Complex semantic-oriented business process multi-agent modeling and intelligent error correction system

Abstract

The invention relates to the technical field of data processing, in particular to a complex semantic-oriented business process multi-agent modeling and intelligent error correction method and system. The method comprises the following steps: firstly, based on a BPMN meta-model, disassembling business process elements, analyzing BPMN specifications, forming a structured framework of business process modeling, processing unstructured texts in parallel through an intelligent agent group, identifying business process key concepts, mapping the business process key concepts into semi-structured information, and generating a preliminary extraction result; verifying the integrity of the preliminary extraction result based on a BPMN specification, and generating an information integrity verification report; based on an information integrity verification report, missing information of the preliminary extraction result is retrieved through a knowledge base and complemented; performing secondary verification on the complemented extraction result, and correcting the complemented extraction result based on business logic and user requirements; and generating an XML code and a visual flow chart conforming to the BPMN specification based on the corrected extraction result. According to the method, the problems of semantic fracture and structure deficiency of the complex semantic text in business process modeling can be effectively solved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, system, and computer-readable storage medium for multi-agent modeling and intelligent error correction of business processes with complex semantics. Background Technology

[0002] Business process modeling, as a core supporting technology for enterprise digital transformation, is widely used in the fields of process automation, knowledge management, and system integration.

[0003] As business complexity increases, existing technology systems primarily rely on UML / MOF meta-modeling frameworks, BPMN 2.0 standardization specifications, and single-agent-driven modeling methods. Tools like BPMN.js have formed a complete technology chain from text parsing to visualization generation. However, existing single-agent-based modeling techniques mainly rely on pre-defined rule bases or finite state machines for process modeling, making it difficult to effectively handle multi-level semantic relationships in complex texts. When parsing business requirements involving multiple causal relationships and cross-departmental collaboration, single agents suffer from limitations such as limited knowledge dimensions and insufficient reasoning depth. For example, when processing complex conditional sentences like "When a customer submits an application and the risk assessment level is higher than the threshold, it requires dual approval from the legal department and the business department," the agent may fail to decompose the task collaboration unit, resulting in a missing decision gateway node. This limitation stems from the lack of task decomposition and collaborative reasoning mechanisms in single agents, making it difficult to establish a multi-dimensional semantic analysis framework.

[0004] Meanwhile, existing information extraction methods based on regular expressions or traditional NLP suffer from information gaps when processing unstructured text. When business descriptions are scattered across different media such as policy documents, meeting minutes, and email correspondence, it is difficult for a single agent to establish a unified standard for information granularity. Summary of the Invention

[0005] The present invention aims to at least partially solve one of the technical problems in the related art.

[0006] Therefore, the first objective of this invention is to propose a multi-agent modeling and intelligent error correction system for business processes with complex semantics, comprising: S1, based on the BPMN meta-model, decomposes business process elements, parses the BPMN specification, forms a structured framework for business process modeling, and uses intelligent agents to process unstructured text in parallel, identifies key business process concepts and maps them into semi-structured information, and generates preliminary extraction results. S2, verify the integrity of the preliminary extraction results based on the BPMN specification, and generate an information integrity verification report; S3, based on the information integrity verification report, retrieves and completes the missing information from the initial extraction results through the knowledge base; S4. Perform a second verification on the completed extraction results. After confirming that the extraction results are complete and conform to the BPMN syntax rules, modify the completed extraction results based on business logic and user needs. S5 generates XML code and a visual flowchart that conform to the BPMN specification based on the corrected extraction results.

[0007] In one embodiment of the present invention, S1 further includes: S11, based on the BPMN meta-model, defines core elements to parse the BPMN specification and forms a structured framework for business process modeling; S12, Based on the BPMN meta-model, define the core elements and identify key concepts of business processes in non-institutionalized texts. S13, set the key concepts of the business process as semi-structured information according to the BPMN specification to form the preliminary collection results.

[0008] In one embodiment of the present invention, the method for correcting the extracted result based on business logic and user requirements in step S4 is as follows: The accuracy of the business logic and the consistency of user requirements are verified based on the semi-structured information. If there is a deviation between the extracted results and the business logic or user requirements, the extracted results will be corrected.

[0009] In one embodiment of the present invention, the method for generating XML code conforming to the BPMN specification in step S5 is as follows: Based on the corrected extraction results, XML code snippets of the BPMN 2.0 specification are generated in modules. The XML code of each module is merged, and the code syntax and element relationships are verified to generate a complete XML file. The modules include active nodes, connectors and swimlane definitions.

[0010] In one embodiment of the present invention, the method for generating the visual flowchart in step S5 is as follows: Based on the complete XML file, the intelligent rendering engine of BPMN.js integrates a genetic algorithm for automatic layout optimization, defines a three-dimensional layout evaluation function including the number of intersections, path length, and node spacing, and generates a clear and readable flowchart.

[0011] In one embodiment of the present invention, step S3 refers to the knowledge base domain concepts, rules, and constraints.

[0012] To achieve the above objectives, a second aspect of the present invention provides a multi-agent collaborative business process semantic parsing and modeling apparatus, comprising: The preliminary extraction result generation module is used to decompose business process elements based on the BPMN meta-model, parse the BPMN specification, form a structured framework for business process modeling, process unstructured text in parallel through intelligent agent swarms, identify key concepts of business processes and map them into semi-structured information, and generate preliminary extraction results. The preliminary verification and information completion module verifies the completeness of the preliminary extraction results based on the BPMN specification, generates an information completeness verification report, and retrieves and completes the missing information of the preliminary extraction results through the knowledge base based on the information completeness verification report. The secondary verification and correction module is used to perform secondary verification on the completed extraction results. After confirming that the extraction results are complete and conform to the BPMN syntax rules, the module corrects the completed extraction results based on business logic and user requirements. The flowchart generation module is used to generate XML code and visual flowcharts that conform to the BPMN specification based on the corrected extraction results.

[0013] In one embodiment of the present invention, the preliminary extraction result generation module is further configured to: Based on the BPMN metamodel, the core elements are defined and the BPMN specification is parsed to form a structured framework for business process modeling; Based on the BPMN meta-model, the core elements are defined to identify key concepts of business processes in non-institutionalized texts. The key concepts of the business process are set as semi-structured information according to the BPMN specification to form preliminary collection results.

[0014] In one embodiment of the present invention, the flowchart generation module is further configured to: The flowchart is modified according to the user's instructions, and an evaluation report is output that includes three dimensions: syntactic compliance, semantic completeness, and visual readability. Finally, a standard BPMN2.0 model file is generated, and the system principle diagram and flowchart are summarized.

[0015] To achieve the above objectives, a third aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect.

[0016] The methods, systems, and storage media of this invention can effectively solve the problems of semantic breaks and structural omissions in business process modeling of complex semantic text. Through multi-agent collaborative parsing and a three-layer progressive verification mechanism, the semantic integrity, logical accuracy, and automated generation efficiency of the model are significantly improved.

[0017] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0018] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart of a multi-agent modeling and intelligent error correction method for business processes with complex semantics according to an embodiment of the present invention; Figure 2 This is a structural diagram of a multi-agent modeling and intelligent error correction system for business processes oriented towards complex semantics, according to an embodiment of the present invention. Detailed Implementation

[0019] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0020] 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 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 skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0021] The following description, with reference to the accompanying drawings, illustrates a multi-agent modeling and intelligent error correction method for complex semantic business processes according to an embodiment of the present invention. Example 1 Figure 1 This is a flowchart of a multi-agent modeling and intelligent error correction method for business processes with complex semantics according to an embodiment of the present invention.

[0022] like Figure 1 As shown, the multi-agent collaborative modeling method for complex text semantic association includes the following steps: S1 decomposes business process elements based on the BPMN meta-model, parses the BPMN specification, forms a structured framework for business process modeling, and uses intelligent agents to process unstructured text in parallel, identify key business process concepts and map them into semi-structured information to generate preliminary extraction results.

[0023] Further, step S1 includes: S11, based on the BPMN meta-model, defines core elements to parse the BPMN specification and forms a structured framework for business process modeling; S12, Based on the BPMN meta-model, define the core elements and identify key concepts of business processes in non-institutionalized texts. S13, set the key concepts of the business process as semi-structured information according to the BPMN specification to form the preliminary collection results.

[0024] Specifically, in some implementations, constructing a multi-agent task decomposition framework based on the BPMN meta-model is one of the core steps in this invention to achieve automated business process modeling. This step involves structurally decomposing business process elements (such as activities, events, gateways, swimlanes, etc.) in the BPMN 2.0 standard and assigning these elements to different groups of agents for collaborative parsing, thereby achieving efficient modeling and semantic association processing of complex business processes.

[0025] At the technical implementation level, this step first parses the structure of the business process based on the BPMN meta-model. The BPMN meta-model defines the types, attributes, and relationships of various elements in the flowchart, such as the input / output flow of activities, the triggering conditions of events, the branching logic of gateways, and the resource allocation relationships of swimlanes. Agent I parses the BPMN 2.0 XML Schema definition to construct a structured template containing all process elements and passes it to multiple agents within the process flow. These agents are responsible for identifying different process elements in the text, such as activity nodes, event types, and gateway logic, and mapping them into semi-structured information according to the BPMN specification, including node types, connection relationships, and attribute values.

[0026] At the application level, this step is widely applicable to scenarios such as enterprise-level business process modeling, cross-departmental collaborative process design, and automated process document parsing. For example, in loan approval process modeling in the financial industry, key activities such as "customer submits application," "risk assessment," and "approval decision" can be parsed and mapped to structured nodes such as service tasks and exclusive gateways in BPMN, and then assigned to the corresponding swimlanes according to departmental responsibilities.

[0027] At the technical level, this step, through a multi-agent collaborative parsing mechanism, effectively solves problems such as semantic fragmentation and information loss in traditional single-agent modeling, improving the accuracy and completeness of business process element identification. Furthermore, the framework supports dynamic expansion and multi-source information fusion, laying a structured foundation for subsequent meta-model validation, knowledge reasoning, and XML generation, thereby significantly enhancing the automation and intelligence level of the entire modeling system.

[0028] Simultaneously, this step first preprocesses the unstructured text using a semantic recognition agent, including natural language processing (NLP) operations such as word segmentation, part-of-speech tagging, and dependency parsing. Subsequently, a conditional analysis agent identifies conditional semantic units in the text (such as "when...", "if..."), while an action mapping agent extracts verb phrases or activity descriptions related to business operations and maps them to activity nodes or task nodes in the BPMN. A constraint verification agent identifies restrictive semantics in the text (such as "must", "must not", "requires") and transforms them into gateway constraints or swimlane assignment rules in the BPMN. The agents exchange information and engage in collaborative reasoning through a semantic association network, ensuring the consistency and integrity of semantic relationships at different granularities (such as sentence level, paragraph level, and document level).

[0029] At the application level, this step is widely applicable to business process modeling in fields such as finance, healthcare, and government. For example, in financial approval processes, intelligent agents can identify complex semantics such as "customer submits application," "risk assessment level exceeds threshold," and "requires dual approval from legal and business departments" from policy documents, and map them to parallel gateways and multi-lane structures in BPMN, thereby achieving automated process modeling and compliance verification.

[0030] The technical effect of this step is that, through multi-agent collaborative parsing, it significantly improves the accuracy and coverage of semantic relationship recognition in complex texts, solves problems such as semantic fragmentation and incomplete rule mapping in traditional single-agent modeling, and lays a solid foundation for subsequent model generation and quality verification.

[0031] S2 verifies the integrity of the initial extraction results based on the BPMN specification and generates an information integrity verification report.

[0032] S3, based on the information integrity verification report, retrieves and completes the missing information from the preliminary extraction results through the knowledge base.

[0033] Specifically, steps S2-3 construct a dual-engine collaborative reasoning mechanism by integrating the structured reasoning capabilities of knowledge graphs with the contextual semantic understanding capabilities of large language models (LLM), thereby effectively solving the problems of semantic fragmentation, information loss, and insufficient contextual understanding in traditional modeling methods.

[0034] At the technical implementation level, this step reads the information integrity verification report and identifies missing or conflicting semantic relationships, such as undefined activity nodes, unconnected gateway paths, or unclear resource allocation relationships. Subsequently, it invokes the knowledge graph reasoning engine to perform structured reasoning based on the domain ontology and entity relationships in the existing business knowledge base (such as "approval-role-department," "condition-action-constraint," etc.) to attempt to complete the missing semantic information. If the knowledge graph cannot provide sufficient contextual support, it further invokes a large language model to predict and complete the missing information through context-aware semantic generation capabilities. For example, for the statement "requires dual approval from the legal department and the business department," if the "legal department" and "business department" do not have corresponding role nodes in the knowledge graph, the LLM can generate reasonable role definitions based on contextual semantics and suggest their collaborative paths in the process.

[0035] At the application level, this step is widely applicable to complex business scenarios such as cross-departmental collaborative process modeling, financial risk control process design, and medical business process automation. For example, in financial approval processes, when the original text description lacks the mapping relationship between "approver role" and "approval conditions," this dual-engine mechanism can automatically identify and complete the relevant nodes, generating semi-structured data conforming to the BPMN 2.0 specification for subsequent XML code generation and visualization rendering.

[0036] At the technical level, this step significantly improves the semantic completeness and reasoning accuracy of business process modeling. Structured reasoning using knowledge graphs ensures that supplementary information conforms to domain ontology logic; contextual understanding through a large language model enhances the ability to capture implicit semantics in unstructured text. The dual-engine collaborative mechanism effectively compensates for the limitations of a single agent in semantic understanding and reasoning depth, providing high-quality input data for subsequent multi-agent collaborative verification and closed-loop error correction, thereby improving the automation level and model consistency of the entire modeling system.

[0037] S4 performs a second verification on the completed extraction results. After confirming that the extraction results are complete and conform to the BPMN syntax rules, the completed extraction results are corrected based on business logic and user requirements.

[0038] Specifically, the three-layer progressive quality verification mechanism proposed in this invention is the core link to achieve high-precision and high-consistency modeling of business process models.

[0039] At the technical implementation level, this step first verifies the structural integrity of the initially extracted semi-structured information based on the BPMN 2.0 metamodel (such as the Bpmn2ModelElementInstance interface defined in the OMG standard). For example, it verifies rules such as "whether start and end events are unique" and "whether the gateway connects to a valid path" to ensure that the model conforms to the BPMN syntax specification. Second, based on the domain ontology and business rule base, it verifies the rationality of the process path and the logical relationships between nodes to ensure closed-loop process logic and avoid deadlocks or unreachable paths. Finally, in conjunction with business compliance standards (such as ISO / IEC 20000-1 service management process specifications, internal corporate documents, etc.), it conducts a compliance review of the activity nodes, role assignments, resource constraints, etc. in the model to ensure that the model conforms to the actual business operation process.

[0040] At the application level, this mechanism is widely applicable to industries with high requirements for process compliance, such as finance, healthcare, and government. For example, in modeling financial credit approval processes, Agent III can verify whether process nodes conform to the BPMN 2.0 specification, Agent V can verify whether the approval conditions are logically consistent, and Agent VI conducts compliance reviews based on regulatory policies and internal risk control rules. Through collaborative verification by multiple agents, the system can automatically identify and correct problems such as "missing approval nodes" and "incorrect role allocation," significantly improving modeling efficiency and accuracy.

[0041] The technical advantage of this step lies in achieving multi-dimensional quality control of the business process model through a layered and progressive verification mechanism and dynamic feedback between agents. Compared to traditional single-agent verification methods, the three-layer verification mechanism of this invention can effectively reduce the error propagation rate, improve the executability and maintainability of the model, and provide a high-quality data foundation for subsequent XML generation and visualization rendering.

[0042] S5 generates XML code and a visual flowchart that conform to the BPMN specification based on the corrected extraction results.

[0043] Specifically, the method for generating XML code conforming to the BPMN specification is as follows: Based on the corrected extraction results, XML code snippets of the BPMN 2.0 specification are generated in modules. The XML code of each module is merged, and the code syntax and element relationships are verified to generate a complete XML file. The modules include active nodes, connectors and swimlane definitions.

[0044] The method for generating visual flowcharts is as follows: Based on the complete XML file, the intelligent rendering engine of BPMN.js integrates a genetic algorithm for automatic layout optimization, defines a three-dimensional layout evaluation function including the number of intersections, path length, and node spacing, and generates a clear and readable flowchart.

[0045] In step 5, the system uses the intelligent agent swarm (VI) to generate XML code snippets from the validated semi-structured data in modules according to the BPMN 2.0 specification. This process is based on the BPMN 2.0 meta-model structure, mapping core elements in the business process such as Activity, Event, Gateway, Sequence Flow, and Swimlane to their corresponding XML structures. The generation of each module must strictly adhere to the BPMN 2.0 XML Schema definition to ensure that the generated code snippets conform to the `BPMN 2.0 XML Interchange Format` standard at the syntactic level.

[0046] This step is commonly used in practical applications for cross-departmental business process modeling scenarios, such as financial approvals and supply chain management. Through modular generation, the system can support parallel development and distributed verification, improving modeling efficiency and maintainability.

[0047] Furthermore, this step lays the structural foundation for subsequent code integration and global verification, ensuring that there are no conflicts at the syntax level among the modules, and providing reliable input for generating a complete business process model file that conforms to the BPMN 2.0 standard.

[0048] The multi-agent collaborative modeling method for complex text semantic association in this invention can effectively parse multi-granularity semantic associations in complex texts, improving the completeness and accuracy of business process modeling. Through multi-agent collaboration and a three-layer progressive verification mechanism, dynamic verification and intelligent error correction of the model are achieved, significantly enhancing the automation and standardization of the modeling process.

[0049] Example 2 Figure 2 This is a structural diagram of a multi-agent modeling and intelligent error correction system for complex semantic business processes according to an embodiment of the present invention.

[0050] like Figure 2 As shown, a multi-agent collaborative business process semantic parsing and modeling device includes: The preliminary extraction result generation module is used to decompose business process elements based on the BPMN meta-model, parse the BPMN specification, form a structured framework for business process modeling, process unstructured text in parallel through intelligent agent swarms, identify key concepts of business processes and map them into semi-structured information, and generate preliminary extraction results. The preliminary verification and information completion module verifies the completeness of the preliminary extraction results based on the BPMN specification, generates an information completeness verification report, and retrieves and completes the missing information of the preliminary extraction results through the knowledge base based on the information completeness verification report. The secondary verification and correction module is used to perform secondary verification on the completed extraction results. After confirming that the extraction results are complete and conform to the BPMN syntax rules, the module corrects the completed extraction results based on business logic and user requirements. The flowchart generation module is used to generate XML code and visual flowcharts that conform to the BPMN specification based on the corrected extraction results.

[0051] Furthermore, the preliminary extraction result generation module is also used for: Based on the BPMN metamodel, the core elements are defined and the BPMN specification is parsed to form a structured framework for business process modeling; Based on the BPMN meta-model, the core elements are defined to identify key concepts of business processes in non-institutionalized texts. The key concepts of the business process are set as semi-structured information according to the BPMN specification to form preliminary collection results.

[0052] Furthermore, the flowchart generation module is also used for: The flowchart is modified according to the user's instructions, and an evaluation report is output that includes three dimensions: syntactic compliance, semantic completeness, and visual readability. Finally, a standard BPMN2.0 model file is generated, and the system principle diagram and flowchart are summarized.

[0053] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned multi-agent modeling and intelligent error correction method for business processes oriented towards complex semantics.

[0054] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0055] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A method for multi-agent modeling and intelligent error correction of business processes with complex semantics, characterized in that, include: S1, based on the BPMN meta-model, decomposes business process elements, parses the BPMN specification, forms a structured framework for business process modeling, and uses intelligent agents to process unstructured text in parallel, identify key concepts of the business process and map them into semi-structured information, generating preliminary extraction results. S2, verify the integrity of the preliminary extraction results based on the BPMN specification, and generate an information integrity verification report; S3, based on the information integrity verification report, retrieves and completes the missing information from the initial extraction results through the knowledge base; S4. Perform a second verification on the completed extraction results. After confirming that the extraction results are complete and conform to the BPMN syntax rules, modify the completed extraction results based on business logic and user needs. S5 generates XML code and a visual flowchart that conform to the BPMN specification based on the corrected extraction results.

2. The method as described in claim 1, characterized in that, S1 further includes: S11, based on the BPMN meta-model, defines core elements to parse the BPMN specification and forms a structured framework for business process modeling; S12, Based on the BPMN meta-model, define the core elements and identify key concepts of business processes in non-institutionalized texts. S13, set the key concepts of the business process as semi-structured information according to the BPMN specification to form the preliminary collection results.

3. The method as described in claim 1, characterized in that, The method for correcting the extracted results based on business logic and user needs in step S4 is as follows: The accuracy of the business logic and the consistency of user requirements are verified based on the semi-structured information. If there is a deviation between the extracted results and the business logic or user requirements, the extracted results will be corrected.

4. The method as described in claim 1, characterized in that, The method for generating XML code conforming to the BPMN specification in step S5 is as follows: Based on the corrected extraction results, XML code snippets of the BPMN 2.0 specification are generated in modules. The XML code of each module is merged, and the code syntax and element relationships are verified to generate a complete XML file. The modules include active nodes, connectors and swimlane definitions.

5. The method as described in claim 4, characterized in that, The method for generating the visual flowchart in step S5 is as follows: Based on the complete XML file, the intelligent rendering engine of BPMN.js integrates a genetic algorithm for automatic layout optimization, defines a three-dimensional layout evaluation function including the number of intersections, path length, and node spacing, and generates a clear and readable flowchart.

6. The method as described in claim 1, characterized in that, Step S3 describes the knowledge base domain concepts, rules, and constraints.

7. A multi-agent collaborative business process semantic parsing and modeling system, characterized in that, include: The preliminary extraction result generation module is used to decompose business process elements based on the BPMN meta-model, parse the BPMN specification, form a structured framework for business process modeling, process unstructured text in parallel through intelligent agent swarms, identify key concepts of business processes and map them into semi-structured information, and generate preliminary extraction results. The preliminary verification and information completion module verifies the completeness of the preliminary extraction results based on the BPMN specification, generates an information completeness verification report, and retrieves and completes the missing information of the preliminary extraction results through the knowledge base based on the information completeness verification report. The secondary verification and correction module is used to perform secondary verification on the completed extraction results. After confirming that the extraction results are complete and conform to the BPMN syntax rules, the module corrects the completed extraction results based on business logic and user requirements. The flowchart generation module is used to generate XML code and visual flowcharts that conform to the BPMN specification based on the corrected extraction results.

8. The system as described in claim 7, characterized in that, The preliminary extraction result generation module is also used for: Based on the BPMN metamodel, the core elements are defined and the BPMN specification is parsed to form a structured framework for business process modeling; Based on the BPMN meta-model, the core elements are defined to identify key concepts of business processes in non-institutionalized texts. The key concepts of the business process are set as semi-structured information according to the BPMN specification to form preliminary collection results.

9. The system as described in claim 7, characterized in that, The flowchart generation module is also used for: The flowchart is modified according to the user's instructions, and an evaluation report is output that includes three dimensions: syntactic compliance, semantic completeness, and visual readability. Finally, a standard BPMN2.0 model file is generated, and the system principle diagram and flowchart are summarized.

10. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method as claimed in any one of claims 1-6.

Images