Statistical-Machine Intelligent Relaxation for Set-Covering Location Models To Identify Locations of Charging Stations for Electric Vehicles

Abstract

Europe strengthens its policies on climate change, green transition, and sustainable energy by addressing the high greenhouse-gas emissions in the transportation sector. Europe aims to reduce such emissions and reach a state of carbon neutrality by 2030 and 2050, respectively. This is feasible only if electric vehicles dominate the transportation sector. Paving the way for electric vehicle deployment on roads is subject to the provision of electric-vehicle-charging stations on the roads such that sufficiently good driving experience without any obstacles can be achieved. To address this timely societal challenge, we proposed a novel methodology by using the well-known facility-location-allocation methodology named set-covering location models with statistical machine learning and developed it for the problem settings of identifying electric-vehicle-charging station locations. Statistical machine learning was employed in the proposed model to more precisely identify and determine feasible coverage sets. We demonstrated the efficiency of the proposed model for the Capital Region of Denmark, where the green transition is part of the political agenda and is of severe societal concern, by using the newly collected main road transportation dataset.