Infrastructure layout optimization method and apparatus

Abstract

This invention discloses a method and apparatus for infrastructure layout optimization, relating to the field of infrastructure layout optimization technology. The method includes: acquiring multi-source data of a target area; extracting population dynamic profile tags, land development gene tags, and facility attribute tags at a standard unit scale; constructing a ternary knowledge graph containing people, land, and facilities; and performing supply and demand matching assessment based on the ternary knowledge graph to identify areas with facility shortages and a list of candidate sites with implementation potential. This invention solves the technical problems of existing infrastructure layouts, such as crude supply and demand calculations and reliance on experience in selecting potential sites, leading to low layout accuracy and implementation efficiency. It achieves accurate quantitative assessment of facility supply and demand and intelligent selection of potential sites, thus improving the accuracy and implementation efficiency of infrastructure layout optimization.

Description

Technical Field

[0001] This invention relates to the field of infrastructure layout optimization technology, and specifically to an infrastructure layout optimization method and apparatus. Background Technology

[0002] Current urban infrastructure planning and layout largely rely on traditional manual statistics and experience-based judgments. Given increasingly frequent population flows and complex and diverse land development conditions, it is difficult to achieve a coordinated analysis of population demand, land conditions, and facility supply. Traditional methods often use static population data for demand estimation, which cannot adapt to the dynamic characteristics of population changes and easily leads to discrepancies between facility supply and demand. At the same time, the assessment of land development potential relies heavily on manual qualitative judgments, lacking quantitative indicators. This results in low efficiency and strong subjectivity in candidate site selection, making it difficult to balance facility service coverage efficiency with project feasibility. Consequently, some areas experience facility shortages or resource imbalances, hindering the refined and efficient advancement of infrastructure layout.

[0003] The existing infrastructure layout suffers from technical problems such as crude supply and demand calculations and reliance on experience in selecting potential land parcels, resulting in low accuracy and efficiency in layout. Summary of the Invention

[0004] This application provides an infrastructure layout optimization method and apparatus to address the technical problems of low layout accuracy and implementation efficiency caused by the extensive supply and demand calculation and reliance on experience in the selection of potential land parcels in existing infrastructure layouts.

[0005] In view of the above problems, this application provides an infrastructure layout optimization method and apparatus.

[0006] A first aspect of this application provides an infrastructure layout optimization method, the method comprising: Acquire multi-source data for the target area, including population activity data, land ownership and current status data, and infrastructure status data; based on the multi-source data, extract population dynamic profile tags, land development gene tags, and facility attribute tags at the standard unit scale; construct a ternary knowledge graph containing people-land-facilities based on the population dynamic profile tags, land development gene tags, and facility attribute tags, with population units, land units, and facility units as nodes, and supply and demand relationships, carrying capacity relationships, and accessibility relationships as edges; perform supply and demand matching assessment based on the ternary knowledge graph to identify areas with facility shortages and a list of candidate sites with implementation potential.

[0007] A second aspect of this application provides an infrastructure layout optimization device, the device comprising: The system includes a multi-source data acquisition module for acquiring multi-source data of the target area, including population activity data, land ownership and current status data, and infrastructure status data; a tag extraction module for extracting population dynamic profile tags, land development gene tags, and facility attribute tags at a standard unit scale based on the multi-source data; a knowledge graph construction module for constructing a ternary knowledge graph containing people-land-facilities based on the population dynamic profile tags, land development gene tags, and facility attribute tags, with population units, land units, and facility units as nodes and supply-demand relationships, carrying capacity relationships, and accessibility relationships as edges; and a matching evaluation module for performing supply-demand matching evaluation based on the ternary knowledge graph to identify areas with facility shortages and a list of candidate sites with implementation potential.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: The system acquires multi-source data for the target area; based on this data, it extracts population dynamic profile tags, land development gene tags, and facility attribute tags at the standard unit scale; it constructs a ternary knowledge graph encompassing people, land, and facilities, with population units, land units, and facility units as nodes, and supply-demand relationships, carrying capacity relationships, and accessibility relationships as edges; and it performs supply-demand matching assessments based on the ternary knowledge graph to identify areas with facility shortages and a list of candidate sites with implementation potential. This achieves precise quantitative assessment of facility supply and demand and intelligent screening of potential land parcels, improving the accuracy and efficiency of infrastructure layout optimization. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This application provides a schematic diagram of an infrastructure layout optimization method. Figure 2 This is a schematic diagram of an infrastructure layout optimization device provided in an embodiment of this application.

[0011] Figure labeling: Multi-source data acquisition module 10, tag extraction module 20, knowledge graph construction module 30, matching evaluation module 40. Detailed Implementation

[0012] This application provides an infrastructure layout optimization method and apparatus to address the technical problems of low layout accuracy and implementation efficiency caused by the extensive supply and demand calculation and reliance on experience in the selection of potential land parcels in the existing infrastructure layout.

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0014] Example 1, as Figure 1 As shown, this application provides an infrastructure layout optimization method, the method comprising: Step S100: Obtain multi-source data for the target area, including population activity data, land ownership and current status data, and infrastructure status data.

[0015] Specifically, in order to obtain multi-source data for the target area, the acquired multi-source data includes three categories: population activity data, land ownership and current status data, and infrastructure status data. This type of multi-source data is the core basic data for subsequent extraction of various labels at the standard unit scale, construction of a human-land-infrastructure ternary knowledge graph, and conducting supply and demand matching assessments and identifying areas with infrastructure shortages and candidate sites, providing comprehensive data support for the advancement of the entire infrastructure layout optimization process.

[0016] Step S200: Based on the multi-source data, extract population dynamic profile tags, land development gene tags, and facility attribute tags at the standard unit scale.

[0017] Specifically, based on the acquired multi-source data, population dynamic profile labels, land development gene labels, and facility attribute labels are extracted at the standard unit scale. Among them, spatiotemporal clustering analysis is carried out on population activity data to extract population dynamic profile labels such as aging rate, school-age child density, and tidal population coefficient, comprehensively characterizing the population structure and spatiotemporal flow characteristics of each unit. Land development gene coding was implemented using a combination of standardized coding and machine learning regression algorithms for land ownership and current status data. A gradient boosting regression tree (GBDT) model was used to predict implementation difficulty scores. This model comprises a three-layer architecture: a label coding layer, a feature modeling layer, and a score prediction layer. The label coding layer classifies and assigns values ​​to property ownership and current utilization status based on the national land spatial planning land use classification standards and land ownership management regulations, generating ownership nature labels and current utilization labels. The feature modeling layer selects ownership complexity, current building volume ratio, and the number of historical issues as core features, performs quantification and normalization processing, and divides the data into training and testing sets in a 7:3 ratio. The score prediction layer uses the actual implementation difficulty value as the supervision label, setting training parameters of 0.1 learning rate, 100 decision trees, and a maximum depth of 6. After training and optimization, it outputs an implementation difficulty score ranging from 0 to 100. Finally, these are integrated to form a land development gene label containing ownership nature labels, current utilization labels, and implementation difficulty scores. Digital processing of infrastructure service capacity is carried out on the current status data of infrastructure. Facility types are defined according to unified classification standards, service scale is quantified and statistically analyzed, and the ratio of actual operating load to design capacity is calculated. Facility attribute tags containing facility type, service scale, and actual load rate are generated to fully characterize the functional positioning, service supply capacity and operating status of the facilities.

[0018] Step S300: Construct a ternary knowledge graph containing people-land-facilities based on the population dynamic profile tags, the land development gene tags, and the facility attribute tags. The ternary knowledge graph uses population units, land units, and facility units as nodes, and supply and demand relationships, carrying capacity relationships, and accessibility relationships as edges.

[0019] Specifically, a ternary knowledge graph of people, land, and facilities is constructed based on extracted standard unit-scale population dynamic profile tags, land development gene tags, and facility attribute tags. First, three entity types—population units, land units, and facility units—are defined according to the three types of tags, and corresponding attribute fields are set for each entity type. Then, relationship edges between the three types of entities are constructed. Between population units and facility units, supply and demand relationship edges are created for matching units within the circle, based on the facility standard service radius, supporting construction standards, and the actual road network calculation isochronous range, and demand matching degree weights are added. Between land units and facility units, carrying capacity relationship edges are created for compatible units based on the compatibility comparison table of facility type and land use nature, and implementation feasibility weights are added based on the land implementation difficulty score. Between population units and land units, reachability relationship edges are created for units whose spatial adjacency meets the preset distance threshold. Finally, a ternary knowledge graph is formed with population units, land units, and facility units as nodes and supply and demand, carrying capacity, and reachability relationships as edges.

[0020] Step S400: Based on the ternary knowledge graph, conduct a supply and demand matching assessment to identify areas with facility shortages and a list of candidate sites with implementation potential.

[0021] Specifically, based on the constructed human-land-facility ternary knowledge graph, supply and demand matching assessments are conducted, and areas with facility shortages are identified, along with a list of potential candidate sites. First, a standard unit to be analyzed is selected, corresponding standard population unit nodes are obtained, and various population distribution data are extracted from dynamic population profile tags. Combined with various facility service allocation standards, simulated facility demand values ​​are calculated. Then, by querying all associated facility unit nodes through the supply and demand relationship edges of the graph and summing the facility scale, the actual supply value is obtained. The difference between the simulated demand value and the actual supply value is calculated. If the difference exceeds a preset threshold, the standard unit is marked as a facility shortage area, and the type and scale of the shortage facilities are recorded. Subsequently, based on the type and scale of the shortage of facilities, the potential of land supply-side management in areas with facility shortages is identified. First, all land unit nodes are obtained and sorted according to the implementation difficulty score. The top N with scores less than a preset threshold are selected as high-potential candidate site list units. Then, accessible population unit nodes within the service radius of each high-potential unit are queried through the reachability relationship edge. The comprehensive index of facility shortage, obtained by weighted summation of the shortage scale of various facilities, is calculated. The high-potential candidate site list units and population unit nodes are matched with spatial adjacency to generate potential land parcel-demand unit matching pairs. Finally, the matching pairs are sorted from high to low according to the comprehensive index of facility shortage to generate a candidate site list with implementation potential.

[0022] In one possible implementation, step S200 further includes: Step S210: Perform spatiotemporal clustering analysis on the population activity data and extract the population dynamic profile labels for each standard unit. The population dynamic profile labels include aging rate, school-age child density, and tidal population coefficient.

[0023] Step S220: Encode the land development genes of the land ownership and current status data to generate land gene tags for each land unit. The land gene tags include ownership nature tags, current use tags, and implementation difficulty scores.

[0024] Step S230: Perform digital processing on the infrastructure status data to generate facility service capacity labels for each facility. The facility attribute labels include facility type, service scale, and actual load rate.

[0025] Specifically, using pre-defined standard units as the spatial analysis basis, spatiotemporal clustering analysis is conducted on the acquired population activity data of the target area. First, based on the spatial activity range, residence, and travel trajectory of the population, the population data is accurately matched and collected with each standard unit. Then, combined with the dynamic changes in population over time, statistical calculations are performed to calculate the proportion of elderly people in the total population of the area within each standard unit, i.e., the aging rate, and the ratio of the number of school-age children to the unit area, i.e., the school-age child density. At the same time, by analyzing the fluctuations and flow patterns of population numbers on weekdays and rest days, and during peak and off-peak periods, the tidal population coefficient, which reflects the temporal flow characteristics of the population, is quantitatively calculated. Finally, these three core indicators are integrated into a unique population dynamic profile label for each standard unit, comprehensively representing the population structure and spatiotemporal flow characteristics of each unit.

[0026] Land development gene coding is implemented by combining standardized coding and gradient boosting regression tree (GBDT) machine learning regression model with land ownership and current status data. The overall structure adopts a three-layer serial architecture, namely the label coding layer, feature modeling layer, and score prediction layer, with forward input-output relationships between the layers.

[0027] First, at the label coding layer, the ownership type of each land unit is standardized and assigned values ​​according to the land use classification standards of national land spatial planning and land ownership management regulations, generating ownership attribute labels. At the same time, classification and coding are performed based on the actual construction and use status of the land, generating current use labels. Then, at the feature modeling layer, ownership complexity, current building volume ratio, and number of historical issues are used as core prediction features. The above three types of features are preprocessed by numericalization and normalization to construct a unified-dimensional implementation difficulty prediction feature dataset, which is then divided into training and test sets in a 7:3 ratio. Finally, in the scoring prediction layer, a GBDT regression model was built, with the following core training parameters set: learning rate 0.1, number of decision trees 100, maximum tree depth 6, and minimum number of sample splits 2. The processed feature dataset was used as input, and the manually labeled actual land development implementation difficulty score was used as the supervision label. The model training and iterative optimization were completed. After the model accuracy was evaluated using the test set and met the requirements, the feature data of each land unit was input into the trained GBDT model. The model output a continuous quantitative value of 0 to 100 as the implementation difficulty score. Finally, the land ownership attribute label, current use label, and implementation difficulty score were integrated and encoded to form a complete land development gene label for each land unit.

[0028] Systematic digital processing of infrastructure service capacity data in the target area is conducted. First, according to the functional attributes and service areas of the infrastructure, various facilities are accurately defined and classified according to unified facility classification standards, generating facility type labels that clearly represent the functional positioning of the facilities. Then, core indicators such as the actual construction scale, service carrying capacity, and coverage of each facility are quantitatively statistically analyzed and digitally assigned values ​​to form service scale labels that reflect the facility's service supply capacity. At the same time, combined with the actual monitoring data of the facility's daily operation, the actual number of service recipients and resource usage are statistically analyzed to calculate the actual operating load, which is then compared with the facility's designed service capacity and rated load to obtain the actual load rate that accurately reflects the facility's operating load status. Finally, the facility type labels, service scale labels, and actual load rates are integrated to generate unique facility attribute labels for each facility, comprehensively and digitally representing the facility's function, supply capacity, and actual operating status.

[0029] In one possible implementation, step S220 further includes: The implementation difficulty score is predicted using a pre-set machine learning model based on the complexity of land unit ownership, current building volume ratio, and the number of historical issues.

[0030] Specifically, the implementation difficulty score is quantitatively predicted using a pre-defined Gradient Boosting Regression Tree (GBDT) machine learning model, combining three core features of the land unit: ownership complexity, current building volume ratio, and number of historical issues. First, ownership complexity is quantified and assigned values ​​based on the number of ownership entities, the clarity of boundaries, and the degree of nature difference. Current building volume ratio is calculated using actual survey data. The number of historical issues is statistically counted based on types such as land acquisition and demolition disputes, planning adjustments, and incomplete land use procedures. Then, the three types of feature data undergo normalization and standardization preprocessing to eliminate dimensional differences. The processed feature dataset is input into the GBDT regression model, which has been trained and validated using samples. The model, relying on the gradient boosting algorithm, iteratively trains multiple regression decision trees to perform weighted fitting calculations on the correlation between each feature and the difficulty of land development implementation. Finally, it outputs a continuous numerical value that accurately quantifies the ease or difficulty of land unit development implementation; this value is the implementation difficulty score for the corresponding land unit.

[0031] In one possible implementation, step S300 further includes: Step S310: Based on the population dynamic profile tag, the land development gene tag, and the facility attribute tag, define entity types, including population units, land units, and facility units, and set corresponding attribute fields for each entity type.

[0032] Step S320: Create supply and demand relationship edges between population units and facility units based on spatial distance and service standards.

[0033] Step S330: Create a bearing relationship edge between the land unit and the facility unit according to the land compatibility rules.

[0034] Step S340: Create reachability edges between population units and land units based on spatial adjacency.

[0035] Step S350: Using population units, land units, and facility units as nodes, and based on the supply and demand relationship edges, carrying relationship edges, and reachability relationship edges, form the ternary knowledge graph.

[0036] Specifically, using the previously extracted population dynamic profile tags, land development gene tags, and facility attribute tags as core data support, the entity type definition and attribute field configuration of the ternary knowledge graph are carried out. First, the three core entity types of population units, land units, and facility units are clearly defined as the basic nodes of the graph. Then, based on the tag system corresponding to each entity type, exclusive and complete attribute fields are matched and set for them. Among them, the attribute fields of the population unit accurately map the relevant indicators of the population dynamic profile tags such as aging rate, school-age child density, and tidal population coefficient. The attribute fields of the land unit fully cover the core content of the land development gene tags such as ownership attribute tags, current use tags, and implementation difficulty scores. The attribute fields of the facility unit fully correspond to the key indicators of the facility attribute tags such as facility type, service scale, and actual load rate. This achieves accurate association and one-to-one matching between the attribute fields of each entity type and the corresponding tags, laying a standardized entity data foundation for the subsequent construction of relationships between graph nodes.

[0037] Between population units and facility units with defined attribute fields, a supply-demand relationship edge is created based on both spatial distance calculation and facility service standards. First, according to industry standards and planning requirements for different facility types, the standard service radius and service allocation standards for various facilities such as education, medical care, and culture and sports are determined. Then, with each facility unit as the spatial center, combined with the actual road network data of the target area, the isochronous service range of each facility unit within the standard service radius is delineated through road network distance calculation. All population units falling within the isochronous range are accurately identified, and a preliminary association is established between the matched facility units and population units, creating a supply-demand relationship edge. At the same time, by combining the population dynamic profile label of the population unit and the facility type label of the facility unit, it is further determined whether the service demand and supply capacity of the two are compatible. If they are compatible, a demand matching degree weight is added to the corresponding supply-demand relationship edge to quantitatively represent the closeness of the supply-demand relationship between population units and facility units.

[0038] Between land units and facility units whose attributes have been defined, the creation of carrying capacity relationship edges is carried out strictly in accordance with land compatibility rules. First, a compatibility comparison table corresponding to various facility types and land use properties in the national land spatial planning is sorted out and obtained. This comparison table clearly divides the adaptation judgment results into three categories: compatible, incompatible, and incompatible. Then, all land units and facility units are traversed one by one, and the land use property field in the land development gene label of the land unit and the facility type field in the facility attribute label of the facility unit are extracted respectively. After the two fields are accurately matched, the compatibility between the two is judged one by one by referring to the compatibility comparison table. That is, carrying capacity relationship edges are created for land units and facility units whose judgment results are compatible. At the same time, based on the implementation difficulty score in the land development gene label of the land unit, a quantitative implementation feasibility weight is added to the corresponding carrying capacity relationship edge, so as to accurately characterize the carrying capacity compatibility and development implementation potential of the land unit to the facility unit.

[0039] Between population units and land units with completed attribute field configuration, reachability edges are created based on spatial adjacency as the core criterion. First, the boundaries are delineated by combining the spatial geographic data of the target area with standard units. A unified spatial distance threshold is preset as the adjacency judgment standard for population units and land units. Then, the spatial position relationship of each population unit and each land unit is accurately calculated one by one to check whether the spatial adjacent distance between the two meets the preset distance threshold requirement. For population units and land units whose spatial adjacency meets the judgment standard, a unique association relationship is established for them and a reachability edge is created to intuitively represent the spatial reachability between population units and land units.

[0040] Using population units, land units, and facility units with defined attribute fields as core nodes, and the previously constructed supply-demand relationship edges, carrying capacity relationship edges, and reachability relationship edges as the links between nodes, a three-dimensional knowledge graph is formed through knowledge graph modeling technology to achieve full-dimensional association integration. During the modeling process, the three types of entity nodes are accurately located and information is attached according to spatial attributes and attribute labels. At the same time, weighted supply-demand relationship edges, carrying capacity relationship edges, and reachability relationship edges are associated with the matching entity nodes, realizing the visualization and structured presentation of the supply-demand association between population units and facility units, the carrying capacity association between land units and facility units, and the reachability association between population units and land units. Finally, a three-dimensional knowledge graph with complete node information, clear edge relationships with quantitative weights, and accurate reflection of the inherent association and spatial coupling characteristics of people, land, and facilities is formed.

[0041] In one possible implementation, step S320 further includes: Step S321: Determine the standard service radius and service allocation standards for each facility type.

[0042] Step S322: Using the facility unit node as the center, calculate the isochronous circle range within the standard service radius based on the actual road network.

[0043] Step S323: Identify population unit nodes that fall within the isochronous circle range, and create supply and demand relationship edges between the facility unit nodes and the population unit nodes.

[0044] Step S324: Determine whether the population dynamic profile label of the population unit and the facility type label of the facility unit match. If they match, add a demand matching weight to the supply and demand relationship edge.

[0045] Specifically, by combining the overall land use plan, relevant national standards and industry norms for the allocation of public service facilities, and adapting to the actual situation of the target area, such as urban functional positioning, population density, and regional development characteristics, different types of infrastructure, such as education, medical care, culture and sports, elderly care, and commerce, are classified and assessed. The standard service radius corresponding to each type of facility is clearly defined, that is, the spatial range threshold that the facility can effectively cover and provide services. At the same time, based on the service attributes of each type of facility and the scale of regional population demand, the service allocation standards are determined, including the construction scale of the facility, service carrying capacity, per capita allocation indicators, and service population base, forming a set of facility service standards that are in line with the actual situation of the target area and clearly classified.

[0046] The isochronous circle range within the standard service radius is calculated with facility unit nodes as the center: First, the actual road network data of the target area, including road level, traffic direction, road segment length, and actual traffic speed, is imported into the GIS spatial analysis platform. Different traffic impedance weights are assigned to road segments of different levels and road conditions. Then, the determined standard service radius of each facility type is converted into road network traffic time thresholds. Starting from each facility unit node, the network isochronous circle generation algorithm of GIS is used to accurately calculate the multi-directional traffic distance and time along the actual road network topology. Finally, the spatial range that each facility unit can cover within the traffic time corresponding to the standard service radius is automatically delineated, that is, the isochronous circle range that conforms to the actual traffic conditions.

[0047] The isochronous circle range vector layer of each facility unit and the spatial point layer of the population unit nodes are imported into the GIS platform. Through the point-area matching algorithm of spatial overlay analysis, all population unit nodes whose spatial points fall within the isochronous circle vector range are accurately retrieved and identified. A list of successfully matched facility unit-population unit node pairs is exported simultaneously. This list of node pairs is then imported into the knowledge graph construction platform. Relying on the node association modeling tool of the graph, and according to the preset graph topology rules, corresponding association edges, i.e. supply and demand relationship edges, are automatically created for each set of matched facility unit nodes and population unit nodes in the list, realizing the direct transformation of geospatial matching results into knowledge graph topology relationships.

[0048] A matching rule algorithm library for facility type and population profile tags is constructed. Core rules are pre-defined, such as matching educational facilities with school-age child density tags, medical facilities with aging rate tags, and elderly care facilities with elderly population percentage tags. The algorithm extracts dynamic population profile tags from population units, such as aging rate and school-age child density, and compares them precisely with the facility type tags of the facility units, outputting a binary "match / mismatch" result. For node pairs determined to be matched, a linear weighted scoring algorithm is used to calculate the demand matching degree weight. First, the core indicator values ​​of the population profile tags, such as aging rate (50%) and school-age child density (20%), are normalized from 0 to 1. Then, differentiated weight coefficients are assigned to corresponding indicators based on the facility type, such as a weight coefficient of 0.8 for the aging rate of medical facilities and a weight coefficient of 0.2 for other auxiliary indicators. The quantification is completed using the formula: demand matching degree weight = Σ(normalized indicator value × weight coefficient). Finally, the calculated weight values ​​within the 0-1 range are written into the attribute fields of the supply and demand relationship edge, completing the construction of the supply and demand relationship edge with quantified weights.

[0049] In one possible implementation, step S330 further includes: Step S331: Obtain a compatibility comparison table between each facility type and land use nature. The compatibility comparison table includes three categories of judgment: compatible, compatible, and incompatible.

[0050] Step S332: Traverse all land units. For each land unit, obtain the land use property field from the corresponding land development gene tag.

[0051] Step S333: Traverse all facility units and obtain the corresponding facility type field for each facility unit.

[0052] Step S334: Based on the land use nature field and the facility type field, determine whether the facility type and land use nature are compatible according to the compatibility lookup table. If they are compatible, create a bearing relationship edge.

[0053] Step S335: For the bearing relationship edge, add an implementation feasibility weight, which is derived from the implementation difficulty score in the land development gene tag.

[0054] Specifically, firstly, based on the spatial land use classification standards, regional control detailed planning, and facility layout special planning, compatibility control clauses for various land use properties and facility types are extracted. Then, through rule structuring, facility types such as education, medical care, municipal administration, and cultural and sports facilities are matched with land use properties such as residential, commercial, public service, and industrial properties one by one, forming a digital comparison table containing three categories of judgment results: compatible, compatible, and incompatible. This table is then stored in the system rule base in the form of a key-value pair rule table, realizing a standardized compatibility judgment basis that can be directly called.

[0055] A full traversal of all land unit nodes included in the ternary knowledge graph is performed. The relevant attribute information of each unit is read sequentially according to the land unit number. The land use nature field is accurately extracted and parsed from the pre-bound land development gene tags to clarify the specific planning use of the land plot, such as residential land, public management and public service land, commercial service facilities land, industrial land, green space and square land, etc. This provides basic land use attribute data for subsequent compatibility matching with facility types.

[0056] A complete traversal of all facility unit nodes in the ternary knowledge graph is performed, and each facility unit is visited in a preset order. The corresponding facility type field is extracted from the associated facility attribute tags to clarify the specific category of the facility, such as education, medical care, elderly care, culture and sports, municipal administration, and transportation.

[0057] The land use nature field of the traversed land units is matched one by one with the facility type field of the facility units. Using a pre-established compatibility comparison table as the judgment standard, the current facility type and the corresponding land use nature are judged to determine whether they belong to the compatible category. If the judgment result is compatible, a carrying relationship edge is established between the land unit node and the facility unit node in the knowledge graph, thereby clarifying that the land can legally carry the construction layout of the corresponding facilities and forming a stable spatial carrying relationship.

[0058] The implementation difficulty score is extracted from the land development gene tag of the land unit, that is, the original score range is set to 0~100 points, where 0 points is the lowest development difficulty and 100 points is the highest development difficulty. The difficulty value is normalized to convert it into a standardized value in the range of 0~1 by normalizing the difficulty value = original difficulty score / 100. Then, the implementation feasibility weight is calculated by inverse mapping logic, the formula is implementation feasibility weight = 1 - normalized difficulty value, to ensure that the lower the development difficulty of the land unit, the higher the corresponding feasibility weight, with the weight range of 0~1. Finally, the edge attribute assignment algorithm of the knowledge graph is called to write the calculated weight value into the exclusive attribute field of the bearing relationship edge, to complete the quantitative binding of the weight, so that the bearing relationship edge has calculable and comparable feasibility quantitative characteristics.

[0059] In one possible implementation, step S340 further includes: For each population unit and each land unit, when the spatial adjacency between them meets a preset distance threshold, a reachability edge is created between the population unit and the land unit.

[0060] Specifically, for each population unit and each land unit in the knowledge graph, the actual spatial distance between their spatial geometric centers or boundaries is calculated one by one, and this distance is compared with a preset spatial adjacency distance threshold. When the spatial distance between the two is less than or equal to the threshold, it is determined that the adjacency reachability condition is met, and then a reachability relationship edge is established between the current population unit node and the land unit node, thereby structurally representing the spatial proximity and reachability relationship between the population unit and the land unit.

[0061] In one possible implementation, step S400 further includes: Step S410: Select the standard unit to be analyzed, obtain the corresponding standard population unit node, and extract the distribution data of various population groups from the population dynamic profile label.

[0062] Step S420: Based on the service configuration standards of various facilities within the standard unit and the distribution data, calculate the simulated demand values ​​for various facilities.

[0063] Step S430: Through the supply and demand relationship edges of the ternary knowledge graph, query all facility unit nodes associated with the standard population unit node, and accumulate the facility scale as the actual supply value.

[0064] Step S440: Calculate the difference between the simulated demand value and the actual supply value. If the difference is greater than a preset threshold, mark the standard unit as the facility shortage area and record the type and scale of the shortage facilities.

[0065] Step S450: Based on the type and scale of the shortage facilities, identify the land supply potential for the area with the shortage facilities and generate the candidate site list.

[0066] Specifically, the standard units to be analyzed are first selected in the spatial analysis platform according to community, grid or administrative boundary. The corresponding standard population unit node is located by spatial affiliation matching algorithm. Then, the quantity, density and spatial distribution data of various population groups such as permanent residents, school-age population, elderly population and tidal population are extracted from the population dynamic profile tag library associated with the node according to the field index, so as to complete the structured extraction of basic population information for demand calculation.

[0067] A service allocation standard library for different facility types is constructed. For example, educational facilities are allocated one kindergarten per 1,000 school-age children, medical facilities are allocated five beds per 1,000 people, and elderly care facilities are allocated ten beds per 1,000 people. Then, the extracted population distribution data of various types are calibrated. The calibrated population base is calculated as the original population size × the regional population mobility correction coefficient to correct for the impact of population dynamic changes. Finally, the core algorithm is used to calculate the simulated demand value of facilities as follows: the simulated demand value of facilities is calculated as the calibrated corresponding population base value × (facility allocation standard / 1000). The theoretical simulated demand values ​​of various facilities such as education, medical care, and elderly care are calculated separately. For example, if a unit has 2,000 school-age children after calibration, and the allocation standard is 60 places per 1,000 people, the simulated demand value of kindergartens is 2,000 × (60 / 1000) = 120 places, thus completing the quantitative calculation of demand values.

[0068] Starting with the standard population unit node, we traverse all supply and demand relationship edges connected to it in the ternary knowledge graph, retrieve all corresponding facility unit nodes through graph association query, and read the quantitative supply indicators such as service scale, number of beds, land area or class capacity of each facility unit one by one. We then arithmetically sum up the scale indicators of similar facilities to finally obtain the actual supply value of various facilities corresponding to the standard population unit.

[0069] For facilities of the same type, calculate the difference between the simulated demand value and the actual supply value, that is, subtract the actual supply value from the simulated demand value; if the difference is positive and greater than the preset shortage judgment threshold, then determine that the current standard unit is a shortage area of ​​this type of facility, mark the unit, and simultaneously record the corresponding shortage facility type and the quantitative shortage scale corresponding to the difference.

[0070] Based on the identified types and scales of infrastructure shortages, and taking the areas with infrastructure shortages as the core scope, the land supply potential is screened by combining the land use nature, feasibility weight, and spatial accessibility of land units in the ternary knowledge graph. Priority is given to land units that are compatible in land use, have low development difficulty, are spatially adjacent, and can meet the scale requirements of infrastructure construction. After comprehensive ranking, a candidate site list that meets the needs of infrastructure improvement is formed.

[0071] In one possible implementation, step S450 further includes: Step S451: Obtain all land unit nodes, sort them according to the implementation difficulty score in the land development gene tag, and select the top N with implementation difficulty scores less than the preset threshold as high-potential candidate site selection list units.

[0072] Step S452: For each high-potential candidate site selection list cell, query the reachable population cell nodes within the service radius through the reachability relation edges.

[0073] Step S453: Calculate the comprehensive index of facility shortage for the reachable population unit node, which is obtained by weighted summation based on the scale of various facility shortages.

[0074] Step S454: Match the high-potential candidate site list units with the population unit nodes according to their spatial adjacency to generate potential land parcel-demand unit matching pairs.

[0075] Step S455: Sort the potential land parcel-demand unit matching pairs from high to low according to the comprehensive facility shortage index to generate a candidate site list.

[0076] Specifically, the process involves traversing and acquiring all land unit nodes within the region, reading the implementation difficulty score from the land development gene tag corresponding to each land unit, sorting the scores in ascending order, and selecting the top N land units with implementation difficulty scores below the preset difficulty threshold and ranking high. These are then used as high-potential candidate site selection units with excellent development conditions and low implementation risks.

[0077] For each high-potential candidate site selection unit, spatial retrieval is performed using the reachability edges in the ternary knowledge graph, with the site as the center and the standard service radius of the corresponding facility as the range. All population unit nodes within the service coverage area that can be effectively served by the site are extracted and used as the service object set of the candidate site.

[0078] Preset weight coefficients are assigned to different types of shortage facilities such as education, medical care, and elderly care. Then, the shortage scale values ​​of each type of facility in each reachable population unit node are read separately. The shortage scale of each type of facility is multiplied by its corresponding weight and then summed. The comprehensive index is calculated by the formula Facility Shortage Comprehensive Index = Σ(Shortage Scale of a Certain Facility × Corresponding Weight). The higher the index value, the more prominent the facility supply gap of that population unit.

[0079] Based on the established spatial adjacency and service coverage relationships between high-potential candidate site selection units and population unit nodes, each candidate site is associated and paired with the corresponding population unit node within its service range to form a set of "potential site-demand unit matching pairs", providing a basic pairing relationship for subsequent site selection based on demand urgency.

[0080] Based on the comprehensive index of facility shortage corresponding to each potential plot-demand unit matching pair, the plots are arranged in descending order of index from high to low, and the candidate plots in the areas with the most urgent service demand are given priority, thus forming an orderly candidate site list that takes into account both development potential and service urgency.

[0081] Example 2, based on the same inventive concept as the infrastructure layout optimization method in the foregoing examples, such as... Figure 2As shown, this application provides an infrastructure layout optimization device. The device and method embodiments in this application are based on the same inventive concept. The device includes: The multi-source data acquisition module 10 is used to acquire multi-source data of the target area, including population activity data, land ownership and status data, and infrastructure status data.

[0082] The tag extraction module 20 is used to extract population dynamic profile tags, land development gene tags, and facility attribute tags at the standard unit scale based on the multi-source data.

[0083] The knowledge graph construction module 30 is used to construct a ternary knowledge graph containing people-land-facilities based on the population dynamic profile tags, the land development gene tags, and the facility attribute tags. The ternary knowledge graph uses population units, land units, and facility units as nodes, and supply and demand relationships, carrying relationships, and reachability relationships as edges.

[0084] The matching evaluation module 40 is used to perform supply and demand matching evaluation based on the ternary knowledge graph, and to identify areas with facility shortages and a list of candidate sites with implementation potential.

[0085] Furthermore, the device is also used to perform the following functions: The population activity data is subjected to spatiotemporal clustering analysis to extract dynamic population profile tags for each standard unit. These dynamic population profile tags include aging rate, school-age child density, and tidal population coefficient. The land ownership and current status data are encoded using land development genes to generate land gene tags for each land unit. These land gene tags include ownership nature tags, current use tags, and implementation difficulty scores. The infrastructure current status data is digitally processed to generate facility attribute tags for each facility. These facility attribute tags include facility type, service scale, and actual load factor.

[0086] Furthermore, the device is also used to perform the following functions: The implementation difficulty score is predicted using a pre-set machine learning model based on the complexity of land unit ownership, current building volume ratio, and the number of historical issues.

[0087] Furthermore, the device is also used to perform the following functions: Based on the dynamic population profile tags, the land development gene tags, and the facility attribute tags, entity types are defined, including population units, land units, and facility units, and corresponding attribute fields are set for each entity type. Supply and demand relationship edges are created between population units and facility units based on spatial distance and service standards; carrying capacity relationship edges are created between land units and facility units based on land compatibility rules; and reachability relationship edges are created between population units and land units based on spatial adjacency. Using population units, land units, and facility units as nodes, and based on the supply and demand relationship edges, carrying capacity relationship edges, and reachability relationship edges, the ternary knowledge graph is formed.

[0088] Furthermore, the device is also used to perform the following functions: Determine the standard service radius and service allocation standard for each facility type; calculate the isochronous circle range within the standard service radius based on the actual road network, with the facility unit node as the center; identify the population unit nodes falling within the isochronous circle range, and create a supply-demand relationship edge between the facility unit node and the population unit node; determine whether the population dynamic profile label of the population unit and the facility type label of the facility unit match, and if they match, add a demand matching degree weight to the supply-demand relationship edge.

[0089] Furthermore, the device is also used to perform the following functions: Obtain a compatibility lookup table between each facility type and land use nature, the compatibility lookup table including three categories: compatible, compatible, and incompatible; traverse all land units, for each land unit, obtain the land use nature field in the corresponding land development gene tag; traverse all facility units, for each facility unit, obtain the corresponding facility type field; based on the land use nature field and the facility type field, according to the compatibility lookup table, determine whether the facility type and land use nature are compatible, if compatible, create a carrying relationship edge; for the carrying relationship edge, attach an implementation feasibility weight, the implementation feasibility weight is derived from the implementation difficulty score in the land development gene tag.

[0090] Furthermore, the device is also used to perform the following functions: For each population unit and each land unit, when the spatial adjacency between them meets a preset distance threshold, a reachability edge is created between the population unit and the land unit.

[0091] Furthermore, the device is also used to perform the following functions: A standard unit to be analyzed is selected, the corresponding standard population unit node is obtained, and the distribution data of various population groups is extracted from the population dynamic profile label; according to the service allocation standards of various facilities within the standard unit, combined with the distribution data, the simulated demand value of various facilities is calculated; through the supply and demand relationship edge of the ternary knowledge graph, all facility unit nodes associated with the standard population unit node are queried, and the facility scale is accumulated as the actual supply value; the difference between the simulated demand value and the actual supply value is calculated, and if the difference is greater than a preset threshold, the standard unit is marked as the facility shortage area, and the shortage facility type and shortage scale are recorded; based on the shortage facility type and shortage scale, the land supply-side potential of the facility shortage area is identified, and the candidate site list is generated.

[0092] Furthermore, the device is also used to perform the following functions: All land unit nodes are acquired and sorted according to the implementation difficulty score in the land development gene tag. The top N nodes with implementation difficulty scores less than a preset threshold are selected as high-potential candidate site selection list units. For each high-potential candidate site selection list unit, accessible population unit nodes within the service radius are queried through reachability relationship edges. The facility shortage comprehensive index of the accessible population unit nodes is calculated, which is obtained by weighted summation based on the scale of various facility shortages. The high-potential candidate site selection list units are matched with the population unit nodes according to spatial adjacency to generate potential land parcel-demand unit matching pairs. The potential land parcel-demand unit matching pairs are sorted from high to low according to the facility shortage comprehensive index to generate a candidate site selection list.

[0093] It should be noted that the order of the embodiments described above is for descriptive purposes only and does not represent the superiority or inferiority of the embodiments. Specific embodiments of this specification have been described above. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0094] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0095] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. An infrastructure layout optimization method, characterized in that, include: Acquire multi-source data for the target area, including population activity data, land ownership and current status data, and infrastructure status data; Based on the multi-source data, population dynamic profile tags, land development gene tags, and facility attribute tags are extracted at the standard unit scale; Based on the aforementioned population dynamic profile tags, land development gene tags, and facility attribute tags, a ternary knowledge graph encompassing people, land, and facilities is constructed. This ternary knowledge graph uses population units, land units, and facility units as nodes, and supply-demand relationships, carrying capacity relationships, and accessibility relationships as edges. Specifically, it includes: Based on the population dynamic profile tags, the land development gene tags, and the facility attribute tags, entity types are defined, including population units, land units, and facility units, and corresponding attribute fields are set for each entity type; Between population units and facility units, supply and demand relationships are created based on spatial distance and service standards; Between land units and facility units, a bearing relationship edge is created according to land compatibility rules; Between population units and land units, reachability edges are created based on spatial adjacency. Using population units, land units, and facility units as nodes, and based on the supply and demand relationship edges, carrying capacity relationship edges, and reachability relationship edges, the ternary knowledge graph is formed. Based on the aforementioned ternary knowledge graph, a supply and demand matching assessment is conducted to identify areas with facility shortages and a list of candidate sites with implementation potential.

2. The infrastructure layout optimization method as described in claim 1, characterized in that, Based on the aforementioned multi-source data, population dynamic profile tags, land development gene tags, and facility attribute tags at the standard unit scale are extracted, including: Spatiotemporal clustering analysis is performed on the population activity data to extract population dynamic profile tags for each standard unit. The population dynamic profile tags include aging rate, school-age child density, and tidal population coefficient. The land ownership and current status data are encoded with land development genes to generate land gene tags for each land unit. The land gene tags include ownership nature tags, current use tags, and implementation difficulty scores. The infrastructure status data is digitally processed to generate facility service capacity labels for each facility. The facility attribute labels include facility type, service scale, and actual load rate.

3. The infrastructure layout optimization method as described in claim 2, characterized in that, The implementation difficulty score is predicted using a pre-set machine learning model based on the complexity of land unit ownership, current building volume ratio, and the number of historical issues.

4. The infrastructure layout optimization method as described in claim 1, characterized in that, Between population units and facility units, supply and demand relationships are created based on spatial distance and service standards, including: Determine the standard service radius and service allocation standards for each facility type; Using facility unit nodes as the center, calculate the isochronous circle range within the standard service radius based on the actual road network; Identify population unit nodes that fall within the isochronous circle range, and create supply and demand relationship edges between the facility unit nodes and the population unit nodes; Based on the population dynamic profile label of the population unit and the facility type label of the facility unit, determine whether the two match. If they match, add a demand matching degree weight to the supply and demand relationship edge.

5. The infrastructure layout optimization method as described in claim 1, characterized in that, Between land units and facility units, carrying capacity edges are created according to land compatibility rules, including: Obtain a compatibility comparison table between each facility type and the land use nature. The compatibility comparison table includes three categories of judgment: compatible, compatible, and incompatible. Iterate through all land units, and for each land unit, retrieve the land use property field from the corresponding land development gene tag; Iterate through all facility units, and for each facility unit, retrieve the corresponding facility type field; Based on the land use nature field and the facility type field, and according to the compatibility lookup table, it is determined whether the facility type and land use nature are compatible. If they are compatible, a bearing relationship edge is created. For the bearing relationship edge, an implementation feasibility weight is added, which is derived from the implementation difficulty score in the land development gene tag.

6. The infrastructure layout optimization method as described in claim 1, characterized in that, Between population units and land units, reachability edges are created based on spatial adjacency, including: For each population unit and each land unit, when the spatial adjacency between them meets a preset distance threshold, a reachability edge is created between the population unit and the land unit.

7. The infrastructure layout optimization method as described in claim 1, characterized in that, Based on the aforementioned ternary knowledge graph, a supply and demand matching assessment is performed to identify areas with facility shortages and a list of candidate sites with implementation potential, including: Select the standard unit to be analyzed, obtain the corresponding standard population unit node, and extract the distribution data of various population groups from the population dynamic profile label; Based on the service allocation standards of various facilities within the standard unit, and combined with the aforementioned distribution data, the simulated demand values ​​for various facilities are calculated. By using the supply and demand relationship edges of the ternary knowledge graph, query all facility unit nodes associated with the standard population unit node, and sum the facility scale as the actual supply value; Calculate the difference between the simulated demand value and the actual supply value. If the difference is greater than a preset threshold, mark the standard unit as the facility shortage area and record the type and scale of the shortage facilities. Based on the types and scale of the shortage facilities, the land supply potential of the shortage areas is identified, and the candidate site list is generated.

8. The infrastructure layout optimization method as described in claim 7, characterized in that, Based on the types and scale of the shortage facilities, land supply-side potential is identified for the areas with the shortage of facilities, and a candidate site list is generated, including: All land unit nodes are obtained and sorted according to the implementation difficulty score in the land development gene tag. The top N implementation difficulty scores with a preset threshold are selected as high-potential candidate site selection list units. For each high-potential candidate site selection list unit, query the reachable population unit nodes within the service radius through the reachability relation edges; Calculate the comprehensive index of facility shortage for the reachable population unit node, which is obtained by weighted summation based on the scale of various facility shortages; The high-potential candidate site selection list units are matched with the population unit nodes according to their spatial adjacency to generate potential land parcel-demand unit matching pairs; The potential land parcel-demand unit matching pairs are sorted from high to low according to the comprehensive facility shortage index to generate a candidate site list.

9. An infrastructure layout optimization device, characterized in that, The apparatus is used to implement the infrastructure layout optimization method according to any one of claims 1-8, the apparatus comprising: The multi-source data acquisition module is used to acquire multi-source data of the target area, including population activity data, land ownership and status data, and infrastructure status data. The tag extraction module is used to extract population dynamic profile tags, land development gene tags, and facility attribute tags at the standard unit scale based on the multi-source data. The knowledge graph construction module is used to construct a ternary knowledge graph containing people-land-facilities based on the population dynamic profile tags, the land development gene tags, and the facility attribute tags. The ternary knowledge graph uses population units, land units, and facility units as nodes, and supply and demand relationships, carrying relationships, and reachability relationships as edges. The matching and evaluation module is used to perform supply and demand matching evaluation based on the ternary knowledge graph, and to identify areas with facility shortages and a list of candidate sites with implementation potential.

Images