Personalized Physical Education Teaching Path Planning and Optimization Based on Deep Reinforcement Learning

Physical education courses have always been an important part of students’ comprehensive quality education, and how to make full use of big data technology to improve the informatisation level of physical education teaching has become an urgent problem to be solved at present. Through an overview of the characteristics of the personalised physical education teaching model based on deep reinforcement learning, the feature types of the collected data are clarified, and a series of data preprocessing operations are taken to remove the interfering information in the data. From the principle of deep reinforcement learning, the deep Q-network algorithm is designed, and the state space, action space, reward function, and selection strategy in the personalised sports teaching path planning model are also defined and designed. Comprehensive simulation experiments and statistical analysis confirm the personalized sports teaching path planning model based on deep reinforcement learning. In contrast, the DQN algorithm can achieve good results in any situation, and can give the most suitable personalised teaching plan according to each student’s problem in order to achieve the purpose of personalised sports teaching path planning and dynamic adjustment. The teaching model in this paper has a significant relationship with the traditional teaching model in terms of physical education learning autonomy, participation in physical education classes, and skill test scores, i.e., the teaching model in this paper has a better optimisation effect than the traditional teaching model.