CHAPTER 4: FORECASTING Suggested Solutions
Summer II, 2009
Question 4.1
(a) 3-week moving average:
(b) 3-week weighted moving average:
Week of
Pints used
Weight
Computations
August 31
360
September 7
389
September 14
410
September 21
381
0.1
381x0.1= 38.1
September 28
368
0.3
368x0.3=110.4
October 5
374
0.6
374x0.6=224.4
October 12 Forecast
372.9
(c) Exponential smoothing (with a smoothing constant, = 0.2):
Week of
Forecast: Ft = Ft-1 + (At-1 Ft-1)
August 31
360
360.00
September 7
389
360.00 = 360.00 + 0.2(360 - 360.00)
September 14
410
365.80 = 360.00 + 0.2(389 - 360.00)
September 21
381
374.64 = 365.80 + 0.2(410 - 365.80)
September 28
368
375.91 = 374.64 + 0.2(381 - 374.64)
October 5
374
374.33 = 375.91 + 0.2(368 - 375.91)
October 12
Actual
374.26 = 374.33 + 0.2(374 - 374.33)
BUS P301:01
Question 4.5
(a) 2-year moving average:
(b) Mean Absolute Deviation (MAD)
Two-year
Moving Average
Error
IErrorI
3400
3500
-100
100
3800
3700
100
100
3700
3600
100
100
Totals
100
300
MAD =
300/3 = 100
Year
Mileage
3000
4000
3
4
5
(c) 2-Year Weighted Moving Average
Two-year Weighted
Moving Average
Error
IErrorI
3400
3600
-200
200
3800
3640
160
160
3700
3640
60
60
Totals
20
420
MAD =
420/3 = 140
Year
Mileage
3000
4000
3
4
5
(d) Exponential Smoothing using =0.5 and an initial forecast of 3000 for year 1.
Year
Mileage
Forecast
3000
3000
4000
Forecast Error
Error x 0.5
New Forecast
3000
3000
1000
500
3500
3400
3500
-100
-50
3450
3800
3450
350
175
3625
3700
3625
75
38
3663
Total
1325
The forecast for year 6 is 3,663 miles.
BUS P301:01
Question 4.6 Raw data set up for trend projections
January
February
March
April
May
June
July
August
September
October
November
December
Sum
Average
Y Sales
X Period
20
21
15
14
13
16
17
18
20
20
21
23
1
2
3
4
5
6
7
8
9
10
11
12
1
4
9
16
25
36
49
64
81
100
121
144
XY
20
42
45
56
65
96
119
144
180
200
231
276
78
6.5
650
1474
218
18.17
(a) Plotting the monthly sales data:
(b) [i] Nave method: The coming January = December = 23
[ii] 3-month moving average: (20 + 21 + 23)/3 = 21.33
[iii] 6-m weighted [(.1 17)+(.1 18)+(.1 20)+(.2 20)+(.2 21)+ (.3 23)]/1.0 = 20.6
[iv] Exponential smoothing with alpha, = 0.3
FOct
18 0.3(20 18)
FNov
18.6 0.3(20 18.6)
FDec
19.02 0.3(21 19.02)
FJan
19.6 0.3(23 19.6)
[v] Trend
78, x
6.5,
18.6
y = 218, y
19.02
19.6
20.62
21
18.17
1474 (12)(6.5)(18.2) 54.4
0.38
650 12(6.5)2
143
a 18.2 0.38(6.5) 15.73
Forecast = 15.73 + .38(13) = 20.67, where next January is the 13th month.
b
(c) Only trend provides an equation that can extend beyond one month.
BUS P301:01
Question 4.7 Weighted Moving Average. Assume that Present = Period (week) 6.
So:
1
A
3 6
F7
1
A
4 5
1
A
4 4
1
A
6 3
1.0
1
1
1
1
(52) +
(63) +
(48) +
(70) = 56.75 patients
3
4
4
6
Where:
1.0 = weights 0.333 + 0.25 + 0.25 + 0.167 or 1/3, , , 1/6
Question 4.13
(a) Exponential smoothing,
Year
1
2
3
4
5
6
Demand
45
50
52
56
58
?
= 0.6:
Exponential
Smoothing = 0.6
41
41.0 + 0.6(4541) = 43.4
43.4 + 0.6(5043.4) = 47.4
47.4 + 0.6(5247.4) = 50.2
50.2 + 0.6(5650.2) = 53.7
53.7 + 0.6(5853.7) = 56.3
Absolute
Deviation
4.0
6.6
4.6
5.8
4.3
= 25.3
MAD = 5.06
Exponential smoothing,
Year
1
2
3
4
5
6
= 0.9:
Demand
Exponential
Smoothing = 0.9
Absolute
Deviation
45
50
52
56
58
?
41
41.0 + 0.9(4541) = 44.6
44.6 + 0.9(5044.6 ) = 49.5
49.5 + 0.9(5249.5) = 51.8
51.8 + 0.9(5651.8) = 55.6
55.6 + 0.9(5855.6) = 57.8
4.0
5.4
2.5
4.2
2.4
= 18.5
MAD = 3.7
(b) 3-year moving average:
Year
1
2
3
4
5
6
Demand
45
50
52
56
58
?
Three-Year
Moving Average
Absolute
Deviation
(45 + 50 + 52)/3 = 49
(50 + 52 + 56)/3 = 52.7
(52 + 56 + 58)/3 = 55.3
7
5.3
= 12.3
MAD = 6.2
BUS P301:01
(c) Trend projection:
Year
Demand
1
2
3
4
5
6
45
50
52
56
58
?
Absolute
Deviation
Trend Projection
42.6 + 3.2
42.6 + 3.2
42.6 + 3.2
42.6 + 3.2
42.6 + 3.2
42.6 + 3.2
1 = 45.8
2 = 49.0
3 = 52.2
4 = 55.4
5 = 58.6
6 = 61.8
0.8
1.0
0.2
0.6
0.6
= 3.2
MAD = 0.64
a bX
XY nXY
b
a
X 2 nX 2
Y bX
XY
X2
1
2
3
4
5
45
50
52
56
58
45
100
156
224
290
1
4
9
16
25
Then: X = 15, Y = 261, XY = 815, X2 = 55,
X=
3,
= 52.2
Therefore:
815 5 3 52.2
b
3.2
55 5 3 3
a 52.20 3.20 3 42.6
Y6 42.6 3.2 6 61.8
(d)
Comparing the results of the forecasting methodologies for parts (a), (b), and (c).
Forecast Methodology
Exponential smoothing,
= 0.6
MAD
5.06
Exponential smoothing,
= 0.9
3.7
3-year moving average
Trend projection
6.2
0.64
Based on a mean absolute deviation criterion, the trend projection is to be
preferred over the exponential smoothing with = 0.6, exponential smoothing
with = 0.9, or the 3-year moving average forecast methodologies.
BUS P301:01
Question 4.14
Week
Actual
Method 1
Error
IErrorI
Error^2
Method 2
Error
IErrorI
Error^2
0.7
0.90
-0.20
0.20
0.04
0.80
-0.10
0.10
0.010
1.0
1.05
-0.05
0.05
0.00
1.20
-0.20
0.20
0.040
1.0
0.95
0.05
0.05
0.00
0.90
0.10
0.10
0.010
1.0
1.20
-0.20
1.11
-0.11
0.11
0.012
0.510
0.072
0.20
0.04
Totals
0.50
0.085
MAD
0.125
<<Better
MSE
0.021
MAD
0.128
MSE
0.018
<<Better
BUS P301:01
Question 4.39 Raw data set up for trend analysis:
Year X
1
2
3
4
5
6
7
8
9
10
55
Patients Y
X2
Y2
XY
36
33
40
41
40
55
60
54
58
61
478
1
4
9
16
25
36
49
64
81
100
385
1,296
1,089
1,600
1,681
1,600
3,025
3,600
2,916
3,364
3,721
23,892
36
66
120
164
200
330
420
432
522
610
2,900
Given: Y = a + bX where:
b
a
XY
nXY
nX 2
X
Y
bX
and
X
= 55,
5.5, Y
= 478,
XY
= 2900,
X2
= 385,
271
82.5
3.28
Y2
= 23892,
47.8,
Then:
b
a
2900 10 5.5 47.8
385 10 5.52
47.8 3.28 5.5 29.76
2900 2629
385 302.5
and Y = 29.76 + 3.28X. For:
11: Y
29.76 3.28 11 65.8
12: Y
29.76 3.28 12
Therefore:
Year 11
Year 12
69.1
65.8 patients
69.1 patients
The model seems to fit the data pretty well. One should, however, be more
precise in judging the adequacy of the model.
Two possible approaches are computation of
(a) the correlation coefficient, or (b) the mean absolute deviation.
BUS P301:01
The correlation coefficient:
n
XY
X2
X
10
10
385
X Y
2
n Y2
2900
552
55 478
10
23892
4782
29000 26290
3850 3025 238920 228484
2710
825 10436
r2
2710
2934.3
0.924
0.853
The coefficient of determination of 0.853 is quite respectableindicating our
original judgment of a good fit was appropriate.
Year
X
1
2
3
4
5
6
7
8
9
10
Patients
Y
36
33
40
41
40
55
60
54
58
61
Trend
Forecast
29.8 + 3.28
29.8 + 3.28
29.8 + 3.28
29.8 + 3.28
29.8 + 3.28
29.8 + 3.28
29.8 + 3.28
29.8 + 3.28
29.8 + 3.28
29.8 + 3.28
1 = 33.1
2 = 36.3
3 = 39.6
4 = 42.9
5 = 46.2
6 = 49.4
7 = 52.7
8 = 56.1
9 = 59.3
10 = 62.6
Deviation
Absolute
Deviation
2.9
3.3
0.4
1.9
6.2
5.6
7.3
2.1
1.3
1.6
2.9
3.3
0.4
1.9
6.2
5.6
7.3
2.1
1.3
1.6
= 32.6
MAD = 3.26
The MAD is 3.26this is approximately 7% of the average number of patients
and 10% of the minimum number of patients. We also see absolute deviations, for
years 5, 6, and 7 in the range 5.67.3. The comparison of the MAD with the
average and minimum number of patients and the comparatively large deviations
during the middle years indicate that the forecast model is not exceptionally
accurate. It is more useful for predicting general trends than the actual number of
patients to be seen in a specific year.
BUS P301:01
Question 4.40 Raw data set up for trend analysis:
Crime
Rate X
Year
1
2
3
4
5
6
7
8
9
10
Column Totals
Patients
Y
58.3
61.1
73.4
75.7
81.1
89.0
101.1
94.8
103.3
116.2
854.0
36
33
40
41
40
55
60
54
58
61
478
X2
Y2
XY
3,398.9
3,733.2
5,387.6
5,730.5
6,577.2
7,921.0
10,221.2
8,987.0
10,670.9
13,502.4
76,129.9
1,296
1,089
1,600
1,681
1,600
3,025
3,600
2,916
3,364
3,721
23,892
2,098.8
2,016.3
2,936.0
3,103.7
3,244.0
4,895.0
6,066.0
5,119.2
5,991.4
7,088.2
42,558.6
Given: Y = a + bX where
XY nXY
b
X 2 nX 2
bX
and X = 854, Y = 478, XY = 42558.6, X2 = 76129.9,
Y2
= 23892,
= 85.4,
= 47.8.
Then:
b
and
For:
42558.6 10 85.4 47.8
76129.9 10 85.42
1737.4
0.543
3197.3
47.8 0.543 85.4 1.43
42558.6 40821.2
76129.9 72931.6
Y = 1.43 + 0.543X
X
131.2 : Y
90.6 : Y
1.43 0.543(131.2)
1.43 0.543(90.6)
72.7
50.6
Therefore:
Crime rate = 131.2
Crime rate = 90.6
72.7 patients
50.6 patients
Note that rounding differences occur when solving with Excel.
BUS P301:01
