Dynamic Inventory Control with Satisfaction-Dependent Demand

In this paper, we consider the discrete multiperiod newsvendor dynamic inventory control problem where customers follow a simple satisfaction-based demand process, where their probability of demand depends on whether their demand was satised the last time they demanded a product, and observe the differences between optimal policies and myopic policies which do not directly consider how inventory policies can affect future demand. We conrm the intuitive result that inventory managers should tend to order more than the myopic policy when satised customers are more likely to demand product, and less than the myopic policy when satised customers are less likely to demand. Moreover, we and that, when choosing a fixed order policy, even an empirically myopic solution with perfect demand distribution information will move away from the optimum towards a suboptimal solution.