Corresponding Author

Wenlong Fu

Document Type



Reinforcement learning is an important branch of machine learning and artificial intelligence. Compared with traditional reinforcement learning, model-based reinforcement learning obtains the action of the next state by the model that has been learned, and then optimizes the policy, which greatly improves data efficiency. Based on the present status of research on model-based reinforcement learning at home and abroad, this paper comprehensively reviews the key techniques of model-based reinforcement learning, summarizes the characteristics, advantages and defects of each technology, and analyzes the application of model-based reinforcement learning in games, robotics and brain science.