A decentralized approach for public fast charging of electric vehicles using delegate multi-agent systems

Electric vehicles (EVs) will play a central role in future road traffic. To cope with the limited range and long charging times of such vehicles, a public fast-charging infrastructure will be deployed to recharge EV batteries. A challenge in such an infrastructure is the high costs associated with electrical peak loads; a charging station along a busy highway may see hundreds of EVs per hour. Driven by this challenge, we take a first step towards a multi-agent solution, which coordinates EVs to select fast charging stations along their route. In this solution, EVs are represented by EV agents, and charging stations by CS (charging station) agents. The coordination between these agents is based on the coordination technique “delegate multi-agent system (dMAS)”, in which ants are used to delegate charging requirements from EV agents towards CS agents. These ants are repeatedly send out in order to cope with a dynamic and uncertain environment. Our approach is evaluated in a simulation of a highway segment in Belgium. In this scenario, the delegate MAS approach is compared to (1) a naive charging strategy and (2) a TMC (traffic message channel) based charging strategy. Results show that our approach achieves significant cost reductions compared to these strategies.