Analysis of admission and inventory control policies for make-to-stock production networks

This article is concerned with the control of inventories and sales in a manufacturing system that produces one product to meet random demand. The objective is to specify critical stock levels and either maximum backlog levels or backordering probabilities so as to maximize the mean profit rate of the system. Four policies are investigated: lost sales, complete backordering, randomized admission, and partly lost sales. The optimal parameters for each policy are determined using queueing theory and convex analysis. Numerical results show that determining the base stock and the base backlog jointly achieves higher profit than the other control policies.