Intelligent control method based on big data

Abstract

The invention relates to the technical field of data management, in particular to an intelligent control method based on big data, and the method comprises the following steps: S1, collecting operation state data and environment information of a target system, constructing a control intention vector model in combination with a task target and a control boundary, and packaging the control intention vector model into a migration data object; s2, in the control execution process, collecting a state flow and a feedback sequence, detecting behavior drift, dynamically correcting a control strategy, and outputting a corrected control sequence; and S3, applying the correction strategy to a target system, collecting an execution response, identifying behavior chain evolution, updating a control intention model, outputting an evolvable control intention chain, and supporting a control task of a next period. According to the method, real-time data acquisition, behavior offset identification and control intention chain evolution are combined, and dynamic adjustment and adaptive optimization of a control strategy are realized, so that the stability, robustness and long-term adaptive capacity of the system are improved.

Description

Technical Field

[0001] This invention relates to the field of data management technology, and in particular to an intelligent control method based on big data. Background Technology

[0002] With the rapid development of industrial automation and intelligence, systems based on big data and intelligent control have become one of the key technologies in modern complex engineering fields. Intelligent control systems usually need to collect a large amount of data in real time, including system operating status, environmental context information, etc., to support decision-making and optimize control strategies. Traditional control methods mostly rely on manually set models and rules, which are difficult to adapt to dynamically changing complex environments and real-time feedback, resulting in a lack of flexibility and adaptability in the system during execution.

[0003] While existing intelligent control methods have made some progress in certain fields, most of them suffer from problems such as slow response, untimely adjustment, and inability to effectively cope with changing environments. Many traditional control systems have fixed model settings and lack adaptive capabilities, which makes the system prone to control drift or unable to correct its strategy in time when facing environmental changes. In addition, existing technologies usually rely on a single control objective and constraint, and fail to effectively integrate dynamic adjustment mechanisms, making it difficult to achieve continuous optimization and long-term stable control effects. Summary of the Invention

[0004] This invention provides an intelligent control method based on big data, which can detect and correct behavioral deviations in real time during control execution and dynamically adjust control strategies, thereby improving the adaptability and accuracy of the control system. At the same time, through the evolution and updating of the control intent chain, the system can continuously optimize control tasks in different cycles, improve system stability and robustness, and meet the intelligent control needs in complex environments.

[0005] A big data-based intelligent control method includes the following steps:

[0006] S1, collect the target system's operational status data and environmental context information in the early stage of control, combine the task objective and control boundary, construct a control intent vector model including the desired control direction, allowable disturbance range and control action boundary, and encapsulate the control intent vector model as a transfer data object;

[0007] S2, during the control execution process, collects the state flow and feedback behavior sequence of the controlled object in real time, and measures the offset between the current behavior sequence and the control intention vector model through the behavior offset recognition mechanism. If a behavior drift event is detected, the control strategy path is dynamically corrected based on the intention boundary constraints, and the updated corrected control strategy sequence is output.

[0008] S3 applies the modified control strategy sequence to the target system, collects its execution response and boundary triggering behavior, identifies the evolution trend of the control behavior chain based on the causal response inference mechanism, updates the structure of the control intent vector model, and outputs the control intent chain to support the control task of the next cycle.

[0009] Optionally, S1 includes:

[0010] S11: Collect the operational status data and environmental context information of the target system in the early stage of control. The operational status data includes key control variables, state feedback sequences and boundary response records. The environmental context information includes external disturbance sources, operational constraints and historical event labels. Through a unified time axis alignment mechanism and context event indexing mechanism, a time-series-structure composite dataset reflecting the initial behavioral background of the system is constructed.

[0011] S12, based on the operating status data and environmental context information, parse the current control task objective and extract the control constraint boundary conditions. The task objective includes the expected system state and its range of achievement. The control constraint boundary conditions include the system adjustment frequency, amplitude, and time delay. Through the objective-boundary joint coding mechanism, a structured control objective expression unit is generated.

[0012] S13 combines the constructed time-series-structured dataset with structured control target expression units and utilizes a control semantic alignment mechanism to generate a control intent vector model that includes the desired control direction, permissible disturbance range, and control action boundary.

[0013] Optionally, S11 includes:

[0014] S111, collect multi-source data assets in the early stage of control and divide them into operational status data and environmental context information, among which;

[0015] The operational status data includes key control variables (such as temperature, current, and position). State feedback sequence Boundary response record ε bnd ;

[0016] The environmental context information includes external disturbance sources. Operational constraints and historical event tags

[0017] S112 employs a centralized sliding window mechanism to construct a unified time axis τ, and in each time slice t i The convergence of normalized data representations forms a time synchronization vector. Generate a synchronization time series matrix M sync ;

[0018] S113, Construct an event-driven structured index graph, and use an event-state mapping function to associate time synchronization data with event triggering logic to establish a context event index set G. evt and compare it with the time series matrix M sync Combined, a structural map of time-event coupling is formed.

[0019] Optionally, S12 includes:

[0020] S121, based on operational status data and environmental context data, uses a status index mapping function to identify the control objective within the current cycle.

[0021] S122, parse the execution boundary constraints related to the control task, extract the maximum adjustment frequency, adjustment magnitude, and response delay under the current context, and construct the control boundary set.

[0022] S123, control target With control boundary set Joint encoding is performed to generate structured control target representation units g. k .

[0023] Optionally, S13 includes:

[0024] S131, the time series matrix M sync With structured control target expression unit g k Perform pairings to construct a variable-state association graph.

[0025] S132, based on the variable-state association graph Generate control semantic labels, including desired control direction. (Upward, downward, or unchanged), permissible disturbance range and control boundary parameters (adjustment frequency limit) Amplitude Limitation With response latency limitations );

[0026] S133, merge control semantic labels into a control intent vector q i Ultimately, a control intent vector model Q is generated. intent .

[0027] Optionally, S2 includes:

[0028] S21, During the execution of the control strategy, the continuous state flow of the controlled object is collected. With feedback behavior sequence The data is then time-aligned and normalized according to a uniform sampling period to form the control behavior observation matrix M. beh ;

[0029] S22, based on the control intent vector model Q intent The behavioral offset metric function is used to calculate the behavioral offset between the current state and the intended expectation. k ,like Then it is determined that the k-th variable has experienced a drift event, where, The threshold for determining behavioral deviation;

[0030] S23, for the key control variables that cause drift, read the control boundary parameters (adjustment frequency limit) from their control intent vector. Amplitude Limitation With response latency limitations It dynamically adjusts the control frequency and adjustment range of the control strategy path, generates a corrected control strategy path, and finally outputs an updated corrected control strategy sequence.

[0031] Optionally, S21 includes:

[0032] S211, During the execution of the control strategy, the real-time state variables of the controlled object are continuously collected to form a state flow. Simultaneously, the actual execution results of the control strategy response are collected to form a feedback behavior sequence.

[0033] S212, state flow With feedback behavior sequence Synchronization is performed using a uniform sampling period Δt. If time offset or data loss exists, an interpolation function is used. Complete state flow With feedback behavior sequence Alignment;

[0034] S213, for all aligned state flows With feedback behavior sequence After normalization, the final control behavior observation matrix M is formed. beh .

[0035] Optionally, S23 includes:

[0036] S231, for the i-th key control variable that has detected drift, from the corresponding control intent vector q i Extract control boundary constraint parameters to form the control boundary parameters, including adjustment frequency limits. Amplitude Limitation With response latency limitations

[0037] S232, in the control strategy path Based on the control boundary parameters, the control strategy path is modified, including frequency limits (control step limits) and amplitude limits (control action limits), ultimately yielding the modified control strategy path.

[0038] S233, the revised control strategy path As input, combined with response latency constraints Generate modified control strategy sequence

[0039] Optionally, S3 includes:

[0040] S31, modify the control strategy sequence from the previous cycle. The j-th key control variable applied to the target system is used to collect the state feedback sequence in real time during control execution. And record the set of boundary behavior events triggered during execution.

[0041] S32, construct the behavior evolution chain based on the collected state feedback sequence and the triggered boundary behavior event set. And combined with control intent vector With state feedback sequence Calculate the evolution trend vector of the control behavior chain

[0042] S33, based on the evolutionary trend vector Update the control intent vector based on the boundary trigger result. Forming the control intent version for the next cycle Ultimately, all updated control intentions are constructed into a control intention chain in chronological order.

[0043] The beneficial effects of this invention are:

[0044] This invention, through a big data-based intelligent control method, can effectively achieve real-time control and adaptive adjustment of a target system. By collecting real-time operational status data and environmental context information of the target system, and combining the task objectives and control boundaries, a control intent vector model is constructed, which can comprehensively reflect the system's behavior during execution. The update mechanism of this model provides strong support for the adjustment of control strategies, ensuring that the system can make precise adjustments based on feedback data and changes in control tasks.

[0045] This invention, through the introduction of a behavior drift recognition mechanism, can detect potential behavior drift in the system in real time, and then dynamically correct the control strategy path. This behavior drift-based control strategy correction method not only ensures the robustness of the control system in the face of external disturbances or changes in internal parameters, but also adaptively adjusts the strategy according to the actual deviation of the control target, greatly improving the stability and adaptability of the control system.

[0046] This invention provides a control framework for continuous learning and optimization through structured updating and chain evolution of the control intent vector model. The control intent for each cycle is iteratively updated based on historical responses and boundary triggering results, ensuring the long-term stability and accurate execution of the control task. In addition, the evolution of the control intent chain not only improves the automation and intelligence of the control task, but also enables the control system to have cross-cycle continuity and flexibility, and better cope with dynamic changes in complex environments. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a schematic diagram of the control method flow according to an embodiment of the present invention;

[0049] Figure 2 This is a schematic diagram illustrating the control intent generation in an embodiment of the present invention. Detailed Implementation

[0050] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. Those skilled in the art may employ other alternative methods to implement some well-known technologies; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0051] like Figures 1-2 As shown, an intelligent control method based on big data includes the following steps:

[0052] S1, collect the target system's operational status data and environmental context information in the early stage of control, combine the task objective and control boundary, construct a control intent vector model including the desired control direction, allowable disturbance range and control action boundary, and encapsulate the control intent vector model as a transfer data object;

[0053] S2, during the control execution process, collects the state flow and feedback behavior sequence of the controlled object in real time, and measures the offset between the current behavior sequence and the control intention vector model through the behavior offset recognition mechanism. If a behavior drift event is detected, the control strategy path is dynamically corrected based on the intention boundary constraints, and the updated corrected control strategy sequence is output.

[0054] S3 applies the modified control strategy sequence to the target system and collects its execution response and boundary triggering behavior. Based on the causal response inference mechanism, it identifies the evolution trend of the control behavior chain and updates the structure of the control intent vector model, outputting an evolvable control intent chain to support the control task of the next cycle.

[0055] S1 includes:

[0056] S11: Collect the operational status data and environmental context information of the target system in the early stage of control. The operational status data includes key control variables, state feedback sequences and boundary response records. The environmental context information includes external disturbance sources, operational constraints and historical event labels. Through a unified time axis alignment mechanism and context event indexing mechanism, a time-series-structure composite dataset reflecting the initial behavioral background of the system is constructed.

[0057] S12, based on the operating status data and environmental context information, parse the current control task objective and extract the control constraint boundary conditions. The task objective includes the expected system state and its range of achievement. The control constraint boundary conditions include the system adjustment frequency, amplitude, and time delay. Through the objective-boundary joint coding mechanism, a structured control objective expression unit is generated.

[0058] S13 combines the constructed time-series-structured dataset with structured control target expression units and utilizes a control semantic alignment mechanism to generate a control intent vector model that includes the desired control direction, permissible disturbance range, and control action boundary.

[0059] S11 includes:

[0060] S111, collect multi-source data assets in the early stage of control and divide them into operational status data and environmental context information, among which;

[0061] Operating status data includes key control variables (such as temperature, current, position, etc.). State feedback sequence Boundary response record ε bnd ;

[0062] Environmental context information includes external disturbance sources. Operational constraints and historical event tags

[0063] S112 employs a centralized sliding window mechanism to construct a unified time axis τ, and in each time slice t i The convergence of normalized data representations forms a time synchronization vector. Generate a synchronization time series matrix M sync , represented as:

[0064]

[0065] in, This is a time alignment function based on interpolation and moving average. On a standardized unified time axis τ, for each time point The generated normalized state vector;

[0066] S113, Construct an event-driven structured index graph, and use an event-state mapping function to associate time synchronization data with event triggering logic to establish a context event index set G. evt and compare it with the time series matrix M sync Combined, a structural map of time-event coupling is formed. Represented as:

[0067]

[0068] Among them, e j For the j-th event, In relation to event e j Time of occurrence t j The corresponding system state vector.

[0069] S12 includes:

[0070] S121, based on operational status data and environmental context data, uses a status index mapping function to identify the control objective within the current cycle. Represented as:

[0071]

[0072] Among them, s k For the k-th key control variable, This is the lower limit of the target for the key control variables (set as 90%-95% of the expected values ​​of the key control variables). Set the target upper limit for the key control variables (set as 105%-110% of the expected value of the key control variables);

[0073] S122, parse the execution boundary constraints related to the control task, extract the maximum adjustment frequency, adjustment magnitude, and response delay under the current context, and construct the control boundary set. Represented as:

[0074]

[0075] in, To the maximum allowable adjustment frequency, The maximum adjustment range in a single instance (set as 5%-10% of the key control variable), For the maximum tolerable response latency, K is the total number of objective dimensions involved in the control task;

[0076] S123, control target With control boundary set Joint encoding is performed to generate structured control target representation units g. k , represented as:

[0077]

[0078] S13 includes:

[0079] S131, the time series matrix M sync With structured control target expression unit g k Perform pairings to construct a variable-state association graph. Represented as:

[0080]

[0081] Among them, s i For the i-th key control variable, g i To control variable s i Associated structured control target representation units;

[0082] S132, based on the variable-state association graph Generate control semantic labels, including desired control direction. (Upward, downward, or unchanged), permissible disturbance range and control boundary parameters (adjustment frequency limit) Amplitude Limitation With response latency limitations ), represented as:

[0083]

[0084] Where, μ i For the key control variable s i The expected target value, For from M sync The current or historical average value of this dimension is extracted. These are direction labels: 1 indicates an upward adjustment, -1 indicates a downward adjustment, and 0 indicates maintaining the current position. Permissible disturbance range;

[0085] S133, merge control semantic labels into a control intent vector q i Ultimately, a control intent vector model Q is generated. intent , represented as:

[0086]

[0087] Q intent ={q1,q2,…,q K};

[0088] Where q1,q2,…,q K These are the 1st, 2nd, ..., Kth control intent vectors, respectively.

[0089] S2 includes:

[0090] S21, During the execution of the control strategy, the continuous state flow of the controlled object is collected. With feedback behavior sequence The data is then time-aligned and normalized according to a uniform sampling period to form the control behavior observation matrix M. beh ;

[0091] S22, based on the control intent vector model Q intent The behavioral offset metric function is used to calculate the behavioral offset between the current state and the intended expectation. k ,like Then it is determined that the k-th variable has experienced a drift event, where, The threshold for determining behavior offset is expressed as:

[0092]

[0093] in, For the k-th variable in the observation sequence at time τ j The value, Let k be the expected value of the k-th variable in the intention vector;

[0094]

[0095] in, Let λ be the average behavioral offset of the k-th control variable over the historical execution cycles. k σ is the adjustment coefficient. k is the standard deviation of the k-th control variable in the historical behavior offset sequence;

[0096] S23, for the key control variables that cause drift, read the control boundary parameters (adjustment frequency limit) from their control intent vector. Amplitude Limitation With response latency limitations It dynamically adjusts the control frequency and adjustment range of the control strategy path, generates a corrected control strategy path, and finally outputs an updated corrected control strategy sequence.

[0097] S21 includes:

[0098] S211, During the execution of the control strategy, the real-time state variables of the controlled object are continuously collected to form a state flow. Simultaneously, the actual execution results of the control strategy response are collected to form a feedback behavior sequence.

[0099] S212, state flow With feedback behavior sequence Synchronization is performed using a uniform sampling period Δt. If time offset or data loss exists, an interpolation function is used. Complete state flow With feedback behavior sequence Alignment, represented as:

[0100]

[0101] in, The aligned state-behavior joint vector. For time interpolation and alignment functions, Let be the state observation vector (e.g., temperature, current, velocity, etc.) at the i-th time point. For the corresponding time point, n represents the behavioral feedback result (such as action execution status, response amplitude, etc.), where n is the number of sampling points;

[0102]

[0103] Where, x i x i+1 Given the known observations, t i t i+1 t represents the adjacent known sampling time points, and t represents the current sampling time point to be aligned;

[0104] S213, for all aligned state flows With feedback behavior sequence Normalization is performed to unify the dimensions and scales of different indicators, ultimately forming the control behavior observation matrix M. beh , represented as:

[0105]

[0106] in, For the normalized function, μ w σ w They are respectively for The mean and standard deviation vectors calculated for each dimension The normalized observation vector, These are the times t1, t2, ..., t n The normalized control behavior observation vector constructed on top.

[0107] S23 includes:

[0108] S231, for the i-th key control variable that has detected drift, from the corresponding control intent vector q i Extract control boundary constraint parameters to form the control boundary parameters, including adjustment frequency limits. Amplitude Limitation With response latency limitations

[0109] S232, in the control strategy path Based on the control boundary parameters, the control strategy path is modified, including frequency limits (control step limits) and amplitude limits (control action limits), ultimately yielding the modified control strategy path. Represented as:

[0110]

[0111] Where clip is a constraint function, δ j Limited to the range Inside, m ′ This represents the number of control steps allowed to be executed after frequency restrictions, where m is the total number of control actions in the original policy path. The amplitude of the control action in step j after correction;

[0112] S233, the revised control strategy path As input, combined with response latency constraints Generate modified control strategy sequence Represented as:

[0113]

[0114] in, To control the sequence scheduling function and ensure that the interval between each adjustment action is not less than [amount missing]

[0115] S3 includes:

[0116] S31, modify the control strategy sequence from the previous cycle. The j-th key control variable applied to the target system is used to collect the state feedback sequence in real time during control execution. And record the set of boundary behavior events triggered during execution.

[0117] S32, construct the behavior evolution chain based on the collected state feedback sequence and the triggered boundary behavior event set. And combined with control intent vector With state feedback sequence Calculate the evolution trend vector of the control behavior chain Represented as:

[0118]

[0119] Where T is the number of sampling points within the control execution time window. For time τ k The state feedback sequence vector below, To control the desired target value of the j-th variable;

[0120] S33, based on the evolutionary trend vector Update the control intent vector based on the boundary trigger result. Forming the control intent version for the next cycle Ultimately, all updated control intentions are constructed into an evolvable chain of control intentions in chronological order. Represented as:

[0121]

[0122] Where, η j Update the step size coefficient for the intention corresponding to the j-th variable. This is the control intent vector from the previous cycle. This is the updated control intent vector.

[0123] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0124] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A smart control method based on big data, characterized in that, Includes the following steps: S1, collect the target system's operational status data and environmental context information in the early stage of control, combine the task objective and control boundary, construct a control intent vector model including the desired control direction, allowable disturbance range and control action boundary, and encapsulate the control intent vector model as a transfer data object; S2, during the control execution process, collects the state flow and feedback behavior sequence of the controlled object in real time, and measures the offset between the current behavior sequence and the control intention vector model through the behavior offset recognition mechanism. If a behavior drift event is detected, the control strategy path is dynamically corrected based on the intention boundary constraints, and the updated corrected control strategy sequence is output. S3 applies the modified control strategy sequence to the target system, collects its execution response and boundary triggering behavior, identifies the evolution trend of the control behavior chain based on the causal response inference mechanism, updates the structure of the control intent vector model, and outputs the control intent chain to support the control task of the next cycle.

2. The intelligent control method based on big data according to claim 1, characterized in that, S1 includes: S11: Collect the operational status data and environmental context information of the target system in the early stage of control. The operational status data includes key control variables, state feedback sequences and boundary response records. The environmental context information includes external disturbance sources, operational constraints and historical event labels. Through a unified time axis alignment mechanism and context event indexing mechanism, a time-series-structure composite dataset reflecting the initial behavioral background of the system is constructed. S12, based on the operating status data and environmental context information, parse the current control task objective and extract the control constraint boundary conditions. The task objective includes the expected system state and its range of achievement. The control constraint boundary conditions include the system adjustment frequency, amplitude, and time delay. Through the objective-boundary joint coding mechanism, a structured control objective expression unit is generated. S13 combines the constructed time-series-structured dataset with structured control target expression units and utilizes a control semantic alignment mechanism to generate a control intent vector model that includes the desired control direction, permissible disturbance range, and control action boundary.

3. The intelligent control method based on big data according to claim 2, characterized in that, S11 includes: S111, collect multi-source data assets in the early stage of control and divide them into operational status data and environmental context information, among which; The operational status data includes key control variables. State feedback sequence Boundary response record ε bnd ; The environmental context information includes external disturbance sources. Operational constraints and historical event tags S112 employs a centralized sliding window mechanism to construct a unified time axis τ, and in each time slice t i The convergence of normalized data representations forms a time synchronization vector. Generate a synchronization time series matrix M sync ; S113, Construct an event-driven structured index graph, and use an event-state mapping function to associate time synchronization data with event triggering logic to establish a context event index set G. evt and compare it with the time series matrix M sync Combined, a structural map of time-event coupling is formed.

4. The intelligent control method based on big data according to claim 3, characterized in that, S12 includes: S121, based on operational status data and environmental context data, uses a status index mapping function to identify the control objective within the current cycle. S122, parse the execution boundary constraints related to the control task, extract the maximum adjustment frequency, adjustment magnitude, and response delay under the current context, and construct the control boundary set. S123, control target With control boundary set Joint encoding is performed to generate structured control target representation units g. k .

5. The intelligent control method based on big data according to claim 4, characterized in that, S13 includes: S131, the time series matrix M sync With structured control target expression unit g k Perform pairings to construct a variable-state association graph. S132, based on the variable-state association graph Generate control semantic labels, including desired control direction. Permissible Disturbance Range And control boundary parameters; S133, merge control semantic labels into a control intent vector q i Ultimately, a control intent vector model Q is generated. intent .

6. The intelligent control method based on big data according to claim 5, characterized in that, S2 includes: S21, During the execution of the control strategy, the continuous state flow of the controlled object is collected. With feedback behavior sequence The data is then time-aligned and normalized according to a uniform sampling period to form the control behavior observation matrix M. beh ; S22, based on the control intent vector model Q intent The behavioral offset metric function is used to calculate the behavioral offset between the current state and the intended expectation. k ,like Then it is determined that the k-th variable has experienced a drift event, where, The threshold for determining behavioral deviation; S23: For the key control variables that cause drift, read the control boundary parameters in their control intention vector, dynamically adjust the control frequency and adjustment range of the control strategy path, generate the corrected control strategy path, and finally output the updated corrected control strategy sequence.

7. The intelligent control method based on big data according to claim 6, characterized in that, S21 includes: S211, During the execution of the control strategy, the real-time state variables of the controlled object are continuously collected to form a state flow. Simultaneously, the actual execution results of the control strategy response are collected to form a feedback behavior sequence. S212, state flow With feedback behavior sequence Synchronization is performed using a uniform sampling period Δt. If time offset or data loss exists, an interpolation function is used. Complete state flow With feedback behavior sequence Alignment; S213, for all aligned state flows With feedback behavior sequence After normalization, the final control behavior observation matrix M is formed. beh .

8. The intelligent control method based on big data according to claim 7, characterized in that, S23 includes: S231, for the i-th key control variable that has detected drift, from the corresponding control intent vector q i Extract control boundary constraint parameters to form the control boundary parameters, including adjustment frequency limits. Amplitude Limitation With response latency limitations S232, in the control strategy path Based on the control boundary parameters, including frequency and amplitude limits, the corrected control strategy path is obtained. S233, the revised control strategy path As input, combined with response latency constraints Generate modified control strategy sequence 9. The intelligent control method based on big data according to claim 8, characterized in that, S3 includes: S31, modify the control strategy sequence from the previous cycle. The j-th key control variable applied to the target system is used to collect the state feedback sequence in real time during control execution. And record the set of boundary behavior events triggered during execution. S32, construct the behavior evolution chain based on the collected state feedback sequence and the triggered boundary behavior event set. And combined with control intent vector With state feedback sequence Calculate the evolution trend vector of the control behavior chain S33, based on the evolutionary trend vector Update the control intent vector based on the boundary trigger result. Forming the control intent version for the next cycle Ultimately, all updated control intentions are constructed into a control intention chain in chronological order.

Images