Maximizing Tourist Satisfaction: A Personalized Route Planning Model Based on Multi-objective Q-learning

With the continuous expansion of the tourism industry, there is an increasing demand for personalized and efficient travel experiences. Effective itinerary planning not only enhances tourist satisfaction but also optimizes time and resource allocation. However, conventional tourism route planning methods often rely on static rules or manual recommendations, which lack flexibility and real-time optimization capabilities. To address these limitations, this study proposes a reinforcement learning-based approach for tourism route optimization, aiming to identify the optimal visiting sequence to maximize the overall travel experience. The contributions of this study are: 1) modeling the route planning problem as a Finite MDP using Q-learning; 2) designing a multi-objective optimization framework that dynamically balances attraction attractiveness, visit duration, and commuting time; and 3) incorporating real-world road network data to ensure the accuracy and practical relevance of the proposed method. Experimental results show that the proposed Q-learning-based method increases tourist satisfaction by 18.5%, reduces commuting time by 20%, and effectively determines the optimal tour sequence within a fixed time constraint.