A multi-objective evolutionary approach for the electric vehicle charging stations problem

This paper presents a framework that integrates a mobility and traffic simulator with evolutionary algorithms to solve the location problem of electric vehicle charging infrastructure. The framework presented in this research is configured with a real-life scenario that considers a maximum number of charging stations and a specific number of electric vehicles that are driven according to a movement distribution model for the urban city under consideration. The concerning scenario is formulated as a multi-objective optimization problem and is approximated by two multi-objective evolutionary algorithms based on two evolutionary approaches for multi-objective optimization, namely, NSGA-II and MOEA/D-gen. In this way, the adopted algorithms search for optimal station locations and capacities in terms of travel time and the number of charging stations. The analysis of the results obtained by the multi-objective approaches used in the proposed framework is two-fold. On the one hand, solutions confirm that both objectives considered in this study are in conflict. That is, the better the solution for one objective is, the more deteriorated it is for the other. On the other hand, it was found that NSGA-II obtained a better approximation to the optimal solutions according to two multi-objective performance indicators. These results give an idea of the type of algorithms that could best solve the formulated problem in terms of efficiency and effectiveness and understand the behavior of evolutionary algorithms in the multi-objective electric vehicle charging stations problem. Therefore, this research traces the route to study different scenarios and analyze the characteristics of the problem. Furthermore, it opens the door to exploring other meta-heuristics and their components to deal with this problem adequately.