A land resource optimization allocation decision-making method and system based on artificial intelligence

Abstract

This invention discloses an artificial intelligence-based land element optimization allocation decision-making method and system, specifically relating to the field of land resource management data processing technology. It addresses the problem of low feasibility of existing land allocation methods due to the disconnect between macro-optimization and micro-constraints. The method involves establishing a multi-scale optimization model, first obtaining macro-optimization objectives and plot-level micro-constraints, then calculating the entropy-based mixed-use potential parameters for each plot, generating a plot usability intensity distribution map, identifying key conflict plots by simulating the chain reaction of intensity deviations, and finally constructing an optimization model integrating macro and micro conditions to generate land element allocation schemes. This achieves full-process optimization of land resources from macro-strategy to micro-implementation.

Description

Technical Field

[0001] This invention relates to the field of land resource management data processing technology, and more specifically, to an artificial intelligence-based method and system for optimizing the allocation of land resources. Background Technology

[0002] In the field of land resource management, artificial intelligence technology has been applied to assist in the optimal allocation of land resources. Existing methods typically construct optimization models based on macro-level, environmental, and social benefit objectives at the regional level, and generate land use allocation schemes through machine learning and multi-objective optimization algorithms. These methods can provide idealized planning suggestions for the allocation of land resources at a macro scale, serving administrative management and decision support.

[0003] However, existing AI-based land allocation methods have significant drawbacks: their optimization process focuses on achieving macro-level goals, failing to effectively incorporate micro-level constraints at the plot level, such as property boundaries, geological conditions, and infrastructure layout. This results in generated allocation schemes that, while exhibiting optimization characteristics at the macro level, often conflict with micro-level physical or rule constraints during actual implementation, reducing the feasibility of the schemes and making them difficult to directly apply to actual land management and implementation processes. Summary of the Invention

[0004] In order to overcome the above-mentioned deficiencies of the prior art, the present invention provides a land element optimization allocation decision method and system based on artificial intelligence to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] An artificial intelligence-based land resource optimization decision-making method includes:

[0007] S1. Obtain the macro-optimization objectives and plot-level micro-constraints for the target area;

[0008] S2. Based on micro-constraints and the attributes of neighboring plots, calculate the probability distribution of each plot adapting to different land use types, and use the information entropy algorithm to quantify the uncertainty of mixed use of the probability distribution in order to generate entropy-based mixed use potential parameters for each plot.

[0009] S3. Using entropy-based mixed utilization potential parameters as the core input, the usability intensity of each plot is preprocessed and analyzed through a multi-objective optimization algorithm to generate a plot usability intensity distribution map.

[0010] S4. Combining the land parcel availability intensity distribution map with macro-optimization objectives, simulate the chain impact effect of intensity deviation of each parcel on the functions of key areas and assess its spatial diffusion risk, and identify key conflict parcels.

[0011] S5. Taking key conflict plots as the core optimization area, construct a multi-scale optimization model that integrates macro-optimization objectives and micro-constraints.

[0012] S6. Generate optimal allocation schemes for land elements through a multi-scale optimization model and output the allocation results.

[0013] Furthermore, obtain the macro-optimization objectives for the target area and the micro-constraints at the plot level, including:

[0014] The ecological protection indicators, spatial structure indicators, and environmental quality indicators of the target area are obtained as macro-optimization targets; the property boundary data, geological condition data, and infrastructure layout data at the plot level are obtained as micro-constraints.

[0015] The property boundary data includes land use boundary vector data and ownership registration information; the geological condition data includes soil type distribution and terrain slope data; and the infrastructure layout data includes road network and energy pipeline distribution vector data.

[0016] Furthermore, based on micro-constraints and the attributes of neighboring plots, the probability distribution of each plot's suitability for different land use types is calculated. The information entropy algorithm is then used to quantify the uncertainty of mixed-use development in the probability distribution to generate entropy-based mixed-use potential parameters for each plot, including:

[0017] Based on property boundary data, geological condition data, and infrastructure layout data, the suitability levels of each plot for residential, commercial, and ecological land use are analyzed.

[0018] Based on the existing land use type distribution in the attributes of neighboring land parcels, calculate the conditional probability of each land parcel adapting to different land use types;

[0019] The conditional probabilities of different land use types corresponding to each plot are combined into a probability distribution vector;

[0020] The information entropy algorithm is used to calculate the information entropy value of the probability distribution vector of each plot, and the information entropy value is used as the entropy-based mixed utilization potential parameter to characterize the uncertainty of mixed utilization.

[0021] Furthermore, by combining the existing land use type distribution in the attributes of neighboring plots, the conditional probability of each plot adapting to different land use types is calculated. This is done by statistically analyzing the area proportion of each land use type within the buffer zone around the target plot, and then weighting and correcting the area proportion based on the property boundary data and geological condition data in the micro-constraints of the plot, thus obtaining the conditional probability value of the corresponding plot adapting to each land use type.

[0022] Furthermore, using entropy-based mixed-use potential parameters as the core input, a multi-objective optimization algorithm is used to preprocess and analyze the usability intensity of each plot, generating a plot usability intensity distribution map, including:

[0023] The entropy-valued hybrid utilization potential parameter is used together with the ecological protection index and spatial structure index in the macro-optimization objective as the optimization objective;

[0024] Pareto front solution set search is performed on the usability intensity of each parcel based on a multi-objective optimization algorithm;

[0025] The solution with the highest available strength that satisfies the micro-constraints is selected from the solution set to generate a distribution map of available strength of the land parcel.

[0026] Furthermore, the Pareto front solution set search for the usability intensity of each land parcel based on the multi-objective optimization algorithm is to take the entropy-based mixed utilization potential parameter of each land parcel and the ecological protection index and spatial structure index in the macro-optimization objective as the optimization objective, and use the multi-objective optimization algorithm to search for the Pareto optimal solution set of usability intensity under the condition of satisfying all micro-constraints.

[0027] Furthermore, by combining the land parcel availability intensity distribution map with macro-optimization objectives, the cascading impact effects of intensity deviations in each parcel on key regional functions are simulated and their spatial diffusion risks are assessed, identifying key conflict parcels, including:

[0028] Based on the intensity values ​​of each plot in the land use intensity distribution map and the expected values ​​of the corresponding indicators in the macro-optimization objectives, the intensity deviation value of each plot is calculated.

[0029] Construct a key regional function dependency network, where nodes include ecological integrity maintenance functions and transportation connectivity guarantee functions, and edges represent the interdependencies between functions;

[0030] Simulate the cascading failure process caused by the strength deviation of each plot in the functionally dependent network, and calculate its impact on the global stability of the network.

[0031] Simultaneously assess the spatial transmission path and attenuation degree of intensity deviation in each plot along the road network and energy pipeline distribution defined in the infrastructure layout data;

[0032] Based on a comprehensive assessment of the intensity value affecting the overall stability of the network and the degree of spatial transmission attenuation, key conflict areas were identified.

[0033] Furthermore, taking key conflict parcels as the core optimization area, a multi-scale optimization model integrating macro-optimization objectives and micro-constraints is constructed, including:

[0034] The property boundary data and geological condition data of key conflicting land parcels are defined as immutable elements as rigid constraints, serving as hard boundaries for optimization solutions.

[0035] Adjustable elements in the infrastructure layout and geological condition data of key conflict sites are transformed into flexible constraints and added as penalty terms to a multi-objective function constructed from ecological protection indicators and spatial structure indicators in the macro-optimization objective.

[0036] A multi-scale optimization model is constructed based on hard boundaries and a multi-objective function with a penalty term.

[0037] Furthermore, a multi-scale optimization model is used to generate optimal allocation schemes for land elements, and the allocation results are output, including:

[0038] A multi-objective optimization algorithm is used to solve the multi-scale optimization model to obtain the Pareto optimal solution set of land use type and utilization intensity of key conflict plots;

[0039] By combining the original intensity data of non-critical conflict sites in the site availability intensity distribution map, the optimal solution of critical conflict sites is integrated with the original intensity data of non-critical conflict sites.

[0040] Generate a land feature optimization scheme that includes a vector layer containing the final land use type and utilization intensity allocation results for each plot; output the vector layer file of the land feature optimization scheme.

[0041] On the other hand, the present invention provides a land resource optimization allocation decision system based on artificial intelligence, comprising:

[0042] The information acquisition module is used to acquire the macro-optimization objectives of the target area and the micro-constraints at the plot level.

[0043] The parameter generation module is used to calculate the probability distribution of each plot adapting to different land use types based on micro-constraints and the attributes of neighboring plots. It uses the information entropy algorithm to quantify the uncertainty of mixed use of the probability distribution in order to generate entropy-based mixed use potential parameters for each plot.

[0044] The distribution generation module is used to generate a distribution map of the usable intensity of land parcels by taking entropy-valued mixed utilization potential parameters as the core input, and preprocessing and analyzing the usable intensity of each parcel through a multi-objective optimization algorithm.

[0045] The conflict identification module is used to combine the land use intensity distribution map with macro-optimization objectives to simulate the chain impact effect of intensity deviations of various land parcels on the functions of key areas and assess their spatial diffusion risk, thereby identifying key conflict land parcels.

[0046] The model building module is used to construct a multi-scale optimization model that integrates macro-optimization objectives and micro-constraints, with key conflict plots as the core optimization area.

[0047] The results output module is used to generate optimal allocation schemes for land elements through a multi-scale optimization model and output the allocation results.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] 1. By establishing a multi-scale optimization model that combines macro-optimization objectives with micro-constraints, the problem of the disconnect between macro and micro levels in traditional land allocation methods is effectively solved. By introducing entropy-based mixed utilization potential parameters, the utilization uncertainty at the plot level can be quantified, providing a scientific basis for subsequent optimization. By simulating the chain reaction effect and spatial diffusion risk of intensity deviation, the accurate identification of key conflict plots is achieved, providing clear target areas for subsequent optimization and significantly improving the practical feasibility and implementability of land allocation schemes.

[0050] 2. By constructing a multi-scale optimization model that integrates macro-level optimization objectives and micro-level constraints, the entire process of land resource allocation, from macro-strategy to micro-level implementation, was optimized. This model not only considered the development strategy needs at the regional level but also fully incorporated the actual constraints at the plot level. As a result, the final allocation scheme not only meets the macro-planning objectives but also satisfies the micro-level implementation conditions, greatly improving the scientific nature and practicality of land resource allocation and providing more reliable technical support for land resource management. Attached Figure Description

[0051] Figure 1 This is a flowchart of a land element optimization allocation decision-making method based on artificial intelligence according to the present invention;

[0052] Figure 2 This is a schematic diagram of the structure of a land element optimization allocation decision system based on artificial intelligence according to the present invention. Detailed Implementation

[0053] 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 some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] Example 1: Figure 1 This invention presents an artificial intelligence-based land resource optimization allocation decision-making method, comprising:

[0055] S1. Obtain the macro-optimization objectives and plot-level micro-constraints for the target area;

[0056] S2. Based on micro-constraints and the attributes of neighboring plots, calculate the probability distribution of each plot adapting to different land use types, and use the information entropy algorithm to quantify the uncertainty of mixed use of the probability distribution in order to generate entropy-based mixed use potential parameters for each plot.

[0057] S3. Using entropy-based mixed utilization potential parameters as the core input, the usability intensity of each plot is preprocessed and analyzed through a multi-objective optimization algorithm to generate a plot usability intensity distribution map.

[0058] S4. Combining the land parcel availability intensity distribution map with macro-optimization objectives, simulate the chain impact effect of intensity deviation of each parcel on the functions of key areas and assess its spatial diffusion risk, and identify key conflict parcels.

[0059] S5. Taking key conflict plots as the core optimization area, construct a multi-scale optimization model that integrates macro-optimization objectives and micro-constraints.

[0060] S6. Generate optimal allocation schemes for land elements through a multi-scale optimization model and output the allocation results.

[0061] S1. Obtain the macro-optimization objective and the micro-constraints at the plot level for the target area. The specific implementation is as follows:

[0062] When obtaining macro-optimization targets for the target area, ecological protection indicators are quantitatively extracted by collecting satellite remote sensing images and ecological survey reports of the target area. These indicators specifically include vegetation cover, biodiversity index, and the proportion of ecologically sensitive areas. Vegetation cover is obtained by calculating the proportion of pixels with a Normalized Difference Vegetation Index (NDVI) greater than 0.3. For example, NDVI values ​​are based on Landsat... The reflectance of the red band (0.64-0.67μm) and near-infrared band (0.85-0.88μm) of satellite imagery was calculated. The biodiversity index was calculated based on the Shannon-Wiener index from regional species survey data. This index was obtained by substituting the species number and their relative abundance into the formula. The area ratio of ecologically sensitive areas was calculated by overlaying the vector boundaries of legally protected areas such as nature reserves and water conservation areas with the target area. Spatial structure indicators were extracted from urban planning outlines and territorial spatial planning texts, including the intensity of construction land development, the proportion of green space and open space, and the connectivity of transportation networks. The intensity of construction land development was determined by the ratio of existing construction land area to the total area in the target area. For example, the existing construction land area was extracted from the construction land layer of the land use status map. The proportion of green space and open space was obtained by extracting land use data. The area ratio of green space, plazas, and water bodies in the current map is calculated by summing their areas. Traffic network connectivity is characterized by the road network density and the accessibility index between major traffic nodes within the target area. The accessibility index uses the shortest path algorithm in spatial network analysis to calculate the average travel time; for example, Dijkstra's algorithm is used to calculate the shortest time path from each traffic node to other nodes. Environmental quality indicators are obtained from monitoring data released by environmental protection departments, including the annual average concentration of PM2.5, the surface water quality compliance rate, and the average regional environmental noise. The annual average concentration of PM2.5 is obtained from hourly data from atmospheric monitoring stations, calculated as an annual arithmetic average; for example, the annual average of the daily average concentrations of each station is taken and then averaged regionally. The surface water quality compliance rate is calculated by comparing the percentage of times the water quality category at the monitoring section meets the target water category; for example, the water quality category is based on the "Surface Water Environmental Quality Standard" GB. The classification criteria in 3838-2002 stipulate that the average value of regional environmental noise is obtained by spatial interpolation of the equivalent sound level Leq measured by the grid point method. For example, the average value is calculated after generating a continuous surface from discrete point noise data using the Kriging interpolation method.

[0063] When obtaining micro-level constraints at the land parcel level, property boundary data is extracted from the real estate registration database of the natural resources department, including land use boundary vector data and ownership registration information. The land use boundary vector data is a polygon layer with a coordinate system, where each polygon corresponds to a land parcel boundary. For example, the coordinate system is the CGCS2000 national geodetic coordinate system. Ownership registration information is linked to each land parcel in the form of an attribute table, recording the right holder, land use, and term of use. Geological condition data is obtained by integrating geological exploration reports and geographic information data, including soil type distribution and topographic slope data. Soil type distribution data comes from a 1:50000 soil type map vector layer, such as soil type... The soil types include brown soil, red soil, and paddy soil. Topographic slope data is generated by calculating slope values ​​based on the elevation difference of neighboring pixels using a digital elevation model (DEM). The slope values ​​are expressed in degrees. For example, a 3×3 pixel window is used to calculate the elevation change rate. Infrastructure layout data is extracted from special planning maps of urban planning and construction departments. This includes vector data of road network and energy pipeline distribution. Road network data is a line layer that includes road grade, width, and traffic capacity attributes. For example, road grades are divided into expressways, arterial roads, and secondary arterial roads. Energy pipeline distribution data is a combination of line and point layers that records the planar location, diameter, and burial depth of water supply and distribution pipelines, gas pipelines, and power cables. For example, the pipe diameter is recorded in millimeters.

[0064] All data was obtained from government open data platforms, professional geographic information databases, or authorized third-party data services. The data formats were Shapefile, GeoJSON, or raster images, and the coordinate system was CGCS2000 National Geodetic Coordinate System to ensure spatial data consistency. After acquisition, the data was preprocessed using geographic information system software, including coordinate transformation, data cropping, and attribute field alignment. For example, ArcGIS software was used to convert data from different coordinate systems to the CGCS2000 coordinate system. A mask extraction tool was used to crop the data to the target area. A field calculator was used to align field names and data types in the attribute tables, ensuring that macro-optimization target indicators and micro-constraint data matched in spatial scope and attribute structure, providing standardized input for subsequent steps.

[0065] S2. Based on micro-constraints and the attributes of neighboring plots, calculate the probability distribution of each plot's suitability for different land use types. Use the information entropy algorithm to quantify the uncertainty of mixed use in the probability distribution to generate entropy-based mixed use potential parameters for each plot. The specific implementation is as follows:

[0066] First, based on the property boundary data, geological condition data, and infrastructure layout data obtained in step S1, the suitability level of each plot for residential, commercial, and ecological use is analyzed. The suitability level for residential land is determined by comprehensively assessing the permitted use types in the property boundary data (ownership registration information), whether the terrain slope in the geological condition data is less than 15 degrees, and the number of educational and medical facilities within 500 meters of the road network in the infrastructure layout data. For example, when the property registration information allows residential construction, the terrain slope is less than 8 degrees, and there are more than three public service facilities within 500 meters, the residential land suitability level is rated as the highest. The suitability level for commercial land is determined by analyzing the traffic flow data of the surrounding road network and the distance from the commercial center. The suitability level of commercial land is rated as the highest level based on the comprehensive assessment of the foundation bearing capacity in the geological condition data. For example, when a plot of land is adjacent to a main road, less than 1000 meters away from the regional commercial center, and has a foundation bearing capacity greater than 150 kPa, the suitability level of commercial land is rated as the highest level. The suitability level of ecological land is determined based on whether the distribution of soil types in the geological condition data belongs to the ecological protection type and its connectivity with existing ecological land. For example, when the soil type is wetland or forest and the distance from the boundary of a nature reserve is less than 300 meters, the suitability level of ecological land is rated as the highest level. The suitability level of each land use type adopts a scoring system of 1 to 5 points, where 5 points represent the most suitable and 1 point represents the least suitable. In the scoring process, if a certain land use type is explicitly prohibited in the property boundary data, a score of 1 point is directly given.

[0067] Next, combining the existing land use type distribution in the attributes of adjacent plots, the conditional probability of each plot adapting to different land use types is calculated. Specifically, a circular buffer zone with a radius of 500 meters is established with the geometric center point of the target plot as the center. The area proportions of existing residential, commercial, and ecological land within this buffer zone are statistically analyzed. For example, using GIS spatial analysis tools to calculate the area proportions of various land use types within the buffer zone, when residential land accounts for 45%, commercial land accounts for 30%, and ecological land accounts for 25%, the initial area proportions are recorded as 0.45, 0.30, and 0.25, respectively. Then, based on the property boundary data and geological condition data in the micro-constraints of the plot, the area proportions are weighted and corrected. The weights of the property boundary data are based on the ownership registration information. The strictness of land use restrictions is set. For example, the weight is 0 when the restriction is absolutely prohibited, 0.5 when the restriction is conditionally permitted, and 1.0 when there is no restriction. The weight of geological condition data is set according to the development difficulty of terrain slope and soil type. For example, the weight is 0.3 when the slope is greater than 25 degrees, 0.6 when the slope is between 5 and 25 degrees, and 0.8 when the slope is less than 5 degrees. The area proportion is multiplied by the weight coefficient to obtain the weighted value. Then, the weighted values ​​of all land use types are normalized so that the sum of the conditional probability values ​​is 1. This gives the conditional probability value of the land plot being suitable for residential, commercial, and ecological use. For example, the conditional probability value is 0.4 for residential use, 0.35 for commercial use, and 0.25 for ecological use.

[0068] It is worth noting that, to ensure the objectivity and repeatability of conditional probability value calculations and reduce over-reliance on expert experience, after obtaining the initial area proportion through buffer analysis, a quantitative rule based on micro-constraints can be used for automatic correction. The correction rule is implemented through a predefined decision matrix, which takes the specific values ​​of the micro-constraints as input and outputs the corresponding weight coefficients. The decision matrix is ​​constructed as follows: For property boundary data, if the property registration information explicitly states "construction prohibited," the weight coefficient is 0; if it states "conditional construction," the weight coefficient ranges from 0.3 to 0.7 depending on the strictness of the restrictions (e.g., 0.3 for a setback of 50 meters or more, and 0.5 for a setback of 20 meters); if it states "construction permitted," the weight coefficient is 1.0. For geological condition data, the weight coefficient is 1.0 when the slope is less than 5 degrees, 0.8 for slopes between 5 and 15 degrees, 0.5 for slopes between 15 and 25 degrees, and 0.1 for slopes greater than 25 degrees. According to the geological exploration report, the soil bearing capacity has a weighting coefficient of 1.0 when it is greater than 150 kPa, 0.6 when it is between 100 and 150 kPa, and 0.2 when it is less than 100 kPa. The above weighting coefficients are multiplied by the initial area proportions obtained from the statistics within the buffer zone, and the results are normalized to obtain the conditional probability values ​​of each plot of land adapted to different land use types based on objective rules.

[0069] Then, the conditional probability values ​​corresponding to different land use types for each plot are combined into a probability distribution vector. The probability distribution vector of each plot contains three dimensions, which correspond to the conditional probability values ​​of residential land, commercial land and ecological land in order. For example, the probability distribution vector of a certain plot is represented as (0.4, 0.35, 0.25). The sum of each element in the vector is strictly equal to 1. When the probability sum is 0.99 or 1.01, it is normalized to 1 through numerical adjustment method to ensure the rigor of mathematical calculation. Finally, the information entropy algorithm is used to calculate the information entropy value of the probability distribution vector of each plot. The information entropy value is obtained by substituting each probability value in the probability distribution vector into the information entropy calculation formula. The information entropy calculation formula is the negative cumulative sum of each probability value multiplied by the base-2 logarithm of that probability value. This information entropy value is used as the entropy-based mixed use potential parameter to characterize the uncertainty of mixed use of the plot. The larger the information entropy value, the higher the possibility that the plot is suitable for mixed use. When the conditional probability of a plot is concentrated in one land use type, the information entropy value is close to 0. When the conditional probability is uniformly distributed, the information entropy value reaches its maximum.

[0070] All calculations were implemented using Python scripts, with spatial buffer analysis using the GeoPandas library and probability and information entropy calculations using the NumPy library, ensuring that the calculation process was repeatable and verifiable. Finally, the entropy-based mixed utilization potential parameters for each plot were stored as attribute fields in the plot vector data, providing input for subsequent steps. All calculation results were retained to four decimal places to ensure the consistency of data precision.

[0071] S3. Using entropy-based mixed-use potential parameters as the core input, the usability intensity of each plot is preprocessed and analyzed through a multi-objective optimization algorithm to generate a plot usability intensity distribution map. The specific implementation is as follows:

[0072] First, the entropy-based mixed utilization potential parameters of each plot calculated in step S2 are combined with the ecological protection indicators and spatial structure indicators from the macro-optimization objectives obtained in step S1 as optimization objectives. The ecological protection indicators include vegetation coverage, biodiversity index, and the proportion of ecologically sensitive areas. The spatial structure indicators include the intensity of construction land development, the proportion of green space and open space, and transportation network connectivity. These indicators are normalized to a value range of 0 to 1 and assigned corresponding weight coefficients. The weight coefficients are determined by scoring from professionals in the field based on the importance of each indicator in the target area planning. For example, five professionals are invited to independently score the importance of each indicator, with scores ranging from 1 to 5. The average score is then used as the weight coefficient. In ecologically sensitive areas, the weight coefficient for ecological protection indicators is set to 0.6, the weight coefficient for spatial structure indicators is 0.3, and the weight coefficient for entropy-based mixed utilization potential parameters is 0.1. In urban built-up areas, the weight coefficient for spatial structure indicators is set to 0.5, the weight coefficient for ecological protection indicators is 0.3, and the weight coefficient for entropy-based mixed utilization potential parameters is 0.2, ensuring a balance among the optimization objectives.

[0073] The Pareto front search is performed on the usability intensity of each land parcel based on a multi-objective optimization algorithm. A fast non-dominated sorting genetic algorithm with an elitist strategy is used as the multi-objective optimization algorithm. The population size is set to 100 individuals, the number of iterations is 200 generations, the crossover probability is 0.9, and the mutation probability is 0.1. During algorithm execution, the usability intensity of each land parcel is used as the decision variable. The intensity value range is set according to the land parcel area and planning requirements. For example, the intensity value range for residential land is 1.0 to 3.0, for commercial land is 2.0 to 5.0, and for ecological land is... The value ranges from 0.1 to 1.0. The micro-constraints obtained in step S1 are used as constraint functions, including the land use boundary in the property boundary data, the slope restriction in the geological condition data, and the road capacity restriction in the infrastructure layout data. In each iteration, the algorithm calculates the objective function value of each solution. The objective function is constructed by weighted summation. Individual selection is performed based on non-dominated sorting and congestion calculation. Finally, the Pareto optimal solution set of available intensity is obtained. This solution set contains multiple non-dominated solutions, each of which represents a available intensity allocation scheme that achieves a balance among various optimization objectives.

[0074] The specific application of the multi-objective optimization algorithm is as follows: Each individual (chromosome) is encoded as a real number array consisting of the utilization intensity values ​​of all land parcels. The fitness function F is defined as the weighted sum of the optimization objectives: Where feco is the normalized comprehensive value of ecological protection indicators, fspa is the normalized comprehensive value of spatial structure indicators, and fent is the normalized entropy-based mixed utilization potential parameter. w1, w2, and w3 are the aforementioned weighting coefficients. The constraint handling adopts the penalty function method, transforming the degree of violation of micro-constraints into a penalty term Penalty, which is subtracted from the fitness value; that is, the final fitness Ffinal = F - Penalty.

[0075] The formula for calculating Penalty is: Where violationi is the violation amount of the i-th constraint (e.g., the difference between the actual utilization intensity and the allowable intensity of the land use boundary), and λi is the penalty coefficient for the corresponding constraint, which is set according to the importance of the constraint. For example, for an immutable rigid constraint, λi is set to a very large number (e.g., 10). 6 To ensure the algorithm automatically discards such solutions, λi is set to a value between 1.0 and 100 depending on the adjustment difficulty for elastic constraints. The algorithm uses simulated binary crossover (SBX) for crossover operations and polynomial mutation operations for mutation operations to better search the solution space.

[0076] The solution with the highest usable intensity that satisfies the micro-constraints is selected from the Pareto optimal solution set. Specifically, each solution in the solution set undergoes a feasibility test. The test criteria include whether it meets the land use boundary limits in the property rights data, the slope limits in the geological conditions data, and the capacity limits in the infrastructure layout data. For example, solutions with a slope greater than 25 degrees but assigned high utilization intensity, or solutions that exceed the land use boundary, are removed. Then, the solution with the largest weighted sum of usable intensity among the remaining feasible solutions is selected as the final solution. The weighted sum of usable intensity is determined based on the utilization intensity and area of ​​different land use types. The proportions are calculated, for example, the weight of residential land use intensity is 0.4, the weight of commercial land use intensity is 0.4, and the weight of ecological land use intensity is 0.2. Finally, a distribution map of the usable intensity of the land parcels is generated. This map is represented in the form of a vector layer. Each land parcel contains a use intensity attribute value. The use intensity value is determined according to the allocation result in the final solution. For example, the use intensity value of a certain land parcel is 2.5, which means that the planning development intensity of the land parcel is at a medium to high level. The layers are stored in the standard geographic information system format, and the coordinate system is consistent with the data in step S1 to provide input for subsequent steps.

[0077] S4. Combining the land parcel's available intensity distribution map with macro-optimization objectives, simulate the chain reaction impact of intensity deviations in various parcels on key area functions and assess their spatial diffusion risk, identifying key conflict parcels. The specific implementation is as follows:

[0078] First, based on the intensity values ​​of each plot in the land use intensity distribution map generated in step S3 and the expected values ​​of the corresponding indicators in the macro-optimization objectives obtained in step S1, the intensity deviation value of each plot is calculated. The intensity deviation value is obtained by dividing the absolute difference between the actual intensity value and the expected intensity value by the expected intensity value. For example, if the actual utilization intensity value of a plot is 2.5 and the expected intensity value of this type of plot in the macro-optimization objectives is 2.0, then the intensity deviation value is (2.5-2.0) / 2.0=0.25, indicating that the utilization intensity of the plot exceeds the expected value by 25%. The determination of the expected intensity value is based on a combination of planning standards and assessments by professionals in the field. Specifically, by collecting historical planning implementation data of the target area, analyzing the distribution patterns of various land use intensities, and combining the independent assessment opinions of more than 5 professionals in the field, the expected intensity value range of various land uses is finally determined. For example, the expected intensity value range for residential land is 1.8 to 2.2, for commercial land it is 3.5 to 4.5, and for ecological land it is 0.5 to 1.0.

[0079] When constructing a functional dependency network for key areas, the functions of maintaining ecological integrity and ensuring transportation connectivity are used as network nodes. The edges between nodes represent the interdependencies between functions. The network construction process includes three steps: node definition, edge connection, and weight assignment. In node definition, the function of maintaining ecological integrity is subdivided into three sub-nodes: vegetation cover function, biodiversity maintenance function, and soil and water conservation function. The function of ensuring transportation connectivity is subdivided into three sub-nodes: road capacity function, transportation hub connectivity function, and emergency passage accessibility function. Edge connection is determined by analyzing historical event data to determine the correlation between functions. For example, by analyzing the transportation construction and ecological impact data of the past 10 years, the impact path of transportation connectivity on ecological integrity is determined. The weight assignment adopts the Delphi method. Ten professionals in the field are invited to score the degree of dependence of each pair of functions from 0 to 1. After three rounds of back-to-back scoring, the average value is taken. For example, the final weight of the impact of ecological integrity on transportation connectivity is determined to be 0.3, and the weight of the impact of transportation connectivity on ecological integrity is 0.2.

[0080] The edge weight assignment for the function-dependent network, based on the Delphi method, incorporates a verification and correction process using objective monitoring data. Specifically, after obtaining the average of preliminary scores from experts in the field, these are used as prior weights. Subsequently, urban operational data from the past 5-10 years for the target area, such as traffic flow data, air quality monitoring data, and remote sensing data on vegetation growth status, are retrieved. Statistical analysis methods, such as Pearson correlation coefficient or Granger causality test, are used to quantitatively analyze the correlation and influence direction between the time series data of different functional nodes. The obtained objective correlation strength is then fused with the prior weights, for example, by taking a weighted average of the two, to finally determine the weight value of each edge in the network, reducing the subjective arbitrariness in network construction.

[0081] This study simulates the cascading failure process caused by the intensity deviation of each plot in a functionally dependent network. A discrete-time step simulation method is used, with a time step of 1 unit and a total simulation duration of 10 units. Within each time step, the impact of the intensity deviation on adjacent nodes is calculated. The degree of impact is determined by the connection weights and deviation values ​​between nodes. The specific formula is: Impact Intensity = Intensity Deviation Value × Connection Weight × Attenuation Coefficient. The attenuation coefficient is set to 0.1 to 0.3 based on the node distance. For example, when the plot intensity deviation is 0.25, the connection weight is 0.3, and the attenuation coefficient is 0.2, the impact on directly connected nodes... The impact strength is 0.25 × 0.3 × 0.2 = 0.015. The formula for updating the state of the affected node is: new state value = original state value × (1 - impact strength). When the node state value is lower than the failure threshold of 0.8, it is determined to be a failed node. Failed nodes will continue to affect their neighboring nodes, forming a cascading effect. The global stability of the network is evaluated by calculating the network efficiency index after the node state changes. The network efficiency is represented by the inverse average of the shortest paths between all node pairs. The specific calculation process is to first calculate the shortest path length between all node pairs in the network, then take the inverse of these path lengths, and finally calculate the arithmetic mean of these inverses.

[0082] The quantitative relationship between the attenuation coefficient and the node distance is determined by the attenuation function. This embodiment uses an exponential attenuation function: Where d is the network distance between nodes (in kilometers), and β and γ are attenuation parameters. Calibrated based on historical data, β is set to 0.3 and γ to 0.5. This means that when the node distance d is 0 (i.e., itself), the attenuation coefficient is 0.3; when d is 2 kilometers, the attenuation coefficient is approximately... This functional relationship ensures that the process of the influence intensity decaying with increasing distance is continuous and quantifiable.

[0083] The failure threshold was determined based on the results of retrospective simulation analysis of historical emergencies (such as major infrastructure failures and natural disasters). By comparing the node functional state values ​​with actual failure scenarios in multiple cases, it was found that when the node state value dropped below 0.8, the probability of actual functional failure exceeded 90%. Therefore, setting the threshold to 0.8 can better reflect the cascading failure critical point in the real world in the simulation.

[0084] In practical implementation, the calculation of influence intensity requires ensuring that all parameters have clear dimensional definitions and consistent units. The intensity deviation value is a dimensionless parameter, obtained by dividing the difference between the actual and expected intensity values ​​by the expected intensity value. The connection weight is a dimensionless parameter, ranging from 0 to 1, representing the degree of dependency between functional nodes. The attenuation coefficient is a dimensionless parameter, determined by the distance attenuation function based on the network distance between nodes, ranging from 0.1 to 0.3. Since all three parameters are dimensionless, their product, the influence intensity, is also dimensionless. In the affected node state update formula, the original state value is a dimensionless parameter, representing the functional state level of the node, with an initial value of 1.0. The influence intensity is also a dimensionless parameter. Therefore, the calculation result of new state value = original state value × (1 - influence intensity) is also dimensionless. The failure threshold of 0.8 is a dimensionless parameter, sharing the same dimensional basis as the node state value.

[0085] Simultaneously, the spatial transmission path and attenuation degree of intensity deviations from various land parcels along the road network and energy pipeline distribution defined in the infrastructure layout data are assessed. Spatial transmission analysis employs a combination of network analysis and buffer analysis. First, based on the infrastructure layout data obtained in step S1, road network and energy pipeline network models are constructed. The road network model includes road grade, width, and capacity attributes, while the energy pipeline network model includes pipeline type, diameter, and transport capacity attributes. The transmission path search uses Dijkstra's algorithm to find the shortest path from the source land parcel to other land parcels. Transmission attenuation is then assessed. The coefficients are determined based on the path type and distance. Specifically, an attenuation coefficient lookup table is established by analyzing the propagation pattern of intensity deviation in historical data. For example, the attenuation coefficient for main roads is 0.1 per kilometer, for secondary roads it is 0.15 per kilometer, and for energy pipelines it is 0.2 per kilometer. The transmission impact range is determined by establishing multi-ring buffer zones. Buffer zones of 100 meters, 300 meters, and 500 meters are established along the road network with the plot as the center. The cumulative intensity deviation of the affected plots within each buffer zone is calculated. The cumulative value is calculated as the weighted sum of the intensity deviation values ​​of each affected plot, with the weights determined based on the distance attenuation coefficient.

[0086] Based on the comprehensive evaluation results of the intensity value affecting the global stability of the network and the degree of spatial transmission attenuation, key conflict sites are identified. The comprehensive evaluation adopts a weighted scoring method, with a network influence intensity weight of 0.6 and a spatial transmission influence weight of 0.4. The scoring threshold is set to 0.7. The network influence intensity value is calculated through cascading failure simulation results, taking the ratio of the final network efficiency to the initial network efficiency. The spatial transmission influence value is obtained by calculating the cumulative intensity deviation within a 500-meter buffer and normalizing it to the 0-1 range. When the comprehensive score of a site exceeds 0.7, it is identified as a key conflict site. For example, if the network influence intensity value of a site is 0.8 and the spatial transmission influence value is 0.6, then the comprehensive score is 0.8×0.6+0.6×0.4=0.72, which exceeds the threshold of 0.7 and is therefore identified as a key conflict site. The spatial location and attribute information of all key conflict sites are recorded in a dedicated layer. The layer is stored in Shapefile format and contains attribute fields such as site number, intensity deviation value, network influence intensity value, spatial transmission influence value, and comprehensive score, providing input for subsequent optimization.

[0087] S5. Taking key conflict areas as the core optimization region, construct a multi-scale optimization model that integrates macro-optimization objectives and micro-constraints. The specific implementation is as follows:

[0088] First, the immutable elements in the property boundary data and geological condition data of the key conflict plots identified in step S4 are defined as rigid constraints. The immutable elements in the property boundary data include the land use boundary and the use restrictions that are explicitly stated as unadjustable in the ownership registration information. For example, the land use boundary is determined based on the land parcel boundary in the real estate registration database. The use restrictions that are explicitly stated as unadjustable in the ownership registration information include construction restrictions within the cultural relic protection land area or use restrictions within the basic farmland protection area. The immutable elements in the geological condition data include steep slope areas with a slope greater than 25 degrees, high-risk areas for geological disasters, and underground cultural relic burial areas. The determination of these areas is based on the geological disaster risk assessment report and cultural relic survey data. The slope data is calculated through a digital elevation model. For example, the elevation change rate is calculated using a 3×3 pixel window, and the slope value is expressed in degrees. These rigid constraints serve as hard boundaries for the optimization solution and must be strictly adhered to during the optimization process. No solution may exceed these boundary conditions.

[0089] Adjustable elements in the infrastructure layout and geological condition data of key conflict sites are transformed into flexible constraints. The adjustable elements in the infrastructure layout data include the expandable range of the road network, the expandable capacity of energy pipelines, and the increaseable capacity of public service facilities. For example, the expandable range of the road network is determined by analyzing the difference between the road right-of-way width and the existing road width; the expandable capacity of energy pipelines is calculated based on the difference between the maximum allowable capacity in pipeline design specifications and the existing capacity; and the increaseable capacity of public service facilities is determined based on the difference between the population capacity within the service radius and the current service population. Adjustable elements in the geological condition data... The adjustment factors include developable areas with slopes between 8 and 25 degrees, areas where soil bearing capacity can be improved through engineering measures, and areas where groundwater conditions can be optimized through drainage facilities. These flexible constraints are added to the optimization model by setting penalty terms. The weight coefficient of the penalty terms is determined according to the importance and difficulty of the constraints. For example, the penalty weight for road widening is set to 0.3, the penalty weight for pipeline expansion is set to 0.2, and the penalty weight for engineering geological improvement is set to 0.5 by using the evaluation method of professionals in the field. The weight determination process is carried out by independent scoring by professionals in the field, and the average value is taken as the final weight.

[0090] A multi-objective function is constructed based on ecological protection indicators and spatial structure indicators in the macro-optimization objectives. Ecological protection indicators include vegetation coverage, biodiversity index, and the proportion of ecologically sensitive areas. Spatial structure indicators include the intensity of construction land development, the proportion of green space and open space, and transportation network connectivity. These indicators are normalized to a value range between 0 and 1, and assigned corresponding weight coefficients. The weight coefficients are determined using the analytic hierarchy process (AHP). For example, by having professionals in the field compare the importance of each indicator pairwise, a judgment matrix is ​​constructed, and the weight of the ecological protection indicator is calculated to be 0.6, and the weight of the spatial structure indicator is 0.4. Then, a penalty term for the elastic constraint is added to the multi-objective function. The penalty term adopts a quadratic penalty function form. When the solution violates the elastic constraint, a penalty value is generated. The penalty coefficient increases linearly according to the degree of violation. For example, when the road widening exceeds 10% of the widenable range, a penalty value of 0.1 is added for every 1% exceeding the limit.

[0091] A multi-scale optimization model is constructed based on hard boundaries and a multi-objective function with a penalty term. The model adopts a multi-level structure. The first level handles the micro-optimization within each plot, with decision variables including the land use type and utilization intensity of each key conflict plot. Land use types are divided into three categories: residential, commercial, and ecological. The utilization intensity range is determined according to the land use type; for example, the utilization intensity ranges from 1.0 to 3.0 for residential land, 2.0 to 5.0 for commercial land, and 0.1 to 1.0 for ecological land. The second level handles the coordination optimization between plots, with decision variables including the functional coordination and infrastructure sharing degree of adjacent plots. Functional coordination is calculated by considering the adjacent plots. The land use type compatibility score is determined, and the degree of infrastructure sharing is calculated by analyzing the coverage and service capacity of public service facilities. The optimization algorithm adopts a constrained multi-objective genetic algorithm with a population size of 200 individuals, 500 iterations, a crossover probability of 0.85, and a mutation probability of 0.15. When the algorithm runs, it first checks whether the solution meets all rigid constraints and removes individuals that do not meet them. Then, it calculates the objective function value and the penalty term value. Finally, it selects the optimal individual through non-dominated sorting and crowding calculation and outputs the Pareto optimal solution set. Each solution in the solution set represents a planning scheme that achieves a balance among multiple optimization objectives.

[0092] S6. Generate optimal allocation schemes for land elements through a multi-scale optimization model, and output the allocation results. The specific implementation is as follows:

[0093] First, a fast non-dominated sorting genetic algorithm with an elitist strategy is used to solve the multi-scale optimization model constructed in step S5. The algorithm parameters are set as follows: population size of 200 individuals, number of iterations of 500 generations, crossover probability of 0.85, and mutation probability of 0.15. These parameter values ​​are determined through pre-experiments, for example, by conducting 10 experiments with different parameter combinations and selecting the parameter combination with the best convergence effect. During the solution process, each individual represents a planning scheme, including the land use type and utilization intensity allocation of key conflict plots. The land use type is encoded using integers, 1 1 represents residential land, 2 represents commercial land, and 3 represents ecological land. The utilization intensity is encoded using real numbers, and the value range is determined according to the land use type. For example, the utilization intensity of residential land ranges from 1.0 to 3.0, the utilization intensity of commercial land ranges from 2.0 to 5.0, and the utilization intensity of ecological land ranges from 0.1 to 1.0. After multiple generations of evolution, the Pareto optimal solution set of land use type and utilization intensity of key conflicting plots is obtained. This solution set contains multiple non-dominated solutions, each of which represents a planning scheme that achieves a balance among multiple optimization objectives.

[0094] By combining the existing intensity data of non-critical conflict plots in the land use intensity distribution map, the optimal solution for critical conflict plots is integrated with the existing intensity data of non-critical conflict plots. The integration process is achieved through spatial connection and attribute field merging. First, the land use intensity distribution map generated in step S3 is overlaid with the critical conflict plot range identified in step S4. The spatial query function of the geographic information system is used to distinguish between critical and non-critical conflict plots. Then, the solution with the highest comprehensive score is selected from the Pareto optimal solution set as the final solution. The comprehensive score is calculated by weighted summation, and the weight coefficients are determined according to the importance of the optimization objective. For example, the ecological protection index has a weight of 0.6, and the spatial structure index has a weight of 0.4. Finally, the optimization results of critical conflict plots are merged with the existing intensity data of non-critical conflict plots to form a complete land element configuration dataset. During the merging process, the consistency of attribute fields and the integrity of data are ensured.

[0095] A land feature optimization configuration scheme is generated, which includes a vector layer containing the final land use type and utilization intensity allocation results for each plot. The vector layer adopts the standard Shapefile format and includes attribute fields such as plot number, land use type, utilization intensity, and optimization identifier. The land use type field stores integer codes, where 1 represents residential land, 2 represents commercial land, and 3 represents ecological land. The utilization intensity field stores real numbers. The optimization identifier field is used to distinguish the optimization results of key conflict plots from the original data of non-key conflict plots. The layer coordinate system is consistent with the data in step S1, using the CGCS2000 national geodetic coordinate system to ensure the consistency of spatial data. At the same time, a legend file and a symbolic configuration file are generated for easy visualization.

[0096] The system outputs vector layer files of optimized land feature allocation schemes. File output is batch-processed, generating complete vector data including point, line, and polygon feature types. It also outputs attribute data tables, spatial index files, and metadata files. The attribute data tables contain complete attribute information for all land parcels. The spatial index file uses an R-tree index structure to improve query efficiency. The metadata file contains information such as data source, processing time, coordinate system, and field descriptions. All files are packaged in ZIP compressed format and provided through a data interface, supporting reading and use by standard geographic information system software. This provides directly applicable results data for land planning and management. Data quality checks are performed during the output process to ensure data integrity and accuracy.

[0097] Example 2: Figure 2 A schematic diagram of a land element optimization allocation decision system based on artificial intelligence is provided according to the present invention. The land element optimization allocation decision system based on artificial intelligence includes:

[0098] The information acquisition module is used to acquire the macro-optimization objectives of the target area and the micro-constraints at the plot level.

[0099] The parameter generation module is used to calculate the probability distribution of each plot adapting to different land use types based on micro-constraints and the attributes of neighboring plots. It uses the information entropy algorithm to quantify the uncertainty of mixed use of the probability distribution in order to generate entropy-based mixed use potential parameters for each plot.

[0100] The distribution generation module is used to generate a distribution map of the usable intensity of land parcels by taking entropy-valued mixed utilization potential parameters as the core input, and preprocessing and analyzing the usable intensity of each parcel through a multi-objective optimization algorithm.

[0101] The conflict identification module is used to combine the land use intensity distribution map with macro-optimization objectives to simulate the chain impact effect of intensity deviations of various land parcels on the functions of key areas and assess their spatial diffusion risk, thereby identifying key conflict land parcels.

[0102] The model building module is used to construct a multi-scale optimization model that integrates macro-optimization objectives and micro-constraints, with key conflict plots as the core optimization area.

[0103] The results output module is used to generate optimal allocation schemes for land elements through a multi-scale optimization model and output the allocation results.

[0104] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0105] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0106] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center containing one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0107] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0108] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0109] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0110] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0111] If the aforementioned functions are implemented as software functional modules 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 application, in essence, 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 application. 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.

[0112] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0113] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An artificial intelligence-based land element optimization configuration decision method, characterized in that, The method comprises the following steps: S1, obtaining macroscopic optimization targets of a target region and microcosmic constraint conditions at a plot level, comprising: obtaining ecological protection indexes, spatial structure indexes and environmental quality indexes of the target region as the macroscopic optimization targets; and obtaining property boundary data, geological condition data and infrastructure layout data at the plot level as the microcosmic constraint conditions; wherein the property boundary data comprises red line vector data and property registration information, the geological condition data comprises soil type distribution and terrain slope data, and the infrastructure layout data comprises road network and energy pipeline distribution vector data; S2, calculating probability distribution of each plot adapting to different land use types based on the microcosmic constraint conditions and adjacent plot attributes, and using an information entropy algorithm to quantize the mixed use uncertainty of the probability distribution to generate an entropy- valued mixed use potential parameter of each plot; S3, taking the entropy- valued mixed use potential parameter as a core input, preprocessing and analyzing the available intensity of each plot by a multi- objective optimization algorithm to generate a plot available intensity distribution map; S4, combining the plot available intensity distribution map with the macroscopic optimization targets, simulating the chain impact effect of intensity deviation of each plot on key region functions and evaluating the spatial diffusion risk thereof, and identifying key conflict plots, comprising: calculating an intensity deviation value of each plot based on the intensity value of each plot in the plot available intensity distribution map and the expected value of the corresponding index in the macroscopic optimization targets; constructing a key region function dependency network, wherein the nodes include an ecological integrity maintenance function and a traffic connectivity guarantee function, and the edges represent the interdependence relationship between functions; simulating the cascading failure process caused by the intensity deviation of each plot in the function dependency network, and calculating the influence intensity thereof on the global stability of the network; simultaneously evaluating the path and attenuation degree of the spatial conduction of the intensity deviation of each plot along the road network and energy pipeline distribution defined in the infrastructure layout data; identifying key conflict plots according to the comprehensive evaluation results of the influence intensity on the global stability of the network and the spatial conduction attenuation degree; S5, taking the key conflict plots as a core optimization region, and constructing a multi- scale optimization model integrating the macroscopic optimization targets and the microcosmic constraint conditions; S6, generating a land element optimization configuration scheme through the multi- scale optimization model, and outputting a configuration result.

2. The land element optimization arrangement decision method based on artificial intelligence according to claim 1, characterized in that, S2, calculating probability distribution of each plot adapting to different land use types based on the microcosmic constraint conditions and adjacent plot attributes, and using an information entropy algorithm to quantize the mixed use uncertainty of the probability distribution to generate an entropy- valued mixed use potential parameter of each plot, comprising: based on the property boundary data, the geological condition data and the infrastructure layout data, analyzing the suitability grades of each plot for residential land, commercial land and ecological land; combining the existing land use type distribution in the adjacent plot attributes, calculating the conditional probability of each plot adapting to different land use types; combining the conditional probability of each plot corresponding to different land use types into a probability distribution vector; using an information entropy algorithm to calculate the information entropy value of each plot probability distribution vector, and taking the information entropy value as an entropy- valued mixed use potential parameter representing the mixed use uncertainty.

3. The land element optimization arrangement decision method based on artificial intelligence according to claim 1, characterized in that, In combination with the existing land use type distribution in the attribute of the adjacent plots, the conditional probability of each plot adapting to different land use types is calculated by statistically determining the area proportion of each land use type in the buffer zone set around the target plot, and the area proportion is weighted and corrected according to the property boundary data and the geological condition data in the micro-constraint conditions of the plot, so as to obtain the conditional probability value of the plot adapting to each land use type.

4. The land element optimization arrangement decision method based on artificial intelligence according to claim 2, characterized in that, The entropy-mixed utilization potential parameter is taken as the core input, and the available intensity of each plot is preprocessed and analyzed by a multi-objective optimization algorithm to generate an available intensity distribution map of the plot, including: The entropy-mixed utilization potential parameter is taken as the core input, and the available intensity of each plot is preprocessed and analyzed by a multi-objective optimization algorithm to generate an available intensity distribution map of the plot, including: The entropy-mixed utilization potential parameter is taken as the core input, and the available intensity of each plot is preprocessed and analyzed by a multi-objective optimization algorithm to generate an available intensity distribution map of the plot, including: The entropy-mixed utilization potential parameter is taken as the core input, and the available intensity of each plot is preprocessed and analyzed by a multi-objective optimization algorithm to generate an available intensity distribution map of the plot, including:

5. The land element optimization arrangement decision method based on artificial intelligence according to claim 4, characterized in that, The entropy-mixed utilization potential parameter is taken as the core input, and the available intensity of each plot is preprocessed and analyzed by a multi-objective optimization algorithm to generate an available intensity distribution map of the plot, including:

6. The land element optimization arrangement decision method based on artificial intelligence according to claim 4, characterized in that, The key conflict plot is taken as the core optimization area, and a multi-scale optimization model integrating the macro-optimization target and the micro-constraint condition is constructed, including: The property boundary data and the geological condition data of the key conflict plot are defined as rigid constraint conditions and used as the hard boundary of the optimization solution; The infrastructure layout data and the adjustable elements in the geological condition data of the key conflict plot are converted into elastic constraint conditions, and are added as penalty terms into the multi-objective function constructed by the ecological protection index and the spatial structure index in the macro-optimization target; A multi-scale optimization model is constructed based on the hard boundary and the multi-objective function with penalty terms.

7. The land element optimization arrangement decision method based on artificial intelligence according to claim 6, characterized in that, The land element optimization configuration scheme is generated by the multi-scale optimization model, and the configuration result is output, including: The multi-objective optimization algorithm is used to solve the multi-scale optimization model to obtain the Pareto optimal solution set of the land use type and the utilization intensity of the key conflict plot; The optimal solution of the key conflict plot is integrated with the original intensity data of the non-key conflict plot in the plot available intensity distribution map; The land element optimization configuration scheme including the final land use type and the utilization intensity distribution result of each plot is generated, and the vector layer file of the land element optimization configuration scheme is output.

8. An artificial intelligence-based land element optimal allocation decision system for implementing the artificial intelligence-based land element optimal allocation decision method of any one of claims 1-7. The information acquisition module is used to acquire the macro-optimization target of the target area and the micro-constraint condition at the plot level; The parameter generation module is used to calculate the probability distribution of each plot adapting to different land use types based on the micro-constraint condition and the attribute of the adjacent plots, and to quantify the mixed utilization uncertainty of the probability distribution by using the information entropy algorithm to generate the entropy-mixed utilization potential parameter of each plot; ​ The distribution generation module is configured to take the potential parameter as a core input in an entropy value mixing manner, preprocess and analyze the available intensity of each plot by a multi-objective optimization algorithm, and generate a plot available intensity distribution map; The conflict identification module is configured to combine the plot available intensity distribution map with the macro-optimization target, simulate the chain impact effect of the intensity deviation of each plot on the function of the key area and evaluate the spatial diffusion risk thereof, and identify key conflict plots; The model construction module is configured to take the key conflict plots as a core to optimize the area and construct a multi-scale optimization model that integrates the macro-optimization target and the micro-constraint condition; The result output module is configured to generate a land element optimization configuration scheme by the multi-scale optimization model and output a configuration result.

Images