Task workflow orchestration method and apparatus, storage medium, and electronic device

Abstract

The application discloses a task workflow arrangement method and device, a storage medium and an electronic device, and relates to the technical field of data processing. The method comprises the following steps: obtaining service data of a target service at a target time, and generating a standardized task to be processed based on the service data; determining at least one target standardized task matched with the standardized task from a preset task library, and taking an optimal workflow of each target standardized task as a candidate workflow of the standardized task; determining the execution cost of each candidate workflow based on the working state information of the execution personnel of each process node in the enterprise data lake at the target time; and taking the candidate workflow with the lowest execution cost as the optimal workflow of the standardized task. The application solves the technical problem that the related art uses a static workflow architecture to arrange a task workflow, and lacks pre-deployment, resulting in low management efficiency and lagging management and control.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically, to a task workflow orchestration method, apparatus, storage medium, and electronic device. Background Technology

[0002] As enterprises deepen their digital transformation, task management platforms (such as Software as a Service systems) have become core infrastructure for supporting business processes. However, when faced with complex and ever-changing business scenarios, mainstream task management platforms exhibit the following significant shortcomings:

[0003] First, since most mainstream task management platforms adopt a static workflow architecture based on finite state machines, their task generation and process orchestration mainly rely on manual configuration or simple rule triggers. When faced with a new task, they cannot automatically associate it with the workflows of similar historical tasks, resulting in planning starting from the initial state each time. This makes it impossible to use past experience to optimize the process, leading to low generation efficiency.

[0004] Secondly, most mainstream task management platforms are reactive, only making adjustments after an error or delay occurs during task execution. This "execute first, correct later" model makes it difficult for managers to avoid process blockages caused by resource squeezing or dependency conflicts in advance, resulting in serious lag in control.

[0005] There is currently no effective solution to the above problems. Summary of the Invention

[0006] This invention provides a task workflow orchestration method, apparatus, storage medium, and electronic device to at least solve the technical problems of related technologies that use static workflow architecture to orchestrate task workflows and lack prior simulation, resulting in low management efficiency and lagging control.

[0007] According to one embodiment of the present invention, a task workflow orchestration method is provided, comprising: acquiring business data of a target business at a target time, and generating standardized tasks to be processed based on the business data; determining at least one target standardized task matching the standardized task from a preset task library, and using the optimal workflow of each target standardized task as a candidate workflow for the standardized task, wherein the preset task library includes multiple executed standardized tasks and the optimal workflow corresponding to each standardized task; determining the execution cost of each candidate workflow based on the work status information of the executors at each process node in the enterprise data lake at the target time, wherein the enterprise data lake stores at least the real-time work status information of the executors at process nodes on multiple workflow paths; and using the candidate workflow with the lowest execution cost as the optimal workflow for the standardized task.

[0008] Optionally, generating standardized tasks to be processed based on business data includes: preprocessing the business data, wherein the preprocessing includes: consistency verification, text extraction, and standardization processing, wherein the consistency verification includes at least one of the following: time conflict verification, responsible person conflict verification, approval authority conflict verification, and delivery standard conflict verification; constructing a target task information relationship diagram based on the preprocessed business data; and converting the target task information relationship diagram into standardized tasks, wherein the standardized tasks include multiple task fields, and the task fields include at least: task name field, task type field, personnel field, time field, and deliverable field.

[0009] Optionally, constructing a target task information relationship graph based on the preprocessed business data includes: encoding the preprocessed business data into task information feature vectors; constructing an initial task information relationship graph based on the task information feature vectors, wherein the graph nodes in the initial task information relationship graph include at least: task nodes, personnel nodes, permission nodes, approval nodes, deliverable nodes, and time nodes, and the edges in the task information relationship graph include at least: responsible person edges connecting task nodes and personnel nodes, deliverable edges connecting task nodes and deliverable nodes, approval dependency edges connecting approval nodes and task nodes, date constraint edges connecting time nodes and task nodes, and task dependency edges connecting different task nodes; retrieving multiple target business logics related to the target business from a preset business logic library, and writing the multiple target business logics as constraints into the initial task information relationship graph to obtain the target task information relationship graph, wherein the business logic library includes: multiple business logics related to different businesses.

[0010] Optionally, the execution cost of each candidate workflow is determined based on the work status information of the executors at each process node in the enterprise data lake at the target time. This includes: for each candidate workflow, determining the compliance risk level of the candidate workflow based on the node information of each process node on the candidate workflow, wherein the node information includes at least: node type and the permission information of the executors; when the compliance risk level is low, determining the multi-dimensional risk indicators of the candidate workflow based on the personnel information of the executors at each process node on the candidate workflow and their work status information at the target time, and determining the execution cost of the candidate workflow based on the multi-dimensional risk indicators, wherein the personnel information includes at least: unit labor cost, the work status information includes at least: the available time window from the target time to the end time of the executor's work, the list of remaining tasks to be processed, the estimated processing time and the estimated waiting time for each remaining task, and the multi-dimensional risk indicators include at least: real-time personnel conflict index, estimated overdue probability, estimated delivery time and estimated overdue cost; when the compliance risk level is high, setting the execution cost of the candidate workflow to the maximum value.

[0011] Optionally, the compliance risk level of the candidate workflow is determined based on the node information of each process node in the candidate workflow, including: mapping the node information of each process node in the candidate workflow to a preset business rule base, wherein the business rule base includes multiple business rules, and the business rules include at least one of the following: job permission rules, node approval rules, and process integrity rules; if the node information of any process node in the candidate workflow does not meet the business rules in the business rule base, the compliance risk level of the candidate workflow is determined to be high; if the node information of all process nodes in the candidate workflow meets the business rules in the business rule base, the compliance risk level of the candidate workflow is determined to be low.

[0012] Optionally, based on the personnel information of the executors at each process node of the candidate workflow and their work status information at the target time, multi-dimensional risk indicators for the candidate workflow are determined, including: determining the backlog workload of the executors at each process node of the candidate workflow at the target time based on the remaining task list and the estimated processing time of each remaining task, and determining the unit-time productivity of the executors at each process node of the candidate workflow in conjunction with the available time window; calculating the ratio of backlog workload to unit-time productivity to obtain the real-time load of the executors at each process node of the candidate workflow at the target time; statistically analyzing the number of target process nodes in the candidate workflow whose real-time load at the target time exceeds a preset saturation threshold and the overall overload degree, and determining the real-time personnel conflict index of the candidate workflow in conjunction with the skill complementarity and resource sharing relationship among the executors of the target process nodes; and determining the real-time personnel conflict index of the candidate workflow based on the execution of each process node of the candidate workflow. The estimated probability of delay for candidate workflows is determined by considering the real-time workload of personnel at the target time, the estimated waiting time of personnel at each process node, and the pre-set waiting time threshold. Based on the task dependencies of the candidate workflows, at least one parallel sub-path is identified, where each sub-path includes multiple non-parallel nodes. For each sub-path, the estimated processing time and estimated waiting time of personnel at each non-parallel node within the sub-path for completing the remaining tasks at the target time are summed sequentially to obtain the estimated completion time of the sub-path. The estimated completion times of at least one sub-path are compared, and the latest estimated completion time is taken as the estimated delivery time of the candidate workflow. The estimated delay between the estimated delivery time and the expected delivery time of the candidate workflow is determined, and combined with the unit labor cost of personnel at each process node in the candidate workflow and the pre-set delay penalty coefficient of the target business, the estimated delay cost of the candidate workflow is determined.

[0013] Optionally, after selecting the candidate workflow with the lowest execution cost as the optimal workflow for the standardized task, the method further includes: executing the standardized task according to the optimal workflow, and periodically acquiring the status information of the executors at each process node on the optimal workflow, wherein the status information includes at least one of the following: whether they are on duty, personnel load; if the executors of the target process node are not on duty during the target period, or if the personnel load of the executors of the target process node during the target period is higher than a preset load threshold, determining multiple candidates based on the personnel level of the executors of the target process node, and determining a recommended score for each candidate based on the personnel information of each candidate, and replacing the executors of the target process node with the candidate with the highest recommended score, wherein the personnel information includes at least: professional matching degree, historical processing efficiency, personnel load, and historical processing experience index.

[0014] According to one embodiment of the present invention, a task workflow orchestration apparatus is also provided, comprising: a task generation module, configured to acquire business data of a target business at a target time, and generate standardized tasks to be processed based on the business data; a first determination module, configured to determine at least one target standardized task matching the standardized task from a preset task library, and to use the optimal workflow of each target standardized task as a candidate workflow for the standardized task, wherein the preset task library includes multiple executed standardized tasks and the optimal workflow corresponding to each standardized task; a second determination module, configured to determine the execution cost of each candidate workflow based on the work status information of the executors at each process node in the enterprise data lake at the target time, wherein the enterprise data lake stores at least the real-time work status information of the executors at process nodes on multiple workflow paths; and a third determination module, configured to use the candidate workflow with the lowest execution cost as the optimal workflow for the standardized task.

[0015] Optionally, the above-mentioned apparatus is further configured to generate standardized tasks to be processed based on business data through the following method: preprocessing the business data, wherein the preprocessing includes: consistency verification, text extraction and standardization processing, wherein the consistency verification includes at least one of the following: time conflict verification, responsible person conflict verification, approval authority conflict verification, and delivery standard conflict verification; constructing a target task information relationship diagram based on the preprocessed business data; and converting the target task information relationship diagram into standardized tasks, wherein the standardized tasks include multiple task fields, and the task fields include at least: task name field, task type field, personnel field, time field and deliverable field.

[0016] Optionally, the above-mentioned apparatus is further configured to construct a target task information relationship graph based on the preprocessed business data using the following method: encoding the preprocessed business data into task information feature vectors; constructing an initial task information relationship graph based on the task information feature vectors, wherein the graph nodes in the initial task information relationship graph include at least: task nodes, personnel nodes, permission nodes, approval nodes, deliverable nodes, and time nodes, and the edges in the task information relationship graph include at least: responsible person edges connecting task nodes and personnel nodes, deliverable edges connecting task nodes and deliverable nodes, approval dependency edges connecting approval nodes and task nodes, date constraint edges connecting time nodes and task nodes, and task dependency edges connecting different task nodes; retrieving multiple target business logics related to the target business from a preset business logic library, and writing the multiple target business logics as constraints into the initial task information relationship graph to obtain the target task information relationship graph, wherein the business logic library includes multiple business logics related to different businesses.

[0017] Optionally, the aforementioned device is further configured to determine the execution cost of each candidate workflow based on the work status information of the executors at each process node in the enterprise data lake at the target time, using the following method: For each candidate workflow, the compliance risk level of the candidate workflow is determined based on the node information of each process node in the candidate workflow, wherein the node information includes at least: node type and the permission information of the executors; When the compliance risk level is low, the multi-dimensional risk indicators of the candidate workflow are determined based on the personnel information of the executors at each process node in the candidate workflow and their work status information at the target time, and the execution cost of the candidate workflow is determined based on the multi-dimensional risk indicators, wherein the personnel information includes at least: unit labor cost, the work status information includes at least: the available time window from the target time to the end time of the executor's work, the list of remaining tasks to be processed, the estimated processing time and the estimated waiting time for each remaining task, and the multi-dimensional risk indicators include at least: real-time personnel conflict index, estimated overdue probability, estimated delivery time and estimated overdue cost; When the compliance risk level is high, the execution cost of the candidate workflow is set to the maximum value.

[0018] Optionally, the aforementioned device is further configured to determine the compliance risk level of a candidate workflow based on the node information of each process node in the candidate workflow using the following method: mapping the node information of each process node in the candidate workflow to a preset business rule base, wherein the business rule base includes multiple business rules, and the business rules include at least one of the following: job authority rules, node approval rules, and process integrity rules; if the node information of any process node in the candidate workflow does not meet the business rules in the business rule base, the compliance risk level of the candidate workflow is determined to be high; if the node information of all process nodes in the candidate workflow meets the business rules in the business rule base, the compliance risk level of the candidate workflow is determined to be low.

[0019] Optionally, the aforementioned device is further configured to determine the multidimensional risk indicators of the candidate workflow based on the personnel information of the executors at each process node of the candidate workflow and their work status information at the target time, using the following methods: determining the backlog workload of the executors at each process node of the candidate workflow at the target time based on the remaining task list and the estimated processing time of each remaining task, and determining the unit-time productivity of the executors at each process node of the candidate workflow in conjunction with the available time window; calculating the ratio of the backlog workload to the unit-time productivity to obtain the real-time load of the executors at each process node of the candidate workflow at the target time; statistically analyzing the number of target process nodes in the candidate workflow whose real-time load at the target time exceeds a preset saturation threshold and the overall overload level, and determining the real-time personnel conflict index of the candidate workflow in conjunction with the skill complementarity and resource sharing relationships among the executors of the target process nodes; and determining the real-time personnel conflict index of the candidate workflow based on the personnel information of the executors at each process node of the candidate workflow. The estimated probability of delay for a candidate workflow is determined by considering the real-time workload of personnel at each process node at the target time, the estimated waiting time of personnel at each process node, and the relationship between these values ​​and the preset waiting time threshold. Based on the task dependencies of the candidate workflow, at least one parallel sub-path is identified, where each sub-path includes multiple non-parallel nodes. For each sub-path, the estimated processing time and estimated waiting time of personnel at each non-parallel node within the sub-path for completing the remaining tasks at the target time are summed sequentially to obtain the estimated completion time of the sub-path. The estimated completion times of at least one sub-path are compared, and the latest estimated completion time is taken as the estimated delivery time of the candidate workflow. The estimated delay between the estimated delivery time and the expected delivery time of the candidate workflow is determined, and combined with the unit labor cost of personnel at each process node in the candidate workflow and the preset delay penalty coefficient of the target business, the estimated delay cost of the candidate workflow is determined.

[0020] Optionally, the above-mentioned device is further configured to, after selecting the candidate workflow with the lowest execution cost as the optimal workflow for the standardized task, execute the standardized task according to the optimal workflow, and periodically obtain the status information of the executors at each process node on the optimal workflow, wherein the status information includes at least one of the following: whether they are on duty, personnel load; if the executor of the target process node is not on duty during the target period, or if the personnel load of the executor of the target process node during the target period is higher than a preset load threshold, determine multiple candidates based on the personnel level of the executor of the target process node, determine the recommended score of each candidate based on the personnel information of each candidate, and replace the executor of the target process node with the candidate with the highest recommended score, wherein the personnel information includes at least: professional matching degree, historical processing efficiency, personnel load, and historical processing experience index.

[0021] According to one embodiment of the present invention, a computer-readable storage medium is also provided, wherein the storage medium stores a computer program, wherein the computer program is configured to execute the task workflow orchestration method described above when running on a computer or processor.

[0022] According to one embodiment of the present invention, an electronic device is also provided, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the task workflow orchestration method of any of the above.

[0023] According to one embodiment of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements the task workflow orchestration method described in any of the above claims.

[0024] In this embodiment of the invention, business data of the target business at a target time is acquired, and standardized tasks to be processed are generated based on the business data. Then, at least one target standardized task matching the standardized task is determined from a preset task library, and the optimal workflow of each target standardized task is used as a candidate workflow for the standardized task. The execution cost of each candidate workflow is determined based on the work status information of the executors at each process node in the enterprise data lake at the target time, so that the candidate workflow with the lowest execution cost is selected as the optimal workflow for the standardized task. This achieves the technical effect of pre-quantitative evaluation and path optimization of the workflow through multi-dimensional data before the task is formally executed. It transforms the passive post-event adjustment of traditional static workflows into proactive pre-event prediction and precise resource matching based on real-time data-driven approaches, thereby effectively avoiding personnel bottlenecks, approval delays, and resource conflicts, and improving task delivery efficiency and the accuracy of full-process control. Furthermore, it solves the technical problem of related technologies that use static workflow architecture to orchestrate task workflows and lack pre-event simulation, resulting in low management efficiency and lagging control. Attached Figure Description

[0025] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0026] Figure 1 This is a flowchart of a task workflow orchestration method according to one embodiment of the present invention;

[0027] Figure 2 This is a flowchart illustrating the generation of a standardized task according to one embodiment of the present invention;

[0028] Figure 3 This is a flowchart of constructing a target task information relationship diagram according to one embodiment of the present invention;

[0029] Figure 4 This is a flowchart illustrating the execution cost of determining candidate workflows according to one embodiment of the present invention;

[0030] Figure 5 This is a structural block diagram of a task workflow orchestration apparatus according to one embodiment of the present invention;

[0031] Figure 6 This is a structural block diagram of a computer terminal for implementing a task workflow orchestration method according to one embodiment of the present invention. Detailed Implementation

[0032] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0033] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. In the description of these embodiments, unless otherwise stated, "a plurality of" means two or more. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0034] According to an embodiment of this application, a task workflow orchestration method is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0035] Figure 1 This is a flowchart of a task workflow orchestration method according to one embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps:

[0036] Step S101: Obtain the business data of the target business at the target time, and generate standardized tasks to be processed based on the business data.

[0037] Optionally, in this embodiment, the aforementioned target business refers to various business documents or data entities included in the full lifecycle management. For example, enterprise-level business entities with clearly defined template fields, review processes, and release blocking requirements, such as modular parts manuals, parts procurement requisition forms, production and manufacturing work orders, and after-sales project task books in vehicle R&D.

[0038] Optionally, in this embodiment, the target time is a specific time node or the instant of an event that triggers a change in the state of a business object or the generation of a new business object. For example: the time when a new part or module solution is identified during the solution selection phase, the time when a field is found to be missing or the format is abnormal, the time when a task is overdue or the executor is overloaded, the time when a review is rejected or a rectification is completed, and the time when a downstream system notifies of completion.

[0039] Optionally, in this embodiment, the aforementioned business data is a set of raw data used to drive task generation and status updates, formed by manual input by business personnel, external system event triggers, or multimodal unstructured data input. Therefore, this business data has the characteristics of being multi-source, heterogeneous, and unstructured. Specifically: multi-source refers to data originating from various channels such as manual interaction, automated system events, and unstructured media; heterogeneous refers to data formats covering vastly different forms such as structured fields, semi-structured text, and unstructured images / audio; and unstructured refers to data that has not undergone standardization and processing, exhibiting characteristics such as free semantic expression, uneven information density, and missing fields. For example, business data can be structured field data from forms manually entered by business personnel; business event trigger signals from external systems such as Enterprise Resource Planning (ERP) systems or Customer Relationship Management (CRM); text extracted from speech-to-text, natural language text, contract screenshots, and structured information after image recognition of paper documents.

[0040] Optionally, in the embodiments of this application, the above-mentioned standardized tasks are executable work units that drive the flow of business trends, and their types include, but are not limited to: confirmation tasks, review tasks, rectification tasks, release tasks, etc.

[0041] Step S102: Determine at least one target standardized task that matches the standardized task from the preset task library, and take the optimal workflow of each target standardized task as the candidate workflow of the standardized task.

[0042] Optionally, in this embodiment of the application, the aforementioned preset task library is a structured database that stores multiple standardized tasks that have been executed in the past and the optimal workflow corresponding to each standardized task, such as the optimal workflow corresponding to the module manual compilation task, the optimal workflow corresponding to the parts procurement approval task, the optimal workflow corresponding to the production and manufacturing work order approval task, etc.

[0043] Optionally, in the embodiments of this application, the aforementioned target standardized task is a historical or templated task instance extracted from a preset task library that is similar or comparable to the standardized task to be processed in terms of task type, business object attributes, complexity, and resource requirements.

[0044] Step S103: Determine the execution cost of each candidate workflow based on the work status information of the executors at each process node in the enterprise data lake at the target time.

[0045] Optionally, in this embodiment, the aforementioned enterprise data lake is a unified data asset foundation for centrally storing, aggregating, and accumulating all enterprise task data, business object data, and system interaction logs. It includes, but is not limited to, real task execution logs, execution costs, approval time records, deliverable return records, resource usage records, personnel workload data, external system event data, historical task data, and real-time work status information of personnel at each process node. The aforementioned personnel are individuals or roles responsible for handling the tasks at each process node. For example, for the task of compiling a module manual, the relevant personnel could be proofreaders, auditors, procurement auditors, financial auditors, logistics auditors, and candidate responsible persons with corresponding job authority.

[0046] Optionally, in the embodiments of this application, the above-mentioned execution cost is a comprehensive risk and resource consumption index that may be generated during the actual execution of the candidate workflow path.

[0047] Step S104: Select the candidate workflow with the lowest execution cost as the optimal workflow for the standardized task.

[0048] Optionally, in the embodiments of this application, the optimal workflow is the best execution plan selected to guide the execution of standardized tasks.

[0049] Through the above steps, business data of the target business at the target time is acquired, and standardized tasks to be processed are generated based on the business data. Then, at least one target standardized task matching the standardized task is determined from a preset task library, and the optimal workflow of each target standardized task is used as a candidate workflow for the standardized task. The execution cost of each candidate workflow is determined based on the work status information of the executors at each process node in the enterprise data lake at the target time, and the candidate workflow with the lowest execution cost is selected as the optimal workflow for the standardized task. This achieves the technical effect of pre-quantitative evaluation and path optimization of the workflow through multi-dimensional data before the task is officially executed. It transforms the passive post-event adjustment of traditional static workflows into proactive pre-event prediction and precise resource matching based on real-time data-driven approaches, thereby effectively avoiding personnel bottlenecks, approval delays, and resource conflicts, and improving task delivery efficiency and the accuracy of full-process control. Furthermore, it solves the technical problem of low management efficiency and lagging control caused by related technologies that use static workflow architectures to orchestrate task workflows and lack pre-event simulations.

[0050] The steps of the task workflow orchestration method provided in this application embodiment will be described below with reference to the specific implementation process.

[0051] As an optional solution, in step S101, based on business data, through... Figure 2 The process shown generates standardized tasks to be processed, including:

[0052] Step S101a: Preprocess the business data.

[0053] Optionally, in this embodiment, since the business data encompasses multimodal unstructured data such as voice, natural language text, contract screenshots, and photos of paper documents, and data from different sources and formats typically exhibit significant differences in semantic expression, data structure, and information density, semantic conflicts or omissions in fields such as time, responsible person, and delivery standards are highly likely to occur when inputting from multiple sources. Therefore, before generating standardized tasks, this embodiment needs to preprocess the unstructured data into a unified format and resolve multi-source information contradictions through consistency checks.

[0054] Specifically, the preprocessing includes: consistency verification, text extraction, and standardization. Consistency verification is the process of comparing the logical consistency and compliance of multi-source input information using a conflict discrimination rule base. This includes: time conflict verification, where recommended values ​​are given according to the priority of official documents, business system events, and verbal statements when deadlines, approval periods, or delivery windows from different sources are inconsistent; responsible person (i.e., executor) conflict verification, where candidate responsible persons are suggested when the responsible person lacks the necessary authority, project authority, or has exceeded the workload threshold; approval authority conflict verification, where the approval level is completed based on a preset rule base when the task type triggers confidentiality, budget, or quality-mandated nodes and the corresponding approver is not included in the input; and delivery standard conflict verification, where conflicts are marked and correction suggestions are generated when the screenshot contract differs from the manually entered deliverable format, version, or acceptance criteria.

[0055] Furthermore, text extraction is the process of parsing multimodal unstructured data (such as speech, images, and documents) and converting it into structured text information. This includes, but is not limited to, speech recognition transcription, optical character recognition, and natural language text extraction. Standardization, on the other hand, is the process of converting external system event fields into a unified data format within the task management platform. For example, parsing JSON or XML data returned from external system interfaces and mapping it to the internal system's task trigger source, business object number, event type, and timestamp ensures consistency and relevance of data across systems.

[0056] Step S101b: Construct a target task information relationship diagram based on the preprocessed business data.

[0057] Optionally, in the embodiments of this application, the above-mentioned target task information relationship diagram is a graph structure data used to represent task elements and their mutual logical relationships.

[0058] Regarding the aforementioned target task information relationship diagram, it can be obtained through... Figure 3 The process shown yields:

[0059] Step 1: Encode the preprocessed business data into task information feature vectors.

[0060] Specifically, the aforementioned task information feature vector is generated by converting preprocessed business data into a computer-processable numerical vector representation using a specific encoding algorithm. Specifically, a pre-trained natural language processing model (such as BERT, Transformer, etc.) or a multimodal fusion model is used to semantically embed the preprocessed business data, extracting multidimensional semantic features (such as key entities, intents, and contextual information). These extracted multidimensional semantic features are then concatenated, weighted, or fused using attention mechanisms to generate a fixed-dimensional or dynamic-dimensional numerical vector that can represent the overall semantics and logical relationships of the task—that is, the task information feature vector.

[0061] Step 2: Construct an initial task information relationship graph based on the task information feature vector.

[0062] The graph nodes in the initial task information relationship diagram mentioned above include at least the following: task nodes, personnel nodes, permission nodes, approval nodes, deliverable nodes, and time nodes. Specifically: task nodes are data entities representing specific business execution units, such as module manual creation tasks, parts replenishment tasks, review tasks, and rectification tasks; personnel nodes are data entities representing business entities that undertake task execution, approval, or collaboration roles, such as task leaders, executors, collaborators, approvers at various levels, and downstream notification recipients; permission nodes are rule-based data entities defining the legality of personnel node operations, such as job permissions, project permissions, confidentiality levels, budget limits, and approval level restrictions; approval nodes are data entities used for confirmation, verification, or approval in the business process, such as procurement approval nodes, financial approval nodes, logistics approval nodes, and quality review nodes; deliverable nodes are data entities representing outputs generated during task execution, such as module manual documents, parts drawings, interface confirmation forms, cost breakdown tables, and acceptance reports; and time nodes are time-dimensional data entities that constrain the timeliness of task execution, such as task start time, deadline, approval period, delivery window, and historical average processing time.

[0063] The edges in the aforementioned task information relationship graph include at least the following: responsible person edges connecting task nodes and personnel nodes (used to represent the task execution subject, such as assignment relationships, joint responsibility relationships, and permission mapping relationships); deliverable edges connecting task nodes and deliverable nodes (used to identify task output associations, such as generation relationships, version iteration relationships, and acceptance associations); approval dependency edges connecting approval nodes and task nodes (used to identify process control logic, such as pre-validation relationships, forced blocking relationships, and condition triggering relationships); date constraint edges connecting time nodes and task nodes (used to identify timeliness constraints, such as deadline constraints, approval duration threshold constraints, and resource availability time window constraints); and task dependency edges connecting different task nodes (used to identify the order of different business logic, such as pre- and post-sequential dependencies, parallel subtask dependencies, deliverable version dependencies, and blocking condition dependencies).

[0064] Specifically, the above steps involve first extracting key entity information from the task information feature vector using an entity recognition algorithm, and then mapping it to the node types defined in the schema to complete node instantiation. Specifically: if the task information feature vector contains personnel name, employee ID, and role description, it is instantiated as a personnel node; if the task information feature vector contains task name, type, and status, it is instantiated as a task node; if the task information feature vector contains time information such as deadline and approval time limit, it is instantiated as a time node; if the task information feature vector contains deliverable information such as contract attachments and document names, it is instantiated as a deliverable node; if the task information feature vector contains specific business permission identifiers or compliance levels, it is instantiated as a permission node; if the task information feature vector contains a specific process step name (such as "financial audit"), it is instantiated as an approval node.

[0065] Then, based on the semantic association strength and logical relationship between entities in the task information feature vector, edge connection logic is executed. Specifically: if an assignment semantic association is detected between a personnel node and a task node, a person in charge edge is established from the personnel node to the task node; if an output semantic association is detected between a task node and a deliverable node, a deliverable edge is established from the task node to the deliverable node; if a task node is detected to require a specific approval process, an approval dependency edge is established from the approval node to the task node; if a task node is detected to be subject to specific time constraints, a date constraint edge is established from the time node to the task node; if a pre- or post-delay or parallel dependency relationship is detected between different task nodes, a task dependency edge connecting the different task nodes is established.

[0066] Finally, the instantiated graph nodes and connected edges are aggregated to form an initial task information relationship graph containing multi-source data fusion information.

[0067] Step 3: Retrieve multiple target business logics related to the target business from the preset business logic library, and write these multiple target business logics as constraints into the initial task information relationship diagram to obtain the target task information relationship diagram.

[0068] The aforementioned business logic library is a knowledge base storing enterprise-level business operation logic, industry standards, and historical experience data. Therefore, the business logic library includes multiple business logics formed based on the business characteristics, data associations, and process dependencies of different business scenarios. For example, the module manual compilation logic in the vehicle R&D scenario involves parsing module parts composition, matching interface information, associating with vehicle models, calculating cost composition, and tracing reasons for abandonment; the supplier qualification verification logic in the parts procurement scenario involves supplier rating, associating with historical delivery records, retrieving quality breach records, and conducting compliance reviews; the process route scheduling logic in the manufacturing scenario involves constraints on the sequence of processes, matching equipment resource usage, setting time quota standards, and checking the completeness of materials; and the fault code diagnosis logic in the after-sales service scenario involves mapping fault phenomena to codes, retrieving common fault cases, recommending repair steps, and querying spare parts inventory.

[0069] Specifically, the above steps involve selecting multiple target business logics strongly related to the target business from the business logic library based on the business type of the target business, and mapping and solidifying the structured requirements, integrity standards, and compliance restrictions contained in these target business logics onto the corresponding nodes and edges of the constructed initial task information relationship graph. That is, these logics are transformed into mandatory constraints on the integrity of task node fields (such as mandatory field checks), restrictions on personnel node permissions (such as specific approval roles), standardization of approval node processes (such as mandatory financial audits), and timeliness requirements for time nodes. This corrects the initial graph, which was originally generated based solely on multimodal data and may contain missing or conflicting information, into a target task information relationship graph that conforms to industry standards and enterprise specifications and contains complete business constraint information, ensuring that the subsequently generated task objects are executable and compliant.

[0070] Step S101c: Convert the target task information relationship diagram into a standardized task.

[0071] The aforementioned standardized task includes multiple task fields, and the task fields include at least the following: task name field, task type field, personnel field, time field, and deliverable field.

[0072] Specifically, the process involves traversing all nodes and edges in the target task information relationship graph to extract key business elements. Specifically: from task nodes, extract task name, task type, and current status; from personnel nodes, extract information on responsible persons and participants; from approval nodes, extract the approval chain and compliance requirements; from deliverable nodes, extract deliverable name, format, and acceptance criteria; from time nodes, extract deadline, start time, and approval time limit; and from permission nodes, extract confidentiality level and budget restrictions.

[0073] Then, based on a pre-defined standardized task template, the extracted information is mapped into structured fields to obtain standardized tasks. This standardized task template defines the standard fields that a task must include, such as task name, task type, responsible person, participants, deadline, deliverables, acceptance criteria, approval chain, resource requirements, prerequisite tasks, and compliance tags. Simultaneously, based on the edges in the target task information relationship graph, the specific values ​​of each field are determined. For example, the value of the "Responsible Person" field is determined by the "Responsible Person" edge, the value of the "Deliverables" field is determined by the "Deliverables" edge, the "Approval Chain" list is constructed by the "Approval Dependency" edge, and the "Deadline" is determined by the "Date Constraint" edge.

[0074] Finally, the standardized tasks undergo integrity and compliance checks to obtain those that pass the checks, ensuring that they meet the mandatory field requirements and business rule constraints. Specifically: if the target task information relationship graph lacks necessary nodes (such as missing approval nodes or time nodes), or if the edge connections violate business logic (such as missing necessary approval steps), the standardized task is deemed incomplete, and a missing field prompt message is generated. Simultaneously, based on the written target business logic constraints, the standardized task is checked to see if it conforms to industry standards (such as module manuals must include component composition and interface information). If it does not conform, it is marked as non-compliant.

[0075] It should be noted that if a standardized task fails the verification, rectification suggestions can be generated and returned to the manual processing stage until a standardized task that meets the requirements is generated.

[0076] The above technical solution enables automated conversion from multi-source unstructured data to standardized task objects, effectively solving the problems of semantic conflicts from multi-source input, missing task elements, and high manual input costs in traditional systems. It ensures high standardization, high accuracy, and strong compliance of task generation, and significantly reduces the workload of manual secondary verification.

[0077] Furthermore, from multiple historical standardized tasks (i.e., executed standardized tasks) in the preset task library, at least one target standardized task with the highest similarity (i.e., the most similar business scenarios, complexity, and constraints) is selected, and the optimal workflow (i.e. the most optimized and efficient workflow) of each target standardized task is selected as the candidate workflow of the standardized task.

[0078] Furthermore, for each candidate workflow, based on the work status information of the executors at each process node in the enterprise data lake at the target time, it can be configured according to... Figure 4 The process shown determines the execution cost of candidate workflows, including:

[0079] Step S106a: Determine the compliance risk level of the candidate workflow based on the node information of each process node in the candidate workflow.

[0080] Among them, node information is the attribute data of process nodes, which includes, but is not limited to: node type (such as approval node, execution node, parallel gateway, etc.), role information of executors (such as permission information), average processing time threshold, historical compliance pass rate, and associated mandatory compliance rule identifiers (such as confidentiality approval, budget grading, and mandatory quality verification).

[0081] The aforementioned compliance risk level assesses the likelihood and severity of a candidate workflow violating the company's mandatory compliance regulations (such as unauthorized approval or missing necessary review steps) during execution. It can be a specific quantitative risk value or a risk level label (such as low risk, medium risk, high risk, or veto level).

[0082] Specifically, the compliance risk level of the aforementioned candidate workflows can be determined through the following steps:

[0083] First, the node information of each process node in the candidate workflow is mapped to a preset business rule base. The business rule base includes multiple business rules, and the business rules include at least one of the following: job permission rules (used to verify whether the executor has the job level or departmental permissions required to execute the node), node approval rules (used to verify whether a specific task type must go through a specific approval process), and process integrity rules (used to verify whether the candidate workflow contains all necessary end-to-end nodes).

[0084] If the node information of any process node within a candidate workflow does not meet the business rules in the business rule base, the compliance risk level of the candidate workflow is determined to be high.

[0085] If the node information of all process nodes in the candidate workflow meets the business rules in the business rule base, the compliance risk level of the candidate workflow is determined to be low.

[0086] Therefore, the above solution can be understood as follows: extract the key attributes of each process node in the candidate workflow (including node type, designated handler or role, pre- / post-dependencies, and business attributes), and compare them item by item with the hard compliance rules stored in the preset business rule base. For example: map the "handler role" of a node to the "position authority rule" in the business rule base to verify whether the role has the legal authority to execute the task of that node; map the "process step" of a node to the "node approval rule" to confirm whether the step is an approval node required by a specific type of task (such as financial audit or quality review) and whether the approval elements are complete; map the "sequence order" of a node to the "process integrity rule" to verify whether the workflow contains all the necessary steps stipulated by the business and without any missing or illegal skipping steps. If the verification finds that the attribute of any node conflicts with the mandatory requirements in the business rule base (such as lack of approval authority or missing mandatory approval steps), the compliance risk level of the candidate workflow will be immediately determined to be "high" and may trigger a blocking mechanism; the compliance risk level will be determined to be "low" only when the attributes of all nodes fully comply with all business rules.

[0087] By using this pre-emptive automated rule mapping and verification, potential compliance loopholes (such as unauthorized operations and missing processes) can be accurately identified and blocked before the task is actually executed, ensuring the mandatory compliance of workflow execution, reducing management and legal risks caused by violations, and providing a clear basis for subsequent manual intervention or automatic adjustment through clear level determination.

[0088] Step S106b: When the compliance risk level is low, determine the multi-dimensional risk indicators of the candidate workflow based on the personnel information of the executors of each process node in the candidate workflow and their work status information at the target time, and determine the execution cost of the candidate workflow based on the multi-dimensional risk indicators.

[0089] The personnel information mentioned above shall include at least the unit labor cost, and the work status information shall include at least the following: a list of remaining tasks to be processed within the available time window from the target time to the end time of the executor's work (if the executor's work cycle is an eight-hour work system, then the end time of the executor's work is the start time of the work and the end time after eight hours of effective work time), the estimated processing time and the estimated waiting time for each remaining task, and the multi-dimensional risk indicators shall include at least the following: real-time personnel conflict index, estimated overdue probability, estimated delivery time and estimated overdue cost.

[0090] Specifically, when the compliance risk level is low, by analyzing the personnel information (unit hourly cost) of the executors at each node in the candidate workflow and their work status information at the target time (available time window, list of tasks to be processed and their estimated processing time and estimated waiting time), the resource occupancy rate and time conflict degree of the executors when executing each node of the candidate workflow are calculated. Then, the real-time personnel conflict index (reflecting the degree of resource crowding when multiple people are working in parallel), the estimated overdue probability (combining historical processing efficiency and current load to predict the possibility of failure), the estimated delivery time (accumulating the estimated time of each node and the blocking waiting time), and the estimated overdue cost (quantified economic loss based on unit hourly cost and overdue time) are derived. Finally, the execution cost is obtained by weighting these multi-dimensional indicators to achieve the optimal path selection with low cost and high efficiency.

[0091] Specifically, the multidimensional risk indicators for the aforementioned candidate workflows can be determined through the following steps:

[0092] Step 1: Based on the list of remaining tasks and the estimated processing time of each remaining task, determine the backlog of work for the executors at each process node of the candidate workflow at the target time, and combine it with the available time window to determine the unit time productivity of the executors at each process node of the candidate workflow; calculate the ratio of backlog to unit time productivity to obtain the real-time load of the executors at each process node of the candidate workflow at the target time.

[0093] Specifically, for each process node in the candidate workflow, based on the list of remaining tasks to be processed by the executor of that process node and the estimated processing time of each remaining task, the total backlog of work for that executor at the target time (i.e., the total processing time corresponding to completing all remaining tasks) is calculated. Then, combined with the available time window between the target time and the end time of the work for that executor, the maximum available capacity per unit time is calculated. Finally, the ratio of the total backlog of work to the capacity per unit time is calculated to obtain the real-time load value reflecting the resource saturation level of that executor when executing the corresponding process node.

[0094] Step 2: Statistically analyze the number of target process nodes in the candidate workflow whose real-time load exceeds the preset saturation threshold at the target time and the overall overload level. Combine this with the skill complementarity and resource sharing relationship among the executors of the target process nodes to determine the real-time personnel conflict index of the candidate workflow.

[0095] Specifically, the number of target process nodes in the candidate workflow whose real-time load exceeds a preset saturation threshold is counted to quantify the overall scale and degree of overload of the nodes. Then, the skill complementarity (i.e., whether they can be replaced by other personnel with similar skills) and resource sharing relationship (i.e., whether they share the same physical or digital resources) among the executors of these overloaded nodes are combined to assess the intensity of collaboration obstacles and resource competition caused by executor overload, thereby comprehensively calculating the real-time personnel conflict index.

[0096] Step 3: Based on the real-time load of the executors at each process node in the candidate workflow at the target time, and the relationship between the estimated waiting time of the executors at each process node and the preset waiting time threshold, determine the estimated overdue probability of the candidate workflow.

[0097] Specifically, for each process node in the candidate workflow, the estimated waiting time for the executors of that node is compared with a preset waiting time threshold. If the estimated waiting time exceeds the preset waiting time threshold, the node is determined to have a high risk of queuing congestion. Finally, based on the degree to which the waiting time exceeds the preset waiting time threshold and the criticality of the process node in the candidate workflow, a weighted calculation is performed to determine the estimated probability of delay due to waiting in the candidate workflow.

[0098] Step 4: Based on the task dependencies of the candidate workflows, determine at least one parallel sub-path, where each sub-path includes multiple non-parallel nodes; for each sub-path, sequentially accumulate the estimated processing time and estimated waiting time for each non-parallel node to complete its remaining tasks at the target time to obtain the estimated completion time of the sub-path; compare the estimated completion times of at least one sub-path and use the latest estimated completion time as the estimated delivery time of the candidate workflow.

[0099] Specifically, based on the task dependencies of the candidate workflow, at least one parallel sub-path containing multiple non-parallel nodes is first identified. For each sub-path, the estimated processing time and estimated waiting time for each non-parallel node's executors to complete each remaining task at the target time are summed to obtain the estimated completion time of the sub-path. Finally, the estimated completion times of all parallel sub-paths are compared, and the latest time point is taken as the estimated delivery time of the entire candidate workflow to reflect the constraint of the critical path on the overall project duration.

[0100] Step 5: Determine the estimated delay between the estimated delivery time and the expected delivery time of the candidate workflow, and combine this with the unit labor cost of the executors at each process node of the candidate workflow and the preset overdue penalty coefficient of the target business to determine the estimated overdue cost of the candidate workflow.

[0101] Specifically, the difference between the estimated delivery time and the expected delivery time (a preset value) of the candidate workflow is first calculated to obtain the estimated delay duration. Then, the unit labor cost of the executors at each process node in the candidate workflow (reflecting the economic value of human resource input) and the preset overdue penalty coefficient of the target business (reflecting the additional loss weight of default or delay) are combined, and the estimated delay duration is multiplied by the unit labor cost and the penalty coefficient and accumulated to quantify the estimated overdue cost that may be caused by the overdue.

[0102] Furthermore, by weighted summation or weighted average of the multi-dimensional risk indicators, the execution cost of the candidate workflow can be obtained. The weights of each indicator can be dynamically adjusted based on the task type of the standardized task (e.g., "urgent R&D task," "routine administrative task," or "high compliance risk task"). For example, if the standardized task type is "delivery timeliness priority" (e.g., urgent R&D task), the weights of the estimated probability of delay and estimated delivery time are automatically increased, while the weights of the estimated delay cost or unit labor cost are decreased, so that the fastest completion path is prioritized in the simulation, even if its labor cost is slightly higher or there is a minor risk. If the standardized task type is "cost control priority" (e.g., routine administrative task), the weights of the estimated delay cost and unit labor cost are automatically increased, reducing sensitivity to delivery time, so as to select the lowest cost path. If the standardized task type is "resource balancing priority" (e.g., long-term project), the weight of the real-time personnel conflict index is automatically increased to avoid overloading key personnel and causing bottlenecks in subsequent tasks.

[0103] Step S106c: When the compliance risk level is high, set the execution cost of the candidate workflow to the maximum value.

[0104] Specifically, when the compliance risk level is high, regardless of whether personnel resources are available or costs are low, as long as the workflow has the risk of violating hard compliance rules (such as insufficient permissions or lack of mandatory approval), the system will determine that the path is unacceptable and directly set the execution cost to the maximum value (or infinity), thereby automatically eliminating it in the subsequent scoring and ranking, ensuring that any candidate workflow must first meet the compliance bottom line before execution, and preventing violations.

[0105] In this embodiment, a differentiated execution cost determination strategy is adopted by distinguishing compliance risk levels. When the compliance risk is high, the maximum value is directly set to trigger a "one-vote veto," ensuring that the strict compliance red line cannot be crossed and fundamentally avoiding legal, security, and major business risks. When the compliance risk is low, multi-dimensional risk indicators such as personnel workload conflict, probability of overdue delivery, delivery cycle, and economic cost are quantitatively assessed, and the execution cost of candidate workflows is determined accordingly. This breaks through the limitations of traditional workflows that rely solely on static rules or a single time dimension, achieving a technological leap from "reactive adjustment after the fact" to "proactive prediction before the fact." It achieves optimal resource allocation and efficiency maximization under the premise of compliance, ensuring the absolute security and compliance of business operations, and avoiding talent waste or process rigidity caused by excessive conservatism through refined quantitative assessment, thus achieving the best balance between security constraints and efficiency optimization.

[0106] Ultimately, the candidate workflow with the lowest execution cost is selected as the optimal workflow for the standardized task, and this optimal workflow is then sent to the real workflow engine for execution.

[0107] Therefore, the optimal workflow arrangement method described above in this application effectively avoids the risks of personnel overload, approval congestion, or resource conflicts caused by information asymmetry in manual scheduling, ensuring that the selected path is in a globally optimal state in terms of delivery cycle, human resource costs, and compliance risks, thereby significantly improving the on-time delivery rate of tasks, reducing operating costs, and realizing the refined and efficient allocation of enterprise human resources and business resources.

[0108] Furthermore, during the execution of the optimal workflow in a real workflow engine, issues such as sudden staff absences, overload, or status changes may occur, leading to process blockages and delays. To address this, this application provides a closed-loop mechanism of "real-time monitoring - dynamic evaluation - automatic replacement" to achieve runtime self-healing and dynamic optimization of the workflow. This ensures that tasks maintain efficient flow even under complex and changing human resource conditions, avoiding delivery delays caused by delayed human intervention or information asymmetry. The specific implementation process is as follows:

[0109] First, execute standardized tasks according to the optimal workflow, and periodically (e.g., every minute or every hour) poll or listen to the status information of the executors at each process node of the optimal workflow. The status information includes at least one of the following: whether they are on duty (e.g., on leave, on a business trip, or offline) and personnel load (e.g., the number of tasks remaining or the current load rate).

[0110] If the personnel executing the target process node are not on duty during the target period, or if the personnel load of the personnel executing the target process node during the target period exceeds the preset load threshold, multiple candidates are determined based on the personnel level of the personnel executing the target process node, and a recommendation score is determined based on the personnel information of each candidate. The candidate with the highest recommendation score replaces the personnel executing the target process node. The personnel information includes at least: professional matching degree, historical processing efficiency, personnel load, and historical processing experience index.

[0111] In other words, if the personnel executing the target process node are marked as "off-duty" during the target period, or if their workload exceeds a preset threshold (e.g., workload rate greater than 1), the exception handling logic is triggered. When an exception is triggered, the personnel executing the target process node are first screened based on their job level to find other on-duty personnel with the same or matching job level from the enterprise organizational structure or permission library, generating multiple candidate lists to ensure that the qualifications of the replacement personnel meet business compliance requirements (e.g., high-level tasks must be performed by personnel of the corresponding job level). Next, the personnel information of each candidate in the candidate list is obtained, and the professional matching degree (i.e., the degree of matching between the candidate and the skills and module domain required for the task), historical processing efficiency (i.e., the average time or completion rate of the candidate in handling similar tasks in the past), personnel workload (i.e., the candidate's current idleness or workload rate, prioritizing those with lower workloads to balance the workload), and historical processing experience index (i.e., the candidate's historical operation frequency or success experience value in the target business) are weighted and summed to determine the recommendation score for each candidate. Finally, by comparing the recommendation scores of all candidates, the candidate with the highest score is selected as the new executor. The executor field of the process node in the optimal workflow of the standardized task is automatically updated to reassign the process task to the new executor. At the same time, a task reminder is sent to the new executor and a personnel change notification is sent to the original executor to ensure seamless process connection and achieve runtime self-healing without manual intervention.

[0112] Therefore, this application embodiment effectively solves the pain points of traditional task management platforms, such as multi-source data conflicts, rigid and lagging processes, passive risk control, and data silos, by constructing a closed-loop management system covering the entire lifecycle from multimodal business data verification, standardized task generation, pre-deduction and optimization to dynamic self-healing and automatic rule optimization during actual execution. It achieves high automation and accuracy in task creation, proactive and refined risk control, adaptive and intelligent process scheduling, and continuous iterative optimization of business rules, significantly improving the task execution efficiency, resource utilization, and compliance control capabilities of enterprises in complex collaborative scenarios.

[0113] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0114] This embodiment also provides a task workflow orchestration apparatus for implementing the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0115] Figure 5 This is a structural block diagram of a task workflow orchestration apparatus according to one embodiment of the present invention, such as... Figure 3 As shown, the device includes:

[0116] The task generation module 501 is used to obtain the business data of the target business at the target time and generate standardized tasks to be processed based on the business data.

[0117] The first determining module 502 is used to determine at least one target standardized task that matches the standardized task from the preset task library, and to take the optimal workflow of each target standardized task as the candidate workflow of the standardized task. The preset task library includes multiple standardized tasks that have been executed and the optimal workflow corresponding to each standardized task.

[0118] The second determining module 503 is used to determine the execution cost of each candidate workflow based on the work status information of the executors of each process node in the enterprise data lake at the target time. The enterprise data lake stores at least the real-time work status information of the executors of the process nodes on multiple workflow paths.

[0119] The third determination module 504 is used to select the candidate workflow with the lowest execution cost as the optimal workflow for the standardized task.

[0120] Optionally, the task generation module 501 is further configured to generate standardized tasks to be processed based on business data using the following methods: preprocessing the business data, wherein the preprocessing includes: consistency verification, text extraction, and standardization processing, wherein the consistency verification includes at least one of the following: time conflict verification, responsible person conflict verification, approval authority conflict verification, and delivery standard conflict verification; constructing a target task information relationship diagram based on the preprocessed business data; and converting the target task information relationship diagram into standardized tasks, wherein the standardized tasks include multiple task fields, and the task fields include at least: task name field, task type field, personnel field, time field, and deliverable field.

[0121] Optionally, the task generation module 501 is further configured to construct a target task information relationship graph based on the preprocessed business data using the following method: encoding the preprocessed business data into task information feature vectors; constructing an initial task information relationship graph based on the task information feature vectors, wherein the graph nodes in the initial task information relationship graph include at least: task nodes, personnel nodes, permission nodes, approval nodes, deliverable nodes, and time nodes, and the edges in the task information relationship graph include at least: responsible person edges connecting task nodes and personnel nodes, deliverable edges connecting task nodes and deliverable nodes, approval dependency edges connecting approval nodes and task nodes, date constraint edges connecting time nodes and task nodes, and task dependency edges connecting different task nodes; retrieving multiple target business logics related to the target business from a preset business logic library, and writing the multiple target business logics as constraints into the initial task information relationship graph to obtain the target task information relationship graph, wherein the business logic library includes multiple business logics related to different businesses.

[0122] Optionally, the second determining module 503 is further configured to determine the execution cost of each candidate workflow based on the work status information of the executors at each process node in the enterprise data lake at the target time, using the following method: For each candidate workflow, the compliance risk level of the candidate workflow is determined based on the node information of each process node on the candidate workflow, wherein the node information includes at least: node type and executor permission information; When the compliance risk level is low, the multi-dimensional risk indicators of the candidate workflow are determined based on the personnel information of the executors at each process node on the candidate workflow and their work status information at the target time, and the execution cost of the candidate workflow is determined based on the multi-dimensional risk indicators, wherein the personnel information includes at least: unit labor cost, the work status information includes at least: available time window from the target time to the executor's work completion time, a list of remaining tasks to be processed, the estimated processing time and estimated waiting time for each remaining task, and the multi-dimensional risk indicators include at least: real-time personnel conflict index, estimated overdue probability, estimated delivery time and estimated overdue cost; When the compliance risk level is high, the execution cost of the candidate workflow is set to the maximum value.

[0123] Optionally, the second determining module 503 is further configured to determine the compliance risk level of the candidate workflow based on the node information of each process node in the candidate workflow using the following method: mapping the node information of each process node in the candidate workflow to a preset business rule base, wherein the business rule base includes multiple business rules, and the business rules include at least one of the following: job permission rules, node approval rules, and process integrity rules; if the node information of any process node in the candidate workflow does not meet the business rules in the business rule base, the compliance risk level of the candidate workflow is determined to be high; if the node information of all process nodes in the candidate workflow meets the business rules in the business rule base, the compliance risk level of the candidate workflow is determined to be low.

[0124] Optionally, the second determining module 503 is further configured to determine the multidimensional risk indicators of the candidate workflow based on the personnel information of the executors at each process node of the candidate workflow and their work status information at the target time, using the following methods: determining the backlog workload of the executors at each process node of the candidate workflow at the target time based on the remaining task list and the estimated processing time of each remaining task, and determining the unit-time productivity of the executors at each process node of the candidate workflow in conjunction with the available time window; calculating the ratio of the backlog workload to the unit-time productivity to obtain the real-time load of the executors at each process node of the candidate workflow at the target time; statistically analyzing the number of target process nodes in the candidate workflow whose real-time load at the target time exceeds a preset saturation threshold and the overall overload level, and determining the real-time personnel conflict index of the candidate workflow in conjunction with the skill complementarity and resource sharing relationship among the executors of the target process nodes; and determining the real-time personnel conflict index of the candidate workflow based on the candidate workflow. The estimated probability of delay for a candidate workflow is determined by considering the real-time load of personnel at each process node at the target time, the estimated waiting time of personnel at each process node, and the relationship between these values ​​and the preset waiting time threshold. Based on the task dependencies of the candidate workflow, at least one parallel sub-path is identified, where each sub-path includes multiple non-parallel nodes. For each sub-path, the estimated processing time and estimated waiting time of personnel at each non-parallel node within the sub-path for completing the remaining tasks at the target time are summed sequentially to obtain the estimated completion time of the sub-path. The estimated completion times of at least one sub-path are compared, and the latest estimated completion time is taken as the estimated delivery time of the candidate workflow. The estimated delay between the estimated delivery time and the expected delivery time of the candidate workflow is determined, and combined with the unit labor cost of personnel at each process node in the candidate workflow and the preset delay penalty coefficient of the target business, the estimated delay cost of the candidate workflow is determined.

[0125] Optionally, the device further includes an adjustment module 505, which is used to execute standardized tasks according to the optimal workflow and periodically acquire the status information of the personnel executing each process node on the optimal workflow. The status information includes at least one of the following: whether they are on duty, and their workload. If the personnel executing the target process node are not on duty during the target period, or if the workload of the personnel executing the target process node during the target period is higher than a preset workload threshold, multiple candidates are determined based on the personnel level of the personnel executing the target process node, and a recommended score is determined based on the personnel information of each candidate. The candidate with the highest recommended score replaces the personnel executing the target process node. The personnel information includes at least: professional matching degree, historical processing efficiency, personnel workload, and historical processing experience index.

[0126] It should be noted that the above modules can be implemented by software or hardware. For the latter, they can be implemented in the following ways, but are not limited to: all the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0127] According to another aspect of the embodiments of this application, a computer-readable storage medium is provided, the computer-readable storage medium including a stored program, wherein the program executes the steps in any of the above method embodiments when it is run.

[0128] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, ROMs, RAMs, portable hard drives, magnetic disks, or optical disks.

[0129] Optionally, in this embodiment, the computer program in the storage medium can be configured to perform the following steps when executed by the processor:

[0130] Step S1: Obtain the business data of the target business at the target time, and generate standardized tasks to be processed based on the business data;

[0131] Step S2: Determine at least one target standardized task that matches the standardized task from the preset task library, and take the optimal workflow of each target standardized task as the candidate workflow of the standardized task. The preset task library includes multiple standardized tasks that have been executed and the optimal workflow corresponding to each standardized task.

[0132] Step S3: Determine the execution cost of each candidate workflow based on the work status information of the executors at each process node in the enterprise data lake at the target time. The enterprise data lake stores at least the real-time work status information of the executors at the process nodes on multiple workflow paths.

[0133] Step S4: Select the candidate workflow with the lowest execution cost as the optimal workflow for the standardized task.

[0134] According to another aspect of the embodiments of this application, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor is configured to perform the steps of any of the method embodiments described above via the computer program. In an exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0135] Specific examples in this embodiment can be found in the examples described in the above embodiments and exemplary implementations, and will not be repeated here.

[0136] According to another aspect of the embodiments of this application, a computer program product is also provided, the computer program product including a computer program / instructions containing program code for performing the method shown in the flowchart.

[0137] Optionally, in this embodiment, the computer program in the above-described computer program product can be configured to perform the following steps when executed by a processor:

[0138] Step S1: Obtain the business data of the target business at the target time, and generate standardized tasks to be processed based on the business data;

[0139] Step S2: Determine at least one target standardized task that matches the standardized task from the preset task library, and take the optimal workflow of each target standardized task as the candidate workflow of the standardized task. The preset task library includes multiple standardized tasks that have been executed and the optimal workflow corresponding to each standardized task.

[0140] Step S3: Determine the execution cost of each candidate workflow based on the work status information of the executors at each process node in the enterprise data lake at the target time. The enterprise data lake stores at least the real-time work status information of the executors at the process nodes on multiple workflow paths.

[0141] Step S4: Select the candidate workflow with the lowest execution cost as the optimal workflow for the standardized task.

[0142] As an alternative implementation, the electronic device described above may exist in the form of a mobile terminal, a computer terminal, or a similar computing device. Figure 6 A block diagram of a computer terminal for implementing a task workflow orchestration method is shown. Figure 6 As shown, the computer terminal 60 may include a processor 602, a memory 604 for storing data, and a transmission device 606 for communication functions. The processor 602 may include one or more processors. For ease of explanation, Figure 6 The diagram illustrates multiple processors, including a first processor 602a, a second processor 602b, ..., an nth processor 602n. The processor 602 may be a processing device including, but not limited to, a microprocessor (MCU) or a programmable gate array (FPGA).

[0143] In addition, the computer terminal 60 may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 6 The structure shown is for illustrative purposes only and does not limit the structure of the computer terminal described above. For example, computer terminal 60 may also include... Figure 6 The more or fewer components shown, or having the same Figure 6 The different configurations shown.

[0144] Processor 602 may include one or more processing units. For example, processor 602 may include processing devices such as central processing unit (CPU), graphics processing unit (GPU), digital signal processing (DSP) chip, microcontroller unit (MCU), field-programmable gate array (FPGA), neural network processing unit (NPU), tensor processing unit (TPU), artificial intelligence (AI) type processor, etc. Different processing units may be independent components or integrated into one or more processors.

[0145] It should be noted that the aforementioned one or more processors 602 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 60. As described in the embodiments of this application, the data processing circuits serve as processor control (e.g., selection of a variable resistor termination path connected to an interface).

[0146] The memory 604 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the task workflow orchestration method in this embodiment. The processor 602 executes various functional applications and data processing by running the software programs and modules stored in the memory 604, thereby implementing the aforementioned task workflow orchestration method. The memory 604 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 604 may further include memory remotely located relative to the processor 602, and these remote memories can be connected to the computer terminal 60 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0147] The transmission device 606 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 60. In one example, the transmission device 606 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 606 may be a Radio Frequency (RF) module, used for data interaction with external networks or devices via wired and / or wireless network connections.

[0148] The display can be, for example, a liquid crystal display (LCD) or a touchscreen display (also known as a "touchscreen" or "touch display screen"). This LCD allows the user to interact with the user interface of the mobile terminal. Furthermore, the computer terminal 60 can be connected to a cursor control device and a keyboard via an input / output interface. The cursor control device (e.g., a mouse, trackball, or touchpad) is primarily used to transmit directional information and command selections to the terminal and control the movement of the cursor on the display; the keyboard is primarily used to receive letters, numbers, and other control commands input by the user, thereby enabling interaction between the user and the computer terminal 60.

[0149] It should be noted here that, in some optional embodiments, the above... Figure 6 The computer terminal shown may include hardware elements (including circuitry), software elements (including computer code stored on a computer-readable medium), or a combination of both hardware and software elements. It should be noted that... Figure 6 This is only one instance of a specific particular instance, and is intended to illustrate the types of components that may exist in the aforementioned computer terminal.

[0150] Obviously, those skilled in the art should understand that the modules or steps of this application described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. They can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, this application is not limited to any particular combination of hardware and software.

[0151] The above are merely preferred embodiments of this application and are not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

[0152] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0153] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0154] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0155] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0156] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0157] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0158] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A task workflow orchestration method, characterized in that, include: Acquire the business data of the target business at the target time, and generate standardized tasks to be processed based on the business data; At least one target standardized task matching the standardized task is determined from the preset task library, and the optimal workflow of each target standardized task is used as the candidate workflow of the standardized task. The preset task library includes multiple standardized tasks that have been executed and the optimal workflow corresponding to each standardized task. The execution cost of each candidate workflow is determined based on the work status information of the executors at each process node in the enterprise data lake at the target time. The enterprise data lake stores at least the real-time work status information of the executors at process nodes on multiple workflow paths. The candidate workflow with the lowest execution cost is selected as the optimal workflow for the standardized task.

2. The method according to claim 1, characterized in that, Based on the aforementioned business data, standardized tasks to be processed are generated, including: The business data is preprocessed, including consistency verification, text extraction and standardization, wherein the consistency verification includes at least one of the following: time conflict verification, responsible person conflict verification, approval authority conflict verification, and delivery standard conflict verification. Construct a target task information relationship diagram based on the preprocessed business data; The target task information relationship diagram is converted into the standardized task, wherein the standardized task includes multiple task fields, and the task fields include at least: task name field, task type field, personnel field, time field and deliverable field.

3. The method according to claim 2, characterized in that, Based on the preprocessed business data, a target task information relationship diagram is constructed, including: The preprocessed business data is encoded into a task information feature vector; An initial task information relationship graph is constructed based on the task information feature vector. The graph nodes in the initial task information relationship graph include at least: task nodes, personnel nodes, permission nodes, approval nodes, deliverable nodes, and time nodes. The edges in the task information relationship graph include at least: responsible person edges connecting task nodes and personnel nodes, deliverable edges connecting task nodes and deliverable nodes, approval dependency edges connecting approval nodes and task nodes, date constraint edges connecting time nodes and task nodes, and task dependency edges connecting different task nodes. The system retrieves multiple target business logics related to the target business from the preset business logic library, and writes these multiple target business logics as constraints into the initial task information relationship graph to obtain the target task information relationship graph. The business logic library includes multiple business logics related to different businesses.

4. The method according to claim 1, characterized in that, The execution cost of each candidate workflow is determined based on the work status information of the executors at each process node in the enterprise data lake at the target time, including: For each candidate workflow, the compliance risk level of the candidate workflow is determined based on the node information of each process node on the candidate workflow, wherein the node information includes at least: node type and permission information of the executor; When the compliance risk level is low, based on the personnel information of the executors at each process node of the candidate workflow and their work status information at the target time, a multi-dimensional risk indicator for the candidate workflow is determined, and the execution cost of the candidate workflow is determined based on the multi-dimensional risk indicator. The personnel information includes at least: unit labor cost, and the work status information includes at least: the available time window from the target time to the work completion time of the executor, the list of remaining tasks to be processed, the estimated processing time and the estimated waiting time for each remaining task, and the multi-dimensional risk indicator includes at least: real-time personnel conflict index, estimated overdue probability, estimated delivery time and estimated overdue cost. If the compliance risk level is high, the execution cost of the candidate workflow will be set to the maximum value.

5. The method according to claim 4, characterized in that, Based on the node information of each process node in the candidate workflow, the compliance risk level of the candidate workflow is determined, including: The node information of each process node in the candidate workflow is mapped to a preset business rule base, wherein the business rule base includes multiple business rules, and the business rules include at least one of the following: job permission rules, node approval rules, and process integrity rules. If the node information of any process node in the candidate workflow does not meet the business rules in the business rule base, the compliance risk level of the candidate workflow is determined to be high. If the node information of all process nodes in the candidate workflow meets the business rules in the business rule base, the compliance risk level of the candidate workflow is determined to be low.

6. The method according to claim 4, characterized in that, Based on the personnel information of the executors at each process node of the candidate workflow and their work status information at the target time, multidimensional risk indicators of the candidate workflow are determined, including: Based on the remaining task list and the estimated processing time of each remaining task, the backlog of work for the executors at each process node of the candidate workflow at the target time is determined, and combined with the available time window, the unit time productivity of the executors at each process node of the candidate workflow is determined; the ratio of the backlog of work to the unit time productivity is calculated to obtain the real-time load of the executors at each process node of the candidate workflow at the target time. The number of target process nodes in the candidate workflow whose real-time load at the target time exceeds a preset saturation threshold and the overall overload level are statistically analyzed. Combined with the skill complementarity and resource sharing relationship among the executors of the target process nodes, the real-time personnel conflict index of the candidate workflow is determined. Based on the real-time load of the executors at each process node of the candidate workflow at the target time, and the relationship between the estimated waiting time of the executors at each process node and the preset waiting time threshold, the estimated overdue probability of the candidate workflow is determined. Based on the task dependencies of the candidate workflows, at least one parallel sub-path is determined, wherein the sub-path includes multiple non-parallel nodes; for each sub-path, the estimated processing time and estimated waiting time for the executors of each non-parallel node in the sub-path to complete each remaining task at the target time are sequentially accumulated to obtain the estimated completion time of the sub-path; the estimated completion times of the at least one sub-path are compared, and the latest estimated completion time is taken as the estimated delivery time of the candidate workflow; Determine the estimated delay between the estimated delivery time and the expected delivery time of the candidate workflow, and combine the unit labor cost of the executors at each process node of the candidate workflow and the preset overdue penalty coefficient of the target business to determine the estimated overdue cost of the candidate workflow.

7. The method according to claim 1, characterized in that, After selecting the candidate workflow with the lowest execution cost as the optimal workflow for the standardized task, the method further includes: The standardized task is executed according to the optimal workflow, and the status information of the executors at each process node of the optimal workflow is periodically obtained, wherein the status information includes at least one of the following: whether they are on duty, and the workload of the personnel. If the personnel executing the target process node are not on duty during the target period, or if the personnel load of the personnel executing the target process node during the target period exceeds a preset load threshold, multiple candidates are determined based on the personnel rank of the personnel executing the target process node, and a recommendation score is determined for each candidate based on the personnel information of each candidate. The candidate with the highest recommendation score replaces the personnel executing the target process node. The personnel information includes at least: professional matching degree, historical processing efficiency, personnel load, and historical processing experience index.

8. A task workflow orchestration device, characterized in that, include: The task generation module is used to acquire the business data of the target business at the target time and generate standardized tasks to be processed based on the business data. The first determining module is used to determine at least one target standardized task that matches the standardized task from a preset task library, and to take the optimal workflow of each target standardized task as the candidate workflow of the standardized task. The preset task library includes multiple standardized tasks that have been executed and the optimal workflow corresponding to each standardized task. The second determining module is used to determine the execution cost of each candidate workflow based on the work status information of the executors of each process node in the enterprise data lake at the target time, wherein the enterprise data lake stores at least the real-time work status information of the executors of process nodes on multiple workflow paths. The third determining module is used to select the candidate workflow with the lowest execution cost as the optimal workflow for the standardized task.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein the computer program is configured to execute the task workflow orchestration method according to any one of claims 1 to 7 when run on a computer or processor.

10. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program and the processor being configured to run the computer program to perform the task workflow orchestration method according to any one of claims 1 to 7.

Images