Scholarlink

<p>Rapid urban traffic growth increases congestion and emissions, highlighting the urgent need for effective traffic light control to improve mobility and network performance. This paper proposes a novel approach to traffic light management based on polling systems with non-exhaustive service. Two queueing models are investigated: M/M/1 with Binomial Gated Service and Multiple Vacation, and M/M/1 with Gated Service and Single Vacation, both designed to optimize signal timing at isolated urban intersections. In the proposed system, the number of vehicles permitted to pass during each green phase is dynamically determined according to queue length and average waiting time, allowing green time to be extended on the most congested approaches. Key performance metrics were identified and analyzed through numerical evaluation to compare models efficiency. Furthermore, a bi-criteria optimization problem was formulated to determine the optimal traffic light plan under varying urban traffic conditions, solved using a genetic algorithm. The models were validated using the SUMO (Simulation of Urban MObility) platform. Results show that the M/M/1 (Binomial Gated, Multiple Vacation) model consistently outperforms both the alternative polling model and a conventional fixed-time traffic signal plan, by achieving better balance between phases, reducing queue lengths, and significantly decreasing vehicle sojourn time. This research provides a practical framework for improving traffic efficiency at isolated intersections, contributing to more sustainable urban mobility and lower environmental impacts.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Innovative Traffic Light Management: Leveraging Polling Systems for Smart Control

  • 期刊论文

  • 发表日期：

    2026年4月24日

查看全文
  • Sabiha Larbi,
  • Fazia Rahmoune,
  • Mohammed Said Radjef,
  • Zohra Aoudia

摘要

Rapid urban traffic growth increases congestion and emissions, highlighting the urgent need for effective traffic light control to improve mobility and network performance. This paper proposes a novel approach to traffic light management based on polling systems with non-exhaustive service. Two queueing models are investigated: M/M/1 with Binomial Gated Service and Multiple Vacation, and M/M/1 with Gated Service and Single Vacation, both designed to optimize signal timing at isolated urban intersections. In the proposed system, the number of vehicles permitted to pass during each green phase is dynamically determined according to queue length and average waiting time, allowing green time to be extended on the most congested approaches. Key performance metrics were identified and analyzed through numerical evaluation to compare models efficiency. Furthermore, a bi-criteria optimization problem was formulated to determine the optimal traffic light plan under varying urban traffic conditions, solved using a genetic algorithm. The models were validated using the SUMO (Simulation of Urban MObility) platform. Results show that the M/M/1 (Binomial Gated, Multiple Vacation) model consistently outperforms both the alternative polling model and a conventional fixed-time traffic signal plan, by achieving better balance between phases, reducing queue lengths, and significantly decreasing vehicle sojourn time. This research provides a practical framework for improving traffic efficiency at isolated intersections, contributing to more sustainable urban mobility and lower environmental impacts.