An improved variable neighborhood search for bi-objective mixed-energy fleet vehicle routing problem

Abstract

With increasing emphasis on environmental protection, the green vehicle routing problem has been a research hotspot. This paper investigates a mixed-energy fleet green vehicle routing problem with time windows, where the mixed-energy fleet is presented for the situation in which transport companies introduce electric vehicles into their homogeneous fleet with only gasoline and diesel vehicles (GDVs) due to environmental consideration. Since multiple pollutants are emitted by GDVs into the environment, this paper minimizes five dominating types of pollutant emissions. The other objective is to minimize the total delay time, which is defined with a newly designed optimal service period to improve customer satisfaction. To solve this bi-objective NP-hard problem, an improved variable neighborhood search (VNS) with a selection mechanism is presented to find the Pareto frontier of the model. The computational results based on Solomon benchmark show that the proposed VNS with a selection mechanism has better capability and efficiency than the original VNS. Using numerical experiments, we found that delivery with only GDVs can better minimize the delay time in cluster-distributed customers with narrow time windows, while the mixed-energy fleet performs better in both optimization objectives in wider time windows. Our research conclusion can provide a more reasonable decision-making basis of delivery for transportation companies.