Event triggered non-cooperative game theory for demand side management in smart grids

Over the past decade, a new concept for the power grid called the smart grid emerged, to enable two way energy transport and communication. With this new smart grid architecture many opportunities for control methods arise. While this can improve the efficiency of the power grid, all these methods require communication, which is a limited resource. In this thesis I focus on one particular concept of the smart grid called demand side management. One of the possible ways to apply demand side management is by use of game theory. By molding the game into a specific kind of game called potential game, similarities between the convergence of these potential games and the methods used in event triggered control, which is a method to reduce communication in networked control systems, can be observed. By using these similarities the principles of event triggered control can be applied to potential games. This way I create an event triggered potential game, which reduces the amount of communication required to reach the Nash equilibrium. To accomplish this I design two algorithms for event triggered potential games. One in which the decision on whether or not to communicate is made in a centralized way and one in which this occurs in a decentralized way, where all players decide for themselves when to communicate.Mechanical Engineering | Systems and Contro