Urban and rural spatial layout dynamic optimization system based on multi-source data fusion

Abstract

The invention relates to the technical field of data processing and optimization decision making, in particular to a dynamic optimization processing system for multi-source heterogeneous data, and aims to solve the problems that the integration efficiency of the multi-source heterogeneous data is low, and the real-time performance of multi-target collaborative optimization is poor. Comprising the steps that a multi-source data acquisition and preprocessing module realizes standardized fusion; constructing a multi-objective function and constraint conditions based on the dynamic optimization model of the spatial-temporal characteristics; an NSGA-II algorithm is improved to generate a Pareto optimal solution set; and closed-loop optimization is formed by real-time monitoring and a feedback adjustment mechanism. Through a data-model-scheme-feedback system, dynamic collaborative optimization of economy, ecology, livelihood and space targets is realized. The method is suitable for complex decision support scenes such as urban and rural planning, resource allocation and environment monitoring, and improves the multi-source heterogeneous data processing efficiency and optimization precision.

Description

Technical Field

[0001] This invention relates to the field of data processing and optimization decision-making technology, specifically to a dynamic optimization processing system for multi-source heterogeneous data. This system integrates structured data, unstructured data, and real-time streaming data, combining spatiotemporal feature analysis and multi-objective collaborative optimization algorithms to construct a closed-loop system of "data-model-solution-feedback." It is suitable for scenarios requiring dynamic optimization decision-making, such as urban and rural planning, resource allocation, and environmental monitoring. It is particularly suitable for decision support in complex systems that require comprehensive processing of multi-source heterogeneous data and the achievement of multi-objective (economic, ecological, livelihood, spatial) collaborative optimization. Background Technology

[0002] Current urban and rural layout planning faces three major pain points: First, fragmented and heterogeneous storage from multiple sources (remote sensing imagery, population statistics, traffic flow, ecological monitoring, etc.) makes it difficult to unify planning data support; second, traditional planning relies on static data to formulate fixed schemes, failing to respond to dynamic changes such as population flow and industrial relocation; and third, optimization models often focus on a single objective (land use efficiency), making it difficult to balance diverse needs such as the economy, ecology, and people's livelihoods. Existing technologies often limit data fusion to single types of data (such as only merging remote sensing and GIS data), dynamic adjustments rely on manual iteration, and multi-objective collaboration lacks a scientific weighting mechanism, resulting in planning schemes being out of touch with actual development and low implementation efficiency.

[0003] While existing technologies have made breakthroughs in data fusion, dynamic modeling, and multi-objective optimization, they still fail to simultaneously address core issues such as unified support for multi-source heterogeneous data, dynamic response, multi-objective scientific collaboration, and closed-loop feedback. There is an urgent need for a technical solution that can deeply integrate multi-source data, dynamically adapt to urban and rural changes, balance multiple objectives, and achieve closed-loop optimization in order to improve the accuracy and adaptability of urban and rural spatial layout planning. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a dynamic optimization system for urban and rural spatial layout based on multi-source data fusion, which solves the problems of lack of a unified processing mechanism for multi-source heterogeneous data, limitations of static schemes, and insufficient multi-objective collaboration.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a dynamic optimization system for urban and rural spatial layout based on multi-source data fusion, comprising: A standardized fusion framework for multi-source heterogeneous data, used to uniformly process structured data, unstructured data, and real-time streaming data; The dynamic optimization model based on spatiotemporal characteristics includes a multi-objective collaborative algorithm module and a real-time feedback mechanism module, wherein the real-time feedback mechanism module is connected to the multi-objective collaborative algorithm module. The data-model-solution-feedback closed-loop system is connected to the multi-source heterogeneous data standardization and fusion framework and the dynamic optimization model based on spatiotemporal features; The multi-source data acquisition and preprocessing module includes a data acquisition unit, a cleaning unit, a spatiotemporal alignment unit, and a format conversion unit, which are connected to the multi-source heterogeneous data standardization and fusion framework. The multi-objective function and constraint module of the dynamic optimization model, including economic objectives, ecological objectives, livelihood objectives, spatial objectives, spatial constraints, scale constraints and functional constraints, is connected to the dynamic optimization model based on spatiotemporal characteristics. An improved NSGA-II algorithm solution module, including a Pareto optimal solution set generation unit and an AHP filtering unit, is connected to the multi-objective function and constraint condition module of the dynamic optimization model. The real-time monitoring and feedback adjustment mechanism module includes an integrated air-ground-space monitoring unit and a trigger threshold iteration unit, which are connected to the data-model-scheme-feedback closed-loop system. The trigger threshold iteration unit is connected to the improved NSGA-Ⅱ algorithm solution module.

[0006] Preferably, the data acquisition unit includes a basic geographic data acquisition unit, a socio-economic data acquisition unit, a real-time dynamic data acquisition unit, and a planning constraint data acquisition unit; the cleaning unit uses outlier detection and missing value filling to process the data; the spatiotemporal alignment unit uses the CGCS2000 coordinate system as the spatial reference and matches the data update frequency with the year-quarter-month time dimension; the format conversion unit realizes the mutual conversion between raster data and vector data through the GDAL library, and standardizes unstructured data through JSON / CSV format.

[0007] Preferably, the multi-source heterogeneous data standardization and fusion framework adopts a three-layer fusion architecture, including a data layer, a feature layer, and a decision layer; the data layer stores data through a distributed database and a Kafka message queue; the feature layer uses principal component analysis and attention mechanism for dimensionality reduction; and the decision layer generates a comprehensive evaluation matrix of urban and rural space through a weighted fusion algorithm.

[0008] Preferably, the multi-objective collaborative algorithm module constructs economic, ecological, livelihood, and spatial objective functions; the real-time feedback mechanism module dynamically adjusts the objective weights based on monitoring data.

[0009] Preferably, the constraints include spatial constraints, scale constraints, and functional constraints; the spatial constraints prohibit development within the ecological protection red line; the scale constraints limit the total scale of urban and rural construction land; and the functional constraints require that the distance between industrial land and residential land be ≥500m.

[0010] Preferably, the AHP screening unit selects the optimal solution from the Pareto optimal solution set using expert scoring and the analytic hierarchy process.

[0011] Preferably, the integrated air-ground-space monitoring unit includes a satellite remote sensing unit, an unmanned aerial vehicle (UAV) aerial photography unit, and a ground sensor unit; the trigger threshold iteration unit automatically triggers model iteration when the monitoring index deviates from the target by ±10%.

[0012] Preferably, the solution output module in the data-model-solution-feedback closed-loop system includes a spatial layout map generation unit, an indicator description generation unit, and an implementation suggestion generation unit; dynamic simulation and interactive operation of the solution are realized through the WebGIS visualization platform.

[0013] This invention provides a dynamic optimization system for urban and rural spatial layout based on multi-source data fusion. It has the following beneficial effects: 1. This invention unifies the processing of structured, unstructured, and real-time streaming data through a standardized fusion framework for multi-source heterogeneous data. It combines a multi-objective collaborative algorithm module and a real-time feedback mechanism module within a spatiotemporal-feedback-based dynamic optimization model, along with the connection between the data-model-solution-feedback closed-loop system and the multi-source data acquisition and preprocessing module, to achieve dynamic optimization processing of multi-source heterogeneous data. Simultaneously, by integrating economic, ecological, livelihood, spatial, spatial, scale, and functional constraints within the multi-objective function and constraint module of the dynamic optimization model, along with the Pareto optimal solution set generation unit and AHP filtering unit in the improved NSGA-II algorithm solution module, and the integrated air-ground-space monitoring unit and trigger threshold iteration unit in the real-time monitoring and feedback adjustment mechanism module, and connecting these with the "data-model-solution-feedback" closed-loop system and the improved NSGA-II algorithm solution module, this invention addresses the difficulties in fusion of multi-source heterogeneous data, the static nature of optimization models, and insufficient multi-objective collaboration. This enables dynamic optimization of urban and rural spatial layout, improving the accuracy and adaptability of planning schemes. Detailed Implementation

[0014] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Example 1: This invention provides a dynamic optimization system for urban and rural spatial layout based on multi-source data fusion. The system includes a standardized fusion framework for heterogeneous multi-source data, a dynamic optimization model based on spatiotemporal characteristics, a data-model-solution-feedback closed-loop system, a multi-source data acquisition and preprocessing module, a module for the multi-objective function and constraints of the dynamic optimization model, a module for solving the improved NSGA-II algorithm, and a real-time monitoring and feedback adjustment mechanism module. The structure and function of each module will be described in detail below.

[0016] Multi-source heterogeneous data standardization and fusion framework This framework employs a three-layer fusion architecture (data layer, feature layer, and decision layer) to uniformly process structured data (such as socioeconomic statistics), unstructured data (such as text descriptions of remote sensing images), and real-time streaming data (such as sensor monitoring data). The data layer stores historical data using a distributed database (such as Hadoop HDFS) and receives real-time streaming data via a Kafka message queue, ensuring efficient data storage and transmission. The feature layer uses Principal Component Analysis (PCA) to reduce the dimensionality of high-dimensional features and introduces an attention mechanism (such as the Transformer model) to automatically learn key feature weights, improving feature extraction efficiency. The decision layer generates a comprehensive urban-rural spatial evaluation matrix using a weighted fusion algorithm (such as the entropy weight method), providing data support for subsequent optimization.

[0017] Dynamic optimization model based on spatiotemporal features The model comprises a multi-objective collaborative algorithm module and a real-time feedback mechanism module. The multi-objective collaborative algorithm module constructs functions for economic objectives (e.g., maximizing GDP), ecological objectives (e.g., maximizing green space coverage), livelihood objectives (e.g., public service facility coverage), and spatial objectives (e.g., spatial compactness), and achieves multi-objective collaborative optimization through Pareto front analysis. The real-time feedback mechanism module dynamically adjusts the objective weights by comparing monitoring data with target values. For example, when ecological indicators deviate from the target, the weight of the ecological objective is automatically increased to ensure the model's adaptability.

[0018] Data-Model-Solution-Feedback Closed-Loop System This system connects a standardized fusion framework for multi-source heterogeneous data with a dynamic optimization model based on spatiotemporal characteristics, forming a complete closed loop for data processing and optimization. The solution output module includes a spatial layout map generation unit (based on the ArcGIS platform), an indicator description generation unit (automatically generating PDF reports), and an implementation suggestion generation unit (such as land use adjustment suggestions). It also enables dynamic simulation and interactive operation of the solution through a WebGIS visualization platform, allowing users to view the optimization results in real time by adjusting parameters.

[0019] Multi-source data acquisition and preprocessing module This module comprises four core data acquisition units (basic geographic data, socioeconomic data, real-time dynamic data, and planning constraint data), a cleaning unit, a spatiotemporal alignment unit, and a format conversion unit. The cleaning unit uses the 3σ method to detect outliers and employs mean imputation to handle missing values. The spatiotemporal alignment unit uses the CGCS2000 coordinate system as the spatial reference and matches data update frequencies using year-quarter-month time dimensions (e.g., aligning annual socioeconomic data with monthly remote sensing data). The format conversion unit uses the GDAL library to convert between raster data (e.g., DEM) and vector data (e.g., administrative divisions) and standardizes unstructured data (e.g., text-described planning policies) using JSON / CSV formats.

[0020] Multi-objective function and constraint module of dynamic optimization model This module includes economic objectives (such as maximizing the benefits of industrial land use), ecological objectives (such as maximizing the area of ​​ecological protection), livelihood objectives (such as the coverage rate of educational facilities), spatial objectives (such as spatial agglomeration), spatial constraints (prohibition of development within ecological protection red lines), scale constraints (total scale of urban and rural construction land ≤ upper limit), and functional constraints (distance between industrial land and residential land ≥ 500m). Specific parameters for these constraints can be adjusted according to regional planning; for example, the scope of ecological protection red lines can be obtained through remote sensing interpretation.

[0021] Improved NSGA-II algorithm solution module This module includes a Pareto optimal solution set generation unit and an AHP screening unit. The Pareto optimal solution set generation unit generates non-dominated solution sets using an improved NSGA-II algorithm (introducing crowding distance sorting and elite retention strategies); the AHP screening unit selects the optimal solution from the Pareto solution set using expert scoring (such as the 1-9 scale method) and the Analytic Hierarchy Process (AHP), ensuring the scientific validity and operability of the solution.

[0022] Real-time monitoring and feedback adjustment mechanism module This module includes an integrated air-ground-space monitoring unit and a trigger threshold iteration unit. The integrated air-ground-space monitoring unit combines satellite remote sensing (such as Landsat 8), UAV aerial photography (resolution up to 0.1m), and ground sensors (such as PM2.5 monitors) to achieve multi-scale monitoring; the trigger threshold iteration unit automatically triggers model iteration when the monitored index deviates from the target by ±10%. For example, when the green coverage rate is lower than the target value by 10%, the optimization model is rerun to adjust the land use layout.

[0023] Further describing the multi-source data acquisition and preprocessing module, the data acquisition unit includes a basic geographic data acquisition unit, a socio-economic data acquisition unit, a real-time dynamic data acquisition unit, and a planning constraint data acquisition unit. The basic geographic data acquisition unit acquires vector data such as administrative divisions and topography through a GIS platform, with a scale of 1:10000. The socio-economic data acquisition unit acquires structured data such as GDP and population from statistical yearbooks, with a time span of the past 5 years. The real-time dynamic data acquisition unit acquires real-time data such as traffic flow and air quality through an IoT sensor network, with a sampling frequency of once per hour. The planning constraint data acquisition unit extracts constraint information such as ecological red lines and basic farmland from the land spatial planning text and parses the unstructured text using NLP technology.

[0024] The cleaning unit uses the isolated forest algorithm to detect outliers and fills in missing values ​​using the multiple interpolation (MICE) method to ensure data integrity.

[0025] The spatiotemporal alignment unit uses the CGCS2000 coordinate system as the spatial reference and unifies the spatial resolution through spatial interpolation (such as Kriging interpolation); the time dimension matching adopts the sliding window method, which breaks down annual data into quarterly data to match the high-frequency update requirements.

[0026] The format conversion unit converts raster data to vector data using GDAL's gdal_translate and ogr2ogr tools; unstructured data is converted into JSON format after key information is extracted using the BERT model.

[0027] The three-layer fusion architecture is further described in detail. It consists of a data layer, a feature layer, and a decision layer. The data layer uses a distributed database, Hadoop HDFS, with a storage capacity of petabytes (PB), supporting both structured and unstructured data. A Kafka message queue is configured with three partitions and two replicas to ensure stable transmission of real-time streaming data. The feature layer uses PCA dimensionality reduction implemented through the scikit-learn library, preserving 95% of the variance. The attention mechanism employs a multi-head attention model with a hidden layer dimension of 128, and feature weights are normalized using the Softmax function. The decision layer's weighted fusion algorithm combines entropy weighting (objective weights) and AHP (subjective weights) to generate a comprehensive evaluation matrix with dimensions of (number of samples × number of indicators).

[0028] The multi-objective collaborative algorithm module constructs economic, ecological, livelihood, and spatial objective functions; the economic objective function is... , If For the area of ​​industrial land of type i, For The output per unit area; the ecological objective function is , Eg For the area of ​​Class j ecological land, Total Area The total urban and rural area is denoted as ; the objective function for improving people's livelihood is . , Tk For the average reachability time of the k-th type of facility, Qk To serve population weights; the spatial objective function is: .

[0029] The real-time feedback mechanism module calculates the deviation rate between the monitoring data and the target value through a sliding window. When the deviation rate exceeds 5%, an adaptive weight adjustment algorithm (such as fuzzy logic) is used to dynamically update the target weight. For example, the ecological target weight is increased from 0.3 to 0.5.

[0030] The constraints include spatial constraints, scale constraints, and functional constraints. Spatial constraints are obtained from the ecological protection red line area through remote sensing interpretation with an accuracy of over 90%, and prohibited development areas are stored as polygon vector data. Scale constraints are determined from the upper limit of the total scale of urban and rural construction land according to the national land space plan, for example, the upper limit for a certain area is 500 km². 2 The actual scale is calculated through the land use type field; the functional constraints are calculated by spatial analysis of the distance between industrial land and residential land, with a minimum distance threshold of 500m, which can be met by adjusting the land use layout.

[0031] The AHP screening unit invites five domain experts to compare the solutions in the Pareto solution set pairwise. A judgment matrix is ​​constructed using the 1-9 scaling method. The eigenvectors and the largest eigenvalue of the judgment matrix are calculated. After passing the consistency test (CR<0.1), the weight of the solution is determined. Finally, the solution with the highest comprehensive score is selected as the optimal solution.

[0032] The integrated air-ground-space monitoring unit comprises a satellite remote sensing unit, a UAV aerial photography unit, and a ground sensor unit. The satellite remote sensing unit uses Landsat 8 data with a spatial resolution of 30m and a revisit period of 16 days. The UAV aerial photography unit is equipped with a five-lens camera with a resolution of 0.1m and a flight altitude of 100m. The ground sensor unit deploys sensors for PM2.5, noise, etc., with a sampling frequency of once per hour. The trigger threshold iteration unit recalculates the improved NSGA-II algorithm via API calls when the monitored indicator deviates from the target by ±10%, with no more than 5 iterations to avoid wasting computational resources.

[0033] The solution output module in the data-model-solution-feedback closed-loop system includes a spatial layout map generation unit, an indicator description generation unit, and an implementation suggestion generation unit. The spatial layout map generation unit generates a land use planning map based on ArcGIS, including elements such as legend and scale. The indicator description generation unit automatically generates a PDF report using a LaTeX template, which includes an indicator comparison table. The implementation suggestion generation unit generates land use adjustment suggestions based on a rule engine, such as adjusting industrial land area A to residential land.

[0034] The WebGIS visualization platform uses the open-source platform GeoServer to publish map services, allowing users to perform interactive operations through a browser, such as dragging sliders to adjust the proportion of industrial land and viewing the optimization results in real time.

[0035] Example 2: This invention provides a method for dynamic optimization of urban and rural spatial layout based on multi-source data fusion, comprising the following steps: S1. Multi-source data acquisition and standardized preprocessing Multi-dimensional data collection Four types of core data are collected through basic geographic data collection units, socio-economic data collection units, real-time dynamic data collection units, and planning constraint data collection units: Basic geographic data: 0.5m-2m resolution annual remote sensing imagery, quarterly updated GIS vector maps (roads, water systems, administrative divisions); Socioeconomic data: 5-year updated census structured forms, monthly industry output data, and quarterly data on the coordinates and service radius of public service facilities (schools / hospitals); Real-time dynamic data: minute-level roadside radar + floating car GPS traffic flow data, hour-level mobile phone signaling temporary population movement data, and hour-level ecological monitoring sensor (air quality / soil moisture) data; Planning constraint data: vector files of ecological protection red lines, permanent basic farmland, and urban development boundaries (updated synchronously when policies are adjusted).

[0036] b. Data standardization processing Data preprocessing is completed through a cleaning unit, a spatiotemporal alignment unit, and a format conversion unit. Data cleaning: Z-score outlier detection method was used to remove abnormal data such as traffic flow peaks, and KNN interpolation method was used to complete missing population data in townships to ensure data quality; Spatiotemporal alignment: All data coordinates are aligned using the 2000 National Geodetic Coordinate System (CGCS2000) as a unified spatial reference; data with different update frequencies are matched using year-quarter-month as the time dimension (e.g., annual remote sensing data and monthly industry data are associated with the quarterly dimension). Format conversion: The GDAL library enables the conversion between raster data (remote sensing imagery) and vector data (GIS map), and standardizes unstructured data such as mobile phone signaling through JSON / CSV format, laying the foundation for subsequent integration.

[0037] S2. Multi-source heterogeneous data fusion processing Based on a framework-driven three-layer architecture (data layer, feature layer, and decision layer), data collaborative integration is achieved. Data layer storage and access HBase distributed database is used to store preprocessed static data (such as population census data, GIS maps). By building a low-latency interface through Kafka message queues, we can achieve rapid access to real-time streaming data (traffic flow, ecological monitoring) and avoid data update delays. b. Feature layer extraction and dimensionality reduction Extract core features from each data source: land use type features of remote sensing images (arable land / construction land / ecological land), congestion index features of traffic data, and accessibility features of public service data; Principal component analysis (PCA) combined with attention mechanism is used to reduce the dimensionality of multi-dimensional features, retaining key information (such as reducing the dimensionality to 6-8 core features) and reducing the computational load of subsequent models; c. Weighted fusion of decision-making levels The entropy weight method is used to calculate the weights of each feature (such as ecological sensitivity weight 0.3, land use efficiency weight 0.25, and facility accessibility weight 0.2) to avoid subjective bias. A comprehensive evaluation matrix of urban and rural space is generated through a weighted fusion algorithm, which includes core indicators such as land use efficiency, ecological sensitivity, and accessibility of public services, forming a unified planning data support. S3. Setting Multi-Objective Functions and Constraints Construction of a multi-objective optimization function A quantitative function is established around the four major goals of economy, ecology, people's livelihood, and space: Economic objective: Maximize industrial land value density, formula: , If For the area of ​​industrial land of type i, For Output per unit area; Ecological goal: Maximize the proportion of ecological land use, formula: , Eg For the area of ​​Class j ecological land, TotalArea Total urban and rural area; Public welfare objective: Minimize the average access time to public service facilities, as shown in the formula. , Tk For the average reachability time of the k-th type of facility, Qk To serve the population weight; Spatial Objective: The formula for the ratio of urban and rural construction land is as follows: .

[0038] b. Constraint Definition Define three types of insurmountable planning boundaries: Spatial constraints: Development and construction are prohibited within the ecological protection red line, and the area of ​​permanent basic farmland shall not be reduced; Scale constraints: The total scale of urban and rural construction land shall not exceed the upper limit approved by the superior planning authority; Functional constraints: The spatial distance between industrial land and residential land shall be ≥500m to avoid industrial pollution affecting residents' lives.

[0039] S4. Dynamic Optimization Model Construction and Algorithm Solution Construction of a multi-objective collaborative dynamic optimization model Activate the multi-objective collaborative algorithm module within the model to balance target conflicts through a weight linkage mechanism (e.g., when economic targets are improved, automatically verify whether ecological targets meet the requirement of ≥40%). An embedded real-time feedback mechanism module is provided, and an interface is reserved for subsequent monitoring data to ensure that the target weight can be dynamically adjusted according to actual development (such as automatically increasing the weight of the people's livelihood target when the time for public services to be available exceeds the standard).

[0040] b. Solving with an improved NSGA-II algorithm The solution generation and selection were accomplished by improving the NSGA-II algorithm solution module: Algorithm optimization: The selection operator of the traditional NSGA-II algorithm is improved by introducing an elite retention strategy to improve the convergence speed of the optimal solution; Parameter settings: population size 200, number of iterations 100, crossover probability 0.8, mutation probability 0.05; Pareto optimal solution set generation: Output 10-15 candidate layout schemes that satisfy multiple objectives and constraints; AHP Screening: The Analytic Hierarchy Process (AHP) screening unit within the module is activated, and combined with expert scores (economic objectives weighted at 0.3, ecology at 0.3, and people's livelihood at 0.4), a unique optimal solution is selected from the solution set.

[0041] S5. Optimization Solution Output and Visualization a multi-form scheme output Outputs are generated through the spatial layout map generation unit, indicator description generation unit, and implementation suggestion generation unit: Spatial layout map: includes vector layers of land use zoning (residential / industrial / ecological / public service), road network optimization, and facility layout adjustments; Indicator Explanation: Covers the predicted values ​​and optimization ranges of core indicators such as economy, ecology, and people's livelihood; Implementation Recommendations: Clearly define construction priorities in the short, medium, and long term (e.g., prioritize the construction of primary schools in the west in the short term, and promote ecological restoration in the east in the medium term).

[0042] bWebGIS Visualization An interactive visualization system built on the WebGIS platform supports: Dynamic simulation of the plan (simulating the trend of land use type changes within 5 years); Interactive query at the land parcel level (click on any land parcel to view detailed indicators such as land use, expected output value, etc.); Comparison of multiple options (showing the differences in metrics among different candidate options side by side); The output format is compatible with commonly used tools in the planning industry, supporting PDF, CAD, and SHP file export.

[0043] S6. Real-time monitoring and dynamic feedback iteration Integrated monitoring of space, air, and ground Activate the integrated air-ground-space monitoring unit to achieve full-dimensional tracking of implementation effects: Satellite remote sensing monitoring (quarterly): Verify changes in land use types (such as whether cultivated land has been illegally converted into construction land); Drone aerial monitoring (monthly): Focus on verifying the construction progress of key areas such as industrial parks and public service facilities; Ground sensor monitoring (real-time): Monitors dynamic indicators such as traffic congestion index, efficiency of public service facilities, and ecological environment quality.

[0044] b. Construct a closed-loop optimization mechanism by triggering threshold iteration units: Threshold setting: When the monitored indicator deviates from the planning target by ±10% (such as the public service access time increasing from 12 minutes to 18 minutes), an iteration is automatically triggered; Iteration process: Data update: The monitoring data is fed into the multi-source heterogeneous data standardization and fusion framework to update the urban and rural spatial comprehensive evaluation matrix; Model adjustment: The dynamic optimization model based on spatiotemporal characteristics automatically adjusts the target weights (e.g., the weight of the livelihood target is increased from 0.4 to 0.45). Resolve: Start the improved NSGA-II algorithm solution module to generate a new optimal solution; Solution Update: The iterative solution is output through the WebGIS platform to replace the old solution, completing one closed loop.

[0045] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A dynamic optimization system for urban and rural spatial layout based on multi-source data fusion, characterized in that: include: A standardized fusion framework for multi-source heterogeneous data, used to uniformly process structured data, unstructured data, and real-time streaming data; The dynamic optimization model based on spatiotemporal characteristics includes a multi-objective collaborative algorithm module and a real-time feedback mechanism module, wherein the real-time feedback mechanism module is connected to the multi-objective collaborative algorithm module. The data-model-solution-feedback closed-loop system is connected to the multi-source heterogeneous data standardization and fusion framework and the dynamic optimization model based on spatiotemporal features; The multi-source data acquisition and preprocessing module includes a data acquisition unit, a cleaning unit, a spatiotemporal alignment unit, and a format conversion unit, which are connected to the multi-source heterogeneous data standardization and fusion framework. The multi-objective function and constraint module of the dynamic optimization model, including economic objectives, ecological objectives, livelihood objectives, spatial objectives, spatial constraints, scale constraints and functional constraints, is connected to the dynamic optimization model based on spatiotemporal characteristics. An improved NSGA-II algorithm solution module, including a Pareto optimal solution set generation unit and an AHP filtering unit, is connected to the multi-objective function and constraint condition module of the dynamic optimization model. The real-time monitoring and feedback adjustment mechanism module includes an integrated air-ground-space monitoring unit and a trigger threshold iteration unit, which are connected to the data-model-scheme-feedback closed-loop system. The trigger threshold iteration unit is connected to the improved NSGA-Ⅱ algorithm solution module.

2. The urban and rural spatial layout dynamic optimization system based on multi-source data fusion according to claim 1, characterized in that: The data acquisition unit includes a basic geographic data acquisition unit, a socio-economic data acquisition unit, a real-time dynamic data acquisition unit, and a planning constraint data acquisition unit; the cleaning unit uses outlier detection and missing value filling to process the data; the spatiotemporal alignment unit uses the CGCS2000 coordinate system as the spatial reference and matches the data update frequency with the year-quarter-month time dimension; the format conversion unit realizes the mutual conversion between raster data and vector data through the GDAL library, and standardizes unstructured data through JSON / CSV format.

3. The urban and rural spatial layout dynamic optimization system based on multi-source data fusion according to claim 1, characterized in that: The multi-source heterogeneous data standardization and fusion framework adopts a three-layer fusion architecture, including a data layer, a feature layer, and a decision layer. The data layer stores data through a distributed database and a Kafka message queue. The feature layer uses principal component analysis and attention mechanism for dimensionality reduction. The decision layer generates a comprehensive evaluation matrix of urban and rural space through a weighted fusion algorithm.

4. The urban and rural spatial layout dynamic optimization system based on multi-source data fusion according to claim 1, characterized in that: The multi-objective collaborative algorithm module constructs economic, ecological, livelihood, and spatial objective functions; the real-time feedback mechanism module dynamically adjusts the objective weights based on monitoring data.

5. The urban and rural spatial layout dynamic optimization system based on multi-source data fusion according to claim 1, characterized in that: The constraints include spatial constraints, scale constraints, and functional constraints; the spatial constraints prohibit development within the ecological protection red line; the scale constraints limit the total scale of urban and rural construction land; and the functional constraints require that the distance between industrial land and residential land be ≥500m.

6. The urban and rural spatial layout dynamic optimization system based on multi-source data fusion according to claim 1, characterized in that: The AHP screening unit selects the optimal solution from the Pareto optimal solution set using expert scoring and the analytic hierarchy process.

7. The urban and rural spatial layout dynamic optimization system based on multi-source data fusion according to claim 1, characterized in that: The integrated air-ground-space monitoring unit includes a satellite remote sensing unit, an unmanned aerial vehicle (UAV) aerial photography unit, and a ground sensor unit; the trigger threshold iteration unit automatically triggers model iteration when the monitoring index deviates from the target by ±10%.

8. The urban and rural spatial layout dynamic optimization system based on multi-source data fusion according to claim 1, characterized in that: The solution output module in the data-model-solution-feedback closed-loop system includes a spatial layout map generation unit, an indicator description generation unit, and an implementation suggestion generation unit; dynamic simulation and interactive operation of the solution are realized through the WebGIS visualization platform.