基于Hadoop平台的实时多路径交通流分配算法

发布时间：2018-01-13 19:07

本文关键词：基于Hadoop平台的实时多路径交通流分配算法　出处：《中国公路学报》2014年09期 　论文类型：期刊论文

【摘要】：为解决采用传统方法进行交通流分配时难以保证交通流均衡性的问题,基于Hadoop大数据计算平台,设计并实现了一种实时多路径交通流分配算法。当需要进行交通流分配的OD对之间存在多条阻抗最小路径时,该算法首先能够搜索到给定OD对的所有阻抗最小路径集合,然后将该OD对的交通流均匀分配到搜索到的路径集合上,从而实现交通流的均衡分配,避免各个路段实际交通量超过其流量阈值,有效预防后继拥堵的发生,最后对基于Hadoop平台的实时多路径交通流分配算法进行了仿真试验。结果表明:仿真路网中各路段在各个时刻的交通量均值相对于传统方法平均减少了42.3%,交通量方差相对于传统方法平均减少了35.8%,交通量最大值相对于传统方法平均减少了19.4%。
[Abstract]:In order to solve the problem that it is difficult to ensure the equilibrium of traffic flow when using the traditional method to distribute the traffic flow, based on the Hadoop big data computing platform. A real-time multipath traffic flow assignment algorithm is designed and implemented. When there are multiple impedance minimum paths between OD pairs that need to be allocated traffic flow. The algorithm can first search all the minimum impedance path sets of a given OD pair, and then distribute the traffic flow of the OD pair to the search path set uniformly, so as to achieve the balanced assignment of traffic flow. Avoid the actual traffic volume of each section exceed its flow threshold, and effectively prevent the occurrence of subsequent congestion. Finally, the simulation experiment of real-time multi-path traffic flow assignment algorithm based on Hadoop platform is carried out. The results show that:. Compared with the traditional method, the average traffic volume of each section in the simulation road network is reduced by 42.3%. Compared with the traditional method, the mean variance of traffic volume is 35.80.The maximum value of traffic volume is 19.4knuckles less than that of the traditional method.
【作者单位】：长安大学信息工程学院;长安大学陕西省道路交通智能检测与装备工程技术研究中心;
【基金】：国家自然科学基金项目(51278058,61303041) 中央高校基本科研业务费专项资金项目(2013G2241020,2013G1241119) 交通运输部应用基础研究项目(2014319812150) 陕西省工业攻关项目(2014K05-28)
【分类号】：U491
【正文快照】： 0引言交通诱导是城市路网交通流分配的一种重要手段,Dijkstra算法是交通诱导过程中经常采用的一种最短路径搜索算法。实际的城市道路交通网络非常复杂,对于1个OD(Origin Destination,OD)对,通常存在多条最短路径。当有大规模并发的路径搜索请求时,对于1个OD对Dijkstra算法只能

【参考文献】