基于Hadoop平台的实时多路径交通流分配算法  

Real-time Multi-path Traffic Flow Assignment Algorithm Based on Hadoop Platform

作　　者：段宗涛[1,2] 李莹[1] 郑西彬 康军[1,2] 程豪[1] 

机构地区：[1]长安大学信息工程学院,陕西西安710064 [2]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,陕西西安710064

出　　处：《中国公路学报》2014年第9期98-104,共7页China Journal of Highway and Transport

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)

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%。The equilibrium of traffic flow based on the traditional traffic flow assignment algorithms is hard to guarantee. To solve this problem, Hadoop big data computing platform was adopted for designing and implementing a real-time multi-path traffic flow assignment algorithm. If there were more than one paths with the minimum impedance between the purposed OD pairs, a set of all minimum impedance paths between the OD pairs would be found out firstly. Then, the traffic flow on the OD pairs would be uniformly assigned to the set of paths so as to avoid exceeding of actual traffic flow in each section beyond its capacity threshold, and prevent the delayed congestion. At last, a simulation of the real-time multi-path traffic flow assignment algorithm based on Hadoop big data computing platform was made. The results show that compared with traditional methods, the traffic flow of each section at each moment in the simulated road network decreases by 42.3%averagely, the mean square deviation decreases by35.8%averagely, and the maximum traffic flow decreases by 19.4% averagely.

关 键 词：交通工程 交通流分配 模型仿真 HADOOP平台 交通诱导 交通流均衡 OD对 

分 类 号：U495[交通运输工程—交通运输规划与管理]