基于多智能体分群同步的城市路网交通控制

在城市交通网络中,经常出现某些路段较为拥堵,而其它路段内仍有空间未被充分利用.为了解决该问题,本文提出一种路网分群一致算法.首先,以路段的空间占有率为状态,建立交通网络状态空间模型,描述路网中车流的传递关系;进一步,将路网中的各条路段抽象为智能体,提出路网多智能体分群一致算法,将路段集划分为若干个非空不交子集,每个子集内的路段状态可达一致,不同子集的一致状态不同.从而,该算法可使各路段的空间占有率达到均衡,减轻局部拥堵,减少车辆延误时间.最后,采用VISSIM交通仿真软件对北京市望京地区的实际网络进行仿真,仿真结果说明了本文方法的可行性和有效性.

Urban traffic network control based on cluster consensus of multi-agent systems

Traffic jam often occurs in some links within urban traffic network(UTN), however, there exist much space in other links not being utilized sufficiently. Motivated by this problem, cluster consensus algorithm of UTN is proposed in the paper. Firstly, the space occupancy of the link is chosen as state of UTN, and then the state-space model of UTN is established, which describes transition relation of vehicles on links. Furthermore, each of links within UTN is abstracted as an agent, and then cluster consensus algorithm of multi-agent system of UTN is provided, i.e., the set of links is divided into some nonempty and disjoint subsets, in each of which the space occupancy of links can reach consensus, and the consensus states are not necessarily equal. Thus, the proposed algorithm can balance the space occupancy on links, which relieves traffic jam, and decreases delay of vehicles. At last, traffic simulation software VISSIM is used to analyze the real-world road network in Wangjing district in Beijing, and simulation results show the feasibility and effectiveness of the proposed methods.