Closed-loop monitoring and control method, device and equipment based on target driving and medium

Abstract

The invention discloses a closed-loop monitoring and control method and device based on target driving, equipment and a medium. Comprising the steps of obtaining a system observation vector and a prediction vector, and obtaining overall deviation based on deviation of the observation vector or the prediction vector relative to a monitoring target; decomposing the total deviation based on the context state of the system to obtain a controllable deviation; issuing an action strategy and an effect specification based on the context state and the controllable deviation; and executing the action strategy, performing process monitoring and effect monitoring based on the effect specification, and driving maintenance or adjustment of the strategy based on a monitoring result until the controllable deviation is smaller than a preset threshold value. According to the target-driven closed-loop monitoring and control method provided by the invention, a closed loop of target-perception-evaluation-decision-making-monitoring execution-regeneration-trace reserving is formed, and continuous monitoring and active intervention of the system are realized.

Description

Technical Field

[0001] This application relates to the field of intelligent monitoring and control technology, and more specifically, to a target-driven closed-loop monitoring and control method, device, equipment, and medium. Background Technology

[0002] With the deepening of digital transformation, key areas such as intelligent transportation, data center operation and maintenance, and industrial process control have significantly increased their requirements for system stability, efficiency, and compliance. These systems typically operate under high concurrency, long durations, and across multiple scenarios, continuously generating massive amounts of heterogeneous data. Any deviation in state that cannot be monitored and intervened in a timely and accurate manner may be amplified into service interruptions, performance degradation, or compliance risks.

[0003] Currently, the monitoring paradigm commonly used in the industry is mainly based on data visualization and static threshold alarms. Historical averages or empirical thresholds are compared with real-time data, and an alarm is triggered when the threshold is exceeded. On-duty personnel then manually handle the situation using SOPs (Standard Operating Procedures).

[0004] Current monitoring systems rely solely on static thresholds for early warning, making it difficult to adaptively generate action strategies for control and adjust targets based on real-time system operation. Summary of the Invention

[0005] This application provides a target-driven closed-loop monitoring and control method, apparatus, device, and medium to at least solve the technical problem in related technologies that it is difficult to perform closed-loop monitoring and control of a system based on a target.

[0006] According to one aspect of the embodiments of this application, a target-driven closed-loop monitoring and control method is provided, comprising: The monitoring target is determined, the system observation vector and prediction vector are obtained, and the overall deviation is obtained based on the deviation of the observation vector or prediction vector relative to the monitoring target; The overall deviation is decomposed based on the system context state to obtain the controllable deviation; Based on the aforementioned context state and effective risk value, release the action strategy and effect specifications; The action strategy is executed, and the execution process of the action strategy is monitored. The effect of the action strategy is monitored based on the effect specifications. The monitoring results drive the maintenance or adjustment of the strategy until the controllable deviation is less than a preset threshold.

[0007] According to another aspect of the embodiments of this application, a target-driven closed-loop monitoring and control device is also provided, comprising: The gap assessment module is used to determine the monitoring target, obtain the system observation vector and prediction vector, and obtain the overall deviation based on the deviation of the observation vector or prediction vector relative to the monitoring target. The context attribution module is used to decompose the overall deviation based on the system context state to obtain controllable deviations; The strategy publishing module is used to publish action strategies and effect specifications based on the context state and the controllable deviation. The execution closed-loop module is used to execute the action strategy, monitor the execution process of the action strategy, monitor the execution effect based on the effect specifications, and drive the maintenance or adjustment of the strategy based on the monitoring results until the controllable deviation is less than a preset threshold.

[0008] According to another aspect of the embodiments of this application, an electronic device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the above-described target-driven closed-loop monitoring and control method through the computer program.

[0009] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, wherein a computer program is stored in the computer program, and the computer program is configured to execute the above-described target-driven closed-loop monitoring and control method at runtime.

[0010] The technical solutions provided in this application embodiment may include the following beneficial effects: This application provides a target-driven closed-loop monitoring and control method, which constructs a closed-loop control system integrating target management, perception estimation, gap assessment, context processing, strategy decision-making and dual-loop monitoring.

[0011] The system first obtains the overall deviation based on a preset detection target. Then, it introduces contextual information to decompose the overall deviation, obtaining controllable deviations that can be adjusted. Monitoring and execution strategies are made only for controllable deviations. Based on the controllable deviations, the optimal action strategy is selected under contextual constraints, and the strategy execution process is monitored in a dual-loop manner, focusing on process compliance and execution effectiveness. The monitoring results drive the strategy to maintain or adjust until the controllable deviation falls below a preset threshold, achieving the preset monitoring target. This realizes an automated closed loop from deviation perception to control decision-making to control execution and achievement of control targets, ensuring the quantifiable evaluation and safe controllability of the strategy. Furthermore, by tracking process compliance and effectiveness achievement in parallel, it provides objective and credible evidence to support the handling actions, effectively solving core problems such as target absence, closed-loop incompleteness, environmental insensitivity, and unverifiable execution. This significantly improves the intelligence level, intervention accuracy, and operational efficiency of monitoring and control in complex systems. Attached Figure Description The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart of a target-driven closed-loop monitoring and control method according to an embodiment of this application; Figure 2 This is a flowchart of another target-driven closed-loop monitoring and control method according to an embodiment of this application; Figure 3 This is a schematic diagram of a target-driven closed-loop monitoring and control device according to an embodiment of this application; Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0012] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0013] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0014] Currently, most existing monitoring systems are still at the stage of data visualization + static threshold alarms. The current solutions have the following shortcomings: First, it only "looks at the data" instead of "achieves the goal," lacks a KPI-oriented target trajectory / tolerance zone as a reference, lacks target alignment and adaptive capabilities, and cannot dynamically adjust the monitoring target according to the real-time operating situation.

[0015] Secondly, the lack of strategic closed-loop control means that the process from "discovering deviations" to "selecting strategies, executing them, verifying them, and reviewing them" is not closed-loop. The effectiveness of the strategies cannot be quantified, and it is difficult to automatically generate and securely release action strategies.

[0016] Furthermore, contextual factors such as weather, holidays, construction, and supply changes were not systematically modeled, leading to false alarms and misdiagnosis.

[0017] Furthermore, the process and its effects are unverifiable, lacking an objective assessment of the compliance and effectiveness achieved during the handling process, resulting in a lack of basis for post-event review and strategy optimization. These pain points urgently require a goal-driven closed-loop monitoring and control method to address.

[0018] Finally, the lack of process and outcome evidence regarding whether the contingency plans were implemented step-by-step and achieved the expected results makes post-mortem analysis and optimization difficult. These deficiencies directly lead to high invalid alarm rates, excessively long response times, SLA (Service Level Agreement) breaches, and cascading effects.

[0019] Based on this, the embodiments of this application provide a target-driven closed-loop monitoring and control method, which has the following beneficial effects: This application achieves a paradigm shift from traditional static threshold alarms to dynamic goal achievement by constructing an intelligent closed-loop control system based on "goal-driven, environment-aware, and dual-loop monitoring." Using clear goals and duty tolerances as anchors, the system continuously evaluates goal achievement and corrects deviations; it suppresses boundary jitter and short-term noise through strategies such as hysteresis, significantly reducing false alarms and frequent alarms; it automatically selects more likely effective actions based on "expected risk reduction," employing a progressive release mechanism to reduce secondary risks from ineffective or harmful strategies, shorten recovery time, and improve service achievement rates; simultaneously, it implements dual-loop monitoring, ensuring "verifiable execution and verifiable results."

[0020] By incorporating contextual factors such as weather, holidays, construction, and supply changes into the calculations, targets, thresholds, and control limits are dynamically adjusted according to the scenario. Deviations are broken down into "environmentally explainable parts" and "controllable parts," triggering alarms and policies only for the controllable parts, reducing false alarms and focusing resources. The entire process records criteria, versions, contexts, policies, actions, execution receipts, and effect evaluations, supporting compliance audits and cross-regional and cross-time period comparative analysis, providing a reusable evidentiary basis for continuous optimization.

[0021] The target-driven closed-loop monitoring and control method of this application, as described below with reference to the accompanying drawings, will be described in detail. Figure 1 As shown, the method mainly includes the following steps: S101 determines the monitoring target, obtains the system observation vector and prediction vector, and obtains the overall deviation based on the deviation of the observation vector or prediction vector relative to the monitoring target.

[0022] Considering that real systems have fluctuations and measurement noise, in one embodiment of this application, the monitoring target is defined by an interval method. The monitoring target is an interval with a certain fluctuation range. When the actual value or predicted value of the monitoring indicator falls within the interval, it is considered to be "meeting the standard". The part outside the interval is considered to be "deviation". This monitoring target is the target of system monitoring and control. Through control, the monitoring indicators of the system are made to reach the monitoring target interval range.

[0023] In one implementation, the monitoring target can be adjusted according to the scenario. Under the constraints of the environmental context, the monitoring target can be modified and adjusted to obtain a monitoring target that is adapted to the context.

[0024] In one implementation, before acquiring the system observation vector and prediction vector, the method further includes pre-defining monitoring targets and indicator weights. This includes defining monitoring targets and indicator weights for each monitoring indicator; before acquiring the system observation vector and prediction vector, the method further includes pre-setting hierarchical thresholds to classify risks, setting entry thresholds, exit thresholds, and time conditions for each risk level, wherein the entry threshold is greater than the exit threshold.

[0025] In one implementation, a risk value is determined based on a controllable deviation, and a grading threshold is further determined, wherein the mapping relationship between the controllable deviation and the risk value is preset.

[0026] Specifically, the system monitoring objectives are first defined. Considering that real systems have fluctuations and measurement noise, this application uses an interval-based approach to define the monitoring objectives of the monitoring indicators.

[0027] In one implementation, a monitoring target for a certain monitoring indicator of a certain monitoring object at time t is defined. :

[0028] in, For the time domain, To monitor the lower boundary of the target, To monitor the upper limit of the target, and have .

[0029] Optionally, It can be a constant range, such as consistent throughout the day, meaning the target's upper and lower bounds remain constant throughout the day; it also supports time-segmented interval modes that switch regularly according to preset time periods, work shifts, or seasons; and a curve mode that can continuously and smoothly adjust with higher precision, such as every minute or every 5 minutes. Furthermore, it can also define differentiated modes based on specific scenarios. This application does not impose any restrictions.

[0030] When the actual or predicted value of the monitoring indicator falls within the monitoring target, it is considered "meeting the standard"; the portion exceeding the range is considered "deviation," and the deviation is recorded as follows: Deviations usually bring risks, and the risk value is denoted as . .

[0031] Furthermore, when the system needs to jointly evaluate multiple indicators, weights need to be assigned to each indicator to achieve fusion, aggregation, and horizontal comparison. Assume the system has a total of... The first indicator, the The non-negative weights of each indicator are: Define the weight matrix for:

[0032] in, It is a diagonal matrix, which facilitates the weighted aggregation of deviations from various indicators into an overall deviation. Before weighting and aggregating the overall deviation, different indicators need to be aligned in terms of scope and units, or standardized, before weighting. The specific weighting configuration strategy is scenario-adaptive and can be dynamically adjusted according to business logic. For example, in the spatial dimension, the weight of the main road indicator can be assigned higher than that of the secondary road indicator, or in the time dimension, key time periods can be assigned higher weights than general time periods. Finally, the weights are normalized to facilitate horizontal comparison of the aggregated results.

[0033] Furthermore, by mapping controllable deviations to risk values, and to improve alarm robustness, this application addresses risk values. A tiered threshold and hysteresis mechanism were set up for risk values. Set entry threshold and exit threshold ,satisfy:

[0034] Specifically, when First rise exceeding At that time, it was considered to have "entered an alarm event"; later, it needed to be lowered to a lower level. Only then is it considered "restore / exit".

[0035] Entering a threshold higher than exiting a threshold is considered "hysteresis," which can prevent frequent jitter and false alarms caused by small fluctuations near the boundary.

[0036] Furthermore, to achieve robust alarms, the system will continuously update risk values. Mapped to discrete, multi-level alarm states, each alarm level is independently configured with entry and exit thresholds exhibiting hysteresis characteristics, along with an additional time-duration criterion. Specifically, the system requires a risk value... Not only must an alarm state be entered if it continuously exceeds a certain entry threshold for a preset duration, but it must also continuously remain below a lower exit threshold for a specified duration before the state can be restored. This dual criterion of "threshold hysteresis" and "time tolerance" effectively filters out instantaneous risk peaks or troughs caused by random fluctuations, thereby fundamentally avoiding the "sawtooth" phenomenon of alarm states frequently jumping near boundaries.

[0037] In a specific embodiment of this application, the target management module is configured as the system's parameter control center, and its core function is to manage the target set. Weight set and thresholds / control limits These three core parameter categories are uniformly defined and maintained throughout their entire lifecycle. The target set includes monitoring targets and monitoring indicators, while the weight set includes the weights of each monitoring indicator. This module provides enterprise-level parameter governance capabilities, including the creation, tracing, comparison, and archiving of parameter versions. It supports the gradual implementation of new parameter versions to the target range in a canary release mode and allows for quick one-click rollback of parameters when needed. The creation, updating, and activation of all parameters are communicated to downstream modules in real time via event messages, ensuring that all system components always perform calculations and decisions based on the latest and consistent parameter view.

[0038] Furthermore, the current observation vector of the system is collected, and the observation vector, historical data, and localized parameters are input into the pre-trained prediction model to obtain the prediction vector representing the future state of the system.

[0039] Specifically, the observation vector of the acquisition system at the current moment. Characterization system in A standardized core data unit representing multiple monitoring indicators at different times and in multiple dimensions. An array of observations at time points.

[0040] Furthermore, using the currently output standardized observation vector As a baseline for the system's current state, the data, combined with the regularities inherent in historical time-series data and localized parameters, are input into a pre-trained prediction model for forward inference. This model integrates the above inputs to deduce the future step size. The expected state at that point yields the prediction vector. .

[0041] Optionally, the prediction model is implemented based on a machine learning framework that integrates historical time series patterns, real-time observation data, and localized context parameters. It can be trained on a large amount of historical data, such as using a deep learning model. Finally, the trained model is deployed as a microservice.

[0042] The observation vector and the prediction vector satisfy: ,

[0043] The observed and predicted vectors must meet strict "three-same" standards: same dimensions, same sampling period, and same timestamp alignment. Simultaneously, each output includes key metadata such as the model version number, data caliber version identifier, and context state snapshot, ensuring that the entire analysis chain, from gap calculation and risk assessment to strategy decision-making, is based on a unified and traceable data foundation. Furthermore, the system dynamically calculates a range for each time slice's data points. Quality score This score comprehensively reflects the reliability of the data collection and processing process, and is directly used by downstream modules to downweight or isolate low-quality or extremely abnormal samples.

[0044] Furthermore, for each monitoring indicator, the deviation of its observed vector or predicted vector relative to the monitoring target is calculated to obtain the independent deviation of each monitoring indicator; based on the indicator weights, the independent deviations of each monitoring indicator are combined into the overall deviation.

[0045] Specifically, let the first Each indicator at time actual value or predicted value Its deviation relative to the monitoring target is:

[0046] Among them, the predicted value The deviation calculation method is similar and will not be repeated here. The deviation vector is obtained as follows:

[0047] The multidimensional biases are weighted and combined into the total bias:

[0048]

[0049] Furthermore, after obtaining the overall deviation, it also includes mapping the overall deviation to a risk value.

[0050] Optionally, a monotonically non-decreasing function can be used. Mapping the overall deviation to a risk value:

[0051] Furthermore, the risk values ​​are smoothed, and the corresponding risk levels are triggered based on the smoothed risk values ​​and the entry and exit thresholds for each risk level. Active warnings and action strategies are issued based on the overall deviation, risk level, and risk value.

[0052] To further improve the stability of the trigger, it is possible to perform a certain operation before triggering. Execution Exponential Smoothing:

[0053] Based on the set tiered entry and exit thresholds, risk values ​​are mapped to risk levels:

[0054] When a risk value first and continuously exceeds the entry threshold of a certain level, the system automatically maps the risk status at that moment to the corresponding level and triggers a corresponding early warning notification. Thereafter, the early warning for that level is only considered lifted once the risk value falls below a lower exit threshold and remains stable. This achieves a tiered, proactive, and robust early warning system for risk levels, effectively avoiding early warning fluctuations caused by instantaneous volatility.

[0055] In one implementation, the system also includes outputting corresponding data such as overall deviation, risk value, risk level, and early warning events for subsequent traceability auditing, and further includes optimizing and issuing action strategies based on data such as overall deviation, risk value, and risk level.

[0056] S102 decomposes the overall deviation based on the system context state to obtain the controllable deviation.

[0057] In one implementation, the current environmental information and data caliber information of the system are first uniformly encoded to obtain the context state.

[0058] In the specific implementation of this application, multi-source variables related to the external environment and data caliber are systematically integrated. A series of key factors, including weather conditions, holiday information, special events, construction plans, system supply capacity, adjacent system status, and changes in data collection caliber and sampling mechanisms, are collected in real time, standardized, coded, and vectorized to form a dimensional... Context state vector:

[0059] Each time the context state is updated, the system generates a unique version identifier for the currently effective context state and manages the record, thereby ensuring that the environmental context on which risk assessment and strategy decisions depend at any time can be accurately traced and reproduced, providing a solid foundation for end-to-end explainability and auditing.

[0060] Furthermore, in order to distinguish reasonable fluctuations caused by the environment, the overall deviation is decomposed based on the context state to obtain controllable deviation and reasonable deviation.

[0061] Decompose the overall deviation based on the context state:

[0062] in, Indicates in the current context Under the given circumstances, reasonable deviations that occur even without intervention are deviations that cannot be eliminated. The controllable deviation that the strategy should focus on is the part of the deviation that can be reduced or eliminated in the current context through the control actions / strategies available to the system, such as scheduling, rate limiting, parameter adjustment, resource expansion, process reordering, etc.

[0063] Furthermore, the controllable deviation is mapped to a risk value to obtain an effective risk value.

[0064]

[0065] In the embodiments of this application, the execution results of the action strategy are monitored. When a scene switch is detected, a context event is generated, and the monitoring target, threshold, effective risk value, controllable deviation, reasonable deviation, etc., which are dynamically adjusted based on the context event are output for subsequent traceability audit and strategy optimization.

[0066] This application introduces a deviation decomposition mechanism, enabling the system to identify and accept such inherent noise, thus avoiding invalid alarms and unnecessary intervention costs. Controllable deviation clearly defines the target range that the system's control actions can actually influence and that can be reduced, directly targeting truly optimizable parts for resource allocation and action selection, thereby significantly improving the accuracy, effectiveness, and resource utilization efficiency of system intervention.

[0067] S103 releases action strategies and effect specifications based on contextual states and controllable deviations.

[0068] In one implementation, an executable action set is generated under contextual constraints. Action strategies that are expected to reduce controllable deviations or effective risk values ​​are selected preferentially, and effect specifications are generated simultaneously during deployment. The entire deployment process can employ a progressive security mechanism of "shadow verification → canary deployment → automatic rollback," thereby achieving precise risk control while ensuring the overall controllability and rollback capability of the system.

[0069] First, based on the current context state constraints, the set of executable action strategies is dynamically determined.

[0070]

[0071] in, This represents a dynamic set of hard constraints that comprehensively cover the system's safe operation boundaries, available resource limits, business compliance requirements, and mutual exclusion or dependency relationships between different actions. This feasible domain is not static but changes with the real-time collected context state. Dynamic refresh and real-time reconstruction are performed to ensure that the policy action set generated at any time meets the executability and security requirements of the current environment.

[0072] Furthermore, based on the maximization of the controllable deviation reduction as the selection criterion, the optimal action is selected from the action strategy set; Furthermore, based on the criterion of maximizing the reduction of effective risk value, the optimal action is selected from the set of action strategies.

[0073]

[0074] in, As an evaluation window, soft constraints such as cost, fairness, and service level can be added in practice, but the principle of "based on effective risk value" remains unchanged. The principle of "reducing the main objective" is to achieve the goal.

[0075] Furthermore, based on the optimal action and controllable deviation, the effect specification corresponding to the optimal action is generated. The effect specification includes the expected effect trajectory and the action target range corresponding to the optimal action.

[0076] In some implementations, the effect specification includes the actionified target range for that action, generated using a target correction operator.

[0077] In one implementation, let Moments released action In the release At the same time, it also includes generating the expected effect trajectory under this action.

[0078] Assume there is a total common Prediction for each time slice:

[0079] Among them, the function This is an effect prediction model, which can be obtained through historical playback, causal estimation, simulation, or machine learning. This application's embodiments do not impose specific limitations. Model parameters include the current state. Current context ,action and forward time slice The index.

[0080] Furthermore, based on monitoring objectives Generate actionable target range:

[0081] in, It is the target correction operator, in Under this premise, the target range for animation is calibrated a second time to generate the final target range for animation.

[0082] During secondary calibration, the upper and lower bounds of the monitoring target are shifted, relaxed, or adjusted based on the executed actions and the expected changes in system state they cause, while strictly maintaining the hysteresis relationship between the validity of the interval and the threshold. For example, in traffic control scenarios, when the system takes diversion actions, it will temporarily and appropriately relax the upper limit of the congestion index target of the upstream road segment in the initial stage of control based on the upstream traffic accumulation effect that the action may cause, effectively avoiding a large number of invalid "expected alarms".

[0083] The predicted trajectory serves as the direct basis for the recalibration of the target correction operator, revealing in advance the possible changes in the system state after the action is executed. Based on this prediction, the target correction operator adaptively recalibrates the monitoring target, which is adjusted according to the context state, to generate a reasonable monitoring baseline.

[0084] Employ a pre-defined progressive release process and rollback mechanism to release the optimal action and the target range for the action.

[0085] In the specific implementation of this application, the release of policies and action target ranges follows a strict security operation and maintenance process. The system releases the optimal action and its corresponding action target range, employing a gradual deployment mechanism of "shadow verification → canary release → full rollout." This process requires the policy to first be simulated in a shadow environment to verify its logical validity; after passing the test, it is then subjected to canary trial operation within limited system resources or business traffic, with continuous monitoring of its actual effects; only when the canary phase confirms that the expected goals have been achieved and no security red lines have been crossed will the policy be fully rolled out. Throughout the entire release and execution process, if deterioration in effect or violation of security constraints is detected, the system will automatically trigger a preset rollback procedure. This ensures a high degree of controllability and operational security of system intervention.

[0086] S104 executes action strategies, monitors the execution process of action strategies, monitors the execution effect based on effect specifications, and drives the maintenance or adjustment of strategies based on monitoring results until the controllable deviation is less than the preset threshold.

[0087] In one implementation, an action strategy is executed, and the execution process and effects of the action strategy are monitored. The compliance of the execution process and the compliance of the effects of the action strategy are monitored. In one implementation, firstly, an action strategy is executed, and the execution process and effect of the action strategy are monitored. The compliance of the execution process and the compliance of the effect of the action strategy are monitored. Based on the time compliance, work readiness gap value, and schedule gap value during the strategy execution period, the process compliance is calculated.

[0088] Time compliance:

[0089] in, To normalize the expected lateness duration, both the "probability of lateness" and the "magnitude of lateness" are considered. For predicting the "completion time" random variable; For steps The promised completion date This serves as the tolerance level for this step. The definition considers both the probability and magnitude of lateness, allowing for early risk assessment even when delays are anticipated but not yet expected.

[0090] Readiness gap value:

[0091] in, Indicate steps exist The work readiness rate at any given time is calculated based on the availability of dependencies / resources / preconditions, work orders, personnel, materials, upstream receipts, permissions, etc.

[0092] Schedule gap value:

[0093] in, Steps exist The percentage of work progress at any given moment.

[0094] Then the steps exist The process compliance at any given moment is defined as:

[0095] in, , .

[0096] Define the optimal action exist The definition of compliance at any given moment is:

[0097] in For steps The weighting coefficients.

[0098] Furthermore, based on the deviation between the action target interval and the observation vector, the effect compliance is obtained, and based on the process compliance and effect compliance, a joint criterion is obtained.

[0099] In the window Within, the target range of the published action. For reference, calculate the observations The out-of-bounds deviations are aggregated to obtain the compliance of the effect:

[0100]

[0101] in The observation vector was identified. The dimensions. Furthermore, based on process compliance and effect compliance, a joint criterion is obtained:

[0102] Joint criteria Set hysteresis, and let its input and output lines be respectively... and ,have .

[0103] When the joint criterion is greater than or equal to the preset entry threshold ( Once the time conditions are met, the decision-making process is triggered, and the strategy engine executes maintenance or adjustment.

[0104] In one implementation, when the conditions for triggering the decision-making process described above are met, the strategy engine executes a maintenance or adjustment strategy, including freezing the current monitoring target, weight, and threshold parameter status, collecting monitoring data during the execution of the action strategy; sending the frozen parameter status and monitoring data back to the strategy engine, republishing the action strategy and effect specifications; executing the republished action strategy, and monitoring the execution process based on the republished effect specifications.

[0105] Specifically, the process first freezes the current... The system first establishes a view to maintain the integrity of the problem scenario. Then, it packages all monitoring data, including process compliance, effect deviation, joint criteria, and strategy execution snapshots, into a decision evidence package and sends it back to the strategy engine module. Based on this evidence package, the strategy engine performs root cause analysis and then autonomously decides to roll back the strategy, branch out the handling plan, or recalculate the optimal action, thus forming a complete autonomous closed loop from monitoring to decision-making and optimization.

[0106] In one implementation, when the conditions for triggering the decision-making process described above are met, the strategy engine performs adjustment or re-optimization, including: obtaining the current context state and controllable deviation, and determining a new action strategy and effect specifications based on maximizing the reduction of controllable deviation as the optimization criterion. Execute step S104 above until the newly determined action strategy is completed. If the joint criterion is less than or equal to the preset exit threshold (…), then… Once the time condition is met, the system proceeds to the next time slice and continues monitoring. When the joint criterion drops below the exit threshold and remains stable, the system determines that the current strategy execution has returned to normal and automatically switches to the next monitoring cycle. Continue running. This cyclical monitoring process will continue until the policy is acted upon. When the complete execution cycle ends or the system reaches the predetermined termination conditions, the system will exit the special monitoring process for that strategy.

[0107] In some implementations, a predetermined termination condition, such as the controllable deviation of the monitored target reaching a preset threshold, is considered to have achieved the control target and reached the predetermined termination condition.

[0108] In one implementation, maintaining or adjusting a strategy based on monitoring results includes: Collect monitoring data during the execution of action strategies. The monitoring data includes context data, process compliance data, and effect compliance data. The effectiveness of action strategy execution is determined based on process compliance data and effect compliance data; When the execution effect of the action strategy fails to meet the preset effect specification, the monitoring target is adjusted according to the context data to obtain the adjusted monitoring target; The above steps S101-S104 are executed traversally according to the adjusted monitoring target until the controllable deviation is less than the preset threshold.

[0109] In one implementation, the system establishes a full-link data versioning and traceability mechanism for the target set. Weight set Threshold Context state Strategy Actions Action-based monitoring targets Process compliance Effect bias and joint criteria Key parameters and states are recorded in snapshot form and associated with version identifiers. Each system state change or policy execution generates an evidence package containing all the above data, and a unique fingerprint is generated using a cryptographic hash algorithm to ensure data immutability. This mechanism provides complete traceability support for system operation, enabling precise querying, comparative analysis, and review of the decision-making basis, execution effects, and system state at any point in history, while meeting stringent compliance audit requirements.

[0110] To facilitate understanding of the methods in the embodiments of this application, the following description is provided in conjunction with the appendix. Figure 2 Further description.

[0111] like Figure 2 As shown, it includes the following steps: S201: Target Management: Define monitoring targets, set indicator weights, and configure tiered thresholds.

[0112] S202: Perception Assessment: Acquire system observation vectors and estimate prediction vectors.

[0113] S203: Gap and Risk Assessment: Assess overall deviation and calculate risk value.

[0114] S204: Context State: Collect the context state, decompose the overall deviation into reasonable deviation and controllable deviation, and output the effective risk value for strategy decision-making.

[0115] S205: Policy Security Deployment: Generate a set of actionable actions under contextual constraints, select the best actions that are expected to reduce effective risks, and generate effect specifications simultaneously during deployment.

[0116] S206: Strategy Execution: Execute the strategy and monitor whether the handling process is carried out according to the plan and whether the monitoring effect is achieved as expected, and drive branching / upgrading / rollback with joint criteria.

[0117] S207: Audit: End-to-end versioned traceability, generating evidence packages and hash fingerprints, supporting query comparison, review and compliance audit.

[0118] This application transforms business KPIs into quantifiable, dynamic monitoring targets through a target alignment mechanism, enabling computers to automatically assess target achievement and accurately calculate deviations, thus solving the problem of target gaps in traditional monitoring. By introducing an environmental awareness module, the system can intelligently identify deviations caused by environmental factors and focus on controllable deviations for risk synthesis and strategy optimization. Combined with a dual-loop execution monitoring mechanism, it simultaneously tracks process compliance and effectiveness achievement, effectively ensuring the verifiability of strategy execution and closed-loop optimization capabilities. Finally, the system achieves adaptive parameter optimization through target regeneration and provides reliable evidence for audit review through full-link evidence-based record keeping. Overall, it realizes a paradigm shift from static alarms to dynamic target-driven systems, from open-loop handling to closed-loop optimization, and from experience-based decision-making to evidence-based management, significantly improving the operational efficiency and risk response capabilities of complex systems.

[0119] According to another aspect of the embodiments of this application, a target-driven closed-loop monitoring and control device for implementing the above-described target-driven closed-loop monitoring and control method is also provided. For example... Figure 3 As shown, the device includes: The gap assessment module 301 is used to determine the monitoring target, obtain the system observation vector and prediction vector, and obtain the overall deviation based on the deviation of the observation vector or prediction vector relative to the monitoring target.

[0120] The context attribution module 302 is used to decompose the overall deviation based on the system context state to obtain controllable deviation.

[0121] The strategy publishing module 303 is used to publish action strategies and effect specifications based on context state and controllable deviation.

[0122] The closed-loop module 304 is used to execute action strategies, monitor the execution process of action strategies, monitor the execution effect based on effect specifications, and drive the maintenance or adjustment of strategies based on monitoring results until the controllable deviation is less than a preset threshold.

[0123] It should be noted that the target-driven closed-loop monitoring and control device provided in the above embodiments is only illustrated by the division of the above functional modules when executing the target-driven closed-loop monitoring and control method. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the target-driven closed-loop monitoring and control device and the target-driven closed-loop monitoring and control method embodiments provided in the above embodiments belong to the same concept, and the implementation process is detailed in the method embodiments, which will not be repeated here.

[0124] According to another aspect of the embodiments of this application, an electronic device corresponding to the target-driven closed-loop monitoring and control method provided in the foregoing embodiments is also provided, so as to execute the above-described target-driven closed-loop monitoring and control method.

[0125] Please refer to Figure 4 This illustrates a schematic diagram of an electronic device provided by some embodiments of this application. For example... Figure 4 As shown, the electronic device includes: a processor 400, a memory 401, a bus 402, and a communication interface 403. The processor 400, the communication interface 403, and the memory 401 are connected via the bus 402. The memory 401 stores a computer program that can run on the processor 400. When the processor 400 runs the computer program, it executes the target-driven closed-loop monitoring and control method provided in any of the foregoing embodiments of this application.

[0126] The memory 401 may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 403 (which can be wired or wireless), such as the Internet, wide area network, local area network, or metropolitan area network.

[0127] Bus 402 can be an ISA bus, PCI bus, or EISA bus, etc. Buses can be divided into address buses, data buses, control buses, etc. Memory 401 is used to store programs. After receiving execution instructions, processor 400 executes the program. The target-driven closed-loop monitoring and control method disclosed in any of the foregoing embodiments of this application can be applied to processor 400, or implemented by processor 400.

[0128] The processor 400 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of the processor 400 or by instructions in software form. The processor 400 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules may reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 401. The processor 400 reads the information in memory 401 and, in conjunction with its hardware, completes the steps of the above method.

[0129] The electronic device provided in this application embodiment and the target-driven closed-loop monitoring and control method provided in this application embodiment are based on the same inventive concept and have the same beneficial effects as the methods they adopt, operate or implement.

[0130] According to another aspect of the embodiments of this application, a computer-readable storage medium corresponding to the target-driven closed-loop monitoring and control method provided in the foregoing embodiments is also provided, wherein a computer program (i.e., a program product) is stored thereon, and when the computer program is run by a processor, it executes the target-driven closed-loop monitoring and control method provided in any of the foregoing embodiments.

[0131] It should be noted that examples of computer-readable storage media may also include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical and magnetic storage media, which will not be elaborated here.

[0132] The computer-readable storage medium provided in the above embodiments of this application and the target-driven closed-loop monitoring and control method provided in the embodiments of this application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0133] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0134] The above embodiments merely illustrate several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this patent should be determined by the appended claims.

Claims

1. A target-driven closed-loop monitoring and control method, characterized in that, include: S101 Determine the monitoring target, obtain the system observation vector and prediction vector, and obtain the overall deviation based on the deviation of the observation vector or prediction vector relative to the monitoring target; S102 decomposes the overall deviation based on the system context state to obtain the controllable deviation; S103, based on the aforementioned context state and controllable deviation, releases the action strategy and effect specifications; S104 executes the action strategy, monitors the execution process of the action strategy, monitors the effect based on the effect specifications, and drives the maintenance or adjustment of the strategy based on the monitoring results until the controllable deviation is less than a preset threshold.

2. The method according to claim 1, characterized in that, Before obtaining the system observation vector and prediction vector, the following steps are also included: Define target ranges and indicator weights for each monitoring indicator; An entry threshold, an exit threshold, and a time condition are set for each risk level, wherein the entry threshold is greater than the exit threshold, and the risk level is classified according to the risk value.

3. The method according to claim 2, characterized in that, Obtain the system observation vector and prediction vector, and based on the deviation of the observation vector or prediction vector relative to the monitored target, obtain the overall deviation, including: The current observation vector of the acquisition system is collected, and the observation vector, historical data and localization parameters are input into the pre-trained prediction model to obtain the prediction vector representing the future state of the system. For each monitoring indicator, the deviation of its observed vector or predicted vector relative to the monitoring target is calculated to obtain the independent deviation of each monitoring indicator. Based on the indicator weights, the independent deviations of each monitoring indicator are combined into the overall deviation.

4. The method according to claim 1, characterized in that, The overall deviation is decomposed based on the system context state to obtain controllable deviations, including: The current system environment information and data caliber information are uniformly encoded to obtain the context state; Based on the context state, the overall deviation is decomposed to obtain controllable deviation and reasonable deviation; The controllable deviation is mapped to a risk value to obtain an effective risk value.

5. The method according to claim 1, characterized in that, Based on the aforementioned context state and controllable deviations, the action strategy and effect specifications are released, including: Based on the current context state constraints, dynamically determine the set of executable action strategies; Based on the maximization of the controllable deviation reduction as the selection criterion, the optimal action is selected from the set of action strategies; Based on the optimal action and the controllable deviation, the effect specification corresponding to the optimal action is generated, wherein the effect specification includes the expected effect trajectory and the action target range corresponding to the optimal action; Publish the optimal action and effect specifications.

6. The method according to claim 1, characterized in that, Execute the action strategy, monitor the execution process based on the action strategy, monitor the execution effect based on the effect specifications, and drive the maintenance or adjustment of the strategy based on the monitoring results until the controllable deviation is less than a preset threshold, including: Calculate process compliance based on time compliance, work readiness gap, and schedule gap during strategy execution; Based on the deviation between the effect specification and the observation vector, the effect compliance is obtained; based on the process compliance and the effect compliance, a joint criterion is obtained. Once the joint criterion is greater than or equal to the preset entry threshold and the time condition is met, the decision-making process is triggered, and the strategy engine re-selects the best option. Once the joint criterion is less than or equal to the preset exit threshold and the time condition is met, the process proceeds to the next time slice and continues monitoring until the action strategy is completed.

7. The method according to claim 6, characterized in that, The strategy engine will re-select the best option, including: Obtain the current context state and controllable deviation, and determine a new action strategy and effect specifications based on the maximization of the reduction of the controllable deviation as the selection criterion. Execute step S104 until the action strategy is completed.

8. The method according to claim 1, characterized in that, The maintenance or adjustment of strategies is driven by monitoring results; Collect monitoring data during the execution of the action strategy, including context data, process compliance data, and effect compliance data; The execution effect of the action strategy is determined based on the process compliance data and the effect compliance data. When the execution effect of the action strategy fails to meet the preset effect specification, the monitoring target is adjusted according to the context data to obtain the adjusted monitoring target; According to the adjusted monitoring target, steps S101-S104 are executed traversally until the controllable deviation is less than the preset threshold.

9. A target-driven closed-loop monitoring and control device, characterized in that, include: The gap assessment module is used to determine the monitoring target, obtain the system observation vector and prediction vector, and obtain the overall deviation based on the deviation of the observation vector or prediction vector relative to the monitoring target. The context attribution module is used to decompose the overall deviation based on the system context state to obtain controllable deviations; The strategy publishing module is used to publish action strategies and effect specifications based on the context state and controllable deviations. The execution closed-loop module is used to execute the action strategy, monitor the execution process of the action strategy, monitor the execution effect based on the effect specifications, and drive the maintenance or adjustment of the strategy based on the monitoring results until the controllable deviation is less than a preset threshold.

10. An electronic device, characterized in that, It includes a processor and a memory storing program instructions, the processor being configured to execute, when executing the program instructions, the target-driven closed-loop monitoring and control method as described in any one of claims 1 to 8.

11. A computer-readable medium, characterized in that, It stores computer-readable instructions, which are executed by a processor to implement a target-driven closed-loop monitoring and control method as described in any one of claims 1 to 8.

Images