A Decision Support System for a Long-distance Routing Problem based on the Ant Colony Optimization Metaheuristic

AbstractThis paper presents a decision support system for the optimization of long distance freight transport by road in the Iberian Peninsula (Spain and Portugal) between several origins and destinations without warehouses at the origin and the final destination. In every order, there is at least one origin and one destination, and these two positions do not need to coincide with other origin/destination pairs for different orders. This can be considered to be a many-to-many type problem, where the freight is picked up at different customer locations and delivered in other client locations.The aim of the system described consists of helping traffic managers of freight transport companies to reduce operational costs by optimizing the loading of freight in vehicles, optimizing routes and grouping orders following certain load/unload procedures in such a way that pick-up and delivery activities can be previously planned.For the resolution of the problem, it is proposed an algorithm based on ant optimization techniques which has allowed the system to be modeled taking into account its most characteristic peculiarities: no depots at the beginning and end of the route, a maximum driving time per day, the capacity of vehicles, the compatibility of types of goods in the same vehicle, customers with time windows to pickup/deliver goods, and the manner of loading/unloading vehicles (Last Input First Output).The effectiveness of the algorithm has been proved using data from real problems. The basis of our computational experiments is the historical data from a large transport company in the Iberian Peninsula