基于多Agent技术的城市道路交通拥堵协调控制研究
发布时间：2018-09-13 12:55
【摘要】：随着国民经济的飞速发展和城市化进程步伐的加快,交通对城市经济发展和人民生活水平的提高的影响显得愈发重要。由于城市的道路建设难以跟上交通发展的速度,车辆的急剧增加和道路的增长呈现巨大矛盾,交通拥堵的严重程度也随之急剧增大。解决交通拥堵的主要办法是：以交通管理设施为出发点,科学组织、科学管控道路交通流,使其最大程度地有序流动,将路网交通的通行潜力发挥至最大。鉴于传统交通控制方式交通的局限性和交通规模的复杂性,人们运用智能水平更高的技术手段来解决交通拥堵问题成为急切需求。首先,本文介绍了交通拥堵的定义、分类及演化规律,并对交通信号控制的分类、基本参数和效益评价指标进行了阐述。同时指出了当前世界的几种主流信号控制系统及智能控制方法。其次,本文与人工智能技术领域的最新理论成果—多智能体技术(Multi-Agent System,MAS)相结合,从基于多Agent技术的控制策略上来研究探讨城市道路偶发性交通拥堵的协调控制。文章主要分析了Agent的特征属性、结构模型。同时还探讨了基于多Agent的分布式交通控制系统结构,结合该结构的优点,指出拥堵协调控制中引入多Agent技术的可行性与优越性。再次,结合多Agent技术,本文以单路口Age nt为控制单元,对其组成和结构进行了分析。并主要研究以遗传算法(GA)为基础的单路口Agent发生交通拥堵时,其信号优化控制的分布式结构模型,目标函数定为路口车辆的平均延误,运用实数编码方式的GA实现对单路口的拥堵优化控制。分析实例表明,经过优化的配时方案比现控制方式优越,控制效果较好。最后,本文运用博弈论的相关理论知识,通过分析描述两个路口的问题,探索研究了简单的两个路口Agent的博弈过程及协调机理。结合实际交通情况来优化两路口的控制策略,并通过实例将优化后的性能指标与现有控制方式进行分析对比,进而论证Agent间协调机制有效、可行,达到了预期的效果。
[Abstract]:With the rapid development of national economy and the acceleration of urbanization, the influence of traffic on the development of urban economy and the improvement of people's living standard is becoming more and more important. Because the urban road construction is difficult to keep up with the speed of traffic development, the sharp increase of vehicles and the growth of roads present a huge contradiction, and the severity of traffic congestion also increases rapidly. The main solution to traffic congestion is to organize and control the road traffic flow scientifically, to maximize the orderly flow, and to maximize the traffic potential of the road network by taking the traffic management facilities as the starting point, scientifically organizing and managing the road traffic flow scientifically. In view of the limitation of the traditional traffic control mode and the complexity of the traffic scale, it is urgent for people to solve the traffic congestion problem by using the technology with higher intelligence level. Firstly, this paper introduces the definition, classification and evolution of traffic congestion, and expounds the classification, basic parameters and benefit evaluation index of traffic signal control. At the same time, several mainstream signal control systems and intelligent control methods in the world are pointed out. Secondly, combined with the latest theoretical achievement of artificial intelligence technology, multi-agent technology (Multi-Agent System,MAS), the coordinated control of accidental traffic congestion on urban roads is studied from the control strategy based on multi-Agent technology. This paper mainly analyzes the characteristic attribute and structure model of Agent. At the same time, the structure of distributed traffic control system based on multiple Agent is discussed. Combining the advantages of this structure, the feasibility and superiority of introducing multi-Agent technology into congestion coordination control are pointed out. Thirdly, combined with multiple Agent technology, the composition and structure of single intersection Age nt are analyzed in this paper. This paper mainly studies the distributed structure model of signal optimization control in single intersection Agent traffic jam based on genetic algorithm (GA). The objective function is defined as the average delay of intersection vehicle. The real coding GA is used to realize the optimal control of single intersection congestion. The analysis example shows that the optimized timing scheme is superior to the current control method and the control effect is better. Finally, by using the relevant theory of game theory and analyzing and describing the problems of two junctions, this paper explores the game process and coordination mechanism of Agent in two simple junctions. Combined with the actual traffic situation, the control strategy of the two junctions is optimized, and the optimized performance index is analyzed and compared with the existing control methods through an example, which proves that the coordination mechanism between Agent is effective and feasible, and achieves the expected effect.
【学位授予单位】：长沙理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.54
，
本文编号：2241233
[Abstract]:With the rapid development of national economy and the acceleration of urbanization, the influence of traffic on the development of urban economy and the improvement of people's living standard is becoming more and more important. Because the urban road construction is difficult to keep up with the speed of traffic development, the sharp increase of vehicles and the growth of roads present a huge contradiction, and the severity of traffic congestion also increases rapidly. The main solution to traffic congestion is to organize and control the road traffic flow scientifically, to maximize the orderly flow, and to maximize the traffic potential of the road network by taking the traffic management facilities as the starting point, scientifically organizing and managing the road traffic flow scientifically. In view of the limitation of the traditional traffic control mode and the complexity of the traffic scale, it is urgent for people to solve the traffic congestion problem by using the technology with higher intelligence level. Firstly, this paper introduces the definition, classification and evolution of traffic congestion, and expounds the classification, basic parameters and benefit evaluation index of traffic signal control. At the same time, several mainstream signal control systems and intelligent control methods in the world are pointed out. Secondly, combined with the latest theoretical achievement of artificial intelligence technology, multi-agent technology (Multi-Agent System,MAS), the coordinated control of accidental traffic congestion on urban roads is studied from the control strategy based on multi-Agent technology. This paper mainly analyzes the characteristic attribute and structure model of Agent. At the same time, the structure of distributed traffic control system based on multiple Agent is discussed. Combining the advantages of this structure, the feasibility and superiority of introducing multi-Agent technology into congestion coordination control are pointed out. Thirdly, combined with multiple Agent technology, the composition and structure of single intersection Age nt are analyzed in this paper. This paper mainly studies the distributed structure model of signal optimization control in single intersection Agent traffic jam based on genetic algorithm (GA). The objective function is defined as the average delay of intersection vehicle. The real coding GA is used to realize the optimal control of single intersection congestion. The analysis example shows that the optimized timing scheme is superior to the current control method and the control effect is better. Finally, by using the relevant theory of game theory and analyzing and describing the problems of two junctions, this paper explores the game process and coordination mechanism of Agent in two simple junctions. Combined with the actual traffic situation, the control strategy of the two junctions is optimized, and the optimized performance index is analyzed and compared with the existing control methods through an example, which proves that the coordination mechanism between Agent is effective and feasible, and achieves the expected effect.
【学位授予单位】：长沙理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.54
，
本文编号：2241233
本文链接：https://www.wllwen.com/kejilunwen/jiaotonggongchenglunwen/2241233.html
