基于深度 Q 网络的城市轨道交通协同限流方法 |
| |
引用本文: | 王殿元,赵兴东,豆 飞,周 旭. 基于深度 Q 网络的城市轨道交通协同限流方法[J]. 都市快轨交通, 2024, 37(3): 97-102 |
| |
作者姓名: | 王殿元 赵兴东 豆 飞 周 旭 |
| |
作者单位: | 交控科技股份有限公司,北京 100070;北京市地铁运营有限公司,北京 100044 |
| |
基金项目: | 国家重点研发计划资助(2020YFB1600702);北京市科技新星计划项目(Z211100002121098) |
| |
摘 要: | 为解决城市轨道交通高峰小时区间满载率过高的问题,本文提出一种基于深度强化学习的城市轨道交通协同限流控制方法。该方法利用历史客流数据建立线网层面的限流仿真环境和智能体模型,以区间满载率为状态,以限流策略为动作,以客流体验为奖励,通过多轮强化学习训练产生最优的限流方案。随后利用北京地铁线网数据构建仿真实验并验证了该方法的有效性。仿真结果表明,协同限流方法可以有效降低断面客流量,缓解高峰小时区间拥挤程度,提高乘客出行舒适度。
|
关 键 词: | 城市轨道交通;深度强化学习;客流控制;北京地铁 |
Cooperative Passenger Flow Control Method for Urban Rail TransitUtilizing Deep Q-Network |
| |
Affiliation: | Traffic Control Technology Co., Ltd., Beijing 100070;Beijing Subway Operation Co., Ltd., Beijing 100044 |
| |
Abstract: | Rapid urbanization and population growth have led to a continuous increase in passenger flow in urban rail transit,which presents significant challenges to the safety, comfort, and stability of rail transit operations. To solve the problem ofexcessive load rate of urban rail transit during peak hours, we propose a cooperative passenger flow control method for urbanrail transit based on deep reinforcement learning. This method uses the full load rate between intervals as its state, a flow restrictionstrategy as its action, and the passenger flow experience as its reward. It generates an optimal flow restriction scheme throughmulti-round reinforcement learning. We validated the effectiveness of this method by constructing simulation experiments usingdata from the Beijing subway network. The simulation results show that the cooperative passenger flow control method caneffectively reduce passenger flow in a section, relieve congestion during peak hours, and improve passenger travel comfort. |
| |
Keywords: | urban rail transit deep reinforcement learning passenger flow control Beijing Subway |
|
| 点击此处可从《都市快轨交通》浏览原始摘要信息 |
|
点击此处可从《都市快轨交通》下载免费的PDF全文 |
|