【作者】 潘若愚；

【导师】 杨善林；

【作者基本信息】 合肥工业大学 ， 管理科学与工程， 2014， 博士

【摘要】 随着城市经济、规模的不断发展和人民生活水平的不断提高,我国大中城市的交通出行压力在不断上升。作为城市系统的重要组成部分,城市公共交通系统为城市系统内人员、物质的流动提供必不可少的物质条件,城市公交系统内交通需求与供给的不平衡将会给城市经济社会的发展带来非常大的影响。当前,优先发展城市公共交通系统、促进出行需求的多样性已经成为广大学者、政府部门公认的解决大中城市交通问题的最有效途径。因此,传统公交线网优化时仅仅考虑地面公交线路的布设,忽视其他公共交通工具的做法存在很大的局限性,单一的考虑地面公交线网的布设已经难以满足城市交通出行多样化的要求。在实际的公交线网规划时,需要综合考虑城市公交系统内地面交通、轨道交通等多方面的规划,从整体上发挥不同类型公交工具的优势,提高城市公交系统的服务效能。本文的研究主要围绕城市公共汽车交通网络规划的关键理论、技术展开,以学术意义和实用价值并重为原则,在城市公交网络流量预测、城市公交线网优化方法研究、公交线网优化设计模型与算法研究、公交线网优化效能评估研究等方面取得了一定的成果。本文的主要内容如下：(1)公共交通网络是城市发展的基础,本文在深入分析国内外研究现状的基础之上,确定了对合肥市城市公交系统的资源配置与优化应当坚持以局部调整与整体统筹的并举的原则。通过分析城市公交线网交通流量的变化,有效的掌握相应的交通出行规律,从而有利于根据实际情况制定合理的公交优化方案。本文以公交系统内站点的客流量变化为研究对象,通过深入分析其变化规律,针对不同类型的交通流量数据,分别提出了基于经验模态分解的时间序列预测模型和基于经验模态分解的支持向量机模型,并明确了不同模型下的对于公交流量序列预测的预测适用范围。(2)城市公交线网是城市公交系统的基本构成和核心。本文针对城市公交线网的特点,尝试以蚁群优化算法进行公交线网优化研究。通过分析了传统蚁群算法的特点和不足,重点提出了基于信息素递减、半动态候选列表的蚁群算法,并提出了dijkstra-蚁群混合算法用于相应的公交线网优化分析,所提算法解决了传统蚁群算法的不足,并提高了蚁群算法运用于公交线路优化的求解效能。(3)城市公交线路优化绩效评估对于公交线路布设效能的判定至关重要。通过深入分析国内外公交线网评价方法的优点及不足,同时针对公交线路布设不同阶段的特点,提取了相应的评价指标体系,上述评价指标体系可有效的适用于公交线路前期布设及后期运营效能的定量评估。提出了基于分层聚类的主成分分析法运用于公交线路优化效能评价。(4)城市公交线网优化的研究并不是相互分割的部分,需要将公交规划的各个流程合理的融入整个公交线网优化体系当中,才能有效提高公交系统的服务效能。城市公交需求的变化是一个动态发展的过程,我国大城市的发展速度很快,公交规划不仅需要考虑长期的公交发展需求,还应当考虑较短时间范围内的公交需求变化。本文以合肥市一环内公交线优化为例,深入分析了一环内道路交通状况及轨道交通建设情况,提出了基于dijkstra-蚁群混合的公交优化算法,得到了基于距离最短和基于客流量最大的城市公交系统线网优化模型,同时对公交路线优化的前期数据采集、优化问题求解、后期备选方案评价等过程进行了研究。在提出的城市公交线网优化研究体系的基础之上,再以合肥市一环内线路为实例研究了区域内各级线路优化调整方法,并给出了相应备选方案集的效能评定。通过实例验证,证明了本文理论体系的有效性。更多还原

【Abstract】 With the continuous improvement of the urban economy and people’s living standards, the traffic pressure of China’s large and medium-sized cities gradually increased. As an important part of the urban systems, urban public transport system could provide essential material conditions to the personnel, material flows within the city system, the imbalance in demand and supply of ur-ban public transport will bring a big impact on the economic and social development of the city. Currently, give priority to the development of urban public transport systems and promot the di-versity of travel demand has been recognized by the majority of scholars, government depart-ments as the most effective way to solve the traffic problems of cities. Therefore, there are signif-icant limitations of traditional public transportation network optimization which only consider the ground bus lines, neglect other public transport, and can not meet the diverse requirements of ur-ban transportation. For the actual public transportation network planning, there is need to consider the various aspects of urban transit systems planning ground transportation, rail transportation. We should take advantage of different types of public transport as a whole, to improve the service performance of the urban public transport system. This paper mainly focuses on the critical theory of urban bus transport network planning and technology, both academic and practical significance are considered. Series of research achievement have been fullfilled in those aspects such as urban bus network traffic forecasting, urban public transportation network optimization research, urban public transit network optimization model and algorithm, evaluation of public transit network optimization effectiveness, etc. The main contents are as follows:(1) Public transport network is the foundation for urban development. On the basis of in-depth analysis on the current research, this paper identifies the resources configuration and optimization of Hefei urban transport system, which should adhere to the principles of both the local adjustment and overall co-ordination. By analyzing the transformation of the urban public transportation network traffic, it can be effectively grasp the corresponding traffic travel rules, thus conducive to the development of rational public transport optimization program based on the actual situation. Through in-depth analysis of the variation, this article focuses on the changes in the transportation system traffic site. For different types of traffic data, Time series forecasting model based on empirical mode decomposition, and support vector machine model based on em-pirical mode decomposition are presented, respectively. Meanwhile, the scope for bus traffic forecasts predicted sequences under different models are determined.(2)Urban public transportation network is the basic and core of urban transit system. Due to the characteristics of urban public transportation network, this paper put forward the ant colony optimization algorithm to public transportation network optimization. By analyzing the characte-ristics and shortcomings of traditional ant colony algorithm, ant colony algorithm based on phe-romone decreasing and ant colony algorithm based semi-dynamic candidate list are proposed. The mixed dijkstra-ant colony optimization algorithm is applied for the corresponding public trans-portation network analysis. The proposed algorithm could solve the shortcomings of traditional ant colony algorithm, and improve the performance of the ant colony algorithm for solving bus route optimization problem.(3) City bus route optimization evaluation performance is critical for determining the effec-tiveness of bus lines laid. Through in-depth analysis of the advantages and disadvantages of do-mestic and international public transportation network evaluation methods, according to the cha-racteristics of different stages for bus lines laid, the corresponding evaluation index is extracted. The evaluation index system can effectively apply to pre-routed bus lines and quantitative as-sessment of the effectiveness of post-operation. Principal component analysis based on hierar-chical clustering is proposed to bus route optimization performance evaluation.(4) Urban public transportation network optimization are not mutually separated parts, there is need to integrate the various processes rational planning of public transport among the entire public transportation network optimization system. And then, the service performance for bus system will be effectively improved. The variation of urban public transport demand is a dynamic process. The development speed of China’s major cities is very fast. For urban public transport planning, there is need to consider the long-term development of the urban public transport sys-tem, the demand changing within a short time frame should also be considered. This paper takes the Hefei city ring bus line optimization as an example. By depth analysis of the situation within a ring road and rail traffic construction, the mixed dijkstra-ant colony optimization algorithm is presented. After that, urban transit systems line network optimization model based on the shortest distance, and urban transit system traffic line network optimization model based on the maximum traffic are established. At last, the data collection, optimization problem solving, and post-alternative evaluation process for the bus route optimization. On the fundamental of pro-posed urban transit network optimization theoretical system, this paper studies the route optimiza-tion within regional levels adjustment methods for Hefei city ring bus line optimization. And the performance evaluations for corresponding set of options are proposed.Finally, the verified example proves the effectiveness of the theoretical system for Urban public transportation network optimization.更多还原