关键词: 交通工程, 影响分析, 社团结构, 公交专用道, 复杂网络

Abstract:

Designation of bus lanes is an important measure to protect public transport priority. However, it will compress running space of vehicles. So part of the vehicles have to choose other sections in the peak hours, that can affect other vehicles on the other road. In order to research the influencing regions, we transfer it into process of the data sample cluster, this paper proposes here a measure for community structure detection of similarities between vertices based on random walks, which combines topologies of complex network and epidemic dynamic. This algorithm improves the transition probability of random walk that particles transfer between nodes in the network, not only relate to the degree of nodes, but to the distance between nodes. It deduces dissimilarity by transition probabilities and based on it. It uses hierarchical clustering to build a tree for final community structure. A case study on Jingtong Expressway in Beijing shows the urban traffic network proved the feasibility of this method, which can also provide new research ideas and methods to relevant research works.

Key words: traffic engineering, influence analysis, community structure, bus lane, complex networks