Professional Documents
Culture Documents
MENG 445 Production and Inventory Control: Forecasting
MENG 445 Production and Inventory Control: Forecasting
Forecasting
Gaafar 2005
Aggregate planning
Production
Inventory management
Inventory
Vendors
Gaafar 2005
Forecasting
Basic Problem: predict demand for planning purposes. Forecasting Tools: Qualitative: Delphi Analogies Quantitative: Causal models (e.g., regression models) Time series models
Gaafar 2005
Forecasting Laws
1) Forecasts are always wrong! 2) Forecasts always change! 3) The further into the future, the less reliable the forecast!
40% 20%
+10% -10% Start of season
Trumpet of Doom
16 weeks 26 weeks
Gaafar 2005
Why Forecasting?
Most decisions made in PP&C are plans for future How much of a certain product should I make next week? How much inventory should I keep in the May 2001? If you forecast the future (more accurately), you will (more likely) have better plans. Many decisions (plans) in PPC depend on the forecasts of demand. Forecasting demand versus forecasting sales.
Gaafar 2005
Qualitative Forecasting
Delphi Technique: Iteratively tapping multiple experts opinion with continuous comparisons and revisions.
Gaafar 2005
Quantitative Forecasting
Goals: Predict future from past Smooth out noise Standardize forecasting procedure Methodologies: Causal Forecasting: regression analysis Time Series Forecasting: moving average exponential smoothing regression analysis seasonal models
Gaafar 2005
Forecast Generation
Current Observation
Forecast
Forecast Control
Modified Forecast
Production and Inventory Control (8) Gaafar 2005
xt, t = 1, ,T
Observed value at time t
(T )
xt
T
Production and Inventory Control (9)
T+
t
Gaafar 2005
Possible Patterns
Gaafar 2005
MT
V (M T ) s , N
2
T (T ) M T 2 1 x
1 s N
Averages the latest N observations. N ranges from 2 to 12. Larger N provides stability, but less responsive. Assigns equal weights to the latest N observations Does not adapt to error
Production and Inventory Control (12) Gaafar 2005
Mt
t (t 1) x
et
|et|
2.29
Production and Inventory Control (13) Gaafar 2005
T (T ) S T 2 1 x
a
2 a
Combines current observation and previous forecasts for future predictions. a ranges from 0.1 to 0.5. Smaller a provides stability, but less responsive.
t (t 1) x
et
|et|
Gaafar 2005
Model: xt = a + bt + et, et NID(0, s2) Assumptions: Linear trend Noise is normally distributed t Noise is independent Noise has a constant variance (doesnt change with time) Noise averages out to zero
Gaafar 2005
t T N 1 T
t x
t T N 1
2 t (
x t T N 1 t T N 1 T 2 t T N 1
x t t )
t x b a
2 1 ( T t ) (T ) 2 1 ( T ) a b s xT N S tt
Gaafar 2005
M
M
[ 2] T
1 N
T
t T N 1
x,
t
[ 2]
N 1 E[ M T ] E[ xT ] b 2
( N 1) b 2
1 N
t T N 1
M t,
[ 2]
E[ M T ] E[ xT ] ( N 1)b E[ M T ]
T (T ) 2 M T M T , x
T (T ) (2 x
2 2 [ 2] ) M T (1 ) M T , (point estimate) N 1 N 1
2 2 1 (T t )2 [ 2] T (T ) (2 N 1) M T (1 N 1) M T 2 1 N S tt s x
Gaafar 2005
S
S
T
a S T (1 a ) S T 1 , E[S T
[ 2]
[ 2] [ 2]
[ 2]
2 ] E[ xT ] b E[S T ] b a a
2
[ 2]
a S T (1 a ) S T 1 ,
[ 2]
E[ S T ] E[ xT ]
b E[ S T ]
b a
T (T ) 2 S T S T , x
T (T ) (2 x
a a [ 2] ) S T (1 ) S T , (point estimate )
a a [ 2] 1 (T t )2 T (T ) (2 ) S T (1 ) S T 2 1 N S tt s x
Production and Inventory Control (19) Gaafar 2005
Example The quarterly sales of company XYZ over the past five years are given below. Use this data to forecast the sales in every quarter of years 6 and 7. Use the same data to forecast the total sales of years 6 and 7.
Quarter Year 1 2 3 4 5 I 17.51 21.33 25.26 27.38 32.99 II 40.24 47.81 50.04 65.50 67.27 III 35.76 45.45 53.92 56.58 65.94 IV 20.08 24.11 29.65 26.51 34.34
Gaafar 2005
Gaafar 2005
a = 23.55 b = 5.28.
Seasonal Indices
3 4 5 Average
Gaafar 2005
Seasonal Indices
Year
I 34.77 38.09
II 76.05 83.33
IV 37.95 41.58
Forecasted Values
6 7
Gaafar 2005