demand forecasting
Demand Forecasting
Demand forecasting
•
Why is it important
•
How to evaluate
•
Qualitative Methods
•
Causal Models
•
Time-Series Models
•
Summary
Production and operations management
Product
Development
long term medium term short term Product portifolio Purchasing Manufacturing Distribution
Supply network designFacility
Partner selection location
Distribution network design and layout
Derivatuve
Supply Demand forecasting is product developmentcontract the starting
? point of all
?
adaptions design planning and control !
Current
product support Demand fulfillment Materials ordering Demand forecasting
Inventory
management
Production Distribution control planning
Fulfillment
Operations Transport implementation scheduling planning
Wal-mart experience - situation 1996
Wal-Mart
Store1
Warner-Lambert
Other suppliers store2500 Wal-mart experience-situation now
•
Collaborative forecasting and replenishment software installed
•
Initial forecasts generated by Wal-Mart
•
Forecasts refined by Warner-Lambert
•
Inventory cost reduced by 70%
•
Service levels improved from 96% to 99%
•
System adapted by others
Common features of all forecasts
•
Forecasts are usually wrong; Knowledge of the forecast error makes forecasts more meaningful
•
Aggregate forecasts are more accurate than individual forecasts
•
Short-term forecasts are more accurate than long-term forecasts
•
Choosing appropriate aggregation levels, time horizons, and forecasting techniques is crucial
Which forecast is better ?
Forecast quality
•Which forecast is better?
•How can we evaluate the forecasting performance?
Measures of forecast accuracy
Define forecast error
t f t At
A t : actual value in period t; ft : forecast for period t
Mean absolute deviation
1 T
MAD = t
T t 1
Mean squared error
2
1 T
MSE = t
T t 1
Mean absolute percentage error
1 T t
MAPE =
T t 1 At
Measuring forecast accuracy
MAD
= 4.56/5=0.912
MSE
= 7.15/5=1.430
Demand forecasting
•
Why is it important
•
How to evaluate
•
Qualitative Methods
•
Causal Models
•
Time-Series Models
•
Summary
Production and operations management
Product
Development
long term medium term short term Product portifolio Purchasing Manufacturing Distribution
Supply network designFacility
Partner selection location
Distribution network design and layout
Derivatuve
Supply Demand forecasting is product developmentcontract the starting
? point of all
?
adaptions design planning and control !
Current
product support Demand fulfillment Materials ordering Demand forecasting
Inventory
management
Production Distribution control planning
Fulfillment
Operations Transport implementation scheduling planning
Wal-mart experience - situation 1996
Wal-Mart
Store1
Warner-Lambert
Other suppliers store2500 Wal-mart experience-situation now
•
Collaborative forecasting and replenishment software installed
•
Initial forecasts generated by Wal-Mart
•
Forecasts refined by Warner-Lambert
•
Inventory cost reduced by 70%
•
Service levels improved from 96% to 99%
•
System adapted by others
Common features of all forecasts
•
Forecasts are usually wrong; Knowledge of the forecast error makes forecasts more meaningful
•
Aggregate forecasts are more accurate than individual forecasts
•
Short-term forecasts are more accurate than long-term forecasts
•
Choosing appropriate aggregation levels, time horizons, and forecasting techniques is crucial
Which forecast is better ?
Forecast quality
•Which forecast is better?
•How can we evaluate the forecasting performance?
Measures of forecast accuracy
Define forecast error
t f t At
A t : actual value in period t; ft : forecast for period t
Mean absolute deviation
1 T
MAD = t
T t 1
Mean squared error
2
1 T
MSE = t
T t 1
Mean absolute percentage error
1 T t
MAPE =
T t 1 At
Measuring forecast accuracy
MAD
= 4.56/5=0.912
MSE
= 7.15/5=1.430