Value of capacity flexibility in manufacturing systems with seasonal demands

Abstract

In this article, we analyze a manufacturing system subject to seasonal demands. The system is modeled as a single-stage production facility with flexible production rate, and seasonal demands are modeled using a Markov-modulated Poisson process. Using dynamic programming, we develop optimal policies under the infinite-horizon discounted expected cost and the infinite-horizon average expected cost criteria. We prove that the optimal policy is a season-dependent base-stock policy with state-dependent production rates. Furthermore, we identify the monotonic structure of the optimal policy with respect to the net inventory, the demand in a season, and the overall workload on the system. We, then, illustrate the value of joint flexibility in capacity and inventory levels under the seasonal demand settings. Numerical comparisons with the policies used in practice demonstrate the value of capacity flexibility.