Forecasting Computer Usage

Figures & data

Figure 1 MIPS vs Time.

Figure 2 MIPS vs Time.

Table 1. In-class forecasting example.

Time Period	Demand	Forecast, 3-Period Moving Average	Forecast, Exponential Smoothed In-Class Forecasting Example α = 0.2	Deviation, 3-Period Moving Average	Deviation, Exponential Smoothed	Absolute Deviation, 3-Period Moving Average	Absolute Deviation, Exponential Smoothed
1	4.1	4.10
2	3.3	4.10
3	4.0	3.94
4	3.8	3.80 = (4.1 + 3.3 + 4.0) / 3	3.95= (0.2*4.0)+ (0.8*3.94)	0.00 = 3.80 − 3.80	0.15 = 3.95 − 3.80	0.00 = |0.00|	0.15 = |0.15|
5	3.9	3.70	3.92	−0.20	0.02	0.20	0.02
6	3.4	3.90	3.92	0.50	0.52	0.50	0.52
7	3.5	3.70	3.81	0.20	0.31	0.20	0.31
8	3.7	3.60	3.75	−0.10	0.05	0.10	0.05
			CFE	0.40 = 0.00 − 0.20 + 0.50 + 0.20 − 0.10	1.06
			MAD			0.20 = (0.00 + 0.20 + 0.50 + 0.20 + 0.10)/5	0.21

Figure 3 Actual and forecast MIPS.

Figure 4 Optimal double exponential smoothed.

Figure 5 Double exponential smoothed model errors.

Figure 6 Winter's Method.

Figure 7 MIPS vs number of stores.

Figure 8 MIPS vs number of stores.

Figure 9 MIPS vs number of stores residual plot.

Figure 10 Square root MIPS vs number of stores.

Figure 11 Square root MIPS vs number of stores residual plot.

Table

Columns	Description
1 − 11	Date dd-mm-yyyy, August 1996 – July 2000
12 − 19	MIPS usage (MIPS are a measure of computing resources)
20 − 30	Number of stores
31 − 44	Date dd-mm-yyyy, July 2000 – December 2001
45 − 53	Planned Number of Stores through December 2001