基于信息传递效率的地铁网络小世界特性评价
Information transfer efficiency based small-world assessment methodology for metro networks

为了差异化直接相邻和间接相邻的车站对信息传递效率的影响,该文建立了基于信息传递效率的聚类系数模型,构建了地铁网络小世界特性评价方法。通过对全球52个城市的地铁网络样本的小世界特征值计算,得到基于信息传递效率的聚类系数算法的聚类系数值在0.195~0.407之间,平均值为0.29,虽然小于以线路为演化单位的公共交通网络中P空间(Space-of-Stops)下的聚类系数值,仍然远大于相同规模的随机网络聚类系数值(0.01~0.16,平均值为0.06)。故认为基于信息传递效率的聚类系数算法能够更加严格地评价物理网络是否具有小世界特性。在此方法下,52个样本城市地铁网络仍具有小世界特性。
Abstract：This study presents an improved algorithm for the clustering coefficient in a metro network model. The algorithm is based on the information transfer efficiency that considers the differences between the directly and indirectly connected origin-to-destination stations. The algorithm was evaluated using 52 metro networks in the world. The information transfer efficiency based clustering coefficients for the 52 metro networks are between 0.195 and 0.407 (average 0.29), which is lower than the value given by P-Space (Space-of-Stops), but still considerably higher than the values for random networks (0.01 to 0.16, the average is 0.06) with the same size. Therefore, metro networks are small-world networks, although with a stricter evaluation model.