Exploiting elapsed time for managing intermittent demand for spare parts

We present an intermittent demand forecasting method that conditions on the elapsed time since the last demand occurrence to anticipate incoming demand and show, using empirical data, that this can substantially reduce both stock investment and lost revenue for spare parts management. We extensively benchmark our method against existing forecasting and bootstrapping methods on forecast accuracy and inventory performance and demonstrate that its performance is robust under general conditions. Our method is the first to incorporate that activities at the demand side, such as aggregation of demand, preventive and corrective maintenance, can lead to a positive relation between demand size and inter-arrival time of demand occurrences. By anticipating incoming demand, our method offers substantial financial gains.