Hybrid metaheuristic to solve location problem for electric vehicles charging stations
Résumé
Electric vehicles (EV) have recently appeared on the transport market as a solution for reducing greenhouse gas emissions. They are promising means of clean transportation. However, they are obstructed by their limited autonomy that deprives them of completing long trips without a periodic refueling. To encourage their use, the researchers took an interest in the way to optimally deploy charging infrastructure while minimizing investments costs and maximizing the network's coverage as part of the facility location problem (FLP). Several models have been designed to solve the location problem for EV charging stations (CS) (LPEVCS). This work proposes a Mixed Integer Programing (MIP) model for the LPEVCS. Given a set of potential sites, a limited number of stations and charging points (CP) to be located, the LPEVCS seeks to find the optimal location of the CS so as to cover the recharging demand as widely as possible and to ensure continuity in the EV trips. To solve this problem on a large scale, we introduce intelligence into swarms. We propose a hybrid method based on the Particle Swarm Optimization (PSO) and Tabu Search (TS) algorithms. This approach is used to solve several generated sets of data with varying sizes using Java language and proved the possibility to obtain near-optimal solution for large instances