基于多路口动态流量的区域控制研究

发布时间：2018-02-11 17:39

本文关键词： 城市交通分布式多智能体人工代谢算法 Dijkstra VISSIM　出处：《西华大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着城市化建设的日益发展，越来越多的车辆加入到城市交通中，城市交通也因此变得越来越复杂。尤其是大城市，公民的出行已经成为一个很棘手的问题。目前，关于城市智能交通的研究有很多，主要的算法有遗传算法、神经网络和蚁群算法等。但是这些算法所消耗的时间成本和空间成本过大，对于实时性要求高的城市交通来说，这不是最佳选择。本文采用分布式和多智能体的设计方法来设计交通模型，模型中包含五种不同类型的智能体，通过这些不同类型的智能体来描述交通场景中的实体。每个智能体通过与相邻的智能体进行通信，获取自身所需要的交通数据。对于一个包含多个路口的交通区域，，本文采用两种不同的算法进行单路口控制和全局控制。首先对传统的人工代谢算法进行改进并以此计算单个路口的信号方案，然后采用Dijkstra算法对交通区域进行区域协调控制；传统的Dijkstra算法主要用于计算带权图中一个顶点到其他顶点的最短路径，这里对Dijkstra算法进行一定的改进。本文采用交通仿真工具VISSIM和Visual Basic对一个包含九个双向四车道的十字路口的交通区域进行仿真。VISSIM可以模拟真实交通场景并在仿真后获取相关的交通参数：车辆延时、等待时间和车辆速度等，通过分析这些交通参数来对本文提出的控制方法进行评价。仿真结果表明，对于不同的交通流量，本文提出的交通控制方法在解决区域交通的拥堵问题比定时控制显得更为高效。
[Abstract]:With the development of urbanization, more and more vehicles are added to the urban traffic, and the urban transportation becomes more and more complicated. Especially in the big cities, the travel of citizens has become a very difficult problem. There are many researches on urban intelligent transportation, such as genetic algorithm, neural network and ant colony algorithm. This is not the best option. In this paper, a distributed and multi-agent design method is used to design the traffic model, which includes five different types of agents. Entities in traffic scenarios are described by these different types of agents. Each agent communicates with an adjacent agent to obtain the traffic data it needs. For a traffic area with multiple intersections, In this paper, two different algorithms are used for single intersection control and global control. Firstly, the traditional artificial metabolism algorithm is improved to calculate the signal scheme of a single intersection, and then the Dijkstra algorithm is used to control the traffic area. The traditional Dijkstra algorithm is mainly used to calculate the shortest path from one vertex to other vertices in weighted graph. The Dijkstra algorithm is improved in this paper. In this paper, the traffic simulation tools VISSIM and Visual Basic are used to simulate a traffic area with nine two-way four-lane intersections. VisSIM can simulate the real traffic scene and obtain the relevant traffic parameters: vehicle delay. The waiting time and vehicle speed are analyzed to evaluate the control methods proposed in this paper. The simulation results show that, for different traffic flow, The traffic control method proposed in this paper is more efficient than timing control in solving regional traffic congestion.
【学位授予单位】：西华大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.54

【相似文献】