基于深度Q网络的轨道交通客流控制

针对高峰时期城市轨道交通因有限运能，不足以满足乘客出行需求而引发的安全问题，需要采取客流控制策略来调节进入车站的客流量，以缓解车站拥挤。提出一种基于强化学习深度 Q 网络的多站协同控制模型，用来优化每个车站在一定时间内的进站量，以最小化地铁车站乘客的站台超限量、平均等待时间，提高客流控制强度的综合效益。以北京地铁八通线为例进行仿真实验，验证该方法的有效性。仿真结果表明，所提出的模型可以在客流控制强度较低的条件下有效地降低乘客等待时间，提高乘客出行效率，有助于缓解车站的乘客拥堵。

Passenger Flow Control of an Urban Rail Transit Based on a Deep Q-network

In view of the urban rail transit safety problems caused by the limited transport capacity for passenger demand during peak periods, it is necessary to adopt a passenger flow control strategy to adjust the inbound passenger flow and alleviate station congestion. In this study, a multi-station cooperative control model based on a reinforcement learning deep-qnetwork is proposed to optimize the arrival number of each station in a certain period of time. The goal is to minimize the comprehensive benefits of platform overrun, average waiting time, and passenger flow control intensity. Considering the Beijing Metro Batong line as an example, the simulation results verified the effectiveness of the method. The simulation results demonstrate that the model can effectively reduce the waiting time and improve the travel efficiency of passengers under the condition of low passenger flow control intensity. This method is helpful in alleviating passenger congestion at stations without reducing the travel efficiency of passengers.