Agents help schedule efficient charging of electric cars

Researchers in ECS-Electronics and Computer Science at the University of Southampton have designed a new pricing mechanism based on an online auction protocol that makes it possible to charge electric vehicles without overloading the local electricity network.

The new system was designed by a team led by Dr Alex Rogers and presented at AAMAS 2011 – the Tenth Conference on Autonomous Agents and Multiagent Systems - held in Taiwan earlier this week. The researchers have devised a system in which electric vehicle owners use computerised agents to bid for the power to charge the vehicles and also organise time slots when a vehicle is available for charging.

"Plug-in hybrid electric vehicles are expected to place a considerable strain on local electricity distribution networks. If many vehicles charge simultaneously, they may overload the local distribution network; so their charging needs to be carefully scheduled," said Dr Rogers.

To address this issue, Dr Rogers and his team turned to the field of online mechanism design. They designed a mechanism that allows vehicle owners to specify their requirements (i.e. when they need the vehicle, and how far they expect to drive). The system then automatically schedules charging of the vehicles’ batteries. The mechanism ensures that there is no incentive to ‘game the system’ by reporting that the vehicle is needed earlier than is actually the case, and those users who place a higher demand on the system are automatically charged more than those who can wait.

“The mechanism leaves some available units of electricity unallocated. This is counter-intuitive since it seems to be inefficient but it turns out to be essential to ensure that the vehicle owners don’t have to delay plugging-in, or misreport their requirements, in an attempt to get a better deal”, said Dr Enrico Gerding, the lead author of the paper.

In a study based on the performance of currently available electric vehicles, performed by Dr Valentin Robu and Dr Sebastian Stein, the mechanism was shown to increase the number of electric vehicles that can be charged overnight, within a neighbourhood of 200 homes, by as much as 40 per cent.

This research follows on from Dr Rogers’ and Professor Nick Jennings’ work on developing agents that can trade on the stock market and manage crisis communications and Dr Rogers’ iPhone application, GridCarbon for measuring the carbon intensity of the UK grid.