Optimization of wood transportation

The planning of wood supply to pulp and saw mills at Stora Enso in Finland involves multiple optimization models, of which we discuss the short-term transportation models. The inputs to the optimization models are harvest plans on the supply side and mill production plans on the demand side. The overall goal is to transport thousands of batches of timber to dozens of mills as cost-effectively as possible. The planning proceeds in stages: First, harvested batches of timber are combined into units of truckloads. Second, the truckloads are assigned to truck, rail or water transport to mills. Finally, truck routes are planned and scheduled in order to arrange the deliveries of the truckloads. The three optimization problems are combinatorial in nature and are addressed by heuristic tabu search and mixed-integer linear programming. We give an overview of the models and our solution approaches, and discuss experiences from production use.