Land resource planning management method and platform based on space-time big data

Abstract

The invention discloses a land resource planning management method and platform based on space-time big data, and relates to the technical field of land resource planning management, and the method comprises the steps: obtaining multi-source space-time big data, land reserve state variables and land actual supply data in a planning region, and carrying out the space-time weighted fusion of the multi-source space-time big data, constructing a flow space sensing matrix, and extracting a land demand feedforward operator from the flow space sensing matrix; based on the land demand feedforward operator and the land reserve state variable, generating an estimated land demand quantity, and performing difference operation on the estimated land demand quantity and the actual land supply data to generate reserve regulation and control deviation data; the method has the beneficial effects that the land vacancy rate is obviously reduced while the land price is ensured to be stable.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of land resource planning management, and particularly relates to a land resource planning management method and platform based on spatiotemporal big data. BACKGROUND

[0002] The present application relates to the technical field of land resource planning management, and particularly relates to a land resource planning management method and platform based on spatiotemporal big data.

[0003] However, the prior art has significant defects in actual application: its regulation model is often based on linear or idealized supply-demand relationship, and it is difficult to cope with complex and variable external environmental interference. Since land resource planning involves macroeconomic fluctuations, policy adjustments, and market speculation behaviors, and other nonlinear factors with strong randomness, the prior art lacks the ability to identify and separate these external noises, resulting in planning instructions generated with huge lag and deviation, which cannot real-time offset the influence of external disturbance on land supply-demand balance, causing a serious disconnection between land supply plan and actual demand, and affecting the scientificity of land resource allocation and the smooth operation of social economy. SUMMARY

[0004] In view of the above prior art, the present application is proposed. The embodiments of the present application provide a land resource planning management method and platform based on spatiotemporal big data, which can real-time offset the influence of external disturbance on land supply-demand balance, and guarantee the scientificity of land resource allocation and the smooth operation of social economy.

[0005] According to one aspect of the present application, a land resource planning management method based on spatiotemporal big data is provided, comprising:

[0006] Obtaining multi-source spatiotemporal big data, land reserve state variables and land actual supply data in the current period planning area, constructing a flow space perception matrix by spatiotemporal weighted fusion of the multi-source spatiotemporal big data, and extracting a land demand feedforward operator therefrom;

[0007] generate a reserve control deviation data by performing a difference operation on the land estimated demand and the land actual supply data;

[0008] input the reserve control deviation data and a land supply decision instruction of a previous decision cycle into a preset extended state observer, estimate and separate a nonlinear comprehensive disturbance quantity in real time based on a preset gain parameter through the extended state observer;

[0009] compensate and correct the reserve control deviation data by using the land demand feedforward operator, and offset the disturbance of the reserve control deviation data by using the nonlinear comprehensive disturbance quantity, to synthesize a land supply decision instruction of a current cycle;

[0010] feed back land price fluctuation rate and land idle rate after execution, and correct fusion weights of the flow space perception matrix and the preset gain parameter respectively by using the feedback data; and

[0011] use the corrected fusion weights and gain parameters as fusion weights and preset gain parameters of a flow space perception matrix of a next cycle.

[0012] According to another aspect of the present application, a land resource planning and management platform based on spatiotemporal big data is provided, comprising:

[0013] a flow space perception module for acquiring multi-source spatiotemporal big data, land reserve state variables and land actual supply data in a planning area of a current cycle, constructing a flow space perception matrix by performing spatiotemporal weighted fusion on the multi-source spatiotemporal big data, and extracting a land demand feedforward operator therefrom;

[0014] a reserve deviation calculation module for generating a land estimated demand based on the land demand feedforward operator and the land reserve state variables, and performing a difference operation on the land estimated demand and the land actual supply data to generate a reserve control deviation data;

[0015] a disturbance observation module for inputting the reserve control deviation data and a land supply decision instruction of a previous decision cycle into a preset extended state observer, estimating and separating a nonlinear comprehensive disturbance quantity in real time based on a preset gain parameter through the extended state observer;

[0016] a decision synthesis module for compensating and correcting the reserve control deviation data by using the land demand feedforward operator, and offsetting the disturbance of the reserve control deviation data by using the nonlinear comprehensive disturbance quantity, to synthesize a land supply decision instruction of a current cycle;

[0017] An adaptive correction module is configured to send the land supply decision instruction to a land reserve execution terminal, collect feedback data of land price fluctuation rate and land idle rate after execution in real time, correct the fusion weight of the flow space perception matrix and the preset gain parameter respectively by using the feedback data, and use the corrected fusion weight and gain parameter as the fusion weight and preset gain parameter of the flow space perception matrix in the next period.

[0018] According to another aspect of the present application, an electronic device is provided, comprising a memory for storing computer executable instructions and a processor for executing the computer executable instructions, which, when executed by the processor, implement the steps of the method described above.

[0019] According to another aspect of the present application, a computer storage medium is provided, which stores computer executable instructions, which, when executed by a processor, implement the steps of the method described above.

[0020] Compared with the prior art, the land resource planning and management method and platform based on spatiotemporal big data according to the embodiments of the present application can capture potential land demand caused by instantaneous changes such as population flow and traffic flow, and make "predictive compensation" before deviation is formed, thereby greatly improving the response speed of regulation and control. The present application does not need to model the interference, but estimates and offsets it in real time as "total disturbance", so that the decision instruction can penetrate the market "noise" and effectively reduce the abnormal fluctuation rate of land price. When the environment changes in trend, the perception strategy and observation intensity can be automatically optimized, the dependence on expert experience is reduced, the long-term stability of the regulation and control model in different urban areas and different economic cycles is ensured, the reserve regulation and control deviation data are generated by difference operation and disturbance offset, the land idle rate is significantly reduced while ensuring the stability of land price. BRIEF DESCRIPTION OF DRAWINGS

[0021] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description of embodiments of the present application taken in conjunction with the accompanying drawings. The accompanying drawings are intended to provide a further understanding of embodiments of the present application and are incorporated in and constitute a part of this specification, illustrate embodiments of the present application and serve to explain the present application, but do not limit the present application. In the drawings, the same reference numerals refer to the same components or steps throughout the drawings.

[0022] Figure 1 FIG. 1 is a schematic diagram of the overall flow of the land resource planning and management method based on spatiotemporal big data of the present application.

[0023] Figure 2 FIG. 2 is a schematic diagram of the land supply decision instruction correction of the land resource planning and management method based on spatiotemporal big data of the present application.

[0024] Figure 3 A flow space perception matrix construction schematic diagram of the land resource planning management method based on spatiotemporal big data of the application. DETAILED DESCRIPTION

[0025] Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments described herein.

[0026] In the prior art, land resource planning management has experienced evolution from manual experience discrimination to geographic information system assisted decision making, but still faces significant challenges in dealing with multi-dimensional dynamic changes in demand scenarios in the process of urbanization. For the land market affected by macro policy adjustment, market psychological expectation and economic environment change, the supply and demand balance of the land market shows strong nonlinearity and random disturbance. The traditional method based on linear growth prediction or static index allocation is difficult to identify and separate these external noises hidden in the data, resulting in serious lag and deviation of the generated land supply instructions, causing abnormal fluctuations in land prices or land resource idling, affecting the precision of urban governance and the efficiency of resource allocation.

[0027] In order to solve the above problems, the applicant found that the core of land resource planning imbalance lies in the lack of implicit disturbance perception and the open loop control property of the regulation system. Through analysis, it is found that the dynamic data such as population flow and material exchange in multi-source spatiotemporal big data have a significant leading role and can be used as a compensation basis for predicting the instantaneous changes of land demand. At the same time, although the fluctuations of external policies and markets are difficult to model directly, their impact on the system can be estimated inversely by observing the system state deviation.

[0028] Based on this, a technical route combining flow space perception matrix construction, extended state observer nonlinear disturbance separation and closed loop adaptive compensation is proposed: first, the flow space perception matrix is constructed by spatiotemporal weighted fusion to extract the feedforward operator, and a predictive compensation mechanism is established; secondly, the extended state observer is introduced to estimate and separate the nonlinear comprehensive disturbance caused by policy and economic changes in real time, and the disturbance is offset by reverse superposition; finally, the perception weight and observation parameter are dynamically corrected by the feedback data of land price and idle rate. Through the deep coupling of "feedforward + feedback", the nonlinear deviation problem of land resource planning in complex environment is solved, the high synchronization of land supply instruction and actual demand is realized, and the stable operation of land market is ensured.

[0029] Embodiment one:

[0030] Reference Figures 1-3For an embodiment of the present application, a land resource planning management method based on spatio-temporal big data is provided:

[0031] Figure 1 The land resource planning management method based on spatio-temporal big data according to the embodiment of the present application is illustrated, which comprises:

[0032] The multi-source spatio-temporal big data in the current period planning area, land reserve state variables and land actual supply data are acquired, the spatio-temporal weighted fusion is performed on the multi-source spatio-temporal big data, the flow space perception matrix is constructed, and the land demand feedforward operator is extracted therefrom.

[0033] In the present embodiment, the multi-source spatio-temporal big data specifically covers the population moving track data, the material exchange logistics data and the real-time traffic density data of the traffic network in the planning area within a preset period; the land reserve state variables include the location attribute, the development degree and the corresponding ownership restriction information of the current collected land plot; the land actual supply data reflects the land area, the transaction price and the use classification of the land that has been put into market circulation within the current decision window; after the above basic data is acquired, the spatio-temporal weighted fusion is performed on the multi-source spatio-temporal big data, and the flow space perception matrix is constructed; specifically, the management platform divides the planning area into a plurality of standardized space grid units, extracts the interactive features of each space grid unit in time sequence, and uses the spatio-temporal weighted algorithm to distribute the weights according to the timeliness and the spatial correlation coefficient of different types of spatio-temporal big data, so as to map the scattered point-shaped moving track, the logistics flow and the traffic density to a unified grid interactive coordinate system, thereby generating the flow space perception matrix reflecting the spatial correlation intensity.

[0034] It should be noted that the land demand feedforward operator is extracted from the flow space perception matrix; this process identifies the characteristic vector reflecting the regional activity and the resource demand directionality by performing the characteristic decomposition on the flow space perception matrix, converts the characteristic vector into the land demand feedforward operator representing the instantaneous change of land demand, and the feedforward operator is used to describe the advanced demand variable of the planning area relative to the administrative planning index due to the fluctuation of social and economic activities; through the above processing, the high-dimensional dynamic spatio-temporal flow data is converted into the quantifiable control parameter with the predictive compensation function, thereby providing a real-time data foundation for the subsequent dynamic regulation and control of land reserve.

[0035] Based on the land demand feedforward operator and the land reserve state variable, the land estimated demand is generated, and the specific formula is as follows:

[0036] ;

[0037] Among them, the land estimated demand (m2), the base land demand (m2) determined based on the land reserve state variable, is a land demand feed-forward operator, representing the demand response time constant (unit: d), is a socio-economic activity rate of change, representing the land demand area variation rate (m2 / d) in the planning area due to population and material flow within a unit time derived from the flow space perception matrix;

[0038] The estimated land demand and the actual land supply data are then subtracted to generate reserve control deviation data, and the specific formula is as follows:

[0039]

[0040] wherein, is the reserve control deviation data, is the actual land supply data, representing the actual land supply area that has been issued and executed by the management platform in the current decision-making period; the reserve control deviation data calculated can accurately quantify the physical gap between supply and demand in the market, and the management platform transmits this reserve control deviation data as a real-time input signal to the subsequent extended state observer, and through the observer, the deviation component caused by policy fluctuations is separated, thereby providing a pure error input basis for the synthesis of the final decision instruction.

[0041] The reserve control deviation data and the land supply decision instruction of the previous decision-making period stored in advance are input into the preset extended state observer synchronously;

[0042] In the embodiments of the present application, it is necessary to clearly define that for a continuously running management system or platform, the land supply decision instruction of the previous decision-making period is derived from the instruction value recorded in the system internal memory, which has been actually issued and executed in the previous period. For the extreme case where the management platform or system is in the initial start state (i.e., the first decision-making period), since the management system has not yet generated historical execution instructions at this time, the management platform pre-sets an initial instruction value as a substitute input. Specifically, the initial instruction value is initialized and set by the management platform according to the historical average annual land supply plan of the planning area or the current administrative guidance area;

[0043] By introducing this initialization mechanism, even in the first decision-making period where the actual measurement data of the previous decision-making period is missing, the extended state observer can still obtain continuous control gain input items. The management platform uses this initial anchor point in combination with the real-time obtained reserve control deviation data, so that the extended state observer can quickly converge within a very short time (usually within the first 10% of the time points of the first decision-making period) after the system starts. This means that the current decision instruction generated by the management platform in the first period will serve as the cornerstone for the system to enter normal regulation and control, and will be stored as the "land supply decision instruction of the previous decision-making period" for the next decision-making period.​

[0044] The nonlinear composite disturbance is estimated and separated in real time using an extended state observer based on a preset gain parameter, as shown in the following formula:

[0045] ;

[0046] in, To reserve the observed estimate of the control deviation, representing the real-time tracking value of the deviation area, The measurement error represents the difference (m²) between the estimated area and the measured area. express The derivative with respect to time represents the rate of change of the observed area (m² / d). For internal system state variables, represents the velocity term (m² / d) of the supply-demand gap evolution. , representing the acceleration term of the evolution of the supply-demand gap ( ), This is a nonlinear composite disturbance quantity, representing the equivalent impact of an unknown disturbance on the acceleration level of the system. ), for The derivative with respect to time represents the rate of change of the disturbance ( ), This indicates the land supply decision instructions from the previous decision-making cycle. ), , and This is a preset gain parameter, which physically represents the observer bandwidth adjustment factor. Units are , Units are , Units are b is the control gain coefficient, representing the weight of the input command's influence on the system acceleration. );

[0047] Through real-time iteration of the above equations, utilizing , and The dimensional transfer effect transforms the observation error e into corrective signals at the velocity and acceleration levels step by step. Ultimately, the management platform extracts the nonlinear comprehensive disturbance from the expanded state. Due to this nonlinear comprehensive disturbance The dimensions are precisely defined as the acceleration term ( When the management platform synthesizes decision instructions in the future, it can convert the comprehensive acceleration disturbance into an area compensation amount based on the time span of the control cycle, thereby achieving the offsetting of disturbances to the reserve control deviation data.

[0048] The land demand feedforward operator is used to compensate and correct the reserve regulation deviation data, and the nonlinear comprehensive disturbance quantity is used to disturb and offset the reserve regulation deviation data, and the land supply decision instruction of the current period is synthesized.

[0049] Specifically, the disturbance offsetting includes:

[0050] The nonlinear comprehensive disturbance quantity is converted into an equivalent compensation control quantity, and the specific formula is as follows:

[0051] ;

[0052] is the equivalent compensation control quantity, which represents the hedging land area needed to hedge against external disturbances, is the regulation period of the land supply decision instruction;

[0053] According to the directionality characteristics of the time-varying gain coefficient, the nonlinear comprehensive disturbance quantity is subjected to piecewise linearization processing to generate a dynamic compensation item of the same dimension as the reserve regulation deviation data.

[0054] Subsequently, the management platform extracts the time-varying gain coefficient of the equivalent compensation control quantity relative to the reserve regulation deviation data, and the management platform calculates the coefficient in real time to represent the correlation characteristics of disturbance intensity and the current supply-demand gap. According to the directionality characteristics of the time-varying gain coefficient, i.e., whether the disturbance aggravates or alleviates the supply-demand deviation, the management platform performs piecewise linearization processing on the nonlinear comprehensive disturbance quantity; during the piecewise linearization processing, the management platform predefines different regulation thresholds, and when the regulation threshold is exceeded, a high-order compensation function is used, and when the threshold is lower, a smooth linear function is used, thereby generating a dynamic compensation item of the same dimension as the reserve regulation deviation data ( ), which represents the modified area that can be directly used to hedge against external shocks after linearization and smoothing.

[0055] The dynamic compensation item is reversely superimposed on the reserve regulation deviation data to correct the reserve regulation deviation data, and the specific formula is as follows:

[0056] ;

[0057] wherein, is the corrected reserve regulation deviation data;

[0058] Based on the corrected reserve regulation deviation data , combined with the modified results of the feedforward compensation operator, the final land supply decision instruction of the current period is synthesized by weighting, which is delivered to the land reserve execution terminal in the form of a digitalized land parcel list, and through the above-mentioned segmented linearization processing and reverse superposition, the precise hedging of nonlinear fluctuations such as policy and economic changes is realized, and it is ensured that the final land supply decision instruction delivered can truly reflect the supply and demand balance state after being controlled.

[0059] The land supply decision instruction is delivered to the land reserve execution terminal, and the feedback data of the land price fluctuation rate and the land idle rate after execution is collected in real time, and the feedback data is used to modify the fusion weight of the flow space perception matrix and the preset gain parameter.

[0060] Specifically, the modification of the fusion weight and the preset gain parameter includes:

[0061] The deviation value of the land price fluctuation rate and the land idle rate in the feedback data relative to the preset target is calculated, and the distribution density of the deviation value in the spatial dimension is analyzed to generate the deviation spatio-temporal distribution feature quantity, and the specific formula is as follows:

[0062] ;

[0063] Wherein, is the deviation spatio-temporal distribution feature quantity, the comprehensive feedback deviation intensity in the unit space grid, is the land price fluctuation rate deviation of the i-th sampling point, indicating the difference between the measured land price fluctuation rate and the target fluctuation rate, is the land idle rate deviation of the i-th sampling point, indicating the difference between the measured idle rate and the preset red line value, is the area of the grid unit;

[0064] The management platform identifies the influence factor mapping relationship of each dimension data in the multi-source spatio-temporal big data on the land use deviation through the deviation spatio-temporal distribution feature quantity The management platform uses the partial correlation analysis technology to judge whether the prediction inaccuracy is caused by population flow, logistics flow or traffic flow data, and adjusts the fusion weight according to the size of the influence factor mapping relationship. If the influence factor mapping relationship of a certain dimension data is larger, the management platform will correspondingly increase the weight proportion of this dimension in the flow space perception matrix;

[0065] The time sequence change rate of the feedback data is calculated, and the correction step is determined based on the amplitude of the time sequence change rate. The preset gain parameter is modified in real time by using the correction step, and the specific formula is as follows:

[0066] ;

[0067] Wherein, is the modified preset gain parameter, is a preset gain parameter before correction, is a correction step, indicating the gain adjustment intensity under a unit change rate, is a time series change rate of land price fluctuation rate deviation;

[0068] using the correction step The bandwidth of the expansion state observer is adjusted in real time. When the time series change rate amplitude increases due to the dramatic change of the market environment, the management platform automatically increases the correction intensity and increases the value of the preset gain parameter, thereby enhancing the tracking sensitivity of the expansion state observer to the nonlinear comprehensive disturbance quantity. Through this closed-loop feedback correction, the management platform ensures that the flow space perception matrix and the regulation parameter can be iterated with the real-time evolution of the land market.

[0069] The feedback correction mechanism based on land price fluctuation rate and land idle rate is introduced in the implementation of the present application because in the actual regulation process, the land reserve system is not a linear system isolated from the outside world, and it has the characteristics of dynamic drift under the influence of the external macro environment. If only fixed fusion weights and gain parameters are used, the system will gradually lose the perception accuracy of the latest market pulse. Therefore, the present application learns from the parameter identification logic in the adaptive control theory, uses the results after execution as an "error probe" to drive the self-optimization of the regulation algorithm in reverse, thereby realizing the logical closed-loop stability.

[0070] By correcting the fusion weight, the management platform can automatically identify which spatio-temporal data (such as population flow or capital flow) has more significant influence on land price and idle in the current period, and realize dynamic focusing of the perception dimension. By correcting the preset gain parameter, the management platform can automatically adjust the "sensitivity" of the observer according to the deviation degree of the regulation effect, increase the regulation intensity when the market fluctuates dramatically, and converge the parameters in the stable period to avoid over-regulation, thereby realizing flexible switching of the control intensity.

[0071] In the prior art, alternative solutions usually include periodic manual expert evaluation correction or offline parameter adjustment based on regression analysis. However, manual correction has serious lag and is difficult to respond to market pulses on a weekly or even daily basis. Offline parameter adjustment cannot handle instantaneous random disturbances in non-stationary processes. In contrast, the real-time feedback correction mechanism of the present application has significant advantages in response speed and decision continuity.

[0072] In order to further illustrate the actual process of data processing by the management platform, the following will be exemplarily disclosed in combination with a set of consecutive numerical examples:

[0073] Suppose in the current decision-making period , the management platform presets the fusion weight of "population flow data" in the flow space perception matrix to 0.6, the fusion weight of "traffic flow data" to 0.4, and the preset gain parameter of the expansion state observer is 100, the land price fluctuation rate collected in real time is 8% (the preset target is ) and the land idle rate is 12% (the preset target is 5%);

[0074] The management platform identifies that the land price fluctuation rate deviation value is and the land idle rate deviation value is ; in the spatial dimension analysis, the management platform finds that the grid units with violent land price fluctuation are highly coincident with the traffic hub nodes, and the correlation degree with the population density distribution is reduced, based on the mapping relationship, the management platform triggers the fusion weight correction logic, adjusts the weight of the "traffic flow data" from 0.4 to 0.55, and correspondingly adjusts the weight of the "population flow data" to 0.45, so that the flow space perception matrix is more suitable for the current physical influence factor;

[0075] At the same time, the management platform calculates that the time sequence change rate of the feedback data is (namely, the deviation still has an expansion trend in a short time), the management platform determines the correction step based on the amplitude as , the preset gain parameter is compensated and corrected from 100 to 115 in real time, so as to improve the tracking speed of the expansion state observer to the nonlinear comprehensive disturbance, and after entering the next decision cycle , the management platform recalculates by using the updated weight and gain parameter, so that the land price fluctuation rate falls to 3.5% and the land idle rate falls to 6%, thereby proving that the closed-loop correction operation can be self-iterated with the real-time evolution of the land market.

[0076] Figure 2 is a land supply decision instruction correction schematic diagram of the land resource planning management method based on space-time big data of the application;

[0077] The application further proposes that the management method further comprises:

[0078] Before the land supply decision instruction is delivered to the land reserve execution terminal, a plurality of alternative land supply schemes are generated based on the land supply decision instruction of the current period;

[0079] After synthesizing the land supply decision-making instruction of the current period, the next action is not executed immediately, but in the "decision-making audit window period" before the land supply decision-making instruction is issued to the land reserve execution terminal, a plurality of alternative land supply schemes are generated based on the land supply decision-making instruction of the current period through a combinatorial optimization algorithm; the significance of setting this time node is that the virtual environment constructed by the management platform conducts a "stress test" on the preliminary decision, and through simulation, the potential negative social and economic response is predicted before the instruction is formally issued, which can avoid the sharp fluctuations in the land market caused by direct execution of the instruction, and the combinatorial optimization algorithm is used to generate a plurality of alternative land supply schemes that differ in spatial layout or development intensity, with the initially generated land supply decision-making instruction as the benchmark center, within the parameter space of the preset FAR upper and lower limits, land price guidance interval and land use ratio, and through multi-objective genetic algorithm optimization, a plurality of alternative land supply schemes that differ in spatial layout or development intensity are generated, ensuring that the decision set covers different risk preferences;

[0080] For each land supply scheme, based on historical multi-source spatio-temporal big data, the future land price volatility rate, land idle rate and traffic flow change rate that will be triggered in the planning area after the implementation of each scheme are predicted;

[0081] Specifically, the coordinates of the land blocks, the FAR parameters and the land use classification in each land supply scheme are extracted and mapped as perturbation variables into the spatio-temporal evolution topology graph constructed by historical multi-source spatio-temporal big data; the spatio-temporal evolution topology graph is used to analyze the spatio-temporal correlation influence domain of the perturbation variables on adjacent land blocks;

[0082] Specifically, the spatio-temporal evolution topology graph constructed by the management platform takes historical geographic units as nodes and the flow interaction (population, capital, logistics) between units as edge weights. The management platform injects the perturbation variables (such as increasing the FAR of a certain land block from 1.5 to 3.0) in the scheme into the graph as attribute increments of the topology nodes. Using the spatio-temporal evolution topology graph, the management platform analyzes the spatio-temporal correlation influence domain of the perturbation variables on adjacent land blocks. Specifically, the management platform uses the Laplacian matrix of the topology graph to calculate the diffusion distribution of the perturbation energy in the graph, and locks the set of adjacent geographic nodes most affected by identifying the evolution of the weight of the associated edges;

[0083] Then, the social and economic characteristic variables caused by the change of land properties in the spatio-temporal correlation influence domain are calculated, and the specific formula is as follows:

[0084] ;

[0085] Among them, is the social and economic characteristic variable, which represents the relative change rate of local area activity after the implementation of the land supply scheme, and respectively represent the scheme preset volume rate and the historical average volume rate of the region, and respectively represent the scheme planning area and the historical benchmark supply area, and respectively represent the current real-time traffic characteristic value of the associated node j and the historical average value, represent the associated weight of node j and the disturbed plot, is the associated influence domain set;

[0086] According to the social and economic characteristic evolution variable, the corresponding future land price fluctuation rate, land idle rate and traffic flow change rate after the implementation of each scheme are synchronously derived through the preset cross-dimension association operator, and the specific formula is as follows:

[0087] ;

[0088] wherein, respectively represent the future land price fluctuation rate, land idle rate and traffic flow change rate, represent the net change rate of the social and economic characteristics, represent the preset cross-dimension association operator.

[0089] Based on the comparison between the future land price fluctuation rate, land idle rate and traffic flow change rate and the corresponding preset target threshold, the instruction pre-optimization parameter is generated. Specifically, after the corresponding future land price fluctuation rate, land idle rate and traffic flow change rate after the implementation of each scheme are synchronously derived, these prediction indexes are compared with the corresponding preset target threshold to generate the instruction pre-optimization parameter. The preset target threshold is defined by the management platform as the boundary condition for the stable operation of the land market,

[0090] The specific formula is as follows:

[0091] ;

[0092] wherein, represent the instruction pre-optimization parameter, represent the prediction index value, including the future land price fluctuation rate, land idle rate and traffic flow change rate, represent the corresponding preset target threshold, represent the sensitivity weight of each index, used to balance the importance of different control targets, P represents the land price dimension, R represents the land use efficiency dimension, and T represents the land use efficiency dimension. By weighted summation of the relative deviations of the three dimensions, industry indexes with different physical meanings can be converted into instruction pre-optimization parameters in a unified dimension, is the evaluation index dimension index, is ;

[0093] The instruction pre-optimization parameter is used to correct the land supply decision instruction of the current period. The parameter is used to fine-tune the supply area, transfer time or upper limit of the volume rate in the preliminary instruction, so that the corrected scheme can maximize the approximation to the regulation target. The specific formula is as follows:

[0094] ;

[0095] For the corrected land supply decision instruction, represents the land supply decision instruction of the current period;

[0096] The corrected land supply decision instruction is delivered to the land reserve execution terminal. The corrected land supply decision instruction is delivered to the land reserve execution terminal. By introducing this layer of pre-optimization correction based on the predicted index before delivery, the management platform can offset the potential planning deviation caused by the spatial correlation in advance. This mechanism ensures that each instruction received by the execution terminal has undergone a closed-loop verification of "simulation-comparison-correction", thereby improving the accuracy and forward-looking of land resource planning management from the source.

[0097] In the embodiments of the present application, the above-mentioned pre-optimization operation based on "simulation-comparison-correction" is introduced because in traditional land management, decisions are often based only on the current supply-demand gap, ignoring the physical and socio-economic chain reactions to surrounding plots after the decision instruction is issued. In order to control the risk before execution, the management platform refers to the "model predictive control (MPC)" idea in control engineering, and through the construction of a digital twin spatio-temporal topology map, multiple scheme evolution simulations are performed in the "review window period" of instruction issuance, thereby realizing the technical leap from "experience decision" to "simulation-driven decision";

[0098] By perturbing and sampling the volume rate, use, and other parameters, the most insensitive and robust execution scheme to the market can be identified, improving the robustness of the decision. By using the correlation operator to unify land price, idle rate, and traffic flow, the "see-saw effect" of single-target regulation leading to deterioration of other dimension indicators is avoided, and cross-dimension collaborative optimization is achieved;

[0099] In the prior art, alternative schemes usually include empirical coefficient correction based on historical statistical rules, or scheme evaluation through a simple expert review system. However, these alternative technologies cannot quantitatively consider the topological correlation between plots, and cannot provide automatic and closed-loop parameter compensation for nonlinear comprehensive disturbance;

[0100] For example, assume that the land supply decision instruction synthesized by the management platform at present is , the initial preset volume rate is 2.0, the management platform generates an alternative plan in the decision review window, which plans an area is , the preset volume rate is 2.2, the associated weight of a key node j in the adjacent influence domain is 0.8, the current traffic characteristic value of the node collected in real time is , and the historical average value is ;

[0101] ;

[0102] After the implementation of the plan, the local area activity level is increased by about 10.88% compared with the historical level; assuming that the component of the land price dimension in the cross-dimension association operator Γ is 0.5 (representing the sensitivity coefficient of the net increase of activity level to land price fluctuation), then the predicted value of future land price fluctuation rate ΔP = 0.5 × (ΔSec−1) = 0.5 × (1.1088−1) = 0.0544 (i.e. 5.44%), and similarly, assuming that the predicted future land idle rate is 4.2% (i.e. 0.042), and the future traffic flow change rate is 6.5% (i.e. 0.065);

[0103] The management platform sets the land price fluctuation rate target threshold as 3%, the land idle rate target threshold as 5%, and the traffic flow target threshold as 6%, and sets the sensitivity weights of the three as 0.5, 0.3, and 0.2, respectively;

[0104] ;

[0105] ;

[0106] ;

[0107] The original instruction is modified by the instruction pre-optimization parameter σ;

[0108] ;

[0109] Through the above rigorous numerical evolution, the management platform identifies that if the supply of 105,000 m2 is implemented as planned, it will cause the land price fluctuation to be greatly exceeded, so the final supply area is modified to 62,480 m2 by the instruction pre-optimization parameter, and the land supply decision instruction after the modification is issued, thereby avoiding the market risk caused by blind expansion at the physical source. ​

[0110] Figure 3 A flow space perception matrix construction schematic diagram of the land resource planning management method based on spatiotemporal big data of the application;

[0111] The application further proposes the construction of the flow space perception matrix, which further comprises:

[0112] The planning area is divided into a plurality of space grid units;

[0113] Specifically, in the construction of the flow space perception matrix, first, spatial gridding processing is performed, and according to the administrative boundary and geographical features of the planning area, the planning area is divided into a plurality of standardized space grid units by using equidistant sampling method or Hilbert curve indexing method; for example, in a specific city regulation scene, the length of each space grid unit is set to 100 m x 100 m, that is, a rectangular area with a size of 100 m x 100 m, and such a size can balance the sampling density and calculation efficiency of multi-source spatiotemporal big data; the management platform allocates a unique space coordinate ID to each space grid unit as a minimum logical node for subsequent spatial aggregation calculation of multi-source data (such as mobile phone signaling points, logistics unloading points, etc.).

[0114] Based on multi-source spatiotemporal big data, the correlation strength between each space grid unit is identified, wherein in order to make the correlation strength truly reflect the limiting effect of the physical environment on social and economic flow, a geographical constraint factor is introduced in the calculation of the correlation strength, and the geographical constraint factor is determined according to at least one of the topographic connectivity, planning function compatibility and traffic accessibility between the space grid units;

[0115] Specifically, the introduction of the geographical constraint factor comprises:

[0116] The topographic slope change rate, the legal planning land class connection relationship and the road travel time cost between the space grid units are weighted and mapped to generate a geographical damping term representing the degree of space obstruction, and the specific formula is as follows:

[0117] ;

[0118] wherein, is the geographical damping term, representing the comprehensive resistance coefficient when interacting between space units, is the topographic slope change rate between two grid units, is the maximum slope threshold value of the planning area, is the legal planning land class connection relationship, and the physical meaning is the compatibility score of the use of adjacent land blocks, is the actual road travel time cost between the grids, is the theoretical straight-line travel time benchmark between the two grids, , and ​are preset weight coefficients, and ;

[0119] The interaction features between each spatial grid unit are calculated by using a geographic damping term to attenuate, and a geographic correction correlation quantity is obtained, and the specific formula is as follows:

[0120] ;

[0121] Wherein, is a geographic correction correlation quantity, is the original interaction frequency extracted and aggregated to the grid unit by the multi-source spatio-temporal big data;

[0122] The correlation strength with data flow and physical space constraint characteristics is calculated based on the geographic correction correlation quantity;

[0123] Subsequently, the management platform uses the correlation strength after introducing the geographic constraint factor to perform spatio-temporal weighted fusion on the multi-source spatio-temporal big data, and constructs an optimized flow space perception matrix; in the matrix, the management platform defines the spatial coordinate ID of each spatial grid unit as the row and column index of the matrix, realizes the adaptive correction of the complex geographic obstruction information into the matrix value, and extracts the land demand feedforward operator by using the optimized flow space perception matrix, which can fundamentally filter out the invalid flow correlation caused by mountain barriers or planning prohibited areas, and ensure that the feedforward operator can truly reflect the potential land development demand limited by the physical landscape.

[0124] In the embodiments of the present application, the above-mentioned flow space perception matrix optimization operation based on the geographic constraint factor is introduced because in the traditional flow space analysis, only the social and economic element flow frequency mapped by the multi-source spatio-temporal big data is often concerned, and the physical constraints of geographic environment (such as mountain barriers), legal planning (such as red line restrictions) and traffic infrastructure on these flows are ignored. In order to eliminate the "false correlation" at the pure data level, the management platform refers to the "damping motion" principle in physics, maps the geographic obstacles to spatial damping, so that the data flow characteristics can be more truly mapped to the substantive demand for land resources after being filtered by the physical space.

[0125] By the correction of terrain slope and planning compatibility, areas that cannot be developed in substance due to physical or legal restrictions although data interaction is frequent can be effectively identified and removed, the geographical fidelity of land demand prediction is improved, the real accessibility of urban space structure is more accurately described by using road travel time cost to calculate the attenuation of original flow, the robustness of the perception matrix is enhanced, by defining the spatial coordinate ID of each spatial grid unit as the row and column index of the matrix, not only the storage structure of multi-source heterogeneous data is logically unified, but also the interaction relationship between each group of grids is fine-tuned at the operation level. The optimized flow space perception matrix makes the subsequently extracted land demand feedforward operator have high spatial resolution and environmental perception, and lays a solid data foundation for land reserve regulation and control according to local conditions.

[0126] Embodiment two:

[0127] For an embodiment of the application, it is different from the previous embodiment:

[0128] The land resource planning and management platform based on spatiotemporal big data comprises:

[0129] The flow space perception module is used for acquiring multi-source spatiotemporal big data in the current period planning area, land reserve state variables and land actual supply data, constructing a flow space perception matrix by spatiotemporal weighted fusion of the multi-source spatiotemporal big data, and extracting a land demand feedforward operator therefrom;

[0130] The reserve deviation calculation module is used for generating a land estimated demand based on the land demand feedforward operator and the land reserve state variables, and performing difference operation on the land estimated demand and the land actual supply data to generate reserve regulation deviation data;

[0131] The disturbance observation module is used for synchronously inputting the reserve regulation deviation data and a land supply decision instruction of a previous decision period stored in advance into a preset extended state observer, and estimating and separating a nonlinear comprehensive disturbance amount in real time based on a preset gain parameter through the extended state observer;

[0132] The decision synthesis module is used for compensating and correcting the reserve regulation deviation data by using the land demand feedforward operator, and offsetting the reserve regulation deviation data by using the nonlinear comprehensive disturbance amount, to synthesize a land supply decision instruction of the current period;

[0133] The adaptive correction module is configured to deliver the land supply decision instruction to a land reserve execution terminal, collect feedback data of land price fluctuation rate and land idle rate after execution in real time, correct the fusion weight of the flow space perception matrix and the preset gain parameter respectively by using the feedback data, and use the corrected fusion weight and gain parameter as the fusion weight and preset gain parameter of the flow space perception matrix in the next period.

[0134] Embodiment three:

[0135] For an embodiment of the present application, which is different from the previous embodiment, the electronic device includes one or more processors and a memory.

[0136] The processor can be a central processing unit (CPU) or other form of processing unit having data processing and / or instruction execution capabilities, and can control other components in the electronic device to perform desired functions.

[0137] The memory can include one or more computer program products, which can include various forms of computer readable storage media, such as volatile memory and / or non-volatile memory. Volatile memory may, for example, include random access memory (RAM), cache memory, and / or the like. Non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, and / or the like.

[0138] In one example, the electronic device can further include an input device and an output device, which are interconnected by a bus system and / or other forms of connection mechanism (not shown). In addition, the electronic device can also include any other appropriate components according to specific application cases.

[0139] Embodiment four:

[0140] Embodiments of the present application can also be computer readable storage media having computer program instructions stored thereon, which, when executed by a processor, cause the processor to perform the steps described in the above “Exemplary Method” section of the specification according to various embodiments of the present application.

[0141] The computer readable storage medium can include any combination of one or more non-transitory media. The non-transitory medium can be a non-transitory storage medium or a non-transitory signal medium. The non-transitory storage medium can include, for example, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the non-transitory storage medium include the following: an electrical connection having one or more wires, a portable disc, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0142] The above description of the disclosed aspects is merely exemplary in nature and is not intended to limit the present disclosure, application, and uses. The description of the aspects together with the accompanying drawings are intended to illustrate and not to limit the scope of the disclosure. The scope of the disclosure is given by the appended claims, and their equivalents.

[0143] The block diagrams of the devices, apparatuses, systems, etc. involved in the present disclosure are only illustrative examples and are not intended to require or imply that the connections, arrangements, configurations, etc. shown in the block diagrams are required or implied. As will be recognized by one of ordinary skill in the art, the devices, apparatuses, systems, etc. can be connected, arranged, configured, etc. in any manner. Words such as "include," "contain," "have," etc. are used synonymously with each other and mean "including but not limited to." The word "or" is used in the inclusive sense, meaning "and / or." The word "such as" is used in the sense of "such as but not limited to." The word "for" is used in the sense of "for and / or."

[0144] It is also important to note that the devices, apparatuses, and methods described in the present disclosure can be embodied in a variety of other forms; thus, the specific design features are not to be construed as limitations on the scope of the disclosure. Rather, these design features are viewed as illustrative only.

[0145] The above description of the disclosed aspects is merely exemplary in nature and is not intended to limit the present disclosure, application, and uses. The description of the aspects together with the accompanying drawings are intended to illustrate and not to limit the scope of the disclosure. The scope of the disclosure is given by the appended claims, and their equivalents.

[0146] The foregoing description has been presented for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of the application to the forms disclosed herein. Although various example aspects and embodiments have been discussed above, those of ordinary skill in the art will appreciate a variety of modifications, alternatives, permutations, additions, and sub-combinations, which fall within the scope of the application.

Claims

1. A land resource planning and management method based on spatiotemporal big data, characterized in that, The method comprises: acquiring multi-source spatio-temporal big data, land reserve state variables and land actual supply data within a current planning period, constructing a flow space perception matrix by spatio-temporal weighted fusion of the multi-source spatio-temporal big data, and extracting a land demand feedforward operator therefrom; generating a land estimated demand based on the land demand feedforward operator and the land reserve state variables, and performing a difference operation between the land estimated demand and the land actual supply data to generate reserve control deviation data; inputting the reserve control deviation data and a land supply decision instruction of a previous decision period stored in advance into a preset extended state observer, and estimating and separating a nonlinear comprehensive disturbance quantity in real time based on a preset gain parameter through the extended state observer; compensating and correcting the reserve control deviation data by using the land demand feedforward operator, and canceling the disturbance of the reserve control deviation data by using the nonlinear comprehensive disturbance quantity, to synthesize a land supply decision instruction of the current period; delivering the land supply decision instruction to a land reserve execution terminal, and collecting feedback data of land price fluctuation rate and land idle rate after execution in real time, and respectively correcting fusion weights of the flow space perception matrix and the preset gain parameter by using the feedback data; and using the corrected fusion weights and gain parameters as fusion weights and preset gain parameters of a flow space perception matrix of a next period. 2.The spatiotemporal big data-based land resource planning management method according to claim 1, characterized in that, The method further comprises: generating a plurality of alternative land supply schemes based on the land supply decision instruction of the current period before delivering the land supply decision instruction to the land reserve execution terminal; for each of the land supply schemes, predicting future land price fluctuation rate, land idle rate and traffic flow change rate within the planning area that will be caused after executing each scheme based on historical multi-source spatio-temporal big data; comparing the future land price fluctuation rate, land idle rate and traffic flow change rate with corresponding preset target thresholds to generate instruction pre-optimization parameters; correcting the land supply decision instruction of the current period by using the instruction pre-optimization parameters, and delivering the corrected land supply decision instruction to the land reserve execution terminal. 3.The spatiotemporal big data based land resource planning management method according to claim 2, characterized in that, The future land price fluctuation rate, land idle rate and traffic flow change rate within the planning area that will be caused include: extracting plot location coordinates, plot ratio parameters and land use classifications in each of the land supply schemes, and mapping them as disturbance variables to a spatio-temporal evolution topology graph constructed from historical multi-source spatio-temporal big data; analyzing spatio-temporal correlation influence domains of the disturbance variables on adjacent plots by using the spatio-temporal evolution topology graph, and calculating socio-economic characteristic evolution variables within the spatio-temporal correlation influence domains caused by land property changes; synchronously deriving corresponding future land price fluctuation rate, land idle rate and traffic flow change rate after executing each scheme based on the socio-economic characteristic evolution variables by using a preset cross-dimensional correlation operator. 4.The spatiotemporal big data-based land resource planning management method according to claim 1, characterized in that, The construction of the flow space perception matrix further comprises: dividing the planning area into a plurality of spatial grid units; identify correlation strengths between the spatial grid cells based on the multi-source spatio-temporal big data, wherein the correlation strengths are calculated by introducing a geographical constraint factor determined according to at least one of topographic connectivity, planning function compatibility, and traffic accessibility between the spatial grid cells; perform the spatio-temporal weighted fusion of the multi-source spatio-temporal big data by using the correlation strengths after the introduction of the geographical constraint factor, and construct an optimized flow space perception matrix. 5.The spatiotemporal big data-based land resource planning management method according to claim 4, characterized in that, The introduced geographical constraint factor includes: mapping the topographic slope change rate, the legal planning land class connection relationship, and the road travel time cost between the spatial grid cells to generate a geographical damping term representing the degree of spatial obstruction; performing attenuation calculation on the interaction features between the spatial grid cells by using the geographical damping term to obtain a geographical corrected correlation quantity; calculating the correlation strengths with both data flow and physical space constraint characteristics based on the geographical corrected correlation quantity. 6.The spatiotemporal big data based land resource planning management method according to claim 1, characterized in that, The disturbance offsetting includes: converting the nonlinear comprehensive disturbance quantity into an equivalent compensation control quantity, and extracting a time-varying gain coefficient of the equivalent compensation control quantity relative to the reserve control deviation data; performing piecewise linearization processing on the nonlinear comprehensive disturbance quantity according to the directionality characteristics of the time-varying gain coefficient to generate a dynamic compensation term with the same dimension as the reserve control deviation data; reversely superimposing the dynamic compensation term into the reserve control deviation data to modify the reserve control deviation data, and synthesizing the land supply decision instruction of the current period based on the modified reserve control deviation data. 7.The spatiotemporal big data based land resource planning management method according to claim 1, characterized in that, The modification of the fusion weight and the preset gain parameter includes: calculating the deviation value of the land price fluctuation rate and the land idle rate in the feedback data relative to the preset target, and analyzing the distribution density of the deviation value in the spatial dimension to generate a deviation spatio-temporal distribution feature quantity; identifying the influence factor mapping relationship of each dimension data on the land use deviation in the multi-source spatio-temporal big data through the deviation spatio-temporal distribution feature quantity, and adjusting the fusion weight according to the influence factor mapping relationship; calculating the time sequence change rate of the feedback data, determining a modification step based on the amplitude of the time sequence change rate, and using the modification step to perform real-time compensation modification on the preset gain parameter.

8. A land resource planning and management platform based on spatiotemporal big data, characterized in that, It includes: a flow space perception module for obtaining multi-source spatio-temporal big data, land reserve state variables, and land actual supply data in a planning area in a current period, constructing a flow space perception matrix by performing spatio-temporal weighted fusion on the multi-source spatio-temporal big data, and extracting a land demand feedforward operator therefrom; a reserve deviation calculation module for generating a land estimated demand quantity based on the land demand feedforward operator and the land reserve state variable, and performing difference operation on the land estimated demand quantity and the land actual supply data to generate reserve control deviation data; a disturbance observation module for inputting the reserve control deviation data and a land supply decision instruction of a previous decision period stored in advance into a preset extended state observer, and estimating and separating a nonlinear comprehensive disturbance quantity in real time based on a preset gain parameter through the extended state observer. A decision synthesis module is configured to compensate and correct the reserve regulation deviation data by using a land demand feedforward operator, and to disturb and offset the reserve regulation deviation data by using a nonlinear comprehensive disturbance quantity, so as to synthesize a land supply decision instruction of a current period; An adaptive correction module is configured to issue the land supply decision instruction to a land reserve execution terminal, to collect feedback data of land price fluctuation rate and land idle rate after execution in real time, and to correct the fusion weight of the flow space perception matrix and the preset gain parameter respectively by using the feedback data; and The corrected fusion weight and gain parameter are used as the fusion weight and preset gain parameter of the flow space perception matrix of a next period. 9.An electronic device comprising a memory and a processor, the electronic device characterized by: The memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions, so as to implement the steps of the method in any one of claims 1 to 7.

10. A computer storage medium having stored thereon computer- executable instructions, comprising: The computer executable instructions are executed by the processor, so as to implement the steps of the method in any one of claims 1 to 7.

Images