A Multiagent Approach to Variable-Rate Electric Vehicle Charging Coordination

We present a coordination mechanism for variable-rate electric vehicle (EV) charging that combines the benefits of a decentralized decision making approach with a top-down control mechanism based on price functions. This combination allows independent decisions by self-interested EV agents, while ensuring that aggregate power demand from EV charging converges to a desired profile as determined by a control agent. This profile can yield reduced demand volatility, or a better match between the output of renewable energy sources and overall demand. We observe that price signals from the control agent are sufficient to motivate self-interested EV agents to adjust charging rate and produce the desired profile, even given a range of individual user preferences. We show that our hybrid coordination mechanism prevents herding in EV charging, which is typical in populations where all agents receive the same price signals and make similar charging decisions. Specifically, the control agent learns the responses of the EV agents in order to adjust its price signals, and eventually converges to coordinated charging, thereby producing the desired demand profile. We compare our approach with various benchmarks and show that EV charging congestion (herding) is reduced, while peaks and volatility of demand are mitigated.