Forecasting Mean Demand

This chapter focuses on the issue of forecasting mean demand, leaving the forecasting of variance. It discusses the performance of various forecasting methods for intermittent demand, and examines how intermittent demand data is structured. Every practical forecasting application starts with an examination of the real data available. The chapter discusses the implications of modelling assumptions for intermittent demand forecasting. Demand models are important because they help us to conceptualise the underlying structure of the data and examine in detail the performance of forecasting methods. Single exponential smoothing (SES) is very often used in practice to forecast intermittent demand requirements. John Croston showed that SES is biased after a demand occurring period. According to Croston's method, separate exponential smoothing estimates are made of the average size of the demand and the average interval between demand occurrences. Croston's method is implemented in many software packages.