ABSTRACT
The choice of which forecasting method is most appropriate when linked to an inventory control system is predominated by the likely number of different forecasts that have to be made on a routine basis. For many inventory systems, this could be several thousands. Hence, a high degree of sophistication is clearly not required and a family of forecasting models which are cheap to operate but which provide reasonably accurate forecasts is to be preferred. The exponentially weighted average family of forecasting models is one group that fits this criterion most satisfactorily and since its first exposure by Holt3 has been a popular forecasting model selection when linked to inventory control. This chapter details the simple exponentially weighted average forecasting model and two of its derivatives and contrasts their features with the more commonly known moving average and then considers for which type of demand data this model is most suited.
