Smarter Markets for a Smarter Grid: The Why and How
As the world moves towards a greater fraction of renewable energy in the total energy mix, a smarter grid and electricity infrastructure has become imperative. This introduces many new problems at various levels. For one, renewable energy generators can't participate in the future electricity markets without exposing themselves to too much risk. Thus, new market mechanisms and architectures are going to be needed. There would be entities called `aggregators' who would buy power from the renewable energy generators, and then sell it in the future markets, assuming any risk of a shortfall. Motivated by this, we introduce auction mechanisms (that the aggregators can use) for stochastic resources, such as wind power which can only be supplied by the seller with a certain probability. The goal of the `aggregator' is to elicit these probability distributions truthfully. We propose two mechanisms, a stochastic VCG mechanism, and a shortfall penalty mechanism, both of which we show to be incentive-compatible. It seems that mechanisms for stochastic resources have not been studied before in the theory of mechanism design.
In the second part of the talk, we take another look at the Locational Marginal Pricing (LMP) mechanism that is used daily by almost all the Independent System Operators (ISOs) to determine generation and supply schedules. It is part of the folklore in Power System Economics that an equilibrium exists in the LMP mechanism. In this talk, we first show that contrary to conventional wisdom, a Nash equilibrium (NE) may not exist in the LMP mechanism. We then introduce two power network second-price (PNSP) mechanisms, one single-sided and another double-sided. In either case, we show the existence of an efficient NE. The double-sided PNSP mechanism however, is not budget-balanced, and there is room for more suitable mechanisms.
Bio: Rahul Jain is the K. C. Dahlberg Early Career Chair and Associate Professor in the EE & ISE Departments at USC. He received his B.Tech from IIT Kanpur, and an MA in Statistics and a PhD in EECS from the University of California, Berkeley. His interests span stochastic control, game theory, network economics, queueing theory, and statistical learning with applications to communication networks and power systems. And of course, he is the instructor for this EE 598 course.