Generality And Peculiarity Of The Spatial Complexity Of Urban Road Networks

With the global urbanization process and the complication of the city itself,the importance of urban issues to the sustainable development of human society has become more prominent,and the establishment of corresponding urban science has become an era proposition.However,since the current urban research is still in the pre-paradigm period,exploring the generality of urban phenomena has become the primary task at this stage.Among the various attributes of the city,the spatial form is the background element,and the complexity is the essential feature.Therefore,the spatial complexity of these two points should be the key area of urban research.In view of this,the urban road network is the most stable of the various elements of urban space,its shape and structure have a crucial impact on the operation of the city,meanwhile the unique network form of the road in various land uses also gives it the ability be portrayed by various spatial description.It explores complexity on multiple levels,especially on the space complexity as a good"bridge".Therefore,the road network is a suitable perspective for exploring the generality and peculiarity of urban spatial complexity.On the practical level,studying the complexity of road network space can provide a reference for exploring the optimization mode of road network.On the other hand,it can help to clarify the characteristic status of Chinese cities in the forest of world cities in terms of the complexity of road network space,providing a sample city for reference.Specifically,in order to exploring the generality and peculiarity of urban road network spacial complexity,this paper gradually develops in the following four parts:The first part defines the connotation of spatial complexity and establishes an analytical framework.Firstly,based on the review of complexity research process and the summary of its core ideas,the basic meanings of the definition of spatial complexity are clarified,as well as the dual cognitive framework of land spacial complexity in the human-land system is proposed.Furthermore,based on the definition of the"general"concept,the specific analysis framework and measurement system of this paper are proposed.The second part explores the generality of urban road network spacial complexity.Totally 1040 urban samples were selected on a global scale.The spatial complexity characteristics of each sample urban road network are measured by above measurement system.Then the distribution of results of each measurement index and its correlation were investigated.The third part explores the peculiarity of urban road network spacial complexity.Principal component analysis is utilized to extract the feature vectors that describe the spatial complexity of urban road network.Then,the agglomerative clustering is ustilized to classify the sample cities.Finally,the attribute differentiation and geographic differentiation between each type are analyzed.The fourth part explores the possible evolutionary principles behind the above properties.This paper puts forward some hypothesis of evolutionary principles for urban road network generation,carring out corresponding mathematical and physical expressions.Based on this,the evolution model of urban road network is constructed to explore the complexity of road network space under the influence of different evolution principles.Therefore,this provides a reference for explaining the spacial complexity of urban road network in reality.The main conclusions of the paper are as follows:(1)The spatial complexity of land in the urban human-land system has the dual meanings of spatial intrinsic complexity and spatial representation complexity.The spatial intrinsic complexity of land refers to the complexity of the spatial form of land itself beyond the"simplification of space"represented by plastic geometry and the appearance of fractal features.The spatial representation complexity of land refers to the specific morphological or structural characteristics that are present when it is influenced by the complexity characteristics of the human social system it carries.The two prominent dimensions are the organization and purpose of the system.The spatial structure acting on the land is reflected in the degree of planning and function of the latter.(2)In terms of generality,the urban road network shows a strong generality in spatial complexity.Firstly,in terms of spatial intrinsic complexity,morphological fractals and structural fractals are ubiquitous,but either the morphological power-law distribution or the structural power-law distribution exists only in the high-grade road system in cities.Secondly,in terms of the complexity of spatial representation is concerned,the distribution of the self-organization degree and functional level of the road network are in a kurtosis state,with a strong consistency.Finally,the correlation between the two types of spatial complexity is weak,reflecting strong independence.(3)In terms of peculiarity,urban road networks can be divided into five types according to the peculiarity of spatial complexity.Type ? city has the smallest road network,low self-organization degree,high fractal dimension,and prominent hierarchical structure,which is the best in eff-iciency and robustness.Asian economically developed regions such as Japan and South Korea have more distribution in this type.Type ? cities have the largest scale and are limited in planning.Although the coverage level is good,the overall functional level is weak,corresponding to the big cities that are under planning in reality.In type ? cities,the urban road network has low factors in the intrinsic complexity,and the functional level is similar to that of the type ?.The overall development is in underdeveloped state,which corresponds to the urban road network form in the real development stage.These can be seen mostly in Africa and Asia.Type IV cities are cities with strong planning,have the highest structural fractal factor.There are more distributions in China and North America,which may represent the urban road network under the guidance of motor vehicles.Type V cities are 'unmatched' between the length distribution and the degree distribution,with high compliance and robustness.There are more distributions in Europe and Latin America.(4)Furthermore,there are certain differences and connections between the various regions in the distribution of five types'urban road networks.Firstly,the distribution of Europe and Latin America is similar,with Type ?,Type ? and Type ? as the dominant factors.The similarity between two regions may because of the colonization of the former.Secondly,Africa and Asia have similar distributions in the developing regions,all with the domination by Type ?,which may because of the similar development stages.Thirdly,cities in developed Asian regions such as Japan and South Korea are mainly based on Type?.Thus,this region is worth learning in terms of road network form.Finally,there is a certain similarity between the mainland China and North America,both of them dominated by Type ? and Type ?,which may due to the extensive introduction of the latter's planning concepts and experience.(5)The circle structure mechanism and the connection utility principle can provide explanations for some features of the urban road network spacial complexity in reality.The computer simulation experiments show that urban road network fractals and structural fractals are very stable.However,the introduction of the preference attachment mechanism in the generation process of road network will weaken the universality of the structural fractals.This may be the reason why the road network degree of fit is relatively low on the structural fractal in some cities in reality;the scale-free nature of the road length distribution may be related to the agglomeration economy of the city itself;the scale-free distribution of the road's stroke distribution is common in the tree-like growth network.The introduction of the cycle structure will destroy this characteristic,which provides a possible explanation for why only a rare number of high-grade road systems present a scale-free distribution.