Applying Constraint Programming to the Multi-mode Scheduling Problem in Harvest Logistics

In this paper, we present a Constraint Programming (CP) based model for scheduling forage harvesters and transport vehicles during corn harvest. The key aspects are the synchronization of the two resource types and a forage harvester utilization depending on the number of transport vehicles supporting the harvester. The process is modelled as a pre-emptive multi-mode resource-constraint project scheduling problem with fast-tracking, sequence-dependent time lags and synchronization. We use the specialized scheduling features of CP Optimizer for modelling and solving the harvest logistics problem. The results show the suitability of the CP-based approach for modelling the problem in terms of representability of its characteristics. In computational experiments, a solution is found for any of the test instances. Proving optimality, however, is found to be difficult, especially for larger instances. Further variants of the model without pre-emption and fast-tracking and with fewer modes per activity are introduced and tested, showing improvements in computation time and the number of optimal solutions found for the prior variant.