面向多目标协调的细粒度交通信号智能体建模与控制方法

doi:10.16097/j.cnki.1009-6744.2026.03.019

交通运输系统工程与信息 ›› 2026, Vol. 26 ›› Issue (3): 203-213.DOI: 10.16097/j.cnki.1009-6744.2026.03.019

• 智能交通系统与信息技术 • 上一篇下一篇

面向多目标协调的细粒度交通信号智能体建模与控制方法

陈予禾^1,2a，徐新忠³，姜雯文¹，阙恒荣^2b，王屏^*1

1. 同济大学，道路与交通工程教育部重点实验室，上海201804；2.东南大学，a.交通学院， b. 网络空间安全学院，南京211102；3.上海市道路运输事业发展中心，上海200023

收稿日期:2026-02-04 修回日期:2026-03-23 接受日期:2026-03-31 出版日期:2026-06-25 发布日期:2026-06-23
作者简介:陈予禾（2000—），女，海南海口人，博士生。
基金资助:
上海市交通委员会科研项目 (JT2024-KY-004)；宁波市科学技术局项目“科创甬江2035”重大应用示范计划 (2025Z192)。

Fine-Grained Traffic Signal Agent Modeling and Control Method for Multi-objective Coordination

CHEN Yuhe^1,2a, XU Xinzhong³, JIANG Wenwen¹, QUE Hengrong^2b, WANG Ping^*1

1. The Key Laboratory of Road and Traffic Engineering, Ministry of Education, Tongji University, Shanghai 201804, China; 2a. School of Transportation, 2b. School of Cyber Science and Engineering, Southeast University, Nanjing 211102, China; 3. Shanghai Municipal Center for the Development of Road Transport Services, Shanghai 200023, China

Received:2026-02-04 Revised:2026-03-23 Accepted:2026-03-31 Online:2026-06-25 Published:2026-06-23
Supported by:
Shanghai Municipal Commission of Transportation Project (JT2024-KY-004)；Ningbo Municipal Bureau of Science and Technology Project: "Science and Technology Innovation Yongjiang 2035" Major Application Demonstration Program (2025Z192)。

摘要/Abstract

摘要： 为解决城市交通信号控制难以兼顾通行效率、公交优先与环境影响等多目标问题，本文提出一种基于图注意力机制的交通信号智能体建模方法。首先，在路网空间建模方面，构建以“进口道”为图节点的交通路网空间建模方法，并进一步引入图注意力机制来描述路段间的空间依赖关系，增强智能体对上游交通实况的感知能力。其次，在交通流量时序建模方面，比较多层感知机与长短时记忆网络两种策略网络。然后，在多目标优化任务方面，设计根据道路实况数据进行权重自适应动态调整的奖励函数，实现通行效率、公交优先与环境影响的多目标协同优化。最后，选取一个三纵三横，边长为1.2km正方形结构的城市交通信号网格作为实验路网，对交通信号智能体进行训练，并在不同流量及波动场景下进行测试。结果表明，本文提出的交通信号智能体在各目标上均显著优于简单自适应控制。通行效率方面，在高流量下理想车速达成度提升36.21%；环境影响方面，燃油效率提升超过11.0%；公交优先方面，模型使公交通行能力提高约25.0%，车速接近性提升14.5%。因此，本文构建的多目标交通信号智能体模型具备良好的多目标协调能力，为构建高效、公平、环保的智能交通信号控制提供了可行路径。

关键词: 城市交通, 图注意力机制, 强化学习, 交通信号控制, 公交优先, 环境影响

Abstract: To address the multiple objectives challenge of urban traffic signal control that is difficult to simultaneously balance the traffic efficiency, public transport priority, and environmental impact, this paper proposes a modeling method for traffic signal agent based on a graph attention mechanism. First, in terms of road network spatial modeling, an innovative approach centered on "approach road segments" is constructed, and a graph attention mechanism is further introduced to characterize the spatial dependency relationships among road segments. Thereby it enhances the perception of agent in upstream traffic conditions. Second, in traffic flow temporal modeling, this paper compares two types of network architectures: a multilayer perceptron and a long short-term memory network. Third, for the multi-objective optimization task, a reward function with adaptive and dynamic weight adjustment based on real-time road conditions is designed, enabling the coordinated optimization of traffic efficiency, public transport priority, and environmental impact. Finally, a square urban traffic signal grid consisting of three north-south and three east-west corridors with a side length of 1.2 km is selected as the experimental road network, on which the traffic signal agent is trained and tested under different traffic volumes and fluctuation scenarios. The results show that the proposed traffic signal agent significantly outperforms a simple adaptive control across all objectives: in terms of traffic efficiency, the achievement degree of ideal vehicle speed under high traffic demand is improved by 36.21%; in terms of environmental impact, fuel efficiency is improved by more than 11.0%; and in terms of public transport priority, the model increases the number of buses passing through by approximately 25.0%, with bus speed closeness improved by 14.5%. Therefore, the agent model of multi-objective traffic signal developed in this paper demonstrates a strong multi-objective coordination capability and provides a feasible pathway for building efficient, equitable, and environmentally friendly intelligent traffic signal control systems.

Key words: urban transportation, graph attention mechanism, reinforcement learning, traffic signal control, bus priority, environment impact

中图分类号:

U121

陈予禾, 徐新忠, 姜雯文, 阙恒荣, 王屏. 面向多目标协调的细粒度交通信号智能体建模与控制方法[J]. 交通运输系统工程与信息, 2026, 26(3): 203-213.

CHEN Yuhe, XU Xinzhong, JIANG Wenwen, QUE Hengrong, WANG Ping. Fine-Grained Traffic Signal Agent Modeling and Control Method for Multi-objective Coordination[J]. Journal of Transportation Systems Engineering and Information Technology, 2026, 26(3): 203-213.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: http://www.tseit.org.cn/CN/10.16097/j.cnki.1009-6744.2026.03.019

http://www.tseit.org.cn/CN/Y2026/V26/I3/203

参考文献

[1] 公安部。全国机动车保有量达4.69亿辆驾驶人达5.59 亿名 [EB/OL]. (2026-01-27)[2026-05-18]. https:// www.mps.gov.cn/ n2254314/ n6409334/ c10383533/ content.html. [Ministry of Public Security of China. The Number of Motor Vehicles in China Reaches 469 Million and Drivers Reach 559 Million[EB/OL]. (2026-01-27) [2026-05-18]. https://www.mps.gov.cn/ n2254314/n6409334/c10383533/content.html.]

[2]ZHANG K, BATTERMAN S. Air pollution and health risks due to vehicle traffic[J]. Science of the Total Environment, 2013, 450/451: 307-316.

[3]ABDULHAI B, PRINGLE R, KARAKOULAS G J. Reinforcement learning for true adaptive traffic signal control[J]. Journal of Transportation Engineering, 2003, 129(3): 278-285.

[4]CHU T, WANG J, CODECA L, et al. Multi-agent deep reinforcement learning for large-scale traffic signal control[J]. IEEE Transactions on Intelligent Transportation Systems, 2019, 21(3): 1086-1095.

[5]HUANG L, QU X. Improving traffic signal control operations using proximal policy optimization[J]. IET Intelligent Transport Systems, 2023, 17(3): 592-605.

[6]WEI H, CHEN C, ZHENG G, et al. Presslight: Learning max pressure control to coordinate traffic signals in arterial network[C]//Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, New York: ACM, 2019: 1290-1298.

[7]JIANG Q, QIN M, SHI S, et al. Multi-agent reinforcement learning for traffic signal control through universal communication method[J]. arXiv Preprint arXiv: 2204.12190, 2022.

[8]NISHI T, OTAKI K, HAYAKAWA K, et al. Traffic signal control based on reinforcement learning with graph convolutional neural nets[C]//2018 21st International Conference on Intelligent Transportation Systems (ITSC), Piscataway: IEEE, 2018: 877-883.

[9] WEI H, XU N, ZHANG H, et al. Colight: Learning network-level cooperation for traffic signal control[C]// Proceedings of the 28th ACM International Conference on Information and Knowledge Management, New York: ACM, 2019: 1913-1922.

[10] ZHANG Y, YU Z, ZHANG J, et al. Learning decentralized traffic signal controllers with multi-agent graph reinforcement learning[J]. IEEE Transactions on Mobile Computing, 2023, 23(6): 7180-7195.

[11] JUNG J, KIM I, YOON J. EcoMRL: Deep reinforcement learning-based traffic signal control for urban air quality [J]. International Journal of Sustainable Transportation, 2025, 19(8): 720-729.

[12] ZHANG X, FAN X, YU S, et al. Intersection signal timing optimization: A multi-objective evolutionary algorithm[J]. Sustainability, 2022, 14(3): 1506.

[1]	安琨, 贾作宁. 运营扰动下考虑排班一致性的电动公交调度方法研究[J]. 交通运输系统工程与信息, 2026, 26(3): 1-13.
[2]	刘家林, 许志然, 姬浩, 贾斌, 张萌, 苏兵. 考虑灵活车道分配的新型混合交通协同疏散优化研究[J]. 交通运输系统工程与信息, 2026, 26(3): 25-35.
[3]	袁泉, 潘瑞煦, 梁星宇, 李卓雅, 杨超. 货运社区视角下城市货运介观功能定位与形成机理解析[J]. 交通运输系统工程与信息, 2026, 26(3): 60-71.
[4]	宫磊, 黄鹏鹏, 雷天, 罗钦. 出租车轨迹数据驱动的灵活式公交基准线路规划方法[J]. 交通运输系统工程与信息, 2026, 26(3): 124-133.
[5]	陈越, 贾顺平, 季千喜, 代斯薇, 许奇. 地铁站点类型视角下建成环境对共享单车接驳比例的影响[J]. 交通运输系统工程与信息, 2026, 26(3): 134-143.
[6]	张建华, 公佳豪, 张文会. 智能网联汽车复用公交车道协同控制研究[J]. 交通运输系统工程与信息, 2026, 26(3): 226-234.
[7]	胡宝雨, 王宏轩, 景维鹏. 考虑车内换乘的多线路模块化自动驾驶公交调度优化[J]. 交通运输系统工程与信息, 2026, 26(3): 259-273.
[8]	武慧荣, 盛椿婷, 郭方成. 考虑扇区化与紧凑性约束的客运枢纽定制公交线路优化[J]. 交通运输系统工程与信息, 2026, 26(3): 274-285.
[9]	胡三根, 文炫淇, 巫威眺, 李满琳, 韩霜. 基于社交媒体数据的需求响应公交公众认知与态度研究[J]. 交通运输系统工程与信息, 2026, 26(3): 327-337.
[10]	刘少博, 苏蔚. 基于强化学习的地铁站客流动态管控策略研究[J]. 交通运输系统工程与信息, 2026, 26(3): 338-347.
[11]	唐立, 何彪, 唐昕琛, 王焜. 融合速度障碍与人工势场法的电动垂直起降飞行器路径规划[J]. 交通运输系统工程与信息, 2026, 26(3): 348-359.
[12]	齐嫣然, 张翕然, 李正中, 陈绍宽, 赵疆昀. 基于跨线灵活编组的多交路列车运行图优化研究[J]. 交通运输系统工程与信息, 2026, 26(3): 360-370.
[13]	王静, 雷德明, 翟静, 陈淑楣, 刘林凡. 数据与决策协同的长江航运多式联运优化研究[J]. 交通运输系统工程与信息, 2026, 26(2): 11-23.
[14]	阎桑慧宇, 马瑞, 李健. 城市暴雨内涝下考虑地铁联合疏散的公交应急路线优化[J]. 交通运输系统工程与信息, 2026, 26(2): 72-80.
[15]	谷远利, 宇泓儒, 陈龙, 邓社军, 陆文琦. 网联自动驾驶车辆专用车道动态宏微观协同部署方法[J]. 交通运输系统工程与信息, 2026, 26(2): 125-136.

面向多目标协调的细粒度交通信号智能体建模与控制方法

Fine-Grained Traffic Signal Agent Modeling and Control Method for Multi-objective Coordination

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics