预测式交通管理预案研究

发布时间：2018-10-22 18:16

【摘要】：随着城市不断发展,道路资源有限的情况下,机动车数量不断增加,城市交通环境越来越恶化,交通拥堵成了许多城市发展中亟需解决的问题。交通拥堵给交通参与者带来时间和经济上的损失,同时慢速行驶中,汽车排放的尾气和产生的噪音,使得城市环境质量下降。在对待此问题上,越来越多的交通管理措施被提出。预测式交通管理预案可看作一种比较有效的解决城市交通拥堵的方法,该方法利用当前交通流状况,对已经形成或即将形成的交通拥堵进行准确、及时的预测,提前进行交通管理方案的制定,为交通管理部门解决交通拥堵提供策略,可有效地降低交通拥堵带来的负面影响。本文中的预测式交通管理属于ITS(智能交通系统)的子系统,采用中观交通仿真技术,结合动态交通分配理论与交通管理措施思想,借助动态交通软件DYNAMEQ平台,构建仿真模型,进行仿真并推演得到未来时段内的交通运行状况。对未来时段交通状况进行分析,进而提出交通管理预案,提前对拥堵区域进行组织优化。本文主要研究了下面两个方面的内容：一、考虑到目前时变OD矩阵的获取难度,针对开源的中观交通分配软件NEXTA-DTALite中OD反推模型求解方法的不足,首先对NEXTA-DTALite软件OD矩阵反推模型进行变换,构建新的目标函数,使用开源的线性规划软件GLPK进行求解。利用开源数据格式文件CSV文件的通用读取和输出的特点,实现两个软件在数据交换方面达到无缝衔接。提出了利用NEXTA-DTALite和GLPK进行OD反推模型求解的算法步骤,解决了NEXTA软件中利用EXCEL求解遇到的维数爆炸问题,同时提高了NEXTA-DTALite软件中OD反推模块的开源程度。利用此算法求得的OD矩阵作为DYNAMEQ平台的初始矩阵。二、利用DYNAMEQ平台,对长沙某区域路网进行模型的搭建,经过模型校正准确后,对仿真结果进行详细分析,提取重要MOE(绩效指标值),找出拥堵原因。通过改变关键交叉口的信号配时方案进行优化仿真,对比前后仿真结果,可知MOE值得到优化,验证了该交通管理方案的有效性。
[Abstract]:With the continuous development of cities and the limited road resources, the number of motor vehicles is increasing and the urban traffic environment is getting worse and worse. Traffic congestion has become an urgent problem in the development of many cities. Traffic congestion brings time and economic loss to traffic participants, and at the same time, the exhaust gas and noise produced by vehicles in slow speed reduce the quality of urban environment. In dealing with this problem, more and more traffic management measures have been proposed. Predictive traffic management plan can be regarded as an effective method to solve urban traffic congestion. This method makes use of the current traffic flow situation to accurately and timely predict the traffic congestion that has been formed or is about to be formed. The formulation of traffic management plan in advance provides the traffic management department with strategies to solve traffic congestion, which can effectively reduce the negative impact of traffic congestion. The predictive traffic management in this paper belongs to the subsystem of ITS (Intelligent Transportation system). With the help of dynamic traffic assignment theory and traffic management measures, the simulation model is constructed by means of dynamic traffic distribution theory and traffic management measures. Carry on the simulation and extrapolate the traffic running condition in the future time period. The traffic situation in the future period is analyzed and traffic management plan is put forward to optimize the congestion area ahead of time. This paper mainly studies the following two aspects: first, considering the difficulty of obtaining the time-varying OD matrix, the method of solving the OD back-push model in the open source traffic distribution software NEXTA-DTALite is insufficient. Firstly, the inverse model of OD matrix of NEXTA-DTALite software is transformed, and a new objective function is constructed, which is solved by the open source linear programming software GLPK. By using the common reading and output characteristics of open source data format file (CSV), the two softwares are seamlessly connected in data exchange. This paper presents an algorithm for solving OD backstepping model using NEXTA-DTALite and GLPK, solves the dimension explosion problem encountered by using EXCEL in NEXTA software, and improves the open source degree of OD back-push module in NEXTA-DTALite software. The OD matrix obtained by this algorithm is used as the initial matrix of the DYNAMEQ platform. Secondly, using DYNAMEQ platform, the model of a regional road network in Changsha is built. After the model correction is accurate, the simulation results are analyzed in detail, and the important MOE (performance index) is extracted to find out the congestion reasons. By changing the signal timing scheme of the key intersection and comparing the simulation results before and after, we can see that the MOE value is optimized, which verifies the effectiveness of the traffic management scheme.
【学位授予单位】：长沙理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491

【相似文献】