基于实时路况公交换乘算法的研究与实现

发布时间：2018-05-25 17:51

本文选题：公共交通 + 代价值　；参考：《北京邮电大学》2014年硕士论文

【摘要】：近年来随着科学技术的不断进步,人们的生活水平极大提高,各城市的城市化进程不断被推进,但随之而来的问题也日益严重。城市规模的扩大致使城市公交系统变得越来越发达,错综复杂的公交线路,给人们的出行方便带来极大的挑战。我国的大中型城市人口密度和机动车数量迅速增长,给公交系统带来巨大的压力,交通形势尤为严峻。为了提高公共交通系统对人们出行的吸引力,减少私家车的使用,缓解城市交通压力,论文提出了一种有效、高效、完善的基于实时路况的公交换乘算法。论文首先分析了国内外的学者对于人们的出行问题的研究结果,发现提高公共交通系统的服务质量可以使人们更多地使用公共交通工具,于是论文给出了六种不同的换乘方案；之后分析了国内外公交换乘算法的研究现状,提出了相应的算法设计方案,具体分为静态换乘算法与动态换乘算法两部分；静态换乘算法部分利用了数据建模的相关理论进行建模、利用数据库相关技术最大限度的提升该方案的查询速度。在此算法的基础上使用人工神经网络进行方案路程消耗时间的预测,将预测值作为最终的换乘方案评判条件；然后使用大量的历史公交系统数据对人工神经网络进行训练,将实时的道路状况和公交车的运行状态输入网络得到相应时间的预测值,根据预测值对静态方案重新排序从而得到动态换乘方案。论文还对静态换乘算法进行了准确率、覆盖率和查询效率的系统测试,结果表明论文所提出的静态换乘算法不仅可以提供六种换乘模型,还具有高准确率、高覆盖率和高效的查询效率。将动态换乘方案与静态换乘方案进行了对比,结果表明利用神经网络预测的动态换乘算法结果可以根据查询的时间点进行不同程度的调整,提供实时的换乘方案；与百度换乘工具结果对比,说明了动态换乘算法可以根据实时路况对路程消耗时间进行很好预测,对人们的出行有更好的指导作用。最后总结了在研究生期间的工作内容和成果。
[Abstract]:In recent years, with the development of science and technology, the living standard of people has been greatly improved, and the urbanization process of various cities has been pushed forward, but the following problems are becoming more and more serious. With the expansion of the city scale, the urban public transport system becomes more and more developed, and the complicated bus routes bring great challenges to people's travel convenience. The population density and the number of motor vehicles in large and medium-sized cities in China are increasing rapidly, which brings great pressure to the public transport system, especially the traffic situation. In order to improve the attraction of public transport system to people, reduce the use of private cars, and alleviate the pressure of urban traffic, this paper proposes an effective, efficient and perfect bus transfer algorithm based on real-time traffic conditions. Firstly, this paper analyzes the research results of domestic and foreign scholars on people's travel problem, and finds that improving the service quality of public transportation system can make people use more public transport vehicles, so the paper gives six different transfer schemes. After analyzing the research status of bus transfer algorithm at home and abroad, the corresponding algorithm design scheme is put forward, which is divided into static transfer algorithm and dynamic transfer algorithm. In the part of static transfer algorithm, the theory of data modeling is used to model, and the database correlation technology is used to improve the query speed of the scheme. On the basis of this algorithm, the artificial neural network is used to predict the travel time of the scheme, and the predicted value is taken as the final evaluation condition of the transfer scheme, and then the artificial neural network is trained with a large number of historical bus system data. The real-time road condition and the running state of the bus are input into the network to get the prediction value of the corresponding time. According to the forecast value, the static scheme is reordered and the dynamic transfer scheme is obtained. The static transfer algorithm is tested on the accuracy, coverage and query efficiency. The results show that the static transfer algorithm can not only provide six transfer models, but also have a high accuracy. High coverage and efficient query efficiency. The dynamic transfer scheme is compared with the static transfer scheme. The results show that the dynamic transfer algorithm predicted by neural network can be adjusted according to the query time points to provide real-time transfer scheme. Compared with the result of Baidu transfer tool, it is proved that the dynamic transfer algorithm can predict the travel time well according to the real time traffic conditions, and it has better guidance for people to travel. Finally, it summarizes the contents and achievements of the work during the graduate period.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.17

【参考文献】