| Fractality is the optimal structure of nature and a new way of discovering the order and structure of things.However,current research is limited to geometric fractals.The method of fractal dimensions of network structures recently proposed in the field of complex networks provides a new perspective for exploring the interrelationships between geographical elements.Research shows that many geographic objects are both geometric fractals and structural fractals.However,in the geographic space,the complexity of geographic entities changes with the change of scale.The nature of the observations of the same subject at a certain scale,which the principles or laws that are summed up,may be valid,may be similar,or may need revision on another scale.Roads are important geospatial entities,whose distribution and structure are attached to specific surface morphology and geographical environment under the dual constraints of human factors and natural factors and road network is a typical,open,complex giant system.This thesis takes the road network as the research object,and discusses its structural fractality and self-similarity features under multiple geospatial scales.The thesis collected a total of 1,218 road network data for metropolitan statistical area scales,county scales,and county subdivision scales.For each road network,we analyze their morphological characteristics,general network structure characteristics,structural fractality features,small-world and scalefree features and so on.The results show that the road network is a typical structural fractality.And the boxcovering fractal dimensions of Metropolitan Statistical Area,County and County Subdivision are [2.877,5.040] 、 [2.425,4.335] 、 [1.398,4.298];the volume fractal dimensions are [3.531,4.948] 、 [2.457,4.813] 、 [1.932,4.704];the degree-volume fractal dimensions are [4.131,5.298]、[3.181,5.412]、[2.011,5.213],respectively.All of the three kinds of structural fractal dimensions are normally distributed.And the degree-volume fractal dimensions of almost all road networks is higher than the volume fractal dimensions,the latter are again higher than the box-covering fractal dimensions of road networks.This is determined by the definition of the structural fractal dimension.Compared with different spatial scales,the fractal dimensions of the road networks in the metropolitan statistics area are higher than the fractal dimensions of the county-scale road networks which contains in MSA,and the latter are higher than the fractal dimensions of county subdivisions scale which contains in county.In addition,the larger the road network is,the higher the fractal dimension of the structure is.The study also found that the road network exhibits typical small-world and scale-free features over a range of geospatial scales and is therefore self-similarity in structure.The relevant conclusions of the dissertation are of great significance for the in-depth understanding of road network self-organization structure and evolution,intelligent spatial planning,and knowledge-assisted cartographic synthesis. |
