Retailer’s inventory management problem is researched, where both the demand and the lifecycle of the product
are stochastic. Demand is classified according to purchasing behavior and an inventory allocation policy is proposed among
different types of customers, where backorder is allowed. Through dynamic programming, the optimal inventory policy is
derived for the retailer, which consists of replenishment policy as well as inventory allocation policy. Compared to first-
come-first-served allocation policy, the proposed policy can improve the retailer’s profit significantly and reduce the loss
caused by product deterioration.