The specific problem we addressed in this paper, called the Multi-compartment Vehicle Routing and Loading Problem (MCVRLP). It is originated from a real-life application concerning the distribution of fuel. In this problem, we focused on the integration of the loading aspect in the routing problem where vehicles have multiple compartments and several incompatibilities need to be respected. For this problem, we presented a mathematical model, developed a tabu search metaheuristic and proposed a novel criterion based on the Kolmogorov-Smirnov statistic to prune the neighborhood. We evaluated the performance of the approach by means of extensive numerical experiments. Furthermore, we analysed a real case under different data sets to demonstrate to our industrial partner the effectiveness of the proposed approach.
Mots clés : Logistic, Vehicle routing with compartments, Loading, Fuel distribution, Heuristics