车辆自组织网络环境下动态路径诱导系统的建模与优化策略研究

发布时间：2018-01-07 11:09

本文关键词：车辆自组织网络环境下动态路径诱导系统的建模与优化策略研究　出处：《山东大学》2014年博士论文　论文类型：学位论文

【摘要】：随着城市化进程的加快和汽车工业的发展,现有城市道路的通行能力与不断增长的交通需求之间的矛盾变得日益尖锐,交通拥堵现象日益突出。应用智能交通系统对交通流进行有效地控制与诱导,以缓解交通拥堵,提供畅通和有序的交通环境,是目前各个国家面对交通问题所采用的重要手段。车辆自组织网络(Vehicular Ad Hoc Networks,VANET)作为一个新兴的研究领域,逐渐受到人们的关注。现有的交通管理系统大都利用交通控制中心处理路况信息,车辆节点只作为信息的接收方。利用VANET实现车辆间的相互通信,能够使车辆节点获得实时的路况信息,动态调整行驶路线,从而实现动态路径选择,有效地提高了行驶效率。本文以大规模城市复杂路网作为研究背景,采用前景理论、博弈论、蜜蜂群体优化策略等理论和方法,研究VANET环境下动态路径诱导系统的建模与优化策略,通过划分交通诱导小区、限制搜索区域、建立路径选择模型、优化路径寻优算法等一系列手段,建立一套全面高效的动态路径诱导系统。主要研究工作如下：论文首先论述了课题的研究目的与意义,然后分别介绍了车辆自组织网络和动态路径诱导系统的国内外研究现状,分析了车辆自组织网络的发展与应用前景,以及目前动态路径诱导系统存在的问题。在此基础上,提出了一系列的动态路径诱导模型与优化策略。针对交通状况的复杂性、时变性、不确定性,本文建立了一种基于前景理论的路径选择模型。模型考虑了由于出行者的主观能动性而导致的不完全理性交通行为特征,能够一定程度上克服期望效用理论的不足,准确地描述不确定交通条件下出行者的决策过程,更接近于出行者的实际行为模式。结合实际课题研究的需要,为克服Dijkstra算法在复杂路网中难以满足动态路径诱导实时性要求的不足,提出了一种阈值限制搜索区域的动态路径诱导算法,并成功应用于复杂城市交通路网的动态路径诱导模型系统,实现了城市道路交通系统的实时动态最优路径搜索。该算法在Dijkstra算法的基础上,引入阈值限制搜索区域的搜索机制,根据实际路网的空间分布特性,通过合理设置搜索区域的限定阈值,有效缩减了算法的搜索规模,提高了算法的运算效率,适合于复杂城市路网动态诱导系统的最优路径实时搜索。为了克服车辆自组织网络中车辆间的无线通信受到距离等因素的影响,提出了一种基于Shapley值的动态交通小区博弈划分方法,为路径诱导系统提供合理有效的交通子区。该方法以路网中的路段为基本单位,利用相似度确定交通小区核,并将交通小区划分过程抽象为博弈过程,以Shapley值作为博弈收敛判据,进行博弈迭代直至收敛,从而实现交通小区的划分。算法能够准确可靠的实现交通小区的划分,同时降低了交通诱导系统路径寻优的复杂度,提升了诱导系统的实时性能。在本文提出的交通诱导小区基础上,进一步提出了一种蜜蜂启发式分区路径诱导算法,为路径诱导系统提供合理有效的最优路径。算法模拟生物系统中的蜜蜂觅食现象,对各诱导小区同步并行进行路径寻优,从而快速准确获取最优路径,并能保证算法提供全局最优路径。在前述研究的基础上,本文对通勤交通行为带来的城市道路拥堵问题进行了研究。针对通勤行为的特点,以预测信息精度和路况熟悉程度作为参数,提出了一种基于博弈论的路径选择模型,为通勤者提供可靠的路径选择方案。该模型提供的最优路径能够准确描绘通勤者日常的路径选择决策过程,能够提供给通勤者可靠的通勤路径。最后设计并开发了基于中间件的动态路径诱导系统仿真平台,实现了交通信息的发布与动态路径诱导功能,同时也为本文建立的模型和提出的算法提供一个可靠的仿真环境,为验证模型的有效性和算法的先进性提供有力的技术支持。基于上述研究成果,论文最后对全文进行了总结,并对下一步的研究工作进行了展望,为后续的研究提供了可能的方向。
[Abstract]:With the development of city and accelerate the process of automobile industry, the contradiction between the traffic demand of existing city road traffic capacity and increasing the increasingly sharp, traffic congestion has become increasingly prominent. The application of intelligent transportation system to effectively control and guidance on the traffic flow, to alleviate traffic congestion, provide smooth and orderly traffic environment. Is an important means of various countries face the traffic problems.
Vehicle self-organizing network (Vehicular Ad Hoc Networks, VANET) as a new research field, has gradually attracted people's attention. Most of the traffic control center to deal with the traffic information and traffic management system of the existing vehicle nodes only as the receiver of information. To achieve mutual communication between vehicles using VANET, can make the vehicle obtain real-time node traffic information, dynamic adjustment of the route, in order to achieve dynamic path selection, effectively improves the running efficiency.
Based on the large-scale city complex network as the research background, based on prospect theory, game theory, bee colony optimization strategy theory and method of dynamic route guidance system under the environment of VANET modeling and optimization strategy, by dividing the traffic induced area, limiting the search region, establish a route choice model, optimal path optimization algorithm and a series of means to establish a comprehensive and efficient dynamic route guidance system. The main research work is as follows:
This paper first discusses the purpose and significance of the subject, and then introduces the research status of domestic ad hoc networks and dynamic route guidance system of vehicles, analyzes the development and application prospect of vehicular ad hoc networks, and the dynamic route guidance system problems. On this basis, put forward a series of dynamic path induced model and optimization strategy.
According to the traffic condition complexity, time-varying, uncertainty, this paper established a path choice model based on prospect theory. The model considers the irrational traffic behavior caused by the subjective initiative of travelers, to a certain extent overcome the expected utility theory, accurately describe uncertain traffic conditions the decision-making process for passengers, more close to the practice of travelers.
According to the actual needs of this research, to overcome the Dijkstra algorithm in the complex network is difficult to meet the requirement of real-time dynamic route guidance problem, put forward a threshold limit induced by dynamic path searching area algorithm, and successfully applied to the complex dynamic path of city traffic guidance model system, realize the real-time dynamic optimal path of city road traffic the search. The algorithm based on Dijkstra algorithm, by introducing the threshold mechanism of restricted searching area, according to the distribution characteristics of actual network space, by reasonably setting the search area limit threshold, effectively reduces the size of the search algorithm to improve the computing efficiency of the algorithm, suitable for complex dynamic system optimal induction of city road network the real-time path search.
In order to overcome the vehicles affected by factors such as distance from the wireless communication network between vehicles, presents a Shapley dynamic traffic zone division method based on game theory, provide a reasonable and effective traffic sub district for route guidance system. In this method, the road network in the road as the basic unit, determine the traffic area using kernel similarity. And the traffic zone division is the process of abstraction as a game process, using Shapley as the convergence criterion of game, game iteration until convergence, so as to realize the traffic zone division algorithm. The algorithm can accurately and reliably realize the traffic zone division, while reducing the complexity of traffic guidance system path optimization, enhance the real-time performance of the guidance system.
Based on the proposed cell induced traffic, further proposes a heuristic partition bee routing algorithm, provides the optimal path for the reasonable and effective route guidance system. Foraging phenomena in biological systems bee simulation algorithm, the induction of cell synchronization and walk path optimization, which quickly and accurately obtain the optimal path, and to ensure the algorithm provides a global optimal path.
On the basis of previous research, this paper brings to the city road congestion of commuter traffic behavior was studied. According to the characteristics of commuting behavior, in order to predict the information accuracy and familiarity with the road as a parameter, put forward a path choice model based on game theory, provide a reliable path for commuter optimal path selection scheme. This model can accurately describe the decision-making process of the daily commuter route selection, to provide commuter commuter path reliable.
Finally, the design and development of the dynamic route guidance system simulation platform based on middleware, to achieve the traffic information release and dynamic route guidance function, but also for the model and the proposed algorithm provides a reliable simulation environment, provides a powerful technical support for advanced to verify the validity of the model and algorithm.
Based on the above research results, the final part of the paper is summarized, and the next research work is prospected, which provides a possible direction for further research.

【学位授予单位】：山东大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：U495

【参考文献】