A method and system for circuit breaking and restarting data governance intelligent agents to address cognitive degradation.

Abstract

This application provides a method and system for circuit breaking and restarting a data governance agent's cognitive degradation. The method includes: acquiring the execution result of the data governance agent; generating multidimensional statistical data based on the execution result; comparing the multidimensional statistical data with a dynamic threshold set to obtain a comparison result; when the comparison result meets the abnormal conditions of the execution result, determining one or more backtracking paths corresponding to the execution result, where each backtracking path includes sequentially executed sub-agents; determining the sub-agent to be evaluated based on the comparison result and the preset weights of each sub-agent in the backtracking path; calculating the problem evaluation value of the sub-agent to be evaluated; when the problem evaluation value is greater than a preset threshold, determining the target circuit breaking parameter combination corresponding to the problem evaluation value; and performing circuit breaking and restart operations on the sub-agent to be evaluated based on the target circuit breaking parameter combination. This solution can achieve reasonable circuit breaking and restarting of data governance agents, improving data governance efficiency.

Description

Technical Field

[0001] This application relates to the field of electronic digital data processing, and in particular to a method and system for circuit breaking and restarting a data governance intelligent agent's cognitive degradation. Background Technology

[0002] Data governance agents are a type of generative artificial intelligence technology based on large models. They possess autonomy, adaptability, and interactivity, aiming to optimize the data governance process and improve data quality and value release through automation and intelligent means. They can perceive environmental changes, make judgments and decisions based on learned knowledge and algorithms, and thus execute data governance tasks, including the management of data models, data standards, metadata, data quality, and data security. However, when handling data governance tasks, data governance agents may encounter problems in their decision-making capabilities due to issues with the data itself or the model.

[0003] In related technologies, when using data governance agents for data processing, the common practice is to directly shut down and restart the data governance agent when a large number of consecutive erroneous results are detected. This approach lacks rationality and leads to a decrease in the overall data processing efficiency of the data governance agent. Summary of the Invention

[0004] This application provides a method and system for circuit breaking and restarting data governance agents to address the problem of the lack of reasonable circuit breaking and restarting mechanisms in existing technologies. This method enables reasonable circuit breaking and restarting of data governance agents, thereby improving data governance efficiency.

[0005] In a first aspect, embodiments of this application provide a method for circuit breaking and restarting a data governance agent's cognitive degradation, including:

[0006] Obtain the execution results of the data governance intelligent agent, and generate multidimensional statistical data based on the execution results;

[0007] The multidimensional statistical data is compared with the set dynamic thresholds to obtain the comparison result. If the comparison result meets the abnormal conditions of the execution result, one or more backtracking paths corresponding to the execution result are determined. The backtracking path includes each sequentially executed sub-agent.

[0008] The sub-agent to be evaluated is determined based on the comparison processing results and the preset weights of each sub-agent in the backtracking path, and the problem evaluation value of the sub-agent to be evaluated is calculated.

[0009] If the problem evaluation value is greater than a preset threshold, a target circuit breaker parameter combination corresponding to the problem evaluation value is determined, and the circuit breaker and restart operations are performed on the sub-agent to be evaluated based on the target circuit breaker parameter combination.

[0010] Optionally, determining the sub-agent to be evaluated based on the comparison processing result and the preset weights of each sub-agent in the backtracking path includes:

[0011] In the presence of a backtracking path, the sub-agents in the backtracking path are arranged from high to low according to their corresponding preset weights to obtain the candidate sub-agent permutation combination;

[0012] In the case of multiple backtracking paths, the sub-agents are arranged according to the length of each backtracking path and the preset weight of each sub-agent to obtain the candidate sub-agent arrangement combination.

[0013] Based on the comparison processing results, the agent selection ratio is determined, and each of the agent selected at the highest proportion in the candidate agent permutation is identified as the agent to be evaluated.

[0014] Optionally, the step of arranging sub-agents according to the length of each backtracking path and the preset weight of each sub-agent to obtain the candidate sub-agent arrangement combination includes:

[0015] The path weight is determined according to the length of each backtracking path, and the path weight is multiplied by the preset weight of each corresponding sub-agent to obtain the adjustment weight of each sub-agent.

[0016] The adjustment weights of the same sub-agents in each backtracking path are superimposed to obtain the corresponding final weights. The adjustment weights of the remaining sub-agents in each backtracking path are determined as the corresponding final weights. The sub-agents in each backtracking path are arranged from high to low according to their corresponding final weights to obtain the candidate sub-agent permutation combination.

[0017] Optionally, the multidimensional statistical data includes error output rate, frequency of abnormal decisions, and data non-correlation degree. The step of comparing the multidimensional statistical data with a set of dynamic thresholds to obtain the comparison result includes:

[0018] The error output rate, the frequency of abnormal decisions, and the data non-association degree are compared with the corresponding dynamic thresholds in the dynamic threshold set to obtain the comparison results.

[0019] Correspondingly, the abnormal conditions for the execution result include:

[0020] The comparison results contain statistical data that are greater than the corresponding dynamic threshold.

[0021] Optionally, determining the target circuit breaker parameter combination corresponding to the problem evaluation value includes:

[0022] The problem assessment value is compared with each preset value range, and the circuit breaker parameter combination associated with the preset value range into which the problem assessment value falls is determined as the target circuit breaker parameter combination.

[0023] Optionally, the multidimensional statistical data includes error output rate, frequency of anomalous decisions, and data non-association degree; calculating the problem evaluation value of the sub-agent to be evaluated includes:

[0024] The error output rate, the frequency of abnormal decisions, and the degree of data non-association are weighted according to historical failed task information.

[0025] The problem evaluation value of the agent to be evaluated is obtained by multiplying the error output rate, the frequency of abnormal decisions, and the degree of data non-association by their respective assigned weights and then summing them.

[0026] Optionally, before calculating the problem evaluation value of the sub-agent to be evaluated, the method further includes:

[0027] The original input data associated with the sub-agent to be evaluated is obtained, the integrity of the original input data is checked to obtain the data missing rate, and the type of the original input data is checked to see if it is consistent with the predefined type. Based on the data missing rate and the detection result, it is determined whether the original input data meets the preset data processing requirements.

[0028] Accordingly, calculating the problem evaluation value of the sub-agent to be evaluated includes:

[0029] If the original input data meets the preset data processing requirements, calculate the problem evaluation value of the sub-agent to be evaluated.

[0030] Secondly, embodiments of this application also provide a circuit breaker and restart system for cognitive degradation of data governance intelligent agents, including:

[0031] The data generation module is used to obtain the execution results of the data governance intelligent agent and generate multidimensional statistical data based on the execution results;

[0032] The comparison processing module is used to compare the multidimensional statistical data with a set dynamic threshold to obtain the comparison processing result.

[0033] The path determination module is used to determine one or more backtracking paths corresponding to the execution result when the comparison processing result meets the abnormal conditions of the execution result. The backtracking path includes each sequentially executed sub-agent.

[0034] The evaluation agent determination module is used to determine the sub-agent to be evaluated based on the comparison processing result and the preset weights of each sub-agent in the backtracking path.

[0035] The evaluation value calculation module is used to calculate the problem evaluation value of the sub-agent to be evaluated;

[0036] The circuit breaker parameter determination module is used to determine the target circuit breaker parameter combination corresponding to the problem evaluation value when the problem evaluation value is greater than a preset threshold.

[0037] The circuit breaker and restart module is used to perform circuit breaker and restart operations on the sub-agent to be evaluated based on the target circuit breaker parameter combination.

[0038] Thirdly, embodiments of this application also provide a circuit breaker and restart device for cognitive degradation of data governance intelligent agents, the device comprising:

[0039] One or more processors;

[0040] Storage device for storing one or more programs.

[0041] When the one or more programs are executed by the one or more processors, the one or more processors implement the circuit breaker and restart method for cognitive degradation of data governance agents as described in the embodiments of this application.

[0042] Fourthly, embodiments of this application also provide a storage medium for storing computer-executable instructions, which, when executed by a computer processor, are used to perform the circuit breaker and restart method for cognitive degradation of data governance intelligent agents described in embodiments of this application.

[0043] In this embodiment, the execution results of the data governance agent are obtained, multi-dimensional statistical data are generated based on the execution results, and the multi-dimensional statistical data are compared with a set of dynamic thresholds to obtain comparison results. If the comparison results meet the abnormal conditions of the execution results, one or more backtracking paths corresponding to the execution results are determined. The backtracking paths include each sequentially executed sub-agent. Based on the comparison results and the preset weights of each sub-agent in the backtracking paths, the sub-agent to be evaluated is determined, and the problem evaluation value of the sub-agent to be evaluated is calculated. If the problem evaluation value is greater than a preset threshold, the target circuit breaker parameter combination corresponding to the problem evaluation value is determined, and the circuit breaker and restart operations are performed on the sub-agent to be evaluated based on the target circuit breaker parameter combination. This solution determines the sub-agent to be evaluated through multi-dimensional statistical data generated from the execution results of the data governance agent, and reasonably controls its circuit breaker and restart based on the calculated problem evaluation value of the sub-agent to be evaluated. This solves the problem of the lack of reasonable circuit breaker and restart mechanisms in the prior art for data governance agents, and can realize reasonable circuit breaker and restart of data governance agents, thereby improving data governance efficiency. Attached Figure Description

[0044] Figure 1 A flowchart of a circuit breaker and restart method for cognitive degradation of a data governance agent provided in this application embodiment;

[0045] Figure 2 A schematic diagram of the complete dependency graph of a data governance intelligent agent provided in an embodiment of this application;

[0046] Figure 3 A flowchart of a circuit breaker and restart method for determining the cognitive degradation of a data governance agent to be evaluated is provided in an embodiment of this application;

[0047] Figure 4 A flowchart of a circuit breaker and restart method for a data governance agent that determines the cognitive degradation of a comparison processing result, provided in an embodiment of this application;

[0048] Figure 5 A flowchart of a circuit breaker and restart method for a data governance agent that determines a target combination of circuit breaker parameters and is used as an embodiment of this application is provided.

[0049] Figure 6 A flowchart of a circuit breaker and restart method for cognitive degradation of a data governance agent that includes a computational problem evaluation value, provided in an embodiment of this application;

[0050] Figure 7 A flowchart of a circuit breaker and restart method for a data governance agent that detects cognitive degradation of raw input data, provided in an embodiment of this application;

[0051] Figure 8 A block diagram of a circuit breaker and restart system for cognitive degradation of a data governance intelligent agent provided in an embodiment of this application;

[0052] Figure 9 This is a schematic diagram of the structure of a circuit breaker and restart device for cognitive degradation of a data governance intelligent agent, provided in an embodiment of this application. Detailed Implementation

[0053] The embodiments of this application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the embodiments of this application and are not intended to limit the scope of the embodiments. Furthermore, it should be noted that, for ease of description, only the parts relevant to the embodiments of this application are shown in the accompanying drawings, not the entire structure.

[0054] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0055] This application provides a method for circuit breaking and restarting a data governance agent's cognitive degradation, applicable to data governance agent scenarios such as data cleaning, data quality monitoring, and data security. In this method, the server is the executing entity for each step.

[0056] Figure 1 A flowchart of a circuit breaker and restart method for cognitive degradation of a data governance agent provided in this application embodiment is shown below. Figure 1 As shown, it specifically includes:

[0057] Step S101: Obtain the execution result of the data governance intelligent agent, generate multidimensional statistical data based on the execution result, and compare the multidimensional statistical data with the set dynamic threshold set to obtain the comparison result.

[0058] In this context, a data governance agent refers to an automated system that combines artificial intelligence technology and data governance principles. It can sense changes in data within the environment and automatically execute data governance tasks through learning algorithms and data analysis. This data governance agent can contain multiple independently executing sub-agents. The execution result is the output of the data governance agent after completing the data governance task. This execution result can be used to generate multidimensional statistical data. Multidimensional statistical data can be attribute data from multiple dimensions of the statistical results, such as error output rate, frequency of abnormal decisions, etc. After the multidimensional statistical data is generated, it is compared with a set of dynamic thresholds. The dynamic threshold set represents the set of dynamic thresholds used for comparison with statistical data from various dimensions. This dynamic threshold set can be automatically updated after each comparison. For example, after the comparison is completed, the current multidimensional statistical data is superimposed with historical multidimensional statistical data, and the corresponding average statistical data set is calculated. Based on this average statistical data set, the corresponding updated dynamic threshold set is determined. In one embodiment, multidimensional statistical data includes statistical data in multiple dimensions. One comparison processing method is to compare the statistical data in multiple dimensions with the corresponding dynamic thresholds in the dynamic threshold set to obtain the proportion of statistical data exceeding the corresponding dynamic thresholds, and determine the proportion as the comparison result. Accordingly, the abnormal condition for the execution result is that the proportion is greater than a preset proportion.

[0059] Step S102: If the comparison processing result meets the abnormal conditions of the execution result, determine one or more backtracking paths corresponding to the execution result, wherein the backtracking path includes each sequentially executed sub-agent.

[0060] The execution result anomaly condition is used to characterize the conditions for determining whether the execution result is abnormal by comparing the processing results. If the comparison processing result meets the execution result anomaly condition, one or more backtracking paths corresponding to the execution result of the data governance agent are determined. The backtracking path characterizes the execution path of each sub-agent involved in the process of the data governance agent determining the execution result. This backtracking path includes each sequentially executed sub-agent. In one embodiment, one way to determine the backtracking path is to obtain the complete dependency graph of the data governance agent, determine the terminal sub-agent corresponding to the execution result, find the terminal sub-agent in the complete dependency graph, and traverse the upstream nodes of the terminal sub-agent in reverse order according to the dependency relationship to obtain one or more backtracking paths corresponding to the execution result. The complete dependency graph includes the execution links of each data governance task. For example, as shown... Figure 2 As shown, Figure 2This is a schematic diagram of the complete dependency graph of a data governance agent provided in an embodiment of this application. AL represents each sub-agent in the data governance agent. If the terminal sub-agent corresponding to the execution result is determined to be K, then the backtracking path corresponding to the execution result obtained by traversing backward from the sub-agent K according to the dependency relationship is A→C→G→K and B→D→G→K.

[0061] Step S103: Determine the sub-agent to be evaluated based on the comparison processing results and the preset weights of each sub-agent in the backtracking path, and calculate the problem evaluation value of the sub-agent to be evaluated.

[0062] The preset weights can be pre-defined as the importance of each sub-agent in its corresponding backtracking path. Using the comparison results and these preset weights, the sub-agents to be evaluated can be determined. The sub-agent to be evaluated represents the sub-agent awaiting evaluation to determine if it has a problem. After determining the sub-agent to be evaluated, a problem evaluation value is calculated for it. The problem evaluation value represents the severity of the problem encountered by the sub-agent to be evaluated. In one embodiment, one way to determine the sub-agent to be evaluated is to query the preset correlation between each sub-agent in the backtracking path and the comparison results, calculate the corresponding final weight based on the preset weights of each sub-agent in the backtracking path and the preset correlation with the comparison results, and determine the sub-agent with the final weight greater than the preset weight as the sub-agent to be evaluated.

[0063] In one embodiment, the multidimensional statistical data includes error output rate, frequency of abnormal decisions, and data non-association degree. One way to calculate the problem evaluation value is to query the corresponding preset problem value mapping table according to the error output rate, frequency of abnormal decisions, and data non-association degree to obtain the corresponding problem value to be statistically analyzed, and calculate the average value of the problem value to be statistically analyzed corresponding to the error output rate, frequency of abnormal decisions, and data non-association degree to obtain the problem evaluation value of the agent to be evaluated.

[0064] Step S104: If the problem evaluation value is greater than the preset threshold, determine the target circuit breaker parameter combination corresponding to the problem evaluation value, and perform circuit breaker and restart operations on the sub-agent to be evaluated based on the target circuit breaker parameter combination.

[0065] The preset threshold is a pre-set standard problem evaluation value, which can be set according to actual conditions and is not a fixed value. When the problem evaluation value exceeds the preset threshold, a target circuit breaker parameter combination corresponding to that problem evaluation value is determined. The target circuit breaker parameter combination can be a combination of parameter values ​​for various operations during the circuit breaker process, such as a combination of memory release ratio, number of interface disconnections, and process sleep duration. Using this target circuit breaker parameter combination, the circuit breaker of the evaluated sub-agent can be completed, and a restart operation can be performed. One way to determine the target circuit breaker parameter combination is to compare the problem evaluation value with each preset value range corresponding to each circuit breaker parameter, and combine the circuit breaker parameter values ​​associated with the preset value ranges of each circuit breaker parameter into which the problem evaluation value falls to obtain the target circuit breaker parameter combination. By determining the operation parameters during the circuit breaker process using the problem evaluation value of the sub-agent, the overreaction of a full circuit breaker can be avoided, achieving reasonable circuit breaker and restart, and improving data governance efficiency.

[0066] As described above, by obtaining the execution results of the data governance agent, generating multidimensional statistical data based on the execution results, and comparing the multidimensional statistical data with a set of dynamic thresholds to obtain comparison results, if the comparison results meet the abnormal conditions of the execution results, one or more backtracking paths corresponding to the execution results are determined. The backtracking paths include each sequentially executed sub-agent. Based on the comparison results and the preset weights of each sub-agent in the backtracking paths, the sub-agent to be evaluated is determined, and the problem evaluation value of the sub-agent to be evaluated is calculated. If the problem evaluation value is greater than a preset threshold, the target circuit breaker parameter combination corresponding to the problem evaluation value is determined. Based on the target circuit breaker parameter combination, the sub-agent to be evaluated is subjected to circuit breaker and restart operations. This solution determines the sub-agent to be evaluated through multidimensional statistical data generated from the execution results of the data governance agent, and reasonably controls its circuit breaker and restart based on the calculated problem evaluation value of the sub-agent to be evaluated. This solves the problem of the lack of a reasonable circuit breaker and restart mechanism for data governance agents in the prior art, enabling reasonable circuit breaker and restart of data governance agents and improving data governance efficiency.

[0067] Figure 3 A flowchart of a circuit breaker and restart method for determining the cognitive degradation of a data governance agent to be evaluated, provided in an embodiment of this application, is shown below. Figure 3 As shown, it specifically includes:

[0068] Step S201: Obtain the execution result of the data governance intelligent agent, generate multidimensional statistical data based on the execution result, and compare the multidimensional statistical data with the set dynamic threshold set to obtain the comparison result.

[0069] Step S202: If the comparison processing result meets the abnormal conditions of the execution result, determine one or more backtracking paths corresponding to the execution result, wherein the backtracking path includes each sequentially executed sub-agent.

[0070] Step S203: If there is a backtracking path, arrange each sub-agent in the backtracking path from high to low according to the corresponding preset weight to obtain the candidate sub-agent arrangement combination. If there are multiple backtracking paths, arrange the sub-agents according to the length of each backtracking path and the preset weight of each sub-agent to obtain the candidate sub-agent arrangement combination.

[0071] The candidate sub-agent permutation can be a combination of sub-agents arranged in descending order of their own weights along the backtracking path. Given a backtracking path, arranging the sub-agents in the backtracking path according to their corresponding preset weights in descending order yields the candidate sub-agent permutation. For example, if the backtracking path is A→C→G→K, and the preset weights of sub-agent A, C, G, and K are 0.3, 0.2, 0.4, and 0.1 respectively, then the candidate sub-agent permutation is: G-A-C-K.

[0072] When multiple backtracking paths exist, the candidate sub-agent combinations can be obtained by arranging the sub-agents according to the length of each backtracking path and the preset weights of each sub-agent. Optionally, when multiple backtracking paths exist, one method for arranging the sub-agents is to determine the corresponding path weights based on the length of each backtracking path, multiply each path weight by the preset weight of each corresponding sub-agent to obtain the adjusted weight of each sub-agent, sum the adjusted weights of the same sub-agents in each backtracking path to obtain the corresponding final weight, determine the adjusted weights of the remaining sub-agents in each backtracking path as the corresponding final weights, and arrange the sub-agents in each backtracking path from high to low according to their corresponding final weights to obtain the candidate sub-agent combinations. For example, backtracking path 1 is A→B→C→D, with a length of 4 and a corresponding path weight of 0.6. In backtracking path 1, the preset weights of sub-agents A, B, C, and D are 0.1, 0.4, 0.3, and 0.2, respectively. Therefore, the adjusted weights of sub-agents A, B, C, and D in backtracking path 1 are 0.06, 0.24, 0.18, and 0.12, respectively. Backtracking path 2 is A→C→D, with a length of 3 and a corresponding path weight of 0.4. In backtracking path 2, the preset weights of sub-agents A, C, and D are 0.3, respectively. If the adjusted weights of sub-agents A, C, and D in backtracking path 1 are 0.12, 0.2, and 0.08 respectively, then the adjusted weights of sub-agents A, C, and D in backtracking path 1 and backtracking path 2 are superimposed. The final weight of sub-agent A is 0.18, the final weight of sub-agent B is 0.24, the final weight of sub-agent C is 0.38, and the final weight of sub-agent D is 0.2. The sub-agents are sorted from largest to smallest according to their final weights to obtain the candidate sub-agent permutation combination as sub-agent C-sub-agent B-sub-agent D-sub-agent A.

[0073] In another embodiment, when there are multiple backtracking paths, a method for arranging sub-agents can be to superimpose the preset weights of the same sub-agents in each backtracking path to obtain the corresponding final weights, determine the preset weights of the remaining sub-agents in each backtracking path as the corresponding final weights, and arrange the sub-agents in each backtracking path from high to low according to the corresponding final weights to obtain the candidate sub-agent arrangement combination.

[0074] Step S204: Determine the agent selection ratio based on the comparison processing results, identify each agent with the highest agent selection ratio in the candidate agent permutation and combination as the agent to be evaluated, and calculate the problem evaluation value of the agent to be evaluated.

[0075] The agent selection ratio is used to characterize the proportion of sub-agents selected from the possible combinations of sub-agents for evaluation. For example, if the agent selection ratio is determined to be 50% based on the comparison processing results, the top 50% of the possible combinations of sub-agents are selected as the sub-agents to be evaluated. If the possible combinations of sub-agents are sub-agent C-sub-agent B-sub-agent D-sub-agent A, then the sub-agents to be evaluated are sub-agent C and sub-agent B. In one embodiment, an abnormal condition in the execution result is the presence of statistical data in the comparison processing results that exceeds the corresponding dynamic threshold. One agent selection ratio can be calculated by multiplying the difference between the statistical data exceeding the corresponding dynamic threshold and the corresponding dynamic threshold, and then multiplying this difference by the proportional coefficient associated with the statistical data to obtain the agent selection ratio.

[0076] Step S205: If the problem evaluation value is greater than the preset threshold, determine the target circuit breaker parameter combination corresponding to the problem evaluation value, and perform circuit breaker and restart operations on the sub-agent to be evaluated based on the target circuit breaker parameter combination.

[0077] As described above, after determining one or more backtracking paths corresponding to the execution result, if there is only one backtracking path, the sub-agents in the backtracking path are arranged from high to low according to their corresponding preset weights to obtain a candidate sub-agent arrangement. If there are multiple backtracking paths, the sub-agents are arranged according to the length of each backtracking path and the preset weights of each sub-agent to obtain a candidate sub-agent arrangement. The selection ratio of agents is determined based on the comparison processing results. The sub-agents with the highest selection ratio in the candidate sub-agent arrangements are identified as the sub-agents to be evaluated, and the problem evaluation value of the sub-agents to be evaluated is calculated. This scheme determines the sub-agents to be evaluated by comparing the processing results and the weights of the sub-agents in the backtracking paths, which can ensure the rationality of the sub-agents to be evaluated and improve the efficiency of sub-agent evaluation.

[0078] Figure 4 A flowchart of a circuit breaker and restart method for a data governance agent's cognitive degradation, including determining the comparison processing result, is provided as an embodiment of this application. Figure 4 As shown, it specifically includes:

[0079] Step S301: Obtain the execution results of the data governance intelligent agent, generate multidimensional statistical data based on the execution results, and compare the error output rate, abnormal decision frequency, and data non-association degree with the corresponding dynamic thresholds in the dynamic threshold set to obtain the comparison processing results.

[0080] The multidimensional statistical data includes error output rate, frequency of abnormal decisions, and data non-association. Error output rate refers to the ratio of erroneous execution results output by the data governance agent. Frequency of abnormal decisions can be the number of times the data governance agent triggers unexpected decisions when performing data governance tasks per unit time. Data non-association can be the difference between the maximum logical association strength and the logical association strength of multidimensional data during the execution of data governance tasks by the data governance agent. For example, in the dynamic threshold set, the first dynamic threshold corresponding to error output rate is 5%, the second dynamic threshold corresponding to abnormal decision frequency is 3 times / h, the third dynamic threshold corresponding to data non-association is 5%, the error output rate corresponding to the execution result of the data governance agent is 8%, the abnormal decision frequency is 1 time / h, and the data non-association is 2%. Therefore, the comparison result is that the error output rate corresponding to the execution result of the data governance agent is greater than the corresponding dynamic threshold, while the abnormal decision frequency and data non-association are both less than the corresponding dynamic thresholds.

[0081] Step S302: If there are statistical data in the comparison processing results that are greater than the corresponding dynamic threshold, determine one or more backtracking paths corresponding to the execution result, wherein the backtracking path includes each sequentially executed sub-agent.

[0082] Step S303: Determine the sub-agent to be evaluated based on the comparison processing results and the preset weights of each sub-agent in the backtracking path, and calculate the problem evaluation value of the sub-agent to be evaluated.

[0083] Step S304: If the problem evaluation value is greater than the preset threshold, determine the target circuit breaker parameter combination corresponding to the problem evaluation value, and perform circuit breaker and restart operations on the sub-agent to be evaluated based on the target circuit breaker parameter combination.

[0084] As described above, by comparing the error output rate, abnormal decision frequency, and data non-association degree in the multidimensional statistical data with the corresponding dynamic thresholds in the dynamic threshold set, comparison results are obtained. If any statistical data in the comparison results exceeds the corresponding dynamic threshold, one or more backtracking paths corresponding to the execution result are determined. This scheme can reasonably determine whether anomalies occur when the data governance agent outputs execution results.

[0085] Figure 5 A flowchart of a circuit breaker and restart method for a data governance agent's cognitive degradation, including determining a target combination of circuit breaker parameters, is provided as an embodiment of this application. Figure 5 As shown, it specifically includes:

[0086] Step S401: Obtain the execution result of the data governance intelligent agent, generate multidimensional statistical data based on the execution result, and compare the multidimensional statistical data with the set dynamic threshold set to obtain the comparison result.

[0087] Step S402: If the comparison processing result meets the abnormal conditions of the execution result, determine one or more backtracking paths corresponding to the execution result, wherein the backtracking path includes each sequentially executed sub-agent.

[0088] Step S403: Determine the sub-agent to be evaluated based on the comparison processing results and the preset weights of each sub-agent in the backtracking path, and calculate the problem evaluation value of the sub-agent to be evaluated.

[0089] Step S404: If the problem evaluation value is greater than the preset threshold, compare the problem evaluation value with each preset value range, determine the circuit breaker parameter combination associated with the preset value range into which the problem evaluation value falls as the target circuit breaker parameter combination, and perform circuit breaker and restart operations on the sub-agent to be evaluated based on the target circuit breaker parameter combination.

[0090] The preset numerical ranges are pre-defined ranges of problem evaluation values ​​at different levels. Each preset numerical range corresponds to a circuit breaker parameter combination. The circuit breaker parameter combination can be a combination of various types of operational parameter values ​​that trigger the agent's circuit breaker; for example, a circuit breaker parameter combination could be (30% memory release ratio, 10% interface disconnection count, and 5-minute process sleep duration). The circuit breaker parameter combination associated with the preset numerical range into which the problem evaluation value falls is the target circuit breaker parameter combination. For example, the problem evaluation value of the sub-agent A to be evaluated is 75 points. The preset value range 1 is [60, 70], and the corresponding circuit breaker parameter combination is (memory release ratio 30%, interface disconnection count 10%, process sleep duration 5 minutes). The preset value range 2 is (70, 85], and the corresponding circuit breaker parameter combination is (memory release ratio 60%, interface disconnection count 50%, process sleep duration 15 minutes). The preset value range 3 is (85, 100], and the corresponding circuit breaker parameter combination is (memory release ratio 100%, interface disconnection count 100%, process sleep duration 30 minutes). The problem evaluation value of the sub-agent A to be evaluated is in the preset value range 2. Then the target circuit breaker parameter combination for the sub-agent A to be evaluated is (memory release ratio 60%, interface disconnection count 50%, process sleep duration 15 minutes).

[0091] As described above, when the problem assessment value exceeds a preset threshold, the circuit breaker parameter combination associated with the preset numerical ranges to which the problem assessment value falls is determined as the target circuit breaker parameter combination. Based on this target circuit breaker parameter combination, the sub-agent to be evaluated is subjected to circuit breaker and restart operations. This solution determines the operational parameters during the circuit breaker process by using the sub-agent's problem assessment value, thus avoiding overreactions from full circuit breaker actions, achieving reasonable circuit breaker and restart, and improving data governance efficiency.

[0092] Figure 6 A flowchart of a circuit breaker and restart method for cognitive degradation of a data governance agent, including the calculation of problem evaluation values, is provided for embodiments of this application. Figure 6 As shown, it specifically includes:

[0093] Step S501: Obtain the execution result of the data governance intelligent agent, generate multidimensional statistical data based on the execution result, and compare the multidimensional statistical data with the set dynamic threshold set to obtain the comparison processing result.

[0094] Step S502: If the comparison processing result meets the abnormal conditions of the execution result, determine one or more backtracking paths corresponding to the execution result, wherein the backtracking path includes each sequentially executed sub-agent.

[0095] Step S503: Determine the sub-agent to be evaluated based on the comparison processing results and the preset weights of each sub-agent in the backtracking path. Assign weights to the error output rate, abnormal decision frequency, and data non-association degree based on historical failed task information. Multiply the error output rate, abnormal decision frequency, and data non-association degree by their respective assigned weights and sum them up to obtain the problem evaluation value of the sub-agent to be evaluated.

[0096] The multidimensional statistical data includes statistics on error output rate, frequency of abnormal decisions, and data non-association. Historical failed task information can be relevant records of historical failed tasks. This historical failed task information can be used to assign weights to the error output rate, frequency of abnormal decisions, and data non-association. Multiplying the error output rate, frequency of abnormal decisions, and data non-association by their respective assigned weights and summing them yields the problem evaluation value of the agent to be evaluated. In one embodiment, the historical failure information includes the percentage of historical failed tasks within a preset time period caused by the error output rate, frequency of abnormal decisions, and data non-association, respectively. One weighting method is to determine the percentage of historical failed tasks within a preset time period caused by the error output rate, frequency of abnormal decisions, and data non-association as the corresponding assigned weights. For example, if the error output rate causes 20% of historical failed tasks to fail within a preset time period, the frequency of abnormal decisions causes 50% of historical failed tasks to fail within a preset time period, and the data lack of correlation causes 30% of historical failed tasks to fail within a preset time period, then the weights assigned to the error output rate, the frequency of abnormal decisions, and the data lack of correlation are 0.2, 0.5, and 0.3, respectively. The current error output rate is 10%, the frequency of abnormal decisions is 2 times / h, and the data lack of correlation is 6%. Multiplying the error output rate, the frequency of abnormal decisions, and the data lack of correlation by their respective weights and summing them up, we obtain the problem evaluation value of the agent to be evaluated as 1.038.

[0097] Step S504: If the problem evaluation value is greater than the preset threshold, determine the target circuit breaker parameter combination corresponding to the problem evaluation value, and perform circuit breaker and restart operations on the sub-agent to be evaluated based on the target circuit breaker parameter combination.

[0098] As described above, after identifying the sub-agent to be evaluated, weights are assigned to the error output rate, frequency of abnormal decisions, and data non-association based on historical failed task information. The error output rate, frequency of abnormal decisions, and data non-association are then multiplied by their respective weights and summed to obtain the problem evaluation value for the sub-agent. This scheme improves the rationality and accuracy of the problem evaluation value by assigning weights to various statistical data based on historical failed task information and calculating the corresponding problem evaluation value for the sub-agent.

[0099] Figure 7 A flowchart of a circuit breaker and restart method for a data governance agent that detects cognitive degradation of raw input data, provided in an embodiment of this application, is shown below. Figure 7 As shown, it specifically includes:

[0100] Step S601: Obtain the execution result of the data governance intelligent agent, generate multidimensional statistical data based on the execution result, and compare the multidimensional statistical data with the set dynamic threshold set to obtain the comparison processing result.

[0101] Step S602: If the comparison processing result meets the abnormal conditions of the execution result, determine one or more backtracking paths corresponding to the execution result, wherein the backtracking path includes each sequentially executed sub-agent.

[0102] Step S603: Determine the sub-agent to be evaluated based on the comparison processing results and the preset weights of each sub-agent in the backtracking path, obtain the original input data associated with the sub-agent to be evaluated, perform integrity verification on the original input data to obtain the data missing rate, and detect whether the type of the original input data is consistent with the predefined type.

[0103] The original input data refers to the data entered by the sub-agent to be evaluated before performing the data governance task. Performing an integrity check on this original input data yields the data missing rate. The data missing rate characterizes the proportion of missing data in the original input data. The predefined type is a predefined type of input data for the sub-agent to be evaluated. For example, if the original input data associated with the sub-agent to be evaluated is a user table that should contain 100 user IDs, but 10 tables are empty, the data missing rate is 10%. If the predefined type is descriptive information, and the user table is a table type, then the type of the original input data is inconsistent with the predefined type.

[0104] Step S604: Determine whether the original input data meets the preset data processing requirements based on the data missing rate and the detection results. If the original input data meets the preset data processing requirements, calculate the problem evaluation value of the sub-agent to be evaluated.

[0105] The preset data processing requirements are the pre-defined data processing requirements that the original input data must meet for the sub-agent to be evaluated. If the original input data meets these preset data processing requirements, the problem evaluation value of the sub-agent to be evaluated is calculated. In one embodiment, one method of judgment is to compare the data missing rate with a preset missing rate threshold. If the data missing rate is less than the preset missing rate and the detection result is consistent, it is determined that the original input data meets the preset data processing requirements; otherwise, it is determined that the original input data does not meet the preset data processing requirements. For example, the original input data of the sub-agent A to be evaluated has a data missing rate of 3%, the detection result is that the data type is consistent, and the preset missing rate threshold is 5%. Since the data missing rate is less than the preset missing rate and the detection result is consistent, the original input data of the sub-agent A to be evaluated meets the preset data processing requirements.

[0106] Step S605: If the problem evaluation value is greater than the preset threshold, determine the target circuit breaker parameter combination corresponding to the problem evaluation value, and perform circuit breaker and restart operations on the sub-agent to be evaluated based on the target circuit breaker parameter combination.

[0107] As described above, after identifying the sub-agent to be evaluated, the original input data associated with the sub-agent is obtained. The integrity of the original input data is verified to obtain the missing data rate, and the type of the original input data is checked to see if it matches a predefined type. Based on the missing data rate and the detection results, it is determined whether the original input data meets the preset data processing requirements. If the original input data meets the preset data processing requirements, the problem evaluation value of the sub-agent to be evaluated is calculated. This scheme determines whether to calculate the problem evaluation value of the sub-agent to be evaluated based on the detection results of the input data, which can prioritize eliminating the cause of output errors due to the original input data and avoid invalid evaluation of the sub-agent to be evaluated.

[0108] Figure 8 This is a block diagram of a circuit breaker and restart system for cognitive degradation of a data governance agent, provided in an embodiment of this application. This system is used to execute a circuit breaker and restart method for cognitive degradation of a data governance agent provided in the above embodiment, and has corresponding functional modules and beneficial effects for executing the method. Figure 8 As shown, the system specifically includes:

[0109] Data generation module 101 is used to obtain the execution results of the data governance intelligent agent and generate multidimensional statistical data based on the execution results;

[0110] The comparison processing module 102 is used to compare the multidimensional statistical data with a set dynamic threshold to obtain a comparison processing result;

[0111] The path determination module 103 is used to determine one or more backtracking paths corresponding to the execution result when the comparison processing result meets the abnormal conditions of the execution result. The backtracking path includes each sequentially executed sub-agent.

[0112] The evaluation agent determination module 104 is used to determine the sub-agent to be evaluated based on the comparison processing result and the preset weights of each sub-agent in the backtracking path.

[0113] The evaluation value calculation module 105 is used to calculate the problem evaluation value of the sub-agent to be evaluated.

[0114] The circuit breaker parameter determination module 106 is used to determine the target circuit breaker parameter combination corresponding to the problem evaluation value when the problem evaluation value is greater than a preset threshold.

[0115] The circuit breaker restart module 107 is used to perform circuit breaker and restart operations on the sub-agent to be evaluated based on the target circuit breaker parameter combination.

[0116] As described above, this scheme obtains the execution results of the data governance agent, generates multidimensional statistical data based on these results, compares the multidimensional statistical data with a set of dynamic thresholds to obtain comparison results, and determines one or more backtracking paths corresponding to the execution results when the comparison results meet the abnormal conditions. Each backtracking path includes sequentially executed sub-agents. Based on the comparison results and the preset weights of each sub-agent in the backtracking path, the sub-agent to be evaluated is determined, and its problem evaluation value is calculated. If the problem evaluation value is greater than a preset threshold, the target circuit breaker parameter combination corresponding to the problem evaluation value is determined, and the circuit breaker and restart operations are performed on the sub-agent to be evaluated based on the target circuit breaker parameter combination. This scheme determines the sub-agent to be evaluated through multidimensional statistical data generated from the execution results of the data governance agent and reasonably controls its circuit breaker and restart based on the calculated problem evaluation value. This solves the problem of the lack of a reasonable circuit breaker and restart mechanism for data governance agents in existing technologies, enabling reasonable circuit breaker and restart of data governance agents and improving data governance efficiency.

[0117] In one possible embodiment, the evaluation agent determination module 104 is specifically used for:

[0118] In the presence of a backtracking path, the sub-agents in the backtracking path are arranged from high to low according to their corresponding preset weights to obtain the candidate sub-agent permutation combination;

[0119] In the case of multiple backtracking paths, the sub-agents are arranged according to the length of each backtracking path and the preset weight of each sub-agent to obtain the candidate sub-agent arrangement combination.

[0120] Based on the comparison processing results, the agent selection ratio is determined, and each of the agent selected at the highest proportion in the candidate agent permutation is identified as the agent to be evaluated.

[0121] In one possible embodiment, the evaluation agent determination module 104 is further configured to:

[0122] The path weight is determined according to the length of each backtracking path, and the path weight is multiplied by the preset weight of each corresponding sub-agent to obtain the adjustment weight of each sub-agent.

[0123] The adjustment weights of the same sub-agents in each backtracking path are superimposed to obtain the corresponding final weights. The adjustment weights of the remaining sub-agents in each backtracking path are determined as the corresponding final weights. The sub-agents in each backtracking path are arranged from high to low according to their corresponding final weights to obtain the candidate sub-agent permutation combination.

[0124] In one possible embodiment, the comparison processing module 102 is specifically used for:

[0125] The error output rate, the frequency of abnormal decisions, and the data non-association degree are compared with the corresponding dynamic thresholds in the dynamic threshold set to obtain the comparison results.

[0126] Correspondingly, the abnormal conditions for the execution result include:

[0127] The comparison results contain statistical data that are greater than the corresponding dynamic threshold.

[0128] In one possible embodiment, the fuse parameter determination module 106 is specifically used for:

[0129] The problem assessment value is compared with each preset value range, and the circuit breaker parameter combination associated with the preset value range into which the problem assessment value falls is determined as the target circuit breaker parameter combination.

[0130] In one possible embodiment, the evaluation value calculation module 105 is specifically used for:

[0131] The error output rate, the frequency of abnormal decisions, and the degree of data non-association are weighted according to historical failed task information.

[0132] The problem evaluation value of the agent to be evaluated is obtained by multiplying the error output rate, the frequency of abnormal decisions, and the degree of data non-association by their respective assigned weights and then summing them.

[0133] In one possible embodiment, an original data verification module is also included, specifically for:

[0134] The original input data associated with the sub-agent to be evaluated is obtained, the integrity of the original input data is checked to obtain the data missing rate, and the type of the original input data is checked to see if it is consistent with the predefined type. Based on the data missing rate and the detection result, it is determined whether the original input data meets the preset data processing requirements.

[0135] Accordingly, the evaluation value calculation module 105 is also used for:

[0136] If the original input data meets the preset data processing requirements, calculate the problem evaluation value of the sub-agent to be evaluated.

[0137] Figure 9 A schematic diagram of a circuit breaker and restart device for cognitive degradation of a data governance intelligent agent provided in this application embodiment is shown below. Figure 9 As shown, the device includes a processor 201, a memory 202, an input device 203, and an output device 204; the number of processors 201 in the device can be one or more. Figure 9 Taking a processor 201 as an example; the processor 201, memory 202, input device 203, and output device 204 in the device can be connected via a bus or other means. Figure 9 Taking a bus connection as an example, the memory 202, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions or modules corresponding to the circuit breaker and restart method for cognitive degradation of a data governance intelligent agent in this embodiment. The processor 201 executes various functional applications and data processing of the device by running the software programs, instructions, and modules stored in the memory 202, thereby implementing the aforementioned circuit breaker and restart method for cognitive degradation of the data governance intelligent agent. The input device 203 can be used to receive input digital or character information and generate key signal inputs related to user settings and function control of the device. The output device 204 may include a display screen or other display device.

[0138] This application also provides a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform a circuit breaker and restart method for cognitive degradation of a data governance intelligent agent. The method includes:

[0139] Obtain the execution results of the data governance intelligent agent, and generate multidimensional statistical data based on the execution results;

[0140] The multidimensional statistical data is compared with the set dynamic thresholds to obtain the comparison result. If the comparison result meets the abnormal conditions of the execution result, one or more backtracking paths corresponding to the execution result are determined. The backtracking path includes each sequentially executed sub-agent.

[0141] The sub-agent to be evaluated is determined based on the comparison processing results and the preset weights of each sub-agent in the backtracking path, and the problem evaluation value of the sub-agent to be evaluated is calculated.

[0142] If the problem evaluation value is greater than a preset threshold, a target circuit breaker parameter combination corresponding to the problem evaluation value is determined, and the circuit breaker and restart operations are performed on the sub-agent to be evaluated based on the target circuit breaker parameter combination.

[0143] It is worth noting that in the embodiments of the above-mentioned circuit breaker and restart method system for cognitive degradation of data governance intelligent agents, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the protection scope of the embodiments of this application.

[0144] Note that the above are merely preferred embodiments and the technical principles applied in this application. Those skilled in the art will understand that the embodiments of this application are not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made by those skilled in the art without departing from the protection scope of the embodiments of this application. Therefore, although the embodiments of this application have been described in detail through the above embodiments, the embodiments of this application are not limited to the above embodiments. More other equivalent embodiments may be included without departing from the concept of the embodiments of this application, and the scope of the embodiments of this application is determined by the scope of the appended claims.

Claims

1. A circuit breaker and restart method for cognitive degradation of a data governance intelligent agent, characterized in that, include: Obtain the execution results of the data governance intelligent agent, and generate multidimensional statistical data based on the execution results; The multidimensional statistical data is compared with the set dynamic thresholds to obtain the comparison result. If the comparison result meets the abnormal conditions of the execution result, one or more backtracking paths corresponding to the execution result are determined. The backtracking path includes each sequentially executed sub-agent. The sub-agent to be evaluated is determined based on the comparison processing results and the preset weights of each sub-agent in the backtracking path, and the problem evaluation value of the sub-agent to be evaluated is calculated. If the problem evaluation value is greater than a preset threshold, a target circuit breaker parameter combination corresponding to the problem evaluation value is determined, and the circuit breaker and restart operations are performed on the sub-agent to be evaluated based on the target circuit breaker parameter combination. The multidimensional statistical data includes error output rate, frequency of abnormal decisions, and data non-association degree. The frequency of abnormal decisions refers to the number of times the data governance agent triggers unexpected decisions when performing data governance tasks per unit time. The data non-association degree refers to the difference between the maximum logical association strength and the logical association strength of multidimensional data during the data governance agent's execution of data governance tasks. The step of comparing the multidimensional statistical data with the set dynamic threshold set to obtain the comparison result includes: The error output rate, the frequency of abnormal decisions, and the data non-association degree are compared with the corresponding dynamic thresholds in the dynamic threshold set to obtain comparison results; correspondingly, the abnormal conditions of the execution result include: the existence of statistical data in the comparison results that are greater than the corresponding dynamic threshold; The calculation of the problem evaluation value of the sub-agent to be evaluated includes: assigning weights to the error output rate, the frequency of abnormal decisions, and the degree of data non-association based on historical failed task information; multiplying the error output rate, the frequency of abnormal decisions, and the degree of data non-association by their respective assigned weights and summing them to obtain the problem evaluation value of the agent to be evaluated.

2. The circuit breaker and restart method for cognitive degradation of data governance intelligent agents according to claim 1, characterized in that, The step of determining the sub-agent to be evaluated based on the comparison processing result and the preset weights of each sub-agent in the backtracking path includes: In the presence of a backtracking path, the sub-agents in the backtracking path are arranged from high to low according to their corresponding preset weights to obtain the candidate sub-agent permutation combination; In the case of multiple backtracking paths, the sub-agents are arranged according to the length of each backtracking path and the preset weight of each sub-agent to obtain the candidate sub-agent arrangement combination. Based on the comparison processing results, the agent selection ratio is determined, and each of the agent selected at the highest proportion in the candidate agent permutation is identified as the agent to be evaluated.

3. The circuit breaker and restart method for cognitive degradation of data governance intelligent agents according to claim 2, characterized in that, The step of arranging sub-agents according to the length of each backtracking path and the preset weight of each sub-agent to obtain the candidate sub-agent arrangement combination includes: The path weight is determined according to the length of each backtracking path, and the path weight is multiplied by the preset weight of each corresponding sub-agent to obtain the adjustment weight of each sub-agent. The adjustment weights of the same sub-agents in each backtracking path are superimposed to obtain the corresponding final weights. The adjustment weights of the remaining sub-agents in each backtracking path are determined as the corresponding final weights. The sub-agents in each backtracking path are arranged from high to low according to their corresponding final weights to obtain the candidate sub-agent permutation combination.

4. The circuit breaker and restart method for cognitive degradation of data governance intelligent agents according to any one of claims 1-3, characterized in that, Determining the target circuit breaker parameter combination corresponding to the problem evaluation value includes: The problem assessment value is compared with each preset value range, and the circuit breaker parameter combination associated with the preset value range into which the problem assessment value falls is determined as the target circuit breaker parameter combination.

5. The circuit breaker and restart method for cognitive degradation of data governance intelligent agents according to any one of claims 1-3, characterized in that, Before calculating the problem evaluation value of the sub-agent to be evaluated, the method further includes: The original input data associated with the sub-agent to be evaluated is obtained, the integrity of the original input data is checked to obtain the data missing rate, and the type of the original input data is checked to see if it is consistent with the predefined type. Based on the data missing rate and the detection result, it is determined whether the original input data meets the preset data processing requirements. Accordingly, calculating the problem evaluation value of the sub-agent to be evaluated includes: If the original input data meets the preset data processing requirements, calculate the problem evaluation value of the sub-agent to be evaluated.

6. A circuit breaker and restart system for cognitive degradation of a data governance intelligent agent, characterized in that, include: The data generation module is used to obtain the execution results of the data governance intelligent agent and generate multidimensional statistical data based on the execution results; The comparison processing module is used to compare the multidimensional statistical data with a set dynamic threshold to obtain the comparison processing result. The path determination module is used to determine one or more backtracking paths corresponding to the execution result when the comparison processing result meets the abnormal conditions of the execution result. The backtracking path includes each sequentially executed sub-agent. The evaluation agent determination module is used to determine the sub-agent to be evaluated based on the comparison processing result and the preset weights of each sub-agent in the backtracking path. The evaluation value calculation module is used to calculate the problem evaluation value of the sub-agent to be evaluated; The circuit breaker parameter determination module is used to determine the target circuit breaker parameter combination corresponding to the problem evaluation value when the problem evaluation value is greater than a preset threshold. The circuit breaker and restart module is used to perform circuit breaker and restart operations on the sub-agent to be evaluated based on the target circuit breaker parameter combination. The multidimensional statistical data includes error output rate, frequency of abnormal decisions, and data non-association degree. The frequency of abnormal decisions refers to the number of times the data governance agent triggers unexpected decisions when performing data governance tasks per unit time. The data non-association degree refers to the difference between the maximum logical association strength and the logical association strength of multidimensional data during the data governance agent's execution of data governance tasks. The comparison processing module is specifically used to: compare the error output rate, the frequency of abnormal decisions, and the data non-association degree with the corresponding dynamic thresholds in the dynamic threshold set to obtain comparison processing results; The abnormal conditions of the execution result include: the presence of statistical data in the comparison processing result that is greater than the corresponding dynamic threshold; The evaluation value calculation module is specifically used to: assign weights to the error output rate, the frequency of abnormal decisions, and the degree of data non-association based on historical failed task information; and multiply the error output rate, the frequency of abnormal decisions, and the degree of data non-association by their respective assigned weights and sum them to obtain the problem evaluation value of the agent to be evaluated.

7. A circuit breaker and restart device for cognitive degradation of a data governance intelligent agent, characterized in that, The device includes: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement the circuit breaker and restart method for cognitive degradation of data governance agents as described in any one of claims 1-5.

8. A storage medium for storing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a computer processor, are used to perform the circuit breaker and restart method for cognitive degradation of a data governance agent as described in any one of claims 1-5.

Images