Evolutionary bi-objective optimization for the electric vehicle charging stand infrastructure problem

Abstract

This article reports using a bi-objective evolutionary algorithm interacting with a traffic simulator and data exploration methods to analyze the optimal capacity and location of charging infrastructure for electric vehicles. In this work, the focus of the study is the city of Cuenca, Ecuador. We configure a scenario with 20 candidate charging stations and 500 electric vehicles driving according to the mobility distribution observed in this city. We optimize the vehicle's travel time that requires recharging and the number of charging stations distributed in the city. Quality of Service is defined as the ratio of charged vehicles to vehicles waiting for a charge and is considered a constraint. The approximate Pareto set of solutions produced in our experiments includes a number of trade-off solutions to the formulated problem and shows that the evolutionary approach is a practical tool to find and study different layouts related to the location and capacities of charging stations. In addition, we complement the analysis of results by considering Quality of Service, charging time, and energy to determine the city's best locations. The proposed framework that combines simulated scenarios with evolutionary algorithms is a powerful tool to analyze and understand different charging station infrastructure designs.