Intermittent demand forecasting: An empirical study on accuracy and the risk of obsolescence

摘要: Intermittent demand items account collectively for considerable proportions of the total stock value any organization. Forecasting relevant inventory requirements constitutes a very difficult task and most work in this area is based on Croston’s estimator that relies upon exponentially smoothed sizes inter-demand intervals. This method has been shown to be biased number variants have introduced literature, including recently proposed TSB updates probability instead interval doing so reacts faster decreasing demand. The theoretically unbiased (for all points time), but its empirical performance not investigated yet one objectives our work. More generally, we explore forecasting methods used an intermittent context, paying particular attention effects implications smoothing constant values employed updating purposes. We do by means experimentation large datasets from military sector automotive industry. results enable insights gained into sensitivity various methods’ control used. paper concludes with agenda further research.