•New

Delivery companies face the daily challenge of organizing the transportation of a large number of goods from numerous points of origin to many different delivery points. Among many other factors, they have to decide on the route, the loading capacity of the vans or trucks used and the extent to which they can be filled. Above all, they have to guarantee that delivery will be made before the deadline and at the estimated cost, in spite of traffic jams.

At present, this is a logistics problem that is too complicated to be solved using a formula that can be executed rapidly at a desktop computer. However, although it does not give the perfect answer, there is already an easy-to-use tool that offers logistics managers a way of finding a solution that improves on current formulas and lowers costs. This optimization system is described in the doctoral thesis of Miquel Àngel Estrada, PhD in Civil Engineering, lecturer at the Department of Transport and Regional Planning of the Universitat Politècnica de Catalunya (UPC) and research scientist at the Transport Innovation Center, which is run jointly by the UPC and the Government of Catalonia. The thesis has been awarded the fifth Abertis prize for Research on Transport Infrastructure Management by the Abertis Chair at the UPC.

To reach the optimum solution, the system uses a tabu search-based metaheuristic algorithm, or in other words a group of operations that start by analyzing the present solution and go on to refine and improve it. The system defines the route, the size of the vehicles, and the location of the stopping points and then makes an estimate of the costs and chooses a strategy for sending the goods from three possible options.

The first option is to send them directly. The second is the multiple-stop or “peddling” option, in which the vehicle stops on the way, either to complete the original load or to deliver parts of it to different destinations. The third option uses load-transfer centers or hubs, which are storage centers where the trucks stop at unloading bays and the goods are taken out, redistributed and immediately loaded onto other vehicles at the loading bays.

The optimization system draws its conclusions on the basis of the variables involved. For example, the cost of handling goods in a hub affects the ideal-capacity calculations for the vehicles at the loading bays. The study demonstrates that small packages should be transported in large trucks when handling costs are low, whereas it is better to use vans, which fill up quickly, to transport these packages when handling costs are high.

The method proposed by Miquel Àngel Estrada has been applied in situations in which the vehicles used are carrying full loads. In this case, it reduces costs by 7% compared with solutions proposed by other heuristic methods. The result is still better in cases of fractioned loads, for which the saving is of over 12%.

In addition, the new system calculates the extra costs incurred by delivery companies due to traffic congestion on the road network, and estimates that costs rise exponentially in areas where the traffic circulates at an average of less than 40 km/h for at least four hours a day.