基于聚类算法的交通网络节点重要性评价方法研究

Methods of Importance Evaluation of Traffic Network Node Based on Clustering Algorithms

评估交通网络中节点的重要性,识别出对网络效率起着重要作用的关键节点,对于预防和降低交通拥堵和交通事故等事件对路网整体效率的影响具有重要意义。为识别出关键节点,研究了一种基于聚类算法的交通网络节点重要性评价方法:将道路交通网络抽象为无向加权网络,以节点介数、节点交通量和PageRank值作为节点重要性评价指标,利用基于K-Means算法和随机森林加权的改进FCM算法确定交通网络节点重要性,适用于中小城市道路交通网络。实证分析表明,改进算法的聚类性能明显提高,目标函数值和迭代次数分别降低88.70%和61.54%,同时算法误判率也仅为5.50%,验证了所提出的方法可以更为客观地刻画交通网络节点重要性程度,更为准确地动态辨识出关键节点。

Assessing the importance of nodes in a traffic network and identifying key nodes that play an important role in network efficiency is essential for preventing traffic congestion and reducing impacts of local traffic accidents on the overall efficiency of the road network. In order to identify the key nodes,an evaluation method based on clustering algorithm is developed to evaluate the importance of traffic network nodes in the urban road traffic network. In this method,the road traffic network is abstracted as vectorless weighted network with node betweenness,node traffic volume,and PageRank values as evaluation indices of node importance. Based on the three indices,the importance of traffic network nodes is determined by the improved FCM algorithm based on K-means algorithm and Random Forest weighting. This method is applicable to small and medium-sized urban road traffic network.. Results show that the clustering performance of the improved algorithm is significantly improved with the objective function value and the number of iterations reduced by88.70%and61.54%,respectively. The misjudgment rate of this algorithm is only5.50%,verifying that the proposed method can describe the importance degree of nodes in traffic network more objective,and dynamically identify key nodes more accurate.