Enterprise-level, full-scenario, one-stop intelligent collaborative service platform and methodology for workflow

Abstract

This invention relates to the field of enterprise digital collaboration and workflow management, specifically to an enterprise-level, full-scenario, one-stop intelligent workflow collaboration service platform and method. The platform includes a full-scenario workflow semantic modeling module, a cross-system collaboration adaptation engine, a dynamic permission game control module, a workflow intelligent scheduling and optimization module, and a one-stop collaborative interaction front-end module. The method of this invention achieves closed-loop management of the entire enterprise workflow chain through core steps such as standardized semantic modeling across all business scenarios, establishment of cross-system collaborative links, dynamic permission allocation and conflict resolution, intelligent workflow scheduling optimization, and one-stop collaborative service output. Through core algorithms such as adaptation scoring, semantic similarity calculation, multi-role permission game, and constrained task scheduling, this invention can be widely applied to workflow collaboration management in various enterprise full-business scenarios, significantly improving the efficiency of enterprise business process execution and the level of digital management.

Description

Technical Field

[0001] This invention relates to the field of enterprise digital collaboration and workflow management, specifically to an enterprise-level, full-scenario, one-stop intelligent workflow collaboration service platform and method. Background Technology

[0002] With the deepening of enterprise digital transformation, the online and collaborative management of internal business processes has become a core tool for enterprises to improve operational efficiency. As the core carrier for the flow of the entire business process, cross-role collaboration, and cross-system linkage, the workflow collaboration service platform directly determines the execution efficiency and management level of enterprise business processes through its scenario adaptability, cross-system collaboration capability, permission control capability, and scheduling optimization capability.

[0003] In the prior art, Chinese invention patent application CN115907690A discloses an enterprise team collaboration system based on microservices (SaaS). This system specifically includes a business service unit, a process identification unit, an automatic approval unit, and a process instance unit. Its core technical solution is as follows: the business service unit initiates the workflow and calls the workflow engine to achieve automatic approval and system automation; after the workflow is started, the process identification unit obtains information about the process initiator, approver, business scenario, and process description through business process parameters and sends it to the automatic approval unit; the automatic approval unit, based on an artificial intelligence automatic approval algorithm, calculates the probability prediction value of different process selections by collecting historical approval data, outputs approval suggestions to the approver, and completes the automatic approval after the approver's confirmation; the process instance unit generates corresponding process instances based on the process definition, and also supports the signing, retrospective, and flow control of process instances. This prior art mainly addresses the problems of low efficiency and susceptibility to decision-making errors in manual approval processes by using a probabilistic statistical artificial intelligence algorithm to achieve intelligent suggestions and automatic approval of the approval process, thereby improving the intelligence of the workflow engine in the approval stage.

[0004] However, the aforementioned existing technologies still have the following technical shortcomings: First, the technical solution only focuses on optimizing the approval process, without establishing a standardized semantic modeling system for workflows across all business scenarios. It cannot achieve accurate matching, rapid adaptation, and custom configuration of workflow templates for multiple business scenarios. Customized process templates need to be developed separately for different business scenarios, resulting in poor scenario compatibility, high template development costs, and long adaptation cycles. Second, the technical solution only achieves basic integration with conventional business systems such as OA and finance, without setting up a core architecture for cross-system collaborative adaptation. It cannot complete semantic mapping of heterogeneous system interfaces, bidirectional data format conversion, and exception handling, making it difficult to break down data silos and collaborative gaps between different business systems within the enterprise. Cross-system process linkage capabilities and operational stability are insufficient. Third, the technical solution does not systematically design permission control for the entire workflow process. It only completes basic process flow based on the process initiator and approver, failing to achieve dynamic permission allocation based on the business responsibility boundaries of collaborative roles and the security level of process nodes. It also cannot resolve permission conflicts in real time, resulting in insufficient flexibility in permission control and difficulty in balancing business collaboration efficiency and data security control requirements. Fourth, this technical solution does not involve intelligent scheduling and collaborative resource optimization allocation of workflow tasks. It cannot complete the intelligent decomposition and global optimal scheduling of complex process tasks. The process execution efficiency and resource allocation rely on manual configuration, resulting in low resource utilization and difficulty in meeting the collaborative scheduling needs of enterprise-level large-scale parallel workflows. Fifth, this technical solution adopts a decentralized functional unit design, failing to form a closed-loop integrated architecture for the entire workflow from template construction, system integration, permission control, task scheduling to terminal interaction. It does not set up a unified one-stop collaborative interaction entry point, requiring users to switch between multiple systems and entry points. This results in problems such as non-closed-loop process control, low operational efficiency, insufficient adaptability to all scenarios, and inadequate technology reusability.

[0005] Based on the technical deficiencies of the existing technologies, there is an urgent need to develop an enterprise-level, one-stop intelligent workflow collaboration solution with full-scenario semantic modeling, seamless cross-system collaboration, dynamic permission control, and intelligent scheduling optimization capabilities. This solution aims to address the technical problems of poor workflow platform scenario compatibility, high barriers to cross-system collaboration, insufficient flexibility in permission control, low resource scheduling efficiency, and non-closed-loop full-process management in the existing technologies. Summary of the Invention

[0006] The purpose of this invention is to provide an enterprise-level, full-scenario, one-stop intelligent collaborative service platform and method for workflows, in order to solve the problems mentioned in the background art, such as poor workflow scenario compatibility, high barriers to cross-system collaboration, insufficient flexibility in permission control, and low efficiency in resource scheduling.

[0007] To achieve the above objectives, the present invention provides the following technical solution: An enterprise-level, full-scenario, one-stop intelligent collaborative workflow service platform, including a full-scenario workflow semantic modeling module, a cross-system collaborative adaptation engine, a dynamic permission game control module, a workflow intelligent scheduling and optimization module, and a one-stop collaborative interaction front-end module; The full-scenario workflow semantic modeling module is used to perform standardized semantic modeling of process nodes, participating roles, data interaction rules, and execution constraints in all business scenarios of an enterprise, and generate a unified semantic workflow template compatible with multiple business scenarios. The cross-system collaborative adaptation engine is used to connect with heterogeneous business systems within an enterprise, complete interface semantic mapping, bidirectional data format conversion and instruction pass-through, and realize seamless linkage and data interoperability of cross-system process nodes. The dynamic permission game control module is used to perform dynamic permission allocation and real-time resolution of permission conflicts based on the business responsibility boundaries of collaborative roles and the security level attributes of process nodes. The workflow intelligent scheduling and optimization module is used to collect execution data of the entire workflow chain, and to complete the dynamic adjustment of process nodes, intelligent decomposition of parallel tasks, and optimal allocation of collaborative resources. The one-stop collaborative interaction front-end module is used to provide a unified interactive entry point for all users to initiate processes, approve workflows, process tasks, track progress, and visualize data.

[0008] Preferably, the full-scenario workflow semantic modeling module is further used to calculate the compatibility score between the workflow template to be matched and the target scenario based on the feature parameters of the target business scenario. The formula for calculating the compatibility score is as follows: ; In the formula, Rate the fit. These are preset weights for node matching, role matching, data matching, and latency matching, respectively. ; For process node matching degree, To participate in role matching, For data interaction rule matching degree, The maximum allowable execution latency for the target scenario. The baseline execution delay for the template to be matched. This is the preset maximum delay threshold.

[0009] Preferably, the cross-system collaborative adaptation engine incorporates a domain ontology semantic mapping unit. This unit is used to calculate the semantic similarity between heterogeneous system interface parameters and unified semantic template parameters. The formula for calculating the semantic similarity is as follows: ; In the formula, Let be the semantic similarity between parameter a and parameter b. The preset weighting coefficients, Let cosine similarity be the word vectors corresponding to parameters a and b. Let be the least common ancestor node of parameters a and b in the domain ontology tree. The depth of the least common ancestor node. This represents the maximum depth of the domain ontology tree.

[0010] Preferably, the dynamic permission game control module is used to construct a multi-role non-cooperative permission game model, solve for the Nash equilibrium solution to achieve optimal permission allocation, and the utility function of the game model is: ; In the formula, Let i be the utility function for role i. The permission values ​​requested for role i. Strategies for assigning permissions to other roles. The business contribution coefficient for role i. The total permission quota for process nodes. The business revenue value of the process node. For role i, the risk factor The risk loss value for permission leakage at process nodes; the optimal permission allocation is to satisfy... Nash equilibrium solution .

[0011] Preferably, the workflow intelligent scheduling and optimization module is used to construct a constrained task scheduling optimization model and solve for the optimal task allocation scheme. The objective function of the optimization model is: ; The constraints are: ; In the formula, Let m be the total execution time of the workflow, m be the total number of subtasks after decomposition, and n be the total number of available collaborative resource nodes. Let j be the execution time of subtask j on resource node k. This is a 0-1 decision variable; a value of 1 indicates that subtask j is assigned to resource node k for execution. This represents the resource consumption of subtask j. Let k be the maximum available resource quantity for resource node k. This represents the maximum allowed execution time for subtask j.

[0012] Preferably, the full-scenario workflow semantic modeling module has a built-in template customization configuration unit. The template customization configuration unit is used to classify, manage, and customize the nodes of the unified semantic workflow template. The node classification includes manually executed nodes, automatically executed nodes, condition judgment nodes, parallel convergence nodes, abnormal termination nodes, and rollback nodes. The custom configuration includes adding and deleting nodes within the template, adjusting the dependency topology between nodes, configuring branch conditions for condition judgment nodes, and configuring rollback rules for rollback nodes; the dimensions of the branch condition configuration include business data field thresholds, node execution status identifiers, role operation result codes, and cross-system data feedback values. The rollback rule configuration includes rollback trigger conditions, rollback level range, rollback data reverse synchronization rules, and rollback operation permission verification rules; the template customization configuration unit is also used to perform semantic consistency verification, node dependency relationship topology verification, and compliance constraint verification on the configured template to generate customized workflow templates adapted to specific business scenarios.

[0013] Preferably, the cross-system collaborative adaptation engine has a built-in interface monitoring and exception handling unit. The interface monitoring and exception handling unit is used to perform real-time monitoring of the entire lifecycle of the interfaces of the heterogeneous business systems being connected. The monitoring dimensions include interface call response time, interface return code status, data transmission integrity, link connectivity, and interface concurrent capacity. The interface monitoring and exception handling unit has preset exception judgment rules, which include a threshold for consecutive timeouts, a list of non-successful return code types, a data packet loss rate threshold, and a threshold for the duration of link interruption. When the monitored data triggers the exception judgment rules, the interface monitoring and exception handling unit automatically triggers the preset retry mechanism and graded degradation strategy. The retry mechanism includes gradient configuration of the number of retryes, exponential backoff parameter configuration of the retry interval, idempotency verification rules for retry requests, and backup route switching rules for retry links; the tiered degradation strategy includes tiered triggering conditions for full process degradation, branch process degradation, and node function degradation, as well as bypass rules for degraded process links, local caching rules for interactive data, and data resending rules after link recovery; the interface monitoring and exception handling unit is also used to embed and synchronously store the execution logs and context data of the entire interface call process.

[0014] On the other hand, the present invention provides an enterprise-level, full-scenario, one-stop intelligent workflow collaboration service method, applied to the aforementioned enterprise-level, full-scenario, one-stop intelligent workflow collaboration service platform, comprising the following steps: S1. Standardize semantic modeling of process nodes, participating roles, data interaction rules, and execution constraints for all business scenarios of an enterprise, and generate a unified semantic workflow template compatible with multiple business scenarios; S2. Connect with heterogeneous business systems within the enterprise, complete interface semantic mapping, bidirectional data format conversion and instruction pass-through, and establish cross-system collaborative links; S3. Based on the business responsibility boundaries of collaborative roles and the security level attributes of process nodes, dynamic permission allocation and real-time resolution of permission conflicts are performed. S4. Collect execution data across the entire workflow, and perform dynamic adjustment of process nodes, intelligent decomposition of parallel tasks, and optimal allocation of collaborative resources; S5. Through a unified interactive portal, it provides a one-stop collaborative service for all users, including process initiation, approval workflow, task processing, progress tracking, and data visualization.

[0015] Preferably, step S1 specifically includes the following sub-steps: S101. Collect the original feature parameters of the target business scenario. The original feature parameters include the number of process nodes and node type labels, the number of participating roles and role rights and responsibilities labels, the field types and transmission rules of data interaction, and the time delay constraint parameters and compliance constraint parameters of process execution. S102. Standardize the collected raw feature parameters into semantics to generate a standardized semantic feature vector corresponding to the target business scenario; S103. Traverse the pre-built unified semantic workflow template library, and extract the corresponding template semantic feature vector for each workflow template to be matched in the library; S104. Based on the fit score calculation formula, calculate the fit score between the standardized semantic feature vector of the target business scenario and the template semantic feature vector of each workflow template to be matched in turn. S105. Sort all workflow templates to be matched in descending order of fit score, filter out workflow templates with fit scores higher than the preset fit threshold, and generate a candidate template set. S106. For each workflow template in the candidate template set, perform node dependency topology verification, role and responsibility boundary matching verification, and data interaction rule compliance verification in sequence, and output the candidate workflow templates that pass all verifications; S107. Based on the validated candidate workflow templates, complete the standardized semantic modeling of all business scenarios and generate a unified semantic workflow template compatible with multiple business scenarios.

[0016] Preferably, step S3 specifically includes the following sub-steps: S301. Collect the basic attribute parameters of the current workflow process node. The basic attribute parameters include the node security level attribute, the node total permission quota, the node business revenue value, and the node permission leakage risk loss value. S302. Collect the attribute parameters of all associated collaborative roles in this process node. The attribute parameters include role business responsibility boundary label, role business contribution coefficient, and role risk coefficient. S303. Taking each associated collaborative role as the game participant and the permission allocation value of each collaborative role as the decision variable, based on the collected basic attribute parameters and role attribute parameters, and using the utility function, a multi-role non-cooperative permission game model is constructed. S304. Using an iterative approximation algorithm, solve the pure strategy Nash equilibrium solution of the constructed multi-role non-cooperative permission game model to obtain the initial optimal permission allocation value for each cooperating role; S305. For the initial optimal permission allocation values ​​of each collaborative role, perform permission boundary compliance verification, remove permission values ​​that exceed the legal rights and responsibilities boundaries of the role and the security control scope of the process node, generate the final permission allocation scheme and issue it for execution; S306. Collect data on role permission application changes, process node security level adjustments, and collaborative role additions or exits during the process in real time. When the collected data triggers the preset dynamic update conditions, repeat steps S301 to S305 to complete dynamic permission updates and real-time resolution of permission conflicts.

[0017] Compared with the prior art, the beneficial effects of the present invention are: (1) This invention constructs an enterprise-level business process domain ontology library, performs standardized semantic modeling of process elements in all business scenarios, and achieves accurate matching between workflow templates and target business scenarios by combining the adaptation score calculation formula. At the same time, it supports custom configuration and multi-dimensional verification of template nodes, and can quickly generate customized templates adapted to subdivided business scenarios, avoiding the defects of high development costs and long adaptation cycles caused by the repeated development of process templates for different business scenarios in traditional platforms. Through the semantic similarity calculation formula of semantic mapping unit, the automatic mapping and matching of interface parameters of heterogeneous systems and unified semantic template parameters is realized. With the interface full life cycle monitoring and exception handling mechanism, seamless linkage and data interoperability of cross-system process nodes are realized, effectively breaking down the data silos and collaboration gaps between different business systems within the enterprise, and improving the accuracy and operational stability of cross-system workflow collaboration.

[0018] (2) This invention constructs a multi-role non-cooperative permission game model. Based on the attribute parameters of process nodes and collaborative roles, the optimal allocation of permissions is achieved by solving the Nash equilibrium solution through the utility function. With the permission boundary compliance verification and dynamic update mechanism, the dynamic adjustment of permissions and conflict resolution can be completed in real time. This avoids the defects of traditional static permission control that cannot adapt to the dynamic changes of the process, and also takes into account the business efficiency and data security control requirements of workflow execution. By constructing a task scheduling optimization model with constraints, the intelligent decomposition of complex process tasks and the global optimal allocation of collaborative resources are realized. Under the conditions of resource constraints and time delay constraints, the total execution time of the workflow can be minimized. This solves the problems of uneven resource allocation and low process execution efficiency in traditional task scheduling methods, and effectively improves the utilization rate of collaborative resources and the overall execution efficiency of the workflow.

[0019] (3) The various functional modules of the present invention communicate data in real time and coordinate execution logic, realizing one-stop management of the entire workflow from template construction, system docking, permission control, task scheduling to terminal interaction. Through a unified collaborative interaction front-end, it provides a standardized service entry for all users, avoiding the problems of process execution gaps and low operation efficiency caused by switching multiple systems and multiple entry operations. At the same time, the core algorithm model and architecture design of the present invention have good universality and scalability. It can be adapted without large-scale architecture transformation for specific industries or specific business scenarios. It has strong engineering implementation value and technology reusability, providing a creative and complete technical solution for enterprise-level full-scenario workflow collaborative management. Attached Figure Description

[0020] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are explained in detail together with the embodiments of the invention, but do not constitute a limitation thereof.

[0021] Figure 1 This is a block diagram of the enterprise-level, full-scenario, one-stop intelligent collaborative service platform for workflows, as described in this invention. Figure 2 This is a flowchart of the enterprise-level, full-scenario, one-stop intelligent collaborative service method for workflows according to the present invention. Figure 3 This is a detailed diagram of the full-scenario workflow semantic modeling module of the present invention; Figure 4 This is a detailed diagram of the cross-system collaborative adaptation engine and dynamic permission control of the present invention; Figure 5 This is a detailed diagram of the intelligent scheduling and optimization module for workflow in this invention. Detailed Implementation

[0022] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. 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 of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0023] like Figure 1 , Figures 3-5 As shown, the enterprise-level full-scenario one-stop intelligent collaborative service platform and method provided by the present invention achieves standardized management and control of enterprise full-business scenario workflows, seamless cross-system linkage, dynamic permission adaptation and optimal resource scheduling through an integrated architecture of full-scenario semantic modeling, cross-system collaborative adaptation, dynamic permission game control, intelligent scheduling optimization and one-stop interaction. It solves the technical problems of poor workflow scenario compatibility, high cross-system collaboration barriers, insufficient flexibility of permission control and low resource scheduling efficiency in the prior art.

[0024] The full-scenario workflow semantic modeling module completes standardized semantic modeling of process elements across all enterprise business scenarios, generating a unified semantic workflow template compatible with multiple business scenarios. The module first collects process elements from all enterprise business scenarios, including process nodes, participating roles, data interaction rules, and execution constraints. It then performs ontological semantic annotation on the collected process elements, constructs an enterprise-level business process domain ontology library, and completes standardized semantic encoding of the process elements based on this library, generating a unified semantic workflow template.

[0025] The full-scenario workflow semantic modeling module calculates the fit score between the workflow template to be matched and the target scenario based on the feature parameters of the target business scenario. The formula for calculating the fit score is as follows: ; In the formula, Rate the fit. , , , These are preset weights for node matching, role matching, data matching, and latency matching, respectively. . For process node matching degree, To participate in role matching, For data interaction rule matching degree, The maximum allowable execution latency for the target scenario. The baseline execution delay for the template to be matched. This is the preset maximum delay threshold.

[0026] The full-scenario workflow semantic modeling module includes a built-in template customization configuration unit. This unit categorizes and customizes the nodes of a unified semantic workflow template. Node categories include manually executed nodes, automatically executed nodes, conditional judgment nodes, parallel convergence nodes, abnormal termination nodes, and rollback nodes. Custom configurations include adding and deleting nodes within the template, adjusting the dependency topology between nodes, configuring branch conditions for conditional judgment nodes, and configuring rollback rules for rollback nodes. Dimensions of branch condition configuration include business data field thresholds, node execution status identifiers, role operation result encoding, and cross-system data feedback values. Rollback rule configuration includes rollback trigger conditions, rollback level range, rollback data reverse synchronization rules, and rollback operation permission verification rules. The template customization configuration unit performs semantic consistency verification, node dependency topology verification, and compliance constraint verification on the configured template, generating customized workflow templates adapted to specific business scenarios.

[0027] The cross-system collaboration adaptation engine facilitates the integration of heterogeneous business systems within an enterprise, enabling interface semantic mapping, bidirectional data format conversion, and command pass-through. This achieves seamless linkage and data interoperability across cross-system process nodes. The engine supports adaptation to multiple protocol types, including RESTful interfaces, WebService interfaces, RPC interfaces, and direct database connections, facilitating the access and establishment of links between heterogeneous systems.

[0028] The cross-system collaborative adaptation engine has a built-in domain ontology semantic mapping unit. This unit calculates the semantic similarity between heterogeneous system interface parameters and unified semantic template parameters. The formula for calculating semantic similarity is as follows: ; In the formula, Let be the semantic similarity between parameter a and parameter b. The preset weighting coefficients, Let cosine similarity be the word vectors corresponding to parameters a and b. Let be the least common ancestor node of parameters a and b in the domain ontology tree. The depth of the least common ancestor node. This represents the maximum depth of the domain ontology tree. The domain ontology semantic mapping unit filters parameter pairs with semantic similarity higher than a preset threshold, completing the automatic mapping and matching of interface parameters, and achieving seamless adaptation between heterogeneous system interfaces and unified semantic templates.

[0029] The cross-system collaborative adaptation engine incorporates an interface monitoring and exception handling unit. This unit performs real-time monitoring of the entire lifecycle of interfaces of heterogeneous business systems it interfaces with. Monitoring dimensions include interface call response time, interface return code status, data transmission integrity, link connectivity, and interface concurrency capacity. The interface monitoring and exception handling unit has preset exception judgment rules, including thresholds for consecutive timeouts, a list of non-successful return code types, data packet loss rate thresholds, and link interruption duration thresholds. When monitored data triggers an exception judgment rule, the interface monitoring and exception handling unit automatically triggers preset retry mechanisms and tiered degradation strategies. The retry mechanism includes gradient configuration of retry counts, exponential backoff parameter configuration for retry intervals, idempotency verification rules for retry requests, and backup route switching rules for retry links. The tiered degradation strategy includes tiered triggering conditions for full-process degradation, branch process degradation, and node function degradation, as well as rules for bypassing degraded process links, local caching of interactive data, and data resending rules after link recovery. The interface monitoring and exception handling unit also includes full-process execution logs for interface calls and synchronizes them with context data.

[0030] The dynamic permission game control module performs dynamic permission allocation and real-time resolution of permission conflicts based on the business responsibility boundaries of collaborative roles and the security level attributes of process nodes. The module collects basic attribute parameters of process nodes and attribute parameters of collaborative roles, constructs a multi-role non-cooperative permission game model, and solves for the Nash equilibrium to achieve optimal permission allocation.

[0031] The utility function of a multi-role, non-cooperative power game model is: ; In the formula, Let i be the utility function for role i. The permission values ​​requested for role i. Strategies for assigning permissions to other roles. The business contribution coefficient for role i. The total permission quota for process nodes. The business revenue value of the process node. For role i, the risk factor This represents the risk and potential loss due to permission leakage at process nodes. The optimal permission allocation is to satisfy... Nash equilibrium solution .

[0032] The dynamic permission game control module performs permission boundary compliance checks on the initially optimal permission allocation values ​​obtained from the solution, eliminates permission values ​​that exceed the boundaries of role business responsibilities and the security control scope of process nodes, generates the final permission allocation scheme, and issues it for execution. The module also collects data in real time on role permission application changes, process node security level adjustments, and the addition or removal of collaborative roles during process execution. When the collected data triggers preset dynamic update conditions, the game model is reconstructed and solved, completing dynamic permission updates and real-time resolution of permission conflicts.

[0033] The workflow intelligent scheduling and optimization module collects execution data from the entire workflow chain, enabling dynamic adjustment of process nodes, intelligent decomposition of parallel tasks, and optimal allocation of collaborative resources. Based on the workflow's process topology, the module breaks down complex process tasks into multiple independent atomic subtasks, collects the resource requirement parameters of the subtasks and the attribute parameters of available collaborative resource nodes, constructs a constrained task scheduling optimization model, and solves for the optimal task allocation scheme.

[0034] The objective function of the constrained task scheduling optimization model is: ; The constraints are: ; In the formula, Let m be the total execution time of the workflow, m be the total number of subtasks after decomposition, and n be the total number of available collaborative resource nodes. Let j be the execution time of subtask j on resource node k. It is a 0-1 decision variable. When the value is 1, it means that subtask j is assigned to resource node k for execution. This represents the resource consumption of subtask j. Let k be the maximum available resource quantity for resource node k. This represents the maximum allowed execution time for subtask j.

[0035] The workflow intelligent scheduling and optimization module uses genetic algorithms or branch and bound algorithms to solve the task scheduling optimization model, obtaining the globally optimal task allocation scheme. Based on the optimal allocation scheme, it completes the matching and scheduling of subtasks and collaborative resource nodes. The workflow intelligent scheduling and optimization module collects node execution status data, resource node load data, and task execution latency data in real time during the workflow execution process. When the data triggers the preset optimization trigger conditions, it reconstructs and solves the optimization model, completing the dynamic adjustment of process nodes and the dynamic reallocation of resources.

[0036] The one-stop collaborative interaction front-end module provides a unified interaction entry point for all user roles, enabling a seamless service for workflow initiation, approval processes, task processing, progress tracking, and data visualization. It supports multi-terminal adaptation across PCs, mobile devices, and tablets, generating personalized interfaces and function menus based on user roles and permission configurations. The module offers features such as quick selection of workflow templates, customizable workflow initiation forms, online submission of approval opinions, task reminders, full-link visualization of workflow progress, and multi-dimensional statistical analysis of workflow execution data. The front-end module maintains real-time data synchronization with backend modules, ensuring real-time response and status updates for the entire workflow operation.

[0037] like Figure 2 As shown, the enterprise-level full-scenario one-stop intelligent workflow collaboration service method of the present invention is applied to the above-mentioned enterprise-level full-scenario one-stop intelligent workflow collaboration service platform, and the specific implementation steps are as follows: Step S1. Standardize semantic modeling of process nodes, participating roles, data interaction rules, and execution constraints for all business scenarios of the enterprise, and generate a unified semantic workflow template that is compatible with multiple business scenarios.

[0038] Step S1 specifically includes the following sub-steps.

[0039] S101. Collect the original feature parameters of the target business scenario. The original feature parameters include the number of process nodes and node type labels, the number of participating roles and role rights and responsibilities labels, the field types and transmission rules of data interaction, and the time delay constraint parameters and compliance constraint parameters of process execution.

[0040] S102. Standardize the collected raw feature parameters into semantics to generate a standardized semantic feature vector corresponding to the target business scenario.

[0041] S103. Traverse the pre-built unified semantic workflow template library, and extract the corresponding template semantic feature vector for each workflow template to be matched in the library.

[0042] S104. Based on the above adaptation score calculation formula, calculate the adaptation score between the standardized semantic feature vector of the target business scenario and the template semantic feature vector of each workflow template to be matched in turn.

[0043] S105. Sort all workflow templates to be matched in descending order of fit score, filter out workflow templates with fit scores higher than the preset fit threshold, and generate a candidate template set.

[0044] S106. For each workflow template in the candidate template set, perform node dependency topology verification, role and responsibility boundary matching verification, and data interaction rule compliance verification in sequence, and output the candidate workflow templates that pass all verifications.

[0045] S107. Based on the validated candidate workflow templates, complete the standardized semantic modeling of all business scenarios and generate a unified semantic workflow template compatible with multiple business scenarios.

[0046] Step S2. Connect with heterogeneous business systems within the enterprise, complete interface semantic mapping, bidirectional data format conversion and instruction pass-through, and establish a cross-system collaborative link.

[0047] Step S3. Based on the business responsibility boundaries of collaborative roles and the security level attributes of process nodes, perform dynamic permission allocation and real-time resolution of permission conflicts.

[0048] Step S3 specifically includes the following sub-steps.

[0049] S301. Collect the basic attribute parameters of the current workflow process node. The basic attribute parameters include the node security level attribute, the node total permission quota, the node business benefit value, and the node permission leakage risk loss value.

[0050] S302. Collect the attribute parameters of all associated collaborative roles in this process node. The attribute parameters include the role's business responsibility boundary label, the role's business contribution coefficient, and the role's risk coefficient.

[0051] S303. Taking each associated collaborative role as the game participant and the permission allocation value of each collaborative role as the decision variable, based on the collected basic attribute parameters and role attribute parameters, and using the above utility function, a multi-role non-cooperative permission game model is constructed.

[0052] S304. Using an iterative approximation algorithm, solve the pure policy Nash equilibrium solution of the constructed multi-role non-cooperative permission game model to obtain the initial optimal permission allocation value for each cooperating role.

[0053] S305. For the initial optimal permission allocation values ​​of each collaborative role, perform permission boundary compliance verification, remove permission values ​​that exceed the legal rights and responsibilities boundaries of the role and the security control scope of the process node, generate the final permission allocation scheme and issue it for execution.

[0054] S306. Collect data on role permission application changes, process node security level adjustments, and collaborative role additions or exits during the process in real time. When the collected data triggers the preset dynamic update conditions, repeat steps S301 to S305 to complete dynamic permission updates and real-time resolution of permission conflicts.

[0055] Step S4. Collect execution data of the entire workflow to achieve dynamic adjustment of process nodes, intelligent decomposition of parallel tasks, and optimal allocation of collaborative resources.

[0056] Step S5. Through a unified interactive portal, provide all users with a one-stop collaborative service for process initiation, approval flow, task processing, progress tracking, and data visualization.

[0057] Example 1: Collaborative Procurement Process for Large Discrete Manufacturing Enterprises This embodiment is applied to the collaborative scenario of the entire raw material procurement process in a large discrete manufacturing enterprise. The heterogeneous business systems involved in the enterprise include ERP system, SRM system, WMS system, financial system, and OA approval system.

[0058] First, a full-scenario workflow semantic modeling step is executed to collect the original feature parameters of the procurement process. The process has 12 nodes, including: requirement submission, budget verification, supplier screening, price comparison, contract approval, order placement, goods receipt and inspection, warehousing, invoice verification, payment approval, payment execution, and process archiving. Participating roles include production department requirement submitters, procurement department purchasing staff, procurement department supervisors, finance department budget accountants, finance department supervisors, quality department inspection staff, warehouse managers, and finance cashiers. Data interaction rules include the transmission rules for procurement requirement data, budget data, supplier data, order data, inspection data, warehousing data, invoice data, and payment data. The maximum allowable execution delay is also defined. The preset maximum latency threshold is 20 days. It lasts for 30 days.

[0059] The original feature parameters are standardized and semantically encoded to generate a standardized semantic feature vector for the target scene. The unified semantic workflow template library is traversed to extract the semantic feature vectors of the templates to be matched, and a fit score is calculated. Preset weights are used. It is 0.35. It is 0.25. It is 0.25. The value is 0.15. The node matching degree is calculated. The match score is 0.92, indicating a high degree of character compatibility. The data matching score is 0.95. The execution delay of the template to be matched is 0.90. It lasts for 18 days.

[0060] Substitute into the fit score formula to calculate: First, calculate the delay matching term. ; Then calculate ; Step-by-step calculation: ; ; ; ; Summation yields ; The preset adaptation threshold is 0.85. If the template's adaptation score is higher than the threshold, it will be added to the candidate template set. Topology verification, role matching verification, and compliance verification are performed on the candidate templates. Once all verifications are passed, a unified semantic workflow template for the procurement scenario is generated.

[0061] By using a cross-system collaborative adaptation engine to connect with ERP systems, SRM systems, WMS systems, financial systems, and OA approval systems, the semantic similarity of interface parameters is calculated and preset weights are assigned. The maximum depth of the domain ontology tree is 0.6. Given a value of 10, calculate the semantic similarity between the purchase order number parameter and the order ID parameter in the ERP system, and the cosine similarity of the word vectors. The least common ancestor node depth is 0.88. The value is 8. Substituting this into the formula, we get... If the semantic similarity is higher than the preset threshold of 0.75, parameter mapping is completed.

[0062] Dynamic permission allocation is implemented for contract approval nodes, collecting basic attribute parameters of the nodes and the total permission quota of the nodes. The node's business revenue value is 100. The risk of loss due to node permission leakage is 1000. The value is 800. The relevant collaborative roles are purchasing agent, purchasing department manager, and finance department manager, with corresponding business contribution coefficients. It is 0.3. It is 0.4. The risk coefficient is 0.3. It is 0.4. It is 0.2. It is 0.3.

[0063] Construct a multi-role non-cooperative power game model, with the utility functions as follows: ; ; ; Solving for the Nash equilibrium yields the initial optimal power allocation values. It is 10. It is 50. The threshold is 40. Perform permission boundary compliance checks; if all requirements are met, generate the final permission allocation plan and issue it for execution.

[0064] The entire procurement process is broken down into 8 parallel subtasks, with a total of [number missing] subtasks. The total number of available collaborative resource nodes is 8. Given a value of 5, a constrained task scheduling optimization model is constructed, and the optimal task allocation scheme is obtained. The total execution time of the workflow is [value missing]. The time limit has been optimized to 16 days to meet the maximum allowable latency requirement.

[0065] Ultimately, through a one-stop collaborative interaction front-end module, it provides a one-stop collaborative service for all users across all roles throughout the procurement process, enabling full-link control and cross-system collaboration of the procurement process.

[0066] Example 2: Collaborative Scenario for the Entire Credit Approval Process in a Joint-Stock Commercial Bank This embodiment is applied to the collaborative scenario of the entire process of personal business loan approval in a joint-stock commercial bank. The heterogeneous business systems involved in the enterprise include the credit business system, credit inquiry system, risk control system, anti-money laundering system, core accounting system, electronic contract system, and OA approval system.

[0067] First, a full-scenario workflow semantic modeling step is executed to collect the original feature parameters of the credit approval process. The process has 10 nodes, including loan application, document verification, credit inquiry, initial risk control review, secondary risk control review, anti-money laundering verification, contract signing, loan approval, loan execution, and process archiving. Participating roles include account manager, initial risk control specialist, secondary risk control specialist, anti-money laundering verification specialist, contract administrator, loan approval specialist, and cashier. Data interaction rules include the transmission rules for loan application data, customer credit data, risk control assessment data, anti-money laundering verification data, contract data, and loan disbursement data. The maximum allowable execution delay is also specified. The preset maximum latency threshold is 3 working days. It takes 5 business days.

[0068] The original feature parameters are standardized and semantically encoded to generate a standardized semantic feature vector for the target scene. The unified semantic workflow template library is traversed to extract the semantic feature vectors of the templates to be matched, and a fit score is calculated. Preset weights are used. It is 0.3. It is 0.3. It is 0.3. The value is 0.1. The node matching degree is calculated. The match score is 0.96, indicating a high degree of character compatibility. The data matching degree is 0.98. The execution delay of the template benchmark to be matched is 0.95. It takes 2.5 business days.

[0069] Substitute into the fit score formula to calculate: First, calculate the delay matching term. ; Then calculate ; The preset adaptation threshold is 0.88. If the template's adaptation score is higher than the threshold, it will be included in the candidate template set. Topology verification, role matching verification, and compliance verification are performed on the candidate templates. Once all verifications are passed, a unified semantic workflow template for the credit approval scenario is generated.

[0070] By using a cross-system collaborative adaptation engine to connect with credit business systems, credit inquiry systems, risk control systems, anti-money laundering systems, core accounting systems, electronic contract systems, and OA approval systems, the semantic similarity of interface parameters is calculated and preset weights are assigned. The maximum depth of the domain ontology tree is 0.7. Given a value of 12, calculate the semantic similarity between the customer's ID number parameter and the customer's unique identifier parameter in the credit reporting system, and the word vector cosine similarity. The least common ancestor node depth is 0.92. Substituting 10 into the formula yields... If the semantic similarity is higher than the preset threshold of 0.8, the parameter mapping is completed.

[0071] Dynamic permission allocation is implemented for risk control review nodes, collecting basic attribute parameters of the nodes and the total permission quota of the nodes. The node's business revenue value is 100. The risk of loss due to node permission leakage is 1200. The value is 1500. The associated collaborative roles are the initial review risk control specialist, the secondary review risk control specialist, and the risk control department head, with corresponding business contribution coefficients. It is 0.25. It is 0.5. The risk coefficient is 0.25. It is 0.5. It is 0.2. It is 0.3.

[0072] Construct a multi-role non-cooperative power game model, with the utility functions as follows: ; ; ; Solving for the Nash equilibrium yields the initial optimal power allocation values. It is 5. It is 70. Set the value to 25. Perform permission boundary compliance checks. If all requirements are met, generate the final permission allocation plan and issue it for execution.

[0073] The entire credit approval process is broken down into 6 parallel subtasks, with a total of [number missing] subtasks. The total number of available collaborative resource nodes is 6. Given a time limit of 4, a constrained task scheduling optimization model is constructed to obtain the optimal task allocation scheme and the total execution time of the workflow. The time has been optimized to 2.2 working days, meeting the maximum allowable delay requirement.

[0074] Ultimately, through a one-stop collaborative interaction front-end module, it provides a one-stop collaborative service for the entire credit approval process for all users, realizing full-link control and cross-system collaboration of the credit approval process.

[0075] Example 3: Collaborative Scenario for Cross-Team Product Development in Internet Enterprises This embodiment is applied to a collaborative scenario for the entire product iteration and R&D process of Internet companies. The heterogeneous business systems involved in the enterprise include project management system, code hosting system, test management system, defect management system, operation and maintenance release system, OA approval system, and enterprise IM system.

[0076] First, a full-scenario workflow semantic modeling step is executed to collect raw feature parameters of the product development process. The process has 11 nodes, including: requirement submission, requirement review, product design, UI design, development task breakdown, code development, testing and verification, defect fixing, deployment approval, production release, and process archiving. Participating roles include product manager, UI designer, front-end developer, back-end developer, test engineer, operations engineer, project manager, and technical lead. Data interaction rules include transmission rules for requirement data, design draft data, development task data, code submission data, test case data, defect data, and release request data. The maximum allowable execution latency is also determined. The preset maximum latency threshold is 14 calendar days. It lasts for 20 calendar days.

[0077] The original feature parameters are standardized and semantically encoded to generate a standardized semantic feature vector for the target scene. The unified semantic workflow template library is traversed to extract the semantic feature vectors of the templates to be matched, and a fit score is calculated. Preset weights are used. It is 0.32. It is 0.28. It is 0.25. The value is 0.15. The node matching degree is calculated. The match score is 0.94, indicating a high degree of character compatibility. The data matching degree is 0.96. The execution latency of the template to be matched is 0.93. It lasts for 12 calendar days.

[0078] Substitute into the fit score formula to calculate: First, calculate the delay matching term. ; Then calculate ; The preset adaptation threshold is 0.86. This template's adaptation score is higher than the threshold and it is added to the candidate template set. Topology verification, role matching verification, and compliance verification are performed on the candidate templates. Once all verifications are passed, a unified semantic workflow template for product development scenarios is generated.

[0079] By integrating with project management systems, code hosting systems, test management systems, defect management systems, operations and maintenance release systems, OA approval systems, and enterprise IM systems through a cross-system collaborative adaptation engine, semantic similarity of interface parameters is calculated and preset weights are assigned. The maximum depth of the domain ontology tree is 0.65. Given a value of 11, calculate the semantic similarity between the requirement ID parameter and the unique identifier parameter of the project management system requirement, and the word vector cosine similarity. The depth of the least common ancestor node is 0.90. The value is 9. Substituting this into the formula, we get... The semantic similarity is higher than the preset threshold of 0.78, and the parameter mapping is completed.

[0080] Dynamic permission allocation is implemented for online approval nodes, collecting basic attribute parameters of the nodes and the total permission quota of the nodes. The node's business revenue value is 100. The risk of loss due to node permission leakage is 800. The value is 1000. The associated collaborative roles are test engineer, project manager, technical lead, and operations engineer, with corresponding business contribution coefficients. It is 0.2. It is 0.3. It is 0.3. The risk coefficient is 0.2. It is 0.4. It is 0.25. It is 0.2. It is 0.35.

[0081] Construct a multi-role non-cooperative power game model, with the utility functions as follows: ; ; ; ; Solving for the Nash equilibrium yields the initial optimal power allocation values. It is 10. It is 20. It is 50. Set the value to 20. Perform permission boundary compliance checks; if all requirements are met, generate the final permission allocation plan and issue it for execution.

[0082] The entire product development process is broken down into 10 parallel subtasks, with a total of [number missing] subtasks. The total number of available collaborative resource nodes is 10. Given a value of 6, a constrained task scheduling optimization model is constructed, and the optimal task allocation scheme is obtained. The total execution time of the workflow is [value missing]. The timeframe has been optimized to 11 calendar days to meet the maximum allowable delay requirement.

[0083] Ultimately, through a one-stop collaborative interaction front-end module, it provides a one-stop collaborative service for all users across all roles throughout the product development process, enabling full-link control and cross-system collaboration of the product development process.

[0084] This invention constructs an enterprise-level business process domain ontology library, performs standardized semantic modeling of process elements across all business scenarios, and achieves accurate matching between workflow templates and target business scenarios by combining an adaptability scoring formula. It also supports custom configuration and multi-dimensional verification of template nodes, enabling the rapid generation of customized templates adapted to specific business scenarios. This avoids the high development costs and long adaptation cycles caused by the repetitive development of process templates for different business scenarios in traditional platforms. Through the semantic similarity calculation formula of the semantic mapping unit, it achieves automatic mapping and matching between heterogeneous system interface parameters and unified semantic template parameters. Combined with interface lifecycle monitoring and exception handling mechanisms, it achieves seamless linkage and data interoperability between cross-system process nodes, effectively breaking down data silos and collaboration gaps between different business systems within an enterprise, and improving the accuracy and operational stability of cross-system workflow collaboration.

[0085] This invention constructs a multi-role non-cooperative permission game model. Based on the attribute parameters of process nodes and collaborative roles, it achieves optimal permission allocation by solving the Nash equilibrium solution through a utility function. Combined with permission boundary compliance verification and dynamic update mechanisms, it can dynamically adjust permissions and resolve conflicts in real time. This avoids the shortcomings of traditional static permission control in adapting to dynamic process changes, while also taking into account the business efficiency and data security control requirements of workflow execution. By constructing a task scheduling optimization model with constraints, it realizes the intelligent decomposition of complex process tasks and the global optimal allocation of collaborative resources. Under resource constraints and latency constraints, it can minimize the total execution time of the workflow, solving the problems of uneven resource allocation and low process execution efficiency in traditional task scheduling methods, and effectively improving the utilization rate of collaborative resources and the overall execution efficiency of the workflow.

[0086] The various functional modules of this invention achieve real-time data communication and coordinated execution logic, realizing one-stop management of the entire workflow from template construction, system integration, permission control, task scheduling to terminal interaction. Through a unified collaborative interaction front-end, it provides a standardized service entry point for all users, avoiding the problems of process execution gaps and low operational efficiency caused by switching between multiple systems and operating through multiple entry points. At the same time, the core algorithm model and architecture design of this invention have good versatility and scalability, and can be adapted without large-scale architecture modifications for specific industries or business scenarios. It has strong engineering implementation value and technology reusability, providing a creative and complete technical solution for enterprise-level full-scenario workflow collaborative management.

[0087] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. An enterprise-level, full-scenario, one-stop intelligent collaborative workflow service platform, characterized in that: It includes a full-scenario workflow semantic modeling module, a cross-system collaborative adaptation engine, a dynamic permission game control module, a workflow intelligent scheduling and optimization module, and a one-stop collaborative interaction front-end module; The full-scenario workflow semantic modeling module is used to perform standardized semantic modeling of process nodes, participating roles, data interaction rules, and execution constraints in all business scenarios of an enterprise, and generate a unified semantic workflow template compatible with multiple business scenarios. The cross-system collaborative adaptation engine is used to connect with heterogeneous business systems within an enterprise, complete interface semantic mapping, bidirectional data format conversion and instruction pass-through, and realize seamless linkage and data interoperability of cross-system process nodes. The dynamic permission game control module is used to perform dynamic permission allocation and real-time resolution of permission conflicts based on the business responsibility boundaries of collaborative roles and the security level attributes of process nodes. The workflow intelligent scheduling and optimization module is used to collect full-link execution data of the workflow, and to complete the dynamic adjustment of process nodes, intelligent decomposition of parallel tasks, and optimal allocation of collaborative resources. The one-stop collaborative interaction front-end module is used to provide a unified interactive entry point for all users to initiate processes, approve workflows, process tasks, track progress, and visualize data.

2. The enterprise-level, full-scenario, one-stop intelligent collaborative workflow service platform according to claim 1, characterized in that, The full-scenario workflow semantic modeling module is also used to calculate the fit score between the workflow template to be matched and the target scenario based on the feature parameters of the target business scenario. The formula for calculating the fit score is as follows: ； In the formula, Rate the fit. These are preset weights for node matching, role matching, data matching, and latency matching, respectively. ; For process node matching degree, To participate in role matching, For data interaction rule matching degree, The maximum allowable execution latency for the target scenario. The baseline execution delay for the template to be matched. This is the preset maximum delay threshold.

3. The enterprise-level, full-scenario, one-stop intelligent collaborative service platform for workflows according to claim 1, characterized in that, The cross-system collaborative adaptation engine has a built-in domain ontology semantic mapping unit. This unit is used to calculate the semantic similarity between heterogeneous system interface parameters and unified semantic template parameters. The formula for calculating the semantic similarity is as follows: In the formula, Let be the semantic similarity between parameter a and parameter b. The preset weighting coefficients, Let cosine similarity be the word vectors corresponding to parameters a and b. Let be the least common ancestor node of parameters a and b in the domain ontology tree. The depth of the least common ancestor node. This represents the maximum depth of the domain ontology tree.

4. The enterprise-level, full-scenario, one-stop intelligent collaborative service platform for workflows according to claim 1, characterized in that, The dynamic permission game control module is used to construct a multi-role non-cooperative permission game model, solve for the Nash equilibrium solution to achieve optimal permission allocation, and the utility function of the game model is: In the formula, Let i be the utility function for role i. The permission values ​​requested for role i. Strategies for assigning permissions to other roles. The business contribution coefficient for role i. The total permission quota for process nodes. The business revenue value of the process node. For role i, the risk factor The risk loss value for permission leakage at process nodes; the optimal permission allocation is to satisfy... Nash equilibrium solution .

5. The enterprise-level, full-scenario, one-stop intelligent collaborative service platform for workflows according to claim 1, characterized in that, The workflow intelligent scheduling and optimization module is used to construct a constrained task scheduling optimization model and solve for the optimal task allocation scheme. The objective function of the optimization model is: ； The constraints are: In the formula, Let m be the total execution time of the workflow, m be the total number of subtasks after decomposition, and n be the total number of available collaborative resource nodes. Let j be the execution time of subtask j on resource node k. This is a 0-1 decision variable; a value of 1 indicates that subtask j is assigned to resource node k for execution. Let J be the resource consumption of subtask j. Let k be the maximum available resource quantity for resource node k. This represents the maximum allowed execution time for subtask j.

6. The enterprise-level, full-scenario, one-stop intelligent collaborative workflow service platform according to claim 1, characterized in that, The full-scenario workflow semantic modeling module has a built-in template customization configuration unit. The template customization configuration unit is used to classify, manage and customize the nodes of the unified semantic workflow template. The node classification includes manually executed nodes, automatically executed nodes, condition judgment nodes, parallel convergence nodes, abnormal termination nodes and rollback nodes. The custom configuration includes adding and deleting nodes within the template, adjusting the dependency topology between nodes, configuring branch conditions for condition judgment nodes, and configuring rollback rules for rollback nodes; the dimensions of the branch condition configuration include business data field thresholds, node execution status identifiers, role operation result codes, and cross-system data feedback values. The rollback rule configuration includes rollback trigger conditions, rollback level range, rollback data reverse synchronization rules, and rollback operation permission verification rules; the template customization configuration unit is also used to perform semantic consistency verification, node dependency relationship topology verification, and compliance constraint verification on the configured template to generate customized workflow templates adapted to specific business scenarios.

7. The enterprise-level, full-scenario, one-stop intelligent collaborative service platform for workflows according to claim 1, characterized in that, The cross-system collaborative adaptation engine has a built-in interface monitoring and exception handling unit. The interface monitoring and exception handling unit is used to perform real-time monitoring of the entire lifecycle of the interfaces of the heterogeneous business systems being connected. The monitoring dimensions include interface call response time, interface return code status, data transmission integrity, link connectivity, and interface concurrent capacity. The interface monitoring and exception handling unit has preset exception judgment rules, which include a threshold for consecutive timeouts, a list of non-successful return code types, a data packet loss rate threshold, and a threshold for the duration of link interruption. When the monitored data triggers the exception judgment rules, the interface monitoring and exception handling unit automatically triggers the preset retry mechanism and graded degradation strategy. The retry mechanism includes gradient configuration of the number of retryes, exponential backoff parameter configuration of the retry interval, idempotency verification rules for retry requests, and backup route switching rules for retry links; the tiered degradation strategy includes tiered triggering conditions for full process degradation, branch process degradation, and node function degradation, as well as bypass rules for degraded process links, local caching rules for interactive data, and data resending rules after link recovery; the interface monitoring and exception handling unit is also used to embed and synchronously store the execution logs and context data of the entire interface call process.

8. An enterprise-level, full-scenario, one-stop intelligent collaborative workflow service method, applied to the enterprise-level, full-scenario, one-stop intelligent collaborative workflow service platform as described in any one of claims 1-7, characterized in that, Includes the following steps: S1. Standardize semantic modeling of process nodes, participating roles, data interaction rules, and execution constraints for all business scenarios of an enterprise, and generate a unified semantic workflow template compatible with multiple business scenarios; S2. Connect with heterogeneous business systems within the enterprise, complete interface semantic mapping, bidirectional data format conversion and instruction pass-through, and establish cross-system collaborative links; S3. Based on the business responsibility boundaries of collaborative roles and the security level attributes of process nodes, dynamic permission allocation and real-time resolution of permission conflicts are performed. S4. Collect execution data across the entire workflow, and perform dynamic adjustment of process nodes, intelligent decomposition of parallel tasks, and optimal allocation of collaborative resources; S5. Through a unified interactive portal, it provides a one-stop collaborative service for all users, including process initiation, approval workflow, task processing, progress tracking, and data visualization.

9. The enterprise-level, full-scenario, one-stop intelligent collaborative workflow service method according to claim 8, characterized in that, Step S1 specifically includes the following sub-steps: S101. Collect the original feature parameters of the target business scenario. The original feature parameters include the number of process nodes and node type labels, the number of participating roles and role rights and responsibilities labels, the field types and transmission rules of data interaction, and the time delay constraint parameters and compliance constraint parameters of process execution. S102. Standardize the collected raw feature parameters into semantics to generate a standardized semantic feature vector corresponding to the target business scenario; S103. Traverse the pre-built unified semantic workflow template library, and extract the corresponding template semantic feature vector for each workflow template to be matched in the library; S104. Based on the fit score calculation formula, calculate the fit score between the standardized semantic feature vector of the target business scenario and the template semantic feature vector of each workflow template to be matched in turn. S105. Sort all workflow templates to be matched in descending order of fit score, filter out workflow templates with fit scores higher than the preset fit threshold, and generate a candidate template set. S106. For each workflow template in the candidate template set, perform node dependency topology verification, role and responsibility boundary matching verification, and data interaction rule compliance verification in sequence, and output the candidate workflow templates that pass all verifications; S107. Based on the validated candidate workflow templates, complete the standardized semantic modeling of all business scenarios and generate a unified semantic workflow template compatible with multiple business scenarios.

10. The enterprise-level, full-scenario, one-stop intelligent collaborative workflow service method according to claim 8, characterized in that, Step S3 specifically includes the following sub-steps: S301. Collect the basic attribute parameters of the current workflow process node. The basic attribute parameters include the node security level attribute, the node total permission quota, the node business revenue value, and the node permission leakage risk loss value. S302. Collect the attribute parameters of all associated collaborative roles in this process node. The attribute parameters include role business responsibility boundary label, role business contribution coefficient, and role risk coefficient. S303. Taking each associated collaborative role as the game participant and the permission allocation value of each collaborative role as the decision variable, based on the collected basic attribute parameters and role attribute parameters, and using the utility function, a multi-role non-cooperative permission game model is constructed. S304. Using an iterative approximation algorithm, solve the pure strategy Nash equilibrium solution of the constructed multi-role non-cooperative permission game model to obtain the initial optimal permission allocation value for each cooperating role; S305. For the initial optimal permission allocation values ​​of each collaborative role, perform permission boundary compliance verification, remove permission values ​​that exceed the legal rights and responsibilities boundaries of the role and the security control scope of the process node, generate the final permission allocation scheme and issue it for execution; S306. Collect data on role permission application changes, process node security level adjustments, and collaborative role additions or exits during the process in real time. When the collected data triggers the preset dynamic update conditions, repeat steps S301 to S305 to complete dynamic permission updates and real-time resolution of permission conflicts.

Images