A closed-loop self-learning demand response system and method

The present invention discloses a closed-loop self-learning demand response system and method, the system comprising DR server, the client and the smart meter DR; DR service ends utility power distribution date information, service control information to the client DR stimulation power load scheduling apparatus; the DR client uses a multi-input single-output power on input data network information, the power load device power classification, classification treatment, using rule-based programs and power load devices continuous learning and decision-making, power demand information prediction , and brought together to form a closed loop DR server, stimulating demand for power companies to organize information, cleaning, mining, and processing results in the future based on DR planning, scheduling, control, and produce new information. The present invention can timely and effectively balanced power load, and alerts the user to switch the electric energy, thus reducing expenses and improving the advantages of grid stability.