Research on the Application of Reinforcement Learning in Intelligent Recommendation System for Tourist Destinations

With the booming tourism industry, it is crucial to provide tourists with accurate and personalized recommendations for tourist destinations. This paper proposes Dynamic Environment Aware Reinforcement Learning (DEARL) and applies it to the intelligent recommendation system for tourist destinations. The DEARL algorithm can perceive dynamic environmental factors such as changes in tourist preferences and the real-time status of tourist destinations in real time, and continuously optimize the recommendation strategy by building a model based on the Markov decision process. The experiment uses a real tourism data set, covering the historical itinerary and preference information of 5,000 tourists, and compares traditional recommendation algorithms such as collaborative filtering algorithm (CF), content-based recommendation algorithm (CB) and classic reinforcement learning algorithm Q-learning. In terms of recommendation accuracy, the DEARL algorithm reaches <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{8 5. 3 \%}$</tex>, which is significantly higher than CF's <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{6 8. 5 \%}$</tex>, CB's 72.1 % and Q-learning's 79.2 %. In terms of user satisfaction, the DEARL algorithm received a score of 4.2 (out of 5 points), far exceeding other algorithms. The results show that the DEARL algorithm can effectively improve the accuracy and user satisfaction of tourist destination recommendations, and provide new ideas and methods for the development of intelligent tourism recommendation systems.