A hybrid method based on linear programming and tabu search for routing of logging trucks

In this paper, we consider an operational routing problem to decide the daily routes of logging trucks in forestry. The industrial problem is difficult and includes aspects such as pickup and delivery with split pickups, multiple products, time windows, several time periods, multiple depots, driver changes and a heterogeneous truck fleet. In addition, the problem size is large and the solution time limited. We describe a two-phase solution approach which transforms the problem into a vehicle routing problem with time windows. In the first phase, we solve an LP problem in order to find a destination of flow from supply points to demand points. Based on this solution, we create transport nodes each of which defines the origin(s) and destination for a full truckload. In phase two, we make use of a standard tabu search method to combine these transport nodes, which can be considered to be customers in vehicle routing problems, into actual routes. The standard tabu search method is extended to consider some new features. The solution approach is implemented as part of a newly developed decision support system and we report on tests made on a set of industrial cases from major forest companies in Sweden.