Modelling Spare Parts' Demand: an Empirical Investigation

Intermittent demand patterns are very difficult to forecast and they are, most commonly, associated with spare parts’ requirements. Most research in this area focuses on the control of inventories assuming that an appropriate estimator is in place to forecast demand and that the demand is represented by one of the standard statistical distributions. The choice of a demand distribution has an important impact on the stock control performance. In this paper, after a critical review of the literature dealing with the demand distributions considered for stock control of intermittent demand, an empirical investigation is conducted, using a real data set of approximately 13,000 Stock Keeping Units (SKUs) coming from the military sector (Royal Air Force, RAF UK and US Defense Logistics Agency, DLA) and the Electronics industry. The empirical investigation aims at testing statistical significance of various theoretical distributions. The potential linkage between demand distributional assumptions and demand classification schemes is also explored followed by an agenda for further research in this area.