Research on Collaborative Optimization Path Between Beijing Marshalling Yard and Urban Planning Based on Reinforcement Learning

This research employs deep Q-network (DQN) technology to optimize Beijing's rail operations through adaptive marshaling yard scheduling, targeting enhanced track utilization and reduced train dwell times. By modeling real-time interactions between rail logistics and urban traffic patterns, the self-adjusting system dynamically allocates resources while mitigating congestion hotspots. Tests across multiple scenarios demonstrate measurable improvements in scheduling efficiency and traffic flow coordination compared to conventional methods, particularly in managing variable demand and urban infrastructure constraints.