Dynamic transport-lot assignment for the hot-storage area

In this paper, we present a novel transport-lot optimization problem at the boundary between the continuous casters and the hot-rolling in steel production. The problem and the corresponding solver are part of a system that jointly optimizes the cranes in the hot-storage area and the vehicles feeding the rolling mill. The goal is to group steel slabs into transport lots and assign transporters, handover locations, and due dates.
We develop heuristics to iteratively build the full solution by choosing from several promising choices that satisfy a wide variety of safety and performance requirements.
We evaluate the performance of different heuristics to search the tree of possible choices on a diverse collection of real-world problem instances. We find that the pilot method significantly outperforms the five other tree-search heuristics we tried.