Achieving fairness for EV charging in overload

With the emergence of Electrical Vehicles (EVs), there is a growing investment in power infrastructure to provide charging stations. In an EV parking lot, typically not all vehicles can be charged simultaneously, and thus some scheduling must be performed, taking into account the time the users are willing to spend in the system. In this paper, we analyze the performance of several common scheduling policies through a fluid model. We show that in overload, the amount of unfinished work is the same for all policies, but these can distribute the work performed unfairly across users. We also introduce a new policy called Least Laxity Ratio that achieves a suitable notion of fairness across jobs, and validate its performance by simulation.