The Application of DQN in Thermal Process Control
Tao Ao, Jiong Shen, Xichui Liu · 2019
In recent years, deep reinforcement learning has been widely used in games, robots, autonomous driving and other fields, which proves that deep reinforcement learning is powerful in decision making and problem solving. In this paper, according to deep reinforcement learning and the characteristic of thermal process control, a thermal process control method based on DQN (Deep Q-learning Network) is proposed. Through matlab and python mixed programming, the process of DQN controller training and the simulation of control system are completed. The simulation experiment of water level control system of water tank based on DQN controller shows that deep reinforcement learning can be well applied to thermal process.