Blood platelet bank inventory management: an approximate dynamic programming approach

Abstract

Blood platelets are precious and highly perishable; their supply and demand suffer from significant variation. Consequently, the inventory management of platelets is an actual, contemporary prob- lem of considerable human interest. Although many researchers have solved a plethora of inventory models, their solutions have faced various challenges. This dissertation models some of these chal- lenges, alongside expenses and stock levels. This dissertation is based on four key objectives: (1) to develop a blood platelet inventory model that can represent an actual blood bank inventory, while overcoming the problem's curse of dimensionality; (2) to look for the best issuing policy based on the proposed model that can serve different incoming blood platelet demands; (3) to analyze the effect of having a new, artificial blood platelet alongside the existing natural eight blood types; and (4) to enhance the proposed model for a dual-supplied regional blood platelet bank that serves a network of hospitals. Blood platelet inventory management model is a multi-period, multi-product model that considers the eight natural blood types with uncertain demand, and deterministic lead times, alongside the artificial platelet and patients right to refuse it. The study is supported by both a review of literature and a testing data provided by the Canadian Blood Service. The findings show that modeling blood platelet inventory management, including the eight blood types and their ages, represents the actual-life model without any need for downsizing. It also leads to significantly reductions in shortages and outdates while increasing reward gained and maintaining minimal inventory levels. Compared to a single supply model, the dual supply model give less shortage and outdate rates. The regional blood bank inventory model considers the fact that patients have the right to refuse transfusion using artificial blood platelets. Finally, if the percentage of artificial supply in the inventory is more than 30% and the rate of patient acceptance is more than 30%, then both outdate and shortage percentages are below 1%.