Optimization of Train Operation Control Based on Soft Actor-Critic Deep Reinforcement Learning Algorithm

With the rising interest in the application of artificial intelligence methods in society, the focus on the use of deep reinforcement learning in railway transportation networks has grown significantly. This study aims to explore the optimization problem in train operation control and proposes a deep reinforcement learning approach based on the Soft Actor-Critic algorithm as a training framework. The objective is to achieve energy efficiency and on-time arrivals. By considering various factors such as train speed limits, energy consumption, and operational efficiency, a reward function is designed to guide the train in learning appropriate operating strategies. Through the integration of a realistic train dynamics model and mechanical characteristics, simulation experiments validate the feasibility of this approach in achieving energy savings and punctual arrivals.