What happens to the ability to forecast for periods farther into the future? CEO John Goodale, at Southern Illinois Power and Light, has been collecting data on demand for electric power in its western subregion for only the past 2 years. Those data are shown in the table below. To plan for expansion and to arrange to borrow power from neighboring utilities during peak periods, Goodale needs to be able to forecast demand for each month next year. However, the standard forecasting models discussed in this chapter will not fit the data observed for the 2 years.
|
Demand in Megawatts
|
|
Month
|
Last Year
|
This year
|
|
January
|
5
|
17
|
|
February
|
6
|
14
|
|
March
|
10
|
20
|
|
April
|
13
|
23
|
|
May
|
18
|
30
|
|
June
|
15
|
38
|
|
July
|
23
|
44
|
|
August
|
26
|
41
|
|
September
|
21
|
33
|
|
October
|
15
|
23
|
|
November
|
12
|
26
|
|
December
|
14
|
17
|
(a) What are the weaknesses of the standard forecasting techniques as applied to this set of data?
(b) Because known models are not appropriate here, propose your own approach to forecasting. Although there is no perfect solution to tackling data such as these (in other words, there are no 100% right or wrong answers), justify your model.
(c) Forecast demand for each month next year using the model you propose.