The Model for Personalized Path Planning of English Learning Based on DQN

With the rapid development of artificial intelligence technology, personalized learning has become an important research direction in the field of education. Aiming at the problem that the traditional English learning mode is difficult to dynamically adapt to the differentiated needs of learners, this study proposes a personalized path planning model based on Deep Q-Network (DQN). Through the reinforcement learning framework, the model takes the learner's real-time state (knowledge mastery, cognitive load, historical behavior, etc.) as input, combines experience playback and target network mechanism, and dynamically generates learning paths adapted to individual characteristics. Experiments show that compared with Convolutional Neural Network (CNN) and Long-Short-Term Memory Network (LSTM), DQN model performs best in cumulative reward and discounted cumulative reward, and the convergence speed is improved by about 20%. In the actual case verification, the model is deployed to the online English platform to verify its effectiveness in dynamically adapting the cognitive state and optimizing the learning efficiency. This study provides a new technical path for personalized path planning in the field of education, and helps to break through the efficiency bottleneck of traditional learning model.