Integrated location and routing optimization for the application of electric vehicles in on-demand delivery

Recently, electric vehicles have been utilized more often in city logistics for on-demand delivery due to either the financial subsidies from the government or the restrictions of some environmental policies. Among the various types of electric vehicles, electric mopeds play a significant role in the last-mile delivery of city logistics, especially for meal delivery scenarios. However, the limited driving range due to the small battery capacity complicates the application and scheduling of electric mopeds. In this research, we propose a battery-swapping station location-routing problem for the application of electric mopeds in meal delivery scenarios. The purpose is to determine the optimal deployment for battery-swapping stations and the routes for each rider simultaneously. Moreover, the scales of battery-swapping stations are also considered in terms of the number of available batteries. An improved adaptive large neighborhood search algorithm is devised to solve this problem, in which several customized operators are designed to facilitate the convergence of the searching process. Computational experiments demonstrate that the proposed algorithm outperforms the commercial solver Gurobi over small-scale instances in terms of the solving time and has a relatively dominant performance over the other two metaheuristic approaches for solving large instances. Furthermore, extensive sensitivity analysis is conducted regarding the cost settings of battery-swapping stations, the driving range of electric mopeds, and customer demand levels so as to provide practical and valuable managerial insights for city logistics enterprises.