Planning EV Charger Placements with Heterogeneous Charging Technologies

Abstract

When developing public urban charging infrastructure for electric vehicles (EVs), key questions are how many chargers to deploy, where to locate them, and what charger technology to utilize. This paper introduces a facility location model with multiple facility types that jointly optimizes the placement and type of EV chargers to minimize total infrastructure costs while meeting spatially distributed demand. We then study the benefit of a hybrid mix of chargers relative to single-technology solutions, demonstrating that the gains from hybrid solutions depend critically on demand pooling structures. In particular, hybrid solutions yield the highest benefit when there is heterogeneity in the amount of demand served at each location. We complement our theoretical results through a data-driven case study based on the City of Detroit, developing an end-to-end pipeline to solve the problem for real cities. First a computer vision model finds feasible curbside charging locations by analyzing images from Google Street View, and then an optimization model determines the optimal placement and technology of chargers in Detroit. This pipeline is demonstrably more effective than either machine learning or optimization alone. This work provides both analytical insight and a scalable methodology to support cities in designing cost-effective EV charging networks.