A Knowledge Graph-Based Evaluation and Optimization Method for Commercial Outlets in Urban Centers

Abstract

This invention discloses a method for evaluating and optimizing commercial outlets in urban central areas based on knowledge graphs, comprising the following steps: collecting and processing data on urban central areas nationwide to construct a knowledge graph of case urban central areas; based on the knowledge graph, using a random forest algorithm to construct a correlation model between population intensity and commercial outlet composition, and using a genetic algorithm to derive the optimal commercial outlet composition pattern; based on the commercial outlet composition data of the target urban central area, determining whether its commercial outlet composition needs adjustment according to the optimal commercial outlet composition pattern; if adjustment is needed, automatically solving and outputting the optimization adjustment results for various types of commercial outlets through the optimization model; outputting the optimization adjustment results and displaying them interactively on a digital space sandbox of the target urban central area, and outputting an adjustment report. This invention achieves precise adjustment of commercial outlets in central areas based on knowledge graphs, making the commercial composition of different central areas more reasonable.

Description

Technical Field

[0001] This invention relates to the field of urban planning, specifically to a method for evaluating and optimizing commercial locations in urban centers based on knowledge graphs. Background Technology

[0002] As the core area of ​​a city, the urban center is a vital carrier of bustling commercial activities, a hub of traffic flow, and a densely populated area of ​​social interaction. Accurately evaluating and rationally optimizing the development trends and status of commercial outlets in these areas is a key task in urban management and planning. Traditional outlet evaluation methods mainly rely on expert assessments, questionnaires, and simple statistical analysis. These methods are limited by single data sources, inconsistent subjective evaluation standards, and severely delayed response times. Similarly, in optimizing the layout of commercial outlets, although existing technologies provide basic optimization tools and methods, problems remain, including insufficient data integration, incomplete models, poor real-time performance, and excessive human intervention. Comprehensive and accurate evaluation and optimization of commercial outlets not only facilitates efficient resource allocation, sustainable economic growth, and environmentally friendly improvements but also enhances the quality of life, convenience, and well-being of the surrounding population. Furthermore, the acquisition of real-time, multi-source big data and the development of advanced algorithms and machine learning technologies have made the emergence of more accurate, efficient, dynamic, and comprehensive methods for evaluating and optimizing central areas possible. Based on knowledge graphs, random forest models, and genetic algorithms, this study evaluates the current business composition of urban centers and automatically solves and outputs adjustment patterns according to the optimal model, providing targeted layout optimization methods for different urban centers. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a knowledge graph-based method for evaluating and optimizing commercial outlets in urban centers. This method solves the problems of data fragmentation, insufficient correlation analysis, limitations of static evaluation, incomplete optimization suggestions, and difficulty in meeting personalized needs in traditional evaluation and optimization processes.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A knowledge graph-based method for evaluating and optimizing commercial outlets in urban centers includes the following steps:

[0006] Step 1: Data Collection, Processing, and Knowledge Graph Construction in the Urban Central Area

[0007] Data on urban population, commercial outlets, building vector data, and central area boundaries of cities across the country are acquired, input into a geographic information platform for format standardization, and then a database of case city central areas is formed.

[0008] Based on the database, three types of entities—people, commercial outlets, and central area—are identified and numbered. The location and quantity attributes of the people entity, the location, type, and area attributes of the commercial outlet entity, and the location and area attributes of the central area entity are calculated respectively. The relationship between the three types of entities is determined based on spatial location. The above results are input into the NEO4J graph database platform to construct a knowledge graph of the central area of ​​the case city.

[0009] Step Two: Constructing a Correlation Model Between Population Intensity and Commercial Outlets

[0010] Based on the knowledge graph, data on entities in each central area and their associated population and business outlet entities are extracted. The data is then aggregated and processed into a population-business association dataset containing population intensity data and business outlet composition data. A random forest algorithm is used to train an association model between population intensity and business outlet composition based on the dataset, and the optimal association model is determined using a grid search method. Based on the business outlet composition data and the obtained optimal association model, a genetic algorithm is used to solve for the optimal business outlet composition pattern.

[0011] Step 3: Evaluation of the Composition of Commercial Outlets in the City Center

[0012] Obtain a dataset of commercial outlet composition in the central area of ​​the target city, which includes the proportion and area information of various types of commercial outlets. Obtain the actual commercial outlet composition pattern in the central area and compare it with the determined optimal commercial outlet composition pattern. If the results are inconsistent, it is determined that the commercial outlet composition in the central area of ​​the target city needs to be adjusted.

[0013] Step 4: Optimization of the Composition of Commercial Outlets in the City Center

[0014] Based on the optimal proportion of each type of commercial outlet obtained from the optimal commercial outlet composition pattern and the commercial outlet composition data of the target city center, a commercial outlet optimization and adjustment model is established; the optimization and adjustment results of each type of commercial outlet are automatically solved and output through the optimization and adjustment model.

[0015] Step 5: Optimize the interactive display of the solution and adjust the report output.

[0016] The optimization and adjustment results are projected onto a large touch screen for interactive display based on the DataEase data visualization platform, and an adjustment report for commercial outlets in the target city center is generated.

[0017] Preferably, urban population data, commercial outlet data, building vector data, and central area boundary data of urban centers across the country are acquired, and then input into a geographic information platform for format standardization processing to form a database of case urban centers, specifically including:

[0018] Using the high-performance Intel Core i9-10980XE processor, the population point data, commercial outlet POI data, building vector data, and central area boundary data of urban centers across the country were deduplicated, cleaned, and standardized before being merged and input into the ArcGIS 10.8 geographic information platform. The data coordinates and data format types were unified to form a case urban center database, and the results were stored in a Samsung PM1733 NVMe SSD.

[0019] The population point data refers to vector data of population locations in urban centers across the country, obtained at 2-hour intervals from 9:00 to 22:00 over two consecutive weeks through the Baidu Insight Open Data Platform. This includes population information and latitude / longitude coordinates for each location. The commercial outlet POI data refers to POI data of commercial outlets in urban centers across the country, obtained from the official Dianping platform. This includes the commercial outlet name, category, building floor number, and latitude / longitude coordinates. The building vector data refers to building vector data obtained through the OpenStreetMap open data platform, including building area, number of floors, and latitude / longitude coordinates. The central area boundary data refers to existing maps containing central area boundary information obtained from the official websites of city governments, planning bureaus, or natural resources bureaus, and then drawn using spatial correction and drawing tools in the ArcGIS 10.8 geographic information platform, resulting in central area boundary data including area and latitude / longitude information.

[0020] Preferably, based on the database, three types of entities—people, commercial outlets, and central area—are identified and numbered. The location and quantity attributes of the people entity, the location, type, and area attributes of the commercial outlet entity, and the location and area attributes of the central area entity are calculated respectively. The relationships between the three types of entities are determined based on spatial location. The results are then input into the NEO4J graph database platform to construct a knowledge graph of the city's central area, specifically including:

[0021] Based on the population point data, commercial outlet POI data and central area boundary data in the city center of the case, three types of entities—population, commercial outlets and central area—were identified and extracted, and each was assigned a number.

[0022] The location and quantity attributes of population entities were determined by extracting latitude, longitude, and quantity information of population points from the ArcGIS 10.8 geographic information platform; the location attributes of commercial outlets were determined by extracting latitude, longitude, and longitude information from POI data; and the type information of commercial outlets was extracted and reclassified according to the classification standards for commercial service facilities land in the "Classification of Urban Land Use and Standards for Planning and Construction Land". Based on the reclassification results, the type attributes of commercial outlet entities were determined. The reclassification results specifically include seven categories: retail commerce, wholesale markets, catering services, hotel services, entertainment services, health and fitness services, and lifestyle services. The specific classification standards and characteristics are shown in the table below.

[0023]

[0024]

[0025] Extract the number of floors, area, and latitude / longitude information from the building vector data and the building floor information from the commercial outlet POI data. In the ArcGIS 10.8 geographic information platform, use spatial association tools to extract the building area of ​​the floors of buildings that intersect with the commercial outlets in geographic space, and combine this with field surveys for correction. Based on the final results, determine the area attribute of the commercial outlet entities. Extract the area information and latitude / longitude information from the central area boundary data to determine the area and location attributes of the central area entities.

[0026] The specific indicator names, symbol representations, and calculation methods for the three types of entities are shown in the table below:

[0027]

[0028] Based on the latitude and longitude information of the three types of entities, the spatial association tool in ArcGIS 10.8 is used to determine the population entities and commercial outlet entities located within the central area in geospatial space. The association relationships between the population entities, commercial outlet entities, and the central area entities are obtained based on the results. The specific association relationships and calculation methods are shown in the table below:

[0029]

[0030] The numbers of the three successfully associated entities were used as unique identifiers. The data structure was sorted out by combining the location coordinates and corresponding entity attributes. The processed entity, entity attribute and relationship data were imported into the Neo4j4.4 graph database platform using Cypher statements. The node types were determined to be people, commercial outlets and central area. The data was visualized using Neo4jBrowser and Bloom, and the knowledge graph of the city center of the case was constructed.

[0031] Preferably, based on the knowledge graph, data on entities in each central area and their associated population and business outlet entities are extracted. This data is then aggregated and processed into a population-business association dataset containing population intensity data and business outlet composition data, specifically including:

[0032] Based on the entity associations obtained in step two, the population entities associated with each central area entity and their quantity attributes are extracted, as are the commercial outlet entities associated with each central area entity and their type and area attributes. Using a multi-core, high-thread Intel Xeon Platinum 8280 processor, the per capita intensity index I corresponding to each central area entity is calculated. i Data and the area ratio of various types of commercial outlets R ij The data shows the area proportion (R) of each type of commercial outlet. ij Data aggregation forms the data on the composition of business outlets;

[0033] The per capita intensity index I i The specific calculation formula is as follows:

[0034]

[0035] Among them, I i P represents the population intensity index of the i-th central area entity; j This represents the number of people in the j-th population entity of the central area entity, where n is the total number of population entities associated with this central area entity; A 中 This represents the area attribute value of the entity in the central area; D represents the total number of days in the time period for which the population data was acquired; H represents the total number of times the population data was acquired within a single day in the time period of each day.

[0036] The area percentage of each type of commercial outlet R ij The specific calculation formula is as follows:

[0037]

[0038] Among them, R ij S represents the area percentage of the j-th type of commercial outlets in the i-th central area entity, where j represents, in order, retail, wholesale markets, catering, hotels, financial insurance, arts and media, other business, entertainment, health and fitness, gas stations, other public utility outlets, and other service facilities; j-k This represents the area attribute value of the kth commercial entity belonging to the j-th type of commercial outlet associated with the central area, where l is the total number of commercial outlets of that type associated with the central area; S m This represents the sum of the area attribute values ​​of all commercial outlets associated with the central area;

[0039] The above calculation results are transferred to the Samsung PM1733 NVMe SSD storage device to form a human-industry correlation dataset containing population intensity index data and business network composition data.

[0040] Preferably, a random forest algorithm is used to train a correlation model between population intensity and commercial outlet composition based on the dataset, and the optimal correlation model is determined using a grid search method. Based on the commercial outlet composition data and the obtained optimal correlation model, a genetic algorithm is used to solve for the optimal commercial outlet composition pattern, specifically including:

[0041] A random forest association model of population intensity and business outlets was constructed. The population-business association dataset was divided into a training set and a test set in a ratio of 8:2. Based on a workstation equipped with a high-performance NVIDIA A100 Tensor Core GPU, the training set was fed into the random forest model for training. The optimal parameters of the model were determined using a grid search method. The parameters specifically included the number of decision trees, the maximum depth of the decision trees, the minimum number of sample splits for each decision tree, and the minimum number of sample leaves for each decision tree. Based on the optimal parameters, the test set was fed into the trained model for validation, and finally, the optimal association model of population intensity and business outlets was obtained.

[0042] Using a genetic algorithm, the optimal association model obtained above and the commercial outlet composition data of the case city center are input. The configuration pattern of commercial outlets is represented in the form of real number encoding, where each individual (commercial outlet composition pattern) is represented by the proportion R of each type of commercial outlet. x Composition: Multiple individuals are randomly generated to form an initial population, with each individual representing a commercial outlet configuration pattern based on given constraints. Through repeated selection, crossover, and mutation operations, the population is continuously updated, ultimately yielding the optimal commercial outlet composition pattern {C1:C2:…:C n}

[0043] Preferably, the step of obtaining a dataset of commercial outlet composition in the target city center area, containing information on the proportion and area of ​​various types of commercial outlets, to obtain the actual commercial outlet composition pattern in the center area, comparing it with the determined optimal commercial outlet composition pattern, and if the results are inconsistent, determining that the commercial outlet composition in the target city center area needs to be adjusted, specifically includes:

[0044] Download the planning and design public notice documents for the central area of ​​the target city from the official website of the target city government, planning bureau, or natural resources bureau. These documents include a project description, design drawings, and other relevant documents containing economic and technical indicators of the central area planning scheme. Using a Python text analysis library, automatically identify and extract the percentage data and land area data of each type of commercial outlet in the planning scheme from the public notice documents to form a dataset of the commercial outlet composition in the central area of ​​the target city. Transfer this dataset to a Samsung PM1733 NVMe SSD. Based on the commercial outlet types described in step one, reclassify the dataset to ensure that the commercial outlet types in the target city are consistent with those described in step one. Based on the reclassified commercial outlet composition data, using a workstation equipped with a high-performance NVIDIA A100 TensorCore GPU, calculate the percentage R of the i-th type of commercial outlet in the central area of ​​the target city. 0i With land area S 0i ;

[0045] Based on the obtained R, the proportion of all types of commercial outlets in the target city center area 0i The actual business network structure pattern {C} is obtained by summarizing the data. 01 :C 02 :…:C 0n}, and compare it with the optimal business network structure pattern {C1:C2:…:C} determined in step two. n If the results are inconsistent, it is determined that the composition of commercial outlets in the central area of ​​the target city needs to be adjusted.

[0046] Preferably, based on the optimal proportion of each type of commercial outlet obtained from the optimal commercial outlet composition pattern and the commercial outlet composition data of the target city center, a commercial outlet optimization and adjustment model is established; the optimization and adjustment results of each type of commercial outlet are automatically solved and output through the optimization and adjustment model, specifically including:

[0047] Based on the optimal business network configuration pattern {C1:C2:…:C n The optimal proportion R of the i-th type of commercial outlets is determined. i The obtained land area S of the i-th type of commercial outlets in the target city center area 0i A business outlet optimization and adjustment model is constructed, which specifically includes:

[0048]

[0049] Where, k i Let be the adjustment coefficient for the i-th type of commercial outlet, satisfying the following constraints:

[0050]

[0051] When k i If the value is greater than 1, then the number of such commercial outlets needs to be increased; otherwise, they need to be reduced. i For the adjustment of land area for the i-th type of commercial outlet, when C i If the value is greater than 0, then it is determined that the land area for this type of commercial outlet needs to be increased by |C. i | Conversely, it is determined that the land area used by this type of commercial outlet needs to be reduced. |C i |;

[0052] Based on the aforementioned optimization and adjustment model and constraints, using a workstation equipped with a high-performance NVIDIA A100 TensorCore GPU, a greedy algorithm is employed to automatically solve for the optimal solution, and the optimal solution is determined based on the calculated adjustment coefficient k of the i-th type of commercial outlet. i Adjustment amount of land area C i As a result, the magnitude of the increase, decrease, and adjustment of land area for various types of commercial outlets in the central urban area of ​​the target city was finally determined.

[0053] Preferably, the optimization and adjustment results are projected onto a large touchscreen display based on the DataEase data visualization platform for interactive display, and a report on the adjustment of commercial outlets in the target city center is generated, specifically including:

[0054] Using the DataEase data visualization platform, the spatial distribution and attribute information of various commercial outlets in the target city center, as well as the optimization and adjustment results of the commercial outlet composition, are output in multiple formats such as charts, text, and maps and projected onto a large touch screen with a resolution of 3840×2160. Operators can browse and query the results before and after the optimization of the commercial outlet composition in the target city center or the specific attribute information of the commercial outlets through various gesture commands such as clicking and grabbing. The system also outputs a commercial outlet adjustment report, which includes a distribution map of commercial outlets in the target city center, evaluation results of the commercial outlet composition, commercial outlet adjustment plan and basis, and comparison information of indicators before and after the adjustment of the commercial outlet composition.

[0055] Beneficial effects:

[0056] (1) This invention provides a knowledge graph-based method for evaluating and optimizing commercial networks in urban centers. By constructing a case-based knowledge graph of urban centers containing three types of entities—people, commercial outlets, and the central area—it achieves systematic data integration and deep correlation analysis. This knowledge graph not only effectively integrates multi-source data but also provides a comprehensive view of entity relationships, enabling precise analysis of commercial outlet layout optimization based on structured data and complex spatial relationships. This method significantly improves the scientific rigor of commercial outlet evaluation, provides reliable decision support for urban planners, and ensures that optimization schemes meet actual needs and are highly implementable.

[0057] (2) This invention provides a knowledge graph-based method for evaluating and optimizing commercial networks in urban centers. It employs a random forest algorithm to construct a correlation model between population density and the composition of commercial outlets, and optimizes the configuration of these outlets using a genetic algorithm. This method leverages the comprehensive data of the knowledge graph and the accurate predictions of the random forest algorithm to effectively capture the complex relationship between population density and commercial outlets. Furthermore, the genetic algorithm intelligently searches for the optimal configuration within a broad solution space, significantly improving the accuracy and efficiency of commercial outlet evaluation. In addition, the method's dynamic adjustment capability ensures that the configuration of commercial outlets can adapt to market and environmental changes, thus providing an efficient, scientific, and flexible commercial outlet optimization solution.

[0058] (3) This invention provides a knowledge graph-based method for evaluating and optimizing commercial networks in urban centers. Based on the commercial network composition data of the target urban center and compared with the optimal commercial network composition pattern, it accurately determines whether the existing configuration needs to be adjusted. If adjustment is required, the optimization model will automatically solve and generate optimization adjustment results for various types of commercial networks. This method not only improves the accuracy of optimization judgment and reduces manual intervention, but also realizes the dynamic nature of commercial network evaluation and adjustment, enabling it to adapt to changes in the market and environment at any time. It also possesses foresight, ensuring that the configuration of commercial networks always meets future needs, thereby enhancing the efficient utilization of commercial resources and the sustainable vitality of the urban economy.

[0059] (4) This invention provides a knowledge graph-based method for evaluating and optimizing commercial networks in urban centers. Through the DataEase data visualization platform, the optimization results are projected onto a high-resolution touchscreen for interactive display, generating a detailed commercial network adjustment report. Users can easily browse and query data and specific attribute information of commercial networks before and after optimization using gestures such as clicking and grabbing. This visualization and report generation not only enhances the intuitiveness and interactivity of the data but also provides a comprehensive and systematic basis for decision-making, ensuring the transparency and effectiveness of the adjustment plan. Attached Figure Description

[0060] Figure 1 This is a flowchart of a method according to an embodiment of the present invention.

[0061] Figure 2 This is a schematic diagram of the distribution of commercial outlets in the central area before adjustment in an embodiment of the present invention.

[0062] Figure 3 This is a schematic diagram illustrating the identification of problematic business outlets in the central area according to an embodiment of the present invention.

[0063] Figure 4 This is a schematic diagram showing the adjusted distribution of commercial outlets in the central area according to an embodiment of the present invention. Detailed Implementation

[0064] The technical solutions of the embodiments of the present invention 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 the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the protection scope of the present invention.

[0065] like Figure 1 As shown in the figure, this invention discloses a method for evaluating and optimizing commercial outlets in urban centers based on knowledge graphs, including:

[0066] Data on urban populations, commercial outlets, building vector data, and central area boundaries of cities across the country are acquired, input into a geographic information platform for standardized formatting, and then a database of case city central areas is formed.

[0067] Based on the database, three types of entities—people, commercial outlets, and central area—are identified and numbered. The location and quantity attributes of the people entity, the location, type, and area attributes of the commercial outlet entity, and the location and area attributes of the central area entity are calculated respectively. The relationship between the three types of entities is determined based on spatial location. The above results are input into the NEO4J graph database platform to construct the knowledge graph of the city center of the case.

[0068] Based on the knowledge graph, data on entities in each central area and their associated population and business outlet entities are extracted. The data is then aggregated and processed into a population-business association dataset containing population intensity data and business outlet composition data. Using the random forest algorithm, an association model between population intensity and business outlet composition is trained on the dataset, and the optimal association model is determined using a grid search method. Based on the business outlet composition data and the obtained optimal association model, a genetic algorithm is used to solve for the optimal business outlet composition pattern.

[0069] Obtain a dataset of commercial outlet composition in the central area of ​​the target city, which includes the proportion and area information of various types of commercial outlets. Obtain the actual commercial outlet composition pattern in the central area and compare it with the determined optimal commercial outlet composition pattern. If the results are inconsistent, it is determined that the commercial outlet composition in the central area of ​​the target city needs to be adjusted.

[0070] Based on the optimal proportion of each type of commercial outlet obtained from the optimal commercial outlet composition pattern and the commercial outlet composition data of the target city center, a commercial outlet optimization and adjustment model is established; the optimization and adjustment results of each type of commercial outlet are automatically solved and output through the optimization and adjustment model.

[0071] The optimization and adjustment results are projected onto a large touch screen for interactive display based on the DataEase data visualization platform, and an adjustment report for commercial outlets in the target city center is generated.

[0072] Example

[0073] The technical solution of the present invention will be described in detail below using the central area of ​​Nanjing as an example.

[0074] Using the high-performance Intel Core i9-10980XE processor, the population point data, commercial outlet POI data, building vector data, and central area boundary data of urban centers across the country were deduplicated, cleaned, and standardized before being merged and input into the ArcGIS 10.8 geographic information platform. The data coordinates and data format types were unified to form a case urban center database, and the results were stored in a Samsung PM1733 NVMe SSD.

[0075] The population point data refers to vector data of population locations in urban centers across the country, obtained at 2-hour intervals from 9:00 to 22:00 over two consecutive weeks through the Baidu Insight Open Data Platform. This includes population count and latitude / longitude information for each location. The commercial outlet POI data refers to POI data of commercial outlets in urban centers across the country, obtained from the official Dianping platform. This includes the commercial outlet name, category, building floor number, and latitude / longitude information. The building vector data refers to building vector data obtained through the OpenStreetMap open data platform, including building area, number of floors, and latitude / longitude information. The central area boundary data refers to existing maps containing central area boundary information obtained from the official websites of city governments, planning bureaus, or natural resources bureaus, and drawn using spatial correction and drawing tools in the ArcGIS 10.8 geographic information platform, resulting in central area boundary data including area and latitude / longitude information.

[0076] Based on the population point data, commercial outlet POI data and central area boundary data in the city center of the case, three types of entities—population, commercial outlets and central area—were identified and extracted, and each was assigned a number.

[0077] The location and quantity attributes of population entities were determined by extracting latitude, longitude, and quantity information of population points from the ArcGIS 10.8 geographic information platform; the location attributes of commercial outlets were determined by extracting latitude and longitude information from POI data; the type information of commercial outlets was extracted and reclassified according to the classification standard for commercial service facilities land in the "Urban Land Classification and Planning Construction Land Standards," and the type attributes of commercial outlet entities were determined based on the reclassification results. The reclassification results specifically included seven categories: retail, wholesale markets, catering services, hotel services, entertainment services, health and fitness services, and living services; the number of floors, area, and latitude and longitude information of building vector data were extracted, along with the building floor information of the commercial outlet POI data. In the ArcGIS 10.8 geographic information platform, the building area of ​​the building floor intersecting with the commercial outlet in geographic space was extracted based on the building floor number using spatial association tools, and corrected by field surveys. The area attribute of commercial outlet entities was determined based on the final result; the area and latitude and longitude information of the central area boundary data were extracted to determine the area and location attributes of the central area entities.

[0078] Based on the latitude and longitude information of the three types of entities, the spatial association tool in ArcGIS 10.8 geographic information platform is used to determine the population entities and commercial outlet entities located in the central area in geographic space, respectively. Based on the results, the association relationship between the population entities, commercial outlet entities and the central area entities is obtained.

[0079] The numbers of the three successfully associated entities were used as unique identifiers. The data structure was sorted out by combining the location coordinates and corresponding entity attributes. The processed entity, entity attribute and relationship data were imported into the Neo4j4.4 graph database platform using Cypher statements. The node types were determined to be people, commercial outlets and central area. The data was visualized using Neo4jBrowser and Bloom, and the knowledge graph of the city center of the case was constructed.

[0080] Based on the entity associations obtained in step two, the population entities associated with each central area entity and their quantity attributes are extracted, as are the commercial outlet entities associated with each central area entity and their type and area attributes. Using a multi-core, high-thread Intel Xeon Platinum 8280 processor, the per capita intensity index I corresponding to each central area entity is calculated. i Data and the area ratio of various types of commercial outlets R ij The data shows the area proportion (R) of each type of commercial outlet. ij Data aggregation forms the data on the composition of business outlets;

[0081] The per capita intensity index I iThe specific calculation formula is as follows:

[0082]

[0083] Among them, I i P represents the population intensity index of the i-th central area entity; j This represents the number of people in the j-th population entity of the central area entity, where n is the total number of population entities associated with this central area entity; A 中 This represents the area attribute value of the entity in the central area; D represents the total number of days in the time period for which the population data was acquired; H represents the total number of times the population data was acquired within a single day in the time period of each day.

[0084] The area percentage of each type of commercial outlet R ij The specific calculation formula is as follows:

[0085]

[0086] Among them, R ij S represents the area percentage of the j-th type of commercial outlets in the i-th central area entity, where j represents, in order, retail, wholesale markets, catering services, hotels, entertainment services, health and wellness services, and lifestyle service facilities; j-k This represents the area attribute value of the kth commercial entity belonging to the j-th type of commercial outlet associated with the central area, where l is the total number of commercial outlets of that type associated with the central area; S m This represents the sum of the area attribute values ​​of all commercial outlets associated with the central area;

[0087] The above calculation results are transferred to the Samsung PM1733 NVMe SSD storage device to form a human-industry correlation dataset containing population intensity index data and business network composition data.

[0088] A random forest association model of population intensity and business outlets was constructed. The population-business association dataset was divided into a training set and a test set in a ratio of 8:2. Based on a workstation equipped with a high-performance NVIDIA A100 Tensor Core GPU, the training set was fed into the random forest model for training. The optimal parameters of the model were determined using a grid search method. The parameters specifically included the number of decision trees, the maximum depth of the decision trees, the minimum number of sample splits for each decision tree, and the minimum number of sample leaves for each decision tree. Based on the optimal parameters, the test set was fed into the trained model for validation, and finally, the optimal association model of population intensity and business outlets was obtained.

[0089] Using a genetic algorithm, the optimal association model obtained above and the commercial outlet composition data of the case city center are input. The configuration pattern of commercial outlets is represented in the form of real number encoding, where each individual (commercial outlet composition pattern) is represented by the proportion R of each type of commercial outlet. x Composition: Multiple individuals are randomly generated to form an initial population, with each individual representing a commercial outlet configuration pattern based on given constraints. Through repeated selection, crossover, and mutation operations, the population is continuously updated, ultimately yielding the optimal commercial outlet composition pattern {C1:C2:…:C n}

[0090] like Figure 2 Download the planning and design public notice documents for the central area of ​​Nanjing from the official website of the Nanjing Municipal Planning and Natural Resources Bureau. These documents include project descriptions, design drawings, and other relevant documents containing economic and technical indicators for the central area. Using a Python text analysis library, automatically identify and extract the percentage data and land area data of various types of commercial outlets in the central area of ​​Nanjing from the public notice documents, forming a dataset of the commercial outlet composition in the central area of ​​Nanjing. This dataset is then transferred to a Samsung PM1733 NVMe SSD. Based on the commercial outlet types described in step one, the dataset is reclassified to ensure that the commercial outlets in the central area of ​​Nanjing match the types described in step one. Based on the reclassified commercial outlet composition data, using a workstation equipped with a high-performance NVIDIA A100 Tensor Core GPU, calculate the percentage R of the i-th type of commercial outlet in the central area of ​​Nanjing. 0i With land area S 0i ;

[0091] like Figure 3 Based on the obtained R, the proportion of all types of commercial outlets in the central area of ​​Nanjing. 0i The actual business network structure pattern {C} is obtained by summarizing the data. 01 :C 02 :…:C 0n}, and compare it with the optimal business network structure pattern {C1:C2:…:C} determined in step two. n A comparison was made, and based on the comparison results, it was determined that the composition of commercial outlets in the central area of ​​Nanjing needed to be adjusted.

[0092] Based on the optimal business network configuration pattern {C1:C2:…:C n The optimal proportion R of the i-th type of commercial outlets is determined. i The obtained land area S of the i-th type of commercial outlets in the central area of ​​Nanjing 0i A business outlet optimization and adjustment model is constructed, which specifically includes:

[0093]

[0094] Where, k i Let be the adjustment coefficient for the i-th type of commercial outlet, satisfying the following constraints:

[0095]

[0096] When k i If the value is greater than 1, then the number of such commercial outlets needs to be increased; otherwise, they need to be reduced. i For the adjustment of land area for the i-th type of commercial outlet, when C i If the value is greater than 0, then it is determined that the land area for this type of commercial outlet needs to be increased by |C. i | Conversely, it is determined that the land area used by this type of commercial outlet needs to be reduced. |C i |;

[0097] like Figure 4 Based on the aforementioned optimization and adjustment model and constraints, using a workstation equipped with a high-performance NVIDIA A100 Tensor Core GPU, a greedy algorithm is employed to automatically solve for the optimal solution, and the optimal solution is determined based on the calculated adjustment coefficient k of the i-th type of commercial outlet. i Adjustment amount of land area C i As a result, the exact amount of increase or decrease in land area and the amount of adjustment for various types of commercial outlets in the central area of ​​Nanjing were determined.

[0098] Using the DataEase data visualization platform, the spatial distribution and attribute information of various commercial outlets in the central area of ​​Nanjing, as well as the optimization and adjustment results of the commercial outlet composition, are output in multiple formats such as charts, text, and maps and projected onto a large touch screen with a resolution of 3840×2160. Operators can browse and query the results before and after the optimization of the commercial outlet composition in the central area of ​​Nanjing or the specific attribute information of commercial outlets through various gesture commands such as clicking and grabbing. The system also outputs a commercial outlet adjustment report, which includes a distribution map of commercial outlets in the central area of ​​Nanjing, evaluation results of the commercial outlet composition, commercial outlet adjustment plan and basis, and comparison information of indicators before and after the adjustment of the commercial outlet composition.

Claims

1. A method for evaluating and optimizing commercial outlets in urban centers based on knowledge graphs, characterized in that, include: Step 1: Data Collection, Processing, and Knowledge Graph Construction in the Urban Central Area Data on urban population, commercial outlets, building vector data, and central area boundaries of cities across the country are acquired, input into a geographic information platform for format standardization, and then a database of case city central areas is formed. Based on the database, three types of entities—people, commercial outlets, and central area—are identified and numbered. The location and quantity attributes of the people entity, the location, type, and area attributes of the commercial outlet entity, and the location and area attributes of the central area entity are calculated respectively. The relationship between the three types of entities is determined based on spatial location. The above results are input into the NEO4J graph database platform to construct a knowledge graph of the central area of ​​the case city. Step Two: Constructing a Correlation Model Between Population Intensity and Commercial Outlets Based on the knowledge graph, data on entities in each central area and their associated population and business outlet entities are extracted. The data is then aggregated and processed into a population-business association dataset containing population intensity data and business outlet composition data. A random forest algorithm is used to train an association model between population intensity and business outlet composition based on the dataset, and the optimal association model is determined using a grid search method. Based on the business outlet composition data and the obtained optimal association model, a genetic algorithm is used to solve for the optimal business outlet composition pattern. Step 3: Evaluation of the Composition of Commercial Outlets in the City Center Obtain a dataset of commercial outlet composition in the central area of ​​the target city, which includes the proportion and area information of various types of commercial outlets. Obtain the actual commercial outlet composition pattern in the central area and compare it with the determined optimal commercial outlet composition pattern. If the results are inconsistent, it is determined that the commercial outlet composition in the central area of ​​the target city needs to be adjusted. Step 4: Optimization of the Composition of Commercial Outlets in the City Center Based on the optimal proportion of each type of commercial outlet obtained from the optimal commercial outlet composition pattern and the commercial outlet composition data of the target city center, a commercial outlet optimization and adjustment model is established. The optimization and adjustment model automatically solves and outputs the optimization and adjustment results for various types of commercial outlets. Step 5: Optimize the interactive display of the solution and adjust the report output. The optimization and adjustment results are projected onto a large touch screen for interactive display based on the DataEase data visualization platform, and an adjustment report for commercial outlets in the target city center is generated.

2. The method for evaluating and optimizing commercial outlets in urban centers based on knowledge graphs according to claim 1, characterized in that, In step one, urban population data, commercial outlet data, building vector data, and central area boundary data of urban central areas across the country are obtained, and then input into a geographic information platform for format standardization processing to form a database of case urban central areas, which specifically includes: Using the high-performance Intel Core i9-10980XE processor, the population point data, commercial outlet POI data, building vector data, and central area boundary data of urban centers across the country were deduplicated, cleaned, and standardized before being merged and input into the ArcGIS 10.8 geographic information platform. The data coordinates and data format types were unified to form a case urban center database, and the results were stored in a Samsung PM1733 NVMe SSD. The population point data refers to the vector data of population locations in urban centers across the country, obtained through the Baidu Insight Open Data Platform at 2-hour intervals, from 9:00 to 22:00 for two consecutive weeks. This data includes the number of people at each location and their latitude and longitude. The commercial outlet POI data refers to the commercial outlet POI data in urban centers across the country, obtained through the Dianping.com official platform. This data includes the commercial outlet's name, category, building floor number, and latitude and longitude. The building vector data refers to the building vector data obtained through the OpenStreetMap open data platform, specifically including building area, number of floors, and latitude and longitude information; the central area boundary data refers to the current map containing central area boundary information obtained from the official website of the city government, planning bureau, or natural resources bureau, and the central area boundary data containing area and latitude and longitude information is drawn in the ArcGIS 10.8 geographic information platform using spatial correction and drawing tools.

3. The method for evaluating and optimizing commercial outlets in urban centers based on knowledge graphs according to claim 2, characterized in that, In step one, based on the database, three types of entities—people, commercial outlets, and the city center—are identified and numbered. The location and quantity attributes of the people entity, the location, type, and area attributes of the commercial outlet entity, and the location and area attributes of the city center entity are calculated respectively. The relationships between the three types of entities are determined based on spatial location. The results are then input into the NEO4J graph database platform to construct a knowledge graph of the city center in the case study, specifically including: Based on the population point data, commercial outlet POI data and central area boundary data in the city center of the case, three types of entities—population, commercial outlets and central area—were identified and extracted, and each was assigned a number. The location and quantity attributes of population entities were determined by extracting latitude, longitude, and quantity information of population points from the ArcGIS 10.8 geographic information platform; the location attributes of commercial outlets were determined by extracting latitude and longitude information from POI data; the type information of commercial outlets was extracted and reclassified according to the classification standard for commercial service facilities land in the "Urban Land Classification and Planning Construction Land Standards," and the type attributes of commercial outlet entities were determined based on the reclassification results. The reclassification results specifically included seven categories: retail, wholesale markets, catering services, hotel services, entertainment services, health and fitness services, and living services; the number of floors, area, and latitude and longitude information of building vector data were extracted, along with the building floor information of the commercial outlet POI data. In the ArcGIS 10.8 geographic information platform, the building area of ​​the building floor intersecting with the commercial outlet in geographic space was extracted based on the building floor number using spatial association tools, and corrected by field surveys. The area attribute of commercial outlet entities was determined based on the final result; the area and latitude and longitude information of the central area boundary data were extracted to determine the area and location attributes of the central area entities. Based on the latitude and longitude information of the three types of entities, the spatial association tool in ArcGIS 10.8 geographic information platform is used to determine the population entities and commercial outlet entities located in the central area in geographic space, respectively. Based on the results, the association relationship between the population entities, commercial outlet entities and the central area entities is obtained. The numbers of the three successfully associated entities were used as unique identifiers. The data structure was sorted out by combining the location coordinates and corresponding entity attributes. The processed entity, entity attribute and relationship data were imported into the Neo4j4.4 graph database platform using Cypher statements. The node types were determined to be people, commercial outlets and central area. The data was visualized using Neo4jBrowser and Bloom, and the knowledge graph of the city center of the case was constructed.

4. The method for evaluating and optimizing commercial outlets in urban centers based on knowledge graphs according to claim 3, characterized in that, In step two, based on the knowledge graph, data on entities in each central area and their associated population and business outlet entities are extracted. This data is then aggregated and processed into a population-business association dataset containing population intensity data and business outlet composition data. Specifically, this includes: Based on the entity associations obtained in step two, the population entities associated with each central area entity and their quantity attributes are extracted, as are the commercial outlet entities associated with each central area entity and their type and area attributes. Using a multi-core, high-thread Intel Xeon Platinum 8280 processor, the per capita intensity index I corresponding to each central area entity is calculated. i Data and the area ratio of various types of commercial outlets R ij The data shows the area proportion (R) of each type of commercial outlet. ij Data aggregation forms the data on the composition of business outlets; The per capita intensity index I i The specific calculation formula is as follows: Among them, I i P represents the population intensity index of the i-th central area entity; j This represents the number of people in the j-th population entity of the central area entity, where n is the total number of population entities associated with this central area entity; A 中 This represents the area attribute value of the entity in the central area; D represents the total number of days in the time period for which the population data was acquired; H represents the total number of times the population data was acquired within a single day in the time period of each day. The area percentage of each type of commercial outlet R ij The specific calculation formula is as follows: Among them, R ij S represents the area percentage of the j-th type of commercial outlets in the i-th central area entity, where j represents, in order, retail, wholesale markets, catering services, hotels, entertainment services, health and wellness services, and lifestyle service facilities; j-k This represents the area attribute value of the kth commercial entity belonging to the j-th type of commercial outlet associated with the central area, where l is the total number of commercial outlets of that type associated with the central area; S m This represents the sum of the area attribute values ​​of all commercial outlets associated with this central area; The above calculation results are transferred to the Samsung PM1733 NVMe SSD storage device to form a human-industry correlation dataset containing population intensity index data and business network composition data.

5. The method for evaluating and optimizing commercial outlets in urban centers based on knowledge graphs according to claim 4, characterized in that, In step two, a random forest algorithm is used to train a correlation model between population intensity and commercial outlet composition based on the dataset, and the optimal correlation model is determined using a grid search method. Based on the commercial outlet composition data and the obtained optimal correlation model, a genetic algorithm is used to solve for the optimal commercial outlet composition pattern, specifically including: A random forest association model of population intensity and business outlets was constructed. The population-business association dataset was divided into a training set and a test set in a ratio of 8:

2. Based on a workstation equipped with a high-performance NVIDIA A100 Tensor Core GPU, the training set was fed into the random forest model for training. The optimal parameters of the model were determined using a grid search method. The parameters specifically included the number of decision trees, the maximum depth of the decision trees, the minimum number of sample splits for each decision tree, and the minimum number of sample leaves for each decision tree. Based on the optimal parameters, the test set was fed into the trained model for validation, and finally, the optimal association model of population intensity and business outlets was obtained. Using a genetic algorithm, the optimal association model obtained above and the data on the composition of commercial outlets in the case city center are input. The configuration pattern of commercial outlets is represented in the form of real number encoding, where each individual is composed of the proportion R of various types of commercial outlets. x Composition: Multiple individuals are randomly generated to form an initial population, with each individual representing a commercial outlet configuration pattern based on given constraints. Through repeated selection, crossover, and mutation operations, the population is continuously updated, ultimately yielding the optimal commercial outlet composition pattern {C1:C2:…:C n } 6. The method for evaluating and optimizing commercial outlets in urban centers based on knowledge graphs according to claim 5, characterized in that, Step three involves obtaining a dataset of commercial outlet composition in the target city center, containing information on the proportion and area of ​​various types of commercial outlets. This dataset yields the actual commercial outlet composition pattern in the center, which is then compared with the determined optimal commercial outlet composition pattern. If the results are inconsistent, it is determined that the commercial outlet composition in the target city center needs adjustment. Specifically, this includes: Download the planning and design public notice documents for the central area of ​​the target city from the official website of the target city government, planning bureau, or natural resources bureau. These documents include a project description, design drawings, and other relevant documents containing economic and technical indicators of the central area planning scheme. Using a Python text analysis library, automatically identify and extract the percentage data and land area data of each type of commercial outlet in the planning scheme from the public notice documents to form a dataset of the commercial outlet composition in the central area of ​​the target city. Transfer this dataset to a Samsung PM1733 NVMe SSD. Based on the commercial outlet types described in step one, reclassify the dataset to ensure that the commercial outlet types in the target city are consistent with those described in step one. Based on the reclassified commercial outlet composition data, use a workstation equipped with a high-performance NVIDIA A100 Tensor Core GPU to calculate the percentage R of the i-th type of commercial outlet in the central area of ​​the target city. 0i With land area S 0i ; Based on the obtained R, the proportion of all types of commercial outlets in the target city center area 0i The actual business network structure pattern {C} is obtained by summarizing the data. 01 :C 02 :…:C 0n }, and compare it with the optimal business network structure pattern {C1:C2:…:C} determined in step two. n If the results are inconsistent, it is determined that the composition of commercial outlets in the central area of ​​the target city needs to be adjusted.

7. The method for evaluating and optimizing commercial outlets in urban centers based on knowledge graphs according to claim 6, characterized in that, In step four, a commercial outlet optimization and adjustment model is established based on the optimal proportion of each type of commercial outlet obtained from the optimal commercial outlet composition pattern and the commercial outlet composition data of the target city center. The optimization and adjustment model automatically solves and outputs the optimization and adjustment results for various types of commercial outlets, specifically including: Based on the optimal business network configuration pattern {C1:C2:…:C n The optimal proportion R of the i-th type of commercial outlets is determined. i The obtained land area S of the i-th type of commercial outlets in the target city center area 0i A business outlet optimization and adjustment model is constructed, which specifically includes: Where, k i Let be the adjustment coefficient for the i-th type of commercial outlet, satisfying the following constraints: When k i If the value is greater than 1, then the number of such commercial outlets needs to be increased; otherwise, they need to be reduced. i For the adjustment of land area for the i-th type of commercial outlet, when C i If the value is greater than 0, then it is determined that the land area for this type of commercial outlet needs to be increased by |C. i | Conversely, it is determined that the land area used by this type of commercial outlet needs to be reduced. |C i |; Based on the aforementioned optimization and adjustment model and constraints, using a workstation equipped with a high-performance NVIDIA A100 Tensor Core GPU, a greedy algorithm is employed to automatically solve for the optimal solution, and the optimal solution is determined based on the calculated adjustment coefficient k of the i-th type of commercial outlet. i Adjustment amount C of land area i As a result, the magnitude of the increase, decrease, and adjustment of land area for various types of commercial outlets in the central urban area of ​​the target city was finally determined.

8. The method for evaluating and optimizing commercial outlets in urban centers based on knowledge graphs according to claim 7, characterized in that, In step five, the optimization and adjustment results are projected onto a large touchscreen display based on the DataEase data visualization platform for interactive display, and a report on the adjustment of commercial outlets in the target city center is generated, specifically including: Using the DataEase data visualization platform, the spatial distribution and attribute information of various commercial outlets in the target city center, as well as the optimization and adjustment results of the commercial outlet composition, are output in multiple formats such as charts, text, and maps and projected onto a large touch screen with a resolution of 3840×2160. Operators can browse and query the results before and after the optimization of the commercial outlet composition in the target city center or the specific attribute information of the commercial outlets through various gesture commands such as clicking and grabbing. The system also outputs a commercial outlet adjustment report, which includes a distribution map of commercial outlets in the target city center, evaluation results of the commercial outlet composition, commercial outlet adjustment plan and basis, and comparison information of indicators before and after the adjustment of the commercial outlet composition.

Images