Business processing method and device, medium and product

Abstract

The invention relates to the technical field of agent platform architecture, in particular to a business processing method and device, a medium and a product. The method comprises the following steps: linking a pre-established agent platform with an enterprise system of a target enterprise, and constructing a template task tree based on historical system operation data of the enterprise system; receiving a user demand, disassembling the user demand into a task sequence based on the template task tree, and determining a task type of each task in the task sequence; based on the capability metadata of the agent candidate pool and the task type of the task sequence, selecting agents from the agent candidate pool to form an agent list, and allocating the task sequence to the agent list for execution; and determining a plurality of intervention nodes, determining a task execution deviation degree at each intervention node, and when the task execution deviation degree of any intervention node exceeds a preset intervention threshold, adjusting an execution strategy of a subsequent task of the intervention node in the task sequence. According to the invention, accurate generation and adaptive execution of enterprise tasks can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of agent platform architecture, in particular to a business processing method, device, medium and product. BACKGROUND

[0002] In the process of enterprise digital transformation, more and more enterprises introduce AI agent platforms to assist in business processing to improve business efficiency and automation level, but the current related technology still has many shortcomings in multiple dimensions, which is difficult to meet the needs of enterprise fine business processing. On the one hand, the AI agent platform and the enterprise system are often in a relatively independent state, and the two lack a deep collaboration mechanism. Even if some solutions realize coupling through a simple interface, they do not realize systematic integration and value mining of the enterprise system, and cannot form a standardized task decomposition framework that fits the actual business logic of the enterprise, so that the system cannot convert abstract requirements into structured and executable specific tasks when the user publishes tasks through the enterprise system. On the other hand, in the task allocation link, the existing technology generally adopts a single-dimensional screening logic of random selection or manual selection based on experience, which does not take into account the ability information of different agents, nor the individualized task needs of users, resulting in insufficient adaptability of agents and tasks, and ultimately causing the task execution result to not meet the user's expectations, increasing the business rework cost, and seriously affecting the efficiency and quality of enterprise business processing. SUMMARY

[0003] In order to solve the problem that the AI agent platform of the prior art cannot realize accurate generation and adaptive execution of enterprise tasks, the present application provides a business processing method, device, medium and product.

[0004] In a first aspect, the present application provides a business processing method, which adopts the following technical solution: A business processing method, comprising: When receiving an access request of a target enterprise, link a pre-built agent platform and an enterprise system of the target enterprise, obtain historical system operation data of the enterprise system, and construct a template task tree based on the historical system operation data; Receive a user demand uploaded by a target personnel of the target enterprise through the enterprise system, decompose the user demand into a task sequence based on the template task tree, and determine the task type of each task in the task sequence; Obtain the ability metadata of an agent candidate pool of the agent platform, select agents from the agent candidate pool to form an agent list based on the ability metadata and the task type of the task sequence, and assign the task sequence to the agent list for execution; A number of intervention nodes are determined according to the depth of the task sequence, a task execution deviation degree is determined at each intervention node, and when the task execution deviation degree of any intervention node exceeds a preset intervention threshold, an execution strategy of a subsequent task of the intervention node in the task sequence is adjusted.

[0005] By adopting the technical solution, the intelligent agent platform is linked with the enterprise system, a template task tree is constructed by historical data, standardization and structuring of a business process are realized, user requirements are decomposed into a task sequence and a task type is determined, an intelligent agent is selected based on capability metadata to ensure adaptability and efficiency of task execution, an intervention node is set and an execution strategy is dynamically adjusted to correct task deviation in time and ensure accuracy and stability of business processing, and precise generation and adaptive execution of enterprise tasks are realized.

[0006] In a preferred example, the application can be further configured to: constructing the template task tree based on the historical system operation data, comprising: cutting the historical system operation data based on a session timeout rule to obtain business activity data, and performing clustering analysis on the business activity data to obtain a plurality of clustering clusters; For each clustering cluster in the plurality of clustering clusters, performing sequence pattern mining on cluster members of the clustering cluster to obtain a support degree of each activity mode, selecting a target activity mode from the cluster members based on the support degree, and identifying a dependency relationship and a mapping intensity coefficient between the target activity modes; constructing an initial template task tree by taking each target activity mode obtained from the plurality of clustering clusters as a node based on the dependency relationship; For each node in the initial template task tree, determining an intra-cluster frequency of the node in the clustering cluster and a global frequency of the node in the plurality of clustering clusters, and determining an execution tendency weight between the node and a parent node of the node based on the intra-cluster frequency and the global frequency; annotating the execution tendency weight and the mapping intensity coefficient between each two nodes obtained in the corresponding positions of the initial template task tree to obtain the template task tree.

[0007] By adopting the technical solution, session cutting and clustering analysis are used to realize effective classification of business activities and accurately distinguish different business scenarios, sequence patterns are mined and target activity modes are selected to extract core business steps, dependency relationships and mapping intensity coefficients are identified to clearly determine the close degree of association between business steps, execution tendency weights are calculated to quantify the execution preference between steps, and finally the template task tree constructed has both business logic and quantitative guidance significance, thereby providing precise process support for subsequent requirement decomposition and task execution.

[0008] The application can be further configured in a preferred example as follows: the method further comprises: creating a test environment, generating a plurality of test cases containing different types of test tasks and network environment parameter combinations; For each agent in the agent candidate pool, each test case is executed to obtain the comprehensive test results of the agent for each task type under each network environment, and the comprehensive test results include task performance indicators and environment sensitivity indicators; Based on the comprehensive test results of each agent and the preset scoring rules, a performance score matrix is constructed for each agent, the rows of the performance score matrix represent network environments, the columns represent task types, and the elements represent indicator score vectors, which include indicator scores of each task performance indicator and each environment sensitivity indicator; The comprehensive test results and performance score matrix of each agent in the agent candidate pool are stored as the capability metadata.

[0009] By adopting the above technical solution, multi-dimensional test cases are designed to comprehensively cover different task types and network environments, ensuring the integrity of the performance evaluation of the agent. Through the comprehensive test results and the performance score matrix, the capability of the agent is quantified and structured, and the performance of the agent in different scenarios is intuitively presented. The capability metadata is stored, providing objective and accurate data basis for subsequent selection of agents, and avoiding subjective selection bias.

[0010] The application can be further configured in a preferred example as follows: selecting an agent from the agent candidate pool based on the capability metadata and the task type of the task sequence to form an agent list, comprising: Obtaining the task performance indicator lower limit of each task type under the current network environment; For each task in the task sequence, based on the capability metadata, filtering an agent group from the agent candidate pool that can meet all task performance indicator lower limits of the task type of the task under the network environment matching the current network environment; Based on the dependency constraint relationship and the indicator score of the environment sensitivity indicator of each agent, an agent is selected from the agent group of each task to form the agent list, and the dependency constraint relationship includes that the agents selected by any two tasks without a dependency relationship are not the same.

[0011] By adopting the above technical solution, the agent is filtered in combination with the current network environment to ensure the adaptability of the agent in the actual running environment and avoid performance loss caused by environmental inadaptability. The performance indicator lower limit is set to ensure that the selected agent has basic execution capability. The dependency constraint is considered to avoid conflicts between agents of non-dependent tasks, and a more stable agent is selected in combination with the environment sensitivity indicator score.

[0012] The application can be further configured in a preferred example to determine a current task execution deviation degree of a current intervention node, including: determining a task type of each preceding task before the current intervention node in the task sequence, determining a task type composition of the executed task sequence based on the task type of each preceding task; determining a dynamic weight of each deviation dimension based on the task type composition; obtaining a current task execution result of the current intervention node, and calculating a first task execution deviation degree of the current task execution result on each deviation dimension based on the current task execution result and the dynamic weight of each deviation dimension; determining a preceding task and a subsequent task of the current intervention node in the task sequence, updating the first task execution deviation degree based on a mapping intensity coefficient between the preceding task and the subsequent task in the template task tree to obtain a second task execution deviation degree; judging whether the current intervention node is a first intervention node in the plurality of intervention nodes; if the current intervention node is the first intervention node, taking the second task execution deviation degree as the current task execution deviation degree; if the current intervention node is not the first intervention node, calculating a time sequence cumulative deviation based on the task execution deviation degrees of all intervention nodes between the current intervention nodes, and updating the second task execution deviation degree based on the time sequence cumulative deviation to obtain the current task execution deviation degree.

[0013] By adopting the above technical solution, the dynamic weight is determined according to the preceding task type composition, so that the deviation evaluation is more suitable for the business scenario and avoids the limitation of a single weight; the influence of the association between tasks on the deviation is considered in combination with the mapping intensity coefficient, so that the comprehensiveness of the deviation evaluation is improved; the time sequence cumulative deviation is introduced, the superposition effect of the historical deviation is considered, and the finally calculated task execution deviation degree is more accurate.

[0014] The application can be further configured in a preferred example to adjust the execution strategy of the subsequent task of the intervention node in the task sequence, including: starting from the first task of the task sequence, sequentially obtaining an intermediate task execution result after each task is executed, and calculating an intermediate task execution deviation degree of the intermediate task execution result on each deviation dimension; comparing the intermediate task execution deviation degree obtained after each task is executed with a preset alarm threshold, and taking the task corresponding to the intermediate task execution deviation degree exceeding the preset alarm threshold as a target task; adjust the agent corresponding to the target task in the agent list, and execute the target task and the subsequent task of the target task in the task sequence according to the adjusted agent list.

[0015] By adopting the technical solution, the intermediate result of the task and the deviation degree are monitored in real time, the dynamic tracking of the task execution process is realized, the target task exceeding the threshold can be quickly located, the agent corresponding to the target task is adjusted in time, the task deviation problem is solved in a targeted manner, the deviation is avoided from being conducted to the subsequent task, the target task and the subsequent task are re-executed, the business processing is ensured to return to the correct track, the influence of the deviation on the overall business result is effectively reduced, and the business processing quality is ensured.

[0016] In a preferred example, the application can be further configured as follows: the method further comprises: when the task sequence is executed, determining the lowest level node of each task in the template task tree; determining the child nodes of the lowest level node in the template task tree, and arranging each child node in descending order of the execution tendency weight between the child node and the lowest level node to obtain a recommended task sequence; feeding back the recommended task sequence to the target personnel, so as to execute the corresponding task according to the selection of the target personnel for the recommended task sequence.

[0017] By adopting the technical solution, the recommended task sequence is generated based on the template task tree and the execution tendency weight, the recommended task sequence is consistent with the historical business execution habit of the enterprise, the rationality and practicability of the recommendation are improved, the recommended sequence is fed back to the target personnel, clear guidance is provided for subsequent business processing, and the personnel decision cost is reduced; the personnel can select independently, and the flexibility of intelligent recommendation and manual intervention is considered.

[0018] In a second aspect, the application provides an electronic device, which adopts the following technical solution: one or more processors; a memory; at least one application program, wherein the at least one application program is stored in the memory and is configured to be executed by the at least one processor, and the at least one application program is configured to execute the business processing method according to any one of the first aspect.

[0019] In a third aspect, the application provides a computer readable storage medium, which adopts the following technical solution: a computer readable storage medium, which stores a computer program, and when the computer program is executed in a computer, the computer is caused to execute the business processing method according to any one of the first aspect.

[0020] In a fourth aspect, the present application provides a computer program product, which adopts the technical scheme as follows: A computer program product, comprising a computer program, which, when executed by a processor, implements the business processing method according to any one of the first aspect.

[0021] In summary, the present application has the following beneficial technical effects: The present application links the intelligent agent platform with the enterprise system, constructs a template task tree through historical data, realizes the standardization and structuring of the business process, disassembles the user demand into a task sequence and determines the task type, selects the intelligent agent based on the capability metadata, ensures the adaptability and efficiency of the task execution, sets the intervention node and dynamically adjusts the execution strategy, can timely correct the task deviation, guarantees the accuracy and stability of the business processing, and realizes the accurate generation and adaptive execution of the enterprise task. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 is a flowchart of a business processing method provided by an embodiment of the present application; Figure 2 is a structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0023] The following will be described in detail with reference to the accompanying drawings. Figure 1 to the accompanying drawings Figure 2 The present application will be further described in detail.

[0024] The present embodiment is merely an explanation of the present application, and is not a limitation of the present application. Those skilled in the art can make modifications to the present embodiment without creative contribution, as long as the modifications are within the scope of the claims of the present application.

[0025] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme of the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings of the embodiments of the present application. Obviously, the described embodiments are some of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative contribution are within the scope of protection of the present application.

[0026] In addition, the term "and / or" in this paper is merely a description of the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which means that there are three cases of A alone, A and B together, and B alone. In addition, the character " / " in this paper generally represents an "or" relationship between the associated objects, unless otherwise specified.

[0027] It should be noted that, in the optional embodiments of this application, the data related to object information, when applied to specific products or technologies, requires the permission or consent of the object. Furthermore, the collection, use, and processing of this data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. In other words, if the embodiments of this application involve data related to an object, it must be obtained with the object's authorization and consent, the authorization and consent of relevant departments, and in accordance with the relevant laws, regulations, and standards of the country and region. If the embodiments involve personal information, the acquisition of all personal information requires the individual's consent. If sensitive information is involved, the separate consent of the information subject is required. The embodiments also need to be implemented with the object's authorization and consent.

[0028] This application provides a business processing method, such as... Figure 1 As shown, the method provided in this application embodiment is executed by an electronic device, which can be a server or a terminal device. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. The terminal device can be a smartphone, tablet, laptop, desktop computer, etc., but is not limited to these. The terminal device and the server can be directly or indirectly connected via wired or wireless communication. This application embodiment does not impose any limitations on this connection. The method includes steps S101-S104, wherein: S101. When an access request from a target enterprise is received, the pre-built intelligent agent platform is linked with the target enterprise's enterprise system to obtain historical system operation data of the enterprise system, and a template task tree is constructed based on the historical system operation data.

[0029] Specifically, the intelligent agent platform is a pre-built management system integrating multiple intelligent agents with specific functions, enabling agent registration, scheduling, and monitoring. The specific types of intelligent agents can be set according to actual needs; optional intelligent agents may include data extraction agents, text analysis agents, image recognition agents, and decision recommendation agents. Standardized interfaces are developed to support data interaction between the intelligent agent platform and enterprise systems, configure data synchronization rules, and obtain historical system operation data from the enterprise system, such as business documents, approval records, and operation logs from the past three years.

[0030] The obtained historical system operation data is sequentially subjected to business activity cutting, clustering analysis and activity pattern mining to obtain a target activity pattern. The clustering analysis obtains a plurality of clustering clusters, and each clustering cluster represents a business scenario. The target activity pattern represents an effective high-frequency activity pattern, and a target activity pattern represents a single task, such as creating an order or auditing an order as a task. For each clustering cluster, the dependency relationship between the target activity patterns in the clustering cluster is identified, and a subtask tree of the clustering cluster is constructed based on the obtained dependency relationship, and the nodes in the subtask tree represent the target activity patterns. After obtaining the subtask tree of each clustering cluster, the subtask trees of all clustering clusters are integrated to obtain a template task tree; the integration manner can be to connect the root nodes of each subtask tree to the same node, and the same node connected by each root node serves as the root node of the template task tree.

[0031] In S102, a user demand uploaded by a target personnel of a target enterprise through an enterprise system is received, the user demand is disassembled into a task sequence based on the template task tree, and the task type of each task in the task sequence is determined.

[0032] Specifically, the target personnel is any staff in the target enterprise. A development demand submission interface is developed in the enterprise system, and text input, file upload or voice-to-text and the like are supported to submit the user demand. A similarity algorithm is used to calculate the similarity between the user demand and each subtask tree, a subtask tree with the highest similarity is selected as a business scenario matched with the user demand, and the subtask tree matched with the user demand successfully is recorded as a target subtask tree. Optionally, the similarity algorithm can be a Word2Vec or BERT model.

[0033] In the process of constructing the template task tree, a structured feature library is constructed for all nodes in the template task tree. For each node, the node features include: node attributes (node ID, parent node ID, child node ID), function description (business target of the node), input requirement (pre-requisite data required before the node is executed), and output result (data generated after the node is executed). Taking a subtask tree corresponding to a business scenario of generating a financial statement as an example, the task flow of each node in the subtask tree is: collecting financial data, data verification, data analysis, generating a report, and report auditing. Taking the data verification node as an example, the business target is to verify the integrity and format correctness of the data, the input requirement is to obtain the financial data, and the output result is the structured data after the verification.

[0034] The user demand is feature-structured, the user target and the precondition are extracted from the NPL and converted into a vector form, and an example is as follows: the user demand is to directly generate a financial statement according to the current input data, the core target is to generate a financial statement, and the precondition uses the current input data. When the precondition does not exist in the user demand, the root node of the target sub-task tree is directly taken as the mapping node; when the user demand has a precondition, the precondition and each node in the target sub-task tree are matched in terms of vector similarity, the matched node is taken as the mapping node, the matching manner is to compare the similarity of the historical input data of each node in the precondition and the target sub-task tree, and the node with the highest similarity is taken as the mapping node, and the precondition is taken as the constraint condition of the mapping node. The user target and the output content of each node are matched, and the node with the highest similarity is taken as the terminal node. After the mapping node and the terminal node are determined, the path between the mapping node and the terminal node in the target sub-task tree is extracted as the task sequence corresponding to the user demand.

[0035] For the task type, a classification model can be pre-trained, the task sequence is taken as a training set, and the task type of each task in the task sequence is output by the classification model. For example, the task type corresponding to the acquisition of monthly financial data is data collection, the task type corresponding to the verification of financial data is data verification, the task type corresponding to the calculation of profit is data analysis, the task type corresponding to the generation of a report is document generation, and the task type corresponding to the audit of a report is process approval.

[0036] In S103, the capability metadata of the agent candidate pool of the agent platform is acquired, an agent is selected from the agent candidate pool based on the capability metadata and the task type of the task sequence to form an agent list, and the task sequence is assigned to the agent list for execution.

[0037] Specifically, a test environment is created, and test cases containing different types of test tasks and network environment parameter combinations are generated. Each agent in the agent candidate pool is executed with all test cases, and the comprehensive test results of each agent executing each test case are recorded. The comprehensive test results include task performance indicators and environment sensitivity indicators. The task performance indicators represent the pros and cons of the task execution results, including accuracy, processing time, and success rate. The environment sensitivity indicators represent the influence on the network environment parameters, including stability coefficient, attenuation rate, and environment adaptation degree.

[0038] The preset scoring rule includes a weight proportion of each index and a scoring calculation method. The index score vector of each agent in executing each test case is calculated based on the preset scoring rule. For each agent, the index score vector obtained by executing all test cases forms a performance score matrix. The row of the performance score matrix represents a network environment. Different network environment parameter combinations correspond to different network environments. The list is a task type, and the element represents the index score vector obtained by executing the test case of the corresponding network environment and task type. The comprehensive test result of each agent in the agent candidate pool and the performance score matrix are stored as capability metadata.

[0039] Based on the capability metadata and the task type, a corresponding agent is selected for each task in the task sequence. The selected agents are arranged in the order of the corresponding tasks in the task sequence to obtain an agent list. Each agent in the agent list is assigned a corresponding task. The scheduling engine of the agent platform sends execution instructions to the agents in the order of the task sequence, and monitors the execution state in real time.

[0040] S104, according to the depth of the task sequence, a plurality of intervention nodes are determined, and the degree of task execution deviation is determined at each intervention node. When the degree of task execution deviation of any intervention node exceeds a preset intervention threshold, the execution strategy of the subsequent task of the intervention node in the task sequence is adjusted.

[0041] Specifically, the depth of the task sequence is determined, and the depth represents the total number of tasks a in the task sequence. Based on the depth, the intervention nodes are set according to the equal interval principle. The equal interval principle means that an intervention node is determined every interval step. The interval step can be pre-set and stored in the electronic device. For example, the interval step of the intervention node is 3, and when a = 10, intervention nodes are set after the 3rd, 6th and 9th tasks.

[0042] When the degree of task execution deviation of any intervention node exceeds a preset intervention threshold, the target task with the largest deviation is determined from the tasks that have been executed before the intervention node, the agent used for the target task is adjusted, and the target task and the subsequent tasks of the target task in the task sequence are executed according to the adjusted agent list.

[0043] The embodiment links the agent platform and the enterprise system, constructs a template task tree based on historical data, realizes the standardization and structuring of business processes, disassembles user requirements into task sequences and determines task types, selects agents based on capability metadata to ensure the adaptability and efficiency of task execution, sets intervention nodes and dynamically adjusts the execution strategy to correct task deviations in time and ensure the accuracy and stability of business processing, and realizes the accurate generation and adaptive execution of enterprise tasks.

[0044] In a possible implementation of the embodiment, a template task tree is constructed based on historical system operation data, including: The historical system operation data is cut based on a session timeout rule to obtain business activity data, and clustering analysis is performed on the business activity data to obtain a plurality of clustering clusters; For each clustering cluster in the plurality of clustering clusters, sequence pattern mining is performed on cluster members of the clustering cluster to obtain a support degree of each activity mode, a target activity mode is selected from the cluster members based on the support degree, and a dependency relationship and a mapping strength coefficient between the target activity modes are identified; Each target activity mode obtained from the plurality of clustering clusters is taken as a node to construct an initial template task tree based on the dependency relationship; For each node in the initial template task tree, a cluster-in frequency of the node in the clustering cluster and a global frequency of the node in the plurality of clustering clusters are determined, and an execution tendency weight between the node and a parent node of the node is determined based on the cluster-in frequency and the global frequency; The execution tendency weight and the mapping strength coefficient between each two nodes obtained are marked at corresponding positions of the initial template task tree to obtain a template task tree.

[0045] In the embodiment, the historical system operation data is all operation behavior logs recorded by a system when enterprise employees use various business systems, each record containing user ID, operation time, operation module, operation type, input data, output result and the like. The session timeout rule is a data cutting strategy. If a time interval between two continuous operations exceeds a preset time threshold (such as 30 minutes), the two operations are considered to belong to different business sessions or business activities. The historical system operation data is grouped by user ID, and the operations of each user ID are sorted according to time stamp to obtain an operation series. The time difference between adjacent two operations is calculated, and if the time difference exceeds the time threshold set by the session timeout rule, the operation series is cut at the time difference. Finally, the operation series of the user is cut into a plurality of business activity data.

[0046] Further, features are extracted from the business activity data, and the feature types can include: operation sequence encoding (transforming operation names into digital sequences through encoding), module distribution (statistically counting the frequency of each module being accessed in each business activity data to represent in vector form), and key operation flag (whether a key operation such as payment, issuance, etc. is triggered). The features extracted from the business activity data are input into a clustering algorithm, and the type of the clustering algorithm is not limited in the present embodiment. Optionally, if the number of core business processes of the target enterprise can be obtained, K-Means can be used, and the number of clusters K needs to be specified in advance; if the business types are not clear or the number of business processes is not fixed, DBSCAN algorithm can be selected, and the number of clusters K does not need to be specified, and the algorithm can automatically discover clusters of arbitrary shape. The clustering algorithm automatically divides the business activity data into different cluster clusters according to the feature similarity. For example, all business activities related to reimbursement are divided into cluster 1, and all business activities related to employee onboarding are divided into cluster 2.

[0047] Further, a subtask tree is constructed for each cluster cluster. Taking any cluster cluster as an example, the process of constructing the subtask tree of the cluster cluster includes: a minimum support threshold is preset, and a sequence pattern mining algorithm can automatically find all activity patterns with a frequency (support) greater than or equal to the minimum support threshold. In the mined activity patterns, the patterns with a support lower than the minimum support threshold are discarded, and the remaining patterns are taken as target activity patterns. A target activity pattern can contain one or more steps, and the steps of all target activity patterns obtained by traversing the cluster cluster are identified to identify target steps that do not have a previous step, the previous step being a step adjacent to and before the target step. If the target step is one, it is taken as the root node of the subtask tree of the cluster cluster, and if the target step is multiple, each step is taken as a subnode of the root node.

[0048] For each target step, its subsequent steps are determined from the target activity pattern, the subsequent steps being steps adjacent to and after the target step, and the frequency of each subsequent step in all subsequent steps of the target step is determined, and the subsequent steps exceeding an effective frequency threshold are selected as steps having a dependency relationship with the target step, and each subsequent step having a dependency relationship is set as a node as a subnode of the node where the target step is located.

[0049] For each sub-node of the target step, refer to the above embodiment to determine the postorder steps of the sub-node from the target activity pattern, and determine the frequency of each postorder step in all postorder steps of the sub-node, from which the postorder steps exceeding the valid frequency threshold are screened as the steps having a dependency relationship with the sub-node, and each postorder step having a dependency relationship is set as a node as a sub-node of the sub-node (i.e. a grandchild node of the target step). In this way, all steps in the target activity pattern are traversed until the sub-task tree of the cluster is obtained. Finally, the sub-task trees of the various cluster are integrated to obtain the initial template task tree.

[0050] For each pair of parent node A and child node B having a parent-child relationship in the initial template task tree, the output field of the parent node A and the input field of the child node B are extracted from the node feature library, and the field coincidence ratio of the output field and the input field, i.e. the ratio of the coincident fields to the number of input fields of the child node B, is calculated. Then, from the business activity data cut from the historical system operation data, the business activity samples directly executing the child node B after the parent node A is executed are screened. For each sample, the output data of the parent node A and the execution result of the child node B are recorded, and the output disturbance data of the parent node A is randomly generated. The output disturbance data is superimposed on the output data of the parent node A, and the child node B is re-executed to obtain the disturbed execution result. The output result change amplitude = |the previous output result of the child node B - the latter output result of the child node B| / the previous output result of the child node B. The input data disturbance change amplitude = the output disturbance data / the previous output result of the parent node A; the single-sample sensitivity = the output result change amplitude of the child node B / the input data disturbance change amplitude. The average of the single-sample sensitivity of all samples is calculated to obtain the correlation sensitivity. The mapping strength coefficient of the parent node A and the child node B is obtained by weighted summation of the field coincidence ratio and the correlation sensitivity based on a predefined weight. Refer to the above steps to determine the mapping strength coefficient between each pair of parent node and child node in the initial template task tree.

[0051] For each node in the initial template task tree, the total number of times that the step represented by the node appears in the original cluster to which it belongs is determined as the intra-cluster frequency, and the total number of times that the step represented by the node appears in all clusters is determined as the global frequency. The result of weighted summation of the intra-cluster frequency and the global frequency based on a predefined weight is taken as the execution tendency weight, which represents the degree of preference of the path from the parent step to the child step and is used for task recommendation. The execution tendency weight of each node is marked on the edge between the node and its parent node.

[0052] The embodiment realizes effective classification of business activities, accurately distinguishes different business scenarios through conversation cutting and clustering analysis, mines sequence patterns and filters target activity patterns, extracts core business steps, identifies dependency relationships and mapping strength coefficients, and clearly defines the close degree of association between business steps, calculates execution tendency weights, quantifies execution preferences between steps, and finally constructs a template task tree with business logic and quantitative guidance significance, providing precise process support for subsequent requirement disassembly and task execution.

[0053] In a possible implementation of the embodiment of the application, the method further includes: creating a test environment, and generating a plurality of test cases containing different types of test tasks and network environment parameter combinations; For each agent in the agent candidate pool, each test case is executed to obtain a comprehensive test result of the agent for each task type under each network environment, and the comprehensive test result includes a task performance indicator and an environment sensitivity indicator; a performance score matrix is constructed for each agent based on the comprehensive test result of each agent and a preset scoring rule, a row of the performance score matrix represents a network environment, a column represents a task type, and an element represents an indicator score vector, and the indicator score vector includes an indicator score of each task performance indicator and each environment sensitivity indicator; The comprehensive test result and the performance score matrix of each agent in the agent candidate pool are stored as capability metadata.

[0054] In the embodiment, the network environment parameter combination includes bandwidth, delay, packet loss rate, and jitter. The task performance indicator can include accuracy, processing time consumption, and throughput. The accuracy is used to indicate whether the output content meets the expectation, and the matching degree between the standard output and the actual output can be calculated as the accuracy. The processing time consumption represents the difference between the total time consumption and the network delay. The throughput represents the amount of data processed per unit time. According to the business requirements of the target enterprise, the enterprise manager can set the lower limit of the task performance indicator for each task type in advance, including the qualified line of each task performance indicator.

[0055] Any task type is taken as a target task type, and any network environment is taken as a target network environment. A target test case is executed by each agent in the agent candidate pool in combination with the target network environment and the target task type. Any task performance indicator is taken as a target task performance indicator. From the comprehensive test results of the target test case executed by each agent, the best test result of the target task performance indicator is determined, and the best test result is taken as the upper limit of the range. The lower limit of the target task performance indicator corresponding to the target task type is taken as the lower limit of the range. The range formed by the lower limit and the upper limit of the range is taken as the scoring reference range of the target task performance indicator. Referring to the above process, the scoring reference range of each task performance indicator of each test case can be obtained.

[0056] Any agent is selected as a target agent, and the comprehensive test result of the target test case executed by the target agent is obtained. The process of determining the indicator score of the task performance indicator and the environment sensitivity indicator based on the comprehensive test result includes: if the indicator test result of the target task performance indicator in the comprehensive test result is within its scoring reference range, then the first difference value between the indicator test result of the target task performance indicator and the lower limit of the scoring reference range, and the second difference value between the upper limit and the lower limit of the scoring reference range are calculated. The indicator score of the target agent on the target task performance indicator = first difference value / second difference value x 100.

[0057] If the indicator test result of the target task performance indicator in the comprehensive test result is not within the corresponding scoring reference range, it is determined that the target agent is not available in executing the target test case. The element of the target test case in the performance score matrix is set to a null value, i.e. any task performance indicator in the test result does not meet the corresponding performance indicator lower limit, which indicates that the agent is not available in the test case.

[0058] The environmental sensitivity indicators include stability coefficient, decay rate and environmental fitness. For each agent, the performance of the agent in different network environments for the same task type is analyzed to obtain the indicator score of the environmental sensitivity indicators. Taking the target agent and the target task type as an example, the reciprocal of the standard deviation of the accuracy rate in the test results of the target agent performing the target task type in all network environments is calculated. The smaller the standard deviation is, the higher the stability coefficient is. For the decay rate, based on the throughput in different network environments, the difference between the best network environment (with the highest throughput) and the worst network environment (with the lowest throughput) is determined, and the ratio of the difference to the throughput of the best network environment is calculated as the percentage of performance decline. The accuracy rate of the test results and the throughput can be replaced by a comprehensive performance index obtained by weighted summation of multiple task performance indicators, which is not limited in the embodiment. The environmental fitness represents the comprehensive score, and the weighted sum of the stability coefficient and (1-decay rate) can be calculated as the environmental fitness. After obtaining the numerical value of the environmental sensitivity indicators of the target agent for the target task type, the product of the numerical value of each environmental sensitivity indicator and 100 is taken as the corresponding indicator score, so as to map each indicator score to the score range of 0-100.

[0059] The process of determining the task performance indicators and the indicator scores of the environmental sensitivity indicators is performed by the target agent under the combined conditions of any network environment and any task type, and the indicator scores of the target agent under each network environment and each task type are obtained, and then the performance score matrix of the target agent is constructed. Referring to the construction process of the target agent performance score matrix, the performance score matrix of each agent in the agent candidate pool is constructed.

[0060] The embodiment ensures the integrity of the performance evaluation of the agent by designing multi-dimensional test cases to comprehensively cover different task types and network environments; quantifies and structures the agent capability by combining the test results and the performance score matrix, and intuitively presents the performance of the agent in different scenarios; stores the capability metadata to provide objective and accurate data basis for subsequent selection of the agent, and avoids subjective selection bias.

[0061] One possible implementation of the embodiment of the application selects an agent from the agent candidate pool based on the capability metadata and the task type of the task sequence to construct an agent list, which includes: Obtain the lower limit of the task performance indicators of each task type in the current network environment; For each task in the task sequence, based on the capability metadata, an agent group that can meet all the lower limits of the task performance indicators of the task type of the task in the network environment matching the current network environment is screened from the agent candidate pool; Select one agent from the agent group of each task to form an agent list based on the index score of the dependency constraint relationship and the environmental sensitivity index of each agent, and the dependency constraint relationship includes that the agents selected by any two tasks without dependency relationship are not the same.

[0062] In the embodiment, any task in the task sequence is taken as a target task, and the task type and the task performance index lower limit of the target task are determined. The parameters of the current network environment and each network environment in the test case are compared to determine the best matching network environment as the best matching network environment. Taking the target agent as an example, the index score vector of the target agent in the task type of the target task under the best matching network environment is extracted from the performance score matrix of the target agent, and if it is null, the target agent is unavailable. All agents are traversed, and the agents whose index score vectors in the task type of the target task under the best matching network environment are not null are screened to form an agent group that can meet all task performance index lower limits of the task type of the target task.

[0063] For the agent group of the target task, the index score vector of each agent in the agent group in the best matching network environment and the task type of the target task is calculated, the index score of the environmental sensitivity index is extracted from the vector, and the average index score of each environmental sensitivity index is calculated based on the predefined weight. The agent with the highest average index score is selected from the agent group of the target task and placed in the agent list.

[0064] If there is a dependency relationship between two tasks in the task sequence (adjacent in the task sequence), the same agent can be selected to realize resource reuse. If there is no dependency relationship between two tasks in the task sequence, the agents of the two tasks are different to avoid resource conflicts; the agent of the task whose execution agent is determined first in the two tasks without dependency relationship is unchanged, and the agent of the task whose execution agent is determined later automatically selects the agent with the second highest average index score to avoid.

[0065] The embodiment combines the current network environment to screen the agent, ensures the adaptability of the agent in the actual running environment, and avoids performance loss caused by an unsuitable environment; sets the performance index lower limit to guarantee the basic execution ability of the selected agent; considers the dependency constraint to avoid the conflict of the agents of tasks without dependency relationship, and selects a more stable agent in combination with the environmental sensitivity index score.

[0066] One possible implementation manner of the embodiment of the application determines the current task execution deviation degree of the current intervention node, including: Determine the task type of each preceding task before the current intervention node in the task sequence, and determine the task type composition of the executed task sequence based on the task type of each preceding task. determine a dynamic weight of each deviation dimension based on the task type composition; obtain a current task execution result of the current intervention node, and calculate a first task execution deviation degree of the current task execution result in each deviation dimension based on the current task execution result and the dynamic weight of each deviation dimension; determine a preceding task and a subsequent task of the current intervention node in the task sequence, and update the first task execution deviation degree based on a mapping intensity coefficient between the preceding task and the subsequent task in the template task tree to obtain a second task execution deviation degree; determine whether the current intervention node is a first intervention node in the plurality of intervention nodes; if the current intervention node is the first intervention node, the second task execution deviation degree is taken as the current task execution deviation degree; if the current intervention node is not the first intervention node, calculate a time sequence cumulative deviation based on the task execution deviation degrees of all the intervention nodes between the current intervention nodes, and update the second task execution deviation degree based on the time sequence cumulative deviation to obtain the current task execution deviation degree.

[0067] In this embodiment, all preceding tasks before the current intervention node are traversed, the task type of each task is recorded, and a type list of executed tasks is formed. The proportion of each task type in the preceding tasks is recorded, for example, the proportion of data collection tasks is 60%, and the proportion of data verification tasks is 40%, to obtain a task type composition distribution.

[0068] A plurality of deviation dimensions are defined in advance according to an actual business scenario. Optionally, the deviation dimensions include an accuracy deviation (a matching degree deviation between an actual output result and an expected result), a completeness deviation (a proportion of missing fields of output data), and a format deviation (a degree of inconsistency between an output data format and a format requirement). A basic weight of each deviation dimension (the sum of the weights is 1) is preset for each task type, for example, the weight of the format dimension in the data collection task is higher, and the weight of the accuracy dimension in the data verification task is higher. The distribution of the basic weight of each dimension under a specific task type is not limited in this embodiment.

[0069] The basic weights of each deviation dimension are weighted and summed to obtain a dynamic weight of each deviation dimension according to the task type composition. For any deviation dimension k, the proportion of each task type in the task type composition and the basic weight of the task type for the dimension k are calculated as the deviation of the task type, and the sum of the deviations of each task type in the task type composition is calculated as the dynamic weight of the deviation dimension k. The dynamic weight of each deviation dimension is determined by referring to the determination process of the dynamic weight of the dimension k.

[0070] The current task execution result of the current intervention node is collected, and the current task execution result is an actual output result. The actual output result is compared with a preset expected result. The expected result can be a historical output of the current intervention node in the template task tree, or the expected result can be predicted by a pre-trained neural network. This embodiment is not limited. For each deviation dimension, the deviation value of the actual result and the expected result in the current task execution result is calculated. The accuracy deviation value = 1 - the number of matched fields / the total number of fields. When the output result is numerical data, the accuracy deviation value = 1 - the actual result / the expected result. The integrity deviation value = the number of missing fields / the total number of fields. The format deviation value = the number of format error fields / the total number of fields. The obtained deviation values of each dimension are weighted and summed according to the determined dynamic weight to obtain a first task execution deviation degree.

[0071] In the task sequence, the previous task of the current intervention node is the immediately preceding task, and the subsequent task is the immediately following task. From the template task tree, the mapping strength coefficient between the previous task node and the subsequent task node is found, which represents the influence degree of the output result of the previous task on the execution result of the subsequent task. The obtained mapping strength coefficient is used as a deviation transmission coefficient, which represents the influence proportion of the deviation of the previous task on the subsequent task. The product of the mapping strength coefficient and the first task execution deviation degree is calculated as a second task execution deviation degree.

[0072] If the current intervention node is not the first intervention node, the deviation accumulation effect of the historical intervention node is considered.

[0073] Wherein, n is the number of all intervention nodes before the current intervention node, the attenuation coefficient represents the deviation transmission degree between each two adjacent nodes in the task sequence, m is the interval step number, and represents the task sequence number difference between each two adjacent intervention nodes.

[0074] Further, the current task execution deviation degree = the second task execution deviation degree + the time sequence cumulative deviation. Wherein, the product of the time sequence cumulative deviation and the adjustment coefficient can be used as an updated time sequence cumulative deviation, and the sum of the second task execution deviation degree and the updated time sequence cumulative deviation is calculated as the current task execution deviation degree, so as to control the influence weight of the time sequence cumulative deviation by adjusting the coefficient, and avoid over amplification.

[0075] The embodiment determines the dynamic weight according to the pre-task type composition, so that the deviation evaluation is more suitable for the business scene, and the limitation of a single weight is avoided. The influence of the association between tasks on the deviation is considered in combination with the mapping strength coefficient, so that the comprehensiveness of the deviation evaluation is improved. The time sequence cumulative deviation is introduced, and the superposition effect of the historical deviation is considered, so that the finally calculated task execution deviation degree is more accurate.

[0076] In a possible implementation of the embodiment of the present application, the execution strategy of the subsequent task of the intervention node in the task sequence is adjusted, including: Starting from the first task of the task sequence, the intermediate task execution result after each task is executed is obtained in sequence, and the intermediate task execution deviation degree of the intermediate task execution result in each deviation dimension is calculated; The intermediate task execution deviation degree obtained after each task is executed is compared with the preset alarm threshold, and the task corresponding to the intermediate task execution deviation degree exceeding the preset alarm threshold is taken as the target task; The agent corresponding to the target task in the agent list is adjusted, and the target task and the subsequent task of the target task in the task sequence are executed according to the adjusted agent list.

[0077] In the embodiment, the preset alarm threshold can be set according to actual experience. The intermediate task execution result after each task is executed is calculated in the manner of calculating the first task execution deviation degree of the intervention node. The task with the intermediate task execution deviation degree exceeding the preset alarm threshold is taken as the target task. For the target task, from the agent group corresponding to the target task in the process of constructing the agent list, one agent satisfying the dependency constraint relationship and having the highest average index score except the agent matched by the target task in the agent list is selected as the adjusted agent, so that the update of the agent list is completed.

[0078] The embodiment realizes dynamic tracking of the task execution process by monitoring the intermediate result and the deviation degree in real time, can quickly locate the target task exceeding the threshold, timely adjusts the agent corresponding to the target task, solves the task deviation problem in a targeted manner, avoids the deviation from being conducted to the subsequent task, re-executes the target task and the subsequent task, ensures the business processing to return to the correct track, effectively reduces the influence of the deviation on the overall business result, and protects the business processing quality.

[0079] In a possible implementation of the embodiment of the present application, when the task sequence is executed, the lowest level node of each task in the template task tree is determined; The child nodes of the lowest level node in the template task tree are determined, and each child node is arranged in descending order according to the execution tendency weight between the child node and the lowest level node to obtain a recommended task sequence; The recommended task sequence is fed back to the target personnel, so that the target personnel execute the corresponding task according to the selection of the target personnel for the recommended task sequence.

[0080] In the embodiment, the lowest level node is the node corresponding to the last task in the task sequence in the template task tree. The recommended task sequence conforming to the business habit is generated in combination with the child nodes of the node and the execution tendency weight.

[0081] The embodiment is based on a template task tree and an execution tendency weight to generate a recommended task sequence, fits the historical business execution habit of an enterprise, improves the rationality and practicability of the recommendation, feeds back the recommended sequence to target personnel, provides clear guidance for subsequent business processing, reduces personnel decision cost, supports personnel self-selection, and balances intelligent recommendation and manual intervention flexibility.

[0082] An electronic device is provided in the embodiment of the application, such as Figure 2 As shown in the figure, Figure 2 The electronic device 200 shown in the figure includes a processor 201 and a memory 203. The processor 201 and the memory 203 are connected, such as through a bus 202. Optionally, the electronic device 200 can also include a transceiver 204. It should be noted that the transceiver 204 is not limited to one in actual application, and the structure of the electronic device 200 does not constitute a limitation on the embodiments of the application.

[0083] The processor 201 can be a CPU (Central Processing Unit, central processor), a general-purpose processor, a DSP (Digital Signal Processor, data signal processor), an ASIC (Application Specific Integrated Circuit, application specific integrated circuit), an FPGA (Field Programmable Gate Array, field programmable gate array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It can implement or execute various exemplary logical blocks, modules and circuits described in combination with the disclosure content of the application. The processor 201 can also be a combination of computing functions, such as one or more microprocessor combinations, combinations of DSP and microprocessor, etc.

[0084] The bus 202 can include a channel for transmitting information between the above-mentioned components. The bus 202 can be a PCI (Peripheral Component Interconnect, peripheral component interconnect) bus or an EISA (Extended Industry Standard Architecture, extended industry standard architecture) bus, etc. The bus 202 can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, Figure 2 In the figure, only one thick line is used to represent the bus, but it does not mean that there is only one bus or one type of bus.

[0085] The memory 203 can be a ROM (Read Only Memory) or other type of static storage device that can store static information and instructions, a RAM (Random Access Memory) or other type of dynamic storage device that can store information and instructions, an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited thereto.

[0086] The memory 203 is configured to store application program codes for implementing the solutions of the present application, and the processor 201 is configured to control the execution of the application program codes. The processor 201 is configured to execute the application program codes stored in the memory 203 to implement the content shown in the foregoing business processing method embodiments.

[0087] Figure 2 The electronic device shown is merely an example, and should not impose any limitation on the functions and use range of the embodiments of the present application.

[0088] The embodiments of the present application provide a computer readable storage medium, which has stored thereon a computer program, and when the computer program is run on a computer, the computer can execute the content shown in the foregoing business processing method embodiments.

[0089] It should be understood that, although each step in the flowchart of the accompanying drawings is shown in sequence according to the direction of the arrow, these steps are not necessarily executed in sequence according to the direction of the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and they can be executed in other sequences. Moreover, at least part of the steps in the flowchart of the accompanying drawings can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or sub-steps or stages of other steps.

[0090] The embodiments of the present application provide a computer program product, which includes a computer program, and when the computer program is executed by a processor, the content shown in the foregoing business processing method embodiments is implemented.

[0091] The above merely describes some embodiments of the present application, and it should be pointed out that, for those skilled in the art, some improvements and refinements can be made without departing from the principles of the present application, and these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A service processing method characterized by, The method comprises the following steps: linking the pre-built agent platform and the enterprise system of the target enterprise when receiving an access request of the target enterprise, obtaining historical system operation data of the enterprise system, and constructing a template task tree based on the historical system operation data; receiving user demands uploaded by target personnel of the target enterprise through the enterprise system, decomposing the user demands into a task sequence based on the template task tree, and determining the task type of each task in the task sequence; obtaining the capability metadata of the agent candidate pool of the agent platform, selecting agents from the agent candidate pool to form an agent list based on the capability metadata and the task type of the task sequence, and assigning the task sequence to the agent list for execution; determining a plurality of intervention nodes according to the depth of the task sequence, determining the task execution deviation degree at each intervention node, and adjusting the execution strategy of the subsequent task of the intervention node in the task sequence when the task execution deviation degree of any intervention node exceeds a preset intervention threshold.

2. The service processing method according to claim 1, characterized by, The method further comprises the following steps: cutting the historical system operation data based on a session timeout rule to obtain business activity data, and performing cluster analysis on the business activity data to obtain a plurality of cluster clusters; for each cluster cluster in the plurality of cluster clusters, performing sequence pattern mining on the cluster members of the cluster cluster to obtain the support degree of each activity mode, selecting target activity modes from the cluster members based on the support degree, and identifying the dependency relationship and mapping intensity coefficient between the target activity modes; constructing an initial template task tree by taking each target activity mode obtained from the plurality of cluster clusters as a node based on the dependency relationship; for each node in the initial template task tree, determining the intra-cluster frequency of the node in the cluster cluster and the global frequency of the node in the plurality of cluster clusters, and determining the execution inclination weight between the node and the parent node of the node based on the intra-cluster frequency and the global frequency; annotating the execution inclination weight and the mapping intensity coefficient between each two nodes obtained in the corresponding positions of the initial template task tree to obtain the template task tree.

3. The service processing method according to claim 1, characterized by, The method further comprises the following steps: creating a test environment and generating a plurality of test cases containing different types of test tasks and network environment parameter combinations; for each agent in the agent candidate pool, executing each test case to obtain the comprehensive test result of the agent for each task type under each network environment, wherein the comprehensive test result comprises a task performance indicator and an environment sensitivity indicator; constructing a performance score matrix for each agent based on the comprehensive test result of each agent and a preset scoring rule, wherein the rows of the performance score matrix represent network environments, the columns represent task types, and the elements represent indicator score vectors, and the indicator score vectors comprise indicator scores of each task performance indicator and each environment sensitivity indicator; storing the comprehensive test result and the performance score matrix of each agent in the agent candidate pool as the capability metadata.

4. The service processing method according to claim 3, characterized by, The selecting the agent from the agent candidate pool based on the capability metadata and the task type of the task sequence comprises: obtaining a current network environment and a lower limit of a task performance index of each task type; for each task in the task sequence, based on the capability metadata, filtering an agent group from the agent candidate pool that can meet all lower limits of task performance indexes of the task type of the task under a network environment matching the current network environment; selecting one agent from the agent group of each task based on a dependency constraint relationship and an index score of an environment sensitivity index of each agent, the dependency constraint relationship comprising that the agents selected by any two tasks without a dependency relationship are not the same.

5. The service processing method of claim 1, wherein, The determining the current task execution deviation degree of the current intervention node comprises: determining the task type of each preceding task before the current intervention node in the task sequence, and determining a task type composition of an executed task sequence based on the task type of each preceding task; determining a dynamic weight of each deviation dimension based on the task type composition; obtaining a current task execution result of the current intervention node, and calculating a first task execution deviation degree of the current task execution result on each deviation dimension based on the current task execution result and the dynamic weight of each deviation dimension; determining a preceding task and a subsequent task of the current intervention node in the task sequence, and updating the first task execution deviation degree based on a mapping intensity coefficient between the preceding task and the subsequent task in the template task tree to obtain a second task execution deviation degree; judging whether the current intervention node is a first intervention node in the plurality of intervention nodes; if the current intervention node is the first intervention node, the second task execution deviation degree is taken as the current task execution deviation degree; if the current intervention node is not the first intervention node, calculating a time sequence cumulative deviation based on the task execution deviation degrees of all intervention nodes between the current intervention nodes, and updating the second task execution deviation degree based on the time sequence cumulative deviation to obtain the current task execution deviation degree.

6. The service processing method of claim 1, wherein, The adjusting the execution strategy of the subsequent task of the intervention node in the task sequence comprises: starting from a first task of the task sequence, sequentially obtaining an intermediate task execution result after each task is executed, and calculating an intermediate task execution deviation degree of the intermediate task execution result on each deviation dimension; comparing the intermediate task execution deviation degree obtained after each task is executed with a preset alarm threshold, and taking a task corresponding to an intermediate task execution deviation degree exceeding the preset alarm threshold as a target task; adjusting an agent corresponding to the target task in the agent list, and executing the target task and a subsequent task of the target task in the task sequence according to the adjusted agent list.

7. The service processing method of claim 1, wherein, The method further comprises: when the task sequence is executed, determining a lowest level node of each task in the template task tree. determining child nodes of the lowest level node in the template task tree, and arranging the child nodes in descending order of execution tendency weight between each child node and the lowest level node to obtain a recommended task sequence; feeding back the recommended task sequence to the target personnel, so as to execute corresponding tasks according to the selection of the target personnel for the recommended task sequence.

8. An electronic device, comprising: comprise: at least one processor; a memory; at least one application program, wherein the at least one application program is stored in the memory and configured to be executed by the at least one processor, and the at least one application program is configured to execute the business processing method in any one of claims 1-7.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, When the computer program is executed in the computer, the computer is caused to execute the business processing method in any one of claims 1-7.

10. A computer program product, characterised in that, comprise a computer program, which, when executed by a processor, implements the steps of the business processing method in any one of claims 1-7.

Images