Similarity-Based Label Propagation for Electric Vehicle Network Representation Learning

The Electric Vehicle (EV) market is rapidly expanding as more individuals shift from traditional fuel-powered vehicles to EVs. This shift requires the broad deployment of Electric Vehicle Supplement Equipment (EVSE), including charging stations, to support the growing EV infrastructure. This paper introduces a semi-supervised learning approach to optimize EVSE charging station deployment. We propose a label-propagation algorithm to consider structural similarities across distant nodes in EVSE points in the electric vehicle charging networks, enabling more accurate similarity extraction. Our results demonstrate that this approach has potential for improving the accuracy and efficiency of EVSE charging station deployment. Our semi-supervised method can be applied to optimize efficient charging station deployment in a simulation for digital twin use.