Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 45

Demand Forecasting

Production and Operations Management

Judit Uzonyi-Kecskés
Ph.D. Student
Department of Management and Corporate Economics
Budapest University of Technology and Economics
uzonyi-kecskes@mvt.bme.hu
Literature
• Waters, D.: Operations Management – Producing
goods & Services
• Nahmias, S.: Production and Operation Analysis
• Vonderembse, M. A. and White, G. P.: Operations
Management – Concepts, Methods, and Strategies
Topics
• Introduction
• Forecasting methods
• Patterns of demand
• Forecasting stationary series
– Moving average (with example)
– Simple exponential smoothing (with example)
• Evaluating forecasts
– Analyzing the size of errors (with example)
– Analyzing the validity of the forecasting model (with
example)
Forecasting
• Predicting the future
• Application of forecasting results:
– Capacity planning
– Production scheduling
– Inventory control
– Materials requirement planning
Time Horizon in Forecasting
• Short-term: operative decisions – day to day
planning
• Medium term: tactical decisions – production
planning
• Long-term: strategic decisions – investment
decisions
• The longer the time horizon, the less reliable the
forecast is
– Availability and relevance of historical data
– Seriousness of any error
Forecasting Methods
• Subjective methods
• Objective methods
Subjective Forecasting Methods
• Based on expert opinion
– Personal insight
– Panel consensus
– Delphi method
– Historic analogy
• Based on customer opinion
– Indirectly: Sales force composites
– Directly: Market surveys
Objective Forecasting Methods
• Casual methods
– Analyzing the causes of the demand
– Forecasting the demand based on the measure of
the causes
• Time series/projective methods
– Analyzing the demand of previous periods
– Determining the patterns of the demand
– Forecasting the demand based on the information
of previous prior periods
Patterns of Demand

• Randomness
• Constant demand
• Trend
• Seasonality
Symbols
• t: period t (e.g. day, week, month)
• Dt: observation of demand in period t
• Ft,t+τ: forecast in period t for period t+τ
• Ft: forecast for period t
• Other parameters (e.g. time horizon parameter,
smoothing constants)
Forecasting Stationary Series

• Most frequently used methods:


– Moving average
– Simple exponential smoothing
Moving Average
• Forecasting:
1 tN
Ft    Di   Dt 1  Dt  2    Dt  N 
1
N i t 1 N
• N: number of analyzed periods
– Large N:
• more weight on past data
• forecasts are more stable
– Small N:
• more weight on the current observation of demand
• forecasts react quickly to changes in the demand
Example
In a car factory the management observed that the
demand for the factory’s car is nearly constant.
Therefore they forecast the demand with the help of
moving average based on the demand information of
the last 2 months.
Example
The observed demands in the last 7 periods were the
following:
Period Demand
1 200
2 255
3 176
4 189
5 224
6 283
7 308
Example
• The observed demand in the first two periods was
200 and 255 cars:
– D1=200,
– D2=255.
• The forecast is based on the demand information of
the last 2 months: N=2.
• The first period when forecast can be performed is
period 3: t=3
– Dt-1= D3-1 =D2=255
– Dt-N= D3-2 =D1=200
Example
• Forecast for the third period, if N=2:
1 tN
Ft    Di   Dt 1  Dt  2    Dt  N 
1
N i t 1 N

F3   D1  D2  
1
2
  200  255  227,5
1
2
• Forecasts for the following periods:
F4   255  176  215,5 F6   189  224  206,5
1 1
2 2
F5   176  189  182,5 F7   224  283  253,5
1 1
2 2
F8   283  308  295,5
1
2
Example
• Comparison of the observed and the forecasted
demand
– Draw attention to systematic error in forecasting
– Help to identify outlier data
350
300
250 Observed
200 demand
150 Forecast
100
50
0
1 2 3 4 5 6 7 8
Example
• Multiple-step-ahead forecast
– Last known demands: D6=283 and D7=308.
– Last forecast: F8=295,5.
• We assume that demand is constant!
F7 ,8  F7 ,9  F7 , 7 n  295,5

• Suppose that in period 8 we observe a demand of


D8=195, we now need to update the forecasts:

F9  F8,9  F8,10  F8,8n   308  195  251,5


1
2
Moving average defects

• Same weight
• Constant demand
• Large amount of historic data
Exponential Smoothing
• Forecast is a weighted average
• Current forecast is based on:
– Last forecast
– Last value of demand
– Smoothing constant (e.g. α):
0 ≤ α, ≤ 1

New forecast    last demand  1   last forecast


Simple Exponential Smoothing
• Forecast
Ft    Dt 1  1    Ft 1
• α: smoothing constant (0 ≤ α ≤ 1)
– Large α:
• more weight on the current observation of demand
• forecasts react quickly to changes in the demand
– Small α:
• more weight on past data
• forecasts are more stable
Example
In a car factory the management observed that the
demand for the factory’s car is nearly constant.
Therefore they forecast the demand with the help of
simple exponential smoothing, and they use α=0.1
value as smoothing constant. The forecast for the first
period was 250 cars.
Example
The observed demands in the last 7 periods were the
following:
Period Demand
1 200
2 255
3 176
4 189
5 224
6 283
7 308
Example
• The forecast for the first period was 250 cars:
F1=250.
• The observed demand in the first period was 200
cars: D1=200.
• Forecast for the second period, if α=0.1:

Ft    Dt 1  1     Ft 1

F2    D1  1     F1 
 0.1 200  1  0.1  250 
 20  225  245
Example
F3  0.1 255  0.9  245  246 F4  0.1176  0.9  246  239
F5  0.1189  0.9  239  234 F6  0.1 224  0.9  234  233
F7  0.1 283  0.9  233  238 F8  0.1 308  0.9  238  245

400

300 Observed
demand
200
Forecast
100

0
1 2 3 4 5 6 7 8
Example
• More-step-ahead forecast
– Last known demand: D7=308.
– Last forecast: F8=245.
• We assume that demand is constant!
F7,8  F7,9  F7,7n  245
• Suppose that in period 8 we observe a demand of
D8=195, we now need to update the forecasts:
F9  F8,9  F8,10  F8,8n  0.1195  0.9  245  240
Comparison of the Two Methods
• Similarities
– Both assume that demand is stationary
– Both use a single parameter (N or α)
• Differences
– Number of directly used demand data
– Number and weights of indirectly used demand
data
Evaluating Forecasts
• There are almost always errors in forecasts
– Random effects, noises
– Inappropriate forecasting methods
• Analysis of
– the size of forecasting errors
– the validity of forecasting models
Forecast Error
• Difference between the forecasted value for a period
and the actual demand for the same period
et  Ft  ,t  Dt
et  Ft  Dt
• Covers only one period
• Does not give information about the acceptability of
the forecasting method
Mean Error
• The average error during a term of n periods
t

e i
ME t  i t T 1
T
• Positive and negative errors cancel each other
Absolute Error Measures
• Measures of forecasts accuracy during n periods
• Mean absolute error
t
1
MAE t    ei
T i t T 1

• Positive and negative errors cannot cancel each other


• Does not give information about the relative size of
error
Mean Absolute Percentage Error
• Arithmetical average of percentage error of n periods
t
100 %
1 ei
MAPE t   
T i t T 1 Di

• Gives information about the average, relative size of


the absolute error observed during several periods
Example
• We have the following forecast and demand data.
Evaluate the size of forecast errors.

Period Demand Forecast


1 100 110
2 130 169
3 150 135
4 140 168
5 110 121
Example
• First determine the forecast error in each period

Period Demand Forecast Error


1 100 110
2 130 169 39
3 150 135 -15
4 140 168 28
5 110 121 11
Example
• Determine the presented error measures after period 5
(t=5, T=4)

e i
39  (15)  28  11
ME 5  i 5 4 1 2
  15.75
4 4

1 5 39  15  28  11
MAD 5    ei   23.25
4 i 2 4
Example

1 5 2 1
MSE 5    ei   (39   15  282  112 ) 
2 2

4 i 2 4
1
  (1521  225  784  121)  662.75
4

1 5 ei 1  39 15 28 11 
MAPE 5    100       
4 i  2 Di 4  130 150 140 110 
1
  (0.3  0.1  0.2  0.1)  0.175  17.5%
4
Validity of Forecasting Method
• Analyzing the validity of the forecasting method used
• Signs that forecast
– is inappropriate
– will be inappropriate in the immediate future
• Tracking signal will be used
• Monitoring
– the size of tracking signal values
– the tendency of tracking signal values
Tracking Signal
• Moving sum of forecast error in period t
t
MSFE t  e
i t T 1
i

• Mean absolute error in period t


1 t
MAE t   ei
T i t T 1
• Tracking signal in period t
MSFE t
TS t 
MAE t
Monitoring the Tracking Signal
• Monitoring size
 6  TS t  6
 4  TS t  4

• Monitoring tendency
– Tracking signal diagram
– Typical patterns:
• Small-scale, random alternating near to zero
• Increasing trend
• Decreasing trend
• Regular alternating
Example
• We have the following forecast and demand data.
Evaluate the validity of forecast model.

Period Demand Forecast


1 100 110
2 126 130
3 124 120
4 129 125
5 135 115
Example
• Determine the value of tracking signal in each period

Period Dt Ft et MSFEt |et| MAEt TSt


1 100 110
2 126 130 4 4 4 4 1
3 124 120 -4 0 4 4 0
4 129 125 -4 -4 4 4 -1
5 135 115 -20 -24 20 8 -3
Example
• Draw the tracking signal diagram
Tracking signal 2
1 1
Periods
0 0

-1 2 3 4 -1 5
-2
-3 -3

-4

• Evaluate the validity of forecasting method applied


– Only few data were available
– Does not step out of control borders
– Decreasing trend, systematic undervaluation
– There is a negative trend instead of constant demand, there is a
constant demand instead of positive trend, etc.
Possible questions in the exam
• Name subjective forecasting methods
• In which life cycle period are subjective/objective
methods used?
• Name the similarities/differences between moving
average and exponential smoothing.
• Name differences between forecasts made by simple
exponential smoothing(moving average) with a small
and a large α (N) value?
• Name three different forecasting errors
Possible exercises in the exam
• Give forecast using moving average
• Give forecast using exponential smoothing
• Determine the values of simple error / mean error /
absolute mean error

• You can find examples for these in the presentation!


Exercise for extra points
• The demand for a product is constant. Make forecasts
for periods 3 and 4. Use moving average method.
N=2.
Period 1 2 3 4
Demand 140 150 200 220

• Make forecasts for periods 2,3 and 4. Use exponential


smoothing. α=0.3
• Give a multiple-step-ahead forecast for period 7 from
period 4.

You might also like