| The purpose of the thesis consists of the presentation of an economic model of goods distribution which generalizes the vehicle routing problem with time windows (VRPTW), along with a set of heuristic and metaheuristic techniques able to work it out efficiently. This work systematizes the set of methods of heuristic optimization and establishes the state of the technique regarding the used procedures in the solving of the VRPTW problem and its extents. After having verified certain discrepancies among the theoretical models and the real cases, the thesis defines an objective function which measures the economic profitability of the operations, and makes soft time windows using penalties which take into account customers dissatisfaction. In the same way the feasibility of using a heterogeneous fleet of vehicles with differentiated fixed and variable costs is considered, as well as the loading capacity, speed and different working hours, and, furthermore, the feasibility of multiple use. The asymmetry in the length of time of routes, with approximation and removal periods which regulated level of congestion due to traffic and other access problems, is included. It is also possible to adjust different costs derived from scheduling depending on extra-hours and penalties for breaking deadlines on the arrival to the depot.; The thesis presents a new heuristic frame of sequential route building based on economic approaches (HESECOR) able to solve the proposed model which, furthermore, and in the case of the VRPTW basic problem, has reached, in some cases, the best published solution. Also, a set of metaheuristics based on a sequential neighbourhood search is presented. Starting from the analysis of the behaviour of these techniques regarding the basic and general problems presented, conclusions of practical interest, for the heuristic optimization of combinatory problems as well as for the decision making by transportation companies, are given. |
