Problems in Forecasting

1. Integrated Products Corporation (IPC) needs to estimate its sales for next year. The
most recent six years of revenue data for the company’s line of XT computers are
found in the table below:

Sales Revenues Sales Revenues

Year (millions of dollars) Year (millions of dollars)
1   2.4 4 27.8
2   5.9 5 35.9
3 15.5 6 38.1

a. Assuming that the sales data above are representative of sales expected next year,
use time series regression analysis to forecast next year’s (Year 7’s) sales revenues.
b. Determine the correlation coefficient for the data and interpret its meaning.
c. Find the coefficient of determination for the data and interpret its meaning.

Seasonalized Time - Series

2. Chemco, a firm that provides in-transit warehouse space to major chemical

manufacturers, has a facility at Evansbrook, Illinois. This warehouse stores strontium
nitrate for three firms that ship from the Chemco warehouse to local distributors. To
double its capacity if it needs it, Chemco can lease another nearby warehouse that has
recently become available, but the decision must be made now. Chemco will lease the
warehouse for one year if at any time during the year the maximum quarterly inventory is
expected to exceed the present warehouse capacity, which is 28.5 million pounds. The
recent past inventory levels are:
════════════════════════════════════════════════════════════════════════

Inventory Inventory Inventory

(Millions (Millions (Millions

Year Quarter of Pounds) Year Quarter of Pounds) Year Quarter of Pounds)

───────────────────────────────────────────────────────────────────────────

1 1 10 2 1 12 3 1 14

1 2 8 2 2 10 3 2 14

1 3 12 2 3 16 3 3 18

1 4 14 2 4 20 3 4 22

═══════════════════════════════════════════════════════════════════════════

a. Use seasonalized time series analysis to forecast inventory levels for Q1, Q2, Q3,
and Q4 for next year.

b. Assume that when management refers to the maximum quarterly inventory levels,
what it means is the upper limit of the quarterly forecasts with a probability of
 only .10 of exceeding the limit due to random variation. Find the upper limit or
maximum quarterly inventory levels for next year.

c. Should Chemco lease the warehouse?

Moving averages and exponential smoothing

3. The Sporting Charge Company buys large quantities of copper that is used in its
manufactured products. Bill Bray is developing a forecasting system for copper prices. He
has accumulated this historical data:

Copper Copper
Price/ Price/

Month Pound Month Pound

1 $0.99$  9 $0.98$

2 0.97 10 0.91

3 0.92 11 0.89

4 0.96 12 0.94

5 0.93 13 0.99

6 0.97 14 0.95

7 0.95 15 0.92
8 0.94 16 0.97

a. Use exponential smoothing to forecast monthly copper prices. Compute what the
forecasts would have been for all the months of historical data for  = 0.1,  =
0.3, and  = 0.5 if the forecast for all ’s in the first month was $0.99.

b. Which alpha () value results in the least mean absolute deviation over the 16-
month period?

c. Use the alpha () from Part b to compute the forecasted copper price for Month
17

Bill Bray wishes to compare two forecasting systems to forecast copper prices from the data
in Problem above: moving averages (AP = 3) and exponential smoothing ( =
0.3).

a. Compute the two sets of monthly forecasts over the past 10 months (7 through
16). The exponential smoothing forecast in Month 6 was $0.954.

b. Select the best system and forecast the copper prices for the next month (Month
17).
 Exponential Smoothing

4. Use the data in Problem 1 above to develop a forecast for Year 7 using the exponential
smoothing with trend model. Begin your analysis in Year 1 and assume that  = 0.3 and  =
0.2. Estimate FT1 and T1

Regression Analysis
5. A small regional trucking company has experienced steady growth. Use time series
regression to forecast capital needs for the next 2 years. The company's recent capital
needs have been:
══════════════════════════════════════════════

Capital Needs Capital Needs

(Thousands Of (Thousands Of

Year Dollars) Year Dollars)

------------------------------------------- --------------------------------------------

1 100 3 130

2 110 4 140

5 160

══════════════════════════════════════════════

Moving Averages
6. Jane Montgomery is the manager in charge of maintenance for Rest-International, a
large hotel in Las Vegas. She has observed that since the hotel is fairly new and is fully
booked year-round, maintenance calls occur randomly, with almost no trend or
seasonality. She is developing a system to forecast one month ahead the number of
maintenance calls she will receive. She plans to use moving averages but wonders what
AP to use in order to minimize the forecasting error. Two years of historical data are as
follows:

═════════════════════════════════════════════════

Maintenance Maintenance Maintenance

Month Calls Month Calls Month Calls

───────────── ────────────── ────────────────

1 95 9 97 17 102

2 85 10 101 18 89

3 92 11 82 19 82
 4 100 12 92 20 102

5 80 13 97 21 100

6 91 14 95 22 101

7 89 15 93 23 95

8 84 16 105 24 90

═════════════════════════════════════════════════

a. Develop moving average forecasts for the past 10 months (Months 15 - 24) for AP = 2, 4,

6, and 8 months.

b. Which AP results in the lowest mean absolute forecasting error?

c. Use your recommended AP and forecast the number of maintenance calls for the
next month (Month 25).

Least Square Method

Read the article and the example. Find out any problem from any text book on
Statistics/Forecastic/Demand Management and solve it.