A Hybrid Evolutionary Algorithm for Smart Freight Delivery with Electric Modular Vehicles

摘要

This paper proposes a hybrid evolutionary algorithm for smart urban freight distribution using electric modular vehicles. The hybridization of evolutionary algorithms has recently shown promising results in vehicle routing problems with time windows. We propose here to test this technique on a new problem of routing electric vehicles with limited ranges which have the particularity of being modular. This means that the studied vehicles are composed of several modules which can be released at dedicated locations to go on performing the tours in an optimized way. For big instances of the problem which is NP-Hard, metaheuristics and in particular those involving evolutionary computation are suitable. Therefore, in this study, we propose to address the problem with the hybridization of the genetic algorithm. We use the Best Route crossover operator in combination of a Variable Neighborhood Descent algorithm for the mutation phase. An experimental study on benchmark instances from the literature shows the performance and effectiveness of the resulting evolutionary based algorithm.