Method, device, computer device, medium and program product for job processing

Abstract

This application relates to a method, apparatus, computer equipment, medium, and program product for job processing. The method includes: acquiring a job trajectory set, an environmental observation dataset, and a target path baseline corresponding to a target job instance; calculating an environmental uninterpretability value for a step based on the step's environmental records, the step's state conflict value, and the target environmental baseline; determining a target step to be verified from at least one step based on the step's environmental uninterpretability value and the step's state conflict value; performing verification calculations on the target step to be verified based on the target path baseline to obtain the target step's path conflict information and target path conflict type; and determining the verification result of the target job instance based on the target step's environmental uninterpretability value, the target step's path conflict information, and the target path conflict type. This method can reduce the false positive rate and pinpoint the specific path conflict type.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method, apparatus, computer equipment, medium, and program product for job processing. Background Technology

[0002] As enterprises continue to introduce robotic process automation (RPA) into scenarios such as financial sharing, supply chain collaboration, customer service, report collection, and master data maintenance, more and more business processing steps are being completed automatically by robots according to preset scripts.

[0003] Currently, robot operation verification primarily relies on the difference between the baseline before execution and the result after execution to determine the success of the operation. Alternatively, it may combine operational metrics such as resource consumption, page responsiveness, interface response, and execution time to assist in judging the operation status. However, this verification method is not only prone to abnormal misjudgments due to continuous fluctuations in environmental factors, resulting in a high false positive rate, but it also easily misses issues where the process deviates but the result appears normal, and it cannot distinguish the root cause of the difference. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, computer equipment, medium, and program product for handling job conflicts that can not only reduce the false positive rate but also pinpoint the specific type of path conflict, in order to address the aforementioned technical problems.

[0005] Firstly, this application provides a method for job processing, including:

[0006] Obtain the job trajectory set, environmental observation dataset, and target path baseline corresponding to the target job instance. The job trajectory set includes the job trajectory corresponding to the target job instance. The job trajectory includes at least one step. The environmental observation dataset includes environmental records when each step is executed.

[0007] Based on the environmental record of the step, the state conflict value of the step, and the target environmental baseline, the environmental unexplainable value of the step is calculated. The environmental unexplainable value is used to describe the degree to which the behavior of the step deviates from the target environmental baseline under the environmental record. The target environmental baseline is determined from a set of candidate environmental baselines based on the environmental record of the step. The state conflict value of the step is used to describe the state difference of the step in the environmental record and the target environmental baseline.

[0008] Based on the environmental unexplainable value and the state conflict value of the step, the target step to be verified is determined from the at least one step;

[0009] Based on the target path baseline, the target steps to be verified are verified and calculated to obtain the path conflict information and target path conflict type of the target steps to be verified.

[0010] Based on the unexplainable environmental values ​​of the target verification step, the path conflict information of the target verification step, and the target path conflict type, the verification result of the target job instance is determined, and the verification result is used to describe the credibility of the result of the target job instance.

[0011] In one embodiment, based on the environmental record of the step, the state conflict value of the step, and the target environmental baseline, the environmental unexplainable value of the step is calculated, including:

[0012] Based on the action type of each step, extract the running status summary and resource usage ratio of the corresponding step from the environment record of the corresponding step;

[0013] Based on the operational status summary of the step, the resource consumption ratio of the step, and the action type of the step, calculate the environmental pressure value of the step;

[0014] Calculate the actual behavior value of each step based on the action type of each step.

[0015] Based on the environmental pressure value of the step, the actual behavior value of the step, the state conflict value of the step, and the target environmental baseline, the environmental unexplainability value of the step is calculated.

[0016] In one embodiment, the environmental pressure value of the step is calculated based on the step's operational status summary, the step's resource usage ratio, and the step's action type, including:

[0017] The execution performance value of the step is obtained by normalizing the summary of the running status of the step.

[0018] The resource occupancy ratio of the above steps is normalized to obtain the auxiliary terminal pressure value of the above steps;

[0019] Based on the action type of the step, the action type of the preceding step adjacent to the step, the running status summary of the step, and the running status summary of the preceding step, calculate the step switching exposure value of the step;

[0020] The environmental pressure value of the step is obtained by weighted summation of the execution performance value of the step, the auxiliary terminal pressure value of the step, and the step switching exposure value of the step.

[0021] In one embodiment, based on the environmental stress value of the step, the actual behavior value of the step, the state conflict value of the step, and the target environmental baseline, the environmental unexplainability value of the step is calculated, including:

[0022] The difference between the actual behavior value of the step and the first value in the target environment baseline is calculated to obtain the first difference, which is used to describe the upper boundary of the allowed behavior of the target environment baseline.

[0023] Calculate the difference between the second value in the target environment baseline and the actual behavior value of the step to obtain the second difference, which is used to describe the lower boundary of the permissible behavior of the target environment baseline;

[0024] Based on the action type of the step, set the weight of the state conflict item for the step;

[0025] Based on the weight of the state conflict item, the state conflict value of the step is weighted to obtain the weighted state conflict value.

[0026] The environmental unexplainable value of the step is calculated based on the environmental pressure value, the first difference, the second difference, and the weighted state conflict value.

[0027] In one embodiment, determining the target step to be verified from the at least one step based on the environmental unexplainable value of the step and the state conflict value of the step includes:

[0028] Based on the environmental unexplainable value of the steps, a first step to be verified is determined from the at least one step, wherein the environmental unexplainable value of the first step to be verified is greater than a preset unexplainable threshold.

[0029] Based on the state conflict value of the steps, a second step to be verified is determined from the at least one step, wherein the state conflict value of the second step to be verified is not zero;

[0030] The first step to be verified and the second step to be verified are taken as the target steps to be verified.

[0031] In one embodiment, verification calculations are performed on the target step to be verified based on the target path baseline to obtain path conflict information and target path conflict type of the target step to be verified, including:

[0032] Based on the step number of the target to be verified step, the target path node corresponding to the step number is determined from the target path baseline;

[0033] Based on the preset node existence rules, the node existence is verified for the target to be verified steps to obtain the node existence conflict value.

[0034] Based on the node association relationship of the target path node, the relationship verification is performed on the target verification step and the associated steps of the target verification step to obtain the relationship conflict value. The node association relationship is used to describe the node connection relationship between the target path node and the associated nodes of the target path node.

[0035] Based on the target object summary corresponding to the target path node, the running status summary of the target to be verified step is compared to obtain the summary verification conflict value.

[0036] Based on the preset result verification rules corresponding to the target path baseline, the running results of the target steps to be verified are verified to obtain the completion result conflict value.

[0037] Based on the node existence conflict value, the relationship conflict value, the summary verification conflict value, and the completion result conflict value, the path conflict information and target path conflict type of the target verification step are determined.

[0038] In one embodiment, the path conflict information includes path conflict values; based on the unexplainable environmental value of the target verification step, the path conflict information of the target verification step, and the target path conflict type, the verification result of the target job instance is determined, including:

[0039] Based on the criticality of the target verification step, a weight is set for the target verification step to obtain the path conflict weight of the target verification step;

[0040] Based on the path conflict weight of the target verification step, the path conflict value of the target verification step is weighted to obtain the weighted path conflict value of the target verification step.

[0041] Based on the historical interaction weights of the target verification step, the environmental unexplainable value and the path conflict value of the target verification step are weighted to obtain the target interaction value of the target verification step. The historical interaction weights are used to describe the degree of possibility that the environmental unexplainable deviation indicated by the environmental unexplainable value and the path conflict indicated by the path conflict value may occur simultaneously.

[0042] The verification risk value is calculated based on the unexplainable environmental value of the target verification step, the weighted path conflict value of the target verification step, and the target interaction value of the target verification step.

[0043] Based on the verification risk value and the target path conflict type, the verification result of the target job instance is determined.

[0044] In one embodiment, after determining the verification result of the target job instance based on the unexplainable environmental value of the target verification step, the path conflict information of the target verification step, and the target path conflict type, the method further includes:

[0045] Based on the unexplainable environmental values ​​of the target verification step, determine the environmental attribution result of the target verification step;

[0046] Based on the environmental attribution results and the target path conflict type, a target remediation suggestion is generated for the target verification step.

[0047] Based on the proposed modification of the target, the steps to be verified for the target are updated.

[0048] Secondly, this application also provides a job processing apparatus, comprising:

[0049] The acquisition unit is used to acquire the job trajectory set, environmental observation dataset and target path baseline corresponding to the target job instance. The job trajectory set includes the job trajectory corresponding to the target job instance. The job trajectory includes at least one step. The environmental observation dataset includes environmental records when each step is executed.

[0050] A calculation unit is configured to calculate an environmental unexplainable value for a step based on the environmental record of the step, the state conflict value of the step, and a target environmental baseline. The environmental unexplainable value describes the degree to which the behavior of the step deviates from the target environmental baseline under the environmental record. The target environmental baseline is determined from a set of candidate environmental baselines based on the environmental record of the step. The state conflict value describes the state difference of the step between the environmental record and the target environmental baseline.

[0051] A determining unit is configured to determine a target step to be verified from the at least one step based on an environmental unexplainable value and a state conflict value of the step.

[0052] The verification processing unit is used to perform verification calculations on the target steps to be verified based on the target path baseline, and obtain the path conflict information and target path conflict type of the target steps to be verified.

[0053] The determining unit is used to determine the verification result of the target job instance based on the unexplainable environmental value of the target verification step, the path conflict information of the target verification step, and the target path conflict type. The verification result is used to describe the credibility of the result of the target job instance.

[0054] In one embodiment, the computing unit is specifically used for:

[0055] Based on the action type of each step, extract the running status summary and resource usage ratio of the corresponding step from the environment record of the corresponding step;

[0056] Based on the operational status summary of the step, the resource consumption ratio of the step, and the action type of the step, calculate the environmental pressure value of the step;

[0057] Calculate the actual behavior value of each step based on the action type of each step.

[0058] Based on the environmental pressure value of the step, the actual behavior value of the step, the state conflict value of the step, and the target environmental baseline, the environmental unexplainability value of the step is calculated.

[0059] In one embodiment, the computing unit is specifically used for:

[0060] The execution performance value of the step is obtained by normalizing the summary of the running status of the step.

[0061] The resource occupancy ratio of the above steps is normalized to obtain the auxiliary terminal pressure value of the above steps;

[0062] Based on the action type of the step, the action type of the preceding step adjacent to the step, the running status summary of the step, and the running status summary of the preceding step, calculate the step switching exposure value of the step;

[0063] The environmental pressure value of the step is obtained by weighted summation of the execution performance value of the step, the auxiliary terminal pressure value of the step, and the step switching exposure value of the step.

[0064] In one embodiment, the computing unit is specifically used for:

[0065] The difference between the actual behavior value of the step and the first value in the target environment baseline is calculated to obtain the first difference, which is used to describe the upper boundary of the allowed behavior of the target environment baseline.

[0066] Calculate the difference between the second value in the target environment baseline and the actual behavior value of the step to obtain the second difference, which is used to describe the lower boundary of the permissible behavior of the target environment baseline;

[0067] Based on the action type of the step, set the weight of the state conflict item for the step;

[0068] Based on the weight of the state conflict item, the state conflict value of the step is weighted to obtain the weighted state conflict value.

[0069] The environmental unexplainable value of the step is calculated based on the environmental pressure value, the first difference, the second difference, and the weighted state conflict value.

[0070] In one embodiment, the determining unit is specifically used for:

[0071] Based on the environmental unexplainable value of the steps, a first step to be verified is determined from the at least one step, wherein the environmental unexplainable value of the first step to be verified is greater than a preset unexplainable threshold.

[0072] Based on the state conflict value of the steps, a second step to be verified is determined from the at least one step, wherein the state conflict value of the second step to be verified is not zero;

[0073] The first step to be verified and the second step to be verified are taken as the target steps to be verified.

[0074] In one embodiment, the verification calculation unit is specifically used for:

[0075] Based on the step number of the target to be verified step, the target path node corresponding to the step number is determined from the target path baseline;

[0076] Based on the preset node existence rules, the node existence is verified for the target to be verified steps to obtain the node existence conflict value.

[0077] Based on the node association relationship of the target path node, the relationship verification is performed on the target verification step and the associated steps of the target verification step to obtain the relationship conflict value. The node association relationship is used to describe the node connection relationship between the target path node and the associated nodes of the target path node.

[0078] Based on the target object summary corresponding to the target path node, the running status summary of the target to be verified step is compared to obtain the summary verification conflict value.

[0079] Based on the preset result verification rules corresponding to the target path baseline, the running results of the target steps to be verified are verified to obtain the completion result conflict value.

[0080] Based on the node existence conflict value, the relationship conflict value, the summary verification conflict value, and the completion result conflict value, the path conflict information and target path conflict type of the target verification step are determined.

[0081] In one embodiment, the path conflict information includes path conflict values; the determining unit is specifically used for:

[0082] Based on the criticality of the target verification step, a weight is set for the target verification step to obtain the path conflict weight of the target verification step;

[0083] Based on the path conflict weight of the target verification step, the path conflict value of the target verification step is weighted to obtain the weighted path conflict value of the target verification step.

[0084] Based on the historical interaction weights of the target verification step, the environmental unexplainable value and the path conflict value of the target verification step are weighted to obtain the target interaction value of the target verification step. The historical interaction weights are used to describe the degree of possibility that the environmental unexplainable deviation indicated by the environmental unexplainable value and the path conflict indicated by the path conflict value may occur simultaneously.

[0085] The verification risk value is calculated based on the unexplainable environmental value of the target verification step, the weighted path conflict value of the target verification step, and the target interaction value of the target verification step.

[0086] Based on the verification risk value and the target path conflict type, the verification result of the target job instance is determined.

[0087] In one embodiment, the job processing apparatus further includes a generation unit and an update processing unit; wherein,

[0088] The determining unit is further configured to, after determining the verification result of the target job instance based on the environmental unexplainable value of the target verification step, the path conflict information of the target verification step, and the target path conflict type, determine the environmental attribution result of the target verification step based on the environmental unexplainable value of the target verification step.

[0089] The generation unit is specifically used to generate target repair suggestions for the target verification step based on the environmental attribution results and the target path conflict type.

[0090] The update processing unit is specifically used to update the target verification step based on the target modification suggestion.

[0091] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the job processing method described in the first aspect above.

[0092] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the job processing method described in the first aspect above.

[0093] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the job processing method described in the first aspect above.

[0094] The aforementioned methods, apparatus, computer equipment, media, and program products for processing tasks primarily involve, after acquiring the task trajectory set, environmental observation dataset, and target path baseline corresponding to the target task instance, calculating the environmental uninterpretability value of each step based on the step's environmental records, the step's state conflict value, and the target environmental baseline. The environmental uninterpretability value describes the degree to which the step's behavior deviates from the target environmental baseline according to the environmental records. It should be understood that the task trajectory set includes the task trajectory corresponding to the target task instance, the task trajectory includes at least one step, and the environmental observation dataset includes the environmental records for each step. Furthermore, the target environmental baseline is determined from a set of candidate environmental baselines based on the step's environmental records, and the step's state conflict value describes the state difference between the step and the target environmental baseline. Thus, based on the step's environmental uninterpretability value and the step's state conflict value, a target step to be verified is determined from at least one step. Further, the target path baseline is used to perform verification calculations on the target step to be verified, obtaining the path conflict information and target path conflict type of the target step to be verified. Finally, based on the unexplainable environmental values ​​of the target verification step, the path conflict information of the target verification step, and the target path conflict type, the verification result of the target job instance is determined. The verification result is used to describe the credibility of the result of the target job instance.

[0095] By employing the above method, for the same job instance, the unexplainable values ​​of each step in the corresponding environment record can be determined by using the target environment baseline for each step. This determines the degree to which the behavior of the step deviates from the target environment baseline in that environment record. Then, using the target path baseline corresponding to the job instance, path-level automated verification is performed on the unexplainable trajectory deviations (i.e., the target steps to be verified). Thus, by forming a multi-layered verification from environment interpretation to path verification through dual baseline verification, false alarms caused by dynamic environments can be reduced, lowering the false positive rate. Furthermore, it can identify hidden defects where the result appears normal but the generated path is abnormal, and can pinpoint the anomaly to a specific path conflict type, thereby better meeting the actual needs of automated verification of job instance results in complex RPA engineering projects. Attached Figure Description

[0096] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0097] Figure 1 This is an application environment diagram of a job processing method in one embodiment;

[0098] Figure 2 This is a flowchart illustrating a job processing method in one embodiment;

[0099] Figure 3 This is a flowchart illustrating step 202 in one embodiment;

[0100] Figure 4 This is a flowchart illustrating step 204 in one embodiment;

[0101] Figure 5 This is a flowchart illustrating step 205 in one embodiment;

[0102] Figure 6 This is a flowchart illustrating a job processing method in another embodiment;

[0103] Figure 7 This is a structural block diagram of the job processing device in one embodiment;

[0104] Figure 8 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0105] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0106] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0107] As enterprises continue to introduce RPA robots into scenarios such as financial sharing, supply chain collaboration, customer service, report collection, and master data maintenance, more and more business processing steps are being completed automatically by robots according to preset scripts. To confirm whether the robot's operation has achieved the expected results, information such as operation logs, performance indicators, result files, database records, or interface screenshots are usually collected before and after the operation is executed, and verification is carried out by methods such as fixed threshold comparison, historical average comparison, rule validation, or manual sampling.

[0108] However, current solutions primarily determine the success of a robot's operation by comparing the pre-execution baseline with the post-execution result, or by combining operational metrics such as resource usage, page response time, interface response time, and execution time to assist in judging the operation status. In practical implementation, this technology can achieve basic anomaly detection under conditions of relatively stable operating environments, simple business paths, and stable page and interface responses. However, when RPA robots are deployed in complex environments such as shared servers, virtual machines, remote desktops, browser automation environments, and concurrent interactions of multiple business systems, this approach introduces other problems. For example, environmental factors such as processor usage, memory usage, page response time, and interface response time can fluctuate continuously. This can cause the same script to exhibit apparent changes such as longer page wait times, increased partial retries, delayed step transitions, or slower interface returns, even without operational defects. Traditional static baselines or simple dynamic thresholds easily misinterpret these environmentally induced normal fluctuations as operational anomalies. On the other hand, RPA job results are not solely determined by the backend computation logic; they also heavily rely on page state, control visibility, target object identification, step-by-step relationships, and external system responses. In real-world engineering, situations frequently arise where the process has deviated, but the result still appears to exist. For example, the page may jump to a similar but unexpected page, the control's positioning object may be replaced, it may be difficult to distinguish between legitimate and abnormal branches, or necessary validation nodes may be missing before the result file is generated, yet a complete output file is still produced. Many existing solutions, even when detecting discrepancies, typically only provide failure flags or difference fields, making it difficult to determine whether the discrepancy is caused by runtime environment pressure or RPA execution path deviation. This forces operations personnel to review logs, reproduce scripts, and check page objects and interface configurations, resulting in long troubleshooting cycles and low localization efficiency.

[0109] In other words, current verification of robot operations mainly relies on the difference between the baseline before execution and the result after execution to determine whether the robot operation was successful, or combines operational indicators such as resource consumption, page responsiveness, interface response, and execution time to assist in judging the operation status. However, this verification method is not only prone to abnormal misjudgments due to continuous fluctuations in environmental factors, resulting in a high misjudgment rate, but it is also prone to missing problems where the process deviates but the result appears normal, and it cannot distinguish the root cause of the difference.

[0110] To address the aforementioned technical issues, this application provides a method for job processing that employs dual-baseline verification to achieve multiple verifications, from environmental interpretation to path verification. This reduces false alarms caused by dynamic environments and lowers the false positive rate. Furthermore, it can identify hidden defects where the results appear normal but the generated paths are abnormal, and it can pinpoint the anomalies to specific path conflict types. This better meets the actual needs of automated verification of job instance results in complex RPA engineering projects.

[0111] The job processing method provided in this application embodiment can be applied to, for example, Figure 1 The application environment shown. For example... Figure 1 As shown, the application environment provided in this application includes at least: a job processing device 101. Optionally, the job processing device 101 can be connected to an RPA scheduling platform 102, a browser automation driver, an interface call record, and a runtime terminal monitoring agent.

[0112] The job processing device 101, through the RPA scheduling platform 102, invokes RPA robots deployed on the RPA scheduling platform to execute job tasks. For example, as shown... Figure 1 As shown, in S1, the job processing device 101 acquires the job trajectory set, environmental observation dataset, and target path baseline corresponding to the RPA robot executing the target job instance. The job trajectory set includes the job trajectory corresponding to the target job instance, and the job trajectory includes at least one step. The environmental observation dataset includes environmental records for each step. Next, in S2, the job processing device 101 calculates the environmental uninterpretability value of the step based on the environmental records of the step, the state conflict value of the step, and the target environmental baseline. It should be understood that the environmental uninterpretability value describes the degree to which the behavior of the step deviates from the target environmental baseline under the environmental records. The target environmental baseline is determined from a set of candidate environmental baselines based on the environmental records of the step. The state conflict value of the step describes the state difference between the step in the environmental records and the target environmental baseline. In S3, the job processing device 101 determines the target step to be verified from at least one step based on the environmental uninterpretability value and the state conflict value of the step. In S4, the job processing device 101 performs verification calculations on the target step to be verified based on the target path baseline to obtain the path conflict information and the target path conflict type of the target step to be verified. In S5, the job processing device 101 determines the verification result of the target job instance based on the unexplainable environmental values ​​of the target step to be verified, the path conflict information of the target step to be verified, and the target path conflict type. It should be noted that this verification result is used to describe the reliability of the result of the target job instance.

[0113] It should be noted that the processing device 101 in this application can be, but is not limited to, computing devices such as terminal devices, virtual machines, servers, or cloud platforms. For example, terminal devices can include, but are not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart in-vehicle devices, projection devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. Servers can be independent physical servers, server clusters or distributed systems composed of multiple physical servers, cloud servers providing cloud computing services, or other shared servers, etc.

[0114] The described RPA robot can be understood as a software program that executes tasks in different scenarios according to preset scripts. Alternatively, an RPA robot can automatically execute well-defined, highly repetitive business processes by simulating human operational logic. Furthermore, this RPA robot can be deployed on the job processing device 101 to perform job tasks.

[0115] It should be understood that the processing methods provided in this application can be applied to scenarios such as financial sharing, supply chain collaboration, customer service, report collection, and master data maintenance, and are not limited in this application.

[0116] To clearly describe the solution of this application, it will be described in detail with reference to the accompanying drawings. In an exemplary embodiment, as shown... Figure 2 As shown, a method for job processing is provided, which is applied to... Figure 2 Let's take the processing equipment in the middle as an example for explanation. Figure 2 As shown, it includes at least the following steps:

[0117] 201. Obtain the job trajectory set, environmental observation dataset, and target path baseline corresponding to the target job instance. The job trajectory set includes the job trajectory corresponding to the target job instance. The job trajectory includes at least one step. The environmental observation dataset includes environmental records when each step is executed.

[0118] In one or more embodiments, a target job instance can be understood as the current job task being executed by the RPA robot. This target job instance can be identified by a job instance identifier. Specifically, the job processing device invokes the RPA robot and receives job scheduling instructions through the RPA robot, thus enabling the job processing device to generate a job instance identifier corresponding to the current job task based on the job scheduling instructions.

[0119] For example, after generating a job processing identifier, the job processing device can also write the job processing identifier into the robot scheduling log, the operating environment acquisition record, and the job execution log, so that the environmental state and execution trajectory in the same job can be associated with the same timeline. The job instance identifier is used to connect the acquired data, execution nodes, and subsequent verification conclusions of this job instance. The actual content written uses an internal serial number or a hashed number, and the acquired fields are limited to the scope required for verification.

[0120] After determining the target job instance, the job processing device also needs to obtain the job trajectory set corresponding to the target job instance. This job trajectory set includes the job trajectory corresponding to the target job instance. The job trajectory is understood as the sequence of operation steps required to execute the target job instance. For example, the job trajectory includes at least one step.

[0121] As an illustrative description, the job processing device can obtain the job trajectory set from the RPA platform's execution logs, script execution event streams, or the robot scheduling platform interface. This job trajectory set also includes relevant step information for each step, such as step number, action type, target object summary, start time stamp, end time stamp, execution status, and retry flag. Specifically, the step number comes from pre-configured node identifiers in the RPA process designer; the action type distinguishes between click, input, read, interface call, file generation, and other operations; the target object summary is formed by de-identifying and mapping page names, control identifiers, or interface names; the execution status indicates the completion status of the step; and the retry flag indicates whether the step has been executed again.

[0122] For example, for document generation steps, the job processing device can calculate a summary of document generation time based on the start and end timestamps of that step. If no independent document generation time record is generated for that step, the summary of the time spent connecting subsequent key steps is used as a substitute value for that step. For instance, taking the RPA job task "Log in to the business system → Enter the invoice page → Select date → Export file" as an example, the job trajectory set includes the step number, action type, target object summary, start and end timestamps, execution status, and retry flag for the "Enter the invoice page" step. The business field content is expressed in a summary format, enabling subsequent path verification to identify whether the node was executed, whether it was retried, and whether the connections between preceding and following nodes are normal.

[0123] After identifying the target job instance, the job processing equipment also needs to acquire the corresponding environmental observation dataset. This dataset includes environmental records for each step in the job trajectory. These records include processor usage, memory usage, page response time, and interface response time for each step.

[0124] As an illustrative example, the environmental observation dataset mentioned in this application can be generated by a lightweight monitoring agent module in the job processing device running the RPA robot. The job processing device obtains processor and memory usage ratios by calling the operating system performance counting interface, container runtime status interface, or virtual machine monitoring interface through the monitoring agent module. Additionally, page response time can be obtained from browser automation drivers deployed by the job processing device or wait events on the RPA platform page. For example, when the RPA robot executes the "open report page" step, the job processing device records a summary of the time elapsed between the initiating page and the target control entering an operable state to obtain the page response time. Furthermore, the mentioned interface response time is generated by the request start recording and return records when the job processing device calls external business interfaces through the RPA robot script. For example, when the RPA robot calls the "query order status interface," it writes an interface name summary, call status, and response time summary to obtain the interface response time.

[0125] This application can also normalize the environmental records and operational status summaries that need to be compared later. The normalization process mentioned in this application uses the minimum-maximum linear mapping method in statistical data preprocessing. Its original idea is to map the minimum reference value of the same type of data to zero, map the maximum reference value to one, and let the intermediate data fall between the two in a linear proportion.

[0126] This application applies the basic mapping relationships based on normalization to normal historical samples of the same RPA job type, comparing page response time with historical boundaries, interface response time with historical boundaries, and processor utilization with historical boundaries. This ensures that the same type of data has a unified relative expression in subsequent environmental baselines. The calculation method is as follows:

[0127]

[0128] Where x is the raw value of a certain type of environmental record or trajectory time summary collected in this target task instance, specifically from lightweight monitoring agents, browser automation drivers, RPA platform page wait events or interface call records. minIt is the minimum reference value for the corresponding data category in the historical normal samples of the same RPA job type; x max It is the maximum reference value for the corresponding data category in the historical normal samples of the same RPA job type; z is the processed normalized value, used for subsequent environmental baseline construction. If x max With x min Similarly, the system retains this type of data as a fixed reference field and updates the corresponding reference boundary after subsequent sample accumulation.

[0129] For example, in a certain type of "Export Daily Report" job instance, the minimum reference value x for forming a summary of historical normal page response times. min Take 200, maximum reference value x max If we take 1000 and x = 600 for the current job's collected page response time summary, substituting these values ​​into the formula above, the result is 0.5, indicating that the current page response is in the middle of the historical normal boundary for this job type. If x = 1200 for the currently collected page response time summary, a normalized value greater than 1 is calculated in the same way. This result is retained as a high-level environmental performance in the environmental observation set and is included in the environmental baseline determination in the next step. Through this processing, short-term page slowdowns, increased interface response times, or increased resource consumption are all transformed into relative expressions under the same job type. Subsequent steps can then use these relative expressions to determine whether they belong to explainable fluctuations under the current environmental conditions.

[0130] After the above processing, the job processing device outputs an environment observation set and a job trajectory set. The environment observation set includes processor usage ratio, memory usage ratio, page response time summary, and interface response time summary for the same target job instance, and performs job instance association, time alignment, field pruning, and necessary normalization. The job trajectory set consists of step number, action type, target object summary, start time marker, end time marker, execution status, and retry marker for the same target job instance, representing the critical execution path actually traversed by the RPA robot in this job. The environment observation set is used in the next step to construct the environmental baseline corresponding to the current target job instance, and the job trajectory set continues to be used in subsequent environmental attribution and path verification.

[0131] In addition, for this target task instance, the task processing device also needs to obtain the target path baseline corresponding to the target task instance. Through this target path baseline, the planned execution path of the RPA robot when executing this target task instance can be clearly expressed, which is mainly used for the subsequent path verification process.

[0132] 202. Based on the step's environment record, the step's state conflict value, and the target environment baseline, calculate the step's environment unexplainable value. The environment unexplainable value describes the degree to which the step's behavior deviates from the target environment baseline in the environment record. The target environment baseline is determined from the candidate environment baseline set based on the step's environment record. The step's state conflict value describes the state difference between the step in the environment record and the target environment baseline.

[0133] In one or more embodiments, after obtaining the job trajectory set and environmental observation dataset corresponding to the target job instance, the job processing device can also determine the environmental baseline that matches the environmental record of the step from the candidate environmental baseline set, based on the environmental record of the step, as the target environmental baseline.

[0134] Specifically, the environmental pressure value of the current step can be calculated based on the environmental records of the step, thereby determining the environmental pressure range into which this environmental pressure value falls. Then, from the candidate environmental baseline set, a candidate environmental baseline that matches the target operation instance's function type, step number, action type, and the environmental pressure range into which the environmental pressure value falls is selected as the target environmental baseline. It should be understood that through this target environmental baseline, the resource consumption and response performance that the step should exhibit within the ideal environmental pressure range can be understood, thus providing an objective benchmark for quantifying the environmental deviation of step execution.

[0135] For example, the job processing device can bind the same job type, the same step number, the same action type, and the same environmental pressure range, so that the generated target environmental baseline can reflect the normal performance differences of the corresponding steps under different environmental pressures. For instance, for page jump steps, the sorting object is the page response time summary; for interface call steps, the sorting object is the interface response time summary; for file generation steps, the sorting object is the terminal resource pressure-related summary and the step connection time summary. The job processing device takes the sample value near the bottom after sorting as the lower boundary of allowed behavior and the sample value near the top after sorting as the upper boundary of allowed behavior, and appropriately adjusts the upper boundary of cross-page or cross-system steps in combination with the step switching exposure value. If the number of historical normal samples within the same job type, the same step number, the same action type, and the current environmental pressure range is less than the preset minimum number of samples, it first expands to the adjacent environmental pressure range, and then expands to adjacent steps of the same job type and the same action type. If it is still insufficient after expansion, the most recent valid environmental baseline is used and a sample insufficiency mark is written. If only one valid normal sample is found after filtering, that sample value serves as both the lower and upper bounds of the allowed behavior. This correction amount is derived from the difference in the upper bounds of cross-page steps and same-page steps within the same job type in the normal historical samples, and is applied proportionally according to the step switching exposure value of the current step.

[0136] Taking the "Enter Invoice Page" step as an example, if the environmental stress value of this step falls within the medium-to-high stress range, the task processing equipment will select historical normal samples from the historical normal sample boundary table that are of the same task type, the same "Enter Invoice Page" step, the same page jump action type, and fall within the medium-to-high stress range. Assuming that the page response time summary of these historical normal samples, after sorting, has a lower representative value of 0.38 and an upper representative value of 0.82; within the same task type, cross-page steps have a stable difference of 0.06 at the upper boundary compared to same-page operations, and the current step switching exposure value is 0.5, then the system will convert this stable difference to 0.03 according to the step switching exposure value and adjust the upper boundary to 0.85. The resulting target environmental baseline can be interpreted as follows: under the current task type, current step, current action type, and current environmental stress conditions, when the page response time summary falls between 0.38 and 0.85, it can be used as an attribution judgment for the environmentally explainable performance of this step to proceed to the next step. For API call steps, the system performs the same filtering, sorting, and boundary generation based on the API response time summary; for file generation steps, the system performs the same boundary generation based on the terminal resource pressure-related summary and the step connection time summary, so that each type of action obtains an environmental baseline that matches its engineering sensitivities.

[0137] It should be noted that the environmental stress values ​​and step switching exposure values ​​mentioned above can be found in the following sections. Figure 3 The content described will be understood, but will not be elaborated upon here.

[0138] Thus, after determining the target environmental baseline, the job processing equipment can calculate the environmental unexplainable value for each step based on the environmental records of the steps, the state conflict values ​​of the steps, and the target environmental baseline. It should be understood that the environmental unexplainable value of a step can describe the degree to which its behavior deviates from the target environmental baseline under the corresponding environmental records when executing that step, thereby accurately identifying potential execution risks caused by environmental disturbances.

[0139] In addition, the state conflict value of the described steps can be understood as the state difference between the steps in the environmental record and the target environmental baseline. It can characterize the normal time consumption surface, but the execution steps have deviated from the expected.

[0140] For example, the job processing equipment can quantify the state deviation between a step and the expected target environment baseline by comparing the execution status, retry flag, action type, and target object summary recorded for a step in the corresponding environment with the execution status, allowed retry status, action type, and target object summary in the target environment baseline. For instance, if the execution status is complete, the retry flag falls within the allowed range, the action type is consistent, and the target object summary is consistent, then the state conflict value for that step is zero, indicating that the step conforms to the expected behavior specification in the target environment baseline. Conversely, if there are cases where the step is incomplete, the retry status exceeds the allowed range, the action type is inconsistent, or the target object summary is substituted, then the state conflict value for the step is configured as a relative value according to the severity of the corresponding conflict type in the job template, thus forming a quantifiable state conflict value.

[0141] It should be understood that the described job template is a technical template pre-configured by the RPA robot during the job deployment phase. It is used to describe which steps the job should include, the action type of each step, the target object summary, the retry status, the allowed branches, and the severity of different conflict types, and can provide an expected judgment standard.

[0142] Optionally, after the target environment baseline is generated, the job processing equipment can also write the environmental pressure range and permissible behavior boundary of each step back to the step record corresponding to the job trajectory set to form a baseline-related job trajectory set.

[0143] The baseline-associated job trajectory set retains the step number, action type, target object summary, start time marker, end time marker, execution status, and retry marker for each job trajectory in the job trajectory set. It also adds the corresponding environmental pressure range, allowed behavior boundary, and environmental baseline index. Simultaneously, the job processing device also writes the allowed retry status and allowed target object summary set corresponding to historical normal samples under the current environmental pressure range into the corresponding environmental baseline record. If multiple valid target object summaries exist for the same step in normal samples, the set of valid summaries preset in the job template is retained. For example, the "Enter Invoice Page" step carries a page jump-type environmental baseline index, the "Query Order Status Interface" step carries an interface call-type environmental baseline index, and the "Generate Local Report File" step carries a file generation-type environmental baseline index. During subsequent environmental attribution, the job processing device can directly compare the actual response, waiting, connection, or retry performance with the corresponding boundaries based on the environmental baseline index carried by each step.

[0144] Thus, the job processing equipment can also output a target environmental baseline and a baseline-related job trajectory set. The target environmental baseline consists of job type, step number, action type, environmental pressure range, permissible behavior boundaries, and step switching correction information, used to express the normal performance range of each key step in the current target job instance under the corresponding environmental pressure. The baseline-related job trajectory set is extended from the job trajectory set, where each step record carries a corresponding environmental baseline index. The target environmental baseline and the baseline-related job trajectory set are used together to calculate environmentally unexplainable values, so as to directly determine whether the trajectory deviation of a certain step falls within the permissible range of the corresponding environmental baseline, and further distinguish between environmentally explainable deviations and job deviations that require path verification.

[0145] 203. Based on the unexplainable environmental values ​​and conflicting state values ​​of the steps, determine the target step to be verified from at least one step.

[0146] In one or more embodiments, after calculating the environmental unexplainable value and the state conflict value of the step, the job processing device can perform dual-dimensional discrimination on each step based on a preset unexplainable threshold and whether the state conflict value is 0, so as to filter out the target steps to be verified that meet the dual threshold conditions.

[0147] For example, the job processing equipment can compare the environmental unexplainable value of each step with a preset unexplainable threshold. If the environmental unexplainable value of a certain step (e.g., the first step to be verified) is greater than the preset unexplainable threshold, then the first step to be verified is designated as the target step to be verified.

[0148] Similarly, the job processing equipment can also directly determine whether the state conflict value of each step is 0. If it is determined that the state conflict value of a certain step (such as the second step to be verified) is not zero, then the second step to be verified will also be included in the target steps to be verified.

[0149] In other words, the target verification steps in this application include both steps where the unexplainable environmental value exceeds a preset unexplainable threshold and steps where the state conflict value is non-zero. This allows for the accurate capture of unexplainable environmental steps from two dimensions, while also taking into account steps with abnormal states, avoiding the omission of any abnormal steps and providing a basis for subsequent accurate verification.

[0150] 204. Based on the target path baseline, perform verification calculations on the target steps to be verified to obtain the path conflict information and target path conflict type of the target steps to be verified.

[0151] In one or more embodiments, after the job processing device determines the target step to be verified from at least one step based on the environmental unexplainable value and the state conflict value of each step, it can perform verification calculations on the target step to be verified based on the target path baseline to obtain the path conflict information and target path conflict type of the target step to be verified.

[0152] Path conflict information includes path conflict values. Path conflict values ​​characterize the degree of path deviation between the target step to be verified and the target path baseline, that is, the degree of deviation of the target step to be verified from the target path baseline.

[0153] The target path conflict type can reflect the type of conflict that causes the current target path deviation, including but not limited to the addition of abnormal nodes, missing necessary nodes, abnormal predecessor-successor relationships, target object replacement, or missing completion result nodes, etc., which are not limited in this application.

[0154] 205. Based on the unexplainable environmental values ​​of the target verification step, the path conflict information of the target verification step, and the target path conflict type, determine the verification result of the target job instance. The verification result is used to describe the credibility of the result of the target job instance.

[0155] In one or more embodiments, after determining the path conflict information and target path conflict type of the target step to be verified, the job processing device can determine the verification result of the target job instance by combining the environmental uninterpretable values ​​of the target step to be verified. In this way, the reliability of the result of the target job instance can be described through the verification result.

[0156] In this embodiment, for the same job instance, the unexplainable values ​​of the corresponding steps in the corresponding environment record are first determined by the target environment baseline corresponding to each step, thereby determining the degree to which the behavior of the step deviates from the target environment baseline in the environment record. Then, using the target path baseline corresponding to the job instance, path-level automated verification is performed on the trajectory deviations that are unexplained by the environment (i.e., the target steps to be verified). In this way, by forming a multi-level verification from environment interpretation to path verification through dual baseline verification, false alarms caused by dynamic environments can be reduced, and the false judgment rate can be reduced. Moreover, it can identify hidden defects where the result appears normal but the generated path is abnormal, and it can also locate the anomaly to a specific path conflict type, thus better meeting the actual needs of automated verification of job instance results in complex RPA engineering sites.

[0157] In some alternative embodiments, the method for calculating the environmentally unexplainable value in step 202 can be referred to... Figure 3 Use the flowchart shown to understand the process. Figure 3 As shown, step 202 includes steps 301 to 304. Wherein:

[0158] 301. Based on the action type of each step, extract the running status summary and resource usage ratio of the corresponding step from the environment record of the corresponding step.

[0159] In one or more embodiments, since the environmental records corresponding to each step contain a summary of the operating status and the resource usage ratio collected when the RPA robot executes that step, the job processing device can accurately extract the operating status summary and resource usage ratio related to the action type from the environmental records according to the action type of each step. It should be understood that the operating status summary includes page response time summary, interface response time summary, step connection time summary, etc.; the resource usage ratio includes processor usage ratio and memory usage ratio.

[0160] For example, taking steps with action types including "page jump" as an example, a summary of page response time can be extracted; taking "interface call" as an example, a summary of interface response time can be extracted; taking "file generation" as an example, a summary of file generation time and a summary of step connection time can be extracted; and for "normal input" and "same page reading" steps, a summary of step connection time can be extracted.

[0161] 302. Calculate the environmental pressure value of a step based on its operational status summary, resource usage ratio, and action type.

[0162] In one or more embodiments, in order to more accurately characterize the environmental pressure expression of the step when the RPA robot executes the step, the job processing device needs to calculate the environmental pressure value of the step by combining the action type of the step after extracting the operation status summary and resource occupancy ratio of each step.

[0163] It should be understood that the environmental stress value of this step can reflect the level of environmental stress that the step is subjected to when it is performed in the target environment.

[0164] Optionally, the calculation of the environmental pressure value in step 303 can be understood with reference to the following embodiment:

[0165] First, the job processing equipment normalizes the summary of the running status of the step to obtain the execution performance value of the step.

[0166] The job processing equipment can normalize the extracted runtime status summaries based on the action type of each step. For example, for steps involving "page jumps," the extracted page response time summary can be normalized and mapped to an execution performance value within the range of 0 to 1. Similarly, for "interface call" steps, the interface response time summary is normalized to obtain the corresponding execution performance value. For "file generation" steps, the file generation time summary and the step connection time summary are normalized separately and then weighted and merged to obtain a comprehensive execution performance value. For steps such as "normal input" and "same page reading," the extracted step connection time summary can be normalized to obtain the corresponding execution performance value.

[0167] Next, the task processing equipment normalizes the resource occupancy ratio of the step to obtain the auxiliary terminal pressure value of the step.

[0168] For each step, the task processing device can also normalize the resource usage ratios for each step, such as processor usage and memory usage ratios, to obtain an auxiliary terminal pressure value within the range of 0 to 1, ensuring comparability and consistency across different resource dimensions. It should be understood that the auxiliary terminal pressure value for a step directly represents the actual intensity of terminal hardware resource usage during the execution of that step. The higher the value, the heavier the resource consumption in the terminal, thus reflecting the impact of terminal resource pressure on the RPA robot when executing the step.

[0169] Specifically, after normalizing the processor usage ratio and memory usage ratio, the job processing device can select the maximum value from the normalized results as the auxiliary terminal pressure value for this step.

[0170] Furthermore, since the same type of environmental fluctuations have different impacts on different action types, and page switching during execution, switching between different systems, and switching of interface calls are more susceptible to environmental disturbances in actual operation, this application also introduces a step switching exposure item to describe the execution stability fluctuations caused by environmental differences during step switching, in order to form an environmental stress expression suitable for RPA scenarios. Specifically, the job processing device also needs to calculate the step switching exposure value of the step based on the action type of the step, the action type of the preceding step adjacent to the step, the running state summary of the step, and the running state summary of the preceding step.

[0171] Specifically, for switching steps such as cross-page switching, cross-system navigation, or API call switching, the job processing device can select the maximum value between the normalized value of the running status summary of the current step and the normalized value of the running status summary of the previous step as the step switching exposure value for that step. Alternatively, for switching operations within the same page, the job processing device selects the minimum value between the normalized value of the running status summary of the current step and the normalized value of the running status summary of the previous step as the step switching exposure value for that step, in order to accurately quantify the impact of environmental disturbances on the stability of step execution.

[0172] Finally, after calculating the performance value, auxiliary terminal pressure value, and step switching exposure value of each step, the job processing equipment can perform a weighted summation of the performance value, auxiliary terminal pressure value, and step switching exposure value of the step to obtain the environmental pressure value of the step.

[0173] As an illustrative description, the environmental pressure value of the steps in this application can satisfy the following expression, namely:

[0174]

[0175] Where, q i It is the environmental pressure value corresponding to the i-th step in the job trajectory set, used to determine which environmental baseline this step matches; i comes from the i-th step in the job trajectory set; u i It is the performance value of the i-th step; v i It is the auxiliary terminal pressure value of the i-th step; a i λ is the action sensitivity weight of the i-th step, derived from the action type configuration in the job template. Steps involving page jumps and API calls take higher values, while steps involving ordinary input and reading from the same page take lower values. The value for file generation steps is determined based on the file generation time sensitivity configuration in the job template. λ is the configuration coefficient for the step switching exposure value, calibrated by the environmental fluctuation differences between cross-page, cross-interface, and same-page operations in historical normal samples of the same job type. i This is the step-by-step exposure value switch for the i-th step. i v i and h i All use relative expressions, a i Since λ is a proportional configuration, the calculated q i It is also a relative expression of environmental pressure, which can be directly used for the selection of target environmental baselines and the subsequent calculation of environmental unexplainable values.

[0176] For example, taking the "enter the invoice page" page jump step as an example, the job processing device reads the start and end timestamps of this step from the job trajectory set, and extracts the page response time summary, processor usage ratio, and memory usage ratio within this time interval from the environmental observation set. If the normalized value of the page response time summary is 0.75, the normalized value of the processor usage ratio is 0.4, and the normalized value of the memory usage ratio is 0.35, then the execution performance value of this step is taken as 0.75, and the auxiliary terminal pressure value is taken as the higher of the processor usage ratio and the memory usage ratio, i.e., 0.4. Since this step belongs to the page jump step, the action sensitivity weight is taken as 0.7; since this step enters the new invoice page by clicking the menu, the step switch exposure value is taken as 0.5; historical normal samples of the same job type show that page switch steps have stable additional fluctuations compared to the same page operation, so the configuration coefficient of the step switch exposure value is taken as 0.1. Substituting into the formula, it equals 0.7×0.75 + 0.3×0.4 + 0.1×0.5, and the calculated environmental pressure value is 0.695, indicating that the "enter the invoice page" step is under a medium to high environmental pressure.

[0177] By using the above methods and multi-dimensional data such as the execution performance value of the step, the auxiliary terminal pressure value of the step, and the step switching exposure value of the step, the environmental pressure value of the step can be calculated. This not only enables a quantitative assessment of the environmental adaptability of the step in the entire target operation instance, but also accurately identifies the differentiated impact of environmental disturbances on the stability of the workflow, providing a quantifiable decision-making basis for subsequent verification of whether the target as an instance has risks.

[0178] 303. Calculate the actual behavior value of each step based on the action type of each step.

[0179] In one or more embodiments, in addition to calculating the environmental pressure value of each step, the job processing device also needs to calculate the actual behavior value of each step based on the action type of each step. It should be understood that the actual behavior value of a step can characterize the degree of behavioral response of the RPA robot when performing that step.

[0180] For example, the job processing device can determine the corresponding operating status summary based on the action type of the step, and then calculate the actual behavior value of the step based on the operating status summary.

[0181] For example, the actual behavior value of page jump steps can be taken as a summary of page response time; the actual behavior value of interface call steps can be taken as a summary of interface response time; the actual behavior value of file generation steps can be taken as a summary related to terminal resource pressure or a summary of step connection time; and the actual behavior value of ordinary click, input, and read steps can be taken as a summary of the connection time between this step and the preceding and following key steps. For records of the first key step, the last key step, or records lacking adjacent key steps, the step connection time summary is taken as the available value between the duration of the current step and the connection time of the only adjacent key step. If the main behavior value required for the current action type is missing, the step retains the missing test mark, does not participate in continuous deviation calculation, but continues to participate in state conflict judgment and subsequent path verification.

[0182] 304. Based on the environmental stress value of the step, the actual behavior value of the step, the state conflict value of the step, and the target environmental baseline, calculate the environmental unexplainable value of the step.

[0183] In one or more embodiments, after calculating the environmental pressure value of the step, the actual behavior value of the step, the state conflict value of the step, and the target environmental baseline, the job processing device can calculate the environmental unexplainable value of the step based on the environmental pressure value of the step, the actual behavior value of the step, the state conflict value of the step, and the target environmental baseline.

[0184] For example, the work processing equipment can calculate the environmental unexplainable value of a step by calculating the difference between the actual behavior value and the target environmental baseline, and by combining the environmental stress value and the state conflict value.

[0185] It should be noted that the larger the environmental uninterpretability value of this step, the more the performance of this step deviates from the target environment baseline under the corresponding environment record; conversely, the smaller the environmental uninterpretability value of this step, the closer the performance of this step is to the target environment baseline under the corresponding environment record.

[0186] In some alternative embodiments, the calculation of the environmental pressure value in step 304 can be understood with reference to the following embodiments:

[0187] First, the job processing device extracts a first value from the permissible behavior boundary corresponding to the target environmental baseline. This first value characterizes the upper boundary of permissible behavior of the target environmental baseline. The described upper boundary of permissible behavior can be understood as the maximum permissible behavior threshold of the target environmental baseline. Based on this, the job processing device calculates the difference between the actual behavior value of the step and the first value in the target environmental baseline to obtain the first difference.

[0188] Similarly, the job processing equipment also needs to extract a second value from the permissible behavior boundary corresponding to the target environmental baseline. This second value characterizes the lower boundary of the permissible behavior of the target environmental baseline, i.e., the minimum permissible behavior threshold of the target environmental baseline. Thus, the job processing equipment calculates the difference between the second value in the target environmental baseline and the actual behavior value of the step to obtain the second difference.

[0189] Next, the job processing device can also set state conflict item weights for steps based on the action type of each step. As an illustrative example, the job processing device can match preset state conflict items to the action type of a step according to the state conflict item rules in the job template and generate corresponding state conflict values. For example, for steps such as page jumps and API calls, because the action characteristics of these steps are prone to causing state conflicts, a higher state conflict item weight can be configured for these types of steps. Conversely, for steps such as ordinary clicks, input, and reading, because their actions have a weaker impact on the state, a lower state conflict item weight can be configured.

[0190] Thus, the job processing equipment further weights the state conflict values ​​of the steps based on the configured weights of the state conflict items, obtaining a weighted state conflict value. It should be understood that the state conflict value described here can be understood with reference to the content described in step 202 above, and will not be repeated here.

[0191] Finally, the work processing equipment calculates the unexplainable environmental value of the step based on the environmental pressure value, the first difference, the second difference, and the weighted state conflict value.

[0192] As an illustrative description, the uninterpretable environmental values ​​of the steps in this application can satisfy the following expression, namely:

[0193]

[0194] Among them, g i y is the environmental unexplainable intensity value of the i-th step in the job trajectory set, used to form attribution results and determine whether this step is a target to be verified step. i comes from the step number in the baseline-related job trajectory set. i This is the actual behavior value for the i-th step. L i It is the lower boundary of the allowed behavior of the i-th step within the target environment baseline. i q represents the upper boundary of the allowed behavior of the i-th step within the target environment baseline. i c is the environmental pressure value for the i-th step. i y is the state conflict value at the i-th step, and β is the weight of the state conflict term. i L i U i q iand c i All are relative expressions, 1+q i For the correction term of environmental pressure value, β×c i The calculated g is the state conflict correction term. i It is still expressed as relative intensity.

[0195] For example, taking the "enter the invoice page" step as an example, its corresponding target environment baseline's allowed behavior lower boundary L i The value is 0.38, allowing the upper boundary U of the behavior. i The environmental pressure value q is 0.85. i The value is 0.695. If the actual page response time y for this step in this task is summarized as follows: i The value is 0.79, the execution status is successful, the target object summary is consistent with the target object summary in the environmental baseline, the retry mark falls within the allowable state of the current environmental pressure range, then the upper deviation is 0, the lower deviation is 0, and its state conflict value c i If we set it to 0, then we can calculate the unexplainable environmental value g. i If the value is 0, the system records this step as an environment-interpretable step.

[0196] Conversely, if the actual page response time summary for this step in this task is y i If the execution status is successful and the target object summary is consistent, the upper bias is 0.1, the lower bias is 0, the denominator is 1.695, and the continuous bias, after being corrected for environmental stress, is approximately 0.059. If the attribution threshold for page jump steps in the current job template is set to 0.08, the system records in the attribution results that this step has an out-of-bounds bias, but because the environmental unexplainability intensity does not exceed the attribution threshold and the status matches, this step is classified as an environmentally explainable step. If the actual page response time summary y of this step in this job is... i The value is 0.79, but the target object summary shows that an unexpected page has been entered. In this case, the continuity deviation is 0, and the state conflict value is c. i The severity of the target object is configured accordingly. When the weight β of the state conflict item in the page jump step is 0.4, the inexplicability of the environment is dominated by the state conflict item. The system will treat this step as the target step to be verified and write it into the path verification trajectory set.

[0197] By using the above method, the environmental unexplainable value of a step is calculated by using the environmental stress value of the step, the difference between the actual behavior value and the target environmental baseline, and the weighted state conflict value, thereby accurately identifying key steps that are significantly affected by environmental disturbances or have potential state anomalies.

[0198] In some optional embodiments, the determination of the path conflict information and target path conflict type for the target step to be verified in step 204 can be referred to... Figure 4 The illustrated embodiments are used for understanding. Figure 4 As shown, step 204 includes at least the following steps 401 to 406, namely:

[0199] 401. Based on the step number of the target to be verified step, determine the target path node corresponding to the step number from the target path baseline.

[0200] In one or more embodiments, since the target path baseline is a planned trajectory configured for the target job instance, it includes each path node and a path number for each path node. After determining the target verification step, the job processing device can match the step number of the target verification step with the path number in the target path baseline. If a match is found, the path node corresponding to the matched path number is taken as the target path node corresponding to the step number of the target verification step.

[0201] It should be noted that the step number of the target to be verified step can be understood as the step sequence number of the target to be verified step in the actual operation trajectory.

[0202] 402. Based on the preset node existence rules, perform node existence verification on the target steps to be verified to obtain node existence conflict values.

[0203] In one or more embodiments, after determining the target path node, the job processing device can perform node existence verification on the target to be verified step based on preset node existence rules to obtain the corresponding node existence conflict value.

[0204] For example, the job processing equipment checks whether the target step to be verified exists in the standard step set of the target path baseline, and also checks whether the mandatory steps in the target path baseline appear in the actual job trajectory. If a step not allowed by the target path baseline appears in the actual job trajectory, a new abnormal node is recorded; if a mandatory step is missing, a missing node is recorded.

[0205] For example, if the target step to be verified has no corresponding node in the target path baseline, the node existence conflict value is set to 1; if a corresponding node exists, the node existence conflict value is set to 0. It should be noted that the above values ​​of 1, 0, etc., for the node existence conflict value are only illustrative descriptions, and in practical applications, other values ​​may be used, such as 0.7, etc., which are not limited in this application.

[0206] 403. Based on the node association relationship of the target path node, perform relationship verification on the target to be verified step and the associated steps of the target to be verified step to obtain the relationship conflict value. The node association relationship is used to describe the node connection relationship between the target path node and the associated nodes of the target path node.

[0207] In one or more embodiments, the job processing device also needs to extract the node association relationship of the target path node, so as to identify the associated nodes of the target path node, such as predecessor nodes and successor nodes, from the node association relationship. The predecessor node and successor node described are respectively the preceding execution step and the subsequent execution step of the target path node in the target path baseline.

[0208] After extracting the node association relationship of the target path node, the job processing equipment performs relationship verification on the target to be verified step and the associated steps of the target to be verified step based on the node association relationship, and obtains the relationship conflict value.

[0209] For example, the job processing equipment checks whether the preceding and subsequent steps of the target verification step conform to the node association relationship corresponding to the target path node in the target path baseline. If the preceding step is missing or the subsequent step is misaligned in the actual trajectory, the relationship conflict value is set to 1, and the connection offset is recorded; if they match perfectly, the value is set to 0. It should be noted that the above relationship conflict value of 1, 0, etc. is only an illustrative description, and in actual applications it can be other values, such as 0.7, etc., which are not limited in this application.

[0210] 404. Based on the target object summary corresponding to the target path node, compare the running status summary of the target to be verified step to obtain the summary verification conflict value.

[0211] In one or more embodiments, the job processing device performs semantic alignment and key attribute comparison between the target object summary corresponding to the target path node and the running status summary of the target to be verified step, thereby calculating the summary verification conflict value.

[0212] For example, the job processing device checks whether the page response time summary, interface response time summary, and connection time summary corresponding to the target to be verified step fall within the target object summary set allowed by the target path baseline. If not, the summary verification conflict value is set to 1, and the target object is marked as a substitute. If all match, the summary verification conflict value is set to 0. It should be noted that the above-mentioned verification conflict value of 1, 0, etc. is only an illustrative description, and in actual applications, it can also be other values, such as 0.7, etc., which are not limited in this application.

[0213] 405. Based on the preset result verification rules corresponding to the target path baseline, verify the running results of the target steps to be verified and obtain the completion result conflict value.

[0214] In one or more embodiments, the job processing device may also verify the running results of the target steps to be verified according to the preset result verification rules corresponding to the target path baseline, and obtain the completion result conflict value.

[0215] For example, the job processing equipment checks whether the result of the step to be verified is generated by a path node allowed by the target path baseline. For instance, if the job template specifies that a result file can only be generated after passing through the "Export File Complete" node, and the actual job trajectory lacks this node but a result file record appears, the record is considered to be of incomplete completion source.

[0216] 406. Based on the node existence conflict value, relationship conflict value, summary verification conflict value, and completion result conflict value, determine the path conflict information and target path conflict type of the target to be verified step.

[0217] In one or more embodiments, after calculating the node existence conflict value, relationship conflict value, summary verification conflict value, and completion result conflict value, the job processing device can determine the path conflict information and target path conflict type of the target step to be verified based on the conflict values ​​of these four dimensions.

[0218] Specifically, the job processing equipment can calculate the path conflict value based on the node existence conflict value, relationship conflict value, summary verification conflict value, and completion result conflict value to obtain path conflict information.

[0219] In other words, the path conflict value is generated by the conflict severity rules preset in the job template. Specifically, the relative conflict values ​​are as follows: newly added abnormal nodes, missing necessary nodes, abnormal predecessor-successor relationships, target object substitution, and missing completion result nodes. When the same step triggers one or more path conflicts, the job processing device can select the highest relative conflict value from these four types of conflict values ​​as the path conflict value of the target step to be verified, and retain all triggered conflict types in the verification results as the target path conflict type of the target step to be verified.

[0220] Conversely, when no path conflict is triggered in the same step, that is, when all four types of conflict values ​​are 0, it indicates that the verification has passed, and the path conflict value of the target to be verified step is zero.

[0221] By using the above methods, the path consistency verification of the target verification steps is completed from multiple dimensions, which not only avoids the risk of misjudgment from a single dimension, but also significantly improves the robustness and interpretability of path compliance determination.

[0222] In some alternative embodiments, the determination of the verification result of the target job instance in step 205 can be referred to... Figure 5 The illustrated embodiments are used for understanding. Figure 5 As shown, step 205 includes at least the following steps 501 to 505, namely:

[0223] 501. Based on the criticality of the target verification steps, set weights for the target verification steps to obtain the path conflict weights of the target verification steps.

[0224] In one or more embodiments, the job processing device can invoke a job template and dynamically assign path conflict weights to the target verification step based on the step criticality of the target verification step. For example, the higher the step criticality of the target verification step, the greater its path conflict weight.

[0225] 502. Based on the path conflict weight of the target verification step, the path conflict value of the target verification step is weighted to obtain the weighted path conflict value of the target verification step.

[0226] In one or more embodiments, the job processing device weights the path conflict weights with the path conflict values ​​of the target step to be verified to obtain a weighted path conflict value, thereby more accurately reflecting the actual risk contribution of the target step to be verified in the overall job process.

[0227] 503. Based on the historical interaction weights of the target verification step, the unexplainable environmental value and the path conflict value of the target verification step are weighted to obtain the target interaction value of the target verification step.

[0228] In one or more embodiments, the described historical interaction weights are used to describe the likelihood of simultaneous occurrence of environmentally unexplainable biases indicated by environmentally unexplainable values ​​and path conflicts indicated by path conflict values. A larger historical interaction weight indicates a higher proportion of simultaneous occurrence of environmentally unexplainable biases and path conflicts.

[0229] The job processing device can extract the historical interaction weights of each verification step under the corresponding historical anomaly record of the target job instance based on the job type. Thus, based on the step number of the target verification step, the historical interaction weight of the target verification step can be accurately matched from the historical interaction weights of these multiple verification steps. In this way, the job processing device can weight the unexplainable environmental values ​​and path conflict values ​​of the target verification step according to its historical interaction weights to obtain the target interaction value of the target verification step.

[0230] For example, the job processing equipment can directly calculate the product of the environmental unexplainable value of the target verification step, the path conflict value of the target verification step, and the historical interaction weight of the target verification step, as the target interaction value.

[0231] 504. The verification risk value is calculated based on the unexplainable environmental value of the target verification step, the weighted path conflict value of the target verification step, and the target interaction value of the target verification step.

[0232] In one or more embodiments, after calculating the target interaction value of the target step to be verified, the job processing device can combine the environmental unexplainable value of the target step to be verified and the weighted path conflict value of the target step to be verified to calculate the verification risk value.

[0233] It should be noted that this verification risk value describes the reliability of the results caused by the step to be verified in the current target operation instance. The higher the verification risk value, the lower the reliability of the operation result under the combined effects of environmental disturbances and path conflicts.

[0234] As an example, the verification risk value of this application can satisfy the following expression, namely:

[0235]

[0236] Where V is the verification risk value of this target operation instance, used to generate the final verification result. A is the set of steps actually involved in path verification in the path verification trajectory set. i is the step number in set A; g i b is the unexplainable intensity value of the environment in the i-th step. i is the path conflict value at step i. α is the path conflict weight. γ is the historical interaction weight between the environmental unexplainability intensity value and the path conflict value. n is the number of steps in set A that actually participated in path verification. g i b i α and γ are expressed relatively, and V represents the relative risk. If the trajectory set to be verified is empty, the verification risk value is written as the low-order result, and the output result is reliable based on the attribution result.

[0237] For example, taking the "Export Daily Report" task as an example, if the path verification trajectory set contains two target steps to be verified: "Enter the report page" and "Export file", then set A contains two steps, and n is 2. For the "Enter the report page" step, its environment uninterpretability value gi is 0.059. This step verification found that the target object summary is correct and the relationship between the preceding and following steps is correct, but there is one out-of-bounds wait, and the path conflict value b. iThe value is set to 0.25. For the "Export File" step, the uninterpretable environment value gi is 0.4. Verification of this step revealed that it lacks a prerequisite "Result Verification" node, and the final result file was not generated through the completion node specified by the path baseline. The path conflict value is b. i Let's take 0.75. If the path conflict weight α in the current job template is 0.6 and the historical interaction weight γ is 0.8, then the risk contribution of the "Enter Report Page" step is 0.059 + 0.6×0.25 + 0.8×0.059×0.25, which is approximately 0.221; the risk contribution of the "Export File" step is 0.4 + 0.6×0.75 + 0.8×0.4×0.75, which is 1.09. Summing the two and dividing by 2, we get the verification risk value of this job, which is approximately 0.656. If the verification threshold for this job type in the current job template is 0.5, then the output job result is unreliable, and the main problem is located in the "Export File" step and its missing pre-verification node.

[0238] 505. Based on the verification risk value and the target path conflict type, determine the verification result of the target job instance.

[0239] In one or more embodiments, after calculating the verification risk value, the job processing device can determine the verification result of the target job instance based on the verification risk value and the target path conflict type.

[0240] The verification results of this application include the job instance identifier, job type, verification pass status, verification risk value, main abnormal steps, abnormal source type, and result credibility conclusion. Specifically, if the verification risk value is lower than the first verification threshold in the current job template, and the verification steps for each target in the path verification trajectory set only exhibit minor fluctuations explainable by the target environment baseline, the verification result written into the job result is credible. If the verification risk value is between the first and second verification thresholds, and path conflicts are concentrated near non-critical steps or legitimate branches, the verification result written into the job result needs to be reviewed, and the specific steps marked. If the verification risk value is higher than the second verification threshold, and path conflicts involve missing necessary nodes, illegal branch entry, target object substitution, or missing completion result nodes, the verification result written into the job result is unreliable. It should be understood that the second verification threshold is greater than the first verification threshold.

[0241] For example, in the aforementioned "Export Daily Report" job, the system states in the verification results that the "Export File" step has missing pre-verification nodes and incomplete completion node sources. Although the result file has been generated, the generation path does not meet the path baseline requirements. Therefore, the result file is marked as an untrusted job result.

[0242] By using the above method, after calculating and verifying the risk value, the verification result of the target job instance is determined by combining the target path conflict type. This not only achieves a quantitative evaluation of the job execution process, but also accurately locates the root cause of the anomaly.

[0243] In other alternative embodiments, such as Figure 6 As shown, after performing step 205, the method provided in this application may further include the following steps 206 to 208, namely:

[0244] 206. Based on the unexplainable environmental values ​​of the target verification step, determine the environmental attribution results of the target verification step.

[0245] In one or more embodiments, the job processing device may determine the environmental attribution result of the target verification step based on the environmental unexplainable value of the target verification step.

[0246] For example, regarding the attribution of retry status, the job processing device reads the permissible retry status in the target environment baseline corresponding to the current step, action type, and environmental pressure range, and compares it with the retry markers in the baseline-associated job trajectory set. If a page jump step shows in historical normal samples under the current environmental pressure range that a retry after one wait is permissible, and the retry marker for this job falls within the permissible status, the job processing device writes this retry behavior into the environmentally interpretable record. If the same step is retried multiple times, the target object summary is replaced after the retry, or the execution status shows that the step is not completed, the status conflict value is increased, and it is marked as pending path verification in the attribution results. For interface call type steps, the job processing device compares the interface response time summary with the boundary of the target environment baseline for interface call type, and simultaneously reads the interface call status and the target interface summary. When the interface response is within the permissible boundary and the call status is successful, the attribution is environmentally interpretable; when the interface response is within the permissible boundary but the target interface summary is replaced, it is written into the pending path verification trajectory set. For file generation steps, the job processing device compares the terminal resource pressure-related summary and the step connection time summary with the target environment baseline for file generation, and reads the file generation status. When the resource pressure is high and the connection delay falls within the allowable boundary, it is attributed to the environment being interpretable. When the file generation status fails or the generation step is skipped, it is written into the pending path verification trajectory set.

[0247] It should be noted that the path verification trajectory set in this application is obtained by screening the baseline associated operation trajectory set. It includes steps where the environmental unexplainable value is higher than the preset unexplainable threshold (i.e., the first verification step mentioned above), and / or steps where the execution status, retry status, and target object summary do not match the target environment baseline (i.e., the second verification step mentioned above).

[0248] It should be noted that the described path verification trajectory set can be understood as a set containing the target steps to be verified, specifically including steps where the environmental unexplainable value is greater than the preset unexplainable threshold, and steps where the state conflict value is not 0.

[0249] When generating environmental attribution results, the job processing equipment writes continuous boundary violations, execution status, retry status, and target object summary into the same step attribution record. Each step attribution record includes step number, action type, target object summary, target environment baseline index, actual behavior value, permissible behavior boundary, environmental pressure range, environmental unexplainable intensity value, execution status attribution, and retry status attribution. If a step has an insufficient sample flag or a missing key behavior value flag, this flag is also written into the step attribution record to explain the source of the environmental attribution basis for that step. Steps whose environmental unexplainable values ​​are within the current job template configuration threshold, and whose execution status, retry status, and target object summary all match the target environment baseline, are recorded as environmentally explainable steps. Steps whose environmental unexplainable values ​​exceed the preset unexplainable threshold, or whose execution status, retry status, and target object summary do not match the target environment baseline, are written into the pending path verification trajectory set. The attribution threshold is formed by the distribution of unexplained environmental intensity values ​​in historical normal samples of the same job type. In practice, the job processing equipment calculates the attribution intensity of historical normal samples according to the same steps and action types, sorts them, and selects the empirical quantile near the upper position as the attribution threshold for that step, so that different RPA job types have an attribution scale that matches their own fluctuation level.

[0250] 207. Based on the environmental attribution results and the target path conflict type, generate target remediation suggestions for the target verification steps.

[0251] In one or more embodiments, after determining the environmental attribution result, the job processing device can generate a target remediation suggestion for the target verification step based on the target path conflict type. Optionally, the target remediation suggestion can also be written into a remediation suggestion list as a rule-based disposal item, allowing maintenance personnel to execute it with one click or incorporate it into an automated remediation process.

[0252] For example, the remediation suggestions for missing nodes could point to the corresponding process node in the RPA script, prompting a check for missing skip conditions, abnormal branch configurations, or pre-wait conditions. The remediation suggestions for abnormal predecessor-successor relationships could point to jump conditions or page wait conditions between adjacent steps, prompting a check for step connection configurations. The remediation suggestions for target object summary substitutions would point to control positioning rules, page version changes, or interface address configurations, prompting a check of the target page, control selector, or interface name summary. The remediation suggestions for missing completion result nodes would point to the verification and export nodes before result generation, prompting a check if the result file was generated from the standard completion path. When the environment attribution results show a high unexplainable value but a low path conflict value, the remediation suggestions would point to the runtime environment and job scheduling policy, prompting a review of the job in the stable running window and a check of terminal resources or external system responses. Target remediation suggestions only reference the step number, action type, target object summary, and exception type, enabling operations personnel to locate the specific step.

[0253] Optionally, the job processing device can also output verification results and a list of remediation suggestions. The verification results consist of the job instance identifier, job type, verification pass status, verification risk value, anomalous steps, and result reliability conclusion, indicating whether the RPA job result is reliable. The remediation suggestion list consists of anomalous steps, path conflict type, attribution source, suggested inspection objects, and suggested remediation directions, guiding operations personnel to inspect the corresponding script nodes, page objects, interface configurations, or job scheduling environment.

[0254] 208. Based on the target modification suggestions, update the steps to be verified for the target.

[0255] In one or more embodiments, after generating a target modification suggestion, the job processing device can automatically trigger an update operation or perform a hot update after manual confirmation to complete the update processing of the target to be verified steps.

[0256] In this way, the environmental attribution results and the target verification steps can be further projected into the target path baseline, forming a final closed loop from environmental interpretation to path verification and then to result processing, so that the dual baseline comparison ultimately leads to the automated verification and engineering processing of RPA operation results.

[0257] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0258] Based on the same inventive concept, this application also provides a job processing apparatus for implementing the job processing method described above. The solution provided by this apparatus is similar to the solution described in the above method; therefore, the specific limitations in one or more job processing apparatus embodiments provided below can be found in the limitations of the job processing method described above, and will not be repeated here.

[0259] In one exemplary embodiment, such as Figure 7 As shown, a job processing apparatus is provided, including an acquisition unit 701, a calculation unit 702, a determination unit 703, and a verification processing unit 704. Optionally, it may also include a generation unit 705 and an update processing unit 706.

[0260] The acquisition unit 701 is used to acquire the job trajectory set, environmental observation dataset and target path baseline corresponding to the target job instance. The job trajectory set includes the job trajectory corresponding to the target job instance. The job trajectory includes at least one step. The environmental observation dataset includes environmental records when each step is executed.

[0261] The calculation unit 702 is used to calculate the environmental unexplainable value of the step based on the environmental record of the step, the state conflict value of the step, and the target environmental baseline. The environmental unexplainable value is used to describe the degree to which the behavior of the step deviates from the target environmental baseline under the environmental record. The target environmental baseline is determined from a set of candidate environmental baselines based on the environmental record of the step. The state conflict value of the step is used to describe the state difference of the step in the environmental record and the target environmental baseline.

[0262] The determining unit 703 is used to determine the target step to be verified from the at least one step based on the environmental unexplainable value of the step and the state conflict value of the step;

[0263] The verification processing unit 704 is used to perform verification calculations on the target step to be verified based on the target path baseline, and obtain the path conflict information and target path conflict type of the target step to be verified.

[0264] The determining unit 703 is used to determine the verification result of the target job instance based on the unexplainable environmental value of the target verification step, the path conflict information of the target verification step, and the target path conflict type. The verification result is used to describe the credibility of the result of the target job instance.

[0265] In another alternative embodiment, the computing unit 702 is specifically used for:

[0266] Based on the action type of each step, extract the running status summary and resource usage ratio of the corresponding step from the environment record of the corresponding step;

[0267] Based on the operational status summary of the step, the resource consumption ratio of the step, and the action type of the step, calculate the environmental pressure value of the step;

[0268] Calculate the actual behavior value of each step based on the action type of each step.

[0269] Based on the environmental pressure value of the step, the actual behavior value of the step, the state conflict value of the step, and the target environmental baseline, the environmental unexplainability value of the step is calculated.

[0270] In another alternative embodiment, the computing unit 702 is specifically used for:

[0271] The execution performance value of the step is obtained by normalizing the summary of the running status of the step.

[0272] The resource occupancy ratio of the above steps is normalized to obtain the auxiliary terminal pressure value of the above steps;

[0273] Based on the action type of the step, the action type of the preceding step adjacent to the step, the running status summary of the step, and the running status summary of the preceding step, calculate the step switching exposure value of the step;

[0274] The environmental pressure value of the step is obtained by weighted summation of the execution performance value of the step, the auxiliary terminal pressure value of the step, and the step switching exposure value of the step.

[0275] In one embodiment, the computing unit 702 is specifically used for:

[0276] The difference between the actual behavior value of the step and the first value in the target environment baseline is calculated to obtain the first difference, which is used to describe the upper boundary of the allowed behavior of the target environment baseline.

[0277] Calculate the difference between the second value in the target environment baseline and the actual behavior value of the step to obtain the second difference, which is used to describe the lower boundary of the permissible behavior of the target environment baseline;

[0278] Based on the action type of the step, set the weight of the state conflict item for the step;

[0279] Based on the weight of the state conflict item, the state conflict value of the step is weighted to obtain the weighted state conflict value.

[0280] The environmental unexplainable value of the step is calculated based on the environmental pressure value, the first difference, the second difference, and the weighted state conflict value.

[0281] In one embodiment, the determining unit 703 is specifically used for:

[0282] Based on the environmental unexplainable value of the steps, a first step to be verified is determined from the at least one step, wherein the environmental unexplainable value of the first step to be verified is greater than a preset unexplainable threshold.

[0283] Based on the state conflict value of the steps, a second step to be verified is determined from the at least one step, wherein the state conflict value of the second step to be verified is not zero;

[0284] The first step to be verified and the second step to be verified are taken as the target steps to be verified.

[0285] In one embodiment, the verification calculation unit 704 is specifically used for:

[0286] Based on the step number of the target to be verified step, the target path node corresponding to the step number is determined from the target path baseline;

[0287] Based on the preset node existence rules, the node existence is verified for the target to be verified steps to obtain the node existence conflict value.

[0288] Based on the node association relationship of the target path node, the relationship verification is performed on the target verification step and the associated steps of the target verification step to obtain the relationship conflict value. The node association relationship is used to describe the node connection relationship between the target path node and the associated nodes of the target path node.

[0289] Based on the target object summary corresponding to the target path node, the running status summary of the target to be verified step is compared to obtain the summary verification conflict value.

[0290] Based on the preset result verification rules corresponding to the target path baseline, the running results of the target steps to be verified are verified to obtain the completion result conflict value.

[0291] Based on the node existence conflict value, the relationship conflict value, the summary verification conflict value, and the completion result conflict value, the path conflict information and target path conflict type of the target verification step are determined.

[0292] In one embodiment, the path conflict information includes path conflict values; the determining unit 703 is specifically used for:

[0293] Based on the criticality of the target verification step, a weight is set for the target verification step to obtain the path conflict weight of the target verification step;

[0294] Based on the path conflict weight of the target verification step, the path conflict value of the target verification step is weighted to obtain the weighted path conflict value of the target verification step.

[0295] Based on the historical interaction weights of the target verification step, the environmental unexplainable value and the path conflict value of the target verification step are weighted to obtain the target interaction value of the target verification step. The historical interaction weights are used to describe the degree of possibility that the environmental unexplainable deviation indicated by the environmental unexplainable value and the path conflict indicated by the path conflict value may occur simultaneously.

[0296] The verification risk value is calculated based on the unexplainable environmental value of the target verification step, the weighted path conflict value of the target verification step, and the target interaction value of the target verification step.

[0297] Based on the verification risk value and the target path conflict type, the verification result of the target job instance is determined.

[0298] In one embodiment, the determining unit 703 is further configured to determine the environmental attribution result of the target job instance based on the environmental unexplainable value of the target to be verified step, the path conflict information of the target to be verified step, and the target path conflict type after determining the verification result of the target job instance.

[0299] The generation unit 704 is specifically used to generate target repair suggestions for the target to be verified steps based on the environmental attribution results and the target path conflict type.

[0300] The update processing unit 705 is specifically used to update the target verification step based on the target modification suggestion.

[0301] The above Figure 7 The various modules of the task processing device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0302] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 8As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a job processing method. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0303] Those skilled in the art will understand that Figure 8 The structure shown is only a block diagram of a part of the structure related to the present application and does not constitute a limitation on the computer device on which the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0304] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0305] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0306] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the processor to perform the steps in the above method embodiments.

[0307] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0308] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0309] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0310] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for processing jobs, characterized in that, The method includes: Obtain the job trajectory set, environmental observation dataset, and target path baseline corresponding to the target job instance. The job trajectory set includes the job trajectory corresponding to the target job instance. The job trajectory includes at least one step. The environmental observation dataset includes environmental records when each step is executed. Based on the environmental record of the step, the state conflict value of the step, and the target environmental baseline, the environmental unexplainable value of the step is calculated. The environmental unexplainable value is used to describe the degree to which the behavior of the step deviates from the target environmental baseline under the environmental record. The target environmental baseline is determined from a set of candidate environmental baselines based on the environmental record of the step. The state conflict value of the step is used to describe the state difference of the step in the environmental record and the target environmental baseline. Based on the environmental unexplainable value and the state conflict value of the step, the target step to be verified is determined from the at least one step; Based on the target path baseline, the target steps to be verified are verified and calculated to obtain the path conflict information and target path conflict type of the target steps to be verified. Based on the unexplainable environmental values ​​of the target verification step, the path conflict information of the target verification step, and the target path conflict type, the verification result of the target job instance is determined, and the verification result is used to describe the credibility of the result of the target job instance.

2. The method according to claim 1, characterized in that, Based on the environmental records of the steps, the state conflict values ​​of the steps, and the target environmental baseline, calculate the environmental unexplainable values ​​of the steps, including: Based on the action type of each step, extract the running status summary and resource usage ratio of the corresponding step from the environment record of the corresponding step; Based on the operational status summary of the step, the resource consumption ratio of the step, and the action type of the step, calculate the environmental pressure value of the step; Calculate the actual behavior value of each step based on the action type of each step. Based on the environmental pressure value of the step, the actual behavior value of the step, the state conflict value of the step, and the target environmental baseline, the environmental unexplainability value of the step is calculated.

3. The method according to claim 2, characterized in that, Based on the operational status summary of the step, the resource consumption ratio of the step, and the action type of the step, the environmental pressure value of the step is calculated, including: The execution performance value of the step is obtained by normalizing the summary of the running status of the step. The resource occupancy ratio of the above steps is normalized to obtain the auxiliary terminal pressure value of the above steps; Based on the action type of the step, the action type of the preceding step adjacent to the step, the running status summary of the step, and the running status summary of the preceding step, calculate the step switching exposure value of the step; The environmental pressure value of the step is obtained by weighted summation of the execution performance value of the step, the auxiliary terminal pressure value of the step, and the step switching exposure value of the step.

4. The method according to any one of claims 2 to 3, characterized in that, Based on the environmental pressure value of the step, the actual behavior value of the step, the state conflict value of the step, and the target environmental baseline, the environmental unexplainable value of the step is calculated, including: The difference between the actual behavior value of the step and the first value in the target environment baseline is calculated to obtain the first difference, which is used to describe the upper boundary of the allowed behavior of the target environment baseline. Calculate the difference between the second value in the target environment baseline and the actual behavior value of the step to obtain the second difference, which is used to describe the lower boundary of the permissible behavior of the target environment baseline; Based on the action type of the step, set the weight of the state conflict item for the step; Based on the weight of the state conflict item, the state conflict value of the step is weighted to obtain the weighted state conflict value. The environmental unexplainable value of the step is calculated based on the environmental pressure value, the first difference, the second difference, and the weighted state conflict value.

5. The method according to any one of claims 1 to 4, characterized in that, Based on the unexplainable environmental values ​​and the conflicting state values ​​of the steps, the target step to be verified is determined from the at least one step, including: Based on the environmental unexplainable value of the steps, a first step to be verified is determined from the at least one step, wherein the environmental unexplainable value of the first step to be verified is greater than a preset unexplainable threshold. Based on the state conflict value of the steps, a second step to be verified is determined from the at least one step, wherein the state conflict value of the second step to be verified is not zero; The first step to be verified and the second step to be verified are taken as the target steps to be verified.

6. The method according to any one of claims 1 to 5, characterized in that, Based on the target path baseline, verification calculations are performed on the target steps to be verified to obtain path conflict information and target path conflict types for the target steps to be verified, including: Based on the step number of the target to be verified step, the target path node corresponding to the step number is determined from the target path baseline; Based on the preset node existence rules, the node existence is verified for the target to be verified steps to obtain the node existence conflict value. Based on the node association relationship of the target path node, the relationship verification is performed on the target verification step and the associated steps of the target verification step to obtain the relationship conflict value. The node association relationship is used to describe the node connection relationship between the target path node and the associated nodes of the target path node. Based on the target object summary corresponding to the target path node, the running status summary of the target to be verified step is compared to obtain the summary verification conflict value. Based on the preset result verification rules corresponding to the target path baseline, the running results of the target steps to be verified are verified to obtain the completion result conflict value. Based on the node existence conflict value, the relationship conflict value, the summary verification conflict value, and the completion result conflict value, the path conflict information and target path conflict type of the target verification step are determined.

7. The method according to any one of claims 1 to 6, characterized in that, The path conflict information includes path conflict values; based on the unexplainable environmental values ​​of the target verification step, the path conflict information of the target verification step, and the target path conflict type, the verification result of the target job instance is determined, including: Based on the criticality of the target verification step, a weight is set for the target verification step to obtain the path conflict weight of the target verification step; Based on the path conflict weight of the target verification step, the path conflict value of the target verification step is weighted to obtain the weighted path conflict value of the target verification step. Based on the historical interaction weights of the target verification step, the environmental unexplainable value and the path conflict value of the target verification step are weighted to obtain the target interaction value of the target verification step. The historical interaction weights are used to describe the degree of possibility that the environmental unexplainable deviation indicated by the environmental unexplainable value and the path conflict indicated by the path conflict value may occur simultaneously. The verification risk value is calculated based on the unexplainable environmental value of the target verification step, the weighted path conflict value of the target verification step, and the target interaction value of the target verification step. Based on the verification risk value and the target path conflict type, the verification result of the target job instance is determined.

8. A job processing device, characterized in that, The job processing device includes: The acquisition unit is used to acquire the job trajectory set, environmental observation dataset and target path baseline corresponding to the target job instance. The job trajectory set includes the job trajectory corresponding to the target job instance. The job trajectory includes at least one step. The environmental observation dataset includes environmental records when each step is executed. A calculation unit is configured to calculate an environmental unexplainable value for a step based on the environmental record of the step, the state conflict value of the step, and a target environmental baseline. The environmental unexplainable value describes the degree to which the behavior of the step deviates from the target environmental baseline under the environmental record. The target environmental baseline is determined from a set of candidate environmental baselines based on the environmental record of the step. The state conflict value describes the state difference of the step between the environmental record and the target environmental baseline. A determining unit is configured to determine a target step to be verified from the at least one step based on an environmental unexplainable value and a state conflict value of the step. The verification processing unit is used to perform verification calculations on the target steps to be verified based on the target path baseline, and obtain the path conflict information and target path conflict type of the target steps to be verified. The determining unit is used to determine the verification result of the target job instance based on the unexplainable environmental value of the target verification step, the path conflict information of the target verification step, and the target path conflict type. The verification result is used to describe the credibility of the result of the target job instance.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images