A Multi-Source Data-Driven Intelligent Optimization Method for Territorial Spatial Planning Schemes

By combining multi-source data credibility assessment and rule knowledge graph hard constraints with multi-objective evolutionary algorithms, the problem of data quality assessment and legality of control rules in territorial spatial planning was solved, realizing dynamic optimization and monitoring feedback in multiple scenarios, and improving the robustness and practical applicability of the solution.

CN123117451APending Publication Date: 2026-11-17SHANDONG HENGBO REAL ESTATE APPRAISAL SURVEY PLANNING & DESIGN CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611345450.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-01
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing land spatial planning schemes lack quality assessment and conflict resolution of multi-source heterogeneous data, the soft handling of planning control rules leads to insufficient legitimacy, the single dimension of objectives does not take into account future uncertainties, and there is a lack of dynamic update mechanism.

Method used

By conducting multi-source data credibility assessment and conflict arbitration, a rule knowledge graph is constructed as a hard constraint. A multi-objective evolutionary algorithm with constraint dominance is used to perform dynamic inference of multiple scenarios, and incremental re-optimization is carried out through monitoring and feedback during implementation.

Benefits of technology

It improves the accuracy of multi-source data fusion, ensures the legality and compliance of optimization results, enhances the robustness and timeliness of the solution throughout its lifecycle, and outputs a comprehensive benefit optimization solution that meets actual planning needs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123117451A_ABST
    Figure CN123117451A_ABST
Patent Text Reader

Abstract

This invention belongs to the field of planning scheme optimization technology based on knowledge graphs and machine learning. Specifically, it relates to an intelligent optimization method for land spatial planning schemes driven by multi-source data. The scheme ensures the reliability of input information from the data source by assessing the credibility of multi-source heterogeneous data and arbitrating conflicts. Secondly, by explicitly modeling planning control rules into a knowledge graph and transforming them into hard constraints, and embedding the hard constraints into the dominance criteria of a multi-objective evolutionary algorithm, it ensures that the optimization results do not violate rigid control rules. At the same time, by constructing multi-objective functions for planning operations such as farmland protection, land consolidation, and land reclamation, it overcomes the problem of single objective dimensions in existing technologies. Finally, through multi-scenario coupled dynamic deduction and incremental re-optimization based on implementation monitoring feedback, a dynamic closed loop of "planning-implementation-monitoring-re-optimization" is formed, enhancing the robustness of the scheme in the face of future uncertainties and its timeliness throughout the entire life cycle.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of planning scheme optimization technology based on knowledge graphs and machine learning, specifically involving a multi-source data-driven intelligent optimization method for land spatial planning schemes. Background Technology

[0002] The formulation and optimization of national spatial planning schemes is a complex spatial decision-making problem involving multiple objectives, constraints, and scales. With the popularization of digital technology, some machine learning-based optimization methods have emerged in existing schemes, such as: methods based on geographic information system (GIS) overlay analysis and multi-criteria evaluation, spatial simulation methods based on cellular automata (CA), and spatial optimization methods based on multi-objective evolutionary algorithms such as genetic algorithms and particle swarm optimization.

[0003] However, while existing schemes have achieved a certain degree of shift from manual experience-based judgment to quantitative calculation in planning, the following problems still exist: The multi-source heterogeneous data relied upon in existing schemes are often directly superimposed, lacking quantitative assessment of the quality of each data source and effective resolution of data conflicts. Low-quality or contradictory data can easily interfere with the optimization results. In existing schemes, planning control rules such as the "three zones and three lines," permanent basic farmland protection, and ecological protection red lines are often treated as penalty terms in the fitness function, making it difficult to ensure that the optimization results do not violate rigid control rules, resulting in insufficient legality and feasibility of the schemes. Existing schemes have relatively singular optimization objectives, generally limited to a few objectives such as "economy + ecology," failing to fully integrate with actual planning operations such as farmland protection, land consolidation, and land reclamation. Existing schemes mostly adopt single-scenario static optimization, failing to consider the impact of future uncertainties such as climate change, population migration, and policy adjustments on the robustness of the schemes, and lack a monitoring and feedback mechanism after planning implementation, making it difficult to achieve dynamic rolling updates of the schemes. Summary of the Invention

[0004] This invention addresses the technical problems existing in the prior art by providing a multi-source data-driven intelligent optimization method for land spatial planning schemes, effectively solving the problems existing in the prior art.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A multi-source data-driven intelligent optimization method for land spatial planning schemes includes: Acquire multi-source heterogeneous data related to land spatial planning and preprocess them to obtain standardized data; based on a pre-built data credibility assessment model, perform weighted fusion and conflict arbitration on the standardized data to obtain a unified spatial dataset with credibility labels; A rule knowledge graph is constructed based on planning and control rules, control rules are extracted based on the rule knowledge graph, and the extracted control rules are formalized into hard constraints; wherein, the hard constraints are used to limit the feasible solution space for optimization. Within the feasible solution space constrained by hard constraints, a multi-objective evolutionary algorithm with constraint dominance is used to optimize the use of spatial units to obtain a Pareto candidate solution set. The candidate solution set is subjected to multi-scenario coupled dynamic simulation. The robustness of each candidate solution under different scenarios is evaluated based on the simulation results. Recommended solutions are selected based on the robustness evaluation results. During the implementation of the recommended scheme, monitoring data reflecting the actual land use status is collected, and the monitoring data is compared with the recommended scheme. When the deviation exceeds a preset threshold, incremental re-optimization is performed on the local area where the deviation occurred, and a dynamically updated optimized scheme is output.

[0006] Furthermore, in the constrained multi-objective evolutionary algorithm, the purpose of the spatial unit constitutes the decision variable; wherein, the cumulative violation of all hard constraints by the decision variable is expressed as follows: in, For configuration The cumulative number of violations of all hard constraints, hour, If a solution is feasible, in the comparison of the dominance of candidate solutions, the feasible solution dominates the infeasible solution; if both solutions are feasible, the solution with an objective function value that is not inferior to the other and at least one objective function value that is strictly superior is the dominant solution; if both solutions are infeasible, the solution with the smaller cumulative violation is the dominant solution.

[0007] Furthermore, the objective function includes targets for farmland contiguousness, matching degree of land consolidation and reclamation, and suitability for land reclamation, wherein: Target for contiguous arable land : Increase / decrease linkage matching target : Reclamation suitability targets : in, This represents the total number of cultivated land patches. For the first Area of ​​each cultivated land patch For the demolition area unit collection, To build a new district unit collection, , These are the areas of the corresponding units; for Unit and The distance between units; This is the distance attenuation coefficient; Spatial matching coefficients; This is a collection of units to be reclaimed / renovated. For unit Reclamation suitability score For unit The area.

[0008] Furthermore, the hard constraints include, but are not limited to, the three-zone, three-line control rules, the permanent basic farmland protection rules, the ecological protection red line rules, the urban development boundary rules, the land use control rules, and the slope and topsoil thickness rules; the rule knowledge graph is composed of entity-relationship triples, the triples include head entity, relation, and tail entity, the head entity and tail entity include land parcel, land use, control boundary, and planning indicators, and the relations include located, belong to, prohibited, and restricted by.

[0009] Furthermore, the data credibility assessment model calculates the overall credibility of each data source through five dimensions: completeness, accuracy, timeliness, consistency, and authority; the conflict arbitration specifically involves: only including data sources with an overall credibility not lower than a preset threshold into the fusion set, and performing a weighted average based on the weights of each data source after renormalization within the fusion set.

[0010] Furthermore, when the deviation exceeds a preset threshold, incremental re-optimization is performed on the local area where the deviation occurred, specifically including: The deviation is measured by whether the actual land use status of the spatial unit is consistent with the recommended scheme. When any spatial unit is inconsistent or the proportion of inconsistent spatial units in the entire region exceeds a preset average deviation threshold, the incremental re-optimization is triggered. The optimization process is re-executed for the local region consisting of inconsistent spatial units and their neighboring spatial units.

[0011] Furthermore, in the aforementioned multi-scenario coupled dynamic deduction, the state transition probability of the spatial unit is specifically represented as follows: in, for Full-area configuration at all times For unit exist Current state; For suitability items; For neighborhood influence; Context-driven items; , , These are the weighting coefficients. For normalization function, For the intended use type.

[0012] An electronic device includes a memory, a processor, and a computer program stored in the memory and running on it. When the processor executes the program, it implements the aforementioned intelligent optimization method for multi-source data-driven land spatial planning schemes.

[0013] A non-transitory computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned intelligent optimization method for land spatial planning schemes driven by multi-source data.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: (1) The present invention provides a method for intelligent optimization of land and space planning schemes driven by multi-source data. The scheme ensures the reliability of input information from the source of data and improves the accuracy of multi-source data fusion by assessing the credibility of multi-source heterogeneous data and arbitrating conflicts. Secondly, by explicitly modeling planning control rules into knowledge graphs and transforming them into hard constraints, and embedding the hard constraints into the dominance criteria of multi-objective evolutionary algorithms, the optimization results are ensured to not violate rigid control rules and to be legal, compliant and implementable. At the same time, by constructing multi-objective functions for planning businesses such as farmland protection, land consolidation and land reclamation, the problem of single objective dimension in existing technologies is overcome to a certain extent. Finally, by multi-scenario coupled dynamic deduction and incremental re-optimization based on implementation monitoring feedback, a dynamic closed loop of "compilation-implementation-monitoring-re-optimization" is formed, which enhances the robustness of the scheme in dealing with future uncertainties and the timeliness of the whole life cycle.

[0015] (2) The proposed scheme calculates the comprehensive credibility of each data source from five dimensions: completeness, accuracy, timeliness, consistency, and authority. In the conflict arbitration, only high-credibility data sources with a value not lower than the preset threshold are retained for re-normalization and weighted averaging. This effectively resolves the conflicts between heterogeneous data sources and avoids low-quality data from interfering with the optimization results to a certain extent. At the same time, by explicitly constructing the control rules such as the three zones and three lines, permanent basic farmland protection, and ecological protection red lines into a rule knowledge graph and formalizing them into hard constraints that limit the feasible solution space, and by using the constraint domination mechanism to make the feasible solution always dominate the infeasible solution, the scheme avoids the generation of schemes that violate rigid control rules from the source, making the output scheme legal, compliant, and implementable.

[0016] (3) The proposed scheme constructs multiple objective functions such as farmland contiguousness, increase-decrease linkage matching degree, and reclamation suitability, so that the optimization objectives are directly aligned with actual planning operations such as farmland protection, increase-decrease linkage, and land reclamation. At the same time, it uses a constrained multi-objective evolutionary algorithm to obtain the Pareto candidate scheme set in the feasible solution space. Compared with the existing schemes that only adopt a few objectives such as "economy + ecology", it can output a scheme that is more in line with the needs of national land space planning and has better comprehensive benefits.

[0017] (4) The proposed scheme constructs a set of multiple scenarios such as climate change, population migration, and policy adjustment, and couples cellular automata or system dynamics models to dynamically deduce candidate schemes. At the same time, it uses the minimum or mean of the comprehensive benefits of each scenario minus the risk coefficient multiplied by the standard deviation as a robustness indicator to screen recommended schemes, which effectively overcomes the problem of insufficient single-scenario static optimization in existing technologies. Meanwhile, during the implementation of the recommended scheme, the actual land use status is continuously monitored. When the deviation exceeds the preset threshold, only the local area composed of inconsistent units and their neighbors is incrementally re-optimized, realizing the transformation of the planning scheme from one-time static optimization to dynamic rolling optimization of "compilation-implementation-monitoring-re-optimization". Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below: Figure 1 This is a flowchart illustrating a multi-source data-driven intelligent optimization method for land spatial planning, as described in this embodiment. Figure 2 This is a flowchart illustrating the construction process of the unified spatial dataset described in the embodiments. Figure 3 This is a flowchart of the recommended solution selection process described in the embodiment; Figure 4 This is a flowchart of the local incremental re-optimization process described in the embodiment; Figure 5 This is a schematic diagram of an electronic device structure as described in the embodiment. Detailed Implementation

[0019] To better understand the above-mentioned objectives, features and advantages of the present invention, the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0020] Numerous specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways than those described herein, and therefore the invention is not limited to the specific embodiments disclosed in the following specification.

[0021] Example 1: The following detailed description, in conjunction with the accompanying drawings, illustrates a multi-source data-driven intelligent optimization method for land spatial planning schemes as described in Embodiment 1.

[0022] In one or more embodiments, such as Figure 1 As shown, this embodiment provides a multi-source data-driven intelligent optimization method for land spatial planning schemes, which specifically includes the following processing steps: Step S100: Acquire multi-source heterogeneous data related to land spatial planning and preprocess it to obtain standardized data; based on the pre-built data credibility assessment model, perform weighted fusion and conflict arbitration on the standardized data to obtain a unified spatial dataset with credibility labels; Step S200: Construct a rule knowledge graph based on planning and control rules, extract control rules based on the rule knowledge graph, and formalize the extracted control rules into hard constraints; wherein, the hard constraints are used to limit the feasible solution space for optimization. Step S300: Within the feasible solution space constrained by hard constraints, optimize the use of spatial units using a multi-objective evolutionary algorithm with constraint dominance to obtain a Pareto candidate solution set; Step S400: Perform multi-scenario coupled dynamic simulation on the candidate solution set, evaluate the robustness of each candidate solution under different scenarios based on the simulation results, and select recommended solutions based on the robustness evaluation results; Step S500: During the implementation of the recommended scheme, monitoring data reflecting the actual land use status is collected, and the monitoring data is compared with the recommended scheme. When the deviation exceeds a preset threshold, incremental re-optimization is performed on the local area where the deviation occurs, and the dynamically updated optimized scheme is output.

[0023] In one or more embodiments, step S100 involves: acquiring multi-source heterogeneous data related to land spatial planning and preprocessing it to obtain standardized data; based on a pre-built data credibility assessment model, performing weighted fusion and conflict arbitration on the standardized data to obtain a unified spatial dataset with credibility labels; such as Figure 2 As shown, the specific processing procedure is as follows: Step S101: Multi-source data acquisition The multi-source heterogeneous data collected by the scheme described in this embodiment includes, but is not limited to: data from the Third National Land Survey, high-resolution remote sensing imagery, point-of-interest (POI) data, mobile phone signaling data, socio-economic statistical data, meteorological data, soil data, digital elevation models (DEM), geological hazard risk data, historical planning texts and maps, and planning control data (three zones and three lines, permanent basic farmland, ecological protection red lines, and urban development boundaries), totaling M types of data sources, constituting a data set. .

[0024] In the scheme described in this embodiment, the CGCS2000 coordinate system is uniformly adopted, and the side length of the reference grid is taken as... Collection of usage types (The number of its types is K=6).

[0025] Step S102: Data Preprocessing (1) Spatiotemporal alignment Each data source is resampled to a reference raster (the side length of the reference raster is a preset value, which is taken in this embodiment). ), for data source Falling on the reference unit The original pixels inside Aggregate by area intersection weight: Among them, the area overlap weight is: in, For data source No. The attribute values ​​of each original pixel. Area overlap weight (i.e., pixel) With reference unit The overlapping area occupies the pixel (ratio of area) For resampled cells In the data source The attribute values ​​below.

[0026] (2) Semantic normalization Constructing a unified semantic coding dictionary Mapping different definitions of the same feature / attribute from various data sources to a unified code. : in, For data source The original attribute codes (e.g., classification codes for "paddy fields", "arable land", "basic farmland"); This is the semantic mapping rule function for the data source; To unify semantic encoding.

[0027] Step S103: Data credibility assessment and weighted fusion For each data source The overall credibility is calculated by weighting scores based on five dimensions: completeness, accuracy, timeliness, consistency, and authority. in, , , , , The weight coefficients for the five dimensions are respectively, satisfying... ; , , , , Data sources In the scoring across five dimensions, the weighting coefficients are as follows: , , , =0.15、 =0.15.

[0028] When multiple data sources are in the same unit When conflicting values ​​are given for the same attribute, the scheme described in this embodiment arbitrates based on a weighted average of credibility: in, For unit The data source set participating in conflict arbitration Number the data source. The minimum credibility threshold (set in the scheme described in this embodiment) (Data sources with values ​​below this threshold are excluded). For data source In unit The attribute values ​​given above, For data source In the set The weights after internal renormalization ,satisfy , For data source The overall credibility is determined, and then a unified spatial dataset with credibility labels is output.

[0029] It should be noted that, in response to the problem that most existing solutions directly superimpose multi-source data and lack data quality assessment, the solution described in this embodiment explicitly establishes a five-dimensional credibility assessment model and a conflict arbitration mechanism to achieve "credibility-based fusion and conflict resolution," thus ensuring data reliability from the input source.

[0030] In one or more embodiments, step S200 involves: constructing a rule knowledge graph based on planning and control rules, extracting control rules based on the rule knowledge graph, and formalizing the extracted control rules into hard constraints; wherein the hard constraints are used to limit the feasible solution space for optimization; specifically, the following processing steps are performed: Step S201: Construction of Triples and Knowledge Graph Define entity-relationship triples in the planning and control domain: in, The head entity (including but not limited to: land parcel, land use, control boundary, planning indicators); Relationships (including but not limited to: "located in", "belongs to", "prohibited", "restricted by"); As the tail entity, all triples together constitute the knowledge graph of planning and control rules. .

[0031] Step S202: Formalize the rules into constraint functions The natural language control rules, such as the "three zones and three lines," permanent basic farmland protection, ecological protection red lines, urban development boundaries, land use control, slope, and topsoil thickness, are formalized into computable hard constraints: in, For the first Hard constraints on the overall area usage configuration vector Violation of the metric function This indicates satisfaction. Indicates violation; It is a set with hard constraints; Its feasible solution space is defined as: in, For this set of hard constraints, it should be noted that preferred or guiding requirements such as farmland contiguousness, matching degree of land consolidation with land increase / decrease linkage, and suitability for reclamation are not modeled as constraints, but are directly modeled as optimization objectives. In this embodiment, only rigid control rules such as the three zones and three lines and permanent basic farmland are modeled as hard constraints.

[0032] It should be noted that existing solutions typically treat control rules as "penalties" and process them in a soft manner, which may result in outputting solutions that violate rigid control rules. The solution described in this embodiment explicitly models the rules as a knowledge graph and transforms them into hard constraints, forming the boundary of the feasible domain, thus ensuring the legality, compliance and feasibility of the solution from the source.

[0033] In one or more embodiments, step S300: Within the feasible solution space defined by hard constraints, a multi-objective evolutionary algorithm with constraint dominance is used to optimize the use of spatial units to obtain a Pareto candidate solution set; specifically, the following processing procedure is performed: Step S301: Optimize Problem Modeling Based on the usage status of each spatial unit Composition of full-area usage configuration vector For decision variables (where, demolition area) B , new district Cnew reclamation area All are configured by the aforementioned purpose vectors (Confirmed), in the feasible region Internal multi-objective optimal solution: In the scheme described in this embodiment, the objective function is defined as follows, taking into account the planning business: by For example, it includes the contiguousness of cultivated land, the matching degree of land consolidation and land reclamation, and the suitability for reclamation, among which: Target for contiguous arable land (i.e., the target for farmland protection and the contiguousness of high-standard farmland): in, This represents the total number of cultivated land patches. For the first Area of ​​each cultivated land patch The higher the value and the more concentrated and contiguous the area, the more conducive it is to the construction of high-standard farmland.

[0034] Increase / decrease linkage matching target (i.e., the spatial matching target of the project area for increasing or decreasing linkage): in, This is a set of units in the demolition zone, which is composed of the overall area usage configuration vector. The units are located within candidate land parcels such as inefficient hollow villages and abandoned industrial and mining areas, and are currently under construction. The new district unit set is composed of the overall area usage configuration vector. It consists of units located in areas with suitable development conditions and currently planned for construction. , These are the areas of the corresponding units; for Unit and The distance between units; This is the distance attenuation coefficient; Spatial matching coefficients; The larger the value, the better the match between demolition and new construction.

[0035] Reclamation suitability objective function (i.e., land reclamation suitability targets): in, This is a set of proposed reclamation / remediation units, which is determined by the overall area land use allocation vector. The units located within candidate patches such as abandoned mines and damaged land and selected for reclamation are composed of these units. For unit Reclamation suitability score (specifically, a quantitative score can be obtained through expert evaluation based on soil conditions, pollution level, accessibility, and slope; the score is determined by a set of values). ); For unit The area; The higher the value, the higher the overall suitability for reclamation.

[0036] Step S302: Constraint Domination Solution The scheme described in this embodiment uses a constrained, dominated multi-objective evolutionary algorithm (NSGA-II / NSGA-III) to solve the problem. First, the total number of constraint violations is defined: in, For configuration The cumulative number of violations of all hard constraints, hour, This is a feasible solution.

[0037] Constraint and Domination Rules ( It is true if any of the following conditions are met: (1) ; (2) ; (3) ; in, For logical AND, express Superior in terms of dominance , express Not inferior to in any objective Furthermore, at least one objective is strictly better. The above rule means that feasible solutions always dominate infeasible solutions, and hard constraints are embedded as dominance criteria rather than external penalty terms. The Pareto optimal solution set is output through iteration. .

[0038] It should be noted that the scheme described in this embodiment couples "knowledge graph hard constraints" with "multi-objective evolutionary algorithm," embedding the hard constraints as domination criteria through constraint domination to ensure that all output solutions fall within the feasible region. Furthermore, unlike the narrow setting of the two objectives of "economy + ecology" in existing technologies, the objective function of the scheme described in this embodiment is directly aligned with planning operations such as farmland protection, land consolidation, and land reclamation.

[0039] In one or more embodiments, step S400: Perform multi-scenario coupled dynamic simulation on the candidate solution set, evaluate the robustness of each candidate solution under different scenarios based on the simulation results, and select recommended solutions based on the robustness evaluation results; such as... Figure 3 As shown, the specific processing procedure is as follows: Step S401: Scenario Set and State Transition Construct a multi-scenario set including climate change, population migration, policy adjustments, and the implementation of major projects, and couple cellular automata (CA) / system dynamics (SD) models for deduction, unit The state transition probability at time t is specifically represented as follows: in, for Full-area configuration at all times For unit exist Current state; For suitability items; For neighborhood influence; Context-driven items; , , These are the weighting coefficients. For normalization function, For the intended use type.

[0040] Step S402: Robustness Assessment of the Solution For candidate solutions In different scenarios The following deduction is performed to calculate robustness indicators: Alternatively, the mean-standard deviation form can be used: in, For the plan In the context The comprehensive benefit value below It is a collection of multiple scenarios; , The mean and standard deviation of the benefits for each scenario; This is the risk aversion coefficient. A higher value indicates greater robustness, as indicated by... The recommended options are filtered by sorting the values.

[0041] It should be noted that, in view of the fact that existing solutions mostly adopt single-scenario static optimization and do not consider future uncertainties, the solution described in this embodiment introduces multi-scenario coupled extrapolation and robustness screening, which effectively improves the solution's ability to cope with future changes.

[0042] In one or more embodiments, step S500: During the implementation of the recommended scheme, monitoring data reflecting the actual land use status is collected, the monitoring data is compared with the recommended scheme, and when the deviation exceeds a preset threshold, incremental re-optimization is performed on the local area where the deviation occurred, and a dynamically updated optimized scheme is output; for example... Figure 4 As shown, the specific processing procedure is as follows: Step S501: Implement deviation monitoring and trigger judgment During the planning and implementation process, monitoring data is continuously collected, and the computing unit... Implementation deviations: in, For unit The actual land use status at time t (which can be obtained through remote sensing change detection and land use status survey); Status of planned use; This is a measure of deviation.

[0043] In the scheme described in this embodiment, the specific conditions for triggering incremental re-optimization are as follows: in, For logical "OR", For unit Deviation measurement Representation unit The actual use is inconsistent with the planned use. The average deviation threshold for the entire region; This represents the total number of units.

[0044] Step S502: Local Incremental Re-optimization After triggering, only the changed area is re-optimized to determine the local range: in, For the local re-optimization region, For the neighborhood expansion set (used to ensure boundary continuity); the above expression means: only when The process involves re-executing steps S100 to S300 for local optimization to obtain an updated solution, thus forming a closed loop of "planning, implementation, monitoring, and re-optimization".

[0045] It should be noted that, in response to the problem that existing solutions typically employ one-time static optimization, the solution described in this embodiment achieves dynamic rolling optimization throughout the entire lifecycle of the solution by monitoring deviation triggers and local incremental re-optimization.

[0046] Example 2: like Figure 5 As shown, this embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on it. When the processor executes the program, it implements the intelligent optimization method for land spatial planning scheme driven by multi-source data as described in Embodiment 1.

[0047] Example 3: This embodiment provides a non-transitory computer-readable storage medium storing a computer program that, when executed by a processor, implements the intelligent optimization method for land spatial planning schemes driven by multi-source data as described in Embodiment 1.

[0048] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications or equivalent changes made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A multi-source data-driven intelligent optimization method for land spatial planning schemes, characterized in that, include: Acquire multi-source heterogeneous data related to land spatial planning and preprocess them to obtain standardized data; Based on a pre-built data credibility assessment model, the standardized data is weighted, fused, and conflict-arbitrated to obtain a unified spatial dataset with credibility labels. A rule knowledge graph is constructed based on planning and control rules, control rules are extracted based on the rule knowledge graph, and the extracted control rules are formalized into hard constraints; wherein, the hard constraints are used to limit the feasible solution space for optimization. Within the feasible solution space constrained by hard constraints, a multi-objective evolutionary algorithm with constraint dominance is used to optimize the use of spatial units to obtain a Pareto candidate solution set. The candidate solution set is subjected to multi-scenario coupled dynamic simulation. The robustness of each candidate solution under different scenarios is evaluated based on the simulation results. Recommended solutions are selected based on the robustness evaluation results. During the implementation of the recommended scheme, monitoring data reflecting the actual land use status is collected, and the monitoring data is compared with the recommended scheme. When the deviation exceeds a preset threshold, incremental re-optimization is performed on the local area where the deviation occurred, and a dynamically updated optimized scheme is output.

2. The intelligent optimization method for land spatial planning schemes driven by multi-source data as described in claim 1, characterized in that, In the constrained multi-objective evolutionary algorithm, the purpose of the spatial unit constitutes the decision variable; wherein, the cumulative violation of all hard constraints by the decision variable is expressed as follows: in, For configuration The cumulative number of violations of all hard constraints, hour, If a solution is feasible, in the comparison of the dominance of candidate solutions, the feasible solution dominates the infeasible solution; if both solutions are feasible, the solution with an objective function value that is not inferior to the other and at least one objective function value that is strictly superior is the dominant solution; if both solutions are infeasible, the solution with the smaller cumulative violation is the dominant solution.

3. The intelligent optimization method for land spatial planning schemes driven by multi-source data as described in claim 2, characterized in that, The objective function includes targets for farmland contiguousness, matching degree of land consolidation with land increase / decrease linkage, and suitability for land reclamation, wherein: Target for contiguous arable land : Increase / decrease linkage matching target : Reclamation suitability targets : in, This represents the total number of cultivated land patches. For the first Area of ​​each cultivated land patch For the demolition area unit collection, To build a new district unit collection, , These are the areas of the corresponding units; for Unit and The distance between units; This is the distance attenuation coefficient; Spatial matching coefficients; This is a collection of units to be reclaimed / renovated. For unit Reclamation suitability score For unit The area.

4. The intelligent optimization method for land spatial planning schemes driven by multi-source data as described in claim 1, characterized in that, The hard constraints include, but are not limited to, the three-zone, three-line control rules, the permanent basic farmland protection rules, the ecological protection red line rules, the urban development boundary rules, the land use control rules, and the slope and topsoil thickness rules; the rule knowledge graph is composed of entity-relationship triples, the triples include head entity, relation, and tail entity, the head entity and tail entity include land parcel, land use, control boundary, and planning indicators, and the relations include located, belong to, prohibited, and restricted by.

5. The intelligent optimization method for land spatial planning schemes driven by multi-source data as described in claim 1, characterized in that, The data credibility assessment model calculates the overall credibility of each data source through five dimensions: completeness, accuracy, timeliness, consistency, and authority. The conflict arbitration specifically involves including only data sources with an overall credibility not lower than a preset threshold into the fusion set, and then performing a weighted average based on the weights of each data source after renormalization within the fusion set.

6. The intelligent optimization method for land spatial planning schemes driven by multi-source data as described in claim 1, characterized in that, When the deviation exceeds a preset threshold, incremental re-optimization is performed on the local area where the deviation occurred, specifically including: The deviation is measured by whether the actual land use status of the spatial unit is consistent with the recommended scheme. When any spatial unit is inconsistent or the proportion of inconsistent spatial units in the entire region exceeds a preset average deviation threshold, the incremental re-optimization is triggered. The optimization process is re-executed for the local region consisting of inconsistent spatial units and their neighboring spatial units.

7. The intelligent optimization method for land spatial planning schemes driven by multi-source data as described in claim 1, characterized in that, In the aforementioned multi-scenario coupled dynamic simulation, the state transition probability of a spatial unit is specifically represented as follows: in, for Full-area configuration at all times For unit exist Current state; For suitability items; For neighborhood influence; Context-driven items; , , These are the weighting coefficients. For normalization function, For the intended use type.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running thereon, characterized in that, When the processor executes the program, it implements a multi-source data-driven intelligent optimization method for land spatial planning schemes as described in any one of claims 1-7.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements a multi-source data-driven intelligent optimization method for land spatial planning schemes as described in any one of claims 1-7.