File process management method and system based on dynamic distribution and consensus tracking

Abstract

This invention relates to the field of document workflow management technology, specifically to a document workflow management method and system based on dynamic distribution and consensus tracking. The method involves receiving original document project credentials and workflow initiation metadata, performing project scenario context analysis and context association analysis to generate a project scenario feature list; generating standardized project workflow paths through project flow path analysis and multi-objective project optimization strategies; initiating project consensus verification items at each project consensus verification node; compiling project consensus verification minutes and archiving them chronologically using project evidence management standards to form a project consensus traceability dossier; and based on the project consensus traceability dossier, evaluating project workflow efficiency and extracting project knowledge value to generate a project workflow review report and organize project knowledge modules for updates. This invention achieves dynamic intelligent planning of document workflow paths, reliable solidification of consensus processes, and closed-loop reuse of workflow knowledge.

Description

Technical Field

[0001] This invention relates to the field of document workflow management technology, specifically to a document workflow management method and system based on dynamic distribution and consensus tracking. Background Technology

[0002] In the field of organizational collaboration and office automation, the cross-departmental flow of documents is a fundamental activity for ensuring information transmission and driving decision-making. Currently, most common document workflow management systems use predefined workflow templates, distributing documents to pre-defined recipients based on fixed organizational structures or role permissions, and recording simple approval statuses. This type of approach achieves formalization and partial automation of workflows, providing basic support for daily office work.

[0003] As organizational complexity increases and project scenarios diversify, the aforementioned static rule-based management model is gradually revealing its limitations. On the one hand, document flow paths often rely on manual experience to set, making it difficult to dynamically and intelligently adjust based on the specific content, urgency, and current task context, potentially leading to low flow efficiency or omissions of key steps. On the other hand, existing technologies focus more on the state transitions and result recording of process nodes, lacking systematic technical solutions for how participants reach consensus during the flow, how to ensure the authenticity, reliability, and traceability of operational evidence, and how to transform scattered experience into reusable organizational knowledge after the process ends. Although some research has attempted to introduce technologies such as digital twins and knowledge graphs to enhance the modeling and simulation capabilities of data entities or project processes, their core objectives are mostly focused on monitoring and predicting data status or process compliance, and have not yet effectively addressed the comprehensive challenge of how to achieve intelligent path adaptation, reliable consensus solidification, and in-depth value mining in dynamic, multi-node collaborative document flow. Summary of the Invention

[0004] In view of the above problems, the present invention provides a file flow management method and system based on dynamic distribution and consensus tracking. By intelligently planning the flow path, trustworthy solidification of node consensus and extracting process knowledge, it solves the problems of rigid path, difficulty in tracing consensus and difficulty in reusing experience in the file collaboration process.

[0005] To achieve the above objectives, in a first aspect, this application provides a document flow management method based on dynamic distribution and consensus tracing, comprising:

[0006] Receive original documents, project vouchers, and project content analysis to be transferred;

[0007] The project scenario context is analyzed on the original project vouchers, and the context association analysis is performed on the process initiation metadata to generate a project scenario feature list containing the core attributes of the voucher project and the elements of the process initiation scenario.

[0008] Based on the project scenario feature list query, the organization's project resource map and historical project process case ledger are used to generate a standardized project flow path that includes multiple project consensus verification nodes and their logical relationships and project handling criteria through project flow path analysis and multi-objective project optimization strategies.

[0009] Based on the standardized project flow path, the original project documents are synchronously generated into project document flow copies, and the project document flow copies are driven to flow to each project consensus verification node in sequence according to the standardized project flow path. At each project consensus verification node, project consensus verification matters containing project handling specification guidelines are initiated.

[0010] During the process of handling project consensus verification matters at each project consensus verification node, the supporting evidence for project handling and standardized project feedback information are collected. The supporting evidence for project handling, standardized project feedback information, node project identity information and traceability identifiers of previous project events are compiled into project consensus verification minutes. The project consensus verification minutes are then archived chronologically using the project evidence management standard to form a project consensus traceability file.

[0011] After the project document circulation copy has completed the entire circulation of the standardized project circulation path, the project process efficiency is evaluated and the project knowledge value is extracted based on the project consensus traceability dossier, generating a project process review report and organizing project knowledge modules.

[0012] The project process review report and the organization's project knowledge module are fed back to the organization's project resource map and historical project process case ledger to update the project node handling capability file and optimize the project flow path analysis strategy.

[0013] Furthermore, the original project vouchers are analyzed for project scenario context, and the process initiation metadata is analyzed for context association, generating a project scenario feature list containing core attributes of the voucher project and elements of the process initiation scenario, including:

[0014] The project context of the original project documents is analyzed. The document text content is extracted from the original project documents by optical character recognition algorithm, the audio description information is extracted from the original project documents by speech recognition algorithm, and the visual style information is extracted from the original project documents by image recognition algorithm. From this, the document text content, audio description information and visual style information are obtained.

[0015] The text content of the voucher, audio description information and visual style information are input into a pre-trained big language model. The big language model performs semantic understanding and information fusion, and outputs a structured project summary, core project body, key project items and project attitude judgment of the original document project voucher. The structured project summary, core project body, key project items and project attitude judgment together constitute the core attributes of the voucher project.

[0016] Perform contextual analysis on the process initiation metadata to extract the process initiator identifier, preset urgency level, associated project number, and associated meeting number from the process initiation metadata;

[0017] Based on the process initiator identifier, query the organizational structure ledger to obtain the department and position information to which the process initiator identifier belongs;

[0018] Based on the associated project number and associated meeting number, query the project management ledger and meeting management ledger to obtain the historical project document set and historical meeting minutes set related to the original project voucher;

[0019] By integrating departmental information, job information, preset urgency levels, historical project document collections, and historical meeting minutes collections with the core attributes of the voucher project, a project scenario feature list is generated.

[0020] Furthermore, the text content of the voucher, audio description information, and visual style information are input into a pre-trained large language model. The large language model performs semantic understanding and information fusion, outputting a structured project summary, core project content, key project matters, and project attitude judgment of the original document project voucher, including:

[0021] The voucher text content, audio description information and visual style information are preprocessed and aligned. The preprocessing includes word segmentation and stop word removal of the voucher text content, segmentation and noise reduction of the audio description information, and text region detection and chart element recognition of the visual style information.

[0022] The preprocessed voucher text content, audio description information and visual style information are respectively converted into input vector sequences that can be accepted by the pre-trained large language model. The input vector sequences include text word embedding vector sequences, audio feature vector sequences and image feature vector sequences.

[0023] Input the text word embedding vector sequence, audio feature vector sequence, and image feature vector sequence into the multimodal fusion encoder of the pre-trained large language model;

[0024] In the multimodal fusion encoder, the correlation weights between the text word embedding vector sequence, the audio feature vector sequence, and the image feature vector sequence are calculated through a cross-modal attention mechanism. Based on the correlation weights, the text word embedding vector sequence, the audio feature vector sequence, and the image feature vector sequence are weighted and fused to generate a unified context-aware feature representation.

[0025] Based on context-aware feature representation, the decoder of the pre-trained large language model performs the sequence generation task and outputs a structured project summary. The structured project summary summarizes the theme, purpose and core conclusions of the original document project voucher in the form of natural language paragraphs.

[0026] Based on context-aware feature representation, the named entity recognition and relation extraction module of the pre-trained large language model identifies and extracts the institutions, personnel, locations, times and professional terms mentioned in the original document project vouchers, forming the core project body;

[0027] Based on context-aware feature representation, the argument mining module of a pre-trained large language model identifies statements expressing opinions, judgments, or decisions in original document project vouchers, and extracts the core propositions of the statements to form key project items.

[0028] Based on context-aware feature representation, the project attitude analysis module of the pre-trained large language model classifies the overall tone of the original document project vouchers and the emotional color of each claim in key project matters, and outputs the project attitude judgment.

[0029] Furthermore, based on the project scenario feature list query, the project resource map and historical project process case ledger are organized. Through project flow path analysis and multi-objective project optimization strategies, a standardized project flow path is generated, which includes multiple project consensus verification nodes and their logical relationships, as well as project handling criteria.

[0030] Input the project scenario feature list into the path analysis engine. Based on the core attributes of the voucher project and the elements of the process initiation scenario in the project scenario feature list, the path analysis engine retrieves a set of historical process cases with similarity higher than a preset threshold from the historical project process case ledger.

[0031] Extract the final distribution path and process performance indicators for each historical process case in the historical process case set. The process performance indicators include the total consensus reaching time, the average feedback quality score of nodes, and the quality score of the final output data.

[0032] The project scenario feature list is matched with the organizational project resource map, which contains an entity network consisting of personnel nodes, department nodes, and role nodes, as well as relationship edges that represent the collaborative relationships, reporting relationships, and knowledge domain associations between entities.

[0033] Using the project scenario feature list as the query condition, perform attention-based graph relationship node association calculation in the organizational project resource graph to calculate the relevance score of each personnel node in the organizational project resource graph relative to the project scenario feature list.

[0034] Based on the relevance score, personnel nodes with a relevance score higher than the first threshold are selected from the organizational project resource graph and used as a set of candidate project consensus verification nodes.

[0035] Based on the final distribution path and process efficiency indicators of the historical process case set, as well as the topological relationship position and historical performance data of each candidate project consensus verification node in the candidate project consensus verification node set in the organizational project resource map, a multi-objective optimization model is constructed.

[0036] The multi-objective optimization model aims to minimize the prediction consensus time, maximize the prediction feedback quality, and balance the node performance load, while using process compliance rules and node performance status as constraints.

[0037] The multi-objective project optimization strategy is invoked to solve the multi-objective optimization model, generating a set of assignment path schemes containing multiple Pareto optimal solutions. Each assignment path scheme contains a node sequence consisting of multiple candidate project consensus verification nodes, the logical relationship between nodes in the node sequence, and the project disposal criteria specified for each node.

[0038] From the set of assignment path options, an assignment path option is selected as a standardized project flow path according to the preset strategy selection rules. The strategy selection rules include efficiency-first strategy, quality-first strategy, or risk control strategy.

[0039] Furthermore, based on the standardized project workflow, a project voucher workflow copy is generated synchronously from the original project voucher documents, including:

[0040] Create a project voucher flow copy data structure that is uniquely associated with the original project voucher. The project voucher flow copy data structure includes a content kernel module, a process status module, and a relationship network module.

[0041] Store a copy of the original document's project voucher content, the core attributes of the voucher project, and a structured project summary in the content kernel module;

[0042] Write a complete description of the standardized project flow path, including the multiple project consensus verification nodes included in the standardized project flow path, the logical relationship between the multiple project consensus verification nodes, and the project handling criteria specified for each project consensus verification node, into the process status module, and initialize the current processing pointer of the process status module to point to the starting project consensus verification node in the standardized project flow path.

[0043] In the relationship network module, establish the association between the project credential circulation copy and the process initiator identifier, establish the association between the project credential circulation copy and the associated project number, and establish the association between the project credential circulation copy and the candidate project consensus verification node set in the organization project resource graph.

[0044] Generate an initial set of value tags for the project voucher circulation copy. The set of value tags includes a knowledge density score calculated based on the core attributes of the voucher project, a process priority determined based on a preset urgency level, and an expected impact score predicted based on process efficiency indicators from a set of historical process cases.

[0045] Register the project voucher circulation copy to the project voucher circulation copy management service to complete the instantiation of the original project voucher.

[0046] Furthermore, it drives the project credential transfer copy to flow sequentially to each project consensus verification node according to the standardized project transfer path, and initiates project consensus verification matters containing project handling guidelines at each project consensus verification node, including:

[0047] S601: Determine the current unprocessed project consensus verification node based on the project consensus verification node pointed to by the current processing pointer in the process status module;

[0048] S602: Based on the project handling criteria specified for the consensus verification node of the current pending project in the standardized project flow path, and combined with the core attributes of the credential project in the content kernel module and the structured project summary, generate personalized operation guidance for the consensus verification node of the current pending project. The personalized operation guidance includes text paragraphs that need to be focused on, suggested decision options, and relevant historical reference case summaries.

[0049] S603: Push a notification to the terminal system corresponding to the consensus verification node of the currently pending project, which includes the access link to the project certificate circulation copy, the personalized operation guide and the processing time limit for the project consensus verification.

[0050] S604: When the user corresponding to the consensus verification node of the currently pending project accesses the project certificate transfer copy through the terminal system, the content copy of the original file project certificate, the personalized operation guide, and the current process status of the project certificate transfer copy are presented.

[0051] S605: After the consensus verification node of the currently pending project completes the processing, receive the processing result data from the terminal system. The processing result data includes an operation type identifier, operation content text, and an optional structured feedback form.

[0052] S606: Write the processing result data into the process status module of the project certificate transfer copy, and update the current processing pointer to point to the next project consensus verification node in the standardized project transfer path;

[0053] Repeat steps S601-S606 until all project consensus verification nodes in the standardized project flow path have been processed, or the process status module is marked as terminated.

[0054] Furthermore, during the process of handling project consensus verification matters at each project consensus verification node, supporting evidence for project handling and standardized project feedback information are collected. This supporting evidence, standardized project feedback information, node project identity information, and traceability identifiers of preceding project events are compiled into project consensus verification minutes, including:

[0055] When a user performs an operation through the terminal system at the consensus verification node of the current pending project, user interaction events are collected in real time. These user interaction events include annotation operations on the original document project certificate copy, selection operations on decision options in personalized operation guidance, filling operations in the structured feedback form, calling operations of the electronic signature component, and file download or printing operations.

[0056] The collected user interaction events are serialized and feature extracted to generate supporting evidence for project handling. The supporting evidence for project handling includes the type of operation, the operation timestamp, the target content location of the operation, and the specific content data generated by the operation.

[0057] Standardized project feedback information is extracted from the processing results data. The standardized project feedback information includes support or opposition indicators for key project matters, modification suggestion text, and quantitative scores and explanations of reasons filled in the structured feedback form.

[0058] Obtain the unique node project identity information of the consensus verification node of the currently pending project;

[0059] Query the project consensus traceability dossier and obtain the hash value of the most recently successfully generated project consensus verification record as the traceability identifier of the preceding project event. If the project consensus traceability dossier is empty, the genesis block identifier will be used as the traceability identifier of the preceding project event.

[0060] The supporting evidence for project handling, standardized project feedback information, node project identity information, and traceability identifiers of previous project events are serialized and packaged according to the preset block data structure to generate the original block data to be stored.

[0061] Perform cryptographic hash operations on the original block data to be stored to generate the current block hash value;

[0062] The current block hash value is appended to the original block data to be certified, forming a complete project consensus verification record.

[0063] Furthermore, after the standardized project workflow path has been fully traversed through the project credential transfer copy, the project process efficiency is assessed and project knowledge value is extracted based on the project consensus traceability dossier, generating a project process review report and organizing project knowledge modules, including:

[0064] Traverse all project consensus verification minutes in the project consensus traceability dossier, and extract the node project identity information, operation timestamp, operation action type and standardized project feedback information recorded in each project consensus verification minute;

[0065] Based on the extracted operation timestamps, calculate the flow interval between adjacent project consensus verification nodes in the standardized project flow path, the processing time of each project consensus verification node, and the total consensus time from the starting project consensus verification node to the final project consensus verification node.

[0066] Based on the extracted operation action types and standardized project feedback information, the processing quality of each project consensus verification node is quantitatively evaluated, and a node quality score is generated. The quantitative evaluation dimensions include the level of detail of the feedback, the degree of conformity with the personalized operation guidelines, and the adoption rate of the proposed modification suggestions.

[0067] By aggregating and analyzing the total consensus duration, flow interval duration, processing duration, and node quality scores, the process bottleneck nodes and efficient collaboration node sequences in the standardized project flow path are identified.

[0068] Based on the standardized project feedback information recorded in all project consensus verification minutes in the project consensus traceability file, especially the modification suggestion text and reason explanation, the set of dispute points, optimization schemes and decision basis related to the core content of the original project vouchers are extracted through text clustering algorithm and key phrase extraction algorithm.

[0069] The set of points of contention, the set of optimization solutions, and the set of decision-making basis are associated with the identity information of the node project and the location of the corresponding original document project voucher content, and encapsulated into an organizational project knowledge module that can be independently indexed and referenced;

[0070] The system integrates bottleneck nodes, efficient collaboration node sequences, total consensus duration, node quality scores, and organizational project knowledge modules to generate a project process review report according to a preset report template. The project process review report includes a process effectiveness summary, node performance analysis, knowledge contribution statistics, and process optimization suggestions.

[0071] Furthermore, based on the extracted operation action types and standardized project feedback information, the processing quality of each project consensus verification node is quantitatively evaluated, generating a node quality score, including:

[0072] Define scoring rules and weighting coefficients for each dimension of the quantitative assessment;

[0073] Regarding the level of detail in feedback, we analyze the text length, number of suggested modification texts, and completeness of explanations in the feedback information of standardized projects, and calculate the level of detail sub-score based on the preset level of detail grading standards.

[0074] Regarding the fit dimension with personalized operation guidelines, the operation action type and decision options in the standardized project feedback information are compared with the decision options and focus points suggested in the personalized operation guidelines. The matching degree between the operation and the guidelines is calculated, and a fit sub-score is generated.

[0075] Regarding the adoption status of proposed modification suggestions, the modification suggestion text proposed by the current project consensus verification node is tracked in the project consensus traceability dossier. It is checked whether the modification suggestion text is explicitly or partially adopted in the processing of subsequent project consensus verification nodes, and the adoption status sub-score is calculated based on the adoption results.

[0076] The detailedness sub-score, relevance sub-score, and adoption sub-score are multiplied by their respective weight coefficients and then summed in a weighted manner to obtain the node quality score.

[0077] The node project identity information is associated with and stored with the corresponding node quality score, which is used to update the historical performance data of the corresponding personnel nodes in the organization's project resource map.

[0078] In a second aspect, the present invention also provides a document flow management system based on dynamic distribution and consensus tracking, applicable to the method described in the first aspect. The system includes a document injection and perception module, a path deduction and planning module, a digital twin and process execution module, a consensus collection and storage module, a retrospective analysis and knowledge extraction module, and a feedback learning and optimization module. The document injection and perception module is used to receive the original document project vouchers to be transferred and to analyze and interpret the project content. It performs project scenario context analysis on the original document project vouchers and performs context association analysis on the process initiation metadata to generate a project scenario feature list containing the core attributes of the voucher project and the elements of the process initiation scenario. The path deduction and planning module is used to query the organizational project resource map and historical project process case ledger based on the project scenario feature list. Through project flow path analysis and multi-objective project optimization strategies, it generates a standardized project flow path containing multiple project consensus verification nodes and their logical relationships and project disposal criteria. The digital twin and process execution module is used to generate a project voucher flow copy simultaneously from the original document project vouchers according to the standardized project flow path and drive the project voucher flow. The certificate circulation copy flows sequentially to each project consensus verification node according to the standardized project circulation path. At each project consensus verification node, a project consensus verification matter containing project handling specifications is initiated. The consensus collection and evidence storage module is used to collect supporting evidence for project handling and standardized project feedback information during the processing of project consensus verification matters at each project consensus verification node. The supporting evidence for project handling, standardized project feedback information, node project identity information, and traceability identifiers of previous project events are compiled into a project consensus verification minutes. The project consensus verification minutes are then archived chronologically using project evidence storage management specifications to form a project consensus traceability dossier. The review analysis and knowledge extraction module is used to evaluate project process efficiency and extract project knowledge value based on the project consensus traceability dossier after the project certificate circulation copy has completed the entire circulation of the standardized project circulation path. This generates a project process review report and an organizational project knowledge module. The feedback learning and optimization module is used to feed back the project process review report and the organizational project knowledge module to the organizational project resource map and historical project process case ledger to update the project node handling capability archive and optimize the project circulation path evaluation strategy.

[0079] Unlike existing technologies, the above technical solution provides a file flow management method and system based on dynamic distribution and consensus tracking. It receives original file project credentials and flow initiation metadata, performs project scenario context analysis and context association analysis, and generates a project scenario feature list. Based on this vector query, it organizes the project resource map and historical project flow case ledger, and generates a standardized project flow path through project flow path analysis and multi-objective project optimization strategies. According to this path, the file is instantiated into a flow copy of the project credential and its flow is driven, initiating project consensus verification matters at each project consensus verification node. It collects supporting evidence for project handling and standardized project feedback information, compiles it into a project consensus verification summary, and archives it chronologically using project evidence management standards, forming a project consensus traceability dossier. After the flow is completed, it performs project flow efficiency analysis and project knowledge value extraction based on the project consensus traceability dossier, generating a project flow review report and organizing project knowledge modules, and providing feedback updates. This invention achieves dynamic intelligent planning of file flow paths, reliable solidification of consensus processes, and closed-loop reuse of process knowledge.

[0080] The above description of the invention is merely an overview of the technical solution of this application. In order to enable those skilled in the art to better understand the technical solution of this application and to implement it based on the description and drawings, and to make the above-mentioned objectives and other objectives, features and advantages of this application easier to understand, the following description is provided in conjunction with the specific embodiments and drawings of this application. Attached Figure Description

[0081] The accompanying drawings are only used to illustrate the principles, implementation methods, applications, features, and effects of specific embodiments of the present invention and other related contents, and should not be considered as limitations on this application.

[0082] In the accompanying drawings of the instruction manual:

[0083] Figure 1 This is a schematic diagram illustrating steps S101 to S107 of the method described in the specific implementation embodiment;

[0084] Figure 2 This is a schematic diagram illustrating steps S201 to S206 of the method described in a specific implementation.

[0085] Figure 3 This is a schematic diagram illustrating steps S301 to S308 of the method described in a specific implementation.

[0086] Figure 4 This is a schematic diagram illustrating steps S401 to S409 of the method described in a specific embodiment;

[0087] Figure 5This is a schematic diagram of the structure of the document workflow management system described in a specific implementation.

[0088] The reference numerals used in the above figures are explained as follows:

[0089] 1. Document workflow management system;

[0090] 11. File Injection and Detection Module;

[0091] 12. Path deduction and planning module;

[0092] 13. Digital twin and process execution module;

[0093] 14. Consensus collection and evidence storage module;

[0094] 15. Retrospective analysis and knowledge extraction module;

[0095] 16. Feedback Learning and Optimization Module. Detailed Implementation

[0096] To illustrate the possible application scenarios, technical principles, implementable specific solutions, and achievable objectives and effects of this application in detail, the following description, in conjunction with the listed specific embodiments and accompanying drawings, provides a detailed explanation. The embodiments described herein are merely illustrative of the technical solutions of this application and are therefore intended to limit the scope of protection of this application.

[0097] In this document, the term "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The term "embodiment" appearing in various places throughout the specification does not necessarily refer to the same embodiment, nor does it specifically limit its independence or connection with other embodiments. In principle, in this application, as long as there are no technical contradictions or conflicts, the technical features mentioned in each embodiment can be combined in any way to form corresponding implementable technical solutions.

[0098] Unless otherwise defined, the technical terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the use of related terms herein is merely for the purpose of describing particular embodiments and is not intended to limit this application.

[0099] In the description of this application, the term "and / or" is used to describe the logical relationship between objects, indicating that three relationships can exist. For example, A and / or B means: A exists, B exists, and A and B exist simultaneously. Additionally, the character " / " in this document generally indicates that the preceding and following objects have an "or" logical relationship.

[0100] In this application, terms such as “first” and “second” are used only to distinguish one entity or operation from another, and do not necessarily require or imply any actual quantity, hierarchy or order between these entities or operations.

[0101] Without further limitations, the use of terms such as “comprising,” “including,” “having,” or other similar open-ended expressions in this application is intended to cover non-exclusive inclusion, which does not exclude the presence of additional elements in a process, method, or product that includes the stated elements, such that a process, method, or product that includes a list of elements may include not only those defined elements but also other elements not expressly listed, or elements inherent to such a process, method, or product.

[0102] As understood in the Examination Guidelines, in this application, expressions such as "greater than," "less than," and "exceeding" are understood to exclude the stated number; expressions such as "above," "below," and "within" are understood to include the stated number. Furthermore, in the description of the embodiments in this application, "multiple" means two or more (including two), and similar expressions related to "multiple" are also understood in this way, such as "multiple groups" and "multiple times," unless otherwise explicitly specified.

[0103] In the description of the embodiments of this application, the space-related expressions used, such as "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "vertical," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential," indicate the orientation or positional relationship based on the orientation or positional relationship shown in the specific embodiments or drawings. They are only for the purpose of describing the specific embodiments of this application or for the reader's understanding, and do not indicate or imply that the device or component referred to must have a specific position, a specific orientation, or be constructed or operated in a specific orientation. Therefore, they should not be construed as limitations on the embodiments of this application.

[0104] The processor described in the embodiments of this application can be implemented by hardware, firmware, software, or a combination thereof. It can be a circuit, one or more of an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field-programmable gate array (FPGA), a central processing unit (CPU), a controller, a microcontroller, or a microprocessor. It also includes other physical, biological, or chemical structures that can implement the same or equivalent functions as the processors listed above, such as biological neurons, quantum computing units, DNA computing units, etc., so that the processor can execute some or all of the steps in the computer program or method involved in the various embodiments of this application, or any combination of the steps mentioned therein.

[0105] The computer program involved in the embodiments can be stored in a computer device readable storage medium, which includes, but is not limited to, disks, magnetic tapes, magnetic cards, floppy disks, flash memory, optical disks, optical cards, read-only memory (ROM), random access memory (RAM), erasable programmable ROM (EPROM), and electrically erasable programmable ROM (EEPROM), etc., and also includes other biological, physical, or chemical structures that can achieve the same or equivalent functions as the storage media listed above, such as DNA, RNA, proteins, and other units with information storage capabilities. In specific embodiments, the storage medium involved can be one of the above-mentioned media types, or a combination of the above-mentioned media types. In different embodiments, the computer program involved in the embodiments can be centrally stored in a single medium, or distributed and stored in multiple media. The memory containing the computer device readable storage medium can be non-volatile memory or random access memory. These computer device readable storage media can be built into the device, or can be connected to the device involved in the embodiments as an external device or part of an external device. In some embodiments, the memory having a computer device readable storage medium is deployed locally; in other embodiments, the memory may be deployed remotely from the processor, for example, as a network-attached memory accessed via RF circuitry or an external port and a communication network, wherein the communication network may be the Internet, one or more intranets, a local area network (LAN), a wide area network (WLAN), a storage area network (SAN), or a suitable combination thereof, as long as computer device access to the memory is enabled. Furthermore, the computer program involved in the embodiments may be stored in plaintext / ciphertext form, or it may be designed as training data, integrated and recombined through model training and implicitly stored in the parameter states of a deep neural network or other machine learning model.

[0106] Please see Figure 1 In a first aspect, this embodiment provides a file flow management method based on dynamic distribution and consensus tracing, including:

[0107] S101. Receive the original documents, project vouchers, and project content analysis to be transferred;

[0108] S102. Perform project scenario context analysis on the original document project vouchers and perform context association analysis on the process initiation metadata to generate a project scenario feature list containing the core attributes of the voucher project and the elements of the process initiation scenario.

[0109] S103. Based on the project scenario feature list query, organize the project resource map and historical project process case ledger, and through project flow path analysis and multi-objective project optimization strategy, generate a standardized project flow path containing multiple project consensus verification nodes and their logical relationships and project disposal criteria.

[0110] S104. Based on the standardized project flow path, the original project voucher is synchronously generated into a project voucher flow copy, and the project voucher flow copy is driven to flow to each project consensus verification node in sequence according to the standardized project flow path. At each project consensus verification node, a project consensus verification matter containing the project disposal specification guide is initiated.

[0111] S105. During the process of handling project consensus verification matters at each project consensus verification node, collect supporting evidence for project handling and standardized project feedback information. Organize the supporting evidence for project handling, standardized project feedback information, node project identity information and traceability identifiers of previous project events into a project consensus verification record. Then, use the project evidence management standard to archive the project consensus verification record in chronological order to form a project consensus traceability file.

[0112] S106. After the project voucher circulation copy has completed the entire circulation of the standardized project circulation path, the project process efficiency assessment and project knowledge value extraction are carried out based on the project consensus traceability dossier, and a project process review report and project knowledge module are generated.

[0113] S107. Feed the project process review report and the organization's project knowledge module back to the organization's project resource map and historical project process case ledger to update the project node handling capability file and optimize the project flow path analysis strategy.

[0114] In step S101, the format of the original document project certificate may include various types such as document, table, presentation, etc.; the process initiation metadata is auxiliary information submitted together with the original document project certificate to initiate and define this process flow. It is usually provided by the process initiator or automatically attached by the system according to the context. For example, it may include document title, initiator identifier, preset urgency level label, associated project or meeting number, etc.

[0115] In step S102, the original document project voucher undergoes project scenario context analysis to extract the semantics contained in its various modal information, such as text, images, tables, and even audio and video. Contextual association analysis of the process initiation metadata is performed to understand the background and environment of this flow, such as analyzing the initiator's organizational role and previous documents of related projects. The core attributes of the voucher project are structured representations extracted from the document content, such as themes, key entities, and core viewpoints; the process initiation scenario elements represent the attributes and constraints of the flow task. The project scenario feature list is generated by vectorizing and fusing the above features and contextual information, providing a unified, machine-understandable data input for subsequent intelligent decision-making.

[0116] In step S103, the organizational project resource graph is a networked knowledge base used to model the personnel, departments, roles, and their collaborative relationships and knowledge domain associations within an organization; the historical project process case ledger stores past completed document flow instances marked with final paths and performance indicators. By querying the organizational project resource graph and the historical project process case ledger, it is possible to find past experiences similar to the current task and understand the capabilities and relationships of various entities within the organization. Through project flow path analysis, the propagation and aggregation of information or tasks along relational edges can be simulated on the topological structure of the organizational project resource graph, thereby assessing the potential relevance or processing capabilities of different nodes (personnel or departments) to the current task; the multi-objective project optimization strategy is used to comprehensively weigh multiple objectives such as processing efficiency, resource load balancing, and expected output quality, while satisfying process rule constraints, to search for and determine an optimal or near-optimal node sequence and its collaborative logic from candidate nodes. The final standardized project flow path not only clarifies which project consensus verification nodes need to participate, but also specifies their sequential and parallel logical relationships, as well as the specific project handling criteria that each node needs to complete. This step enables the file flow path to be generated adaptively based on file content, organizational status, and historical experience, rather than relying on a fixed template.

[0117] In step S104, a project voucher circulation copy is synchronously generated from the original document project voucher. This creates a digital mirror image corresponding to the physical document, containing its content, status, relationships, and behavioral logic. The project voucher circulation copy serves as the carrier for process execution, and its internal status (such as the current processing node and completed operations) can be updated in real time. Project consensus verification not only notifies the corresponding nodes for processing but also provides targeted project handling guidelines based on the document content and the node's project handling criteria. These guidelines may include key paragraphs requiring review or decision options, thereby improving the efficiency and standardization of node processing.

[0118] In step S105, when a user responds to a project consensus verification request and performs an operation at a project consensus verification node, the system collects the supporting evidence for the project's actions in real time, such as annotations, signatures, and option selections on documents. Simultaneously, it receives standardized project feedback information submitted by the user, such as review comments, scores, and explanations of reasons. Node project identity information is used to uniquely identify which node generated this data. The traceability identifier for preceding project events originates from previously recorded and stored project consensus verification minutes, used to establish temporal and logical connections between blocks. The generated project consensus verification minutes constitute an indivisible data unit recording a single node's processing behavior. When forming a project consensus traceability dossier, the project evidence management standard can be based on hash chain or blockchain technology. The project consensus traceability dossier completely and tamper-proofly records all consensus and operational traces reached at each project consensus verification node throughout the entire lifecycle of the document, achieving credible solidification and full traceability of process evidence.

[0119] In step S106, project process efficiency is assessed based on the project consensus traceability dossier. Indicators such as processing time of each stage, collaboration intervals between nodes, and feedback quality are extracted from the chain to evaluate the overall efficiency, bottlenecks, and node performance. Project knowledge value extraction utilizes text analysis and clustering methods to extract reusable knowledge particles such as points of contention, solutions, and decision-making basis from standardized project feedback information and modification suggestions recorded in the chain. The project process review report is a systematic summary of the above efficiency analysis results, while the organizational project knowledge module structures and encapsulates the extracted knowledge, enabling it to be independently stored, retrieved, and referenced. This step realizes the transformation from process data to management insights and organizational assets.

[0120] In step S107, the data on node performance in the project process review report can be used to update the capability profiles of the corresponding nodes (personnel or departments) in the organizational project resource map, such as adjusting their capability scores or efficiency tags for handling certain types of tasks. The organizational project knowledge module, as well as the standardized project workflow path and final performance indicators generated in this workflow, are stored as new cases in the historical project workflow case ledger. These updates continuously enrich the system's knowledge reserves, enabling the generation of better path planning based on a more accurate capability model and richer success cases during path deduction in step S103, thereby continuously optimizing the project workflow path judgment strategy.

[0121] This embodiment transforms the traditional static, result-oriented document flow into a dynamic, process-transparent, knowledge-accumulating intelligent collaborative process. It comprehensively understands documents and tasks by generating a project scenario feature list, dynamically queries the organizational knowledge base based on this list, and intelligently plans standardized project flow paths using graph neural networks and multi-objective project optimization strategies. It drives the flow by creating project credential copies and initiates concrete project consensus verification items at nodes, achieving automated and precise execution of the process. Furthermore, it collects operations and feedback throughout the process and uses project evidence management standards to form a project consensus traceability dossier, ensuring process auditability. After the process ends, it performs performance analysis and knowledge extraction based on the project consensus traceability dossier, generating a debriefing report and knowledge fragments, which are fed back to the knowledge base to continuously optimize organizational cognition and deduction strategies. This embodiment achieves a closed loop from static execution to dynamic intelligence, from process recording to consensus evidence storage, and from task completion to knowledge feedback, improving the adaptability, credibility, and knowledge transformation efficiency of document flow.

[0122] Please see Figure 2 In some embodiments, the original document project credentials are analyzed for project scenario context, and the process initiation metadata is analyzed for context association to generate a project scenario feature list containing the core attributes of the credential project and the elements of the process initiation scenario, including:

[0123] S201. Analyze the project scenario context of the original project voucher, extract the voucher text content from the original project voucher using optical character recognition algorithm, extract the audio description information from the original project voucher using speech recognition algorithm, and extract the visual style information from the original project voucher using image recognition algorithm, thereby obtaining the voucher text content, audio description information and visual style information.

[0124] S202. Input the voucher text content, audio description information and visual style information into the pre-trained big language model. The big language model performs semantic understanding and information fusion, and outputs the original document voucher's structured project summary, core project body, key project items and project attitude judgment. The structured project summary, core project body, key project items and project attitude judgment together constitute the core attributes of the voucher project.

[0125] S203. Perform contextual analysis on the process initiation metadata and extract the process initiator identifier, preset urgency level, associated project number and associated meeting number from the process initiation metadata.

[0126] S204. Query the organizational structure ledger based on the process initiator identifier to obtain the department and position information to which the process initiator identifier belongs;

[0127] S205. Based on the associated project number and associated meeting number, query the project management ledger and meeting management ledger to obtain the set of historical project documents and the set of historical meeting minutes related to the original project vouchers.

[0128] S206. Integrate department information, job information, preset urgency level, historical project document collection and historical meeting minutes collection with the core attributes of the voucher project to generate a project scenario feature list.

[0129] In step S201, the optical character recognition algorithm processes the scanned images or picture format content contained in the original document project voucher, converting the printed or handwritten text into machine-readable voucher text content; the speech recognition algorithm transcribes the speech portion of the audio or video content attached to the original document project voucher to generate voucher text content; the image recognition algorithm analyzes visual elements such as charts, photos, and diagrams in the original document project voucher, extracting descriptive labels or key object classification information as visual style information.

[0130] In step S202, the pre-trained large language model refers to a deep neural network model pre-trained on massive multimodal corpora (including text, image-text pairs, audio-text pairs, etc.) through self-supervised learning tasks (such as masked language modeling and contrastive learning), possessing cross-modal semantic understanding and generation capabilities. The credential text content, audio-transcribed text, and visual descriptive labels are input into this model, and the model performs semantic understanding and information fusion. The structured project summary is a text paragraph generated by the model that summarizes the core content and conclusions of the document in natural language; the core project body is a list of named entities with specific meanings in the document identified by the model, such as names of people, organizations, and professional terms; key project items are a set of core propositional statements expressing viewpoints or decisions extracted by the model from the document; and the project attitude judgment is the model's classification result of the overall or partial emotional tone of the document.

[0131] In step S203, the process initiator identifier is information that uniquely identifies the submitter of the document, such as user ID or employee number; the preset urgency level is a classification label that is pre-set by the initiator or the system to indicate the urgency of the process; the associated project number and associated meeting number are unique codes used to link to external project management and meeting management ledgers.

[0132] In step S204, the organizational structure ledger stores the affiliation of personnel, departments, and roles. By querying, the name and code of the department to which the initiator belongs, as well as the organizational position information, can be obtained.

[0133] In step S205, the project management ledger is queried based on the associated project number to obtain historical documents, reports, etc. under the project, forming a historical project document set; the meeting management ledger is queried based on the associated meeting number to obtain the minutes, resolutions, and attachments of the meeting, forming a historical meeting minutes set.

[0134] In step S206, vectorization encoding converts the voucher text content, such as department information, job information, and historical document content summaries, into fixed-dimensional numerical vectors. This is achieved by using a bag-of-words model combined with TF-IDF statistics or a pre-trained word embedding model to obtain sentence vectors. Feature concatenation connects the core attributes of the voucher item (which are already or can be converted into vectors) with the vectorized department information vector, job information vector, urgency label vector (which can be one-hot encoded), and the overall semantic vector of the historical document set, combining them into a high-dimensional list of project scenario features.

[0135] This embodiment extracts the original information through a multimodal parsing algorithm and transforms it into structured semantic features using a pre-trained large language model. At the same time, it parses process metadata and queries external systems to obtain organizational background and historical context. Finally, through project feature integration, it merges content semantics and process context to form a unified numerical representation, providing high-quality input that deeply integrates document content and flow environment for subsequent intelligent path planning.

[0136] Please see Figure 3 In some embodiments, the voucher text content, audio description information, and visual style information are input into a pre-trained large language model. The large language model performs semantic understanding and information fusion, outputting a structured project summary, core project content, key project matters, and project attitude judgment of the original document project voucher, including:

[0137] S301. Preprocess and align the voucher text content, audio description information and visual style information. The preprocessing includes word segmentation and stop word removal of the voucher text content, segmentation and noise reduction of the audio description information, and text region detection and chart element recognition of the visual style information.

[0138] S302. Convert the preprocessed voucher text content, audio description information and visual style information into input vector sequences that can be accepted by the pre-trained large language model. The input vector sequences include text word embedding vector sequences, audio feature vector sequences and image feature vector sequences.

[0139] S303. Input the text word embedding vector sequence, audio feature vector sequence and image feature vector sequence into the multimodal fusion encoder of the pre-trained large language model;

[0140] S304. In the multimodal fusion encoder, the cross-modal attention mechanism is used to calculate the correlation weights between the text word embedding vector sequence, the audio feature vector sequence, and the image feature vector sequence, and based on the correlation weights, the text word embedding vector sequence, the audio feature vector sequence, and the image feature vector sequence are weighted and fused to generate a unified context-aware feature representation;

[0141] S305. Based on the context-aware feature representation, the decoder of the pre-trained large language model performs a sequence generation task to output a structured project summary, which summarizes the theme, purpose, and core conclusion of the original document project voucher in the form of a natural language paragraph;

[0142] S306. Based on the context-aware feature representation, the named entity recognition and relation extraction module of the pre-trained large language model identifies and extracts the institutions, personnel, locations, times, and professional terms mentioned in the original document project voucher to form the core project entity;

[0143] S307. Based on the context-aware feature representation, the argument mining module of the pre-trained large language model identifies the sentences expressing viewpoints, judgments, or decisions in the original document project voucher and extracts the core propositions of the sentences to form the key project matters;

[0144] S308. Based on the context-aware feature representation, the project attitude analysis module of the pre-trained large language model classifies the overall tone of the original document project voucher and the emotional color of each claim in the key project matters to output a project attitude determination.

[0145] In step S301, when processing the voucher text content, word segmentation uses a word segmentation tool based on a dictionary or a statistical model to split the continuous text into independent lexical units; stop word removal filters out high-frequency but low-semantic contribution words such as "de" and "shi" according to a predefined common list of non-content words. When processing the audio description information, speech segmentation divides the speech paragraphs through a silence detection or speaker change point detection algorithm, and noise reduction suppresses background noise through a filtering algorithm to improve the speech recognition accuracy. When processing the visual style information, text region detection uses a connected text proposal network or a deep learning-based text detection model to locate the text regions in the image, and chart element recognition uses an object detection model to identify and classify the types of elements such as bar charts, line charts, and pie charts in the chart.

[0146] In step S302, the text word embedding vector sequence is obtained by mapping each word after word segmentation into a high-dimensional vector using a pre-trained word embedding model (such as Word2Vec or BERT). The audio feature vector sequence is obtained by extracting temporal feature vectors from the denoised audio segments using an audio feature extraction network (such as a Mel spectrogram followed by a convolutional neural network). The image feature vector sequence is obtained by extracting visual feature vectors from the detected text regions or chart element regions of the image using an image feature extractor (such as a pre-trained convolutional neural network).

[0147] In steps S303 to S304, the multimodal fusion encoder of the pre-trained large language model is used to process and fuse multiple modal inputs into the neural network layer. The cross-modal attention mechanism obtains relevance weights by calculating the relevance score between any two vectors in the text word embedding vector sequence, audio feature vector sequence, and image feature vector sequence (e.g., by performing a softmax normalization after the dot product of the query vector and the key vector). Based on the relevance weights, the vector sequences from different modalities are weighted and summed or weighted and concatenated to achieve information complementarity and enhancement, generating a unified context-aware feature representation that integrates multimodal information.

[0148] In step S305, the decoder of the pre-trained large language model is usually based on the Transformer architecture and works in an autoregressive manner. It takes the context-aware feature representation as the initial condition, generates the probability distribution of the next word step by step, and finally outputs a coherent natural language paragraph, namely a structured project summary, through iterative prediction. This summary summarizes the theme, purpose and core conclusions of the original document project credential.

[0149] In step S306, the named entity recognition and relation extraction module typically uses a sequence labeling model based on context-aware feature representation (such as a conditional random field combined with a bidirectional long short-term memory network) to label each position in the feature representation sequence with entity type labels (such as organization, personnel, location, time, professional terminology), and extracts consecutive sequences of the same type of labels as entities to form the core project body.

[0150] In step S307, the argument mining module identifies semantic segments containing viewpoints or judgments by analyzing context-aware feature representations. This module can use a classifier to determine whether a sentence is an argument sentence, or extract argument components through sequence labeling, thereby extracting propositional statements that express core claims and forming key project items.

[0151] In step S308, the project attitude analysis module includes a classifier (such as a fully connected neural network layer with a Softmax activation function), which takes the overall pooled vector of context-aware feature representation or the feature vector corresponding to the key claims as input, performs multi-classification (such as positive, negative, neutral), and outputs the overall project attitude judgment of the original document project voucher and the sentiment color classification of each claim in the key project matters.

[0152] This embodiment provides a regular input to the model through refined preprocessing and vectorization of multi-source information; it utilizes a multimodal fusion encoder and cross-modal attention mechanism to achieve deep interaction and fusion of text, audio, and visual style information, forming a context-aware feature representation; based on the context-aware feature representation, the model's various dedicated modules (decoder, named entity recognition module, argument mining module, and project attitude analysis module) work in parallel or serially to generate structured summaries, entity lists, claim sets, and sentiment tags, respectively, ensuring that the semantic features extracted from the original multimodal files are comprehensive and deeply structured.

[0153] Please see Figure 4 In some embodiments, based on a project scenario feature list, the system queries the organizational project resource map and historical project process case ledger. Through project flow path analysis and multi-objective project optimization strategies, a standardized project flow path is generated, containing multiple project consensus verification nodes and their logical relationships, as well as project handling criteria. This includes:

[0154] S401. Input the project scenario feature list into the path analysis engine. Based on the core attributes of the voucher project and the elements of the process initiation scenario in the project scenario feature list, the path analysis engine retrieves a set of historical process cases with similarity higher than a preset threshold from the historical project process case ledger.

[0155] S402. Extract the final distribution path and process efficiency indicators for each historical process case in the historical process case set. The process efficiency indicators include the total consensus reaching time, the average feedback quality score of nodes, and the quality score of the final output data.

[0156] S403. Match the project scenario feature list with the organizational project resource map. The organizational project resource map contains an entity network consisting of personnel nodes, department nodes, and role nodes, as well as relationship edges that represent the collaborative relationships, reporting relationships, and knowledge domain associations between entities.

[0157] S404. Using the project scenario feature list as the query condition, perform graph relationship node association calculation based on attention mechanism in the organization project resource graph, and calculate the relevance score of each personnel node in the organization project resource graph relative to the project scenario feature list.

[0158] S405. Based on the relevance score, select personnel nodes with a relevance score higher than the first threshold from the organization's project resource graph, and use them as a set of candidate project consensus verification nodes.

[0159] S406. Based on the final distribution path and process efficiency indicators of the historical process case set, as well as the topological relationship position and historical performance data of each candidate project consensus verification node in the candidate project consensus verification node set in the organizational project resource map, construct a multi-objective optimization model.

[0160] S407. The multi-objective optimization model aims to minimize the prediction consensus time, maximize the prediction feedback quality, and balance the workload of nodes, with process compliance rules and node performance status as constraints.

[0161] S408. Call the multi-objective project optimization strategy to solve the multi-objective optimization model and generate a set of assignment path schemes containing multiple Pareto optimal solutions. Each assignment path scheme contains a node sequence consisting of multiple candidate project consensus verification nodes, the logical relationship between nodes in the node sequence, and the project disposal criteria specified for each node.

[0162] S409. From the set of assignment path schemes, select an assignment path scheme as a standardized project flow path according to the preset strategy selection rules. The strategy selection rules include efficiency-first strategy, quality-first strategy, or risk control strategy.

[0163] In step S401, the path analysis engine is a software module used to analyze and plan document flow paths. It retrieves cases by calculating the similarity between the project scenario feature list and the context vector corresponding to each case in the historical project process case ledger (e.g., using cosine similarity or matching scores based on Siamese neural networks). The preset threshold can be determined by analyzing the similarity distribution of historical successful cases.

[0164] In step S402, the total consensus time refers to the total time elapsed from process initiation to final node completion; the average node feedback quality score is the average of the feedback quality of all processing nodes in the case; and the final output data quality score is the post-evaluation score of the case's output documents. These process performance indicators can be directly extracted from the metadata records of historical process cases.

[0165] In step S403, personnel nodes, department nodes, and role nodes in the project resource graph are connected by relationship edges. The relationship edges may have type attributes (such as "reported to", "collaborated with", "area of ​​expertise") and weight attributes (such as collaboration frequency).

[0166] In step S404, the graph relation node association calculation based on the attention mechanism is specifically implemented through a graph attention network. The project scene feature list is used as the global query vector, and the feature vectors (such as skill tags and historical processing domains) of each personnel node in the project resource graph are organized as key vectors. By calculating the dot product between the query and each key vector and normalizing it, the attention weight of each node relative to the current file context is obtained, i.e., the relevance score, which quantifies the potential adaptability of personnel nodes in processing the current file.

[0167] In step S405, the first threshold is used to screen out highly relevant candidates from all personnel nodes. It can be set by selecting several nodes with the highest relevance scores, or nodes with scores exceeding the average score of all nodes plus one standard deviation, to form a set of candidate project consensus verification nodes.

[0168] In step S406, when constructing the multi-objective optimization model, the predicted consensus duration is estimated by regression based on the duration of similar paths in the historical process case set, the average processing speed of candidate nodes, and the communication overhead caused by inter-node collaboration. The predicted feedback quality is predicted based on the historical performance data of candidate nodes recorded in the organizational project resource map (such as the average score of similar past tasks) and its matching degree with the core attributes of the current credential project. The node performance load is quantified based on the number of tasks currently assigned to but not completed by the candidate node.

[0169] In step S407, process compliance rules may include: the requirement to include nodes with specific functional roles, the required approval hierarchy, etc.; node performance status may include whether they are online, busy, or on leave. Process compliance rules and node performance status are transformed into constraints for the model, such as using inequalities or equality constraints to limit the node sequence to include a certain node or exclude unavailable nodes.

[0170] In step S408, the multi-objective project optimization strategy invoked is, for example, a non-dominated sorting genetic algorithm, which iteratively evolves the path schemes encoded by decision variables, generates new schemes through crossover and mutation, and performs non-dominated sorting and crowding calculation based on multiple objective function values, ultimately converging to a set containing multiple non-dominated Pareto optimal solutions. Each solution corresponds to a feasible assignment path scheme, and each scheme defines in detail the execution order (serial or parallel), dependencies, and customized project handling criteria (such as review priorities and decision-making authority) for each node.

[0171] In step S409, the efficiency-first strategy selects the scheme with the shortest predicted consensus time from the Pareto solution set; the quality-first strategy selects the scheme with the highest predicted average feedback quality score; and the risk control strategy selects the scheme with the smallest node performance load variance and the most balanced load. Based on the selected strategies, the final standardized project flow path is determined from the allocation path scheme set.

[0172] This embodiment achieves content-based case retrieval through a path analysis engine, accurately quantifies the relevance of personnel and tasks in the organizational graph using graph attention networks, and constructs a multi-objective optimization model integrating historical performance, node workload, and process rules. An evolutionary algorithm is then used to solve for the Pareto optimal path set. This method represents a leap from relying on fixed rules or manual experience to the automatic generation of file flow paths based on data and models, enabling dynamic, intelligent, and multi-objective trade-offs. By generating standardized project flow paths, not only are the participating nodes and their order clarified, but the processing logic and requirements are also defined, laying the foundation for the automated execution and precise control of subsequent processes. This improves the scientific rigor, adaptability, and overall efficiency of file distribution in complex collaborative scenarios.

[0173] In some embodiments, based on a standardized project workflow path, a project voucher workflow copy is synchronously generated from the original project voucher document, including:

[0174] Create a project voucher flow copy data structure that is uniquely associated with the original project voucher. The project voucher flow copy data structure includes a content kernel module, a process status module, and a relationship network module.

[0175] Store a copy of the original document's project voucher content, the core attributes of the voucher project, and a structured project summary in the content kernel module;

[0176] Write a complete description of the standardized project flow path, including the multiple project consensus verification nodes included in the standardized project flow path, the logical relationship between the multiple project consensus verification nodes, and the project handling criteria specified for each project consensus verification node, into the process status module, and initialize the current processing pointer of the process status module to point to the starting project consensus verification node in the standardized project flow path.

[0177] In the relationship network module, establish the association between the project credential circulation copy and the process initiator identifier, establish the association between the project credential circulation copy and the associated project number, and establish the association between the project credential circulation copy and the candidate project consensus verification node set in the organization project resource graph.

[0178] Generate an initial set of value tags for the project voucher circulation copy. The set of value tags includes a knowledge density score calculated based on the core attributes of the voucher project, a process priority determined based on a preset urgency level, and an expected impact score predicted based on process efficiency indicators from a set of historical process cases.

[0179] Register the project voucher circulation copy to the project voucher circulation copy management service to complete the instantiation of the original project voucher.

[0180] In this embodiment, the project credential circulation copy data structure is a composite data object created in computer memory or a database, used to fully represent the original document project credential and its circulation lifecycle. Specifically, the content kernel module persistently stores the content copy of the original document project credential, the core attributes of the credential project, and a structured project summary; the process status module manages and tracks the document circulation process, internally containing a complete description of the standardized project circulation path, including all project consensus verification nodes in the path, the logical relationships between nodes, and the project disposal criteria specified for each node. The current processing pointer maintained by this module is initialized to point to the starting project consensus verification node in the path upon instantiation. Thus, the project credential circulation copy, through the standardized project circulation path written in its process status module, essentially establishes a strong association with a specific node sequence selected from the candidate project consensus verification node set for this circulation; the relationship network module records the association context of the project credential circulation copy in the organizational environment by establishing links with the process initiator identifier, associated project number, and reference relationships with the candidate project consensus verification node set in the organizational project resource graph.

[0181] An initial set of value tags is generated for the project voucher circulation copies, assigning them meta-attributes to assist in process scheduling and resource allocation. Among them, the knowledge density score is calculated based on the core attributes of the voucher project, such as the scale and professionalism of the core project entity and the depth of argumentation of key project matters, and is obtained by weighted summation or by using a pre-trained voucher text content volume evaluation model; the process priority is directly determined by mapping the preset urgency level in the process initiation metadata; the expected impact score is predicted based on the process efficiency indicators (especially the quality score of the final output data) of similar cases in the historical process case set, which can be the average of these historical scores or calculated by weighted average with case similarity as the weight.

[0182] Registering the project credential transfer copy to the project credential transfer copy management service involves registering the instance identifier, access interface, and metadata of the data structure in a centralized service registry or management system. This enables other modules to discover, locate, and drive the twin to complete the instantiation of the original file project credential.

[0183] This embodiment creates a holographic digital mapping for the document. The project credential circulation copy data structure encapsulates the document content and semantics, embeds the circulation path and processing logic, and establishes organizational relationships. By calculating value tags such as knowledge density score, process priority, and expected impact score, the document is given quantifiable meta-attributes. After registering with the project credential circulation copy management service, the project credential circulation copy becomes the core carrier for automated process execution and collaborative interaction, realizing the transformation of the document from a static document to an intelligent entity with state, behavioral logic, and contextual relationships.

[0184] In some embodiments, the project credential transfer copy is driven to flow sequentially to each project consensus verification node according to the standardized project transfer path, and at each project consensus verification node, a project consensus verification item containing project disposal specification guidelines is initiated, including:

[0185] S601: Determine the current unprocessed project consensus verification node based on the project consensus verification node pointed to by the current processing pointer in the process status module;

[0186] S602: Based on the project handling criteria specified for the consensus verification node of the current pending project in the standardized project flow path, and combined with the core attributes of the credential project in the content kernel module and the structured project summary, generate personalized operation guidance for the consensus verification node of the current pending project. The personalized operation guidance includes text paragraphs that need to be focused on, suggested decision options, and relevant historical reference case summaries.

[0187] S603: Push a notification to the terminal system corresponding to the consensus verification node of the currently pending project, which includes the access link to the project certificate circulation copy, the personalized operation guide and the processing time limit for the project consensus verification.

[0188] S604: When the user corresponding to the consensus verification node of the currently pending project accesses the project certificate transfer copy through the terminal system, the content copy of the original file project certificate, the personalized operation guide, and the current process status of the project certificate transfer copy are presented.

[0189] S605: After the consensus verification node of the currently pending project completes the processing, receive the processing result data from the terminal system. The processing result data includes an operation type identifier, operation content text, and an optional structured feedback form.

[0190] S606: Write the processing result data into the process status module of the project certificate transfer copy, and update the current processing pointer to point to the next project consensus verification node in the standardized project transfer path;

[0191] Repeat steps S601-S606 until all project consensus verification nodes in the standardized project flow path have been processed, or the process status module is marked as terminated.

[0192] In this embodiment, the current processing pointer points to the pending project consensus verification node in the standardized project workflow path, thereby determining the current pending project consensus verification node. When generating personalized operation guidance, the project disposal criteria specified for that node in the standardized project workflow path are combined with the core attributes of the voucher project and the structured project summary in the content kernel module. For example, a text matching algorithm is used to locate paragraphs related to the keywords of the project disposal criteria in the original document project voucher content copy, which are then identified as text paragraphs requiring special attention; suggested decision options that match the document topic and project disposal criteria are generated based on historical cases or a predefined rule base; and cases with similar themes or entities are retrieved from the historical project process case ledger, and their summaries are extracted as historical reference case summaries.

[0193] A notification regarding project consensus verification is pushed to the terminal system corresponding to the currently pending project consensus verification node. This can be sent via message queue or application programming interface (API), and includes a link to access the project credential transfer copy, personalized operation instructions, and processing time limits. When the corresponding user accesses the project credential transfer copy through the terminal system, the interface displays a copy of the original project credential file, personalized operation instructions, and the current progress status of the project credential transfer copy.

[0194] After the consensus verification node of the current pending project completes its processing, the terminal system submits the processing result data. The operation type identifier distinguishes processing actions, such as "Agree" or "Reject"; the operation content text records the specific opinions; an optional structured feedback form contains preset scoring items or multiple-choice questions. The received processing result data is written to the process status module of the project credential transfer copy, stored as the processing record for that node; simultaneously, the current processing pointer is updated to point to the next project consensus verification node in the standardized project transfer path; if there are no subsequent nodes in the path, or the process status module is marked as terminated, the loop ends.

[0195] This embodiment achieves automated circulation and precise interaction of project voucher copies. It enhances processing targeting by dynamically generating personalized operation guides that integrate file content and node tasks; ensures task reach by pushing notifications of integrated links and time-limited project consensus verification matters; and achieves coherent recording and driving of the process through standardized result data reception and status updates. This process transforms static paths into dynamic, interactive, and traceable actual process execution.

[0196] In some embodiments, during the process of handling project consensus verification matters at each project consensus verification node, supporting evidence for project handling and standardized project feedback information are collected. The supporting evidence for project handling, standardized project feedback information, node project identity information, and traceability identifiers of preceding project events are then compiled into a project consensus verification summary, including:

[0197] When a user performs an operation through the terminal system at the consensus verification node of the current pending project, user interaction events are collected in real time. These user interaction events include annotation operations on the original document project certificate copy, selection operations on decision options in personalized operation guidance, filling operations in the structured feedback form, calling operations of the electronic signature component, and file download or printing operations.

[0198] The collected user interaction events are serialized and feature extracted to generate supporting evidence for project handling. The supporting evidence for project handling includes the type of operation, the operation timestamp, the target content location of the operation, and the specific content data generated by the operation.

[0199] Standardized project feedback information is extracted from the processing results data. The standardized project feedback information includes support or opposition indicators for key project matters, modification suggestion text, and quantitative scores and explanations of reasons filled in the structured feedback form.

[0200] Obtain the unique node project identity information of the consensus verification node of the currently pending project;

[0201] Query the project consensus traceability dossier and obtain the hash value of the most recently successfully generated project consensus verification record as the traceability identifier of the preceding project event. If the project consensus traceability dossier is empty, the genesis block identifier will be used as the traceability identifier of the preceding project event.

[0202] The supporting evidence for project handling, standardized project feedback information, node project identity information, and traceability identifiers of previous project events are serialized and packaged according to the preset block data structure to generate the original block data to be stored.

[0203] Perform cryptographic hash operations on the original block data to be stored to generate the current block hash value;

[0204] The current block hash value is appended to the original block data to be certified, forming a complete project consensus verification record.

[0205] In this embodiment, user interaction events are collected through a front-end event listener or browser extension, recording the underlying events of user interaction with interface elements. When generating supporting evidence for project handling, serialization converts complex event objects into a standard format (such as JSON strings or Protocol Buffers encoding) for easy transmission and storage; feature extraction extracts key fields from the serialized data: the operation action type is mapped to a predefined operation code based on the collected event type (such as click, input, change); the operation timestamp is obtained from the system clock at the time the event occurred; the target content location of the operation is located through the document object model path, element unique identifier, or character offset relative to the file content; the specific content data generated by the operation directly records the values ​​in the event payload, such as the input text and the selected option identifier.

[0206] Support or opposition markers in standardized project feedback information are annotated after sentiment or intent analysis of the operation content text using natural language processing technology; modification suggestion text is extracted from paragraphs containing specific suggestion sentence structures or keywords from the operation content text; quantitative scores and explanations of reasons are directly read from the data fields submitted in the structured feedback form.

[0207] Node project identity information is typically associated with the login user session of the terminal system or a pre-configured node-user mapping table. The preceding project event tracing identifier is obtained by querying the persistently stored project consensus tracing dossier data to obtain the hash value of the latest block. Preferably, the project consensus tracing dossier specifically refers to a unique chain structure built in chronological order for the current project credential transfer copy. If the project consensus tracing dossier is empty, a predefined genesis block identifier (such as an all-zero hash) is used as the preceding project event tracing identifier.

[0208] The default block data structure includes a block header and a block body. The block header fields include the version number, timestamp, previous project event traceability identifier, and the reserved position of the current block hash value. The block body stores serialized evidence of project actions, standardized project feedback information, and node project identity information. Serialization and packaging use JSON, XML, or a custom binary format to encode the above data items and populate the block body.

[0209] Cryptographic hashing applies a hash function (such as SHA-256) to the entire original block data obtained after serialization and packaging. The output fixed-length hash digest is the hash value of the current block. Writing this current block hash value into a reserved field in the block header constitutes a complete project consensus verification record.

[0210] This embodiment collects all user interactions, serializes and structures them to generate project handling evidence containing operational details and feedback content; it designs a block data structure with a header and body structure and associates it with the preceding hash and node identity to ensure the integrity and chain association of data units; finally, it uses a cryptographic hash function to generate a unique identifier and constructs a tamper-proof project consensus verification record, transforming each node processing into an independently verifiable and time-sequentially rigorous evidence storage unit, providing a standardized data unit for building a complete project consensus traceability dossier, and solidifying the loose consensus in the process into a trustworthy on-chain record.

[0211] In some embodiments, after the project credential transfer copy has completed the entire standardized project workflow, the project process efficiency is assessed and project knowledge value is extracted based on the project consensus traceability dossier, generating a project process review report and an organizational project knowledge module, including:

[0212] Traverse all project consensus verification minutes in the project consensus traceability dossier, and extract the node project identity information, operation timestamp, operation action type and standardized project feedback information recorded in each project consensus verification minute;

[0213] Based on the extracted operation timestamps, calculate the flow interval between adjacent project consensus verification nodes in the standardized project flow path, the processing time of each project consensus verification node, and the total consensus time from the starting project consensus verification node to the final project consensus verification node.

[0214] Based on the extracted operation action types and standardized project feedback information, the processing quality of each project consensus verification node is quantitatively evaluated, and a node quality score is generated. The quantitative evaluation dimensions include the level of detail of the feedback, the degree of conformity with the personalized operation guidelines, and the adoption rate of the proposed modification suggestions.

[0215] By aggregating and analyzing the total consensus duration, flow interval duration, processing duration, and node quality scores, the process bottleneck nodes and efficient collaboration node sequences in the standardized project flow path are identified.

[0216] Based on the standardized project feedback information recorded in all project consensus verification minutes in the project consensus traceability file, especially the modification suggestion text and reason explanation, the set of dispute points, optimization schemes and decision basis related to the core content of the original project vouchers are extracted through text clustering algorithm and key phrase extraction algorithm.

[0217] The set of points of contention, the set of optimization solutions, and the set of decision-making basis are associated with the identity information of the node project and the location of the corresponding original document project voucher content, and encapsulated into an organizational project knowledge module that can be independently indexed and referenced;

[0218] The system integrates bottleneck nodes, efficient collaboration node sequences, total consensus duration, node quality scores, and organizational project knowledge modules to generate a project process review report according to a preset report template. The project process review report includes a process effectiveness summary, node performance analysis, knowledge contribution statistics, and process optimization suggestions.

[0219] In this embodiment, the flow interval is determined based on the difference in operation timestamps recorded in the consensus verification minutes of adjacent projects; the processing time is calculated by the difference between the earliest and latest operation timestamps recorded in the corresponding block of the consensus verification node of the same project. If there is only one timestamp, it can be estimated by combining the preset standard processing time or the timestamps of adjacent nodes; the total consensus time is the time span between the first operation timestamp of the starting node and the last operation timestamp of the final node.

[0220] The level of detail in the feedback is quantified by statistically analyzing the number of characters in the text fields of the standardized project feedback information, the number of unique words after word segmentation, and the number of independent entries in the suggested modification text. The fit with the personalized operation guide is evaluated by calculating the matching degree between the user's actual operation action type, the selected option identifier, and the predefined suggested action and option set in the personalized operation guide. The adoption rate of the suggested modification is determined by searching for keywords or semantically similar expressions in the suggested modification text of the current node in the standardized project feedback information or operation content text of the suffix project consensus verification node.

[0221] The identification criteria for bottleneck nodes in the process are that the processing time of the node significantly exceeds the average or median processing time of all nodes in the standardized project workflow. The efficient collaborative node sequence consists of multiple consecutive nodes, whose processing times are all below the average, whose node quality scores are above the average, and whose workflow intervals between nodes are all short.

[0222] The text clustering algorithm groups the suggested modification text and the text vectors explaining the reasons into groups. Texts with similar semantics are grouped into the same cluster, and the central theme or the most representative text summary of each cluster constitutes a point of contention or an optimization solution entry. The key phrase extraction algorithm analyzes the frequency distribution and co-occurrence relationship of words in each text cluster, selects the words or phrase combinations with the highest weight as key phrases, and fills them into the set of points of contention, optimization solutions, or decision basis.

[0223] Preferably, the data structure for organizing project knowledge modules includes the following fields: knowledge content (extracted focus, solution, or basis), source node project identity information, location of the associated original file project voucher content (such as chapter number, paragraph index), timestamp of knowledge generation, and knowledge type tag. The encapsulation process involves filling in the data according to this structure and serializing it into a storable format.

[0224] The project process review report's process efficiency summary section includes calculated statistical values ​​such as total consensus time, average processing time, and longest workflow interval. The node performance analysis section lists each node's project identity information, processing time, and node quality score in tabular or chart form, highlighting identified bottleneck nodes and sequences of highly efficient collaborative nodes. The knowledge contribution statistics section summarizes the total number of knowledge modules in the organization's project, categorized by knowledge type (points of contention, optimization solutions, decision-making basis). The process optimization suggestion section, based on bottleneck node analysis results and frequently occurring optimization solutions, generates specific suggestion texts such as adjusting the processing order of specific nodes, optimizing personalized operation guidance content for corresponding nodes, and strengthening inter-node collaboration and communication.

[0225] This embodiment extracts time-series, operational, and textual data from project consensus traceability dossiers. Through a series of algorithmic operations, including difference calculation, multidimensional matching, text retrieval, statistical thresholding, clustering, and key phrase extraction, it transforms the original on-chain records into quantifiable performance indicators, node quality assessments, bottleneck identification results, and structured knowledge sets. Standardized knowledge fragment data structures and report templates ensure that the analytical outputs (project process review reports and organizational project knowledge modules) have a clear format, traceable origins, and direct application value. This embodiment achieves deep data mining and knowledge enhancement of the entire document flow process, transforming one-time process execution data into assessable, optimizable, and reusable organizational assets, completing an intelligent closed loop from process recording to performance insights and knowledge accumulation.

[0226] In some embodiments, based on the extracted operation action type and standardized project feedback information, the processing quality of each project consensus verification node is quantitatively evaluated to generate a node quality score, including:

[0227] Define scoring rules and weighting coefficients for each dimension of the quantitative assessment;

[0228] Regarding the level of detail in feedback, we analyze the text length, number of suggested modification texts, and completeness of explanations in the feedback information of standardized projects, and calculate the level of detail sub-score based on the preset level of detail grading standards.

[0229] Regarding the fit dimension with personalized operation guidelines, the operation action type and decision options in the standardized project feedback information are compared with the decision options and focus points suggested in the personalized operation guidelines. The matching degree between the operation and the guidelines is calculated, and a fit sub-score is generated.

[0230] Regarding the adoption status of proposed modification suggestions, the modification suggestion text proposed by the current project consensus verification node is tracked in the project consensus traceability dossier. It is checked whether the modification suggestion text is explicitly or partially adopted in the processing of subsequent project consensus verification nodes, and the adoption status sub-score is calculated based on the adoption results.

[0231] The detailedness sub-score, relevance sub-score, and adoption sub-score are multiplied by their respective weight coefficients and then summed in a weighted manner to obtain the node quality score.

[0232] The node project identity information is associated with and stored with the corresponding node quality score, which is used to update the historical performance data of the corresponding personnel nodes in the organization's project resource map.

[0233] In this embodiment, the scoring rules are pre-set in the form of conditional judgments or piecewise functions, clearly defining the score values ​​corresponding to different data performances under each dimension; the weight coefficients reflect the relative importance of each dimension in the overall node quality score, and can be set by domain expert experience or allocated based on the statistical analysis results of the scores of each dimension in historical high-quality process cases.

[0234] Preferably, for the dimension of the level of detail in the feedback, the preset level of detail grading standard is as follows: multiple threshold ranges are set for the text length in the standardized project feedback information, and each range corresponds to a basic score; each independent suggestion in the modification suggestion text is given an additional fixed score; the completeness of the explanation of reasons is scored according to whether it includes three parts: problem description, reason analysis and specific suggestions, and full marks are obtained for that part if all three parts are included.

[0235] Regarding the relevance dimension to personalized operation guidance, the matching degree between operations and guidance is calculated by comparing the user's actual action type with the suggested action type in the personalized operation guidance. A complete match earns full marks for that action item, while a partial match earns partial marks. The decision options in the standardized project feedback information are compared with the suggested decision options in the personalized operation guidance; a match earns full marks for that item. The relevance sub-score is a weighted average of the scores for each comparison item.

[0236] Regarding the adoption status of proposed modifications, when tracing the project consensus traceability dossier, it is checked whether the standardized project feedback information or operation content text of subsequent project consensus verification nodes explicitly references the modification suggestion text proposed by the current node, or whether keywords expressing the intention of adoption appear. Explicit adoption means that subsequent nodes directly confirm the implementation of the suggestion; partial adoption means that subsequent nodes refer to the suggestion but make modifications or only adopt part of the content. Based on this judgment, explicit adoption is assigned a high score, partial adoption is assigned a medium score, and no mention is assigned a zero or low score.

[0237] The detailedness sub-score, relevance sub-score, and adoption status sub-score are each multiplied by their respective weighting coefficients and then summed in a weighted manner to obtain the node quality score. The node project identity information is then associated with and stored in relation to the corresponding node quality score. This means that, using the node project identity information as the key, the calculated node quality score is appended as a new performance record to the historical performance data list corresponding to that node, thereby updating the capability profile of the corresponding personnel node in the organizational project resource map.

[0238] This embodiment transforms node quality assessment into a repeatable and verifiable quantitative process by defining specific scoring rules, weights, and calculation logic. The calculation of three sub-scores—extensibility, relevance, and adoption—incentivizes detailed feedback, standardized operations, and effective suggestions, respectively; a weighted summation yields the overall quality score; and associated storage provides accurate data for continuous optimization of organizational understanding. This embodiment enables the objective measurement and accumulation of node performance, supporting data-driven path deduction and optimization.

[0239] Please see Figure 5In a second aspect, this embodiment also provides a file flow management system 1 based on dynamic distribution and consensus tracking, applicable to the method described in the first aspect. The system includes a file injection and perception module 11, a path deduction and planning module 12, a digital twin and process execution module 13, a consensus collection and evidence storage module 14, a retrospective analysis and knowledge extraction module 15, and a feedback learning and optimization module 16. The file injection and perception module 11 is used to receive the original file project vouchers to be transferred and analyze the project content, perform project scenario context analysis on the original file project vouchers, and perform context association analysis on the process initiation metadata to generate a project scenario feature list containing the core attributes of the voucher project and the elements of the process initiation scenario. The path deduction and planning module 12 is used to query the organizational project resource map and historical project process case ledger based on the project scenario feature list, and generate a standardized project flow path containing multiple project consensus verification nodes and their logical relationships and project disposal criteria through project flow path analysis and multi-objective project optimization strategies. The digital twin and process execution module 13 is used to synchronously generate a project voucher flow copy based on the original file project vouchers according to the standardized project flow path. The project credential transfer copy is driven to flow sequentially to each project consensus verification node according to the standardized project transfer path. At each project consensus verification node, a project consensus verification matter containing project handling specifications is initiated. The consensus collection and evidence storage module 14 is used to collect project handling supporting evidence and standardized project feedback information during the process of handling project consensus verification matters at each project consensus verification node. The project handling supporting evidence, standardized project feedback information, node project identity information, and previous project event traceability identifiers are compiled into a project consensus verification minutes, and the project consensus is stored in accordance with the project evidence storage management specifications. The verification minutes are archived chronologically to form a project consensus traceability dossier; the debriefing analysis and knowledge extraction module 15 is used to conduct project process efficiency assessment and project knowledge value extraction based on the project consensus traceability dossier after the project voucher circulation copy has completed the entire circulation of the standardized project circulation path, and generate a project process debriefing report and an organizational project knowledge module; the feedback learning and optimization module 16 is used to feed the project process debriefing report and the organizational project knowledge module back to the organizational project resource map and historical project process case ledger, so as to update the project node handling capability file and optimize the project circulation path assessment strategy.

[0240] This system achieves fully closed-loop management of intelligent file flow through modular design. The file injection and perception module 11 provides a deep understanding of the file context for the process; the path deduction and planning module 12 dynamically plans the optimal path accordingly; the digital twin and process execution module 13 transforms the path into an interactive, automated process; the consensus collection and evidence storage module 14 ensures the reliable traceability of process operations; the retrospective analysis and knowledge extraction module 15 mines the value of process data; and the feedback learning and optimization module 16 completes knowledge feedback and strategy iteration. These modules work in tandem to elevate the traditional static file flow into a dynamic, reliable, and continuously evolving intelligent collaborative process, solving systemic problems such as rigid paths, opaque processes, and difficulty in accumulating experience.

[0241] Furthermore, the above technical solution can be understood in conjunction with the following example:

[0242] In a specific implementation scenario, such as the administrative management and collaborative office environment of a university, the method and system described in this invention can effectively solve the problems of accurate distribution, consensus-building, and process tracking in cross-departmental document circulation. In this scenario, the "original document project voucher" can be various documents that need to be communicated, studied, or approved, such as teaching inspection plans, scientific research policy documents, party and government learning materials, or leaders' speeches; the "process initiation metadata" may include the document initiating department, the preset urgency level, and the associated specific project or meeting identifier.

[0243] For example, when the Academic Affairs Office initiates a "Teaching Inspection Work Plan", the system first analyzes the project scenario context of the document, extracts the text through OCR technology, and uses a pre-trained large language model to understand its semantics, outputting core entities and key claims such as "teaching management", "classroom quality assessment", "secondary colleges", and "Academic Affairs Office", which constitute the core attributes of the credential project; at the same time, combined with the metadata that the process initiator is "Academic Affairs Office", the system queries the organizational structure ledger to obtain its department and position information, and associates it with historical project and meeting data to jointly generate a "project scenario feature list" that represents the specific context of the document.

[0244] Based on this vector, the system queries the university's "Organizational Project Resource Map" (the nodes of which include various colleges, functional departments, and responsible persons, with edges representing their affiliation and collaborative relationships) and "Historical Project Process Case Ledger" (storing the circulation paths and effect data of similar teaching management documents in the past). Through project circulation path analysis, the system calculates the relevance scores of "personnel nodes" such as deans of various colleges, vice deans of teaching, and heads of relevant functional departments to the context of the document, and filters out a set of candidate project consensus verification nodes. Then, with the goal of minimizing the total circulation time and maximizing the feedback quality, a multi-objective optimization model is constructed to solve and generate an optimal "standardized project circulation path". This path may specify that the document needs to be approved by the "vice president in charge of teaching" node, verified and filed by the "university office" node, and then distributed to nodes such as "teaching offices of various secondary colleges", "quality monitoring and evaluation center" and "faculty development center" for learning and feedback.

[0245] Based on this path, the system instantiates the original document project voucher into a "project voucher circulation copy." This twin carries the document content, semantic features, and a planned circulation path. The system drives this twin to circulate along the path. For example, first, a "project consensus verification item" is pushed to the terminal of the "vice principal in charge of teaching," which includes the document content and personalized approval operation instructions (such as prompts to pay attention to inspection standard clauses). After the vice principal reviews and electronically signs online, his operation actions, signature evidence, approval opinions, etc., are collected in real time and organized together with his node project identity information to form a "project voucher." The "Consensus Verification Minutes" are linked to the "Project Consensus Tracing Dossier" after being reliably stored. Subsequently, the twin automatically flows to the "School Office" node, repeating the process of initiating and recording the project consensus verification. After the review node is completed, the twin is simultaneously distributed to each learning node (each college, etc.). When the heads of each college read the document, their actions of opening, browsing, annotating, and even finally clicking "learned and confirmed" are all recorded as evidence for project handling and standardized project feedback information (such as support or not, modification suggestions) in subsequent project consensus verification minutes, continuously extending the project consensus tracing dossier.

[0246] Once all nodes have completed their processing, the system conducts a post-mortem analysis based on the complete "Project Consensus Tracing Dossier." This allows for precise calculation of the total time from initiation to final college confirmation, identification of bottlenecks with excessively long processing intervals (e.g., delayed processing by a particular college), and evaluation of processing quality based on the detail of each node's feedback and its alignment with guidelines. Simultaneously, all modification suggestions and discussion focuses are extracted from the blockchain to form an "Organizational Project Knowledge Module" for "Teaching Inspection Scheme Optimization." The final generated project process post-mortem report and these knowledge fragments are fed back into the organization's project resource graph and historical case library. For example, the system can update the historical performance data of a college dean node in processing "Teaching Management Documents," or prioritize recommending efficient collaborative node sequences demonstrated in the current process when encountering similar documents in the future, thereby continuously optimizing the project flow path analysis strategy.

[0247] This invention transforms the traditional, linear, and untraceable document delivery process in universities into a dynamic, interactive, and learnable consensus-building process based on intelligent path planning, digital twin-driven mechanisms, and blockchain-based evidence storage. This ensures the accuracy of document distribution and the immutability of the process. Furthermore, through quantitative analysis and knowledge accumulation throughout the entire process, it provides data-driven decision support for the continuous improvement of organizational operational efficiency.

[0248] By adopting the above technical solutions, this invention differs from existing technologies and has the following beneficial effects: By generating a project scenario feature list that integrates document content and flow context, and dynamically querying the organizational project resource map and historical case library based on this, it intelligently plans standardized project flow paths using graph neural networks and multi-objective project optimization strategies, realizing the transformation of document flow paths from static presets to dynamic adaptations; by creating project credential flow copies to drive flow and initiating concrete project consensus verification items at each node, it ensures the automated and accurate execution of the process. Furthermore, by collecting operations and feedback throughout the process and using project evidence management standards to form an immutable project consensus traceability dossier, it ensures the traceability and authenticity of process evidence; after the process ends, it performs performance analysis and knowledge extraction based on the project consensus traceability dossier, generating a project process review report and organizational project knowledge modules, and feeding them back to the knowledge base to continuously optimize organizational cognition and project flow path judgment strategies, thereby constructing a complete closed loop from intelligent planning and reliable execution to in-depth review and knowledge feedback. This invention effectively solves the problems of rigid paths, opaque processes, difficulty in tracing consensus, and difficulty in accumulating and reusing experience in traditional document workflow management, and significantly improves the adaptability, reliability, and knowledge transformation efficiency of collaborative document flow.

[0249] Finally, it should be noted that although the above embodiments have been described in the text and drawings of this application, this should not limit the scope of patent protection of this application. Any technical solutions that are based on the essential concept of this application and utilize the content described in the text and drawings of this application, resulting in equivalent structural or procedural substitutions or modifications, as well as the direct or indirect application of the technical solutions of the above embodiments to other related technical fields, are all included within the scope of patent protection of this application.

Claims

1. A file flow management method based on dynamic distribution and consensus tracking, characterized in that, include: Receive original documents, project vouchers, and project content analysis for transfer; The original file project voucher is subjected to project scenario context analysis, and the process initiation metadata is subjected to context association analysis to generate a project scenario feature list containing the core attributes of the voucher project and the elements of the process initiation scenario. Based on the project scenario feature list, query the project resource map and historical project process case ledger, and through project flow path analysis and multi-objective project optimization strategies, generate a standardized project flow path that includes multiple project consensus verification nodes and their logical relationships and project handling criteria. Based on the standardized project workflow path, a project voucher workflow copy is generated synchronously from the original project voucher document, including: Create a project voucher flow copy data structure that is uniquely associated with the original project voucher. The project voucher flow copy data structure includes a content kernel module, a process status module, and a relationship network module. Store a copy of the original document's project voucher content, the core attributes of the voucher project, and a structured project summary into the content kernel module; Write a complete description of the standardized project flow path, including the multiple project consensus verification nodes included in the standardized project flow path, the logical relationship between the multiple project consensus verification nodes, and the project handling criteria specified for each project consensus verification node, into the process status module, and initialize the current processing pointer of the process status module to point to the starting project consensus verification node in the standardized project flow path. In the relationship network module, establish the association between the project credential circulation copy and the process initiator identifier, establish the association between the project credential circulation copy and the associated project number, and establish the association between the project credential circulation copy and the candidate project consensus verification node set in the organization project resource graph; Generate an initial set of value tags for the project voucher circulation copy. The set of value tags includes a knowledge density score calculated based on the core attributes of the voucher project, a process priority determined based on a preset urgency level, and an expected impact score predicted based on process efficiency indicators from a set of historical process cases. Register the project voucher circulation copy to the project voucher circulation copy management service to complete the instantiation of the original project voucher; And drive the project certificate circulation copy to circulate sequentially to each project consensus verification node according to the standardized project circulation path, and initiate project consensus verification matters containing project handling specification guidelines at each project consensus verification node; During the process of handling the project consensus verification matters at each project consensus verification node, the supporting evidence for project handling and standardized project feedback information are collected. The supporting evidence for project handling, the standardized project feedback information, the node project identity information and the traceability identifier of the preceding project event are compiled into a project consensus verification record. The project consensus verification record is then archived chronologically using the project evidence management standard to form a project consensus traceability dossier. After the project voucher circulation copy completes the entire circulation of the standardized project circulation path, the project process efficiency is evaluated and the project knowledge value is extracted based on the project consensus traceability dossier, generating a project process review report and organizing project knowledge modules. The project process review report and the organization's project knowledge module are fed back to the organization's project resource map and the historical project process case ledger to update the project node handling capability file and optimize the project flow path analysis strategy. 2.The file flow management method based on dynamic distribution and consensus tracking of claim 1, wherein, The original document project voucher is subjected to project scenario context analysis, and the process initiation metadata is subjected to context association analysis to generate a project scenario feature list containing the core attributes of the voucher project and the elements of the process initiation scenario, including: The original document project voucher is analyzed for project scenario context, and the voucher text content in the original document project voucher is extracted by optical character recognition algorithm, the audio description information in the original document project voucher is extracted by speech recognition algorithm, and the visual style information in the original document project voucher is extracted by image recognition algorithm, thereby obtaining the voucher text content, audio description information and visual style information; The voucher text content, the audio description information, and the visual style information are input into a pre-trained large language model. The large language model performs semantic understanding and information fusion, and outputs a structured project summary, core project body, key project items, and project attitude judgment of the original document project voucher. The structured project summary, core project body, key project items, and project attitude judgment together constitute the core attributes of the voucher project. Perform contextual analysis on the process initiation metadata to extract the process initiator identifier, preset urgency level, associated project number, and associated meeting number from the process initiation metadata; Based on the process initiator identifier, query the organizational structure ledger to obtain the department and position information to which the process initiator identifier belongs; Based on the associated project number and the associated meeting number, query the project management ledger and the meeting management ledger to obtain the set of historical project documents and the set of historical meeting minutes related to the original document project voucher; The department information, job information, preset urgency level, historical project document set, and historical meeting minutes set are integrated with the core attributes of the voucher project to generate the project scenario feature list. 3.The file flow management method based on dynamic distribution and consensus tracking of claim 2, wherein, The voucher text content, the audio description information, and the visual style information are input into a pre-trained large language model. The large language model performs semantic understanding and information fusion, and outputs a structured project summary, core project content, key project matters, and project attitude judgment for the original document project voucher, including: The voucher text content, the audio description information, and the visual style information are preprocessed and aligned. The preprocessing includes word segmentation and stop word removal of the voucher text content, segmentation and noise reduction of the audio description information, and text region detection and chart element recognition of the visual style information. The preprocessed voucher text content, the audio description information, and the visual style information are respectively converted into input vector sequences acceptable to the pre-trained large language model. The input vector sequences include text word embedding vector sequences, audio feature vector sequences, and image feature vector sequences. The text word embedding vector sequence, the audio feature vector sequence, and the image feature vector sequence are input into the multimodal fusion encoder of the pre-trained large language model; In the multimodal fusion encoder, the correlation weights between the text word embedding vector sequence, the audio feature vector sequence, and the image feature vector sequence are calculated through a cross-modal attention mechanism. Based on the correlation weights, the text word embedding vector sequence, the audio feature vector sequence, and the image feature vector sequence are weighted and fused to generate a unified context-aware feature representation. Based on the context-aware feature representation, the decoder of the pre-trained large language model performs a sequence generation task and outputs the structured project summary, which summarizes the theme, purpose and core conclusions of the original document project credential in the form of natural language paragraphs. Based on the context-aware feature representation, the named entity recognition and relation extraction module of the pre-trained large language model identifies and extracts the institutions, personnel, locations, times and professional terms mentioned in the original document project voucher to form the core project body; Based on the context-aware feature representation, the argument mining module of the pre-trained large language model identifies statements expressing opinions, judgments, or decisions in the original document project vouchers, and extracts the core propositions of the statements to constitute the key project items; Based on the context-aware feature representation, the project attitude analysis module of the pre-trained large language model classifies the overall tone of the original document project vouchers and the emotional color of each claim in the key project matters, and outputs the project attitude judgment. 4.The file flow management method based on dynamic distribution and consensus tracking of claim 2, wherein, Based on the project scenario feature list, the organization queries the project resource map and historical project process case ledger. Through project flow path analysis and multi-objective project optimization strategies, a standardized project flow path is generated, containing multiple project consensus verification nodes and their logical relationships, as well as project handling criteria. The project scenario feature list is input into the path analysis engine. Based on the core attributes of the voucher project and the process initiation scenario elements in the project scenario feature list, the path analysis engine retrieves a set of historical process cases with similarity higher than a preset threshold from the historical project process case ledger. Extract the final distribution path and process efficiency indicators for each historical process case in the historical process case set. The process efficiency indicators include the total consensus reaching time, the average feedback quality score of nodes, and the quality score of the final output data. The project scenario feature list is matched with the organizational project resource graph, which includes an entity network consisting of personnel nodes, department nodes, and role nodes, as well as relationship edges representing the collaboration relationships, reporting relationships, and knowledge domain associations between entities. Using the project scenario feature list as the query condition, perform graph relation node association calculation based on attention mechanism in the organization project resource graph, and calculate the relevance score of each personnel node in the organization project resource graph relative to the project scenario feature list; Based on the correlation score, personnel nodes with a correlation score higher than the first threshold are selected from the organization project resource graph and used as a set of candidate project consensus verification nodes. Based on the final distribution path and process efficiency indicators of the historical process case set, as well as the topological association position and historical performance data of each candidate project consensus verification node in the candidate project consensus verification node set in the organization's project resource map, a multi-objective optimization model is constructed. The multi-objective optimization model aims to minimize the prediction consensus time, maximize the prediction feedback quality, and balance the node performance load, while using process compliance rules and node performance status as constraints. The multi-objective project optimization strategy is invoked to solve the multi-objective optimization model, generating a set of assignment path schemes containing multiple Pareto optimal solutions. Each assignment path scheme contains a node sequence consisting of multiple candidate project consensus verification nodes, the logical relationship between the nodes in the node sequence, and the project disposal criteria specified for each node. From the set of assignment path schemes, an assignment path scheme is selected as the standardized project flow path according to a preset strategy selection rule. The strategy selection rule includes an efficiency-first strategy, a quality-first strategy, or a risk control strategy.

5. The file flow management method based on dynamic distribution and consensus tracking of claim 1, wherein, And drive the project credential transfer copy to flow sequentially to each project consensus verification node according to the standardized project transfer path, and initiate project consensus verification matters containing project handling specification guidelines at each project consensus verification node, including: S601: Determine the current unprocessed project consensus verification node based on the project consensus verification node pointed to by the current processing pointer in the process status module; S602: Based on the project handling criteria specified for the consensus verification node of the current pending project in the standardized project flow path, and combined with the core attributes of the credential project in the content kernel module and the structured project summary, generate personalized operation guidance for the consensus verification node of the current pending project. The personalized operation guidance includes text paragraphs that need to be focused on, suggested decision options, and relevant historical reference case summaries. S603: Push a notification to the terminal system corresponding to the consensus verification node of the currently pending project, which includes the access link to the project certificate circulation copy, the personalized operation guide and the processing time limit for the project consensus verification. S604: When the user corresponding to the consensus verification node of the currently pending project accesses the project certificate transfer copy through the terminal system, the content copy of the original file project certificate, the personalized operation guide, and the current process status of the project certificate transfer copy are presented. S605: After the consensus verification node of the current pending project completes the processing, receive the processing result data from the terminal system. The processing result data includes the operation type identifier, operation content text, and structured feedback form. S606: Write the processing result data into the process status module of the project certificate transfer copy, and update the current processing pointer to point to the next project consensus verification node in the standardized project transfer path; Repeat steps S601-S606 until all project consensus verification nodes in the standardized project flow path have been processed, or the process status module is marked as terminated.

6. The file flow management method based on dynamic distribution and consensus tracing according to claim 5, characterized in that, During the process of handling the project consensus verification matters at each project consensus verification node, supporting evidence for project handling and standardized project feedback information are collected. The supporting evidence for project handling, the standardized project feedback information, node project identity information, and traceability identifiers of preceding project events are compiled into a project consensus verification summary, including: When the user corresponding to the consensus verification node of the current pending project performs an operation through the terminal system, user interaction events are collected in real time. The user interaction events include annotation operations on the original document project certificate content copy, selection operations on decision options in the personalized operation guidance, filling operations in the structured feedback form, calling operations of the electronic signature component, and file download or printing operations. The collected user interaction events are serialized and feature extracted to generate supporting evidence for the project handling. The supporting evidence for the project handling includes the operation action type, operation timestamp, target content location of the operation, and specific content data generated by the operation. The standardized project feedback information is parsed from the processing result data. The standardized project feedback information includes support or opposition indicators for key project matters, modification suggestion text, and quantitative scores and reason explanations filled in the structured feedback form. Obtain the unique node project identity information of the consensus verification node of the currently pending project; Query the project consensus traceability dossier and obtain the hash value of the most recently successfully generated project consensus verification record as the traceability identifier of the preceding project event. If the project consensus traceability dossier is empty, the genesis block identifier is used as the traceability identifier of the preceding project event. The supporting evidence for the project disposal, the standardized project feedback information, the node project identity information, and the traceability identifier of the preceding project event are serialized and packaged according to a preset block data structure to generate the original block data to be stored. Perform a cryptographic hash operation on the original block data to be certified to generate the current block hash value; The current block hash value is appended to the original block data to be certified, forming a complete project consensus verification record.

7. The file flow management method based on dynamic distribution and consensus tracing according to claim 6, characterized in that, After the project credential transfer copy completes the entire transfer of the standardized project transfer path, the project process efficiency is assessed and project knowledge value is extracted based on the project consensus traceability dossier, generating a project process review report and an organizational project knowledge module, including: Traverse all project consensus verification minutes in the project consensus traceability dossier, and extract the node project identity information, operation timestamp, operation action type and standardized project feedback information recorded in each project consensus verification minute; Based on the extracted operation timestamps, calculate the flow interval between adjacent project consensus verification nodes in the standardized project flow path, the processing time of each project consensus verification node, and the total consensus time from the starting project consensus verification node to the final project consensus verification node. Based on the extracted operation action type and the standardized project feedback information, the processing quality of each project consensus verification node is quantitatively evaluated to generate a node quality score. The dimensions of the quantitative evaluation include the level of detail of the feedback, the degree of conformity with the personalized operation guidelines, and the adoption of the proposed modification suggestions. By aggregating and analyzing the total consensus duration, the flow interval duration, the processing duration, and the node quality score, the process bottleneck nodes and efficient collaboration node sequences in the standardized project flow path are identified. Based on the standardized project feedback information recorded in all project consensus verification minutes in the project consensus traceability dossier, the modification suggestion text and the reason explanation are extracted using text clustering algorithm and key phrase extraction algorithm to extract the set of disputed points, optimization schemes and decision basis related to the core content of the original document project voucher; The set of disputed points, the set of optimization solutions, and the set of decision-making basis are associated with the location of the node project identity information and the corresponding original file project voucher content, and encapsulated into an organization project knowledge module that can be independently indexed and referenced; Based on the aforementioned bottleneck nodes, the sequence of efficient collaborative nodes, the total consensus duration, the node quality scores, and the organizational project knowledge modules, a project process review report is generated according to a preset report template. The project process review report includes a process performance summary, node performance analysis, knowledge contribution statistics, and process optimization suggestions.

8. The file flow management method based on dynamic distribution and consensus tracing according to claim 7, characterized in that, Based on the extracted operation action types and the standardized project feedback information, the processing quality of each project consensus verification node is quantitatively evaluated, and a node quality score is generated, including: Define scoring rules and weighting coefficients for each dimension of the quantitative assessment; Regarding the level of detail in the feedback, the text length, number of suggested modification texts, and completeness of the explanation of reasons in the standardized project feedback information are analyzed, and a level of detail sub-score is calculated based on a preset level of detail grading standard. Regarding the fit dimension of the personalized operation guidance, the operation action type is compared with the decision options in the standardized project feedback information and the suggested decision options and focus in the personalized operation guidance. The matching degree between the operation and the guidance is calculated, and a fit sub-score is generated. Regarding the adoption status dimension of the proposed modification suggestions, the modification suggestion text proposed by the current project consensus verification node is tracked in the project consensus traceability dossier. It is checked whether the modification suggestion text is explicitly or partially adopted in the processing of subsequent project consensus verification nodes, and an adoption status sub-score is calculated based on the adoption results. The node quality score is obtained by multiplying the detailedness sub-score, the relevance sub-score, and the adoption status sub-score by their respective weight coefficients and then summing them by weight. The node project identity information is associated with and stored with the corresponding node quality score, which is used to update the historical performance data of the corresponding personnel nodes in the organization project resource map.

9. A file workflow management system based on dynamic distribution and consensus tracking, characterized in that, The system applicable to the method of any one of claims 1 to 8 comprises: The file injection and perception module is used to receive the original file project vouchers to be transferred and analyze the project content, perform project scenario context analysis on the original file project vouchers, and perform context association analysis on the process initiation metadata to generate a project scenario feature list containing the core attributes of the voucher project and the elements of the process initiation scenario. The path deduction and planning module is used to query the project resource map and historical project process case ledger based on the project scenario feature list. Through project flow path analysis and multi-objective project optimization strategies, it generates a standardized project flow path that includes multiple project consensus verification nodes and their logical relationships and project disposal criteria. The digital twin and process execution module are used to generate a project voucher flow copy from the original document project voucher according to the standardized project flow path, and drive the project voucher flow copy to flow to each project consensus verification node in sequence according to the standardized project flow path, and initiate project consensus verification matters containing project handling specification guidelines at each project consensus verification node. The consensus collection and evidence storage module is used to collect supporting evidence for project handling and standardized project feedback information during the process of handling project consensus verification matters at each project consensus verification node. The supporting evidence for project handling, the standardized project feedback information, the node project identity information and the traceability identifier of the preceding project event are organized into a project consensus verification record. The project consensus verification record is archived chronologically using the project evidence storage management standard to form a project consensus traceability dossier. The debriefing and knowledge extraction module is used to conduct project process efficiency assessment and project knowledge value extraction based on the project consensus traceability dossier after the project voucher circulation copy has completed the entire circulation of the standardized project circulation path, and to generate a project process debriefing report and organize project knowledge module. The feedback learning and optimization module is used to feed back the project process review report and the organizational project knowledge module to the organizational project resource map and the historical project process case ledger, so as to update the project node handling capability file and optimize the project flow path analysis strategy.

Images