智能交通中动态路径诱导系统的建模与优化算法的研究

发布时间：2018-11-28 17:52

【摘要】：随着城市现在化进程的不断发展,全球汽车保有量逐年上升,从而引发了一系列的社会问题,例如:交通堵塞、环境污染以及噪声污染等等。智能交通系统,简称ITS(Intelligent transportation system)应运而生,智能交通系统是集信息、通信、控制及网络等技术于一体的综合研究学科,可以提供全方位、实时、准确以及高效的服务信息。ITS是具有潜力的研究方向,进一步说,将成为未来相关研究领域的热点。动态路径诱导系统(Dynamic Route Guidance System,DRGS)是ITS一个重要分支,利用计算机、通信等现代技术,为出行者提供实时交通信息以及最优路径。路径选择模型和路径优化算法是动态路径诱导系统的关键技术,模型可以确定动态路径诱导系统的优化目标,优化算法可以提高路径诱导系统的效率,从而满足实时交通环境。因此,本文分别在建模以及优化算法两方面对动态路径诱导系统进行了研究。1.路径选择模型在实时环境下,影响出行者选择路径的因素比较多,例如:费用、时间、距离等,但目前大多数路径选择模型是基于单目标的。鉴于此,本文分别从费用、时间角度出发,构建了实时环境下基于多目标的路径选择模型。在实时环境下,文中具体分析了道路畅通和拥挤的情形下路阻的产生方式,从而构建了基于时间最短的目标函数,在此基础上,构造了基于费用最低的路径选择模型。采用加权求和函数对多维数据聚集得到组合权重,而权重系数可依据出行者需要或喜好设定。为更精确地反馈结果,本文用类似的量值对多目标函数值进行了预处理,从而均衡各决策变量在组合函数中的支配能力。为验证模型的实用价值,本文以C++仿真实验环境,实验数据以VISSIM仿真并结合实际进行了合理的设定。在仿真环境下,本文模型与基于几何最短距离的路径选择模型进行对比实验,实验结果证明实时环境下基于多目标的路径选择模型更具有实用价值。2.路径优化算法在动态路径诱导系统中,由于交通信息具有实时、动态的特性,因此,本文提出了实时环境下基于混合的动态路径优化算法。该算法是在广义自适应A*算法的基础上,结合剪枝算法,同时引入PSO局部最优及全局最优智能存储、模糊时间窗等优化策略。剪枝算法以当前局部最优为阈值,从而有效控制阈值的大小,避免阈值过大或过小的结果;模糊时间窗约束算法优化时间以及控制仿真时间,从而使算法更好的适应系统的实时性。为验证算法,实验数据采用纽约地图数据,并在仿真实验环境下,分别验证了优化策略的有效性,与此同时,本文算法与A*算法的进行了对比实验。实验结果证明:本文算法中,优化策略在动态路径优化算法是有效的且合理的,同时证明了本文算法可适应于动态路径优化系统。
[Abstract]:With the development of the city, the number of cars in the world is increasing year by year, which leads to a series of social problems, such as traffic jam, environmental pollution and noise pollution and so on. Intelligent Transportation system (ITS (Intelligent transportation system) emerges as the times require. Intelligent Transportation system (its) is a comprehensive research subject which integrates information, communication, control and network technologies. It can provide all-round, real-time, and so on. Accurate and efficient service information. ITS is a potential research direction, further, it will become a hot spot in related research fields in the future. Dynamic path guidance system (Dynamic Route Guidance System,DRGS) is an important branch of ITS. It provides real-time traffic information and optimal path for travelers using modern technologies such as computers and communications. Path selection model and path optimization algorithm are the key technologies of dynamic path guidance system. The model can determine the optimization goal of the dynamic path guidance system, and the optimization algorithm can improve the efficiency of the route guidance system and meet the real-time traffic environment. Therefore, the dynamic path guidance system is studied in the modeling and optimization algorithms. 1. In the real-time environment, there are many factors that affect the traveler's choice of path, such as cost, time, distance, etc. However, most of the current path selection models are based on single objective. In view of this, this paper constructs a path selection model based on multi-objective in real time environment from the point of view of cost and time. In the real time environment, the paper analyzes the way of road obstruction in the case of smooth and congested roads, and then constructs the objective function based on the shortest time. On this basis, a path selection model based on the lowest cost is constructed. The weighted summation function is used to aggregate the multidimensional data to obtain the combined weight, and the weight coefficient can be set according to the traveler's needs or preferences. In order to feedback the results more accurately, this paper preprocesses the values of multi-objective functions with similar quantities, so as to balance the dominance of each decision variable in the composite function. In order to verify the practical value of the model, the experiment data is simulated by VISSIM in C simulation environment. In the simulation environment, the model is compared with the path selection model based on geometric shortest distance. The experimental results show that the path selection model based on multi-objective in real-time environment has more practical value. 2. Due to the real-time and dynamic characteristics of traffic information in dynamic path guidance systems, a hybrid dynamic path optimization algorithm is proposed in this paper. The algorithm is based on the generalized adaptive A * algorithm combined with pruning algorithm and introduces PSO local and global optimal intelligent storage fuzzy time window and other optimization strategies. The pruning algorithm takes the current local optimum as the threshold value, thus effectively controlling the size of the threshold value and avoiding the result that the threshold is too large or too small. The fuzzy time window constraint algorithm optimizes the time and controls the simulation time, which makes the algorithm better adapt to the real time of the system. In order to verify the algorithm, the experimental data are based on the New York map data, and the effectiveness of the optimization strategy is verified in the simulation environment. At the same time, the algorithm is compared with the A * algorithm. The experimental results show that the optimization strategy is effective and reasonable in the dynamic path optimization algorithm, and it is also proved that the proposed algorithm is suitable for the dynamic path optimization system.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U495;TP301.6

【相似文献】