Dynamic Stacking Optimization in Unpredictable Environments: A Focus on Crane Scheduling

Dynamic Stacking Optimization holds significant practical relevance for multimillion-dollar industries involved in production and delivery processes. It entails the utilization of cranes to relocate products, with the relocation needing to be scheduled while adhering to various time constraints. This paper addresses the challenge of developing solution approaches for such dynamic stacking problems in uncertain environments, particularly the environment represented by the “CraneScheduling" simulation of the DynStack Competition 2023, which is part of the Genetic and Evolutionary Computation Conference. Leveraging systematic and task-oriented solutions enabled by industrial digitalization, our work focuses on optimal crane scheduling to prevent delivery errors, maximize block handling efficiency, and minimize truck queues and shipment delays. The dynamic stacking problem involves two cranes operating on the same girder, multiple arrival and handover stacks, and various stacks in the buffer area. Our approach employs a role-based solver, efficiently planning crane assignments to manage container movements within the terminal. The performance of the solvers is compared with other solvers submitted in the competition, building on existing solution approaches. The role-based solver leads to reduced unnecessary movements by efficiently planning crane assignments. However, it shows potential for significant improvement, offering benefits for both the industry and the environment.