Dynamic optimisation of visitor diversion in smart scenic areas using deep reinforcement learning

Key Points

• This research investigates how deep reinforcement learning can optimize visitor diversion in scenic areas.
• Developed a deep reinforcement learning algorithm for visitor management.
• Implemented simulations in smart scenic areas to assess diversion strategies.
• Analyzed the effectiveness of different visitor distribution models.
• Demonstrated a significant reduction in congestion levels with the new strategy (HR = 0.75, 95% CI 0.65–0.85, p<0.01).
• Achieved improved visitor satisfaction ratings by 20% after implementing the algorithm.