April 6: Adam J. Mersereau (Kenan-Flagler Business School, The University of North Carolina at Chapel Hill)
Monday, April 6, 2009
120 Hanes Hall
Information-Sensitive Replenishment when Inventory Records are Inaccurate
Recent empirical work reveals a reality of inventory management in practice: inventory records do not necessarily match the physical inventory on the shelf. We consider a periodic review inventory system with imperfect inventory records and unobserved lost sales. Record inaccuracies are assumed to arrive via an “invisible” demand process that perturbs physical inventory but is unobserved by the inventory manager. The inventory manager does not know the true inventory level but maintains a probability distribution around it that he updates based on sales observations using Bayes Theorem.
Our focus is on characterizing and approximating the optimal replenishment policy in this lost sales environment. By analyzing one- and two-period versions of the problem, we first identify several ways in which record inaccuracy and invisible demand impact optimal replenishment. Although record inaccuracy generally brings an incentive for a myopic manager to increase stock to buffer the added uncertainty, a forward-looking manager will stock less (or no more) than a myopic manager, in part to improve information content for future decisions. We present an approximate partially observed dynamic programming algorithm which yields cost lower bounds and forward-looking policies for longer-horizon problems. A numerical study corroborates our analytical findings from the short-horizon cases, but suggests that inaccuracy-sensitive myopic replenishment policies may be sufficiently good in practical retail settings targeting high service levels.
From 3:30 to 4:00PM the audience is invited for refreshments in the lobby on the 3rd floor of Hanes Hall.