基于模糊聚类的城市快速路交通流状态划分  

Classification of Urban Expressway Traffic Flow Situation Based on Fuzzy Clustering

作　　者：殷俊杰[1] 丁宏飞[1] 薄雾 钟媚[1] 

机构地区：[1]西南交通大学交通运输与物流学院,四川成都610031

出　　处：《重庆交通大学学报（自然科学版）》2013年第4期652-655,共4页Journal of Chongqing Jiaotong University(Natural Science)

基　　金：四川省科技支撑计划项目(2011FZ0050)

摘　　要：针对城市快速路交通流状态分类的问题,提出了一种改进的模糊C均值(FCM)算法。结合层次聚类算法和FCM聚类算法,运用层次聚类算法得到最佳聚类数和初始聚类中心,并通过Relief F特征加权对影响交通状态的不同特征指标赋予相应的权值,最终用FCM算法再次聚类得出交通流状态的分类结果。以VISSIM为工具,对该方法进行了模拟。对比分析结果显示,所提出的方法能够提高城市快速路交通流状态分类的效果。In order to solve the problem of urban expressway traffic flow situation judgment,an improved algorithm which uses fuzzy C-means clustering method is proposed.Based on the combination of FCM algorithm and hierarchical clustering algorithm,the hierarchical clustering algorithm is firstly used to get the number of cluster type and initial cluster centers;then Relief F algorithm is used to calculate the weights of the chose features,which affects the division of the traffic conditions;at last FCM is used to perform the cluster to obtain the division results of traffic conditions.A simulation is performed by software VISSIM to confirm the proposed methodology.Through the comparative analysis,it is obtained that the propose method can improve the classification results of urban expressway traffic flow situation.

关 键 词：交通状态 模糊聚类 层次聚类 特征加权 RELIEF F算法 

分 类 号：U491.112[交通运输工程—交通运输规划与管理]