| The coordinated development of built environment and transportation system is the premise to alleviate traffic problems and enhance land use intensification.It is not only the key to promote healthy and sustainable of urban development,but also one of the important topics in urban planning.At present,China is promoting large-scale urbanization construction and new transportation infrastructure construction,which provides an opportunity to study the coordinated development of built environment and transportation system.The new national spatial planning,which is based on the principle of"integrating multiple regulations into one plan"and"sticking to one blueprint",provides an important practical opportunity for studying the coordination and optimization of the built environment and transportation system.The increasingly mature technology of big data analysis and the improvement of the accuracy of traffic geographic information data provide technical support for in-depth research on the coordination between built environment and traffic system.In this context,based on the rich transportation system and built environment big data such as GPS trajectory,traffic accidents and points of interest(POI),this paper focuses on the relationship between built environment and transportation system operation state,the determination of spatial unit scale,the impact of built environment on transportation system operation state,and the coordinated evaluation and optimization of the two systems,The main research contents of this paper are as follows:(1)With the goal of building a comprehensive,multidimensional and simple built environment and transportation system operation status index system,this paper puts forward the index calculation method based on multi-source spatial data in the5DS dimension,and makes full use of spatial data such as land use data,POI data,Open Street Map data and building contour data to make the characterization of built environment more refined.Based on the coordinated development of land use and transportation system and comprehensive evaluation of transportation system,combined with the spatial unit characteristics and data characteristics,drawing lessons from industrial applications,a transportation system index system for coordinated and sustainable development of built environment and operation state of transportation system is constructed.(2)Aiming at the unstable influence of the built environment on the operation state of the transportation system caused by the different selection of spatial units,this paper proposed the double-layer detection of variables and model parameters method.Through the spatial autocorrelation and collinearity test of variables such as population density,land use mixing degree,intersection density,public traffic proximity and business district accessibility,and the fluctuation test of the estimated value of the result parameters of the geographically weighted regression model realizes the detection of the problem of plastic area unit.A multi-objective optimization method for spatial units is proposed in this paper,which effectively reduces the uncertainty of research results caused by different aggregation scales of spatial data.(3)Aiming at accurately depicting the spatial heterogeneity of the impact of the built environment on the transportation system state,a deep neural network expression method of spatial nearest neighbor relationship is proposed,which realizes the nonlinear fusion expression of two-dimensional nearest neighbor relationship in rectangular space and geographic topological space.The geographically weighted atrous convolution neural network regression model(GACNNWR)with dynamic learning ability is designed,and the training framework of hollow neural network is established,and the statistical diagnosis and analysis method is deduced,which breaks through the limitations of the traditional geographically weighted regression model with single spatial measurement,single kernel function and unified bandwidth.Taking Jinan as an example,the experimental design and model verification are carried out.The results show that the geographically weighted hole convolution neural network can more accurately reveal the geographical nearest neighbor relationship,and improve the fitting and new sample adaptability of the relationship between the built environment and the operation state of the traffic system.(4)In order to evaluate the coordination between built environment and traffic system state objectively,the regression coefficient of the geographically weighted hollow convolution neural network model is used to constrain the weight of the built environment index,and the correlation coefficient is used to constrain the weight of the transportation system state index.The corresponding preference constraint cone is established based on C~2WH model,and then an improved data envelopment model is constructed to effectively restrict the weight of traffic index,traffic accident number,traffic comfort and other indicators.For the uncoordinated and effective research units,to realizing the coordination between the built environment and transportation system state,the improvement scheme was proposed by using the distance between the actual index and the index value in the coordinated state.Thus,realized the precise locking of the uncoordinated factors and improved the practicality of the coordination evaluation and optimization.This paper analyzes the interactive relationship between the built environment and the transportation system state,reduces the impact of the plastic area unit problem on the instability of the research results,reveals the spatial heterogeneity of the impact of the built environment and the operation state of the transportation system,and puts forward the coordination evaluation and optimization method of the two.Provide theoretical reference and technical support to improve the level of coordinated and sustainable development of the city to optimize the operation quality of urban transportation system. |