Land space optimization configuration method under terrain gradient constraint

Abstract

The invention discloses a land space optimization configuration method under terrain gradient constraint. The method comprises the following steps: S1, collecting multi-source geographic data to construct a spatial database; s2, factors such as gradient and elevation are extracted, and terrain gradient partitions are divided; s3, constructing a three-dimensional conflict matrix based on suitability evaluation and landscape pattern analysis; s4, constructing a driving index system, and analyzing a driving path of the spatial conflict; s5, setting a multi-objective optimization function, and constructing game rules and spatial conversion constraint conditions of the three types of subjects; s6, integrating the C-FLUS model, the MAOOAM model and the R2D3 model, and constructing a coupling system to execute dynamic optimization; and S7, when the convergence condition is met, outputting an optimal configuration scheme and generating a compensation and control strategy. According to the method, efficient optimization and intelligent decision support of multi-target land configuration under the terrain differentiation condition are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of land space optimization, and in particular to a land space optimization configuration method under terrain gradient constraints. Background Art

[0002] In my country's current land resource development and utilization, faced with the coordinated constraints of multiple objectives, including farmland protection and ecological restoration, achieving scientific, adaptable, and intelligent land spatial allocation has become a core issue in national land space governance. In particular, in areas with complex terrain, such as hilly and mountainous areas, topographic conditions significantly influence land use suitability, accessibility, and ecological sensitivity. Traditional planning methods often fail to fully address topographic heterogeneity and spatial conflicts, resulting in a structural mismatch between land use layout and ecological carrying capacity, severely restricting the efficient allocation and sustainable use of land resources.

[0003] In existing technologies, the more common land spatial optimization allocation methods mainly rely on static land conversion probabilities and single suitability evaluation factors for spatial simulation, ignoring the coupling effect of complex terrain elements such as terrain undulation and slope classification in the spatial structure. For example, some land simulation methods based on Markov chains and CA models are only applicable to areas with flat terrain. They are often unable to achieve accurate simulation and zoning control for mountainous plots with strong heterogeneity in undulation and slope aspect. In addition, suitability evaluation methods are mostly based on static indicator weighting methods, which fail to consider the interaction between plot structure pattern, neighborhood morphological characteristics and ecological risks. As a result, the configuration logic of cultivated land, gardens and woodlands is not systematic enough and cannot support the comprehensive decision-making needs of multi-objective optimization.

[0004] Traditional approaches to spatial conflict resolution often rely on binary overlap analysis or weighted superposition to identify conflict zones. These approaches lack systematic modeling of the underlying drivers of conflict and struggle to identify the structural connections between the multidimensional factors involved. Some studies have introduced conflict matrix analysis models, but these remain at the qualitative grading stage, lacking a closed-loop relationship with driving pathways and policy regulation logic. Furthermore, most current spatial allocation methods still employ single-objective or linearly weighted multi-objective optimization, making it difficult to simultaneously address the dynamic game between the three objectives of food production, ecological protection, and economic benefits. Furthermore, they lack the systematic introduction of multi-agent response and evolution mechanisms.

[0005] At the simulation tool level, while some research has begun to incorporate intelligent algorithms such as genetic algorithms and particle swarm optimization to solve land spatial allocation problems, these approaches lack a deep integration of spatial constraints, policy boundaries, terrain differentiation, and behavioral response mechanisms. Consequently, optimization results are prone to local optima, and spatial patterns lack realistic evolutionary characteristics. Recent studies have attempted to integrate multi-agent models with reinforcement learning frameworks for complex system simulations, but most approaches remain in the experimental verification phase and lack effective coupling mechanisms with land use simulation models, hindering the development of integrated spatial allocation solutions tailored to terrain gradient constraints.

[0006] More importantly, in the real-world scenario with the participation of multiple subjects, government departments, enterprises and farmers have different behavioral goals and response mechanisms for spatial configuration. Existing methods often fail to establish a dynamic response relationship between behavioral driving functions and configuration variables, and lack a unified modeling framework based on farmer behavior probability, enterprise bidding ability and government regulation intensity, resulting in a deviation between the actual output configuration results and policy intentions.

[0007] Overall, the existing land space optimization methods have the following significant deficiencies: First, there is a lack of spatial zoning and suitability linkage mechanism based on terrain gradient, which cannot fully express the comprehensive restrictive relationship between terrain and land use; second, spatial conflict analysis lacks driving path modeling and response factor identification, and cannot effectively support structural adjustment and transformation judgment; third, the multi-objective optimization process ignores the interaction of subject behavior and strategic response, and cannot achieve comprehensive configuration of multiple objectives, multiple subjects, and multiple constraints; fourth, there is a lack of deep coupling and interactive control mechanism between simulation systems, which leads to the separation of optimization execution and pattern evolution, and the spatial results are difficult to truly reflect the effect of policy implementation.

[0008] Therefore, how to provide a land space optimization configuration method under terrain gradient constraints is an urgent problem that technicians in this field need to solve. Summary of the Invention

[0009] One objective of the present invention is to propose a method for optimizing land spatial allocation under terrain gradient constraints. This method integrates terrain zoning analysis, land suitability assessment, spatial conflict modeling, multi-objective optimization, and multi-agent reinforcement learning simulation. It constructs a coupled system that integrates the C-FLUS land use simulation model, the MAOOAM multi-agent behavior simulation model, and the R2D3 reinforcement learning optimization model. The system performs dynamic simulation and intelligent optimization of land spatial allocation. This method describes in detail the complete process for achieving scientific allocation of cultivated land, gardens, and woodlands under complex terrain conditions. It boasts clear structural expression, precise subject behavior modeling, and intelligent configuration strategy generation, making it suitable for spatial planning and land management tasks in areas with significant terrain undulations.

[0010] A method for optimizing land space allocation under terrain gradient constraints according to an embodiment of the present invention includes the following steps:

[0011] S1, collect data and build spatial database;

[0012] S2, extracting slope, elevation, relief and aspect heterogeneity from the spatial database, calculating the terrain comprehensive index, and dividing the terrain gradient into zones;

[0013] S3. Based on the spatial database and terrain gradient zoning, the suitability evaluation method is used to calculate the suitability levels of cultivated land, garden land, and forest land. Morphological spatial pattern analysis and landscape pattern index are used to identify structural and morphological characteristics, and a conflict matrix is constructed.

[0014] S4. Based on the spatial database and conflict matrix, a driving indicator system was constructed, and the driving paths of spatial conflicts were analyzed using structural equation models and logistic regression models.

[0015] S5. Based on the conflict matrix and driving path, a multi-objective optimization function with grain production, ecological services, and economic benefits as targets is set, and the game rules and spatial transformation constraints of the government, enterprises, and farmers are constructed;

[0016] S6: Integrate the C-FLUS land use simulation model, the MAOOAM multi-agent behavior simulation model, and the R2D3 reinforcement learning optimization model to build a PLUS-MAS-DRL coupled system to perform spatial configuration simulation and dynamic optimization.

[0017] S7. Based on the optimization results, when the conflict area change rate and the multi-objective optimization function converge to meet the set threshold, the optimal land space allocation plan that meets the terrain gradient constraint is output, and the slope zoning compensation and spatial layout control strategy is generated through the decision support system.

[0018] Optionally, the multi-source geographic data includes remote sensing images, digital elevation models, land use status, farmer surveys and policy data.

[0019] Optionally, the conflict matrix is a three-dimensional conflict matrix including structural conflict, functional conflict and morphological conflict.

[0020] Optionally, the driving indicator system includes natural factors, economic factors, social factors, policy factors and farmer factors.

[0021] Optionally, the S2 specifically includes:

[0022] S21. Extract the slope, elevation, relief and aspect heterogeneity factors in the study area from the spatial database and construct a terrain factor set;

[0023] S22. Calculate the terrain comprehensive index of each spatial unit based on the terrain factor set:

[0024] TCI = S × E × R × D;

[0025] Among them, TCI is the terrain comprehensive index, S is the slope, E is the elevation, R is the relief, and D is the slope heterogeneity;

[0026] S23. The terrain comprehensive index is classified by the natural breakpoint method according to the set threshold value, and divided into three terrain gradient zones: river valley plain area, low mountain and hilly area, and medium mountain steep slope area;

[0027] S24. Generate a partition code for each terrain gradient partition as a basic input parameter for land suitability evaluation, spatial conflict identification and spatial optimization simulation.

[0028] Optionally, the S3 specifically includes:

[0029] S31. Based on the spatial database and terrain gradient zoning, use the land suitability evaluation method to classify the suitability of cultivated land, garden land and forest land, and generate a suitability grade map;

[0030] S32. Based on the suitability level layer, calculate the landscape pattern index including patch density, boundary density and aggregation, and construct a structural feature layer;

[0031] S33, using the morphological spatial pattern analysis model to identify spatial boundary morphology and connectivity features, and constructing a morphological feature layer;

[0032] S34. Based on the spatial database and policy parameters, the ecosystem service assessment model is called to construct the functional characteristic layer;

[0033] S35. Spatially superimpose the suitability level layer, structural feature layer, morphological feature layer, and functional feature layer to construct a conflict matrix. The conflict matrix contains three types of layer channels: structural conflict, functional conflict, and morphological conflict. The pixel value in each channel represents the intensity level of the corresponding conflict.

[0034] Optionally, the S4 specifically includes:

[0035] S41. Based on the spatial database and conflict matrix, a driving indicator system is constructed. The driving factors include natural factors, economic factors, social factors, policy factors, and farmer factors. Natural factors include slope, soil type, and precipitation; economic factors include land management income and land market price; social factors include transportation accessibility and population density; policy factors include land use restriction level and subsidy policy intensity; and farmer factors include farmer land size and farming willingness.

[0036] S42. Use structural equation modeling to analyze the causal path relationship between driving factors:

[0037] Y=ΛX+ε;

[0038] Where Y represents the conflict intensity response vector, X represents the driving factor observation variable matrix, Λ represents the causal path coefficient matrix, and ε represents the model error vector;

[0039] S43. For each conflict type, a multivariate logistic regression model was used to model its response probability under different driving factors, and the influencing factors and significance coefficients of structural conflict, functional conflict, and morphological conflict were obtained;

[0040] S44. Based on the joint analysis results of the structural equation model and the logistic regression model, the spatial conflict driving paths, including the dominant factor path and the interaction factor path, are extracted for the parameter setting of the multi-objective optimization function and the game rules.

[0041] Optionally, the S5 specifically includes:

[0042] S51. Set a multi-objective optimization function with grain production, ecological services and economic benefits as optimization targets, and construct the objective function:

[0043]

[0044] in, Boolean variables indicating that the location (i, j) is allocated as cultivated land, forest land, and garden land, respectively, y ij Indicates the unit grain output at position (i, j), esv ij Indicates the ecological service value, eco ij Indicates unit economic benefits, P p (i, j, t) represents the probability of a farmer’s response to the land type at location (i, j) at time t, P Dk (i,j,t) represents the government department D k The probability of preference for location (i, j) and land type t, B ij represents the bidding expression value of the enterprise at position (i, j);

[0045] S52. Constructing the coordination and competition function of government entities:

[0046]

[0047] Among them, C t (i, j) represents the coordination competition degree of land type t at location (i, j), ω k represents the regulatory weight of the k-th government department, UD k(i, j, t) represents the regulatory utility of the department on land type t, α represents the farmer's response weight, UP p (i, j, t) represents the farmer's response utility to land type t, η represents the enterprise's response weight, UE E (i, j, t) represents the enterprise response utility, P t represents the policy priority of land type t;

[0048] S53. Construct the bidding expression function of the enterprise entity:

[0049]

[0050] Among them, B ij represents the competitive value of the enterprise at position (i, j), π 预期 represents the expected profit of the enterprise, T represents the level of technological input, λ represents the ecological penalty coefficient, R eco represents the ecological risk value;

[0051] S54. Construct the response probability function of the farmer entity:

[0052]

[0053] Among them, P p (i, j, t) represents the response probability of the farmer at location (i, j), land type t, and time t, π 果 represents the orchard income, π 耕 represents the income from cultivated land, σ 政策 represents policy sensitivity, ρ represents the risk aversion coefficient, and risk represents the expected value of loss;

[0054] S55. Construct spatial preference probability function of government departments:

[0055]

[0056] Among them, P Dk (i,j,t) represents department D k The preference probability of land type t at location (i, j), β represents the spatial clustering coefficient, R neigh represents the proportion of similar land use in the neighborhood, γ represents the suitability coefficient, Department D k Suitability score for designated land types;

[0057] S56. Construct game rules to control the allocation priority of the government subject through the coordination competition function, control the land acquisition ranking of the enterprise subject through the bidding expression function, and drive the land conversion behavior of the farmer subject through the response probability function. The response functions of the three types of subjects jointly participate in the calculation weight of the spatial configuration variable.

[0058] S57. Set space transformation constraints:

[0059] The intensity constraint of coordinating competition not lower than a given threshold is used to eliminate low-regulation regions;

[0060] The ratio of forest land to garden land shall not exceed the set upper limit;

[0061] Cross-level land type conversion operations are not allowed for plots between terrain gradient zones.

[0062] Optionally, the S6 specifically includes:

[0063] S61. Construct a C-FLUS land use simulation model. Input the terrain gradient partition, initial land use map, land class transition probability matrix, suitability factor layer, and spatial constraint layer into the model. Generate a multi-time series land use simulation raster sequence based on the Markov chain, and output a basic configuration raster map containing the spatial status of cultivated land, garden land, and forest land.

[0064] S62. Construct a MAOOAM multi-agent behavior simulation model. Use the coordination competition function of the government, the bidding expression function of the enterprise, and the response probability function of the farmer as agent behavior strategy inputs. Construct a multi-agent state map based on the two-dimensional location of the plot, the land type, and the behavior utility vector. Simulate the competition, response, and adjustment behavior of the government, enterprises, and farmers for the plot during the land allocation process, and generate a plot behavior response layer containing the agent behavior trajectories.

[0065] S63. Build an R2D3 reinforcement learning optimization model. This model takes the current land state, historical configuration actions, and immediate benefits as inputs, constructs an optimization structure consisting of a state encoding network, a strategy generation network, and a Q-value feedback network. Based on the grain production, ecological service, and economic benefit goals set in the multi-objective optimization function, it learns the impact of different behavior sequences on the achievement of long-term goals and outputs the optimal action allocation strategy for the spatial configuration variables.

[0066] S64. Construct a PLUS-MAS-DRL coupled system, using the spatial configuration trend map output by C-FLUS as the initial state of the simulation, the multi-agent response layer output by MAOOAM as the behavior input tensor, and the configuration strategy output by the R2D3 model as the iterative controller. The spatial configuration variables are updated through the state fusion and behavior weighting mechanism, and the land type of each plot is dynamically assigned and adjusted.

[0067] S65. Set iterative update conditions and convergence criteria in the coupled system, record the change range of the spatial configuration variables in each round, and when the land type change rate of all plots and the target difference of the multi-objective optimization function are lower than the preset threshold, the system is judged to have converged, and the final configuration plan grid map containing the spatial distribution of cultivated land, garden land and woodland is output.

[0068] Optionally, the S7 specifically includes:

[0069] S71. Compare the spatial configuration variables output by the PLUS-MAS-DRL coupling system in two consecutive rounds, extract the land type changes of each plot, and calculate the change ratio of the conflict area corresponding to the plot;

[0070] S72, calculating the numerical difference of the multi-objective optimization function in consecutive rounds, and determining whether the preset convergence standard is met;

[0071] S73. When the difference between the conflict area change ratio and the multi-objective optimization function value is lower than a set threshold, a final spatial configuration result layer including the configuration status of cultivated land, garden land, and forest land is generated;

[0072] S74, spatially overlaying the spatial configuration result layer with the terrain gradient partition layer to extract the land type distribution within each slope grade area;

[0073] S75. Construct a conflict risk-intelligent control decision support system, wherein the system has the functions of slope grade identification, land type identification, conflict intensity reading, spatial parameter combination, and control layer construction, and is used to process the structural information contained in the spatial configuration result layer, terrain gradient partition layer, multi-objective optimization function layer, and conflict matrix layer;

[0074] S76. Input the spatial configuration result layer, the terrain gradient partition layer, the multi-objective optimization function layer, and the conflict matrix layer into the conflict risk-intelligent control decision support system to generate a slope grade compensation table and a spatial layout control layer. The slope grade compensation table is used to represent the control parameters for land type configuration in areas of different slope grades, and the spatial layout control layer is used to express land conversion boundaries, control intensity levels, and spatial conversion constraint ranges.

[0075] S77. Export the slope grade compensation table and the spatial layout control layer as a structured policy configuration dataset and load it into the land management and planning module for executing land use configuration, spatial regulation, and management instruction generation operations.

[0076] The beneficial effects of the present invention are:

[0077] By constructing a method for optimizing land spatial allocation under terrain gradient constraints, this paper effectively addresses existing issues such as insufficient response to terrain heterogeneity, one-sided spatial conflict analysis, weak multi-objective coordination, and fragmented simulation systems. Compared to traditional methods that rely solely on static suitability or single transformation rules for land allocation simulation, this paper incorporates multidimensional terrain indicators such as slope, elevation, relief, and aspect heterogeneity during the data collection and processing stages. This system constructs a terrain zoning system with spatial continuity and gradient identifiability, providing a sophisticated foundation for subsequent land suitability evaluation and spatial constraint modeling.

[0078] In terms of spatial conflict management, this paper proposes a three-dimensional conflict matrix structure, combining structural characteristics, functional contradictions, and morphological mismatches to achieve a comprehensive representation of spatial conflict. Furthermore, by constructing a driving indicator system that integrates natural, social, economic, and farmer response factors, and using structural equation models and logistic regression models to identify the multi-path driving mechanisms of conflict, this approach fills a technical gap in traditional conflict analysis, which fails to reveal the underlying causes of behavior.

[0079] In terms of optimization and simulation, this paper sets a multi-objective optimization function with grain production, ecological services, and economic benefits as its goals. It establishes game rules and spatial conversion constraints for the three main entities of government, enterprises, and farmers, and realizes dynamic control modeling based on real behavioral responses. By integrating the C-FLUS, MAOOAM, and R2D3 models, a PLUS-MAS-DRL coupling system is constructed, integrating land use trends, subject behavior simulation, and reinforcement learning strategy optimization. This breaks through the coupling barrier between spatial configuration simulation and behavioral decision optimization, forming a data-driven, model-integrated, feedback-closed-loop spatial intelligent configuration mechanism.

[0080] Ultimately, under the conditions of conflict change rate and optimization function convergence, the present invention outputs an optimal land allocation plan that satisfies terrain gradient conditions. Combined with the allocation results, it generates slope zoning compensation and spatial layout control strategies, providing structured support for policy formulation and spatial management. Overall, the present invention achieves high-precision land allocation optimization under multiple objectives, multiple agents, and multiple constraints in complex terrain areas, enhancing the scientific nature of planning decisions, the intelligence of the control model, and the adaptability of the allocation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0082] Figure 1 This is a flow chart of a land space optimization configuration method under terrain gradient constraints proposed by the present invention;

[0083] Figure 2This is a schematic diagram of the structure of a three-dimensional conflict matrix for a land space optimization configuration method under terrain gradient constraints proposed by the present invention;

[0084] Figure 3 This is a structural diagram of the PLUS-MAS-DRL coupling system of a land space optimization configuration method under terrain gradient constraints proposed in this invention. DETAILED DESCRIPTION

[0085] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0086] refer to Figure 1-3 A land space optimization configuration method under terrain gradient constraints includes the following steps:

[0087] S1, collect data and build spatial database;

[0088] S2, extracting slope, elevation, relief and aspect heterogeneity from the spatial database, calculating the terrain comprehensive index, and dividing the terrain gradient into zones;

[0089] S3. Based on the spatial database and terrain gradient zoning, the suitability evaluation method is used to calculate the suitability levels of cultivated land, garden land, and forest land. Morphological spatial pattern analysis and landscape pattern index are used to identify structural and morphological characteristics, and a conflict matrix is constructed.

[0090] S4. Based on the spatial database and conflict matrix, a driving indicator system was constructed, and the driving paths of spatial conflicts were analyzed using structural equation models and logistic regression models.

[0091] S5. Based on the conflict matrix and driving path, a multi-objective optimization function with grain production, ecological services, and economic benefits as targets is set, and the game rules and spatial transformation constraints of the government, enterprises, and farmers are constructed;

[0092] S6: Integrate the C-FLUS land use simulation model, the MAOOAM multi-agent behavior simulation model, and the R2D3 reinforcement learning optimization model to build a PLUS-MAS-DRL coupled system to perform spatial configuration simulation and dynamic optimization.

[0093] S7. Based on the optimization results, when the conflict area change rate and the multi-objective optimization function converge to meet the set threshold, the optimal land space allocation plan that meets the terrain gradient constraint is output, and the slope zoning compensation and spatial layout control strategy is generated through the decision support system.

[0094] By constructing a complete set of land spatial optimization configuration processes under terrain gradient constraints, the present invention can realize the spatial organization and dynamic simulation of land resources with the support of multi-source data. It breaks through the problems of traditional methods such as insufficient adaptation to terrain heterogeneity, weak multi-objective conflict coordination ability and unintelligent optimization decision-making, and forms an integrated configuration system driven by data, model collaboration and strategy feedback.

[0095] In this embodiment, the multi-source geographic data includes remote sensing images, digital elevation models, land use status, farmer surveys and policy data.

[0096] This invention uses multi-source geographic information such as remote sensing images, digital elevation models, land use status, farmer surveys and policy data as basic input, which improves the comprehensiveness and accuracy of the comprehensive expression of terrain, ecology and human factors, and provides solid data support for building a high-quality spatial database.

[0097] In this embodiment, the conflict matrix is a three-dimensional conflict matrix including structural conflict, functional conflict and morphological conflict.

[0098] The present invention constructs a three-dimensional conflict matrix composed of three types of elements: structural conflict, functional conflict and morphological conflict, which realizes the fine identification of spatial incoherence at the structural layer, ecological layer and morphological layer, breaking through the problems of one-sided conflict identification and insufficient dimensionality in the past.

[0099] In this embodiment, the driving indicator system includes natural factors, economic factors, social factors, policy factors and farmer factors.

[0100] The present invention constructs a driving indicator system including natural factors, economic factors, social factors, policy factors and farmer factors, establishes the causal logic between the spatial conflict occurrence mechanism and behavioral path, and provides a data basis and logical support for the establishment of subsequent optimization models and game rules.

[0101] In this embodiment, S2 specifically includes:

[0102] S21. Extract the slope, elevation, relief and aspect heterogeneity factors in the study area from the spatial database and construct a terrain factor set;

[0103] S22. Calculate the terrain comprehensive index of each spatial unit based on the terrain factor set:

[0104] TCI = S × E × R × D;

[0105] Among them, TCI is the terrain comprehensive index, S is the slope, E is the elevation, R is the relief, and D is the slope heterogeneity;

[0106] S23. The terrain comprehensive index is classified by the natural breakpoint method according to the set threshold value, and divided into three terrain gradient zones: river valley plain area, low mountain and hilly area, and medium mountain steep slope area;

[0107] S24. Generate a partition code for each terrain gradient partition as a basic input parameter for land suitability evaluation, spatial conflict identification and spatial optimization simulation.

[0108] The present invention realizes the classification and zoning of spatial units under multidimensional indicators such as slope, elevation, undulation and slope heterogeneity through the construction of terrain factor sets and calculation of terrain comprehensive index, forming a gradient spatial pattern suitable for suitability evaluation and simulation constraints of cultivated land, gardens and woodlands.

[0109] In this embodiment, S3 specifically includes:

[0110] S31. Based on the spatial database and terrain gradient zoning, use the land suitability evaluation method to classify the suitability of cultivated land, garden land and forest land, and generate a suitability grade map;

[0111] S32. Based on the suitability level layer, calculate the landscape pattern index including patch density, boundary density and aggregation, and construct a structural feature layer;

[0112] S33, using the morphological spatial pattern analysis model to identify spatial boundary morphology and connectivity features, and constructing a morphological feature layer;

[0113] S34. Based on the spatial database and policy parameters, the ecosystem service assessment model is called to construct the functional characteristic layer;

[0114] S35. Spatially superimpose the suitability level layer, structural feature layer, morphological feature layer, and functional feature layer to construct a conflict matrix. The conflict matrix contains three types of layer channels: structural conflict, functional conflict, and morphological conflict. The pixel value in each channel represents the intensity level of the corresponding conflict.

[0115] The present invention combines the land suitability grade layer, landscape pattern index, morphological characteristic layer and ecological function characteristics to establish a complete spatial feature map, and generates a conflict matrix by overlaying layers, thereby enhancing the systematicness and spatial expression ability of land space evaluation.

[0116] In this embodiment, the S4 specifically includes:

[0117] S41. Based on the spatial database and conflict matrix, a driving indicator system is constructed. The driving factors include natural factors, economic factors, social factors, policy factors, and farmer factors. Natural factors include slope, soil type, and precipitation; economic factors include land management income and land market price; social factors include transportation accessibility and population density; policy factors include land use restriction level and subsidy policy intensity; and farmer factors include farmer land size and farming willingness.

[0118] S42. Use structural equation modeling to analyze the causal path relationship between driving factors:

[0119] Y=ΛX+ε;

[0120] Where Y represents the conflict intensity response vector, X represents the driving factor observation variable matrix, Λ represents the causal path coefficient matrix, and ε represents the model error vector;

[0121] S43. For each conflict type, a multivariate logistic regression model was used to model its response probability under different driving factors, and the influencing factors and significance coefficients of structural conflict, functional conflict, and morphological conflict were obtained;

[0122] S44. Based on the joint analysis results of the structural equation model and the logistic regression model, the spatial conflict driving paths, including the dominant factor path and the interaction factor path, are extracted for the parameter setting of the multi-objective optimization function and the game rules.

[0123] The present invention adopts the linkage analysis of the path relationship and impact significance between driving factors by using the structural equation model and the multivariate logistic regression model, realizes the accurate identification of the dominant factors and interaction factors of spatial conflict, and provides a reasonable parameter basis for the construction of the multi-objective optimization function.

[0124] In this embodiment, the S5 specifically includes:

[0125] S51. Set a multi-objective optimization function with grain production, ecological services and economic benefits as optimization targets, and construct the objective function:

[0126]

[0127] in, Boolean variables indicating that the location (i, j) is allocated as cultivated land, forest land, and garden land, respectively, y ij Indicates the unit grain output at position (i, j), esv ij Indicates the ecological service value, eco ij Indicates unit economic benefits, P p (i, j, t) represents the probability of a farmer’s response to the land type at location (i, j) at time t, P Dk(i,j,t) represents the government department D k The probability of preference for location (i, j) and land type t, B ij represents the bidding expression value of the enterprise at position (i, j);

[0128] S52. Constructing the coordination and competition function of government entities:

[0129]

[0130] Among them, C t (i, j) represents the coordination competition degree of land type t at location (i, j), ω k represents the regulatory weight of the k-th government department, UD k (i, j, t) represents the regulatory utility of the department on land type t, α represents the farmer's response weight, UP p (i, j, t) represents the farmer's response utility to land type t, η represents the enterprise's response weight, UE E (i, j, t) represents the enterprise response utility, P t represents the policy priority of land type t;

[0131] S53. Construct the bidding expression function of the enterprise entity:

[0132]

[0133] Among them, B ij represents the competitive value of the enterprise at position (i, j), π 预期 represents the expected profit of the enterprise, T represents the level of technological input, λ represents the ecological penalty coefficient, R eco represents the ecological risk value;

[0134] S54. Construct the response probability function of the farmer entity:

[0135]

[0136] Among them, P p (i, j, t) represents the response probability of the farmer at location (i, j), land type t, and time t, π 果 represents the orchard income, π 耕 represents the income from cultivated land, σ 政策 represents policy sensitivity, ρ represents the risk aversion coefficient, and risk represents the expected value of loss;

[0137] S55. Construct spatial preference probability function of government departments:

[0138]

[0139] Among them, P Dk (i,j,t) represents department Dk The preference probability of land type t at location (i, j), β represents the spatial clustering coefficient, R neigh represents the proportion of similar land use in the neighborhood, γ represents the suitability coefficient, Department D k Suitability score for designated land types;

[0140] S56. Construct game rules to control the allocation priority of the government subject through the coordination competition function, control the land acquisition ranking of the enterprise subject through the bidding expression function, and drive the land conversion behavior of the farmer subject through the response probability function. The response functions of the three types of subjects jointly participate in the calculation weight of the spatial configuration variable.

[0141] S57. Set space transformation constraints:

[0142] The intensity constraint of coordinating competition not lower than a given threshold is used to eliminate low-regulation regions;

[0143] The ratio of forest land to garden land shall not exceed the set upper limit;

[0144] Cross-level land type conversion operations are not allowed for plots between terrain gradient zones.

[0145] This invention establishes a multi-objective optimization function with grain output, ecological services and economic benefits as targets, integrates the behavioral response models of three types of subjects: government, enterprises and farmers, constructs a multi-dimensional coordination and competition mechanism, and improves the configuration process's ability to fit real behaviors and constraints.

[0146] In this embodiment, S6 specifically includes:

[0147] S61. Construct a C-FLUS land use simulation model. Input the terrain gradient partition, initial land use map, land class transition probability matrix, suitability factor layer, and spatial constraint layer into the model. Generate a multi-time series land use simulation raster sequence based on the Markov chain, and output a basic configuration raster map containing the spatial status of cultivated land, garden land, and forest land.

[0148] S62. Construct a MAOOAM multi-agent behavior simulation model. Use the coordination competition function of the government, the bidding expression function of the enterprise, and the response probability function of the farmer as agent behavior strategy inputs. Construct a multi-agent state map based on the two-dimensional location of the plot, the land type, and the behavior utility vector. Simulate the competition, response, and adjustment behavior of the government, enterprises, and farmers for the plot during the land allocation process, and generate a plot behavior response layer containing the agent behavior trajectories.

[0149] S63. Build an R2D3 reinforcement learning optimization model. This model takes the current land state, historical configuration actions, and immediate benefits as inputs, constructs an optimization structure consisting of a state encoding network, a strategy generation network, and a Q-value feedback network. Based on the grain production, ecological service, and economic benefit goals set in the multi-objective optimization function, it learns the impact of different behavior sequences on the achievement of long-term goals and outputs the optimal action allocation strategy for the spatial configuration variables.

[0150] S64. Construct a PLUS-MAS-DRL coupled system, using the spatial configuration trend map output by C-FLUS as the initial state of the simulation, the multi-agent response layer output by MAOOAM as the behavior input tensor, and the configuration strategy output by the R2D3 model as the iterative controller. The spatial configuration variables are updated through the state fusion and behavior weighting mechanism, and the land type of each plot is dynamically assigned and adjusted.

[0151] S65. Set iterative update conditions and convergence criteria in the coupled system, record the change range of the spatial configuration variables in each round, and when the land type change rate of all plots and the target difference of the multi-objective optimization function are lower than the preset threshold, the system is judged to have converged, and the final configuration plan grid map containing the spatial distribution of cultivated land, garden land and woodland is output.

[0152] By coupling the C-FLUS land use simulation model, the MAOOAM multi-agent behavior simulation model and the R2D3 reinforcement learning optimization model, the present invention forms a collaborative system that includes spatial trend prediction, subject behavior simulation and intelligent optimization control, realizing the dynamic evolution of spatial configuration and strategy iteration.

[0153] In this embodiment, the S7 specifically includes:

[0154] S71. Compare the spatial configuration variables output by the PLUS-MAS-DRL coupling system in two consecutive rounds, extract the land type changes of each plot, and calculate the change ratio of the conflict area corresponding to the plot;

[0155] S72, calculating the numerical difference of the multi-objective optimization function in consecutive rounds, and determining whether the preset convergence standard is met;

[0156] S73. When the difference between the conflict area change ratio and the multi-objective optimization function value is lower than a set threshold, a final spatial configuration result layer including the configuration status of cultivated land, garden land, and forest land is generated;

[0157] S74, spatially overlaying the spatial configuration result layer with the terrain gradient partition layer to extract the land type distribution within each slope grade area;

[0158] S75. Construct a conflict risk-intelligent control decision support system, wherein the system has the functions of slope grade identification, land type identification, conflict intensity reading, spatial parameter combination, and control layer construction, and is used to process the structural information contained in the spatial configuration result layer, terrain gradient partition layer, multi-objective optimization function layer, and conflict matrix layer;

[0159] S76. Input the spatial configuration result layer, the terrain gradient partition layer, the multi-objective optimization function layer, and the conflict matrix layer into the conflict risk-intelligent control decision support system to generate a slope grade compensation table and a spatial layout control layer. The slope grade compensation table is used to represent the control parameters for land type configuration in areas of different slope grades, and the spatial layout control layer is used to express land conversion boundaries, control intensity levels, and spatial conversion constraint ranges.

[0160] S77. Export the slope grade compensation table and the spatial layout control layer as a structured policy configuration dataset and load it into the land management and planning module for executing land use configuration, spatial regulation, and management instruction generation operations.

[0161] After the spatial configuration converges, the present invention analyzes and formulates compensation strategies based on terrain gradients, achieving systematic adjustment and structural optimization between the spatial layout and slope conditions of cultivated land, orchards, and woodlands, and providing a quantitative basis for zoning compensation and spatial control for policy formulation.

[0162] Example 1:

[0163] To verify the feasibility of the present invention in practice, the present invention was applied to a typical area in the hilly and mountainous areas of southern Jiangxi Province—the border between Anxi Town and Datangbu Town in Xinfeng County, Ganzhou City, Jiangxi Province. A composite landform sample area of approximately 145 square kilometers was selected as the test area. This area is a typical landform belt with alternating hilly slopes and low mountains, with significant topography and a land use structure dominated by cultivated land, gardens, and woodlands. However, there are problems such as fragmented cultivated land, disorderly expansion of gardens, and poor ecological connectivity of woodlands. Combined with ecological red line constraints and food security requirements, spatial configuration conflicts have not been effectively alleviated for a long time.

[0164] In the application process of the present invention, a spatial database is first constructed through the Gaofen-1 satellite remote sensing image (resolution 2 meters), 1:10000 DEM digital elevation model, land use status data issued by the Department of Natural Resources in 2022, and farmer survey data (covering 520 households), and multi-source data fusion and georeferencing algorithms are used to achieve rasterization and unified processing. Based on this database, four indicators of slope, elevation, undulation and slope heterogeneity are extracted from each spatial unit to construct a terrain factor set and calculate the terrain comprehensive index. According to the natural breakpoint method, the index is threshold graded and divided into three types of terrain gradient zones: river valley plain area, low mountain gentle slope area and medium mountain steep slope area. The areas of each type of zone account for 23.1%, 48.6% and 28.3% respectively.

[0165] Land suitability assessments were conducted for each of the three zones, and a weighted set of land type factors was constructed. Scores were assigned to cultivated land, orchards, and woodlands to generate suitability rating layers. Furthermore, structural and morphological characteristic layers were constructed by integrating landscape pattern indices such as patch density, aggregation, and boundary complexity with a spatial morphological analysis model based on the MCA method. The InVEST model was then used to assess ecosystem service levels, forming a functional characteristic layer. Through spatial overlay, a three-dimensional conflict matrix was constructed, encompassing the three layers of structural conflict, functional conflict, and morphological conflict. Data showed that areas with high conflict intensity were primarily concentrated in the gentle slope transition zone, accounting for 19.4% of the sample area.

[0166] Subsequently, a spatial conflict driver indicator system was constructed, integrating 15 driving factors: natural factors (slope, soil type, precipitation), economic factors (net operating income, land price), social factors (transportation accessibility, population density), policy factors (intensity of land use control, subsidy level), and farmer factors (cultivation willingness, contract term, etc.). A structural equation model was constructed using AMOS software to analyze the path relationships between the driving factors, and a logistic regression model was used to quantify the conflict probability response results. The results showed that the main path for cultivated land conflicts was "increasing slope → decreasing accessibility → decreasing income → decreasing cultivation willingness," while the main path for forest land conflicts was "low ecological priority → insufficient policy support → increased fragmentation."

[0167] Based on conflict path information, a multi-objective optimization function was constructed, targeting grain production, ecological service value, and economic benefits. Three types of subject response functions were established: a government coordination and competitiveness function, an enterprise bidding expression function, and a farmer response probability function. Response behavior expression models were developed using regulation intensity, investment expectations, and economic sensitivity as core parameters. Game rules and spatial conversion constraints were established within this model, including a limit of 42% for the proportion of woodland and gardens, a prohibition on direct land conversion between inter-level regions, and exclusion of low-coordination regions from the optimization sequence.

[0168] During the optimization process, the C-FLUS land use simulation model, integrated with the present invention, was used to predict current configuration trends and generate a three-phase simulated evolutionary layer. A MAOOAM multi-agent behavior simulation model was constructed, using the behavioral functions of the three types of agents as input to generate a spatial response trajectory map for the government, enterprises, and farmers. The R2D3 reinforcement learning optimization model, based on state transitions, revenue feedback, and policy updates, iteratively outputted the optimal action combination for spatial configuration variables. The system performed 200 iterations, with convergence thresholds set at a land type change rate below 1% and an objective function difference below 0.5%, ultimately successfully outputting a stable configuration layer.

[0169] Finally, the allocation results were overlaid with a terrain gradient zoning layer for analysis, revealing a 21.8% increase in the concentration of cultivated land in the valley plains, a 17.5% increase in the proportion of forested land in the mid-mountain steep slopes, and a more coordinated spatial pattern in the low-mountain gentle slopes. Based on these results, a slope compensation table was generated, proposing a 980 yuan per mu (approximately 100 yuan) subsidy for stable forested land allocation areas with slopes greater than 15 degrees, and special support for water conservancy facilities for newly added cultivated land on low slopes. A spatial layout control layer was also generated to serve as the control boundary for county-level land consolidation planning and village-level land use compilation. The entire allocation result was evaluated by an expert panel from the Ganzhou Municipal Natural Resources and Planning Bureau, who concluded that the spatial structure was reasonable, the optimization process was controllable, and suitable for pilot promotion.

[0170] This example fully verifies the adaptability and practicality of the present invention in hilly and complex terrain areas, and reflects the system capability of the method in integrating multiple models, multiple objectives, and multiple subject elements. It breaks through the technical bottleneck of traditional land allocation methods in the linkage between terrain response and subject behavior, demonstrates excellent multi-dimensional coordination and optimization capabilities and operability, and provides effective technical support for mountainous land integration, ecological compensation, etc.

[0171] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A method for optimizing land space allocation under terrain gradient constraints, characterized in that: The steps include: S1. Collect multi-source geographic data and build a spatial database; S2, extracting slope, elevation, relief and aspect heterogeneity from the spatial database, calculating the terrain comprehensive index, and dividing the terrain gradient into zones; S3. Based on the spatial database and terrain gradient zoning, the suitability evaluation method is used to calculate the suitability levels of cultivated land, garden land, and forest land. Morphological spatial pattern analysis and landscape pattern index are used to identify structural and morphological characteristics, and a conflict matrix is constructed. S4. Based on the spatial database and conflict matrix, a driving indicator system was constructed, and the driving paths of spatial conflicts were analyzed using structural equation models and logistic regression models. S5. Based on the conflict matrix and driving path, a multi-objective optimization function with grain production, ecological services, and economic benefits as targets is set, and the game rules and spatial transformation constraints of the government, enterprises, and farmers are constructed; S6: Integrate the C-FLUS land use simulation model, the MAOOAM multi-agent behavior simulation model, and the R2D3 reinforcement learning optimization model to build a PLUS-MAS-DRL coupled system to perform spatial configuration simulation and dynamic optimization. S7. Based on the optimization results, when the conflict area change rate and the multi-objective optimization function converge to meet the set threshold, the optimal land space allocation plan that meets the terrain gradient constraint is output, and the slope zoning compensation and spatial layout control strategy is generated through the decision support system.

2. The method for optimizing land space allocation under terrain gradient constraints according to claim 1, characterized in that: The multi-source geographic data includes remote sensing images, digital elevation models, land use status, farmer surveys and policy data.

3. The method for optimizing land space allocation under terrain gradient constraints according to claim 1, characterized in that: The conflict matrix is a three-dimensional conflict matrix including structural conflict, functional conflict and morphological conflict.

4. The method for optimizing land space allocation under terrain gradient constraints according to claim 1, characterized in that: The driving indicator system includes natural factors, economic factors, social factors, policy factors and farmer factors.

5. The method for optimizing land space allocation under terrain gradient constraints according to claim 1, characterized in that: The S2 specifically includes: S21. Extract the slope, elevation, relief and aspect heterogeneity factors in the study area from the spatial database and construct a terrain factor set; S22. Calculate the terrain comprehensive index of each spatial unit based on the terrain factor set: TCI = S × E × R × D; Among them, TCI is the terrain comprehensive index, S is the slope, E is the elevation, R is the relief, and D is the slope heterogeneity; S23. The terrain comprehensive index is classified by the natural breakpoint method according to the set threshold value, and divided into three terrain gradient zones: river valley plain area, low mountain and hilly area, and medium mountain steep slope area; S24. Generate a partition code for each terrain gradient partition as a basic input parameter for land suitability evaluation, spatial conflict identification and spatial optimization simulation.

6. The method for optimizing land space allocation under terrain gradient constraints according to claim 1, characterized in that: The S3 specifically includes: S31. Based on the spatial database and terrain gradient zoning, use the land suitability evaluation method to classify the suitability of cultivated land, garden land and forest land, and generate a suitability grade map; S32. Based on the suitability level layer, calculate the landscape pattern indices such as patch density, boundary density and aggregation, and construct the structural characteristic layer; S33, using the morphological spatial pattern analysis model to identify spatial boundary morphology and connectivity features, and constructing a morphological feature layer; S34. Based on the spatial database and policy parameters, the ecosystem service assessment model is called to construct the functional characteristic layer; S35. Spatially superimpose the suitability level layer, structural feature layer, morphological feature layer, and functional feature layer to construct a conflict matrix. The conflict matrix contains three types of layer channels: structural conflict, functional conflict, and morphological conflict. The pixel value in each channel represents the intensity level of the corresponding conflict.

7. The method for optimizing land space allocation under terrain gradient constraints according to claim 1, characterized in that: The S4 specifically includes: S41. Based on the spatial database and conflict matrix, a driving indicator system is constructed. The driving factors include natural factors, economic factors, social factors, policy factors, and farmer factors. Natural factors include slope, soil type, and precipitation; economic factors include land management income and land market price; social factors include transportation accessibility and population density; policy factors include land use restriction level and subsidy policy intensity; and farmer factors include farmer land size and farming willingness. S42. Use structural equation modeling to analyze the causal path relationship between driving factors: Y=ΛX+ε; Where Y represents the conflict intensity response vector, X represents the driving factor observation variable matrix, Λ represents the causal path coefficient matrix, and ε represents the model error vector; S43. For each conflict type, a multivariate logistic regression model was used to model the response probability under different driving factors, and the influencing factors and significance coefficients of structural conflict, functional conflict, and morphological conflict were obtained; S44. Based on the joint analysis results of the structural equation model and the logistic regression model, the spatial conflict driving paths, including the dominant factor path and the interaction factor path, are extracted for the parameter setting of the multi-objective optimization function and the game rules.

8. The method for optimizing land space allocation under terrain gradient constraints according to claim 1, characterized in that: The S5 specifically includes: S51. Set a multi-objective optimization function with grain production, ecological services and economic benefits as optimization targets, and construct the objective function: in, Boolean variables indicating that the location (i, j) is allocated as cultivated land, forest land, and garden land, respectively, y ij Indicates the unit grain output at position (i, j), esv ij Indicates the ecological service value, eco ij Indicates unit economic benefits, P p (i, j, t) represents the probability of a farmer’s response to the land type at location (i, j) at time t, Indicates government department D k The probability of preference for location (i, j) and land type t, B ij represents the bidding expression value of the enterprise at position (i, j); S52. Constructing the coordination and competition function of government entities: Among them, C t (i, j) represents the coordination competition degree of land type t at location (i, j), ω k represents the regulatory weight of the k-th government department, UD k (i, j, t) represents the regulatory utility of the department on land type t, α represents the farmer's response weight, UP p (i, j, t) represents the farmer's response utility to land type t, η represents the enterprise's response weight, UE E (i, j, t) represents the enterprise response utility, P t represents the policy priority of land type t; S53. Construct the bidding expression function of the enterprise entity: Among them, B ij represents the competitive value of the enterprise at position (i, j), π 预期 represents the expected profit of the enterprise, T represents the level of technological input, λ represents the ecological penalty coefficient, R eco represents the ecological risk value; S54. Construct the response probability function of the farmer entity: Among them, P p (i, j, t) represents the response probability of the farmer at location (i, j), land type t, and time t, π 果 represents the orchard income, π 耕 represents the income from cultivated land, σ 政策 represents policy sensitivity, ρ represents the risk aversion coefficient, and risk represents the expected value of loss; S55. Construct spatial preference probability function of government departments: in, Indicates department D k The preference probability of land type t at location (i, j), β represents the spatial clustering coefficient, R neigh represents the proportion of similar land use in the neighborhood, γ represents the suitability coefficient, Department D k Suitability score for designated land types; S56. Construct game rules to control the allocation priority of the government subject through the coordination competition function, control the land acquisition ranking of the enterprise subject through the bidding expression function, and drive the land conversion behavior of the farmer subject through the response probability function. The response functions of the three types of subjects jointly participate in the calculation weight of the spatial configuration variable. S57. Set space transformation constraints: The intensity constraint of coordinating competition not lower than a given threshold is used to eliminate low-regulation regions; The ratio of forest land to garden land shall not exceed the set upper limit; Cross-level land type conversion operations are not allowed for plots between terrain gradient zones.

9. The method for optimizing land space allocation under terrain gradient constraints according to claim 1, characterized in that: The S6 specifically includes: S61. Construct a C-FLUS land use simulation model. Input the terrain gradient partition, initial land use map, land class transition probability matrix, suitability factor layer, and spatial constraint layer into the model. Generate a multi-time series land use simulation raster sequence based on the Markov chain, and output a basic configuration raster map containing the spatial status of cultivated land, garden land, and forest land. S62. Construct a MAOOAM multi-agent behavior simulation model. Use the coordination competition function of the government, the bidding expression function of the enterprise, and the response probability function of the farmer as agent behavior strategy inputs. Construct a multi-agent state map based on the two-dimensional location of the plot, the land type, and the behavior utility vector. Simulate the competition, response, and adjustment behavior of the government, enterprises, and farmers for the plot during the land allocation process, and generate a plot behavior response layer containing the agent behavior trajectories. S63. Build an R2D3 reinforcement learning optimization model. This model takes the current land state, historical configuration actions, and immediate benefits as inputs, constructs an optimization structure consisting of a state encoding network, a strategy generation network, and a Q-value feedback network. Based on the grain production, ecological service, and economic benefit goals set in the multi-objective optimization function, it learns the impact of different behavior sequences on the achievement of long-term goals and outputs the optimal action allocation strategy for the spatial configuration variables. S64. Construct a PLUS-MAS-DRL coupled system, using the spatial configuration trend map output by C-FLUS as the initial state of the simulation, the multi-agent response layer output by MAOOAM as the behavior input tensor, and the configuration strategy output by the R2D3 model as the iterative controller. The spatial configuration variables are updated through the state fusion and behavior weighting mechanism, and the land type of each plot is dynamically assigned and adjusted. S65. Set iterative update conditions and convergence criteria in the coupled system, record the change range of the spatial configuration variables in each round, and when the land type change rate of all plots and the target difference of the multi-objective optimization function are lower than the preset threshold, the system is judged to have converged, and the final configuration plan grid map containing the spatial distribution of cultivated land, garden land and woodland is output.

10. The method for optimizing land space allocation under terrain gradient constraints according to claim 1, characterized in that: The S7 specifically includes: S71. Compare the spatial configuration variables output by the PLUS-MAS-DRL coupling system in two consecutive rounds, extract the land type changes of each plot, and calculate the change ratio of the conflict area corresponding to the plot; S72, calculating the numerical difference of the multi-objective optimization function in consecutive rounds, and determining whether the preset convergence standard is met; S73. When the difference between the conflict area change ratio and the multi-objective optimization function value is lower than a set threshold, a final spatial configuration result layer including the configuration status of cultivated land, garden land, and forest land is generated; S74, spatially overlaying the spatial configuration result layer with the terrain gradient partition layer to extract the land type distribution within each slope grade area; S75. Construct a conflict risk-intelligent control decision support system, wherein the system has the functions of slope grade identification, land type identification, conflict intensity reading, spatial parameter combination, and control layer construction, and is used to process the structural information contained in the spatial configuration result layer, terrain gradient partition layer, multi-objective optimization function layer, and conflict matrix layer; S76. Input the spatial configuration result layer, the terrain gradient partition layer, the multi-objective optimization function layer, and the conflict matrix layer into the conflict risk-intelligent control decision support system to generate a slope grade compensation table and a spatial layout control layer. The slope grade compensation table is used to represent the control parameters for land type configuration in areas of different slope grades, and the spatial layout control layer is used to express land conversion boundaries, control intensity levels, and spatial conversion constraint ranges. S77. Export the slope grade compensation table and the spatial layout control layer as a structured policy configuration dataset and load it into the land management and planning module for executing land use configuration, spatial regulation, and management instruction generation operations.

Images