Optimization-based planning for the stochastic lot-scheduling problem

We describe a finite-horizon stochastic optimization model for the stochastic lot-scheduling problem and procedures for finding near-optimal solutions. Several different products are produced by a single-stage process with significant changeover times and costs, and the demand for these products is random. The deterministic version of this problem, the economic lot-scheduling problem, is the subject of a great deal of research. However, the problem with random demand for the products is commonly found in practice but is not as well researched. The models developed in this paper address the problem of dynamically planning the timing and size of production runs in this kind of production environment. We also report some computational results that indicate the quality of the resulting production schedules.