As the assistant production manager for Acme Widget Company you are given the task of forecasting the next month's sales. Based on your forecast decisions will be made regarding staffing, purchasing, marketing and a host of related activities. You have the following record of the previous 30 months of sales:
Month
Sales
Month
Sales
1
11
16
25
2
13
17
24
3
14
18
26
4
15
19
27
5
13
20
24
6
15
21
22
7
17
22
27
8
18
23
32
9
19
24
31
10
15
25
30
11
17
26
28
12
20
27
25
13
19
28
23
14
23
29
27
15
22
30
29
You decide to examine several mathematical models to determine the best next-period forecast. They include:
The naive method
A 4-period moving average
A 3-period weighted moving average, with a weight of 3 for the most recent period, 2 for the next most recent period and 1 for the least recent period
Exponential smoothing, with a smoothing constant of 0.1
Exponential smoothing, with a smoothing constant of 0.5
A linear trend
1) What is the period 31 forecast for each of the above models?
2) What are the period 32 and period 33 forecasts for each of the above models?
3) If you were only interested in choosing the most accurate model, which would you choose, and why?
4) If you had a low tolerance for large deviations which model would you choose, and why?
Suppose you made the decision to choose the most accurate model, without being averse to large deviations. You continued to use the model to forecast monthly sales over the next 10 months. The actual sales were as follows:
Month
Sales
Month
Sales
31
30
36
25
32
37
37
29
33
32
38
28
34
27
39
29
35
26
40
30
5) Develop a control chart for the errors of the chosen model and state the conclusions of your analysis.
