A Bayesian mean field game approach to supply demand analysis of the smart grid

Abstract

We explore a game theoretic framework for multiple energy producers competing in energy market. Each producer, referred to as a player, optimizes its own objective function given the demand utility. The equilibrium strategy of each player depends on the production cost, referred to as type, of the other players. We show that as the number of players increases, the mean of the types is sufficient for finding the equilibrium. For finite number of players, we design a mean field distributed learning algorithm that converges to equilibrium. We discuss extensions of our model to include several realistic aspects of the energy market.