Quantitative Description On Structural Complexity Of Urban Road Networks

Urban road networks is important supporter and it is the basis of development of cities, prosperously, orderly and rapidly. In the process of urbanization, transportation efficiency are improved, but some new problems at the same time, for instance, rapid expansion of urban road network and traffic jam, improvement of transportation convenience and environment pollution, flyover-crossing traffic and poor practical efficiency of part road network. Comprehensive understanding to structural organization of urban road network and transporting function is necessary to solve these problems.Scholars at home and abroad from geo-information science, geography, urban planning, traffic engineering, behavioral science and operational research conducted large research on morphology, structure, operation mechanism, relationship between characteristic and human behavior, pattern of development and evolution from various aspects and dimensionality. Complex network research arise at the end of twenty century and associate with system theory, information theory and cybernetics, complex network offer give a hand of complexity research. In addition, it has been successfully used in the study of road network. Traditional fractal theory is limited to fractal geometry. Integration of complex network and fractal theory can be used to describe the structural complexity of road network quantitatively and reveal the self-organization characteristics and how structure affects efficiency. The paper aims to propose some indicators to describe structural complexity of road network quantitatively, apply them to the study of some real road networks, explore their self-organization and analyze the relationship between structure and function of road network. Paper’s main work as follows:First part, taking Chengdu as a study case, dual graph is built based on stroke and small word and scale-free properties are explored. The functions of loops and annular-radiant roads and how they connect with other roads in whole network are discussed. Degree centrality, clustering coefficient, betweeness centrality and closeness centrality are used to explored complexity of connection in road network and Correlations among them are also studied.Second part, in case of grid pattern, circular and radial pattern, free pattern and mixing pattern, original graph and dual graph are built to compare their difference in connectivity, accessibility and reachability. Studies show that grid pattern road network is convenient and free pattern network performs poorly in all aspects. Annular-radiate and mixed pattern road network have better properties than free pattern network, but worse than grid because of the filling of grid in the two-dimensional plane except their overall features.Third part, boxing-counting approach has been used in study of multi-scale structural fractal. The structural fractal of Chengdu, Hong Kong and New York are checked in different scales. The results show that road networks are fractal and the values of structural fractal dimension range from 2.21 to 3.95. Hubs are mutually exclusive in networks and exclusiveness is not enhanced or weakened with the change of map scale. Namely, the values of structural fractal dimension is not increasing or decreasing consistently following scale change.