Intelligent site selection method, device and equipment for industrial park project based on agent, and medium

Abstract

The present application belongs to the technical field of industrial project planning and site selection and land use, and provides an intelligent site selection method, device, equipment and medium for industrial park projects based on an agent, comprising: acquiring an industry classification dictionary and constructing an industry knowledge base, wherein knowledge source data comprises unstructured data and structured data; using an agent to analyze a park site selection request according to a special work flow, retrieving from the industry knowledge base according to an analysis result and an industry index mapping result to obtain a site selection scheme list; and using a grid growth method to optimize the site selection scheme list to obtain an optimized site selection scheme list, so that the park site selection efficiency and scientificity are improved.

Description

Technical Field

[0001] This invention relates to the field of industrial project planning, site selection, and land use technology, and in particular to an intelligent site selection method, apparatus, equipment, and medium for industrial park projects based on intelligent agents. Background Technology

[0002] Industrial project site selection typically requires a comprehensive consideration of factors such as industrial orientation, land use planning, land supply status, ecological management, geological safety, transportation and logistics, public facilities, and construction conditions. Current practices largely rely on manually collecting park data, planning maps, and land supply information, followed by staff screening candidate sites based on experience. However, this current method for industrial project site selection has the following shortcomings: (1) Poor industry adaptability: Site selection decisions only focus on single economic indicators such as land price and tax revenue per mu, without comprehensively considering the industrial park's industrial development characteristics and the matching degree of the project, resulting in weak industrial layout synergy and prominent contradictions between spatial planning and development construction; (2) Low work efficiency: Traditional site selection relies on manual collection of massive amounts of data and suitability analysis, with a data processing cycle of up to 15 days, and the determination of the park and plots takes more than one month, which is difficult to meet the needs of enterprises to quickly settle down; (3) Insufficient site selection capabilities: There is a lack of cross-modal fusion capabilities of multi-source heterogeneous data, and the site selection results are mostly presented as point elements, which cannot generate area plots that meet the requirements of enterprise land use scale and form, making it difficult to support actual investment decisions; (4) Low decision credibility: There is a lack of scheme comparison process based on quantitative analysis, and the scientificity and rationality of site selection rely on manual experience judgment, which cannot achieve traceability and verification of decisions.

[0003] Therefore, in the context of increasingly stringent land and space planning control and the strengthening demand for industrial agglomeration development, the site selection of industrial projects still faces the following bottlenecks: insufficient semantic analysis of site selection needs and structured rule mapping; lack of quantitative representation and matching of spatial-industrial linkages; insufficient ability to generate multi-constraint coupled schemes; and lack of traceable basis for scheme comparison results. Summary of the Invention

[0004] Aimed at at least in solving one of the technical problems existing in the prior art, the present invention provides a method, apparatus, equipment and medium for intelligent site selection of industrial park projects based on intelligent agents.

[0005] One aspect of the present invention provides an intelligent site selection method for industrial park projects based on intelligent agents, comprising: Based on the user's request for park site selection, obtain the park's industry classification information, map and expand the industry classification information to obtain an industry classification dictionary; Acquire knowledge source data from the park and construct an industry knowledge base based on the knowledge source data, which includes unstructured data and structured data; A specific workflow is constructed. Based on the specific workflow, an intelligent agent is used to analyze the park site selection request. Based on the analysis results and the industry indicator mapping results, the park focusing results are retrieved from the industry knowledge base. The focus results of the park are extracted using a grid growth method to obtain the available site regions. An optimization algorithm is then used to process the available site regions to obtain a list of optimized site selection schemes.

[0006] According to the intelligent site selection method for industrial park projects based on intelligent agents, the method obtains the industrial classification information of the park based on the user's park site selection request, maps and expands the industrial classification information to obtain an industrial classification dictionary, including: Obtain the industrial classification information of the park, which includes major industrial categories, industrial subcategories, and common core needs; Natural language processing is used to map core keywords, synonyms, and upstream and downstream related terms to industry classification information, resulting in an industry classification dictionary.

[0007] According to the intelligent site selection method for industrial park projects based on intelligent agents, knowledge source data of the park is acquired, and an industry knowledge base is constructed based on the knowledge source data. The knowledge source data includes unstructured data and structured data, including: Acquire knowledge source data for the park, including unstructured data such as policy data and encyclopedia entries related to the park, and structured data such as park lists and statistical yearbook data; The knowledge source data is processed by text segmentation and normalization to obtain data slices. Text segmentation includes word segmentation, stop word removal, and part-of-speech tagging of the knowledge source data. Normalization includes mapping of multi-source data and anonymization of privacy data. The obtained data slices are converted into vector data and stored as triples using MySQL to obtain the industry knowledge base.

[0008] According to the intelligent site selection method for industrial park projects based on intelligent agents, a specific workflow is constructed. Based on this workflow, an intelligent agent is used to parse the park site selection request. The parsing results and industry indicator mapping results are then retrieved from the industry knowledge base to obtain the park focusing results, including: Use Dify to build and orchestrate specific workflows; The site selection request for the park was parsed using natural language processing according to a specific workflow, and the parsing results were obtained. The vectorized parsing results are matched with the industry knowledge base for similarity, and prompt words are generated based on the matching results and the vectorized parsing results. The system employs one of the following methods: precise matching based on prompt words, fuzzy demand processing, and industry classification dictionary matching, to obtain the park-focused results.

[0009] According to the intelligent site selection method for industrial park projects based on intelligent agents, the park focusing results are extracted using a grid growing method to obtain a selectable site area. An optimization algorithm is then used to process the selectable site area to obtain a list of optimized site selection schemes, including: The park in the focused results is divided into multiple regular grid units. The basic information of the grid units is overlaid with the park site selection indicators to obtain the attribute information of the grid units. The basic information includes the center coordinates, serial number and row and column number; the attribute information includes land, three zones and three lines, prohibited construction areas, slope, industry category, main road, ultra-high voltage line, geological disaster risk area and pollution source. Based on the basic and attribute information of the grid cells, a multi-objective solution is performed using a multi-objective function and multiple constraints with corresponding weight parameters to obtain the comprehensive location attraction score for each grid cell. The weight parameters are calculated using TOPSIS. Based on the comprehensive location attractiveness score of each grid cell, grid cells that avoid prohibitive constraints are identified, and a cyclic search is performed using a grid growth method to obtain a set of grid cells. Based on the total area and overall site attractiveness score of the park's site selection requests, the grid unit set is filtered and sorted to obtain a list of optimized site selection schemes.

[0010] According to the intelligent site selection method for industrial park projects based on intelligent agents, the grid cell set is filtered and sorted based on the total area of ​​the park site selection requests and the total comprehensive site selection attractiveness score to obtain a list of optimized site selection schemes, including: Using the existing construction land in the park as the site selection plot, name matching is performed based on the site selection plot and the park focus results to obtain the corresponding industrial park development direction area and the actual built-up area of ​​the industrial park. Spatial analysis is conducted based on the development direction zone of the industrial park, the actual built-up area of ​​the industrial park, and the pre-set industrial park planning zones to obtain industrial land plots. Spatial analysis was performed on industrial land parcels, net land registration data, and land already supplied data to obtain candidate parcels; Based on the total area of ​​the park site selection request, the existing construction land parcels in the park are overlaid with the selected parcels for analysis to obtain candidate plots; The comparison index of grid cells of candidate plots is calculated based on attribute information. The candidate plots are then screened based on the sum of the products of the corresponding weights of the comparison indexes to obtain a list of optimized site selection schemes.

[0011] According to the intelligent site selection method for industrial park projects based on intelligent agents, the multi-objective function and the multiple constraints are determined based on attribute information and basic information. The multi-objective function and the multiple constraints include whether it falls within the three zones and three lines, whether it includes existing construction land markers, spatial planning zoning compliance, terrain adaptability, and environmental safety.

[0012] Another aspect of the present invention provides an intelligent site selection device for industrial park projects based on intelligent agents, comprising: The first module is used to obtain the industrial classification information of the park based on the user's park site selection request, and to map and expand the industrial classification information to obtain an industrial classification dictionary; The second module is used to acquire knowledge source data of the park and build an industry knowledge base based on the knowledge source data, which includes unstructured data and structured data. The third module is used to construct a special workflow. Based on the special workflow, an intelligent agent is used to analyze the park site selection request. Based on the analysis results and the industry indicator mapping results, the park focusing results are retrieved from the industry knowledge base. The fourth module is used to extract the focalization results of the park using a grid growth method to obtain the available site areas, and to process the available site areas using an optimization algorithm to obtain a list of optimized site selection schemes.

[0013] Another aspect of the present invention provides an electronic device, including a processor and a memory; The memory is used to store programs; The processor executes the program to implement the method as described above.

[0014] This invention also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the methods described above.

[0015] The beneficial effects of this invention are as follows: It integrates an AI industry knowledge base with site selection, constructing a knowledge base for analyzing the suitability of industrial park site selection. Through AI assistant question-and-answer sessions, it quickly recommends suitable industrial parks for specific types of projects and associates them with site selection rules. By constructing a grid and recording the element attributes associated with the site selection rules within the grid, it extracts suitable grids step-by-step according to a path of first avoiding and then optimizing. Then, it merges continuous grids to generate site selection schemes for planar units, and finally generates site selection schemes that meet the area requirements. Furthermore, it constructs a comprehensive comparative selection model that integrates land use, planning, geological disaster risks, distance from ultra-high voltage lines / schools, land-saving level, intensive utilization, surrounding facilities, and location conditions, thereby improving the scientific nature of industrial park site selection. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the intelligent site selection process for industrial park projects based on intelligent agents, according to an embodiment of the present invention. Figure 2 This is an AI framework diagram for focusing on industrial parks according to an embodiment of the present invention; Figure 3 This is a flowchart illustrating the knowledge base construction process according to an embodiment of the present invention; Figure 4 This is a flowchart of the overall AI assistant in an embodiment of the present invention; Figure 5 This is a schematic diagram of the process for generating the park focusing result according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the process for generating a list of optimized location selection schemes according to an embodiment of the present invention; Figure 7 This is a schematic diagram of the mesh growth method according to an embodiment of the present invention, wherein (a) is the initial state of the mesh, (b) is the connected state marked when the mesh grows for the first time, (c) is the connected state marked when the mesh grows to the right for the second time, (d) is the connected state marked when the mesh grows downward for the third time, and (e) is the connected region composed of the marked meshes obtained after the mesh grows four times. Figure 8 This is a schematic diagram of another optimized location selection scheme list generation process according to an embodiment of the present invention; Figure 9 These are flowcharts illustrating two optimized paths according to embodiments of the present invention; Figure 10 This is a schematic diagram of an intelligent site selection device for industrial park projects based on intelligent agents, according to an embodiment of the present invention. Detailed Implementation

[0017] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings. Throughout the description, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions. In the following description, suffixes such as "module," "part," or "unit" used to denote elements are used only for the purpose of illustrative purposes and have no specific meaning in themselves. Therefore, "module," "part," or "unit" can be used interchangeably. Terms such as "first," "second," etc., are used only to distinguish technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the sequential relationship of the indicated technical features. In the following description, the consecutive reference numerals for method steps are for ease of review and understanding. Adjusting the implementation order of steps, in conjunction with the overall technical solution of the present invention and the logical relationship between the various steps, will not affect the technical effect achieved by the technical solution of the present invention. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0018] Please refer to Figure 1 and Figure 2 , Figure 1 This is a schematic diagram of the intelligent site selection process for industrial park projects based on intelligent agents according to an embodiment of the present invention. It mainly includes parsing the project requirements into standard industry codes and site selection rule vectors; retrieving candidate parks using a park-industry-indicator knowledge base; performing gridding, constraint coding, and indicator scoring on the multi-source spatial layers within the candidate parks; generating candidate areal plots through grid growth paths and existing construction land extraction paths, respectively; and then performing fusion deduplication, comprehensive scoring, and report output on the candidate plots. Figure 2 This is an AI framework diagram for intelligent site selection and scheme comparison of industrial park projects according to an embodiment of the present invention.

[0019] based on Figure 2 The AI ​​framework in this embodiment of the invention utilizes a routing architecture orchestrated by Dify, featuring three types of intent recognition, to achieve precise focusing on industrial parks. This architecture, through a logical pipeline of intent recognition, industry mapping, and dual-path retrieval, achieves accurate conversion from non-standard user queries to standardized park entities. To meet the requirements of downstream system calls and business rules, the workflow employs structured processing logic at the output end, as shown in Table 1. Table 1 Structured Processing Logic Table

[0020] Specific execution steps and logic: (1) NLP word segmentation and intent segmentation: QuestionClassifier identifies user intent; if it is a park recommendation, it enters the industry matching link; industry category determination: LLM4 node extracts industry type and semantically aligns it with the industry category directory in the environment variables; knowledge base retrieval: retrieval in the JSON feature library; for example, if it matches the general equipment manufacturing industry, the system will locate the relevant park with code 34.

[0021] (2) Core strategy domain, including heuristic sorting logic: Provincial capital city priority: When synthesizing responses, parks in a certain city will be automatically listed first (e.g., "High-tech Industrial Development Zone of a certain city"); Level priority: Under the same conditions, national-level parks (level: national) have higher priority than provincial-level parks; Industry orientation: For example, for the silicon steel industry, a mandatory first-choice logic is implemented, and "Industrial Development Zone of a certain city" will be recommended and pinned to the top.

[0022] Structured parameter output: Atomic data cleaning, using the Parameter Extractor to remove redundant descriptions with LLM instructions; clean output, outputting only a JS Array format array of park names, and simultaneously generating an output_type:"park factory" tag for direct use by front-end visualization components.

[0023] (3) Data storage design for site selection scenario optimization: The AI ​​framework of this embodiment adopts an atomic storage approach optimized for site selection scenario in the retrieval layer, deconstructing the park attributes and industry relationships: administrative / level index, storing the city or state to which the park belongs (such as "City A", "City B") and administrative level, as the first-level filtering Boolean condition during retrieval; atomic association mapping, the system decomposes the many-to-many relationship through the logic of Industry_Relation_Graph, for example (a certain city's economic and technological development zone [Subject], leading industry [Predicate], special equipment manufacturing [Object, Code:35]); data mapping constraints, the Object end is forced to map to the standard industry category code, for example, when NLP extracts "automotive parts", the system automatically points to "automotive manufacturing (Code:36)", thus accurately matching "City A Economic and Technological Development Zone" or "City B Economic and Technological Development Zone".

[0024] It can be determined that after focusing on the industrial park, the AI ​​framework of this embodiment can generate a site selection scheme within the park through the grid growth method in subsequent embodiments, and then optimize the site selection scheme through the optimization site selection model, so as to realize intelligent site selection and scheme comparison for the overall industrial park project.

[0025] Further reference Figure 1 ,in Figure 1Including but not limited to steps S100~S400: S100 obtains the industrial classification information of the park based on the user's park site selection request, maps and expands the industrial classification information, and obtains an industrial classification dictionary.

[0026] In some embodiments, the industrial classification information of the park is obtained, which includes major industrial categories, industrial subcategories, and common core needs; natural language processing is used to map the industrial classification information to core keywords, synonyms, and upstream and downstream related terms to obtain an industrial classification dictionary.

[0027] For example, as shown in Table 2 below, the industry classification information includes industry data collection and preprocessing. For instance, an industrial park in a certain province has 43 industry categories. These 43 industries are divided into 8 major categories based on their common core needs, and then further refined to specific industries. Each major industry category corresponds to a set of core indicator templates. Specific industries can add / delete 1-2 exclusive indicators to avoid redundant configuration.

[0028] Table 2 Industry Classification Table

[0029] S200 acquires knowledge source data from the park and constructs an industry knowledge base based on the knowledge source data, which includes unstructured data and structured data.

[0030] In some embodiments, reference Figure 3 The flowchart for knowledge base construction shown below Figure 4 The overall flowchart of the AI ​​assistant shown includes: Acquire knowledge source data for the industrial park, including unstructured data such as policy data and encyclopedia entries related to the park, and structured data such as park lists and statistical yearbook data. Perform text segmentation and normalization on the knowledge source data to obtain data slices. Text segmentation includes word segmentation, stop word removal, and part-of-speech tagging. Normalization includes mapping of multi-source data and anonymizing of privacy data. Convert the obtained data slices into vector data and store them as triples using MySQL to obtain the industry knowledge base.

[0031] In some embodiments, text segmentation includes operations such as word segmentation, stop word removal, and part-of-speech tagging of text data, setting a segmentation strategy based on the content of the text, and extracting key entities such as place names through entity recognition technology.

[0032] In some embodiments, normalization processing includes performing data normalization processing to unify the format, units, and encoding standards of data from different sources; for cross-domain data, mapping relationships need to be established to ensure that terms from different domains can be correctly understood and associated; in addition, sensitive information, such as personal privacy data, needs to be anonymized to ensure data security.

[0033] In some embodiments, the knowledge graph storage portion adopts a MySQL solution, implementing triple (entity-relationship-attribute) storage through entity tables and relation tables. The entity table stores node information (such as entity_id, name, type, and JSON attributes), while the relation table stores edge information (such as source_id, target_id, relation type, and attributes). Single-hop relationships are queried through JOIN operations, and multi-hop path traversal is achieved using recursive CTEs in MySQL 8.0+, supplemented by index optimization for performance. JSON fields support dynamic attribute storage, meeting the storage needs of small to medium-sized knowledge graphs.

[0034] S300 constructs a special workflow, uses an intelligent agent to analyze the park site selection request based on the special workflow, and retrieves the park focus result from the industry knowledge base based on the analysis result and the industry indicator mapping result.

[0035] In some embodiments, reference Figure 5 The diagram showing the process for generating the focused results of the park includes, but is not limited to, steps S310 to S340: S310 uses Dify to build and orchestrate specialized workflows; S320 uses natural language processing to parse the park site selection request according to the specific workflow to obtain the parsing result; S330: The vectorized parsing results are matched with the industry knowledge base for similarity, and prompt words are generated based on the matching results and the vectorized parsing results. S340 employs a large model that uses one of the following methods: precise matching based on prompt words, fuzzy demand processing, and industry classification dictionary matching, to obtain the park's focused results.

[0036] For example, based on the general industrial intelligent site selection requirements, a special workflow is orchestrated through the Dify platform. Following the basic process of "user submits requirements - analyzes and classifies requirements - retrieves knowledge base - organizes and returns results - presents to the front end", multi-branch logic orchestration and refined data flow are realized. Through automated scripts, API calls, service calls, etc., the system is executed in a preset order and logic, ensuring the system's accurate identification and response to complex user intentions.

[0037] S400 uses a grid growth method to extract the focus results of the park to obtain the available site areas. An optimization algorithm is then used to process the available site areas to obtain a list of optimized site selection schemes.

[0038] refer to Figure 6 The flowchart shown above illustrates the process of generating the list of optimized site selection schemes, which includes, but is not limited to, steps S410 to S440: S410 divides the park in the park focusing result into multiple regular grid units. Based on the basic information of the grid units and the park site selection indicators, the attribute information of the grid units is obtained by overlay analysis. The basic information includes the center coordinates, serial number identifier and row and column number. For example, the attribute information includes land, three zones and three lines, prohibited construction areas, slope, industry category, main road, ultra-high voltage line, geological disaster risk area and pollution source.

[0039] S420, based on the basic and attribute information of the grid cells, performs multi-objective solutions using a multi-objective function and multiple constraints with corresponding weight parameters to obtain the comprehensive location attractiveness score of each grid cell, where the weight parameters are calculated using TOPSIS.

[0040] In some embodiments, the multi-objective function and multiple constraints are determined based on attribute information and basic information. The multi-objective function and multiple constraints include whether it falls within the three zones and three lines, whether it includes existing construction land markings, spatial planning zoning compliance, terrain adaptability, and environmental safety.

[0041] S430: Based on the comprehensive location attractiveness score of each grid cell, grid cells that avoid prohibitive constraints are avoided. A grid growth method is used to perform a cyclic search to obtain a set of grid cells.

[0042] For example, refer to Figure 7The diagram illustrates the grid growth method, where (a) represents the initial state of the grid, (b) represents the connected state marked during the first grid growth, (c) represents the connected state marked during the second grid growth to the right, (d) represents the connected state marked during the third grid growth downwards, and (e) represents the connected region formed by the marked grids after four growths. The "2" indicates a network connectivity state. Subsequent area calculations and scheme optimization are performed using the obtained connected regions. This embodiment focuses on generating areal features (plot boundaries): For each preferred location grid, a grid growth algorithm is used to calculate whether the grid is selectable. The algorithm prioritizes avoiding prohibited constraint grids (C_prohibit_i=0, C_existing_i=0). Other grids are extractable units. Adjacent spatial units that meet other constraint conditions (avoidance areas) are extracted, and the extracted units are connected regions. The judgment method is as follows: Starting from the first grid of the region, the algorithm iterates outwards, marking connected grids as 2, until no connected grids are found, at which point growth stops. This process is repeated until all grids at the park boundary have been traversed. Then, the area of ​​the connected grid regions is calculated to determine whether the total area of ​​the extracted cells is within the user-required range [A_min, A_max]. Finally, the comprehensive site selection attractiveness scores (A_i) of the extracted grid sets that meet the conditions are ranked, and the top 5-10 candidate schemes are extracted. Finally, these 5-10 candidate schemes are comprehensively calculated, and the grids of each candidate scheme are merged and integrated to generate vector data.

[0043] S440: Based on the total area of ​​the park's site selection requests and the overall comprehensive site selection attractiveness score, the grid unit set is filtered and sorted to obtain a list of optimized site selection schemes.

[0044] In some embodiments, reference Figure 8 The diagram shown illustrates another process for generating a list of optimized site selection schemes, which includes, but is not limited to, steps S450 to S490: S450 uses the existing construction land in the park as the site selection plot. Based on the site selection plot and the park focus results, name matching is performed to obtain the corresponding industrial park development direction area and the actual built-up area of ​​the industrial park. S460: Based on the development direction area of ​​the industrial park and the actual built-up area of ​​the industrial park, spatial analysis is performed on the pre-set industrial park planning zones to obtain industrial land plots. S470 involves spatial analysis of industrial land parcels with net land registration data and land already supplied data to obtain candidate parcels; S480: Based on the total area of ​​the park site selection request, the existing construction land parcels in the park are overlaid with the selected parcels to obtain candidate plots; S490: Calculate the comparison index of grid cells for candidate plots based on attribute information, and filter candidate plots based on the sum of the products of the corresponding weights of the comparison indexes to obtain a list of optimized site selection schemes.

[0045] In some embodiments, such as Figure 9 The flowcharts shown represent two optimized paths, including path 1 and path 2, where: Path 1: Spatial Modeling: The park boundary space is discretized into a regular grid (10m×10m). Each grid cell (Cell_i) serves as the basic unit of analysis, and the coordinates of the center point (Xi, Yi), the serial number ID, and the row and column numbers (Ri, Ci) of the grid cell are recorded. The grid cells (Cell_i) are overlaid with site selection indicators for analysis, and attribute information of site selection indicators such as land, three zones and three lines, prohibited construction areas, slope, industry category, main road, ultra-high voltage line, geological disaster risk area, and pollution source are configured for each cell.

[0046] Multi-objective optimization solution: The site selection problem is constructed into a mathematical model that includes multiple objective functions (such as: not involving prohibited construction areas, existing construction land areas, and conforming to the national land space planning zoning) and multiple constraints (area requirements, distance to other facilities, logistics convenience, slope suitability, etc.).

[0047] If a unit falls within an ecological protection red line, permanent basic farmland, a major infrastructure protection zone (such as an ultra-high voltage corridor), or a high-risk area for geological disasters, it is marked as 0; otherwise, it is marked as 1, C_prohibit_i∈{0,1}.

[0048] Existing construction land indicator (C_existing_i): If the unit is occupied by existing buildings or land parcels that have been transferred, it is marked as 0; otherwise, it is marked as 1. C_existing_i∈{0,1}.

[0049] Territorial spatial planning zoning compliance (C_prohibit_i): If the unit is occupied by the industrial development zone of the industrial park planning zone, it is marked as 1; otherwise, it is marked as 0.

[0050] Logistics accessibility (S_logistics_i): The accessibility of a unit to logistics nodes such as highway entrances / exits and railway freight stations is calculated and normalized. S_logistics_i∈(0,1).

[0051] Terrain fit (S_terrain_i): Before overlay analysis, slope patches with slope levels of 1-3 are extracted and then overlay analysis is performed. The fit is calculated based on the unit slope (Slope_i), with smaller slopes indicating higher fit, and then normalized.

[0052] Environmental safety level (S_environment_i): The distance between the calculation unit and the pollution source (such as a sewage treatment plant) or the polluted site. The greater the distance, the higher the safety level, and it is normalized.

[0053] Based on the above attributes, the comprehensive site selection attractiveness score (A_i) of each non-prohibited construction unit (C_prohibit_i=1, C_existing_i=1, C_prohibit_i) can be calculated: A_i = ω_1 * S_logistics_i + ω_2 * S_terrain_i + ω_3 * S_environment_i), where the weights are calculated using the TOPSIS method.

[0054] Based on the above attributes, the grid of prohibited units (C_prohibit_i=0, C_existing_i=0, C_prohibit_i=0) is removed. The comprehensive site selection attractiveness score (A_i) of S_logistics_i, S_terrain_i, S_environment_i, and S_facilities_i within each non-prohibited construction unit (C_prohibit_i=1, C_existing_i=1, C_prohibit_i=1) is calculated using the following model: A_i = ω_1 * S_logistics_i + ω_2 * S_terrain_i + ω_3 * S_environment_i + ω_4 * S_facilities_i), where ω_1 to ω_4 are the corresponding weights, which are calculated using the entropy weight method, as shown below: Define the attributes of the indicators and standardize their processing: ; in, Indicates the first The project (sample) in the first The original observations of each indicator, and ; This represents the standardized index value; Indicates the first The minimum value of each indicator across all projects; Indicates the first The maximum value of each indicator across all projects; Then, calculate the first... The first indicator and the first The numerical proportion of each item for: ; calculate Value, number Entropy value of each indicator ; ; ; Calculate the first Weight of each indicator : ; in, The difference coefficient is used to characterize the first... The amount of information contained in each indicator.

[0055] Based on the above calculations, it can be determined that the score of S_logistics_i is higher the closer it is, the score of S_terrain_i is better the lower the slope, the score of S_environment_i is better the farther away it is, and the score of S_facilities_i is better the closer it is. Each indicator is divided into 5 levels, and each level is scored out of 20 points.

[0056] The results are shown in Table 3, which is a quantitative analysis table of scheme comparison. In this embodiment of the invention, the land parcel score is calculated according to the model: A_i = ω_1 * S_logistics_i + ω_2 * S_terrain_i + ω_3 * S_environment_i + ω_4 * S_facilities_i), and then the land parcel score is calculated as follows: Figure 7 The mesh growth method shown generates mesh generation vector data for each alternative scheme.

[0057] Table 3 Quantitative Analysis Table of Scheme Comparison

[0058] Path 2: Primarily extract existing (approved but not yet supplied) construction land within the park as site selection plots. (1) After the AI ​​assistant focuses on a specific park through industry analysis, it matches the corresponding industrial park development direction area and actual built-up area of ​​the industrial park according to the park name recommended by the AI.

[0059] (2) Spatial analysis was performed on the development direction area data and the actual built-up area data with the industrial park planning zoning data to extract the map patches with the planning zoning name of industrial land.

[0060] (3) Based on the extracted industrial land zoning data, spatial analysis was performed with the net land registration data and the land already supplied data to remove the supplied land plots as candidate plots.

[0061] (4) Overlay and analyze the existing data patches that have been approved but not yet supplied with the candidate patches, and generate the available plots in the park based on the area and aspect ratio requirements input by the user.

[0062] (5) Further optimize the available plots, comprehensively analyze whether they are close to highways and railways, far from mining and exploration, have good supporting industries, avoid ecological protection red lines, cultivated land, high-voltage corridors, geological disaster areas, drinking water source protection areas, scenic spots and other factors, extract the plots that meet the optimization conditions, and use them as the generated site selection scheme.

[0063] Furthermore, the quantitative analysis of the site selection scheme in this embodiment of the invention also includes: The selection index system is constructed by creating a pool of selection indexes including land, planning, geological disaster risk, distance from UHV lines / schools, land-saving level, surrounding facilities, and location conditions. The selection method uses the TOPSIS method to calculate the weight of each index, and then uses the weighted summation method to calculate the comprehensive score of each scheme and rank them.

[0064] In some embodiments, the requirements analysis (NLP identification and result output) is as follows: NLP recognition expands the industry classification dictionary to include core keywords, synonyms, and upstream and downstream related terms for 43 industries, ensuring accurate demand identification.

[0065] 1) Example of precise matching: "Invest in and build a lithium battery material plant" → Keyword matching "Non-metallic mineral products industry" (major category) → Refinement to "Lithium battery material manufacturing" (specific industry); 2) Handling vague requirements: "Want to set up a processing plant in Hunan" → The system prompts "Please specify the processing type: agricultural and sideline food processing / textile processing / machinery processing / electronic processing..." (linking 8 major industry categories); 3) Example of a keyword dictionary (expandable): Agriculture-related industries: planting, breeding, processing, food, beverages, tobacco, cold chain, raw materials. Industrial manufacturing industries: automobiles, electronics, equipment, machinery, parts, industrial land, production capacity… Output results: A unified output structure automatically associates the core indicators and weights of the industry, eliminating the need for separate design. Core advantages: - [Indicator 1]: XX indicator value (such as "natural endowment - average annual temperature 18℃, soil pH 6.0"), meeting the core needs of the industry; - [Indicator 2]: XX indicator value (e.g., "Industry Clustering - 12 Auto Parts Companies"), to reduce implementation costs; - Policy support: XX subsidies (such as "equipment purchase subsidy of 3 million yuan"), XX approval simplification (such as "industrial land approval completed in 15 working days").

[0066] Adaptation criteria: Based on the "XX Industry" indicator system (8 core indicators), the comprehensive score ranks in the top 5% in the province, and no exclusion rules are triggered.

[0067] Important notes: You need to obtain the required permits in advance (such as "environmental pollution discharge permit") and pay attention to the risks (such as "staggered production during peak industrial electricity consumption periods").

[0068] Figure 10 This is a schematic diagram of an intelligent site selection device for industrial park projects based on intelligent agents, according to an embodiment of the present invention. The device includes a first module 1010, a second module 1020, a third module 1030, and a fourth module 1040.

[0069] The system comprises four modules: The first module retrieves the industrial classification information of the park based on the user's park site selection request, maps and expands this information to obtain an industrial classification dictionary; the second module acquires the park's knowledge source data and constructs an industrial knowledge base based on this data, which includes both unstructured and structured data; the third module constructs a specialized workflow, uses an intelligent agent to parse the park site selection request according to the workflow, and retrieves the park focus results from the industrial knowledge base based on the parsing results and the industrial indicator mapping results; the fourth module extracts the park focus results using a grid growth method to obtain selectable site areas, and uses an optimization algorithm to process these selectable site areas to obtain a list of optimized site selection schemes.

[0070] For example, with the cooperation of the first, second, third, and fourth modules in the device, the embodiment device can implement any of the aforementioned intelligent site selection methods for industrial park projects based on intelligent agents. Specifically, based on the user's park site selection request, it obtains the park's industry classification information, maps and expands the industry classification information to obtain an industry classification dictionary; it obtains the park's knowledge source data, constructs an industry knowledge base based on the knowledge source data, where the knowledge source data includes unstructured and structured data; it constructs a special workflow, uses an intelligent agent to parse the park site selection request according to the special workflow, retrieves the park focusing result from the industry knowledge base based on the parsing result and the industry indicator mapping result; it extracts the park focusing result using a grid growth method to obtain selectable site areas, and uses an optimization algorithm to process the selectable site areas to obtain a list of optimized site selection schemes. The beneficial effects of this invention are as follows: It integrates an AI industry knowledge base with site selection, constructing a knowledge base for analyzing the suitability of industrial park site selection. Through AI assistant question-and-answer sessions, it quickly recommends suitable industrial parks for specific types of projects and associates them with site selection rules. By constructing a grid and recording the element attributes associated with the site selection rules within the grid, it extracts suitable grids step-by-step according to a path of first avoiding and then optimizing. Then, it merges continuous grids to generate site selection schemes for planar units, and finally generates site selection schemes that meet the area requirements. Furthermore, it constructs a comprehensive comparative selection model that integrates land use, planning, geological disaster risks, distance from ultra-high voltage lines / schools, land-saving level, intensive utilization, surrounding facilities, and location conditions, thereby improving the scientific nature of industrial park site selection.

[0071] This invention also provides an electronic device, which includes a processor and a memory; The memory stores the program; The processor executes a program to perform the aforementioned intelligent site selection method for industrial park projects based on intelligent agents; the electronic device has the function of carrying and running the software system for intelligent site selection of industrial park projects based on intelligent agents provided in the embodiments of the present invention, such as a personal computer, minicomputer, mainframe, workstation, network or distributed computing environment, standalone or integrated computer platform, or communicating with charged particle tools or other imaging devices, etc.

[0072] This invention also provides a computer-readable storage medium storing a program that is executed by a processor to implement the intelligent site selection method for industrial park projects based on intelligent agents as described above.

[0073] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented in the embodiments of this invention. Alternative embodiments are contemplated, in which the order of various operations is changed and sub-operations described as part of a larger operation are executed independently.

[0074] This invention also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the aforementioned intelligent site selection method for industrial park projects based on intelligent agents.

[0075] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the described functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, considering the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed in the embodiments of the invention, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.

[0076] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0077] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can include, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0078] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0079] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0080] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0081] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0082] The above is a detailed description of the preferred embodiments of the present invention, but the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention, and these equivalent modifications or substitutions are all included within the scope defined by the claims of this application.

Claims

1. A method for intelligent site selection of industrial park projects based on intelligent agents, characterized in that, include: Based on the user's request for park site selection, obtain the park's industry classification information, map and expand the industry classification information to obtain an industry classification dictionary; Acquire knowledge source data from the park and construct an industry knowledge base based on the knowledge source data, which includes unstructured data and structured data; A specific workflow is constructed. Based on the specific workflow, an intelligent agent is used to analyze the park site selection request. Based on the analysis results and the industry indicator mapping results, the park focusing results are retrieved from the industry knowledge base. The focus results of the park are extracted using a grid growth method to obtain the available site regions. An optimization algorithm is then used to process the available site regions to obtain a list of optimized site selection schemes.

2. The intelligent site selection method for industrial park projects based on intelligent agents according to claim 1, characterized in that, The process involves obtaining the park's industry classification information based on the user's park site selection request, mapping and expanding the industry classification information to obtain an industry classification dictionary, including: Obtain the industrial classification information of the park, which includes major industrial categories, industrial subcategories, and common core needs; Natural language processing is used to map core keywords, synonyms, and upstream and downstream related terms to industry classification information, resulting in an industry classification dictionary.

3. The intelligent site selection method for industrial park projects based on intelligent agents according to claim 1, characterized in that, The process involves acquiring knowledge source data from the park and constructing an industry knowledge base based on this data. The knowledge source data includes both unstructured and structured data, including: Acquire knowledge source data for the park, including unstructured data such as policy data and encyclopedia entries related to the park, and structured data such as park lists and statistical yearbook data; The knowledge source data is processed by text segmentation and normalization to obtain data slices. Text segmentation includes word segmentation, stop word removal, and part-of-speech tagging of the knowledge source data. Normalization includes mapping of multi-source data and anonymization of privacy data. The obtained data slices are converted into vector data and stored as triples using MySQL to obtain the industry knowledge base.

4. The intelligent site selection method for industrial park projects based on intelligent agents according to claim 3, characterized in that, The construction of the specialized workflow involves using an intelligent agent to analyze the park site selection request based on the workflow, and retrieving the park focusing results from the industry knowledge base based on the analysis results and industry indicator mapping results. This includes: Use Dify to build and orchestrate specific workflows; The site selection request for the park was parsed using natural language processing according to a specific workflow, and the parsing results were obtained. The vectorized parsing results are matched with the industry knowledge base for similarity, and prompt words are generated based on the matching results and the vectorized parsing results. The system employs one of the following methods: precise matching based on prompt words, fuzzy demand processing, and industry classification dictionary matching, to obtain the park-focused results.

5. The intelligent site selection method for industrial park projects based on intelligent agents according to claim 4, characterized in that, The focusing results of the park are extracted using a grid growing method to obtain a siteable area. An optimization algorithm is then used to process the siteable area to obtain a list of optimized site selection schemes, including: The park in the focused results is divided into multiple regular grid units. The basic information of the grid units is overlaid with the park site selection indicators to obtain the attribute information of the grid units. The basic information includes the center coordinates, serial number and row and column number; the attribute information includes land, three zones and three lines, prohibited construction areas, slope, industry category, main road, ultra-high voltage line, geological disaster risk area and pollution source. Based on the basic and attribute information of the grid cells, a multi-objective solution is performed using a multi-objective function and multiple constraints with corresponding weight parameters to obtain the comprehensive location attractiveness score of each grid cell. The weight parameters are calculated using TOPSIS. Based on the comprehensive location attractiveness score of each grid cell, grid cells that avoid prohibitive constraints are identified, and a cyclic search is performed using a grid growth method to obtain a set of grid cells. Based on the total area of ​​the park's site selection requests and the overall comprehensive site selection attractiveness score, the grid unit set is filtered and sorted to obtain a list of optimized site selection schemes.

6. The intelligent site selection method for industrial park projects based on intelligent agents according to claim 5, characterized in that, The process involves filtering and sorting the grid cell set based on the total area and overall site attractiveness score of the park's site selection requests, resulting in a list of optimized site selection schemes, including: Using the existing construction land in the park as the site selection plot, name matching is performed based on the site selection plot and the park focus results to obtain the corresponding industrial park development direction area and the actual built-up area of ​​the industrial park. Spatial analysis is conducted based on the development direction zone of the industrial park, the actual built-up area of ​​the industrial park, and the pre-set industrial park planning zones to obtain industrial land plots. Spatial analysis was performed on industrial land parcels, net land registration data, and land already supplied data to obtain candidate parcels; Based on the total area of ​​the park site selection request, the existing construction land parcels in the park are overlaid with the selected parcels for analysis to obtain candidate plots; The comparison index of grid cells of candidate plots is calculated based on attribute information. The candidate plots are then screened based on the sum of the products of the corresponding weights of the comparison indexes to obtain a list of optimized site selection schemes.

7. The intelligent site selection method for industrial park projects based on intelligent agents according to claim 5, characterized in that, The multi-objective function and the multiple constraints are determined based on attribute information and basic information. The multi-objective function and the multiple constraints include whether it falls within the three zones and three lines, whether it includes existing construction land markings, spatial planning zoning compliance, terrain adaptability, and environmental safety.

8. An intelligent site selection device for industrial park projects based on intelligent agents, characterized in that, include: The first module is used to obtain the industrial classification information of the park based on the user's park site selection request, and to map and expand the industrial classification information to obtain an industrial classification dictionary; The second module is used to acquire knowledge source data of the park and build an industry knowledge base based on the knowledge source data, which includes unstructured data and structured data. The third module is used to construct a special workflow. Based on the special workflow, an intelligent agent is used to analyze the park site selection request. Based on the analysis results and the industry indicator mapping results, the park focusing results are retrieved from the industry knowledge base. The fourth module is used to extract the focalization results of the park using a grid growth method to obtain the available site areas, and to process the available site areas using an optimization algorithm to obtain a list of optimized site selection schemes.

9. An electronic device, characterized in that, Including the processor and memory; The memory is used to store programs; The processor executes the program to implement the intelligent site selection method for industrial park projects based on intelligent agents as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The storage medium stores a program, which is executed by a processor to implement the intelligent site selection method for industrial park projects based on intelligent agents as described in any one of claims 1-7.

Images