A Two-phase Heuristic for Crane Scheduling in Steel Slab Yards

In this paper, a two-phase heuristic algorithm is proposed for an expanded crane scheduling problem that combines a Resource Constraint Project Problem and a Block World Problem in order to develop a guidance system for minimizing crane handlings in slab yards of steel works. Due to its NP hardness, it is difficult to obtain the theoretically optimal solution of the problem. Therefore, a two-phase heuristic algorithm was developed to obtain an approximate solution in a practical calculation time. The first phase in this approach utilizes a genetic algorithm that is employed to solve a relaxed scheduling problem of rearranging steel slabs in an approximate manner. Next the partial solution is iterated upon by a rule-based algorithm to obtain a feasible solution. Computational experiments are conducted with operation data of JFE Steel, allowing a comparison to be made between actual and theoretical crane handling operations. The resulting data shows that this paper's proposal can reduce the number of handlings by 30%.