Game Theoretic Control of Thermal Loads in Demand Response Aggregators

This paper presents a game theoretic control strategy for optimal thermal load management of multiple buildings in a demand response aggregator. The electricity cost of the aggregator consists of two parts: an energy charge and a demand cost, the latter based on the aggregate peak demand over a billing horizon. We propose a non-cooperative game framework where an internal pricing mechanism is imposed with a Shapley value-based function to distribute the demand charge among the players. A Nash equilibrium problem is formulated and then transformed into a system of nonlinear equation solving problem based on the Nikaido-Isoda function. A generalized Newton-based algorithm is developed to find a Nash equilibrium of the formulated game. Simulation tests were carried out for a six-building aggregator to evaluate the performance of the proposed control strategy along with three baseline control methods. Test results showed that the proposed strategy could result in performance very close to the social optimum, with the Price of Anarchy of 1.002. The alliance stability could also be achieved in the test results.