Title: A serial inventory system with leadtime dependent demands: A reduced state approximation by
Emre Berk, Faculty of Business Administration
Friday 13:40, December 2
Customer response to stockouts has been traditionally modeled as either full backordering or lost sales regardless of the length of wait that a customer may experience until an order is received by the retailer. There is convincing evidence that customers do indeed consider such waits (the remaining lead times) when they make purchase decisions during a stockout.
In this talk, we consider an (S-1,S) serial inventory system facing such lead time dependent backordering decisions. In such a system, the purchase decisions depend on the waiting tolerance distribution of the customer population and the state of the system, which may be described through the times for a given number of most recent orders. We model the system via the method of supplementary variables, establish certain theoretical properties, and propose a reduced state approximation that utilizes only the age of the oldest item in the system. In an extensive numerical study, we demonstrate the efficacy of the proposed approximation in terms of its quality and the goodness of the optimal solution thereby obtained. We also develop some conjectures that are supported by our numerical results and are consistent with the full backordering counterpart.