This study introduces CycLight, a novel cycle-level deep reinforcement learning (RL) approach for network-level adaptive traffic signal control (NATSC) systems. Unlike most traditional RL-based traffic controllers that focus on step-by-step decision making, CycLight adopts a cycle-level strategy, optimizing cycle length and splits simultaneously using Parameterized Deep Q-Networks (PDQN) algorithm. This cycle-level approach effectively reduces the computational burden associated with frequent data communication, meanwhile enhancing the practicality and safety of real-world applications. A decentralized framework is formulated for multi-agent cooperation, while attention mechanism is integrated to accurately assess the impact of the surroundings on the current intersection. CycLight is tested in a large synthetic traffic grid using the microscopic traffic simulation tool, SUMO. Experimental results not only demonstrate the superiority of CycLight over other state-of-the-art approaches but also showcase its robustness against information transmission delays.
A Multi-intersection Vehicular Cooperative Control based on End-Edge-Cloud Computing
Beijing University of Posts and Telecommunications(北京邮电大学);Beijing Information Science and Technology University(北京信息科技大学);University of Technology Sydney(悉尼科技大学)
Network-wide traffic signal control optimization using a multi-agent deep reinforcement learning
University of Macau(澳门大学);Southeast University(东南大学);University of Hawaii at Manoa(夏威夷大学马诺阿分校);University of Delaware(特拉华大学)
低效的交通控制可能会导致许多问题,例如交通拥堵和能源浪费。提出了一种新的多智能体强化学习方法KS-DDPG(Knowledge Sharing Deep Deterministic策略梯度),通过增强交通信号灯之间的协作来实现最优控制。通过引入支持知识共享的通信协议,每个代理可以访问所有代理收集的交通环境的集体表示。该方法通过两个实验分别使用合成和真实世界的数据集进行评估。与最先进的基于强化学习和传统交通方法的比较表明,所提出的KS-DDPG在控制大规模交通网络和应对交通流量波动方面具有显着的效率。此外,引入的通信机制也被证明可以在不显著增加计算负担的情况下加快模型的收敛速度。
Communication Strategy on Macro-and-Micro Traffic State in Cooperative Deep Reinforcement Learning for Regional Traffic Signal Control
School of Advanced Technology, Xi’an Jiaotong-Liverpool University(西交利物浦大学先进技术学院);University of Liverpool(利物浦大学);Department of Engineering and Design, University of Sussex(萨塞克斯大学工程与设计系);International Business School Suzhou, Xi’an Jiaotong-Liverpool University(西交利物浦大学苏州国际商学院);Xidian University(西安电子科技大学);Department of Computer Science, City University of Hong Kong(香港城市大学计算机科学系);City University of Hong Kong Shenzhen Research Institute(香港城市大学深圳研究院)
针对区域交通信号控制中的智能体协作问题,提出基于 GAT 聚合的 GA2-Naive 与 GA2-Aug 通信模块,融合宏观/微观及区域内/间交通状态关联,并验证其可稳定提升现有框架性能。
Heterogeneous Multi-Agent Reinforcement Learning for Distributed Channel Access in WLANs
National Mobile Communications Research Laboratory, Southeast University(东南大学国家移动通信研究中心);Purple Mountain Laboratories(紫金山实验室);College of Electronics and Information Engineering, Shenzhen University(深圳大学电子与信息工程学院);Wireless Technology Lab, 2012 Laboratories, Huawei Technologies Company Ltd.(华为技术有限公司无线技术实验室);ITP Lab, the Department of Electrical and Electronic Engineering, Imperial College London(伦敦帝国理工学院电子与电气工程系ITP实验室)