It is described in a paper entitled "Decentralized Control of Micro-Storage in the Smart Grid", just delivered at the AAAI-11 conference.

ECS-Electronics and Computer Science University of Southampton researchers will describe how they used a novel algorithm that controls a decentralized control mechanism to manage micro-storage in the smart grid.

The mechanism uses agent-based techniques to enable energy suppliers to manage the demand from their consumers. This allows them to reduce their wholesale purchasing costs, yielding savings of up to 16 per cent in energy cost for consumers using devices with an average capacity of 10 kWh, according to the researchers.

The researchers’ approach involves using a real-time pricing scheme that is broadcast to consumers in advance of each daily period.

Computerized agents then buy, sell, and store energy on behalf of the homeowners in order to minimize their net electricity costs.

By adjusting the pricing scheme to match the conditions on the wholesale market, the supplier is able to ensure that, as a whole, consumer agents converge to a stable and efficient equilibrium where costs and carbon emissions are minimized.

The researchers note that “Using the techniques described in this paper, we can envisage energy suppliers providing new tariffs that will incentivize consumers to buy affordable small scale storage devices. In turn this will allow suppliers to manage aggregate load profiles, improve efficiency and reduce carbon output.”

This work iss carried out as part of the IDEAS project, led by Alex Rogers and Nick Jennings of the Agents, Interaction and Complexity research group at ECS.