Charging Electric Vehicles Fairly and Efficiently

Ramsundar Anandanarayanan, Swaprava Nath, Rohit Vaish

Abstract: Motivated by electric vehicle (EV) charging, we formulate the problem of and efficient allocation of a divisible resource among agents that arrive and depart over time and consume the resource at different rates. The agents (EVs) derive utility from the amount of charge gained, which depends on their own charging rate as well as that of the charging outlet. The goal is to allocate charging time at different outlets among the EVs such that the final allocation is envy-free, pareto optimal, and in certain contexts, group-strategyproof. The differences in the charging rates of the outlets and the EVs, and a continuous time-window where the arrivals and departures occur make this a non-trivial combinatorial optimization problem. We show possibilities and impossibilities of achieving a combination of properties such as envy-freeness, pareto optimality, leximin, and group-strategyproofness under different operational settings, e.g., when the EVs have (dis)similar charging technology, or when there are one or more dissimilar charging outlets. We complement the positive existence results with polynomial-time algorithms.