Heuristics and sensitivity analyses to guide replenishment decisions for red blood cells with random transfer

作者:

Highlights:

• We introduce two near-optimal heuristics to address the fixed lifetime perishable inventory problem with random transfers.

• The first heuristic (H1) employs a base stock policy using the convolution of demand and transfer distributions. The second heuristic (H2) employs a transfer adjusted base stock policy.

• Our heuristics are straightforward to calculate and implement and perform quite well.

• We validate our heuristics using inventory-related cost components (i.e., order, holding, wastage, and underage costs) and demand and transfer distributions based on data provided by a large university medical center.

• We conduct a variety of sensitivity analyses considering the effects of different (1) levels of holding and stockout costs, as these costs are difficult to quantify and can vary dramatically, and (2) percentile transfer adjustments in H2.

• We discuss a number of management challenges encountered as part of this project and offer a five-step process to help decision makers with blood replenishment decisions.

摘要

While the transfer of blood is common in practice, research addressing replenishment strategies in the context of random or unscheduled transfers is scarce and the few papers that consider ordering policies with multiple independent sources of supply offer mathematically complex solution approaches that are difficult to implement in practice. There are no simple heuristic approaches to guide ordering decisions in these situations. In this paper, we introduce two near-optimal heuristics to guide replenishment decisions for the fixed lifetime perishable inventory problem with random transfers. Much like the traditional base stock policy, the heuristics are straightforward to calculate and implement. The first heuristic (H1) employs a base stock policy using the convolution of demand and transfer distributions. The second heuristic (H2) employs a transfer adjusted base stock policy. We validate our heuristics using inventory-related cost components (i.e., order, holding, wastage, and underage costs) and demand and transfer distributions based on data provided by a large university medical center. Via simulation, we evaluate the performance of the heuristics by comparing the total costs obtained using each heuristic to the total costs obtained using the optimal ordering policy. Both heuristics perform quite well. We conduct a variety of sensitivity analyses considering the effects of different (1) levels of holding and stockout costs, as these costs are difficult to quantify and can vary dramatically, and (2) percentile transfer adjustments in H2. We observe that the H1 policy consistently results in slightly larger orders than the optimal policy, which leads to a slightly higher inventory, holding costs and wastage rates, yet slightly lower stockout costs. The H2 policy, on the other hand, consistently results in slightly smaller orders than the optimal policy, which leads to a slightly smaller inventory, lower holding costs and wastage rates, yet slightly higher stockout costs. To guide decision making, we discuss a number of management challenges that we encountered as part of this project and offer a five-step process to help decision makers with replenishment decisions.

论文关键词:Perishable inventory,Blood management,Health care operations,Supply chain,Transfers,Heuristic

论文评审过程:Received 29 January 2021, Revised 9 August 2021, Accepted 12 October 2021, Available online 24 October 2021, Version of Record 24 January 2022.

论文官网地址:https://doi.org/10.1016/j.dss.2021.113685