Forecasting and Stock Control for Intermittent Demands

Abstract

Exponential smoothing is frequently used for the forecasts in stock control systems. The analysis given shows that intermittent demands almost always produce inappropriate stock levels. Demand for constant quantities at fixed intervals may generate stock levels of up to double the quantity really needed. A method of overcoming these difficulties is described, using separate estimates of the size of demand, and of the demand frequency. The rules for setting the safety stock levels have also to be adjusted before consistent protection can be obtained against being out of stock.