考虑连接性的路网划分算法

针对传统K均值聚类算法在非均质路网划分应用中的不足,将路网连接性融入算法,解决其在路网划分应用中聚类结果不连续的问题.先使用最大最小距离算法确定初始聚类中心和路段差异性,并以聚类评价指标ANSK确定K值;然后统计连续时间间隔下路网划分结果的动态频数,合并和拆分不稳定的“噪声”路段,提高划分子区内路网的紧凑性.最后,基于现实路网中的车牌照自动识别实测数据,对改进的聚类方法进行了验证.将算法得到的划分效果与K均值聚类算法和Ncut算法进行对比,并对子区做宏观基本图分析.结果表明,改进后的K均值聚类算法在保证自身原有聚类优势下,可以有效实现连接性约束下的路网划分.

Road Network Partition Algorithm Considering Connectivity

Aiming at the deficiency of the traditional K-means clustering algorithm in the application of heterogeneous road network partition, the connectivity of road network is incorporated into the proposed algorithm to solve the problem of discontinuous clustering results. First, we use the max-min distance algorithm to determine the initial clustering center and links’difference, and use the clustering evaluation indexANSK to determine the value of K. Then, we count the dynamic frequency of road network partition results in continuous time intervals, and the unstable "noise" links are merged or split for improving the compactness of the road network in subareas. Last, based on the measured data of automatic number plate recognition in real road network, the improved clustering method is tested in this paper. The partition results of the improved algorithm are compared with the traditional K-means clustering algorithm and Ncut algorithm. Also, the macroscopic fundamental diagram is created to analyze the subarea. The results show that the improved K-means clustering algorithm can effectively implement the road network partition with the connectivity constraint.