Managing Hospital Platelets Inventory with Mid-Cycle Returns and Refill

Motivated by the need of small and medium-sized hospitals to improve their platelet inventory management, we develop a finite-horizon dynamic programming model to optimize inventory control for perishable product (e.g., platelet) with 3-period shelf-life. In addition to identifying the best cycle order quantity at the beginning of each cycle, we propose an optimal inventory-adjustment policy involving returns and refill at the mid-cycle, which is characterized by two thresholds and offers three guidelines: (i) if the mid-cycle inventory is above the upper threshold, then return the inventory to the upper threshold level; (ii) if the inventory is below the lower threshold, then restock the inventory to reach the lower threshold; otherwise, (iii) no action is needed. We derive optimal closed-form solutions and show that the cycle order depends on a pre-specified threshold, a state-independent target, and a state-dependent dynamic target. Namely, if the initial inventory level is below the pre-specified threshold, the inventory after cycle order will be raised to the state-independent target; otherwise, it will be state-dependent. When using fresh platelets in the expedited order, or facing an even more complex scenario with shelf-life longer than 3-period, we identify their structural properties. These properties provide insights for making optimal inventory control decisions and for understanding the economic substitution relationship between the replenished quantity and the existing inventories of various ages. Our model outperforms those in the literature as it is more responsive and reacts more quickly to signals of imminent stockout or pending obsolescence