Case Summary
A local coffee shop delivers their coffee in urns twice each morning so the coffee is kept hot. Connie Andrews, the office manager, likes this practice and wants to continue it for her customers, but is concerned about coffee being wasted. Analysis shows that over the long term, there are no seasonal patterns or trend, but in the short term, there appears to be periods where demand fluctuates.
1.
a. Exponential smoothing puts more weight in more recent historical periods, therefore a period two (2) months ago has more weighting than a period six (6) months ago. With the alpha at 0.1 forecasts follow demand less closely because a low alpha means historical data is more relevant to the forecast.
When the alpha is increased from 0.1 to 0.3, the forecast follows demand more closely. This is because a higher alpha, results in fewer historical periods used to create the forecast. For example, a higher alpha givers higher weighting to recent periods and demand from periods for a year are weighted so lightly that they have little to no effect on the overall forecast. This explains the response of the forecast in periods 3 and 8 with the change in alpha because the higher the alpha the more of a response to demand spikes.
b. When increasing the demand at period nine (9), the forecast for period 10 increases to 67. This shows that the forecast still follows demand but at a smaller increment. When increasing the alpha (keeping demand at 70), the immediate forecast becomes more responsive to the changing demand.
c. When reducing the demand at period seven (7), the forecast for period eight (8), it is once again apparent that the greater the alpha, the more responsive the immediate forecast is to the changing demand.
2.
a. With alpha at 0.1, the MFE = 48.6401 and the MAD = 11.05. Mean Forecast Error (MFE) is a measure of forecast model bias and Mean Absolute Deviation (MAD) indicates the absolute size of the errors. The ideal...