Cloud native space-time fusion analysis method based on space-time asset directory

Abstract

The invention discloses a cloud native space-time fusion analysis method based on a space-time asset directory, and belongs to the technical field of data indexing and fusion analysis, and the method comprises the steps: S1, obtaining a core element of a source asset, and constructing a space-time influence potential model, so as to solve and obtain a space-time influence potential; s2, calculating an influence gradient based on the space-time influence potential, and comparing the modulus length of the influence gradient with a preset threshold value to generate a coupling relationship evaluation result; s3, in response to a coupling relationship evaluation result, performing evolution calculation on the current core index of the target asset by using a state evolution model to predict and generate a future core index; and S4, based on future core indexes, an intervention strategy is generated, and after the intervention strategy is executed, adaptive calibration is carried out on parameters in the state evolution model by using prediction errors, and cognitive crossing from fuzzy correlation to quantitative causality is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of data indexing and fusion analysis, and specifically to a cloud-native spatiotemporal fusion analysis method based on a spatiotemporal asset catalog. Background Art

[0002] When operating and managing complex urban systems, traditional data analysis and decision-making methods mainly rely on isolated, static correlation analysis of historical data. This method usually regards spatiotemporal data as independent, static information points, and uses post-analysis to find fuzzy connections between different events or assets. When intervention decisions are needed, they often rely on manual experience to judge and deal with them, forming a passive emergency response mode.

[0003] The above status quo and shortcomings are mainly due to the limitations of the analysis model and technical links:

[0004] Limitations of analytical models: Traditional technologies struggle to quantify the potential impact of events, and can only draw vague conclusions, such as that event A is related to asset B, without accurately describing the intensity and direction of the impact. Furthermore, their predictive models are often simple linear extrapolations, unable to simulate saturation effects or bottleneck limitations commonly found in the real world, such as parking lot capacity. This can lead to unrealistic predictions of unlimited growth under strong impacts.

[0005] Limitations of the technical chain: Existing technologies lack an automated closed-loop system from data perception, causal reasoning, state prediction to decision-making intervention; analytical models are usually rigid and static, unable to self-correct and evolve based on real-world feedback, resulting in the inability to continuously improve their cognitive and predictive accuracy.

[0006] As a result, when faced with emergencies in cities, managers are unable to transform isolated data into predictable and intervention-enabled dynamic causal relationships; this leads to insufficient efficiency and foresight in emergency responses, and decision-making models that rely on human experience find it difficult to quickly generate scientific and cost-optimal intervention strategies in complex situations.

[0007] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0008] The purpose of the present invention is to provide a cloud-native spatiotemporal fusion analysis method based on a spatiotemporal asset catalog to solve the problems raised in the above background technology.

[0009] The technical solution of the present invention is that the specific steps include:

[0010] S1. Obtain the core elements of the source assets and build a spatiotemporal influence model to obtain the spatiotemporal influence potential;

[0011] S2. Based on the spatiotemporal influence potential, the influence gradient is calculated, and the modulus of the influence gradient is compared with a preset threshold to generate a coupling relationship evaluation result;

[0012] S3. In response to the coupling relationship evaluation results, using the state evolution model, perform evolution calculations on the current core indicators of the target asset to predict and generate future core indicators;

[0013] S4. Generate intervention strategies based on future core indicators, and after executing the intervention strategies, use the prediction errors to adaptively calibrate the parameters in the state evolution model.

[0014] Preferably, S1 includes:

[0015] Collect the core elements of the source assets from the spatiotemporal asset catalog, including spatiotemporal coordinates and core business indicators;

[0016] Normalize the core business indicators to obtain the event source strength; build a spatiotemporal impact potential model by combining the event source strength, the preset time decay coefficient, and the preset regularization length; use the spatiotemporal impact potential model to solve and obtain the spatiotemporal impact potential.

[0017] Preferably, the initial calibration method of the regularization length includes calibrating the core influence range or equivalent radius of the asset to the regularization length based on the type and physical properties of the asset; the initial calibration method of the time attenuation coefficient includes extracting the influence attenuation data of historical similar events and fitting the data using an exponential decay model to solve and obtain the time attenuation coefficient.

[0018] Preferably, S2 includes:

[0019] Calculate the negative gradient of the spatiotemporal impact potential to obtain the impact gradient; collect the gradient modulus generated by events that have not caused changes in the core indicators of the target assets in history, and calibrate the high percentile points as the monitoring response threshold; collect the gradient modulus generated by events that have caused sudden changes in the core indicators of the target assets in history, and calibrate the low percentile points as the critical warning threshold.

[0020] Preferably, the step of generating a coupling relationship evaluation result includes:

[0021] When the mode length of the influencing gradient is less than or equal to the monitoring response threshold, the coupling relationship evaluation result is determined to be weak coupling;

[0022] When the modulus of the impact gradient is greater than the monitoring response threshold and less than or equal to the critical warning threshold, the coupling relationship assessment result is determined to be strong coupling;

[0023] When the modulus of the influencing gradient is greater than the critical warning threshold, the coupling relationship evaluation result is determined to be critical coupling.

[0024] Preferably, S3 includes:

[0025] For asset pairs whose coupling relationship assessment results are strong coupling or critical coupling, the state evolution model is used to perform evolutionary calculations on the current core indicators of the target assets based on the impact gradient, the sensitivity vector and capacity parameters of the target assets to predict and generate future core indicators.

[0026] Preferably, the capacity parameter is calibrated by obtaining the extreme values ​​of the physical properties or historical data of the target asset; the calibration method of the sensitivity vector includes: performing a weighted average calculation based on the historical data of the functional layout of the target asset to determine the direction; and determining the modulus by performing a linear regression analysis on the historical impact gradient and the change in the core indicator.

[0027] Preferably, S4 includes:

[0028] Compare future core indicators with preset business alarm thresholds to identify risky assets. Calculate the system instability index based on the risk asset's business weight, impact gradient, and monitoring response threshold. Generate an intervention strategy based on the system instability index and preset business security thresholds, with the principle of minimizing total intervention cost.

[0029] Business weights are quantified through the analytic hierarchy process or contribution analysis driven by historical data.

[0030] Preferably, the step of adaptive calibration includes:

[0031] After the intervention strategy is executed, the actual core indicators of the target assets are tracked; the prediction error between the actual core indicators and the future core indicators is calculated; and the prediction error is used to correct the sensitivity vector in the state evolution model.

[0032] The present invention provides a cloud-native spatiotemporal fusion analysis method based on a spatiotemporal asset catalog through improvements. Compared with the existing technology, it has the following improvements and advantages:

[0033] 1. This invention achieves a cognitive leap from fuzzy correlation to quantitative causality. Its core lies in the construction of a unified spatiotemporal influence potential model. By introducing a regularization length, this model resolves the singularity problem at the source of traditional physical field models, ensuring computational robustness. This enables the system to automatically classify complex interactions into weak coupling, strong coupling, and critical coupling, providing clear and quantifiable trigger conditions for subsequent analysis and intervention at different levels.

[0034] 2. This invention goes beyond the simple linear extrapolation predictions of existing technologies and introduces a state evolution model that is more consistent with the laws of nonlinear dynamics, ensuring that the prediction logic closely matches the actual response characteristics of the assets. The system automatically solves and generates the optimal intervention strategy based on the principle of minimizing total intervention cost. This marks a shift from a passive, manual emergency response model to an active, automatic intelligent operation model. Closed-loop feedback enables the entire analysis system to have learning capabilities, drawing experience from each practice, and continuously improving its understanding of complex spatiotemporal systems and prediction accuracy. This distinguishes it from all static, fixed analysis models and makes it an intelligent decision-making executor that can evolve with the environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The present invention will be further explained below in conjunction with the accompanying drawings and examples:

[0036] Figure 1 It is a flowchart of the method of the present invention. DETAILED DESCRIPTION

[0037] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to specific embodiments.

[0038] Example 1:

[0039] See also Figure 1 The present invention provides a cloud-native spatiotemporal fusion analysis method based on a spatiotemporal asset catalog, the specific steps of which include:

[0040] S1. Obtain the core elements of the source assets and build a spatiotemporal influence model to obtain the spatiotemporal influence potential;

[0041] S2. Based on the spatiotemporal influence potential, the influence gradient is calculated, and the modulus of the influence gradient is compared with a preset threshold to generate a coupling relationship evaluation result;

[0042] S3. In response to the coupling relationship evaluation results, using the state evolution model, perform evolution calculations on the current core indicators of the target asset to predict and generate future core indicators;

[0043] S4. Generate intervention strategies based on future core indicators, and after executing the intervention strategies, use the prediction errors to adaptively calibrate the parameters in the state evolution model.

[0044] The cloud-native spatiotemporal fusion analysis method based on the spatiotemporal asset catalog disclosed in this embodiment is constructed as a closed-loop system with four core steps to ensure automation and intelligence from data perception to decision execution.

[0045] The method begins by capturing the core elements of source assets and constructing a spatiotemporal impact model. Core elements refer to structured data that characterizes the key characteristics of a spatiotemporal asset, providing the foundational input for subsequent quantitative analysis. This data is sourced from a pre-built spatiotemporal asset catalog, a database that provides a unified digital management of various physical or logical assets in a city.

[0046] In a preferred embodiment, the spatiotemporal asset catalog can be constructed as a relational database or a graph database; the core data table or node attributes should at least include: asset unique identifier, asset type, such as traffic events, commercial facilities, sensors, four-dimensional spatiotemporal coordinates, , core business indicators such as congestion index, passenger flow and related physical or business attributes such as road grade, parking capacity The directory's data can be imported in real time through standard application programming interfaces (APIs) from data sources such as the city's IoT sensor network, public transportation information platforms, and meteorological services. This data is supplemented by manual entry and regular maintenance of static assets such as building information to ensure accuracy and timeliness.

[0047] In this scenario, the system extracts traffic incidents, the spatiotemporal coordinates of source asset A, and its core business indicators, such as the congestion index, from the directory. Based on these elements, a spatiotemporal impact potential model is constructed to describe the scope and intensity of its impact. The purpose of this model is to transform isolated events into a measurable and dynamically changing impact field in the four-dimensional spatiotemporal continuum, thereby obtaining a quantified spatiotemporal impact potential.

[0048] Based on the spatiotemporal influence potential obtained in the previous step, the system calculates the influence gradient and compares the modulus of the influence gradient with the preset threshold to generate a coupling relationship assessment result. The influence gradient, as a measure of the rate of change and direction of the spatiotemporal influence potential in space, is used to accurately quantify the intensity and directionality of the impact exerted by the source asset on the target asset. By performing vector operations on the spatiotemporal influence potential, an influence gradient pointing to shopping center B can be obtained. The preset thresholds referred to here are a set of calibration values ​​used to assess the strength of the impact. The internal logic of their establishment is to objectively distinguish the boundaries between background noise impact, significant effective impact, and critical dangerous impact through historical data statistics. By comparing the calculated influence gradient modulus with these thresholds, the system can objectively determine the strength of the coupling relationship between the traffic accident and shopping center B, such as strong coupling or critical coupling.

[0049] In response to the coupling relationship assessment results, the system uses a state evolution model to perform evolutionary calculations on the target asset's current core indicators to predict and generate future core indicators. When the assessment results indicate a significant impact, a state evolution model is activated. This model aims to simulate the dynamic changes in the target asset's own state over time under external influences. Rather than a simple linear accumulation, it comprehensively considers the impact intensity, the target asset's sensitivity, and its inherent carrying capacity. In this scenario, the model receives the impact gradient as input and deduces core indicators such as Shopping Center B's current passenger flow and parking lot occupancy rate, thereby predicting future core indicators such as a 60% drop in main entrance passenger flow within the next 15 minutes.

[0050] Based on the predicted future core indicators, the system generates an intervention strategy and, after the strategy is executed, uses the prediction error to adaptively calibrate the parameters in the state evolution model. An intervention strategy is a set of optimized operational instructions designed to avoid risks or exploit opportunities. Based on the disruptive prediction generated in the previous step, such as if the parking lot congestion rate exceeds the warning line, the system automatically solves for the most cost-effective intervention combination. After the strategy is executed, the system continuously tracks the actual customer flow changes at Shopping Center B and compares the actual value with the model's predicted value, generating a prediction error. This error is used as a feedback signal to drive the adaptive calibration of the internal parameters of the state evolution model, thus forming a complete learning closed loop.

[0051] Through the above steps, this method constructs a complete technical chain from quantification of spatiotemporal impacts to coupling relationship assessment, and then to future state prediction and closed-loop adaptive intervention; it transforms the originally isolated and static spatiotemporal data into predictable and interventionable dynamic causal relationships, realizing the automated, precise and intelligent operation and management of complex urban system events, greatly improving the efficiency and foresight of emergency response.

[0052] Example 2

[0053] S1 includes:

[0054] Collect the core elements of the source assets from the spatiotemporal asset catalog, including spatiotemporal coordinates and core business indicators;

[0055] Normalize the core business indicators to obtain the event source strength; build a spatiotemporal impact potential model by combining the event source strength, the preset time decay coefficient, and the preset regularization length; use the spatiotemporal impact potential model to solve and obtain the spatiotemporal impact potential.

[0056] The construction and solution process of the spatiotemporal impact potential model is described;

[0057] The core elements of the source assets are collected from the spatiotemporal asset catalog. The core elements include spatiotemporal coordinates and core business indicators. In this scenario, the system accurately extracts the spatiotemporal coordinates of accident A from the spatiotemporal asset catalog database. and its core business indicators This indicator is the quantitative value that best represents the status of an asset, such as the congestion index here, and serves as the raw input for impact intensity.

[0058] Normalize the core business indicators to obtain the event source strength; combine the event source strength, the preset time decay coefficient and the preset regularization length to build a spatiotemporal influence potential model; use the spatiotemporal influence potential model to solve and obtain the spatiotemporal influence potential; event source strength Core business indicators The dimensionless parameters obtained after normalization enable the comparison of the influence of different types of assets within a unified framework, ensuring the universality of the model.

[0059] The normalization process preferably adopts the minimum-maximum normalization method. For a certain type of asset, the event source intensity It can be calculated by the following formula:

[0060]

[0061] in, is the actual value of the core business indicator currently collected, and and are the maximum and minimum values ​​of the core business indicators of this type of asset recorded in the historical database; through this method, business indicators of different dimensions can be uniformly mapped to to In the interval of , the standardized event source intensity is obtained ;

[0062] To build the model, two preset core parameters are introduced: time decay coefficient and regularization length ; Used to describe the natural decline in the impact of an event over time; It is used to solve the problem of numerical singularities in the physical model near the source point due to the distance approaching zero. Based on the above parameters, the space-time influence potential model is constructed, and the calculation is given by the following formula:

[0063]

[0064] in, is the spatiotemporal influence potential, and the input parameters come from: Core business indicators Normalized calculated event source intensity, is the spatial distance between the source asset and the target asset, is the time difference between the two, is the preset time attenuation coefficient, is the preset regularization length; is the base of natural logarithms;

[0065] By substituting the Value, and the spatial distance between it and shopping center B and time difference , we can solve and get an accurate Value, which quantifies the impact of accident A on shopping center B at the current moment;

[0066] This embodiment introduces the normalized event source intensity , time attenuation coefficient and regularization length , provides an explicit and computable definition of the spatiotemporal influence potential model; in particular The introduction of solves the singularity problem that occurs in traditional physical models during close-range calculations through smoothing, ensuring that the model has good computational robustness at any spatial scale. This makes the quantification process of the influence field not only unified and standardized, but also more stable and reliable in terms of computation.

[0067] Compared with the existing technology that usually stays in the isolated and static correlation analysis of historical data, this invention realizes the cognitive leap from fuzzy association to quantitative causality; the core lies in the construction of a unified spatiotemporal influence model; this model introduces regularization length , which solves the problem of traditional physical field model at the source point The singularity problem at ensures the robustness of the calculation; its mathematical expression is:

[0068]

[0069] in, :Space-time influence potential, : Event source intensity, : spatial distance, : time difference, : time decay coefficient, : Regularization length; The physical meaning of this formula is that it abstracts the core business indicators of any source asset into a potential field source that decays continuously over time and space. It provides a unified and singular-free mathematical basis for all subsequent analyses; and Initial calibration based on physical properties and historical data eliminates the black box of the source of model parameters and makes the construction of the influence field source objective and explainable.

[0070] Example 3

[0071] The initial calibration method for the regularization length includes calibrating the core influence range or equivalent radius of the asset as the regularization length based on the type and physical properties of the asset; the initial calibration method for the time decay coefficient includes extracting the influence decay data of historical similar events and fitting the data using an exponential decay model to solve and obtain the time decay coefficient.

[0072] To ensure the objectivity and interpretability of the model parameters, the initial calibration method of two key preset parameters in the model is further disclosed;

[0073] The initial calibration method of the regularization length is described as follows: according to the type and physical properties of the asset, the core range or equivalent radius of the asset is calibrated as the regularization length ; Regularization length Physically, it represents the core scope of the asset's influence, that is, a meaningful minimum interaction distance. The calibration principle is to anchor abstract mathematical parameters with the physical reality of the asset. For example, for point-like event assets such as traffic accidents, the core area of ​​physical influence is relatively small. It can be objectively calibrated to a smaller value, such as 5 meters; for surface assets such as large shopping malls, The equivalent radius can be calculated based on the building area, such as 100 meters;

[0074] The initial calibration method of the time decay coefficient is described as extracting the influence decay data of historical similar events and fitting the data with an exponential decay model to obtain the time decay coefficient. The working principle is to establish a mathematical model for the natural law that the influence of different types of events fades over time; for a certain type of event, multiple samples are extracted from the historical database; for each sample, its data at different time points after the event is collected. Influence indicators on the actual observed congestion index ; These observation data points Fitting to an exponential decay model Among them, : Initial impact indicator; : decay rate, which is the time decay coefficient after being solved by fitting; : Subscript, used to identify different data collection time points, such as ; The attenuation rate solved by the least squares method and other optimization algorithms , which is marked as the initial event of this type value; Generally, the influence of traffic accidents decays quickly. The value is large; while the influence of environmental pollution events decays slowly. The value is smaller;

[0075] Through and This embodiment provides an objective calibration method based on physical properties and historical data statistics. It ensures the transparency and reproducibility of the basic parameters of the model. This method closely integrates the construction of the model with the physical laws and historical experience of the real world, significantly improving the accuracy, interpretability and credibility of the model in practical applications, and avoiding the model uncertainty problem caused by unknown parameter sources.

[0076] Example 4

[0077] S2 includes:

[0078] Calculate the negative gradient of the spatiotemporal impact potential to obtain the impact gradient; collect the gradient modulus generated by events that have not historically caused changes in the core indicators of the target assets, and calibrate the high percentile points as the monitoring response threshold; collect the gradient modulus generated by events that have historically caused sudden changes in the core indicators of the target assets, and calibrate the low percentile points as the critical warning threshold;

[0079] The steps to generate coupling relationship evaluation results include:

[0080] When the mode length of the influencing gradient is less than or equal to the monitoring response threshold, the coupling relationship evaluation result is determined to be weak coupling;

[0081] When the modulus of the impact gradient is greater than the monitoring response threshold and less than or equal to the critical warning threshold, the coupling relationship assessment result is determined to be strong coupling;

[0082] When the modulus of the influencing gradient is greater than the critical warning threshold, the coupling relationship evaluation result is determined to be critical coupling.

[0083] The impact gradient calculation and coupling relationship evaluation process are specified, forming a complete evaluation and grading process;

[0084] The process begins by calculating the negative gradient of the spatiotemporal influence potential to obtain the influence gradient; the influence gradient It is a vector pointing to the direction of the fastest change in the potential, and the modulus is Indicates the intensity of the influence; by the spatiotemporal influence potential obtained in the previous step Applying the gradient operator And take the negative value to solve;

[0085]

[0086] in, To influence the gradient vector, the input of its calculation is the spatiotemporal influence potential obtained in the previous step ;

[0087] Calculation results The direction of the impact of traffic accident A on shopping mall B is precisely indicated, for example, the main vehicle entrance and intensity of B;

[0088] In order to objectively grade the impact intensity, two core thresholds need to be calibrated. First, the gradient modulus generated by events that have not historically caused changes in the core indicators of the target assets is collected, and the high percentile point is calibrated as the monitoring response threshold. This is the monitoring response threshold. The calibration process is based on identifying and filtering out the impact of background noise. The system analyzes all historical events that have affected Shopping Center B and selects those that did not cause significant changes in its core indicators such as customer flow, for example, events with fluctuations less than the preset business sensitivity. is set to the gradient modulus produced by these invalid impact events The 95th percentile of the effect, thus ensuring that only statistically significant effects are systematically addressed;

[0089] Second, collect the gradient modulus generated by historical events that caused a sudden change in the core indicators of the target assets, and mark the low percentile point as the critical warning threshold; this is the critical warning threshold The system identifies high-risk couplings that could cause chain reactions in business systems. It screens out all historical events that have caused sudden changes in Shopping Center B's core indicators, such as a month-on-month change in customer flow exceeding 30%. is set to the gradient modulus corresponding to these strong impact events The 10th percentile; this low percentile is used to ensure the sensitivity of the warning, ensuring that even the smallest impact intensity that has caused mutations in history is sufficient to trigger a high-level warning;

[0090] Based on the above threshold, the step of generating the coupling relationship evaluation result is performed; the influence gradient modulus is calculated Comparison with two objectively calibrated thresholds:

[0091] like , the coupling relationship evaluation result is determined to be weak coupling, indicating that the impact can be ignored and the system is not processed;

[0092] like , the coupling relationship assessment result is determined to be strong coupling, indicating that there is a clear impact transmission relationship, and the system will activate the subsequent causal deduction module;

[0093] like , the coupling relationship assessment result is determined to be critical coupling, indicating that the target asset may have a state mutation. The system will trigger an advanced warning while activating causal deduction;

[0094] This embodiment converts the abstract influence into a gradient vector that can be accurately calculated , and established a set of dual thresholds based on the asset's own historical data ( , ) Calibration and evaluation system; this enables the system to rigorously and objectively quantify the intensity and direction of interactions between assets, providing clear and unambiguous input conditions for subsequent prediction and intervention levels, achieving a cognitive upgrade from fuzzy correlations to quantitative transmission relationships;

[0095] The present invention further defines the influence gradient ,in, : influence gradient vector, : gradient operator, : Spatiotemporal impact potential, which converts the scalar impact potential into a vector, accurately describing the intensity and direction of the impact; this is in sharp contrast to the existing technology that can only draw vague conclusions about the relationship between A and B; more valuable is that the present invention establishes a dual evaluation system based on monitoring response thresholds and critical warning thresholds; these two thresholds are not set manually, but are objectively calibrated through statistical analysis of the historical data of the target assets, corresponding to the minimum impact that produces a meaningful response and the critical impact that causes a sudden change in state; this design enables the system to automatically classify complex interactions into weak coupling, strong coupling and critical coupling, providing clear and quantifiable trigger conditions for the subsequent activation of different levels of analysis and intervention.

[0096] Example 5

[0097] S3 includes:

[0098] For asset pairs whose coupling relationship assessment results are strong coupling or critical coupling, the state evolution model is used to perform evolutionary calculations on the current core indicators of the target assets based on the impact gradient, the sensitivity vector and capacity parameters of the target assets to predict and generate future core indicators.

[0099] The state evolution and prediction process are explained;

[0100] For asset pairs whose coupling relationship is assessed as strongly coupled or critically coupled, the state evolution model is used to perform an evolution calculation on the current core indicators of the target asset based on the impact gradient, the sensitivity vector of the target asset, and the capacity parameter to predict and generate future core indicators. This step is only triggered when there is a significant impact between assets to improve system efficiency. The state evolution model is a nonlinear dynamic model used to predict the future state of the target asset. The model introduces two key parameters: the sensitivity vector and capacity parameters ; Describes the tendency and severity of the target asset's response to impacts from a specific direction; Represents the physical or business upper limit that the core indicators of the target asset can reach; based on these parameters, the core indicators of the target asset at the next moment The prediction is made by the following nonlinear evolution formula:

[0101]

[0102] in, It is the core indicator for prediction at the next moment, and its input parameters come from: The core indicators of the target assets collected at the current moment, is the influence gradient vector calculated in the previous step, and are the pre-calibrated target asset sensitivity vector and capacity parameter, is the preset time step;

[0103] Time step The value of can be determined according to the emergency response timeliness requirements of the business scenario or the data refresh frequency of the core indicators. For example, in a traffic congestion prediction scenario that requires a quick response, The optimal setting is 1 to 5 minutes. In commercial area planning scenarios where passenger flow changes are relatively slow, it can be set to 15 minutes or longer to balance computing resource consumption and prediction accuracy.

[0104] In this scenario, the model uses the calculated gradient Shopping Center B preset and Value, for current passenger flow Perform iterative calculations to accurately predict nonlinear phenomena such as increased parking lot congestion and abnormal increases in customer flow in dining areas in the future;

[0105] This embodiment introduces capacity constraints by constructing a The nonlinear state evolution model greatly improves the authenticity of predictions. This model can scientifically simulate the saturation effects and bottleneck limitations that are common in the real world, avoiding the conclusion of unlimited growth that linear models often draw under strong influences. This enables the system to advance from static relationship strength analysis to dynamic causal effect prediction that is more in line with physical laws, providing a more accurate and reliable decision-making basis for risk warning and resource allocation.

[0106] The present invention goes beyond the simple linear extrapolation prediction of the prior art and introduces a state evolution model that is more in line with the laws of nonlinear dynamics; its mathematical form is ,in, : Prediction core indicators, :Current core indicators, : influence gradient, : sensitivity vector, : Capacity parameter, : Time step; the essence of this formula is the introduction of the capacity parameter , which represents the physical or operational upper limit of the core asset indicator, such as the total number of parking spaces in a parking lot; this parameter enables the model to accurately predict saturation effects or bottleneck limitations that are common in the real world. For example, it predicts that road congestion will not increase indefinitely, but will increase when it reaches its maximum capacity. saturated state when ; This is completely beyond the reach of traditional linear models; At the same time, the sensitivity vector in the model Calibration is performed through regression analysis of historical data and weighted calculation of functional layout to ensure that the prediction logic is closely aligned with the actual response characteristics of the assets.

[0107] Example 6

[0108] The capacity parameters are calibrated by obtaining the extreme values ​​of the physical properties or historical data of the target assets. The calibration method of the sensitivity vector includes: performing weighted average calculation based on the historical data of the functional layout of the target assets to determine the direction; and determining the modulus by performing linear regression analysis on the historical impact gradient and the change in core indicators.

[0109] In this embodiment, the two core parameters in the state evolution model, capacity parameter and sensitivity vector The initial calibration method is specified to eliminate the uncertainty of the model parameters.

[0110] Capacity parameters are calibrated by obtaining the physical properties of the target asset or the extreme values ​​of historical data; Represents the upper or lower limit of the core indicator of the target asset; the calibration method is directly derived from the objective data of the asset, ensuring the physical meaning of the parameter; for example, for shopping center B, the capacity parameter It could be the maximum number of people allowed under fire safety regulations, the total number of parking spaces in its supporting parking lot, or the highest single-day passenger flow record in history;

[0111] The calibration method of the sensitivity vector is decomposed into the calibration of direction and modulus respectively;

[0112] First, direction The internal logic of the calibration is to perform weighted average calculation based on the historical data of the functional layout of the target asset to determine the direction; this method converts abstract attributes such as functional zoning into computable vector directions; for example, the sensitivity direction of shopping center B is not set subjectively, but is obtained by weighted average calculation of historical passenger flow data of multiple entrances, such as the main entrance and the parking lot entrance; each entrance Corresponding to a unit vector pointing from the outside to the inside , its weight is the historical average daily traffic volume of the entrance; the final sensitivity direction Calculated by the following formula:

[0113]

[0114] in, is the sensitivity direction vector, and its input parameter comes from: is the weight of each entry calculated based on historical data, is the unit direction vector corresponding to each entrance; : Subscript, referring to each entry of the target asset;

[0115] Second, mold length The technical solution is to determine the modulus by performing linear regression analysis on the historical impact gradient and the change in core indicators; the modulus quantifies the severity of the asset response; the calibration process collects historical impact-response data pairs, that is, the gradient vector each time asset B is affected. and the actual changes in core indicators caused by this ; According to the linear approximation of the state evolution formula , the independent variable can be , dependent variable ; By analyzing these data points Perform linear regression analysis without intercept , the obtained regression coefficient That is, the initial modulus length objectively calibrated as the sensitivity vector ;

[0116] This embodiment is the two core parameters of the state evolution model and It provides a fully transparent and reproducible quantitative calibration solution; it deeply binds model parameters to the physical specifications, functional layout and historical response behavior of the assets, completely eliminating the subjectivity and arbitrariness of parameter settings, and ensuring the rationality and accuracy of the prediction model during the cold start phase.

[0117] Example 7

[0118] S4 includes:

[0119] Compare future core indicators with preset business alarm thresholds to identify risky assets. Calculate the system instability index based on the risk asset's business weight, impact gradient, and monitoring response threshold. Generate an intervention strategy based on the system instability index and preset business security thresholds, with the principle of minimizing total intervention cost.

[0120] Business weights are quantified through the analytic hierarchy process or contribution analysis driven by historical data.

[0121] The process of generating intervention strategies was specified;

[0122] Compare future core indicators with preset business alarm thresholds to identify risky assets; the system compares the future core indicators of shopping center B predicted by the evolution model with the preset business alarm thresholds to identify risky assets. For example, if the parking lot congestion rate is predicted to reach 95% and the business alarm threshold , such as 85% for comparison; the threshold The determination of the threshold value is usually based on statistical analysis of the critical points of indicators that lead to a significant decline in service quality in historical data, or directly citing the operational safety standards of relevant industries. Once the predicted value exceeds the threshold, Shopping Center B is identified as a risky asset.

[0123] The system instability index is calculated by combining the business weight of risk assets, the impact gradient and the monitoring response threshold; in order to quantify the severity of the overall risk, the system introduces a system instability index ; The calculation of this index requires business weights This weight is a parameter quantified through the Analytic Hierarchy Process (AHP) or contribution analysis driven by historical data, reflecting the importance of a single asset in the entire business network.

[0124] Contribution analysis driven by historical data may include: defining a system-level key performance indicator that can measure the overall operating status of the system, such as the average traffic efficiency of the regional transportation network; constructing a multivariate linear regression model with this system-level KPI as the dependent variable and the core indicators of each risk asset as the independent variable; and directly calibrating the absolute value of the standardized regression coefficient corresponding to the core indicator of each asset in the model as the business weight of the asset. , the coefficient value objectively reflects the degree of influence of the change of a single asset status on the overall status of the system;

[0125] Instability Index The following formula is used to calculate the excess impact of all risk assets, weighted by their importance:

[0126]

[0127] in, is the system instability index, and its input parameters come from: is the total number of risky assets identified, Pre-calibrated business weights for each risk asset, is the impact gradient of each risk asset, The monitoring response threshold for each risk asset; : Subscript, representing each risky asset from 1 to N;

[0128] Based on the system instability index and the preset business security threshold, and with the principle of minimizing the total intervention cost, the system generates the intervention strategy; Value and a business security threshold Compare; the threshold The basis for setting the value is to ensure that the system returns to an acceptable low-risk state after intervention. The value can be obtained by analyzing the highest risk in history without chain reaction. Worth it; if The system will start the intervention strategy optimization process, the optimization goal is to meet the instability index after intervention down to Under the following premise, find a total intervention cost Minimum combination of measures; cost of each measure Derived from a cost base based on historical resource consumption or expert assessments;

[0129]

[0130] in, : the cost of each intervention measure in the measure library, as shown by the subscript j; : Subscripts refer to the various intervention measures in the measure library; : instability index after intervention; : Preset business security thresholds; the system automatically generates and recommends cost-effective intervention strategies by simulating the impact of various operations in the measure library on source asset strength;

[0131] To ensure the objectivity of costs, the cost library should be established based on quantifiable data; for interventions involving material consumption, such as placing traffic cones, the cost The product of the market price of the material and the quantity used; for measures involving manpower scheduling, such as adding traffic coordinators, the cost The product of the standard man-hour cost and the estimated man-hours; for measures that require the use of third-party services, such as the use of towing services, the cost The market will provide public quotations for them; by storing these quantifiable cost data in a structured manner, an objective cost database can be formed;

[0132] In this embodiment, the system instability index is introduced and business weight , transforming fuzzy system risks into quantifiable decision-making indicators that reflect business priorities. Furthermore, by constructing an optimization problem with the goal of optimizing cost-effectiveness, it achieves automated and optimized generation of intervention strategies, replacing the traditional decision-making model that relies on manual experience, making emergency response faster, more scientific, and more economical.

[0133] The progress of this invention is that it does not stop at prediction, but builds a complete closed loop of prediction-decision-intervention-learning; it uses the system instability index ,in, : System instability index, : Total amount of risky assets, : Business weight, : influence gradient, : The monitoring response threshold quantifies multi-point, concurrent future risks into a single decision-making indicator; based on this indicator, the system automatically solves and generates the optimal intervention strategy with the principle of minimizing total intervention cost; this marks the transition from a passive, manual emergency response model to an active, automatic intelligent operation model.

[0134] Example 8

[0135] The steps of adaptive calibration include:

[0136] After the intervention strategy is executed, the actual core indicators of the target assets are tracked; the prediction error between the actual core indicators and the future core indicators is calculated; and the prediction error is used to correct the sensitivity vector in the state evolution model.

[0137] This embodiment concretizes the adaptive calibration step after executing the intervention strategy, and forms a complete and self-evolving learning closed loop through a dimensionally rigorous feedback mechanism.

[0138] After the intervention strategy is executed, the system continues to track the actual core indicators of the target asset. After the system executes the intervention strategy, its process does not terminate, but continues to track and record the actual core indicators of Shopping Center B in the subsequent time through sensors or data interfaces. .

[0139] Calculate the prediction error between the actual core indicators and the future core indicators. The system will collect the real data The future core indicators generated by the state evolution model before intervention , that is, the predicted value is compared to obtain a quantitative prediction error ;in, : Prediction error; this error accurately reflects the degree of deviation between the model prediction and the real world;

[0140] The prediction error is used to modify the sensitivity vector in the state evolution model; this prediction error It is used as a feedback signal to drive the automatic calibration of the core parameters of the state evolution model. To ensure the rigor of the calibration process in mathematical and physical dimensions, the sensitivity vector Updates are performed according to the following gradient descent correction rule:

[0141]

[0142] in, To be the updated sensitivity vector, the input for the calculation is: is the sensitivity vector before updating, The actual core indicators tracked after the intervention, It is the core indicator of model prediction. is the time step used in the state evolution model, is the influence gradient vector calculated in the previous step, is the square of the gradient modulus, and It is a dimensionless machine learning hyperparameter preset to balance convergence speed and stability, namely the learning rate;

[0143] In a preferred embodiment, the learning rate The initial value can be set to to The conventional range; technicians in the relevant technical field can use cross-validation methods or monitor the prediction error during the actual operation of the model. Dynamically adjust the convergence of The value of is used to seek the optimal model convergence speed and stability;

[0144] The physical significance of this correction formula is that it will predict the error As the core driver of correction and make the sensitivity vector The adjustment amount and the gradient vector that causes the impact proportional to; this means that the stronger the influence The resulting error will cause A larger correction is consistent with a logically robust learning mechanism;

[0145] This embodiment introduces a dimensionally self-consistent feedback calibration mechanism based on the prediction error rate, transforming the entire analysis system from a static, one-time prediction tool into a self-evolving intelligent learning system. This adaptive calibration capability ensures that the model parameters can continuously and automatically approximate the physical laws of the real world. This is not only more mathematically rigorous, but also makes the model learning process more physically interpretable, thereby continuously improving the accuracy of future predictions and the effectiveness of intervention decisions, achieving true closed-loop adaptive control.

[0146] The method of the present invention achieves self-evolution through an adaptive calibration mechanism. After the intervention strategy is executed, the system uses the prediction error between the observed actual core indicators and the model prediction value to correct the model parameters. For example, the correction of the sensitivity vector is completed through a dimensionally self-consistent gradient descent rule: ,in, is the new sensitivity vector, is the old sensitivity vector, is the dimensionless learning rate, For actual indicators, As a prediction indicator, is the time step, To influence the gradient; This correction method converts the original prediction value error into a more physically meaningful rate of change error, and ensures the learning rate by normalizing the gradient term. It is a standard dimensionless hyperparameter, which solves the dimensional inconsistency problem of the model. This closed-loop feedback enables the entire analysis system to have the ability to learn, draw experience from each practice, and continuously improve its cognition and prediction accuracy of complex spatiotemporal systems, making it different from all static and fixed analysis models and becoming an intelligent decision-making executor that can evolve with the environment.

[0147] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A cloud-native spatiotemporal fusion analysis method based on spatiotemporal asset catalog, characterized by: The specific steps include: S1. Obtain the core elements of the source assets and build a spatiotemporal influence model to obtain the spatiotemporal influence potential; S2. Based on the spatiotemporal influence potential, the influence gradient is calculated, and the modulus of the influence gradient is compared with a preset threshold to generate a coupling relationship evaluation result; S3. In response to the coupling relationship evaluation results, using the state evolution model, perform evolution calculations on the current core indicators of the target asset to predict and generate future core indicators; S4. Generate intervention strategies based on future core indicators, and after executing the intervention strategies, use the prediction errors to adaptively calibrate the parameters in the state evolution model; S1 includes: collecting the core elements of the source assets from the spatiotemporal asset catalog, the core elements including spatiotemporal coordinates and core business indicators; Normalize the core business indicators to obtain the event source strength; build a spatiotemporal impact potential model by combining the event source strength, the preset time decay coefficient, and the preset regularization length; use the spatiotemporal impact potential model to solve and obtain the spatiotemporal impact potential.

2. The cloud-native spatiotemporal fusion analysis method based on spatiotemporal asset catalog according to claim 1 is characterized in that: The initial calibration method of the regularization length includes calibrating the core range or equivalent radius of the asset as the regularization length based on the type and physical properties of the asset; The initial calibration method of the time decay coefficient includes extracting the influence decay data of historical similar events and fitting the data with an exponential decay model to obtain the time decay coefficient.

3. The cloud-native spatiotemporal fusion analysis method based on spatiotemporal asset catalog according to claim 1 is characterized in that S2 include: Calculate the negative gradient of the spatiotemporal influence potential to obtain the influence gradient; Collect the gradient modulus generated by events that have not caused changes in the core indicators of the target assets in history, and calibrate the high percentile points as the monitoring response threshold; collect the gradient modulus generated by events that have caused sudden changes in the core indicators of the target assets in history, and calibrate the low percentile points as the critical warning threshold.

4. The cloud-native spatiotemporal fusion analysis method based on spatiotemporal asset catalog according to claim 3 is characterized in that: The steps to generate coupling relationship assessment results include: When the mode length of the influencing gradient is less than or equal to the monitoring response threshold, the coupling relationship evaluation result is determined to be weak coupling; When the modulus of the impact gradient is greater than the monitoring response threshold and less than or equal to the critical warning threshold, the coupling relationship assessment result is determined to be strong coupling; When the modulus of the influencing gradient is greater than the critical warning threshold, the coupling relationship evaluation result is determined to be critical coupling.

5. The cloud-native spatiotemporal fusion analysis method based on spatiotemporal asset catalog according to claim 1 is characterized in that S3 include: For asset pairs whose coupling relationship assessment results are strong coupling or critical coupling, the state evolution model is used to perform evolutionary calculations on the current core indicators of the target assets based on the impact gradient, the sensitivity vector and capacity parameters of the target assets to predict and generate future core indicators.

6. The cloud-native spatiotemporal fusion analysis method based on spatiotemporal asset catalog according to claim 5 is characterized in that: The capacity parameters are calibrated by obtaining the extreme values ​​of the physical properties or historical data of the target assets. The calibration method of the sensitivity vector includes: performing weighted average calculation based on the historical data of the functional layout of the target assets to determine the direction; and determining the modulus by performing linear regression analysis on the historical impact gradient and the change in core indicators.

7. The cloud-native spatiotemporal fusion analysis method based on spatiotemporal asset catalog according to claim 1 is characterized in that: S4 includes: Compare future core indicators with preset business alarm thresholds to identify risky assets. Calculate the system instability index based on the risk asset's business weight, impact gradient, and monitoring response threshold. Generate an intervention strategy based on the system instability index and preset business security thresholds, with the principle of minimizing total intervention cost. Business weights are quantified through the analytic hierarchy process or contribution analysis driven by historical data.

8. The cloud-native spatiotemporal fusion analysis method based on spatiotemporal asset catalog according to claim 1 is characterized in that: The steps of adaptive calibration include: After the intervention strategy is executed, the actual core indicators of the target assets are tracked; the prediction error between the actual core indicators and the future core indicators is calculated; and the prediction error is used to correct the sensitivity vector in the state evolution model.

Images