A Group Control Method of Central Air-Conditioning System Based on Reinforcement Learning with the Aim of Energy Saving

This paper proposes a group control method for central air conditioning system based on reinforcement learning to save energy. This method does not depend on the mathematical model of the system, and can automatically improve the control strategy through continuous interaction between the controller and the system. Find the optimal working point of each device under different external conditions by continuous online learning. Even in the case that system parameters changes, the devices can still be guaranteed to work at its optimal performance point. Simulation experiments verify the effectiveness of the proposed method.