Iterated Tabu Search for the Mix Fleet Vehicle Routing Problem with Heterogenous Electric Vehicles
Résumé
In this paper, we address the vehicle routing problem with mixed fleet of conventional and heterogenous electric vehicles, denoted VRP-MFHEV. This problem is motivated by a real-life industrial application and it is defined by a mixed fleet of heterogenous Electric Vehicles (EVs) having distinct battery capacities and operating costs, and identical Conventional Vehicles (CVs) that could be used to serve a set of geographically scattered customers. The EVs could be charged during their trips at the depot and in the available charging stations, which offer charging with a given technology of chargers and propose different charging costs. EVs are subject to the compatibility constraints with the available charging technologies and they could be partially charged. The objective is to minimize the number of employed vehicles and to minimize the total travel and charging costs. To solve the VRP-MFHEV, we propose a Multi-Start Iterated Tabu Search (ITS) based on Large Neighborhood Search (LNS). The LNS is used in the tabu search of the intensification phase and the diversification phase of the ITS. Different implementation schemes of the proposed method including best-improvement and first-improvement strategies, are tested on generalized benchmark instances. The computational results show that ITS produces competitive results, with respect to results obtained in previous studies, while the computational time remains reasonable for each instance. Moreover, using LNS in the intensification phase of ITS seems improving the generated solutions compared to using other neighborhood search procedures such as 2opt.