Facility location models for electric vehicle charging infrastructure

Abstract

This thesis deals with the study of current charging infrastructure availability in highways, as well as proposing optimal allocations for new stations. First, a Machine Learning model is trained in order to estimate the actual range of an electric vehicle. This model will be constructed using heterogeneous data sources and variables that influence the total autonomy, such as speed, temperature, degradation or elevation, among others. Second, this model is used in combination with geospatial data regarding French highway and charging infrastructure locations, in order to propose a methodology for analyzing the availability level of charging stations in highways for electric vehicles. Finally, an optimization framework is implemented to decide the opening of several charging stations inside a highway, providing as possible locations rest and service areas already built, and considering current highway operational charging points.