February 1, 12:00-12:50PM, CS Conference Room
Abstract: Unlike data networks, electricity networks have the particularity that consumption and production (demand and capacity) must balance at all times. In traditional electricity grids, this has led to peaks in demand causing failures within the grid. Indeed, consumer behavioral patterns are such that the overall demand on the grid may exhibit high peaks. We consider two methods of alleviating such peaks. For both we use dynamic prices as signals from the grid to consumer on the load present on the grid. With the forthcoming liberalization of energy markets, dynamic pricing is not unforeseen. We first consider the case where consumers are unwilling to change behaviour and equip themselves with batteries. These energy storage devices will serve to charge electricity when the price is low and provide electricity to the home when price is high. We model the problem as one of optimal control and show results on optimal charging and discharging policies. We then move to the scenario where users are willing to change behaviours and consider demand-scheduling policies. We model the problem as a game where the utility providers and consumers react to one another by adjusting prices and shifting demand, respectively. We show the effectiveness of the resulting policies on load shedding.
Speaker Biography: Nidhi Hegde is a research in the Paris Research Lab of Technicolor. Her research interests are in algorithmic design and performance evaluation of distributed systems and computer networks. Prior to Technicolor she has held research positions at France Telecom R&D, CWI (The Netherlands), INRIA Sophia Antipolis (France), and Bell Labs (USA).