A precedence-based monolithic approach to lot-sizing and scheduling of multiproduct batch plants

Batch scheduling is a highly combinatorial problem involving two major components: the lot-sizing or batching problem (P1) defining the set of batches to be scheduled, and the "pure" short-term batch scheduling problem (P2) assigning resources to batches and sequencing batches at every resource item. Due to the large computational requirements to cope with the whole problem at once, precedence-based optimization strategies have traditionally solved subproblems P1-P2 in a sequential manner. In contrast, this work presents an effective precedence-based approach that integrates both subproblems into a unique MILP formulation and solves the problem in a single step. A pair of examples involving the scheduling of multistage, multiproduct batch facilities carrying out linear processes have been solved. Comparison of the results found with the ones reported by other authors leads to conclude that the proposed approach shows a much better computational perfomance.