CP and Hybrid Models for Two-Stage Batching and Scheduling

Abstract

Batch scheduling is a common problem faced in industrial scheduling when groups of related jobs must be processed consecutively or simultaneously on the same resource. Motivated by the composites manufacturing industry, we present a complex batch scheduling problem combining two-stage bin packing with hybrid flowshop scheduling. We propose five solution approaches: a constraint programming model, a three-phase logic-based Benders decomposition model, an earliest due date heuristic, and two hybrid heuristic/constraint programming approaches. We then computationally test these approaches on generated problem instances modelled on real-world instances. Numeric results show that the heuristic approaches perform as well as or better than the exact models, especially on large instances. The relative success of a simple heuristic suggests that such problems pose an interesting challenge for further research in mathematical and constraint programming.