A method and system for optimizing analysis of territorial space planning based on big data

Abstract

The application discloses a kind of big data-based optimization analysis method and system of territorial space planning, it is related to big data and optimization analysis technical field of territorial space planning, including real-time data acquisition and dynamic preprocessing, dynamic regional division, multi-model integrated space-time prediction, multi-objective dynamic optimization planning, real-time feedback and adaptive adjustment.The big data-based optimization analysis method of territorial space planning provided in the application adopts space-time weighted fuzzy clustering method, divides the territorial space into several sub-regions according to the spatial coordinates and time attributes of data, constructs the weighted distance regular objective function between data points and regional center, and updates the membership and the position of regional center iteratively, realizes the self-adaptation and space-time smoothing of regional segmentation, effectively captures the dynamic change characteristics in region.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of big data and optimization analysis of land space planning, in particular to a land space planning optimization analysis method and system based on big data. BACKGROUND

[0002] In recent years, with the rapid development of information technologies such as remote sensing, GIS, Internet of Things, cloud computing and artificial intelligence, the application of big data in land space planning has become mature. Various data collection, transmission and processing technologies are constantly improving, making it possible to obtain massive spatio-temporal data in real time. At the same time, multi-source data fusion, fuzzy clustering, spatio-temporal prediction and multi-objective optimization methods have been gradually formed and applied, providing a theoretical basis and practical foundation for the transformation from traditional static planning to dynamic real-time optimization.

[0003] However, the existing technology still has many deficiencies. First, most methods rely on static data or batch processing, lack the ability of continuous collection and dynamic preprocessing of real-time data streams, and cannot reflect the instantaneous changes of the environment and resources in the land space in a timely manner. Second, there is heterogeneity between different data sources, and the data cleaning, missing completion and normalization process is rigid and difficult to automatically adjust to adapt to data quality fluctuations, resulting in unstable subsequent analysis results. Third, the regional division algorithm often only relies on spatial information, ignoring the time factor, resulting in that the partitioning results cannot reflect the spatio-temporal continuity. In addition, a single prediction model cannot take into account errors and uncertainties, lacks multi-model integration strategy, and results in insufficient prediction accuracy. Finally, the multi-objective optimization method cannot realize real-time feedback and dynamic weight adjustment, and the planning scheme is difficult to calibrate deviation in the implementation process, and cannot meet the comprehensive optimization requirements of economic, ecological and social benefits. The present application effectively solves the above problems by constructing a dynamic preprocessing based on big data, spatio-temporal weighted fuzzy clustering, multi-model integrated prediction and real-time adaptive optimization mechanism, and realizes the whole process closed-loop dynamic planning. SUMMARY

[0004] In view of the above problems, the present application is proposed.

[0005] Therefore, the technical problem solved by the present application is that the existing land space planning method has the problems of non-real-time data collection and preprocessing, lack of spatio-temporal continuity in regional division, insufficient accuracy of single prediction model and non-timely optimization feedback, and how to realize real-time dynamic optimization of land space planning scheme.

[0006] To solve the above technical problems, the application provides the following technical scheme: a land space planning optimization analysis method based on big data, which comprises collecting space-time data, storing the space-time data through a sliding window, performing abnormality detection, missing data completion and smoothing filtering, dynamically adjusting normalization parameters, fusing overlapping data according to weights, and continuously optimizing preprocessing according to data quality feedback; adopting space-time weighted fuzzy clustering, partitioning multi-source data collected in real time according to spatial coordinates and time attributes, constructing a regular objective function by using weighted distances between data points and region centers, and updating membership degrees and center positions through iteration; constructing a space-time prediction model, updating prediction errors and uncertainties in real time, dynamically adjusting weights, and outputting prediction results through weighted average integration; converting land planning targets into a multi-objective optimization problem, constructing a comprehensive objective function covering economic cost, environmental protection and social benefit indicators, dynamically adjusting weights of each target according to real-time feedback, and using an augmented Lagrange and meta-heuristic algorithm to realize real-time adaptive optimization and dynamic adjustment of planning schemes; after the planning is implemented, real-time data are collected, and the real-time data are compared with and error analyzed based on the pre-prediction values.

[0007] As a preferred scheme of the land space planning optimization analysis method based on big data, the collection of space-time data comprises obtaining original data from remote sensing images, GIS maps, Internet of Things sensors, social and economic statistics and environmental monitoring data sources.

[0008] As a preferred scheme of the land space planning optimization analysis method based on big data, the fusion of overlapping data according to weights comprises giving different weights to each data source according to the stability and quality of each data source, and fusing data at the same position and time by using a weighted average method.

[0009] As a preferred scheme of the land space planning optimization analysis method based on big data, the updating of membership degrees and center positions through iteration comprises dividing space-time data collected in real time in a land space, constructing a fuzzy clustering model with time regularization, and defining space-time distances of data points and regions.

[0010] The membership degrees are updated, the gradient of the regularization term is included in the solving iteration in actual updating, and the center is updated.

[0011] The membership degrees and the region center are updated online, and the regions are divided in space and time.

[0012] As a preferred scheme of the land space planning optimization analysis method based on big data, the construction of a space-time prediction model, the real-time updating of prediction errors and uncertainties, the dynamic adjustment of weights, and the output of prediction results through weighted average integration comprise predicting future states of indicators of each region.

[0013] The spatio-temporal prediction model is constructed, and a spatio-temporal prediction model prediction value is outputted; the spatio-temporal prediction model is integrated; the weight of the spatio-temporal prediction model is dynamically adjusted; and the weight of the spatio-temporal prediction model is constructed according to a recent prediction error;

[0014] Considering the uncertainty of the spatio-temporal prediction model prediction, the error variance estimation is used to define the dynamic weight of the spatio-temporal prediction model.

[0015] As a preferred scheme of the land space planning optimization analysis method based on big data, the comprehensive target function covering economic cost, environmental protection and social benefit indicators is constructed, the target weight is dynamically adjusted according to real-time feedback, and the planning scheme is dynamically optimized and adjusted in real time by using the augmented Lagrange and meta-heuristic algorithm, including multi-objective dynamic optimization of planning decision variables according to the spatio-temporal prediction model prediction result and the land planning target.

[0016] The multi-objective optimization model is constructed, the dynamic weighting is adopted, the weight is dynamically updated according to real-time feedback, and the comprehensive target is defined.

[0017] The strategy of alternately updating is adopted to output the dynamic optimization problem.

[0018] As a preferred scheme of the land space planning optimization analysis method based on big data, after the planning is implemented, real-time data is collected, and the actual data is compared with the pre-prediction value and error analysis is performed, including starting the feedback mechanism, re-collecting the latest data and adjusting the spatio-temporal prediction model parameters, and dynamically optimizing the planning decision when the actual data deviates from the prediction by more than a set threshold.

[0019] Through continuous updating and closed-loop feedback, real-time adaptive adjustment of the planning scheme is realized, and the error feedback parameters are adjusted according to the updated actual data.

[0020] Another object of the present application is to provide a land space planning optimization analysis system based on big data, which can use spatio-temporal weighted fuzzy clustering to divide the multi-source data collected in real time according to spatial coordinates and time attributes, use the weighted distance between data points and region centers to construct a regular target function, and update the membership degree and center position through iteration, thereby solving the problem that the current land space planning method lacks spatio-temporal continuity in region division.

[0021] As a preferred scheme of the big data-based land space planning optimization analysis system, the system comprises a real-time data acquisition and dynamic preprocessing module, a dynamic region division module, a multi-model integrated space-time prediction module, a multi-target dynamic optimization planning module, and a real-time feedback and adaptive adjustment module; the real-time data acquisition and dynamic preprocessing module is used for collecting real-time data and performing cleaning, normalization and fusion; the dynamic region division module is used for adaptively dividing the land space according to the space-time characteristics of the data, so that the data in the same region are similar; the multi-model integrated space-time prediction module is used for constructing a space-time prediction model for each region, and predicting future land use, environment and traffic indexes through dynamic weight integration; the multi-target dynamic optimization planning module is used for generating an optimal planning decision in real time according to a multi-target optimization model of the prediction result and multiple planning targets; and the real-time feedback and adaptive adjustment module is used for continuously monitoring the comparison between actual operation data and predicted values.

[0022] A computer device comprises a memory and a processor, the memory stores a computer program, and the processor executes the computer program to implement the steps of the big data-based land space planning optimization analysis method.

[0023] A computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the big data-based land space planning optimization analysis method.

[0024] The application provides a land space planning optimization analysis method based on big data. BRIEF DESCRIPTION OF DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0026] Figure 1 The overall flowchart of a land space planning optimization analysis method based on big data provided by the first embodiment of the present application.

[0027] Figure 2A real-time data collection and dynamic preprocessing flowchart of a big data-based land space planning optimization analysis method is provided for the first embodiment of the present application.

[0028] Figure 3 A whole flowchart of a big data-based land space planning optimization analysis system is provided for the third embodiment of the present application. DETAILED DESCRIPTION

[0029] In order to make the above objectives, characteristics and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should fall within the scope of protection of the present application.

[0030] Embodiment 1, refer to Figure 1 - Figure 2 For an embodiment of the present application, a big data-based land space planning optimization analysis method is provided, comprising:

[0031] S1: Collecting spatio-temporal data, storing through a sliding window, performing abnormality detection, missing data completion and smoothing filtering, dynamically adjusting normalization parameters, fusing overlapping data according to weights, and continuously optimizing preprocessing according to data quality feedback.

[0032] Further, raw data is obtained from multiple data sources such as remote sensing images, GIS maps, Internet of Things sensors, social and economic statistics, and environmental monitoring. These data are continuously transmitted in a streaming manner, including the spatial position, collection time and observation value of each data. The collected data are preliminarily grouped according to data source and collection time, and stored in a sliding time window within a certain time range to provide a continuous data basis for subsequent processing.

[0033] It should be noted that each data is checked to determine whether it is outside the reasonable range (for example, the normal fluctuation range is determined by historical data or preset standards), and the abnormal data outside the range is marked. For missing data caused by transmission or collection problems, the mean or median of adjacent spatio-temporal data is used for interpolation to make the data sequence continuous and complete. The data is smoothed to reduce the interference of random noise, so that the subsequent analysis can more accurately reflect the true situation.

[0034] Further, since the observation values of different data sources may have different dimensions and numerical ranges, it is necessary to normalize each source data to convert the data to a unified standard scale. In the sliding window, the normalization parameters are dynamically adjusted according to the latest collected data to ensure that the data always reflects the latest distribution. This step can be updated in a step-by-step manner to make the normalization process adaptive to data fluctuations.

[0035] It should be noted that in the same or overlapping space-time region, the normalized data from different data sources are aligned to ensure the consistency of the data in time and space. According to the stability and quality evaluation of each data source, different weights are given to each data source, and the weighted average method is used to fuse the data at the same position and time to obtain a data value that comprehensively reflects the situation of the region. According to the quality of the fused data, the parameters used in the data collection and normalization process are adjusted to ensure that the subsequent output data meets the requirements of subsequent analysis and decision-making.

[0036] S2: Adopting spatio-temporal weighted fuzzy clustering, the real-time collected multi-source data is divided according to spatial coordinates and time attributes, and a regular objective function is constructed by using the weighted distance between data points and region center, and the membership and center position are updated by iteration.

[0037] Further, in the national space, real-time spatio-temporal data is collected:

[0038]

[0039] wherein, represents the spatio-temporal data, N represents the total number of data points, x i , y i are the spatial coordinates of the i-th data point, t i is the collection time of the i-th data point

[0040] is divided into K regions, so that the data in the same region has high similarity in space and time.

[0041] To enhance the adaptability and continuity of region division, a fuzzy clustering model with time regularization is adopted. Let the membership of each data point i to region k be u ik (satisfying ), the spatial center of region k be μ k = μ kx , μ ky , and the time center be τ k .

[0042] The spatio-temporal distance between data point i and region k is defined as:

[0043]

[0044] where d ik represents the integrated spatio-temporal distance between data point i and region k, (x i , y i ) represents the spatial coordinates of data point i, x i is the horizontal coordinate, y i is the vertical coordinate, μ k represents the spatial center of region k, which is usually denoted as μ k = (μ kx , μ ky ), i.e., the horizontal and vertical coordinates of the region center. t i represents the temporal attribute of data point i, i.e., the time of data acquisition, τ k represents the temporal center of region k, which is usually the central tendency of the temporal distribution of data within the region, and λ represents a positive constant, used to balance the relative weights of temporal and spatial differences in the calculation of the integrated distance. By adjusting λ, the degree of influence of the temporal dimension on the overall distance can be changed.

[0045] Let the membership of each data point i to region k be u ik (satisfying and u ik ∈ [0, 1]), where m > 1 is the fuzzy index. λ > 0 is the weighting factor between time and space scales.

[0046] The fuzzy membership and the objective function are constructed as:

[0047]

[0048] where J represents the objective function, m > 1 is the fuzzy index controlling the fuzziness of membership, i represents the index of data points, and k represents the index of clusters or regions.

[0049] The second term is the temporal regularization term, and γ > 0 adjusts the continuity constraint to ensure that the clustering membership of data at the same location changes smoothly over time.

[0050] It should be noted that the iterative solution can use the following rules:

[0051] Update membership:

[0052]

[0053] where j represents the index of traversing all regions, d ij represents the distance between data point i and cluster k.

[0054] In actual updating, the gradient of the regularization term can be included in the solution iteration.

[0055] Update center:

[0056]

[0057] where μ k denotes the spatial center of cluster k, denotes the membership degree u ik with m-th power as weight, τ k denotes the temporal center of cluster k, reflecting the concentration trend of the cluster data in time, t i denotes the collection time of data point i;

[0058] By updating the membership degree and region center online, the adaptive and smooth region division in space and time is realized.

[0059] S3: Constructing the spatio-temporal prediction model, updating the prediction error and uncertainty in real time, dynamically adjusting the weight, and integrating the output prediction result by weighted average.

[0060] Further, the future state of the key indicators (such as land utilization rate, environmental quality, traffic flow, etc.) of each region k is predicted. Assuming that the actual indicator of region k at future time t is Y k (t).

[0061] M prediction models are constructed, each model giving a prediction value In order to integrate the advantages of multiple models and dynamically adjust the model weight, the model weight is constructed according to the recent prediction error.

[0062] Let the error index of model m adopt the exponentially weighted mean square error:

[0063]

[0064] where E m (t) denotes the cumulative prediction error of model m at time t, λ e ∈(0, 1) is the smoothing coefficient, Δt is the time step, Y k (t-Δt) denotes the actual measurement value of region k at time t-Δt, denotes the prediction value of model m for region k at time t-Δt, E m (t-Δt) denotes the cumulative prediction error of model m at the previous time step t-Δt.

[0065] At the same time, considering the uncertainty of model prediction, the error variance V m (t) is estimated as:

[0066]

[0067] where V m(t) represents the prediction error variance of model m at time t, V m (t - At) represents the error variance of model m at the previous time t, λ v ∈ (0, 1) is an adjustment parameter, V m (t - At) is the variance value at the previous time.

[0068] For integrated prediction, the dynamic weight of model m is defined as:

[0069]

[0070] where w m (t) represents the weight of model m at time t, reflecting the pros and cons of the current prediction performance of model m, β, η > 0 are adjustment parameters, E m (t) represents the cumulative prediction error of model m at time t, usually expressed as exponential weighted mean square error (EW-MSE), reflecting the prediction accuracy of the model, V m (t) represents the prediction error variance of model m at time t, used to measure the instability or volatility of the prediction.

[0071] The final integrated prediction value of region k is represented as:

[0072]

[0073] where, represents the final integrated prediction value of region k at time t, which is a weighted average of the prediction results of each model, M represents the total number of models participating in prediction, w m (t) represents the weight of model m at time t as explained above, represents the prediction value of model m at time t in region k;

[0074] By updating the error and variance online and dynamically adjusting the weights of each model, the adaptive integration of multi-model prediction results is realized.

[0075] S4: Convert the land planning target into a multi-objective optimization problem, construct a comprehensive objective function covering economic cost, environmental protection and social benefit indicators, dynamically adjust the weights of each target according to real-time feedback, and use augmented Lagrangian and meta-heuristic algorithms to realize real-time adaptive optimization and dynamic adjustment of planning scheme.

[0076] Further, according to the prediction results and land planning targets, the multi-objective dynamic optimization of planning decision variables x (such as the proportion of land use in each region, resource allocation, etc.) is carried out to balance multiple objectives such as economic benefit, ecological protection and resource utilization.

[0077] Let x is the planning decision vector; n is the dimension of decision variables (e.g. land use proportion in each region, resource allocation, etc.); p objective functions f i (x,t) are defined (e.g. f1 represents economic cost, f2 represents environmental impact, f3 represents social benefit, etc.), q inequality constraints g j (x,t)≤0, and r equality constraints h k (x,t) = 0.

[0078] The multi-objective optimization problem is represented as:

[0079] m x in F(x,t) = [f1(x,t), f2(x,t), …, f p (x,t)]

[0080] s.t.g j (x,t)≤0,j = 1, …, q

[0081] h k (x,t) = 0, k = 1, …, r

[0082] where x is the planning decision vector, t is the time parameter, F(x,t) is the objective function vector, f i (x,t) is the i-th objective function, which describes the evaluation index of economic, environmental, social, etc. aspect, depends on decision variable x and time t, g j (x,t) is the j-th inequality constraint function, which requires to satisfy g j (x,t)≤0 (such as resource constraints, environmental constraints, etc.), h k (x,t) is the k-th equality constraint function, which requires to satisfy h k (x,t) = 0 (such as balance conditions, conservation relations, etc.), p, q, r represent the total number of objective functions, inequality constraints and equality constraints.

[0083] For easy solving, the multi-objective problem is converted into a single objective problem, and a dynamic weighting method is adopted. Let the weight ω i (t) be dynamically updated according to real-time feedback (such as the deviation between the predicted value and the target value), and define the comprehensive objective as:

[0084]

[0085] where L(x,t) represents the comprehensive objective function value, which is the overall evaluation index after converting the multi-objective problem into a single objective problem, ω i (t) represents the dynamic weight of the i-th objective function at time t, φ j > 0 is the penalty coefficient, which is used to strengthen the punishment of inequality constraints.

[0086] The dynamic updating weight can adopt the following formula:

[0087]

[0088] where ω i (t) represents the dynamic weight of the ith objective function at time t, which is used to reflect the importance of the objective in the current overall optimization, exp(-β i Δ i (t)) represents the value of the exponential decay of the relative deviation of the ith objective, β i represents a positive adjustment parameter, which is used to control the sensitivity of the influence of the ith objective deviation on the weight, Δ i (t) represents the relative deviation of the ith objective at time t, is the ith objective value, β i > 0 is an adjustment parameter, and Δ i (t) represents the relative error between the actual objective and the ideal objective.

[0089] To solve the dynamic optimization problem, an alternating updating strategy (similar to the augmented Lagrange method) can be adopted:

[0090] The fixed weight ω i (t) is represented as:

[0091]

[0092] where x * (t) represents the optimal planning decision at time t, i.e., the decision variable vector that minimizes the comprehensive objective function L(x, t), and x belongs to the feasible solution set X, i.e., all decision vectors that satisfy the planning constraints.

[0093] According to the actual operation feedback, ω i (t) and the penalty parameter (or Lagrange multiplier) are updated, and the solution is re-solved until convergence.

[0094] Or the following augmented Lagrange function is represented as:

[0095]

[0096] where, represents the augmented Lagrange function, λ represents the Lagrange multiplier vector, ω i (t) represents the weight of the ith objective function at time t, f i (x, t) represents the ith objective function, g j (x, t) represents the jth inequality constraint function, λ j (t) represents the corresponding g jthe Lagrange multiplier of inequality constraint of (x, t), λ j (t) is the Lagrange multiplier, and ρ > 0 is the penalty parameter and is updated as follows:

[0097]

[0098] wherein, represents the updated Lagrange multiplier corresponding to the jth inequality constraint at the k+1th iteration.

[0099] By dynamically adjusting the weights and penalty terms, adaptive optimization of multi-objective programming is achieved, which can output planning decisions closer to actual needs under real-time feedback.

[0100] S5: After implementing the planning, real-time data is collected and compared with the pre-predicted values for error analysis.

[0101] Furthermore, after the implementation of the planning scheme, the system will monitor key operation indicators such as traffic flow, environmental quality, resource utilization, etc., and continuously collect the latest data from sensors, real-time data platforms, and related data sources. These actual operation data will be compared with the expected values obtained through multi-model integrated prediction on a hourly basis. The specific process is as follows:

[0102] It should be noted that the system automatically acquires the latest data every fixed time (such as every hour or every half hour) and pre-processes it to ensure stable and reliable data quality. Then, through the difference analysis module, the deviation between the actual value and the predicted value is calculated, and statistical methods are used to evaluate whether the deviation exceeds the preset tolerance range (for example, when the error percentage or cumulative deviation reaches the set threshold).

[0103] Furthermore, if the deviation is detected to be excessive, the system will automatically trigger a feedback mechanism. First, it re-collects and processes the latest data, and then uses these data to update the parameters and weights of the original prediction model, so that the model can more accurately reflect the current operating state. Next, the updated prediction results will be used as input to re-run the multi-objective dynamic optimization algorithm to adjust the planning decision variables and generate a new optimal planning scheme.

[0104] It should be noted that the new planning scheme will be immediately fed back to the execution link or decision maker to ensure that the subsequent execution is highly matched with the current actual situation. The entire process forms an automatic closed-loop feedback system, from data collection, prediction, optimization to scheme adjustment, each link realizes real-time adaptation, ensuring that the planning scheme is continuously dynamically optimized throughout the implementation process.

[0105] In Example 2, an embodiment of the present application provides a land space planning optimization analysis method based on big data. In order to verify the beneficial effects of the present application, economic benefit calculation and simulation experiments are used for scientific demonstration.

[0106] First, by multi-channel real-time acquisition of spatio-temporal data of territorial space, the data sources include high-resolution remote sensing images, GIS map data, various Internet of Things sensor data, social and economic statistical data, and environmental monitoring data. First, the sliding window technology is used to store the data collected in the past 24 hours, and each data is detected for abnormalities. The data that exceeds the preset reasonable interval is removed, and the missing part of the data is completed by using the mean value of the adjacent spatio-temporal data. Then, the smoothing filter algorithm is used to eliminate noise. After data cleaning, dynamic normalization processing is performed on each data source. The system automatically adjusts the normalization parameters according to the current data quality, so that each source data is converted to a unified standard. Subsequently, the overlapping data in the same spatial range is weighted and fused according to the quality index, and the preprocessing parameters are continuously optimized according to the data quality feedback to ensure the continuity and accuracy of the subsequent data analysis input. Next, the spatio-temporal weighted fuzzy clustering algorithm is used to divide the real-time collected multi-source data according to their spatial coordinates and collection time. In this process, by constructing the weighted spatio-temporal distance between the data points and the regional center, a regular objective function is formed, and the membership of each data point and the spatial and temporal coordinates of the regional center are continuously updated, so as to obtain adaptive and smooth regional division results. Based on the divided regional data, the embodiment further constructs a spatio-temporal prediction model to predict the future 1 hour of land utilization rate, environmental quality and traffic flow and other key indicators in each region. In the prediction process, multiple prediction models are used, and the model weights are adjusted in real time according to the recent prediction error and uncertainty. The final prediction value is output by weighted average integration. Subsequently, the territorial planning target is converted into a multi-objective optimization problem, and a comprehensive objective function is constructed, which includes economic cost, environmental protection and social benefit indicators, and the weights of each indicator are dynamically adjusted according to real-time feedback. The augmented Lagrangian method and meta-heuristic algorithm are used to realize real-time adaptive optimization of the planning scheme. After the implementation of the planning, the implementation area data is collected in real time and compared with the prediction results. The planning implementation effect is judged by error analysis, and the data and model parameters are automatically updated according to the deviation, forming a whole-process closed-loop feedback, realizing dynamic adjustment and continuous optimization of the planning scheme.

[0107] Table 1 experimental data table

[0108]

[0109] Through analysis of the test data table, it can be seen that each step of the embodiment has obvious innovation and superiority. First, in the data acquisition and preprocessing stage, the amount of collected data reaches more than 100,000 in each area, ensuring the sufficiency of the data; the accuracy of anomaly detection is more than 94%, and the completion rate of missing data is more than 90%, indicating that the processing of abnormal and missing data is mature. The error value after smoothing filtering is controlled within 1, and the normalized dynamic adjustment coefficient is basically between 0.65 and 0.74, showing that the normalization process has good adaptability; and the consistency of data fusion is as high as 97.5% to 98.3%, indicating that the quality of the fused data is very high and can provide a stable foundation for subsequent analysis. In addition, the real-time feedback response time is controlled between 12 and 20 seconds, which is much faster than the response speed of traditional planning methods, which provides sufficient time guarantee for dynamic adjustment of the planning scheme.

[0110] In the regional division stage, after using the spatiotemporal weighted fuzzy clustering, the internal spatiotemporal characteristics of the data of each region tend to be consistent, which is conducive to the subsequent fine spatiotemporal prediction. By iteratively updating the membership and regional center, the partition is smoothed and adapted, solving the defect of the existing method that only relies on spatial information and ignores the time dimension. The prediction model uses multi-model integration, and dynamically adjusts the weight according to the online error and uncertainty, so that the prediction result accuracy is significantly improved; this provides more scientific data support for land planning decision-making. Finally, in the multi-objective dynamic optimization planning link, multiple indicators such as economy, environment and society are considered through dynamic weights, and real-time adaptive optimization is carried out by using the augmented Lagrange and meta-heuristic algorithm, effectively making up for the problems of traditional methods in planning feedback and target imbalance. The overall closed-loop feedback mechanism enables the planning scheme to be continuously corrected and optimized during the implementation process, thereby realizing long-term stable and actual demand-conforming land space planning. Based on the above data analysis, the embodiment shows significant advantages over the prior art in data quality, regional division accuracy, prediction accuracy and response timeliness, proving the innovation and practical value of the invention in the field of land space planning optimization analysis.

[0111] Embodiment 3, refer to Figure 3 An embodiment of the present application provides a land space planning optimization analysis system based on big data, which comprises a real-time data acquisition and dynamic preprocessing module, a dynamic regional division module, a multi-model integrated spatiotemporal prediction module, a multi-objective dynamic optimization planning module, and a real-time feedback and adaptive adjustment module.

[0112] The real-time data acquisition and dynamic preprocessing module is used for acquiring data in real time and performing cleaning, normalization and fusion, the dynamic region division module is used for self-adaptive partitioning of the national space by using the space-time characteristics of the data, so that the data in the same region is similar, the multi-model integrated space-time prediction module is used for constructing a space-time prediction model for each region, predicting future land use, environment and traffic indicators by dynamic weight integration, the multi-objective dynamic optimization planning module is used for generating optimal planning decisions in real time according to the prediction results and a multi-objective optimization model of multiple planning objectives, and the real-time feedback and adaptive adjustment module is used for continuously monitoring the comparison between actual operation data and prediction values.

[0113] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the parts of the present application that essentially contribute to the prior art or the parts of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device) to execute all or part of the steps of the embodiments of the present application. The storage medium mentioned above includes a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk and various program code storage media.

[0114] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered a list of executable instructions for implementing logic functions, and can be specifically embodied in any computer-readable medium for use by an instruction execution system, apparatus or device, such as a computer-based system, a system including a processor or other system that can fetch and execute instructions from the instruction execution system, apparatus or device, or in conjunction with these instructions. For the purpose of this specification, the "computer-readable medium" can be any device that can contain, store, communicate, propagate or transport programs for use by an instruction execution system, apparatus or device, or in conjunction with these instruction execution systems, apparatus or devices.

[0115] More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection (electronic) having one or more wires, a portable computer diskette (magnetic), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can also be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, for example, via an optical scanner, then compiled, interpreted, or otherwise processed, and stored in a computer memory in a form that is then reproducible into a computer readable medium.

[0116] It should be understood that aspects of the application can be implemented in hardware, software, firmware or combinations thereof. In the above embodiments, various steps or methods can be implemented in software or firmware that is stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, and in another embodiment, any of the following techniques, which are well known in the art, can be used to implement the application: a hybrid of the techniques mentioned above; a combination of one or more of the techniques mentioned above; or one or more other techniques suitable for

[0117] It should be understood that aspects of the application can be implemented in hardware, software, firmware or combinations thereof. In the above embodiments, various steps or methods can be implemented in software or firmware that is stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, and in another embodiment, any of the following techniques, which are well known in the art, can be used to implement the application: a hybrid of the techniques mentioned above; a combination of one or more of the techniques mentioned above; or one or more other techniques suitable for

Claims

1. A method for optimizing land spatial planning based on big data, characterized in that, include: Spatiotemporal data is collected, stored through a sliding window, and subjected to anomaly detection, missing data completion, and smoothing filtering. The normalization parameters are then dynamically adjusted, overlapping data are fused according to weights, and the preprocessing is continuously optimized based on data quality feedback. Real-time collection of spatiotemporal data within the national territory: in, Represents spatiotemporal data, Indicates the total number of data points. The first Spatial coordinates of data points For the first The time required to collect data points; Divided into One region; Let each data point For the region The membership degree is ,area The spatial center is and time center ; Define data points With the region The spatiotemporal distance is expressed as: in, Representing data points With the region The combined spatiotemporal distance between them Representing data points spatial coordinates, It is the horizontal coordinate. It is the vertical axis. Indicates the region The spatial center is denoted as , Representing data points The time attribute, i.e., the time of data collection. Indicates the region The time center is usually the central tendency of the temporal distribution of data within a region. Represents a positive constant. This is a trade-off factor between time and spatial scales, used to balance the relative weights of time difference and spatial difference in comprehensive distance calculations. It is achieved by adjusting... The degree to which changing the time dimension affects the overall distance; The fuzzy membership degree and objective function are constructed as follows: in, Describe the objective function. This is a fuzzy index that controls the degree of fuzziness in the membership degree. Indicates the index of the data point. Indexes representing clusters or regions; The second term is the time regularization term. Adjust the continuity constraint to ensure that the cluster membership degree of data at the same location changes smoothly at consecutive time points; Update membership: in, This indicates an index that iterates through all regions. Representing data points Clustering The distance; In actual updates, the gradient of the regularization term is incorporated into the solution iteration; Update Center: in, Clustering The space center, Indicates membership degree by The power of a number is used as the weight. Clustering The time center reflects the central tendency of clustered data over time. Representing data points The collection time; By updating membership and regional centers online, a spatially and temporally adaptive and smooth regional division can be achieved; Spatiotemporal weighted fuzzy clustering is adopted to partition the real-time collected multi-source data according to spatial coordinates and time attributes. A regular objective function is constructed using the weighted distance between data points and the center of the region, and the membership degree and center position are updated iteratively. Construct a spatiotemporal prediction model, update the prediction error and uncertainty in real time, dynamically adjust the weights, and output the prediction results through weighted average integration. The land planning objectives are transformed into a multi-objective optimization problem. A comprehensive objective function is constructed that includes economic cost, environmental protection and social benefit indicators. The weights of each objective are dynamically adjusted based on real-time feedback. Augmented Lagrange and metaheuristic algorithms are used to adaptively optimize and dynamically adjust the planning scheme in real time. After the plan is implemented, real-time data is collected and compared with the pre-predicted values ​​for error analysis.

2. The method for optimizing land spatial planning based on big data as described in claim 1, characterized in that: The spatiotemporal data collected includes raw data obtained from remote sensing images, GIS maps, IoT sensors, socioeconomic statistics, and environmental monitoring data sources.

3. The method for optimizing land spatial planning based on big data as described in claim 2, characterized in that: The weighted fusion of overlapping data includes assigning different weights to each data source based on the stability and quality assessment of each data source, and using a weighted average method to fuse data at the same location and time.

4. The method for optimizing land spatial planning based on big data as described in claim 3, characterized in that: The iterative update of membership degree and center location includes dividing the spatiotemporal data collected in real time in the national territory, constructing a fuzzy clustering model with time regularization, and defining the spatiotemporal distance between data points and regions; To update the membership degree, the gradient of the regularization term is incorporated into the solution iteration during the actual update, and the center is updated accordingly. Regional divisions are performed both spatially and temporally by updating membership and regional centers online.

5. The method for optimizing land spatial planning based on big data as described in claim 4, characterized in that: The spatiotemporal prediction model is constructed, and the prediction error and uncertainty are updated in real time. The weights are dynamically adjusted, and the prediction results are output through weighted average integration, including future state predictions for indicators of each region. Construct a spatiotemporal prediction model, output the predicted values ​​of the spatiotemporal prediction model, integrate the spatiotemporal prediction model, dynamically adjust the weights of the spatiotemporal prediction model, and construct the weights of the spatiotemporal prediction model based on recent prediction errors. Considering the uncertainty of the spatiotemporal prediction model, error variance estimation is used to perform integrated prediction and define the dynamic weights of the spatiotemporal prediction model.

6. The method for optimizing land spatial planning based on big data as described in claim 5, characterized in that: The constructed objective function encompasses economic costs, environmental protection, and social benefits. Based on real-time feedback, the weights of each objective are dynamically adjusted. The augmented Lagrange and metaheuristic algorithms are used to adaptively optimize and dynamically adjust the planning scheme in real time, including multi-objective dynamic optimization of planning decision variables based on the prediction results of the spatiotemporal prediction model and the land planning objectives. A multi-objective optimization model is constructed, employing dynamic weighting to update the weights dynamically based on real-time feedback, and a comprehensive objective is defined. The dynamic optimization problem is output using an alternating update strategy.

7. The method for optimizing land spatial planning based on big data as described in claim 6, characterized in that: After the implementation plan is completed, real-time data is collected and compared with the pre-predicted values ​​for error analysis. This includes activating a feedback mechanism if the actual data deviates from the prediction by more than a set threshold, re-collecting the latest data and adjusting the spatiotemporal prediction model parameters, while dynamically optimizing the planning decision. Through continuous updates and closed-loop feedback, the planning scheme can be adjusted in real time, and the error feedback parameters can be adjusted according to the updated actual data.

8. A system employing the big data-based land spatial planning optimization analysis method as described in any one of claims 1 to 7, characterized in that: It includes a real-time data acquisition and dynamic preprocessing module, a dynamic region division module, a multi-model integrated spatiotemporal prediction module, a multi-objective dynamic optimization planning module, and a real-time feedback and adaptive adjustment module. The real-time data acquisition and dynamic preprocessing module is used to acquire data in real time and perform cleaning, normalization and fusion. The dynamic region division module is used to adaptively partition the national land space using the spatiotemporal characteristics of the data, so that the data in the same region is similar. The multi-model integrated spatiotemporal prediction module is used to construct spatiotemporal prediction models for each region and predict future land use, environmental and transportation indicators through dynamic weight integration. The multi-objective dynamic optimization planning module is used to generate the optimal planning decision in real time based on the prediction results and the multi-objective optimization model with multiple planning objectives. The real-time feedback and adaptive adjustment module is used to continuously monitor the comparison between actual operating data and predicted values.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the big data-based land spatial planning optimization analysis method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the big data-based land spatial planning optimization analysis method as described in any one of claims 1 to 7.

Images