Efficiency of past stock movement simulation in intermittent demand stock control

Abstract

In this paper we present the efficiency measurement study of past stock movement simulation designed for the stock management of products with intermittent demand. The proposed simulation is able to provide the combination of controlled parameters (reorder stock level + order quantity/order-up-to level) included in two frequently used stock management policies (Q-system, PQ-system) which guarantees the minimal stock holding and ordering costs and the required fill rate. To obtain reorder stock level single exponential smoothing, Croston’s method, the modification of Croston’s method made up by Syntetos&Boylan and Levén&Segerstedt, the method of Smart and the total enumeration are considered. When tested on 5730 real intermittent demand series from automotive industry we conclude that the total enumeration used in the past stock movement simulation to calculate reorder stock level leads to the lowest stock holding and ordering costs and represents robust approach in intermittent demand stock control in term of increasing intermittence and lumpiness.