Intelligent Planning And Management Methods For Metro-led Underground Space Based On Multi-source Data

The large-scale utilization of urban underground space(UUS)in modern society originated with the development of metro systems,which have been in operation in 59 countries and 187 cities worldwide.Experience has proved that metro development can effectively alleviate traffic congestion and promote the optimization of spatial configuration.In recent years,the metro construction in major Chinese cities has also stepped into a rapid development phase,and the metro-led underground space(MUS)has gradually become an urban catalyst for public-transport-oriented development.In this study,MUS is defined as the sum of all underground connected spaces and underground spaces with substantial connecting value within metro-led area,with the metro station as the core and the underground pedestrian network as the connecting skeleton.In the context of high-quality urban development in the new era,MUS use is expected to alleviate the scarcity of spatial resources for inventory planning,promote the compact and high-level spatial development,optimize the functional layout and spatial efficiency,and has become the new frontier of UUS use in China.Concerning the planning system for UUS,the master planning for UUS proposes general strategies for MUS development,and the regulatory detailed planning for UUS raises more specific regulations for MUS development in key areas.To date,researchers have actively probed into the technology and system of regulatory detailed planning for MUS during the planning practices in Chinese metropolises.As the complexity of spatial configuration and frequency of spatial activity within MUS increases,the planning and management system of MUS encounters many serious issues,including insufficient recognition of spatial law,lack of decision-making toolkit,and absence of digital planning management,which urgently calls for theoretical reconstruction and technical optimization from the inside out.Nowadays,the emergence of new multi-source data concerning spatial planning has not only created external conditions for the reconstruction of MUS planning,but also stimulated the endogenous motive of datadriven intelligent planning.Multi-source spatiotemporal data implies both opportunities and challenges on MUS planning.The establishment of an intelligent planning system for MUS from a data-driven perspective is likely to restructure the current planning paradigm and rationale,thus supporting the decision-making of planning.Confronted with the foregoing background,this study adopts various multi-source data and data-driven research methods to investigate on the intelligent planning and management methods for MUS.The research contents cover several critical aspects of MUS,including the multi-scale spatiotemporal evolution laws analysis,multi-level traffic network analysis,intelligent layout planning,spatial efficacy assessment,and digital planning management platform construction.The toolkit covering the holistic process of MUS planning that involves knowledge mining,planning compiling,planning implementation and planning management,has been developed in this study,which is supposed to provide robust scientific evidence towards an intelligent planning and management system of MUS.Specifically,the main findings of this study comprise the following parts.(1)Based on POI data,an automatic extraction method of five MUS characteristic parameters including spatial category,geometric form,functional use,traffic organization,and development intensity was proposed.Both the Pearson correlation analysis in classical statistics and geographical detector in spatial statistics were applied to identify the forces driving MUS development.Besides,the indicator system of explanatory and characterization factors at both the macro level for socioeconomic development and micro level for built environment were established.Using 53.34 million records of full POI big data in China at the end of 2021,the multi-scale spatiotemporal evolution laws of 3179 MUSs in China from 1969 to 2021 was analyzed.The results show that the driving mechanism of MUS development in China has differentiated characteristics under distinct space-time conditions,but the development intensity of ground space,the diversity of land use,public transport,and static transport play a significant role in promoting the MUS use in most scenarios,which is consistent with TOD principles.At the macro level,the economic scale and built environment scale are supposed to boost MUS use positively.(2)The features of rail transit travel mode were analyzed using space syntax,and an assessment model for spatial accessibility of macro network of the metro-led unit was thereby established from the perspective of spatial configuration.The model adopts the simplified axis line segment of rail transit network as the unit,the axis length proportion as the weight,and the global integration as the indicator.Meanwhile,the modeling methods of micro network of the metro-led unit for planning implementation evaluation and planning scheme generation were established respectively based on complex network theory.The network analysis indicators such as accessibility,efficiency,robustness,and perceptibility were also proposed.The case study results of Shanghai Wujiaochang-Jiangwan Stadium MUS pedestrian network reveals that the large-scale utilization of MUS would improve the efficiency of metro-led pedestrian network by 30%,which confirms the significance of MUS in traffic connection.Based on the analysis results of network efficiency and vulnerability,the renewal and optimization strategy of spatial configuration characterized by strong connectivity in shallow MUS was proposed.(3)A law-informed intelligent layout planning model for metro-led underground space(LPMUS)was constructed.The land parcel was taken as the basic unit of planning decision vector.The main layout principles were obtained based on MUS development laws,and the principles were digitally interpretated.Multiple optimization objective functions and boundary constraint conditions were defined for ordinary and complex MUS respectively.Therefore,the layout planning issue was transformed into a typical many-objective optimization problem.LPMUS adopts NSGA-III algorithm to find the Pareto front.With the aid of the reference point elite selection strategy,the individuals in non-inferior solution set were evenly distributed in the high-dimensional space.To cope with the solution of Pareto front,a layout scheme generation strategy based on pseudo weight vector method and fuzzy analytic hierarchy process(FAHP)was further proposed,which considers the fuzzy experience judgment of planners and the objective data characteristics of Pareto front.Ultimately,the model was applied to the complex MUS of Daqiao block in Jinan prior zone for replacing old growth drivers with new ones.The layout generated by LPMUS has clear logic and reasonable configuration,which verifies the feasibility and applicability of the model.(4)The concept and characterization of MUS efficacy were defined.An evaluation indicator system for MUS efficacy,which covers three factors(spatial accessibility,spatial coordination,spatial development intensity)and ten quantitative indicators,was put forward based on multi-source data.Hereafter,the MUS efficacy was quantitatively evaluated by technique for order preference by similarity to an ideal solution(TOPSIS).Taking MUS in Shanghai Inner Ring area as an example,this study analyzed the spatial distribution pattern of the efficiency,and adopted location-based service data(LBS)and UUS transaction data to observe and verify the efficiency.The adjusted R-square between the observation and the evaluation results is 0.786,which confirms the reliability of the model.In addition,this study proposed the analysis methods for both global and local MUS efficacy optimization,which can assist planners in determining potential renewal unit and region,thus supporting the overall layout optimization of MUS in key areas.(5)Confronted with the diversified development concepts such as “sustainability,ecology,resilience,and humanism”,the paradigm shift trend of MUS planning with regard to “technical methods,result forms,control modes,and control elements” was analyzed.Drawing lessons from the genealogy theory,the features of MUS planning involving built environment,planning control,planning result,operation and management,and socioeconomic conditions,were digitally interpreted.Moreover,the data standard and management requirements of each feature,such as data type,data format,coding scheme,and authority design,were put forward.Based on the digital pedigree and planning management demand,a conceptual design of the metro-led underground space planning management system(MUS-PMS)was developed,which covers eight modules and four stakeholders.Finally,using HTML5-CSS3 standard specification,combined with j Query,Bootstrap,and other frameworks,the design of planning management platform was completed,which provides experiences for efficient and refined management in planning practice.The innovation of this study can be concluded as follows.(1)A data mining algorithm based on multi-source data was constructed to extract the characteristic parameters of MUS.For the first time,the spatiotemporal evolution pattern of MUS development in China was analyzed using full samples.(2)The multi-level traffic network analysis model,intelligent layout planning model,and spatial efficacy assessment model were proposed,thus forming an intelligent data-driven planning toolkit for holistic process.(3)The paradigm shifts of MUS planning that derive from multi-source data environment were analyzed,and the corresponding digital planning pedigree and planning management platform were constructed.