Application of GRASP methodology to Vehicle Routing Problem (VRP) R. Pino 1 , C. Martínez 1 , V. Villanueva 1 , P. Priore 1 , and I. Fernández 1 1 Department of Business Management, University of Oviedo, Gijón, Asturias, Spain Abstract - In this work the development and implementation of a Decision Support System is described. It will assist in the route calculation process and in optimizing the filling of the trucks that have to transport a considerably number of vehicles from near of 8 origins to more than 3,000 possible destinations spread over Spain and Portugal, and some in France, Germany, etc. The problem can be classified as VRP (Vehicle Routing Problem), been a hybrid between the VRPTW (VRP with Time Windows) and MDVRP (Multi Depot VRP). For its resolution, a Web application using a GRASP methodology as an optimization tool has been developed. The experiments carried out prove an improvement in the results of the freights/routes planning when the DSS is used. The improvement is measured trough the percentage of daily vehicles dispatched, the decrease of the shipment delay, and truck use, among others. Keywords: VRP, Heuristic Optimization, GRASP 1 Introduction The aim of the this work is to present the results obtained in the development of a research project in which heuristics techniques are applied to solve vehicle routing problems. This research was developed as an independent subproject within the project called SITIM 1 , whose aims are to study, evaluate and develop methodologies and systems to improve the Intermodal Transportation through scientific and technological knowledge. For this purpose, a number of actors participate in the project, among which are: Research Groups of various Spanish universities (Universities of Valladolid and Oviedo), public Logistics and Transportation-related institutions and several private companies of these sectors. The subproject aims at analyzing how operational optimization and routing calculation techniques may help a logistics operator to be more effective and efficient. Specifically, the case of a logistics company specialized in the automotive sector has been studied. It provides distribution services, mainly by road transport, for those companies that have the need for vehicle shipments to / from any point of the 1 Análisis, desarrollo y evaluación de Sistemas Inteligentes de Transporte Intermodal. Proyecto MFOM-08-E12/08 del Subprograma para la Movilidad Sostenible y el Cambio Modal en el Transporte del Ministerio de Fomento, Gobierno de España. Iberian Peninsula as well as some strategic points in France and Germany. The planning necessary to carry out these services is currently performed by a team of technicians that make an optimization between service demand and available resources (fleet of trucks). On the basis of their experience, they decide what vehicles travel in each of the available trucks, in order to meet all requirements (specially delivering freight on time), objectives and/or restrictions that the business strategy. These technicians use an ERP platform that manages ordering, billing,... but does not implement any support for transportation planning. The above mentioned procedure presents a series of drawbacks. Basically the search for optimal routes causes a delay in shipments, resulting in having to face penalties more often than desirable. The achievement of an optimized planning after evaluating all (or most) possible combinations of transport which satisfy the objectives and/or restrictions would be necessary. When the dimension of the problem is large as in the present case, a person is not able to evaluate all possible combinations, unless he makes a thorough and detailed study for each transport mean, a fact that becomes functionally and economically impractical. A Support System for Decision Making (DSS: Decision Support Systems) is a software application that, using artificial intelligence algorithms, will provide the user with a better loads/routes planning. As a result, the technicians work would be more efficient and effective, and the objectives achieved to a greater extent. The paper is organized as follows. In the remaining of this section a literature review is made to analyze different options for solving the addressed problem. In second section, the problem and current operational process is presented, emphasizing those aspects that should be improved for achieving greater efficiency work levels. Then, the proposed solution, the chosen heuristics techniques (GRASP methodology) and the developed Web application are described. In the fourth section, the solution is analyzed and compared with the current situation. The paper ends with a section summarizing the final conclusions of the work. From the point of view of operational optimization, the problem tackled in this paper is a Vehicle Routing Problem (VRP), as first established Dantzig and Ramser [1], and