Generating order policies by SDP: non-stationary demand and service level constraints

Abstract

Inventory control implies dynamic decision making. Therefore, dynamic programming seems an appropriate approach to look for order policies. For finite horizon planning, the implementation of service level constraints provides a big challenge. This paper illustrates with small instances the implementation of stochastic dynamic programming (SDP) to derive order policies in a straightforward way for systems with non-stationary demand and service level constraints. The small instances allow to perform a full enumeration of possible policies and show that the SDP derived policies are not necessarily optimal.