Investigating Several Alternatives for Estimating the Compound Lead Time Demandin an (s,Q) Inventory Model

摘要

Using Monte Carlo investigations, the paper analyzes the cost differences betweenseveral alternative approximations for the lead time demand distribution (LTDD) in a continuous review (s,Q) inventory model. The information on LTDD is assumed to be built up by two components: demand per time unit and lead time. Enumeration methods, simulation and parametric approaches are used to obtain compound information on LTDD with help of the above individual components. Three important conclusions result: The simulation approach is simple and still able to take into account certain peculiarities in the lead time distribution in the most proper way; a lack of lead time information should be avoided as much as possible by a good information system. It is shown that extending the lead time information leads to drastical cost reductions in the inventory model; and the gamma distribution appears to be a good approximation for the LTDD in many cases.