On the Evaluation of a Class of Inventory Policies for Perishable Products Such as Blood

A class of inventory and distribution policies for scheduled deliveries of perishable products subject to variable demand is modeled as a Markov chain with a manageable number of states. From its stationary distribution, such measures as probability of shortage, the average age of the inventory, and the average number discarded per time period are obtained. Easily computable bounds on these measures that indicate how they behave as functions of demand, and inventory policy parameters are developed.