One day ahead demand forecasting in the utility industries: Two case studies

Abstract

This paper describes two case studies of short-term demand forecasting for the utilities, water and gas, linked to earlier research in similar contexts. In both cases the forecast of demand has important consequences for the operations and control of productive capacity. It is shown that in these two cases extrapolative methods based on the past data history alone are outperformed by more complex multivariate approaches that include information on the effects of weather. The paper concludes with a discussion of how an organization with an important short-term forecasting problem should go about selecting an appropriate forecasting method.