The effects of forecasting errors on the total cost of operations

The industrial customer faces the need to forecast utility demand. This paper demonstrates how the utility can help the industrial customer forecast more accurately and thus reduce costs. Using the actual business conditions of a public utility, an extensive examination of traditional time series forecasting techniques under various simulated or demand patterns has been carried out. In this paper, traditional forecasting error measures such as bias, mean absolute deviation (MAD) and mean squared error (MSE) have been analysed in correlation with a much more relevant error measure for a business environment: the total cost of a time series model relative to the organization. Finally, some previously held assumptions of autocorrelation and 'model fit' are examined.