Real-Time Perishable Inventory Management with Mid-Cycle Returns and Reordering

Kebing Chen,Jing-Sheng Jeannette Song,Jennifer Shang,Tiaojun Xiao

DOI: https://doi.org/10.2139/ssrn.3032926

2017-01-01

Abstract:Motivated by the need of small and medium-sized hospitals to improve their platelet inventory management and some best practices, we consider an inventory system for a perishable product with a 3-period shelf-life over a finite horizon, in which the regular replenishment of each cycle is supplemented by an inventory adjustment opportunity in mid-cycle involving expedited orders and returns. We show that optimal regular replenishment policy is an (s, S, S(x)) policy: If the initial net inventory x is below the threshold s, it is optimal to place a regular order to raise the inventory to a state-independent target S; otherwise, raise the inventory to a state-dependent target S(x). We also show that the mid-cycle inventory adjustment policy is a control-band policy with two thresholds. If the mid-cycle net inventory is above the upper threshold, it is optimal to return the inventory in excess to that threshold to a central bank; if the mid-cycle net inventory is below the lower threshold, it is optimal to expedite replenishment up to that threshold; otherwise, do nothing. If non-fresh platelets are used in the expedited order, we obtain closed-form expressions for the optimal policy parameters. If fresh platelets are used for the expedited order, we apply antimultimodularity techniques to establish structural properties of the optimal policy to guide practice and reveal the substitution relationship between the replenishment quantity and the existing inventory of various ages. These structural properties also hold for the case of general shelf lives. Our numerical study shows that our more flexible policy outperforms those studied in the literature.