Decomposition algorithms for market clearing with large-scale demand response

This paper is concerned with large-scale integration of demand response (DR) from small loads such as residential smart appliances into modern electricity systems. These appliances have intertemporal consumption constraints, while the satisfaction of the end-user from operating them is captured through utility functions. Incorporation of the appliance scheduling flexibility and user satisfaction in the system optimization points to maximizing the social welfare. In order to solve the resultant very large optimization problem with manageable complexity, dual decomposition is pursued. The problem decouples into separate subproblems for the market operator and each aggregator. Each aggregator addresses its local optimization aided by the end-users' smart meters. The subproblems are coordinated with carefully designed information exchanges between the market operator and the aggregators so that user privacy is preserved. Numerical tests illustrate the benefits of large-scale DR incorporation.