of Moving Average Method




 

Increasing n makes forecast less sensitive to changes Do not forecast trend well Require much historical data
1984-1994 T/Maker Co.

POM - J. Galvn

28

Linear Trend Projection




Used for forecasting linear trend line Assumes relationship between response variable, Y, and time, X, is a linear function  Yi ! a  bX i Estimated by least squares method

Minimizes sum of squared errors


POM - J. Galvn 29

Linear Trend Projection Model Yi !abX i


Y
a b<0 a Time, X
POM - J. Galvn 30

b>0

Scatter Diagram
Sales 4 3 2 1 0 92
Sales vs. Time

93

94

95

96

Time
POM - J. Galvn 31

Least Squares Equations


Equation:
Yi ! a  bx i
n

Slope:

b ! i ! n  x i  n x 
i !

x i y i  nx y

Y-Intercept:

a ! y  bx
POM - J. Galvn 32

Multiplicative Seasonal Model




 

Find average historical demand for each season by summing the demand for that season in each year, and dividing by the number of years for which you have data. Compute the average demand over all seasons by dividing the total average annual demand by the number of seasons. Compute a seasonal index by dividing that seasons historical demand (from step 1) by the average demand over all seasons. Estimate next years total demand Divide this estimate of total demand by the number of seasons then multiply it by the seasonal index for that season. This provides the seasonal forecast. 33 forecast. POM - J. Galvn

Linear Regression Model




Shows linear relationship between dependent & explanatory variables


Example:

Sales & advertising (not time) (not


Slope

Y-intercept

^ Yi = a + b X i
Dependent (response) variable
POM - J. Galvn

Independent (explanatory) variable


34

Linear Regression Model


Y
Yi = a + b X i + Error
Error Regression line

^ =a +b X Yi i

X
Observed value
POM - J. Galvn 35

Linear Regression Equations


Equation:
Yi ! a  bx i
n

Slope:

b ! i ! n  x i  n x 
i !

x i y i  nx y

Y-Intercept:

a ! y  bx
POM - J. Galvn 36

Interpretation of Coefficients


Slope (b) (b
Estimated
 If

Y changes by b for each 1 unit increase in X


b = 2, then sales (Y) is expected to (Y increase by 2 for each 1 unit increase in advertising (X) (X

Y-intercept (a) (a
Average
 If

value of Y when X = 0

a = 4, then average sales (Y) is expected (Y to be 4 when advertising (X) is 0 (X


POM - J. Galvn 37

Correlation


Answers: how strong is the linear how relationship between the variables? Coefficient of correlation Sample correlation coefficient denoted r

Values range from -1 to +1 Measures degree of association

Used mainly for understanding


POM - J. Galvn 38

Coefficient of Correlation Values


Perfect Negative Correlation No Correlation Perfect Positive Correlation

-1.0

-.5

+.5

+1.0

Increasing degree of negative correlation


POM - J. Galvn

Increasing degree of positive correlation


39

Coefficient of Correlation and Regression Model


Y

r=1
^ Yi = a + b Xi X

r = -1

^ Yi = a + b Xi X

r = .89
^ Yi = a + b Xi X

r=0
^ Yi = a + b Xi X
40

POM - J. Galvn

Guidelines for Selecting Forecasting Model




You want to achieve:


No

= (Yi - Yi) = (Actual - Forecast) (Y  Seen in plots of errors over time


 Error

pattern or direction in forecast error ^ forecast error

Smallest
 Mean

square error (MSE)  Mean absolute deviation (MAD)

POM - J. Galvn

41

Pattern of Forecast Error


Trend Not Fully Accounted for Error 0 Time (Years) Error 0 Time (Years)

Desired Pattern

POM - J. Galvn

42

Tracking Signal


Measures how well forecast is predicting actual values Ratio of running sum of forecast errors (RSFE) to mean absolute deviation (MAD)
Good

tracking signal has low values

Should be within upper and lower control limits


POM - J. Galvn 43

Tracking Signal Plot


3 2 1 TS 0 -1 -2 -3
1 2 3 4 5 6
44

Time
POM - J. Galvn

Forecasting in the Service Sector




Presents unusual challenges


need for short term records needs differ greatly as function of industry and product issues of holidays and calendar unusual events
special

POM - J. Galvn

45

Forecasting example
SALES DURING LAST YEAR
LAST YEAR Spring Summer Fall Winter TOTAL ANNUAL SALES ESTIMATION: Real sales 200 350 300 150 1000 Annual increase of sales 10,00%

What are the estimated seasonal sales amount for next year?
POM - J. Galvn 46

Forecasting example (II)


LAST YEAR Past sales Average sales for each season Total past annual sales/ n of seasons Spring Summer Fall Winter TOTAL ANNUAL SALES 200 350 300 150 250 250 250 250 Seasonal factor Past sales/ Avg. Sales 0,8 1,4 1,2 0,6

1000
POM - J. Galvn

1000
47

Forecasting example (III)


NEXT YEAR NEXT YEAR SALES 1100 Average sales for each season Total estimated annual sales/n of seasons Spring Summer Fall Winter TOTAL ANNUAL SALES ? ? ? ? 1100 275 275 275 275 1100
POM - J. Galvn

(10% increase) Seasonal factor As calculated 0,8 1,4 1,2 0,6 Next year's seasonal forecast Avg.sales* Factor 220 385 330 165 1100
48