0% found this document useful (0 votes)
7 views6 pages

Demand Forecasting with Exponential Smoothing

The document presents two forecasting methods: Single Exponential Smoothing and Exponential Smoothing with Trend, applied to demand data over 10 months with an alpha of 0.3. The results indicate that Single Exponential Smoothing has a lower mean absolute deviation (MAD) of 2.613 compared to 2.655 for the trend method, suggesting it is more accurate for this dataset. The calculations for forecasts and errors for each month are detailed, demonstrating the application of both methods.

Uploaded by

tranphumpt2005
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
7 views6 pages

Demand Forecasting with Exponential Smoothing

The document presents two forecasting methods: Single Exponential Smoothing and Exponential Smoothing with Trend, applied to demand data over 10 months with an alpha of 0.3. The results indicate that Single Exponential Smoothing has a lower mean absolute deviation (MAD) of 2.613 compared to 2.655 for the trend method, suggesting it is more accurate for this dataset. The calculations for forecasts and errors for each month are detailed, demonstrating the application of both methods.

Uploaded by

tranphumpt2005
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

a)

The single exponential smoothing forecast


The demand for the last 10 months
=0.3
F t=F t −1 +α ( A t −1 −F t −1)

Month Actual demand Forecast


1 31 31.000
2 34 31.000
3 33 31.900
4 35 32.230
5 37 33.061
6 36 34.243
7 38 34.770
8 40 35.739
9 40 37.017
10 41 37.912

F 2=F1 +α ( A 1−F 1 ) = 31 + 0.30 x (31 - 31) = 31

F 3=F2 + α ( A 2−F 2 ) = 31 + 0.30 x (34 - 31) = 31.9

F 4=F 3 +α ( A 3−F 3 ) = 31.9 + 0.30 x (33 - 31.9) = 32.23


F 5=F 4 +α ( A 4−F 4 ) = 32.23 + 0.30 x (35 - 32.23) = 33.061

F 6=F 5+ α ( A5 −F 5 ) = 33.06 + 0.30 x (37 - 33.06) = 34.243

F 7=F 6+ α ( A6 −F6 ) = 34.24 + 0.30 x (36 - 34.24) = 34.77

F 8=F 7+ α ( A7 −F7 ) = 34.77 + 0.30 x (38 - 34.77) = 35.739

F 9=F 8+ α ( A 8−F 8 ) = 35.74 + 0.30 x (40 - 35.74) = 37.017

F 10=F9 + α ( A 9−F 9 ) = 37.02 + 0.30 x (40 – 37.02) = 37.912

b)
The exponential smoothing with trend forecast
The demand for the last 10 months
=0.3
=0.3
FIT t=F t +T t

F t=F t −1 +α ( A t −1 −FIT t −1)

T t=T t −1 + δ ( Ft −FIT t−1 )

FIT 1=F1 +T 1 = 30 + 1 = 31

Month Actual demand Ft Tt FIT t


1 31 30.000 1.000 31.000
2 34 30.000 0.700 30.700
3 33 30.990 0.787 31.777
4 35 31.357 0.661 32.018
5 37 32.252 0.731 32.983
6 36 33.457 0.873 34.330
7 38 33.958 0.762 34.719
8 40 34.942 0.828 35.770
9 40 36.211 0.961 37.171
10 41 37.059 0.927 37.986

F 2=F1 +α ( A 1−FIT 1 ) = 30 + 0.3 x (31 - 31) = 30

T 2=T 1 + δ ( F2 −FIT 1 ) = 1 + 0.3 x (30 - 31) = 0.7


FIT 2=F2 +T 2 = 30 + 0.7 = 30.7

F 3=F2 + α ( A 2−FIT 2 ) = 30 + 0.3 x (34 – 30.7) = 30.99

T 3=T 2+ δ ( F 3−FIT 2 ) = 0.7 + 0.3 x (30.99 – 30.7) = 0.787

FIT 3=F 3 +T 3 = 30.99 + 0.787 = 31.777

F 4=F 3 +α ( A 3−FIT 3 ) = 30.99 + 0.3 x (33 – 31.78) = 31.357

T 4=T 3 +δ ( F 4 −FIT 3 ) = 0.787 + 0.3 x (31.356 – 31.78) = 0.661

FIT 4=F 4 +T 4 = 31.36 + 0.661 = 32.018

F 5=F 4 +α ( A 4−FIT 4 ) = 31.36 + 0.3 x (35 – 32.02) = 32.252

T 5=T 4 +δ ( F 5−FIT 4 ) = 0.661 + 0.3 x (32.25 – 32.02) = 0.731

FIT 5=F 5 +T 5 = 32.25 + 0.73 = 32.983

F 6=F 5+ α ( A5 −FIT 5 ) = 32.25 + 0.3 x (37 – 32.98) = 33.457

T 6=T 5+ δ ( F 6−FIT 5 ) = 0.73 + 0.3 x (33.46 – 32.98) = 0.873

FIT 6=F 6 +T 6 = 33.46 + 0.874 = 34.33

F 7=F 6+ α ( A6 −FIT 6 ) = 33.46 + 0.3 x (36 – 34.33) = 33.96

T 7=T 6+ δ ( F 7−FIT 6 ) = 0.874 + 0.3 x (33.96 – 34.33) = 0.763

FIT 7=F 7 +T 7 = 33.96 + 0.763 = 34.72

F 8=F 7+ α ( A7 −FIT 7 ) = 33.96 + 0.3 x (38 – 34.72) = 34.94

T 8=T 7+ δ ( F 8−FIT 7 ) = 0.763 + 0.3 x (34.94 – 34.72) = 0.829

FIT 8=F 8 +T 8 = 34.94 + 0.829 = 35.77


F 9=F 8+ α ( A 8−FIT 8 ) = 34.94 + 0.3 x (40 – 35.77) = 36.21

T 9=T 8 +δ ( F 9−FIT 8 ) = 0.829 + 0.3 x (36.21 – 35.77) = 0.961

FIT 9=F 9 +T 9 = 36.21 + 0.961 = 37.17

F 10=F9 + α ( A 9−FIT 9 ) = 36.21 + 0.3 x (40 – 37.17) = 37.06

T 10=T 9+ δ ( F 10−FIT 9 ) = 0.961 + 0.3 x (37.06 – 37.17) = 0.928

FIT 10=F10 + T 10 = 37.06 + 0.928 = 37.99

c)
The mean absolute deviation (MAD)
The demand for the last 10 months
=0.3
=0.3
e t = y t− ^y t

MAE=average(|et|¿

Single Exponential Smoothing Exponential smoothing with trend


Forecast forecast
Mont Actual Ft et FIT t et
h demand
1 31 31.000 0.000 31.000 0.000
2 34 31.000 3.000 30.700 3.300
3 33 31.900 1.100 31.777 1.223
4 35 32.230 2.770 32.018 2.982
5 37 33.061 3.939 32.983 4.017
6 36 34.243 1.757 34.330 1.670
7 38 34.770 3.230 34.719 3.281
8 40 35.739 4.261 35.770 4.230
9 40 37.017 2.983 37.171 2.829
10 41 37.912 3.088 37.986 3.014
MAE 2.613 MAE 2.655

Based on the calculation results, the Single Exponential Smoothing forecasting method gives
higher accuracy than the Exponential Smoothing with Trend method, as shown by the
significantly lower mean absolute deviation (MAD) value (2.613 vs. 2.655). This shows that
incorporating the trend into the forecasting model helps to more closely reflect the increasing
trend of the actual data over the months.

You might also like