Controlling make-to-stock manufacturing systems: a large deviations approach

We consider a model of a multiclass make-to-stock manufacturing system. External demand for each product class is met from the available finished goods inventory; unsatisfied demand is backlogged. The objective is to devise a production policy that minimizes inventory costs subject to guaranteeing stockout probabilities to stay bounded above by given constants ε_j, for each product class j (service level guarantees). Approximating the original system, we analyze a corresponding fluid model to take sequencing decisions and employ large deviation techniques to take idling ones, under various inventory cost structures. Our model is more realistic than most in the literature, since it can accommodate autocorrelated demand and service processes, both critical features of modern failure-prone manufacturing systems