Aggregated scheduling of a multiproduct batch plant by two-stage stochastic integer programming

We consider the mid-term scheduling of a multiproduct batch plant. Market requirements for the final products are uncertain but given in terms of a probability distribution. We introduce a two-stage stochastic integer programming model for the problem and solve it by a scenario decomposition method based on Lagrangian relaxation. We describe heuristics and preprocessing for both the single- and multi-scenario models. Preliminary computational results are presented.