Optimizing multi-agent based urban traffic signal control system

Mingtao Xu, Kun An, Hai Le Vu, Zhirui Ye, Jiaxiao Feng, Enhui Chen

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Agent-based approach is a popular tool for modelling and developing large-scale distributed systems such as urban traffic control system with dynamic traffic flows. This study proposes a multi-agent-based approach to optimize urban traffic network signal control, which utilizes a mathematical programming method to optimize the signal timing plans at intersections. To improve the overall network efficiency, we develop an online agent-based signal coordination scheme, underpinned by the communication among different intersection control agents. In addition, the initial coordination scheme that pre-adjusts the offsets between the intersections is developed based on the historical demand information. Comparison and sensitivity analysis are conducted to evaluate the performance of the proposed method on a customized traffic simulation platform using MATLAB and VISSIM. Simulation results indicate that the proposed method can effectively avoid network oversaturation and thus reduces average travel delay and improves average vehicle speed, as compared to rule-based multi-agent signal control methods.

Original languageEnglish
Pages (from-to)357-369
Number of pages13
JournalJournal of Intelligent Transportation Systems: technology, planning, and operations
Volume23
Issue number4
DOIs
Publication statusPublished - 2019

Keywords

  • mathematical programming
  • multi-agent system
  • signal control
  • traffic network

Cite this

@article{2ecb0fac8b2f4b24836e99dcf7f9a2af,
title = "Optimizing multi-agent based urban traffic signal control system",
abstract = "Agent-based approach is a popular tool for modelling and developing large-scale distributed systems such as urban traffic control system with dynamic traffic flows. This study proposes a multi-agent-based approach to optimize urban traffic network signal control, which utilizes a mathematical programming method to optimize the signal timing plans at intersections. To improve the overall network efficiency, we develop an online agent-based signal coordination scheme, underpinned by the communication among different intersection control agents. In addition, the initial coordination scheme that pre-adjusts the offsets between the intersections is developed based on the historical demand information. Comparison and sensitivity analysis are conducted to evaluate the performance of the proposed method on a customized traffic simulation platform using MATLAB and VISSIM. Simulation results indicate that the proposed method can effectively avoid network oversaturation and thus reduces average travel delay and improves average vehicle speed, as compared to rule-based multi-agent signal control methods.",
keywords = "mathematical programming, multi-agent system, signal control, traffic network",
author = "Mingtao Xu and Kun An and Vu, {Hai Le} and Zhirui Ye and Jiaxiao Feng and Enhui Chen",
year = "2019",
doi = "10.1080/15472450.2018.1501273",
language = "English",
volume = "23",
pages = "357--369",
journal = "Journal of Intelligent Transportation Systems: technology, planning, and operations",
issn = "1547-2450",
publisher = "Taylor & Francis",
number = "4",

}

Optimizing multi-agent based urban traffic signal control system. / Xu, Mingtao; An, Kun; Vu, Hai Le; Ye, Zhirui; Feng, Jiaxiao; Chen, Enhui.

In: Journal of Intelligent Transportation Systems: technology, planning, and operations, Vol. 23, No. 4, 2019, p. 357-369.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Optimizing multi-agent based urban traffic signal control system

AU - Xu, Mingtao

AU - An, Kun

AU - Vu, Hai Le

AU - Ye, Zhirui

AU - Feng, Jiaxiao

AU - Chen, Enhui

PY - 2019

Y1 - 2019

N2 - Agent-based approach is a popular tool for modelling and developing large-scale distributed systems such as urban traffic control system with dynamic traffic flows. This study proposes a multi-agent-based approach to optimize urban traffic network signal control, which utilizes a mathematical programming method to optimize the signal timing plans at intersections. To improve the overall network efficiency, we develop an online agent-based signal coordination scheme, underpinned by the communication among different intersection control agents. In addition, the initial coordination scheme that pre-adjusts the offsets between the intersections is developed based on the historical demand information. Comparison and sensitivity analysis are conducted to evaluate the performance of the proposed method on a customized traffic simulation platform using MATLAB and VISSIM. Simulation results indicate that the proposed method can effectively avoid network oversaturation and thus reduces average travel delay and improves average vehicle speed, as compared to rule-based multi-agent signal control methods.

AB - Agent-based approach is a popular tool for modelling and developing large-scale distributed systems such as urban traffic control system with dynamic traffic flows. This study proposes a multi-agent-based approach to optimize urban traffic network signal control, which utilizes a mathematical programming method to optimize the signal timing plans at intersections. To improve the overall network efficiency, we develop an online agent-based signal coordination scheme, underpinned by the communication among different intersection control agents. In addition, the initial coordination scheme that pre-adjusts the offsets between the intersections is developed based on the historical demand information. Comparison and sensitivity analysis are conducted to evaluate the performance of the proposed method on a customized traffic simulation platform using MATLAB and VISSIM. Simulation results indicate that the proposed method can effectively avoid network oversaturation and thus reduces average travel delay and improves average vehicle speed, as compared to rule-based multi-agent signal control methods.

KW - mathematical programming

KW - multi-agent system

KW - signal control

KW - traffic network

UR - http://www.scopus.com/inward/record.url?scp=85055197038&partnerID=8YFLogxK

U2 - 10.1080/15472450.2018.1501273

DO - 10.1080/15472450.2018.1501273

M3 - Article

VL - 23

SP - 357

EP - 369

JO - Journal of Intelligent Transportation Systems: technology, planning, and operations

JF - Journal of Intelligent Transportation Systems: technology, planning, and operations

SN - 1547-2450

IS - 4

ER -