基于社区发现的交通控制子区优化方法研究

摘要

传统区域交通信号控制系统对控制子区的划分未充分考虑交通网络拓扑结构的复杂特性,由工程师根据交叉口物理距离及现场交通流特性来确定,难以保证其客观性和统一性.本文以社区模块度为评价指标,利用凝聚社区发现算法实现了区域交通信号控制系统的控制子区划分；进而,以北京市望京地区为实验案例进行应用测试,利用VISSIM仿真平台比较了本文方法和传统方法在SCOOT系统控制条件下的控制效果.仿真结果表明,在高峰、平峰、低峰等不同交通需求下,区域主干道的平均旅行时间有明显下降,区域内车均停车延误、车均停车次数、车均延误、平均车速等指标均有改善,验证了方法的有效性和可行性,为区域交通信号控制系统结构优化提供了新方法.%Urban traffic control network is typical complex network. Traditional regional traffic signal control system such as SCOOT, SCATS are normally divided into sub-control areas according to link length between nodes and traffic flow. However, it does not take full consideration of the complex network topology characteristics. In this paper, with the community module degree as evaluation indexes, traffic signal control sub-areas of urban traffic signal network are divided using the community discovery algorithm. Then the PC-SCOOT system and VISS1M simulation platform are used as test-bed based on Wangjing District in Beijing, China. Traditional method and the proposed method are both used to compare traffic indexes of some main urban road and the whole region. Simulation results show that the proposed method could improve average stop delay, average stops, average delay and the average speed per vehicle.