Intelligent management and control method and system for document circulation based on distributed architecture

Abstract

The invention discloses an intelligent document circulation management and control method and system based on a distributed architecture, and relates to the field of data processing. In the method, an official document to be transferred of a first node is obtained, and the first node is any one of a plurality of nodes included in a DAG network topology; according to the document to be transferred, transaction metadata is generated, and the transaction metadata comprises a transaction hash value, a node reputation signature, a semantic tag and a dynamic priority; based on the DAG network topology, determining a transaction dependency relationship between the second node and the first node; according to the transaction dependency relationship, controlling the second node to add a transaction to the transaction metadata to obtain transaction addition data; a lightweight federated optimization algorithm is adopted, modeling is performed on the approval duration feature, the node load feature and the historical performance feature of the second node, and a global circulation strategy is generated through a differential privacy aggregation strategy; and performing official document circulation on the transaction addition data according to the global circulation strategy. By implementing the scheme, the overall document circulation efficiency can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of data processing, and specifically to a method and system for intelligent management and control of document circulation based on a distributed architecture. Background Art

[0002] In traditional document circulation systems, a centralized architecture is usually used for design, that is, all business logic processing, data storage and authority control rely on a central server for unified scheduling and management.

[0003] Although this type of architecture is simple to implement and easy to deploy and maintain, with the rapid growth of business volume, the centralized flow architecture needs to carry a large number of concurrent requests, which can easily lead to computing resource bottlenecks and increased response delays, thereby affecting the overall flow efficiency.

[0004] Therefore, there is an urgent need for an intelligent document circulation management and control method and system based on a distributed architecture. Summary of the Invention

[0005] This application provides a method and system for intelligent document circulation management based on a distributed architecture, which is convenient for improving the overall document circulation efficiency.

[0006] In a first aspect of the present application, a method for intelligent document circulation management based on a distributed architecture is provided, the method comprising: obtaining a document to be circulated from a first node, the first node being any one of a plurality of nodes included in a DAG network topology; generating transaction metadata based on the document to be circulated, the transaction metadata comprising a transaction hash value, a node reputation signature, a semantic label, and a dynamic priority; determining a transaction dependency relationship between a second node and the first node based on the DAG network topology, the second node being a transferable node of the first node in the DAG network topology; controlling the second node to add a transaction to the transaction metadata based on the transaction dependency relationship to obtain transaction addition data; modeling the approval time characteristics, node load characteristics, and historical performance characteristics of the second node using a lightweight federated optimization algorithm, and generating a global circulation strategy through a differential privacy aggregation strategy; and performing document circulation on the transaction addition data according to the global circulation strategy.

[0007] By adopting the above technical solutions and introducing a distributed node coordination mechanism based on the DAG network topology, the performance bottlenecks and disaster recovery limitations of the traditional centralized document circulation architecture have been broken, and the dynamic and intelligent circulation of documents between multiple nodes can be realized. By generating transaction metadata containing semantic tags and dynamic priorities, and combining the transaction dependencies between nodes, the accuracy and adaptability of the circulation path are effectively improved; the introduction of the second node's transaction addition mechanism ensures the continuity and traceability of the operation chain. At the system intelligence level, the combination of lightweight federated optimization algorithms and differential privacy strategies not only achieves dynamic optimization of approval efficiency and load balancing, but also ensures the secure isolation and global coordination of data at each node, thereby overall improving the system's processing capabilities, stability, and intelligent decision-making level in complex business environments. Therefore, it is easy to improve the overall efficiency of document circulation.

[0008] Optionally, obtaining the official document to be circulated in the first node specifically includes: traversing and searching for the circulation label in the first node; obtaining the original official document corresponding to the circulation label; binding the original official document with the acquisition time to obtain the official document status; generating the official document to be circulated based on the original official document and the official document status.

[0009] By adopting the above technical solution, by traversing the first node to search for the transfer tag and obtaining the corresponding original document, the local node's pending transfer information is accurately identified and dynamically extracted, avoiding unnecessary full scans and redundant operations, and improving processing efficiency. Binding the original document to the acquisition time and generating a pending transfer document with an associated status not only enhances the timeliness and contextual awareness of the transaction, but also provides a clear operational starting point and audit basis for subsequent processes, facilitating full-process traceability, status supervision, and compliance control, thereby improving the system's execution transparency and data governance level.

[0010] Optionally, generating transaction metadata based on the official document to be circulated specifically includes: extracting keywords from the title and summary paragraphs of the official document to be circulated, generating a unique ID, a timestamp and the semantic tag of the official document; assembling the unique ID, the semantic tag and the timestamp into a description string, and calling a local hash service to calculate the transaction hash value; signing the transaction hash value and the timestamp with a private key to obtain a verifiable node reputation signature; generating the dynamic priority based on the node load of the first node; and obtaining the transaction metadata based on the transaction hash value, the node reputation signature, the semantic tag and the dynamic priority.

[0011] By adopting the above technical solution, by extracting keywords from circulating documents to generate semantic tags, unique IDs for documents, and timestamps, the semantic recognition capability and timeliness identification of transactions in the system are ensured, and the distinguishability and semantic perception of circulating objects are enhanced from the source. By assembling key information into a description string and calculating the transaction hash value, data tampering and duplicate transactions can be effectively prevented, and the integrity and consistency of circulating transactions can be improved; using the node private key to generate a reputation signature further guarantees the verifiability of the operation source and the credibility of the node identity. At the same time, the transaction priority is dynamically generated in combination with the current load of the node, making the subsequent circulation scheduling more flexible and intelligent, thereby building a transaction metadata foundation with security, credibility, and scheduling sensitivity, which greatly improves the adaptability and operational efficiency of the entire system to complex document circulation scenarios.

[0012] Optionally, determining the transaction dependency between the second node and the first node based on the DAG network topology specifically includes: obtaining the role, queue length, memory usage and acceptable resource type of the second node; matching and obtaining a compliant flow path according to the built-in pluggable rule library in the DAG network topology; if it is determined that the compliant flow path and the role, queue length, memory usage and acceptable resource type all meet preset conditions, establishing a transaction dependency between the second node and the first node.

[0013] By adopting the above technical solution, through comprehensive collection of the role attributes, current queue length, memory usage and acceptable resource types of the second node, dynamic perception and evaluation of the processing capacity and adaptability of the target node are achieved, avoiding the risk of flow failure caused by resource congestion or role mismatch. Combined with the pluggable rule base built into the DAG network topology for compliance path matching, the flow process can flexibly adapt to different business rules and organizational policies, and is highly scalable and configurable. Transaction dependencies are established only when role matching and resource constraints are met, effectively improving the accuracy, efficiency and stability of flows between nodes, thereby providing a reliable decision-making basis and operational guarantee for the intelligent scheduling and steady-state transmission of official documents between multiple nodes.

[0014] Optionally, according to the transaction dependency, controlling the second node to add a transaction to the transaction metadata to obtain the transaction addition data specifically includes: encapsulating the transaction metadata and the transaction dependency into an addition package; pushing the addition package to the second node through a point-to-point encrypted channel; controlling the second node to verify the signature and hash in the addition package, and obtaining the transaction addition data after verification.

[0015] By adopting the above technical solution, this step achieves complete transfer of transaction context by uniformly encapsulating transaction metadata and transaction dependencies into a structured additional package, avoiding information fragmentation and data loss. The use of a point-to-point encrypted channel for transmission effectively ensures the confidentiality and tamper resistance of data during transmission, enhancing the system's communication security level. Controlling the second node to verify the signature and hash in the additional package not only ensures the credibility and integrity of the data source, but also prevents forgery or playback attacks, improving the overall system's risk resistance. This mechanism enhances the consistency, reliability, and security of cross-node transaction processing, laying a solid foundation for the stable and trusted flow of subsequent official documents.

[0016] Optionally, the lightweight federated optimization algorithm is used to model the approval time characteristics, node load characteristics and historical performance characteristics of the second node, and generate a global flow strategy through a differential privacy aggregation strategy, specifically including: obtaining the historical approval time, concurrent task volume and performance score of the second node; determining the approval time characteristics, the node load characteristics and the historical performance characteristics based on the historical approval time, concurrent task volume and performance score; determining the gradient of each edge model through the lightweight federated optimization algorithm, and performing FedAvg weighted averaging to generate a global flow model; according to the global flow model, using differential privacy aggregation to obtain a global flow strategy.

[0017] By adopting the above technical solution, by extracting features from key indicators such as the historical approval time, concurrent task volume, and performance score of the second node, a multi-dimensional characterization of the node's processing capacity and behavioral performance is achieved, providing a high-quality input foundation for the intelligent optimization of the flow path. A lightweight federated optimization algorithm is used to locally train the edge model of each node, and a global flow model is generated through FedAvg weighted aggregation, which not only significantly reduces the load on the central server, but also protects the local privacy of each node's data from being leaked. At the same time, a differential privacy aggregation mechanism is introduced to enhance the privacy of global model parameters, ensuring that the global flow strategy has good generalization ability and dynamic adaptability under the premise of data security and compliance, thereby realizing efficient, secure, and intelligent document flow strategy formulation and execution.

[0018] Optionally, the document circulation of the transaction additional data according to the global circulation strategy specifically includes: determining a third node from the global circulation strategy; and sending the transaction additional data and the document to be circulated to the third node.

[0019] By adopting the above technical solution, the most suitable third node for receiving tasks is precisely determined based on the global flow strategy, achieving intelligent scheduling and dynamic optimal selection of document flow paths, avoiding blind path selection and uneven resource allocation. Sending transaction data along with the documents to be circulated to the third node ensures the complete synchronization of the contextual information and data resources required for task execution, improving the continuity and accuracy of node processing. This mechanism not only improves the system's collaborative efficiency in complex flow networks, but also enhances the goal-orientedness and resource adaptability of the document delivery process, thereby driving the overall system's evolution towards a more efficient, intelligent, and robust operating mode.

[0020] In a second aspect of the present application, an intelligent document circulation management and control system based on a distributed architecture is provided, the system comprising an acquisition module and a processing module, wherein the acquisition module is used to acquire a document to be circulated from a first node, where the first node is any one of multiple nodes included in a DAG network topology; the processing module is used to generate transaction metadata based on the document to be circulated, where the transaction metadata includes a transaction hash value, a node reputation signature, a semantic label, and a dynamic priority; the processing module is further used to determine a transaction dependency relationship between a second node and the first node based on the DAG network topology, where the second node is a transferable node of the first node in the DAG network topology; the processing module is further used to control the second node to add a transaction to the transaction metadata based on the transaction dependency relationship to obtain transaction addition data; the processing module is further used to use a lightweight federated optimization algorithm to model the approval time characteristics, node load characteristics, and historical performance characteristics of the second node, and generate a global circulation strategy through a differential privacy aggregation strategy; the processing module is further used to circulate the document for the transaction addition data in accordance with the global circulation strategy.

[0021] In a third aspect of the present application, an electronic device is provided, which includes a processor, a memory, a user interface, and a network interface. The memory is used to store instructions, the user interface and the network interface are both used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device performs the method described above.

[0022] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores instructions. When the instructions are executed, the method described above is executed.

[0023] In summary, one or more technical solutions provided in this application have at least the following technical effects or advantages: By introducing a distributed node collaboration mechanism based on the DAG network topology, the performance bottlenecks and disaster recovery limitations of the traditional centralized document circulation architecture have been broken, and dynamic and intelligent document circulation between multiple nodes can be achieved. By generating transaction metadata containing semantic tags and dynamic priorities, and combining the transaction dependencies between nodes, the accuracy and adaptability of the circulation path are effectively improved; the introduction of a second node transaction addition mechanism ensures the continuity and traceability of the operation chain. At the system intelligence level, the combination of lightweight federated optimization algorithms and differential privacy strategies not only achieves dynamic optimization of approval efficiency and load balancing, but also ensures the secure isolation and global collaboration of data at each node, thereby overall improving the system's processing capabilities, stability, and intelligent decision-making level in complex business environments. Therefore, it is easy to improve the overall efficiency of document circulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A flowchart of an intelligent document circulation management method based on a distributed architecture provided in an embodiment of the present application; Figure 2 Another flowchart of a method for intelligent document circulation management and control based on a distributed architecture provided in an embodiment of the present application; Figure 3 A schematic diagram of a module of an intelligent document circulation management and control system based on a distributed architecture provided in an embodiment of the present application; Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0025] Explanation of the reference numerals: 31, acquisition module; 32, processing module; 41, processor; 42, communication bus; 43, user interface; 44, network interface; 45, memory. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0027] In the description of the embodiments of this application, words such as "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "for example" or "for instance" in the embodiments of this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "for example" or "for instance" is intended to present the relevant concepts in a concrete manner.

[0028] In the description of the embodiments of the present application, the term "multiple" means two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0029] Traditional document circulation systems generally adopt a centralized architecture design, where business logic processing, data storage, and permission control are uniformly deployed on a central server for centralized scheduling and management. Although this architecture is relatively simple to deploy and maintain initially, as business scale continues to expand and the number of concurrent requests surges, central nodes face enormous computing and scheduling pressures, which can easily lead to resource bottlenecks, slowing system response speeds and significantly limiting overall circulation efficiency.

[0030] In order to solve the above technical problems, this application provides an intelligent management and control method for document circulation based on a distributed architecture. Figure 1 , Figure 1 This is a flow chart of a method for intelligent document circulation management based on a distributed architecture provided in an embodiment of the present application. The method is applied to a server and includes steps S110 to S160, which are as follows: S110: Obtain an official document to be transferred of a first node, where the first node is any one of a plurality of nodes included in a DAG network topology.

[0031] Specifically, the system doesn't have a single central node to manage all document flows. Instead, the entire network is viewed as a directed acyclic graph (DAG) structure composed of multiple peer nodes. When a server needs to obtain a document to be circulated, it can extract the document data for which that node is responsible and which has not yet been circulated from any node in the DAG network (the "first node"). In this way, each node locally holds copies of the documents and pending lists within its scope of responsibility. The server can flexibly centrally or decentralizedly obtain tasks from different nodes, avoiding the risk of excessive pressure or failure at a single point in the traditional centralized architecture, which can cause the entire flow to stagnate.

[0032] For example, suppose that in a cross-provincial and municipal government document collaboration system, there are four office locations, nodes A, B, C, and D, each with an edge instance deployed, forming a DAG topology between them. One day, node C received a budget approval document issued by the Municipal Finance Bureau and marked it as "pending circulation." At this time, if the back-end server wants to continue processing this document, it only needs to initiate a request to node C to obtain the pending document stored locally, without having to return to the central database or other nodes for centralized retrieval. This design not only improves the acquisition speed, but also enhances the fault tolerance of the system. Even if some nodes or central services fail, the server can still obtain new pending documents from other healthy nodes to ensure the continuity of the overall business flow.

[0033] In one possible implementation, obtaining the official document to be circulated of the first node specifically includes: traversing and searching for the circulation label in the first node; obtaining the original official document corresponding to the circulation label; binding the original official document with the acquisition time to obtain the official document status; and generating the official document to be circulated based on the original official document and the official document status.

[0034] Specifically, the first node will add a special "circulation tag" to each official document that needs to be circulated in its own local storage to indicate that the file is in a pending state. When the system performs the acquisition operation, it first traverses the tag list of the node and retrieves all official document entries with "pending circulation" or similar tags one by one; then, it locates the corresponding original official document file based on the tag index, such as a document titled "Annual Audit Report", and reads the file from the storage system into the memory. Next, the system will record the time point of this reading, and logically bind the timestamp to the document itself to form an information unit with status attributes, that is, the status of the document that has been "acquired" and entered the "pending circulation" link at that moment.

[0035] For example, let's assume the first node is the Municipal Finance Bureau's edge server. Within its file storage directory, several documents are labeled "draft_pending." The system initiates a search at 10:15 AM and finds the document "2025 Budget Adjustment Plan.docx" corresponding to the tag and loads it. Simultaneously, the system appends the time "2025-05-13 10:15:00" to the document's metadata, indicating that the document has just entered the "pending approval" phase of the document circulation process. Finally, based on the document's content and this status tag, the system generates a complete "pending document" object for subsequent approval routing and transaction processing steps.

[0036] S120. Generate transaction metadata based on the official document to be circulated. The transaction metadata includes a transaction hash value, a node reputation signature, a semantic tag, and a dynamic priority.

[0037] Specifically, the server takes the previously acquired "official document to be transferred" as input and, through multi-dimensional information extraction and secure computation, constructs a complete "transaction metadata" record. First, the system performs semantic analysis on the document's content, extracting key terms from the title and core sentences from the summary paragraph to generate a series of "semantic tags." These tags help the subsequent routing engine quickly identify the document's subject and processing method. Next, the system concatenates the document's unique identifier (such as an internal number or path), the current timestamp, and the aforementioned semantic tags into a descriptive string. This string is then computed using a built-in hashing service to generate an irreversible and verifiable "transaction hash." To ensure the authenticity of the node initiating the transaction, the server digitally signs the hash and timestamp using the node's private key, creating a "node reputation signature" to prevent malicious tampering and request forgery. Finally, the system dynamically assigns a priority tag based on the first node's real-time load and the document's urgency, packaging the priority along with other metadata. This comprehensive set of metadata ensures security and non-repudiation while providing essential context for subsequent scheduling and auditing.

[0038] For example, suppose an edge node at a municipal government office receives a document titled "2025 Budget Adjustment Application." After semantic analysis, the system extracts tags such as "budget adjustment" and "annual finance." It then combines the document number "DOC-20250513-001" and the current time "2025-05-13 14:30:00" with these tags to form a descriptive string, generating a transaction hash "0xB7E3F9..." This hash and timestamp are then signed with the node's private key to produce a verifiable signature string. Finally, based on the node's current workload of 75% and the impending submission deadline for the budget adjustment, the transaction priority is set to "high." This completes the generation of transaction metadata, including the transaction hash, node reputation signature, semantic tags, and dynamic priority, laying a solid foundation for the document's subsequent circulation and security audits.

[0039] In one possible implementation, transaction metadata is generated based on the official document to be circulated, specifically including: extracting keywords from the title and summary paragraphs of the official document to be circulated, and generating a unique ID, timestamp, and semantic tag for the official document; assembling the unique ID, semantic tag, and timestamp of the official document into a description string, and calling a local hash service to calculate a transaction hash value; signing the transaction hash value and timestamp with a private key to obtain a verifiable node reputation signature; generating a dynamic priority based on the node load of the first node; and obtaining transaction metadata based on the transaction hash value, node reputation signature, semantic tag, and dynamic priority.

[0040] Specifically, the system will call the embedded natural language processing module to perform keyword extraction on the title and summary paragraphs of the official document, extract core terms such as "annual budget" and "cross-departmental coordination", and classify these terms into a semantic tag list in the form of tags. At the same time, in order to uniquely identify the document across the entire network, the system will generate a unique ID for the official document based on the official document's own internal number or storage path, and append the current system timestamp on this basis to record the exact moment of generation. In this way, we obtain three basic data: the unique ID of the official document, the set of semantic tags, and the generation time, which provide structured input for constructing the subsequent transaction description string.

[0041] The system then assembles these three pieces of data into a sequential description string and passes them to a local cryptographic hash service for a one-way hash operation, outputting a fixed-length, unalterable "transaction hash." This hash not only prevents subsequent modification of the description string but also allows for rapid comparison and deduplication. The system then digitally signs the transaction hash and timestamp using the node's private key, generating a "node reputation signature" that verifies that the transaction was initiated by that node. Any subsequent modifications will cause the signature verification to fail. To further enhance the real-time and fairness of scheduling decisions, the system dynamically assesses the current processing load of the first node and assigns a "dynamic priority" to the transaction based on the document's importance or urgency. Ultimately, these elements, including the transaction hash, node reputation signature, semantic tag, and dynamic priority, are integrated into a complete set of transaction metadata, ensuring data security and auditability while also providing a rich basis for subsequent intelligent routing and priority scheduling.

[0042] S130: Determine a transaction dependency relationship between a second node and a first node based on a DAG network topology, where the second node is a transferable node of the first node in the DAG network topology.

[0043] Specifically, the server first selects neighboring nodes that meet business rules, role responsibilities, and current load conditions based on the first node's position in the DAG and its connectivity with neighboring nodes, as potential transfer targets. The server then checks the previous transaction links recorded in the DAG to determine whether these candidate nodes are in a receivable state. Only when node roles, resource availability, and topological dependencies are all met will the server formally establish the "first node - second node" transfer dependency and mark the new transaction path on the DAG.

[0044] For example, in a multi-department collaborative approval network, node A has completed processing a "Contract Approval" document. It is connected to multiple downstream nodes in the DAG (such as node B: Legal Department, node C: Finance Department). The server first identifies node B in the Legal Department as the next candidate for transfer according to the approval rules. It then further checks whether node B's current task queue length and role match. If node B is maintaining a low load and has "Contract Review" permissions, the server creates a new dependency edge from A to B in the DAG, indicating that node B should continue to review this document. This ensures that the transfer path is both business-compliant and takes into account real-time resource availability, achieving efficient and reliable distributed document flow.

[0045] In one possible implementation, based on the DAG network topology, the transaction dependency relationship between the second node and the first node is determined, specifically including: obtaining the role, queue length, memory usage, and acceptable resource type of the second node; matching and obtaining a compliant flow path according to the built-in pluggable rule library in the DAG network topology; if it is determined that the compliant flow path and the role, queue length, memory usage, and acceptable resource type all meet the preset conditions, then establishing a transaction dependency relationship between the second node and the first node.

[0046] Specifically, the system screens and evaluates candidate "second nodes" through multi-dimensional indicators to ensure that the subsequent flow of official documents complies with business rules and can efficiently utilize system resources. First, the system dynamically collects four types of key information for each potential receiving node: the node's "role" positioning in the organizational structure or approval chain (such as legal, financial, or comprehensive management); the current "queue length", that is, the number of to-do items, to measure the processing pressure of the node; runtime resource usage such as "memory usage" to determine whether the node still has the ability to accept new tasks; and "acceptable resource types", such as whether the node only processes specific types of files or is only responsible for certain business areas. Through these real-time indicators, the system can have a comprehensive and accurate understanding of the health status and business adaptability of each candidate node.

[0047] Next, the system uses a predefined pluggable rule base to verify the compliance of the flow paths within the DAG network. For example, documents for "contract approval" must first flow to the legal node and then to the finance node; documents for "budget adjustment" may also require additional audit or management approval. Only nodes that meet the business process requirements in the rule base and meet preset thresholds for role matching, queue length, and resource conditions are officially selected as "second nodes." Once all these conditions are met, the system adds a transaction dependency edge from the "first node" to the "second node" in the DAG structure, clearly indicating that the document should continue to be processed by this node.

[0048] For example, suppose node A has just completed an initial review of an "Annual Financial Report" document. The system first retrieves the roles and performance indicators of nodes B (Legal Department), C (Audit Department), and D (Finance Department) from its neighbor list. It discovers that node B's current queue length is only 10 entries, its memory utilization is 40%, and its "Legal" role, which is required for contract review, is valid. A rule base verification confirms that the "Financial Report" requires legal review first. Once all conditions are met, the system marks the A→B flow dependency in the DAG, ensuring that the document is seamlessly and compliantly transferred to node B for further processing.

[0049] S140 . Control the second node to add a transaction to the transaction metadata according to the transaction dependency relationship to obtain transaction added data.

[0050] Specifically, first, the server will package the previously generated transaction metadata (including hash values, signatures, semantic tags, priorities, etc.) together with the dependencies into an "added transaction package" to ensure that all necessary context is transmitted along with the package body. Next, the server sends this package to the second node through a point-to-point encrypted channel. After receiving it, the second node will verify the legitimacy of the signature and hash to confirm that the transaction was indeed initiated by a trusted node and that the data has not been tampered with. Once the verification is passed, the second node will append a new transaction node to its local DAG structure and save the entire package content as "transaction added data", thereby continuing the complete flow link.

[0051] For example, suppose that on a distributed official document platform, Node A has packaged "Annual Budget Adjustment.docx" and generated transaction metadata, and has determined that the next node is Node B in the legal department. The server then encapsulates all metadata information, along with the "A→B" dependency, into an additional transaction package and sends it to Node B via an encrypted channel. Upon receipt, Node B first verifies the signature and hash. If the verification is successful, a new transaction node is created in the local DAG graph, the document status is updated to "pending legal review," and the entire package is persisted as "transaction addition data." This not only ensures the continuity and integrity of the circulation chain, but also provides accurate operational traces for subsequent audits and tracing.

[0052] In one possible implementation, based on the transaction dependency, the second node is controlled to add the transaction to the transaction metadata to obtain the transaction addition data, specifically including: encapsulating the transaction metadata and the transaction dependency into an addition package; pushing the addition package to the second node through a point-to-point encrypted channel; controlling the second node to verify the signature and hash in the addition package, and obtaining the transaction addition data after the verification is passed.

[0053] Specifically, the server first organically integrates the previously constructed transaction metadata (including hash values, reputation signatures, semantic tags, and priorities) with the established flow dependency (i.e., "first node → second node") to form a complete "added transaction package." This package not only carries all the security verification information for the official document but also clearly identifies the upstream and downstream nodes in the flow chain, ensuring that any subsequent party can clearly understand the source and destination of the transaction. After packaging, the server securely pushes the entire package to the second node via a point-to-point encrypted channel designed specifically for inter-node communication (such as a TLS-based two-way authenticated link), thereby preventing any risk of data leakage or tampering in the middle.

[0054] After receiving the "added transaction package", Node B will immediately verify the digital signature and hash summary in the package: it will use the public key retained by the first node to verify whether the signature is consistent with the hash and timestamp, and recalculate the data description string corresponding to the hash value to confirm that the two are completely matched. Only after these two steps are passed, Node B will consider the contents of this package to be authentic and write it as a new local transaction to its own DAG storage structure. This is the so-called "transaction addition data". Through this strict encapsulation, encrypted transmission and multiple verification mechanisms, the server not only ensures the complete continuation of the circulation link, but also makes any malicious tampering or replay attacks impossible to escape, providing a solid security barrier for the entire distributed document circulation process.

[0055] S150 uses a lightweight federated optimization algorithm to model the approval time characteristics, node load characteristics, and historical performance characteristics of the second node, and generates a global flow strategy through a differential privacy aggregation strategy.

[0056] Specifically, the server does not aggregate the raw approval data from each node for training. Instead, it uses a lightweight federated optimization algorithm to allow each edge node to independently train its own model locally. Model inputs include key metrics such as the node's past approval time distribution, the current task queue length (node load), and performance scores from previous approvals. This not only fully utilizes each node's local computing resources, reducing the computational pressure on the central server, but also avoids directly uploading sensitive approval records or performance data to the central server, reducing the risk of privacy leaks. Each node then uploads the model parameters (rather than the raw data) locally to the aggregation center. The aggregation center then weights and aggregates these parameters based on strategies such as the federated average (FedAvg) to derive a global flow model that represents the overall network behavior pattern.

[0057] After the model aggregation is completed, the aggregation center will also apply a differential privacy aggregation strategy, such as adding a certain amount of random noise when averaging parameters, so as to ensure that even if someone intercepts the uploaded model update, it is impossible to reverse the original data of any single node. When the global model is generated, the server can output a set of "flow strategies" applicable to the entire network based on the model. The strategy will define the priority, timeout threshold and load balancing scheme of different types of documents among each node. For example, if after a certain federated training, the global model finds that the average approval time of the financial report node during peak hours is significantly higher than that of the legal node, and its queue length is often in a "saturated" state, the generated global strategy may recommend that more low-priority documents be allocated to idle legal nodes or other auxiliary nodes in the future to balance the overall flow speed and improve network throughput.

[0058] In one possible implementation, a lightweight federated optimization algorithm is used to model the approval time characteristics, node load characteristics, and historical performance characteristics of the second node, and a global flow strategy is generated through a differential privacy aggregation strategy, specifically including: obtaining the historical approval time, concurrent task volume, and performance score of the second node; determining the approval time characteristics, node load characteristics, and historical performance characteristics based on the historical approval time, concurrent task volume, and performance score; determining the gradients of each edge model through a lightweight federated optimization algorithm, and performing FedAvg weighted averaging to generate a global flow model; according to the global flow model, using differential privacy aggregation to obtain the global flow strategy.

[0059] Specifically, the server first collects three types of key indicators locally from the second node: historical approval time, concurrent task volume, and performance score. The historical approval time reflects the efficiency of the node in processing different types of documents; the concurrent task volume reveals the node's carrying capacity under high load conditions; and the performance score integrates the node's accuracy, compliance rate, and other assessment results over the past several cycles. The server converts these raw data into features that are more usable for modeling, such as the average approval time distribution, peak concurrent processing capacity, and weighted performance index, which correspond to approval time features, node load features, and historical performance features, respectively. Through such feature extraction, it can be ensured that when the federated algorithm is trained locally on each node, it not only covers the full picture of the node's operation, but also avoids directly exposing sensitive approval records.

[0060] Subsequently, the server uses a lightweight federated optimization algorithm to iteratively update the local model of each node. The node trains the model gradients locally based on the above features, and then transmits these gradients rather than the original data to the aggregation center; the central node uses the FedAvg (federal average) strategy to perform a weighted average of all gradients to obtain a global flow model that represents the overall network scheduling rules. To further protect privacy, the center introduces a differential privacy aggregation strategy during the merging process. By adding noise before and after parameter aggregation, it prevents third parties or malicious nodes from inferring the specific characteristics of any single node from the updated data. After completing the model and noise processing, the server can generate a set of global flow strategies that take into account both efficiency and load balancing. For example, after discovering that a node often causes delays when processing financial reports during peak hours, the strategy will automatically recommend temporarily allocating some low-priority documents to more idle approval nodes, thereby improving the overall approval throughput and smoothing out peak and valley fluctuations.

[0061] S160. According to the global circulation strategy, the document circulation is performed for the transaction added data.

[0062] Specifically, the server will combine the global flow strategy generated and issued in advance with the additional data of the current pending transactions to determine the next destination and processing method of each document in the distributed network. Specifically, the server will first read the various parameters regarding priority scheduling, node load balancing, and timeout reminders in the global strategy, and then dynamically select and sort the flow paths based on the document type, urgency, and current node information contained in the transaction additional data. In this way, whether it is a high-priority urgent document or an ordinary daily notification, it can be reasonably assigned according to the globally unified optimization goals, thereby ensuring the timely advancement of the core process and avoiding excessive congestion or idleness of certain nodes during peak hours.

[0063] For example, suppose the global circulation strategy is clarified after the latest federated model update: the financial review node is prone to overload during the afternoon peak period, and the priority of contract approval documents should be raised to "highest"; the legal review node is more idle in the morning period and is suitable for temporarily sharing some work. At this time, when the server receives the additional data of the "contract approval" transaction forwarded from node B, it will immediately mark the transaction as "high priority" according to the strategy and push it to the legal review node first, so that it can use its idle time window for preliminary compliance inspection; after the legal review is completed, the document will be routed to the financial node for subsequent processing. In this way, the server not only achieves load balancing and timeliness at the global level, but also enhances the reliability and flexibility of the entire distributed document circulation system.

[0064] In one possible implementation, refer to Figure 2 , Figure 2Another flow chart of an intelligent document circulation control method based on a distributed architecture provided in an embodiment of the present application, wherein the transaction-added data is circulated for document circulation according to the global circulation strategy, specifically including steps S210 to S220. The above steps are as follows: S210, determining a third node from the global circulation strategy; S220, sending the transaction-added data and the document to be circulated to the third node.

[0065] Specifically, the server first reads the currently effective global routing policy, which contains the priority sorting and selection rules for each node based on different document types and load conditions. By matching the rules in the policy, the system automatically determines the most appropriate third node, taking into account the document's attributes (such as document type and urgency), network topology, and the real-time load status of each node. This step ensures dynamic optimization of the routing path, allowing each document to be assigned to the node with the greatest processing power and best suited to business requirements, thereby improving overall approval efficiency and reducing resource waste between nodes.

[0066] After determining the third node, the server will send the previously encapsulated transaction addition data (including hash check value, reputation signature, priority and other metadata) to the node together with the original or necessary official document content. This push-to-all method ensures that the third node has complete contextual information when receiving it: it can not only verify the legitimacy and source of the transaction, but also immediately obtain the document itself for subsequent operations. For example, if the global policy determines that the "contract approval" link should be taken over by the legal review node C, the server will push the transaction addition package containing the contract summary, automatically extracted semantic tags and dynamic priority identifiers together with the contract document to node C. After receiving it, node C can directly enter the review process without the need to pull or verify other information, thus achieving seamless connection and efficient execution of official document flow.

[0067] This application also provides an intelligent document circulation management and control system based on a distributed architecture, referring to Figure 3 , Figure 3A module diagram of an intelligent document circulation control system based on a distributed architecture provided in an embodiment of the present application. The system is a server, and the server includes an acquisition module 31 and a processing module 32, wherein the acquisition module 31 acquires the document to be circulated from the first node, where the first node is any one of the multiple nodes included in the DAG network topology; the processing module 32 generates transaction metadata based on the document to be circulated, and the transaction metadata includes a transaction hash value, a node reputation signature, a semantic label, and a dynamic priority; the processing module 32 determines the transaction dependency relationship between the second node and the first node based on the DAG network topology, where the second node is a transferable node of the first node in the DAG network topology; the processing module 32 controls the second node to add a transaction to the transaction metadata based on the transaction dependency relationship to obtain transaction addition data; the processing module 32 uses a lightweight federated optimization algorithm to model the approval time characteristics, node load characteristics, and historical performance characteristics of the second node, and generates a global circulation strategy through a differential privacy aggregation strategy; the processing module 32 circulates the document for the transaction addition data according to the global circulation strategy.

[0068] In a possible implementation, the acquisition module 31 acquires the official document to be circulated of the first node, specifically including: the processing module 32 traverses and searches for the circulation label in the first node; the acquisition module 31 acquires the original official document corresponding to the circulation label; the processing module 32 binds the original official document with the acquisition time to obtain the official document status; the processing module 32 generates the official document to be circulated based on the original official document and the official document status.

[0069] In one possible implementation, the processing module 32 generates transaction metadata based on the official document to be circulated, specifically including: the processing module 32 extracts keywords from the title and summary paragraphs of the official document to be circulated, and generates a unique ID, timestamp and semantic tag for the official document; the processing module 32 assembles the unique ID, semantic tag and timestamp of the official document into a description string, and calls the local hash service to calculate the transaction hash value; the processing module 32 uses a private key to sign the transaction hash value and timestamp to obtain a verifiable node reputation signature; the processing module 32 generates a dynamic priority based on the node load of the first node; the processing module 32 obtains transaction metadata based on the transaction hash value, node reputation signature, semantic tag and dynamic priority.

[0070] In one possible implementation, the processing module 32 determines the transaction dependency between the second node and the first node based on the DAG network topology, specifically including: the acquisition module 31 obtains the role, queue length, memory usage and acceptable resource type of the second node; the processing module 32 matches and obtains a compliant flow path based on the built-in pluggable rule library in the DAG network topology; if the processing module 32 determines that the compliant flow path and the role, queue length, memory usage and acceptable resource type all meet the preset conditions, then a transaction dependency relationship is established between the second node and the first node.

[0071] In one possible implementation, the processing module 32 controls the second node to add a transaction to the transaction metadata based on the transaction dependency to obtain the transaction addition data, specifically including: the processing module 32 encapsulates the transaction metadata and the transaction dependency into an addition package; the processing module 32 pushes the addition package to the second node through a point-to-point encrypted channel; the processing module 32 controls the second node to verify the signature and hash in the addition package, and obtains the transaction addition data after the verification is passed.

[0072] In one possible implementation, the processing module 32 uses a lightweight federated optimization algorithm to model the approval time characteristics, node load characteristics, and historical performance characteristics of the second node, and generates a global flow strategy through a differential privacy aggregation strategy, specifically including: the processing module 32 obtains the historical approval time, concurrent task volume, and performance score of the second node; the processing module 32 determines the approval time characteristics, node load characteristics, and historical performance characteristics based on the historical approval time, concurrent task volume, and performance score; the processing module 32 determines the gradient of each edge model through a lightweight federated optimization algorithm, and performs FedAvg weighted average to generate a global flow model; the processing module 32 uses differential privacy aggregation according to the global flow model to obtain a global flow strategy.

[0073] In a possible implementation, the processing module 32 performs document circulation on the transaction addition data according to the global circulation strategy, specifically including: the processing module 32 determines the third node from the global circulation strategy; the processing module 32 sends the transaction addition data and the document to be circulated to the third node.

[0074] It should be noted that the above embodiments provide devices that implement their functions using only the division of the above functional modules as examples. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiments are based on the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0075] This application also provides an electronic device, referring to Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include: at least one processor 41, at least one network interface 44, a user interface 43, a memory 45, and at least one communication bus 42.

[0076] The communication bus 42 is used to realize the connection and communication between these components.

[0077] The user interface 43 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 43 may also include a standard wired interface and a wireless interface.

[0078] The network interface 44 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0079] The processor 41 may include one or more processing cores. Using various interfaces and circuits, the processor 41 connects to various components within the server. It executes instructions, programs, code sets, or instruction sets stored in the memory 45, as well as accesses data stored in the memory 45, to perform various server functions and process data. Optionally, the processor 41 may be implemented using at least one of the following hardware forms: a digital signal processing (DSP), a field-programmable gate array (FPGA), or a programmable logic array (PLA). The processor 41 may integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. The CPU primarily processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing content displayed on the display screen; and the modem handles wireless communications. It is understood that the modem may also be implemented as a separate chip, rather than integrated into the processor 41.

[0080] Among them, the memory 45 may include a random access memory (RAM) or a read-only memory (Read-Only Memory). Optionally, the memory 45 includes a non-transitory computer-readable storage medium. The memory 45 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 45 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 45 may also be optionally at least one storage device located away from the aforementioned processor 41. As Figure 4 As shown, the memory 45 as a computer storage medium may include an operating system, a network communication module, a user interface module, and an application program for an intelligent document circulation management method based on a distributed architecture.

[0081] exist Figure 4 In the electronic device shown, the user interface 43 is mainly used to provide an input interface for the user and obtain data input by the user; and the processor 41 can be used to call an application stored in the memory 45 for an intelligent document circulation management method based on a distributed architecture. When executed by one or more processors, the electronic device executes one or more methods in the above embodiments.

[0082] It should be noted that for the aforementioned method embodiments, for simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for this application.

[0083] The present application also provides a computer-readable storage medium storing instructions, which, when executed by one or more processors, enable an electronic device to execute one or more of the methods described in the above embodiments.

[0084] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0085] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are merely schematic, such as the division of units, which is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interface, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0086] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0087] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0088] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of this application, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of this application. The aforementioned memory includes various media that can store program code, such as USB flash drives, mobile hard drives, magnetic disks, or optical disks.

[0089] The above is only an exemplary embodiment of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and the truth of practice, those skilled in the art will easily think of other embodiments of the present disclosure. This application is intended to cover any variation, use or adaptive change of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art that are not recorded in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A method for intelligent document circulation management based on a distributed architecture, characterized in that: The method comprises: Obtaining a document to be transferred from a first node, where the first node is any one of a plurality of nodes included in a DAG network topology; Generate transaction metadata based on the document to be circulated, where the transaction metadata includes a transaction hash value, a node reputation signature, a semantic tag, and a dynamic priority; Determining, based on the DAG network topology, a transaction dependency relationship between a second node and the first node, where the second node is a transferable node of the first node in the DAG network topology; controlling the second node to add a transaction to the transaction metadata according to the transaction dependency to obtain transaction added data; A lightweight federated optimization algorithm is used to model the approval time characteristics, node load characteristics, and historical performance characteristics of the second node, and a global flow strategy is generated through a differential privacy aggregation strategy. According to the global circulation strategy, the transaction added data is circulated as an official document.

2. The method for intelligent document circulation management based on distributed architecture according to claim 1 is characterized in that: The obtaining of the official document to be transferred at the first node specifically includes: In the first node, traverse and search for the flow label; Obtain the original document corresponding to the circulation tag; Binding the original document with the acquisition time to obtain the document status; The document to be circulated is generated according to the original document and the document status.

3. The method for intelligent management and control of official document circulation based on a distributed architecture according to claim 1 is characterized in that: Generating transaction metadata according to the document to be circulated specifically includes: Extract keywords from the title and abstract of the document to be circulated, and generate a unique ID, a timestamp, and the semantic tag of the document; Assemble the document's unique ID, the semantic tag, and the timestamp into a description string, and call a local hash service to calculate the transaction hash value; Signing the transaction hash value and the timestamp with a private key to obtain a verifiable node reputation signature; generating the dynamic priority according to the node load of the first node; The transaction metadata is obtained based on the transaction hash value, the node reputation signature, the semantic label, and the dynamic priority.

4. The method for intelligent document circulation management based on distributed architecture according to claim 1 is characterized in that: The determining, based on the DAG network topology, a transaction dependency relationship between the second node and the first node specifically includes: Obtain the role, queue length, memory usage, and acceptable resource type of the second node; According to the built-in pluggable rule base in the DAG network topology, a compliant flow path is matched; If it is determined that the compliant flow path and the role, queue length, memory usage, and acceptable resource type all meet preset conditions, a transaction dependency relationship is established between the second node and the first node.

5. The method for intelligent management and control of official document circulation based on distributed architecture according to claim 1 is characterized in that: The controlling the second node to add a transaction to the transaction metadata according to the transaction dependency to obtain transaction added data specifically includes: Encapsulating the transaction metadata and the transaction dependency into an additional package; Pushing the added package to the second node through a point-to-point encrypted channel; The second node is controlled to verify the signature and hash in the added package, and the transaction added data is obtained after the verification passes.

6. The method for intelligent management and control of official document circulation based on a distributed architecture according to claim 1 is characterized in that: The lightweight federated optimization algorithm is used to model the approval time characteristics, node load characteristics, and historical performance characteristics of the second node, and a global flow strategy is generated through a differential privacy aggregation strategy, specifically including: Obtain the historical approval time, concurrent task volume, and performance score of the second node; Determining the approval time characteristics, the node load characteristics, and the historical performance characteristics based on the historical approval time, the concurrent task volume, and the performance score; The lightweight federated optimization algorithm is used to determine the gradient of each edge model and perform FedAvg weighted average to generate a global flow model. According to the global flow model, differential privacy aggregation is used to obtain the global flow strategy.

7. The method for intelligent document circulation management based on distributed architecture according to claim 1 is characterized in that: The document circulation of the transaction added data according to the global circulation strategy specifically includes: Determining a third node from the global flow strategy; The transaction additional data and the document to be circulated are sent to the third node.

8. An intelligent document circulation management and control system based on a distributed architecture, characterized by: The system comprises an acquisition module (31) and a processing module (32), wherein: The acquisition module (31) is used to acquire the document to be circulated of the first node, where the first node is any one of the multiple nodes included in the DAG network topology; The processing module (32) is used to generate transaction metadata based on the official document to be circulated, and the transaction metadata includes a transaction hash value, a node reputation signature, a semantic tag, and a dynamic priority; The processing module (32) is further configured to determine a transaction dependency relationship between a second node and the first node based on the DAG network topology, wherein the second node is a transferable node of the first node in the DAG network topology; The processing module (32) is further configured to control the second node to add a transaction to the transaction metadata according to the transaction dependency relationship, thereby obtaining transaction addition data; The processing module (32) is further configured to use a lightweight federated optimization algorithm to model the approval time characteristics, node load characteristics, and historical performance characteristics of the second node, and generate a global flow strategy through a differential privacy aggregation strategy; The processing module (32) is further used to perform document circulation on the transaction added data according to the global circulation strategy.

9. An electronic device, characterized in that: The electronic device comprises a processor (41), a memory (45), a user interface (43) and a network interface (44), wherein the memory (45) is used to store instructions, the user interface (43) and the network interface (44) are both used to communicate with other devices, and the processor (41) is used to execute the instructions stored in the memory (45) so that the electronic device executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores instructions, and when the instructions are executed, the method according to any one of claims 1 to 7 is performed.

Images