Blockchain-based university organizational structure digital twin system and construction method

Abstract

The application discloses a university organizational structure digital twin system based on a blockchain and a construction method, and comprises a data acquisition layer, a data resource layer, a digital twin layer, an application service layer and a blockchain base layer; a secondary organization digital twin and a multi-dimensional relationship network thereof are constructed; an organization adjustment digital twin path comprising an adjustment proposal, a twin pre-play, influence evaluation, scheme optimization, on-chain evidence, physical execution and effect monitoring is constructed; a personnel tag forest and a tag-privilege dynamic mapping mechanism are established; a document intelligent analysis and adjustment model based on an AI large model is constructed, and an 'AI assistance + manual confirmation' mechanism is followed; a consortium chain network and six types of core intelligent contracts are used to realize full-process credible evidence storage and automatic execution; and the application solves technical problems such as a university organizational structure data island, high adjustment decision risk, difficulty in responsibility tracing and low cross-departmental collaboration efficiency, and realizes computable, predictable, traceable and credible university organizational structure management.

Description

Technical Field

[0001] This invention relates to the intersection of digital twin, blockchain and artificial intelligence technologies, and in particular to a blockchain-based digital twin system for university organizational structure and its construction method, which can be applied to the modernization and informatization of governance in higher education institutions. Background Technology

[0002] The organizational structure of universities is the foundation of their governance system. However, current university organizational structure management primarily relies on traditional human resource management systems and office automation systems. These systems reveal numerous technical bottlenecks when dealing with complex relationships such as the coexistence of internal organizational structures, joint offices, and affiliations, as well as diverse personnel identities and widespread cross-campus management. This is because the internal organizational structure, job positions, personnel appointments, access control, and business collaboration all depend on the accurate representation and dynamic maintenance of the organizational structure.2 However, in current university IT infrastructure development, the organizational structure is often simplified to a static "department tree + personnel affixation" structure, mainly used for basic address book display, organizational hierarchy maintenance, and simple permission configuration. While this approach may have met early IT needs, it is ill-suited to the current state of university IT systems. Increasingly complex real-world scenarios in governance; for example: secondary organizations and secondary units correspond to each other but are not entirely equivalent; subordinate, joint, and affiliated complex organizational relationships coexist; multi-campus extended management and local management coexist; positions and personnel are separated, and personnel may hold multiple positions; job titles, professional titles, and ranks coexist, resulting in complex identity attributes; organizational adjustments, position adjustments, and personnel appointments and removals require cross-system linkage. Under these circumstances, traditional static organizational models are no longer able to support the needs of modern, refined management, and intelligent application of university governance; and the current organizational structure management of universities mainly relies on traditional human resource management systems (HRMS) and office automation systems (OA). These systems have undergone a continuous evolution from stand-alone applications to network collaboration, from decentralized construction to integrated consolidation, and from local optimization to global governance.

[0003] In recent years, digital twin technology has expanded from industrial manufacturing to educational management, with some universities beginning to explore the digital mapping of physical spaces such as campus facilities and laboratories. At the data management level, universities have generally established data platforms to integrate multi-source data from academic affairs, personnel, and research, enabling the digital definition and related queries of "people, places, events, and things." As artificial intelligence technology is applied to document processing and decision support, it has also improved the level of management automation. Despite these advancements, the following core technical challenges remain in the field of university organizational structure management: (1) Data silos and version fragmentation: Organizational structure data is scattered across multiple independent systems, lacking unified standards and real-time synchronization, resulting in "one organization, multiple versions"; universities are large in scale and have many secondary units, and each unit's information system "fights its own battle", with organizational structure data scattered across multiple business systems, lacking unified data standards and real-time synchronization mechanisms; when secondary units are adjusted, the updates of each system are not synchronized, resulting in a chaotic situation of "one organization, multiple versions".

[0004] (2) Lack of digital twin simulation capabilities: Major adjustments such as the establishment, merger, and abolition of institutions lack digital simulation and deduction methods, and decision-making relies on experience, resulting in high trial and error costs; existing systems are mostly post-event recording types and lack pre-event simulation and deduction capabilities; major adjustments such as the establishment, merger, and abolition of secondary institutions cannot be simulated digitally to predict their chain effects on teaching resource allocation, staffing, financial budgets, etc., resulting in high decision-making risks and high trial and error costs.

[0005] (3) Lack of visualization and quantitative analysis of relationships: It is difficult to model the complex relationship network between positions and institutions, and it is difficult to assess the impact of adjustments in a global manner; there is a lack of visualization modeling of job relationships and business connections. Currently, job management is mostly a flat job list, which fails to build a hierarchical relationship network between positions and a cross-departmental business connection map; the impact of job responsibility changes on upstream and downstream business processes is difficult to quantify and assess, resulting in frequent problems of "mismatch between people and positions" and "business breakpoints".

[0006] (4) Disconnection between tags and business systems: Changes to multiple tags such as job title, professional title, and job level cannot automatically trigger the linkage adjustment of permissions and resources in downstream systems; the personnel tag system is disconnected from downstream systems, and the existing personnel tags (such as professional title, education, and job nature) are mostly static attributes, lacking a dynamic mapping mechanism with downstream business systems; changes to tags cannot automatically trigger the adjustment of permissions and resource allocation in related systems, requiring manual maintenance of each system, which is inefficient and prone to errors.

[0007] (5) Difficulty in tracing responsibilities: The historical trajectory and decision-making basis of organizational restructuring lack a credible and tamper-proof evidence storage mechanism, which makes it difficult to meet audit requirements; the level of intelligence in document processing is low, and school documents such as the establishment of institutions, job adjustments, and personnel appointments and removals still rely on manual review and hierarchical approval, lacking intelligent analysis and compliance verification; more importantly, the historical trajectory, decision-making basis, and responsible persons of organizational restructuring lack an tamper-proof evidence storage mechanism, which makes it difficult to meet compliance requirements.

[0008] (6) Lack of trust in data security and sharing: Centralized storage poses a risk of leakage, and cross-departmental data sharing lacks a trustworthy mechanism; organizational structure data involves sensitive personnel information, and the centralized storage model of the existing system poses a risk of data leakage and unauthorized access; cross-departmental data sharing lacks a trustworthy exchange mechanism, data providers worry about data being misused, and data users question the authenticity of the data, forming a trust barrier.

[0009] While digital twin technology is mature in the industrial manufacturing sector, its application in soft systems such as organizational management is still limited. Although blockchain technology offers the potential for trusted evidence storage and automated execution, there is still a lack of systematic solutions for how to deeply integrate it with the complex organizational governance logic of universities. It is evident that existing technological solutions are insufficient to meet the needs of modernizing university governance. Therefore, there is an urgent need for a new technological system that can achieve full digital mapping of organizational structure elements, support pre-adjustment rehearsals, ensure trusted data sharing, and possess intelligent analysis and automated execution capabilities. This system would address the core pain points of existing solutions, such as data silos, delayed adjustments, and difficulties in tracing responsibilities, thereby promoting the modernization of university governance systems and capabilities. Summary of the Invention

[0010] In view of this, the present invention provides a blockchain-based digital twin system for university organizational structure and a construction method thereof, aiming to solve the technical problems existing in the management of existing university organizational structures, such as data silos, high risk of adjustment decisions, difficulty in tracing rights and responsibilities, and low collaborative efficiency.

[0011] In a first aspect, this invention provides a blockchain-based digital twin system for university organizational structures, comprising a data acquisition layer, a data resource layer, a digital twin layer, an application service layer, and a blockchain foundation layer connected sequentially (see [link to relevant documentation]). Figure 1 );in: The data acquisition layer connects to heterogeneous business systems within universities, aggregating organizational data in real time to achieve real-time aggregation of organizational data; The data resource layer consists of an organizational master database, a relational graph library, a historical version library, and a tag mapping library, forming a unified data platform. The digital twin layer, as the core engine, enables the virtual mapping of all elements of the organizational structure, establishing: Secondary Institutional Twin Model: Create an institutional digital twin (ODT) and construct a multidimensional relationship network between institutions; Organizational adjustment digital twin path model (or job relationship twin model): to achieve seven-step closed-loop management of "adjustment proposal - twin rehearsal - impact assessment - solution optimization - on-chain evidence storage - physical execution - effect monitoring"; Personnel tagging application model: Constructing a personnel tag forest (PTF) and a dynamic tag-permission mapping mechanism; Intelligent document analysis and adjustment model: Based on a large AI model, it realizes intelligent document parsing and a decision-making mechanism of "AI assistance + manual confirmation"; The application service layer provides users with intelligent applications such as a twin pre-drill platform, intelligent document processing, and a tag application center. The blockchain foundation layer is built on a consortium blockchain (such as Hyperledger Fabric or FISCO BCOS) to construct the organizational structure of universities, deploy core smart contracts, and provide trusted storage and automatic execution services for the upper layer. Through smart contracts, the trusted execution and full traceability of core functions such as institutional registration, job management, label issuance, document storage, access control and audit traceability are realized.

[0012] In a second aspect, the present invention provides a construction method based on the above-described system, comprising the following steps: S1. Construct digital twins of secondary institutions and their relationship networks; S2. Construct and implement a digital twin path for institutional adjustment; S3. Construct a personnel tagging application model; S4. Construct an intelligent analysis and adjustment model for official documents; S5. Build a trusted foundation for blockchain.

[0013] The "Digital Twin System for University Organizational Structure" proposed in this invention constructs a complete digital management closed loop through five core modules, as shown in Table 1: Secondary organizational relationship model Digital definition of organizational attributes, hierarchical relationship modeling, historical version management, cross-departmental business flow mapping, resource dependency relationships, and quantification of collaboration intensity. Institutional Adjustment of Digital Twin Path Model Adjust the pre-rendering engine, conflict detection algorithm, and blockchain evidence storage. Personnel tagging model Multi-dimensional tag system, tag-permission mapping, dynamic update mechanism Intelligent document processing NLP analysis, compliance verification, and impact prediction. Blockchain Organizational adjustment records are stored on the blockchain for evidence preservation; business contracts are executed intelligently; job change permissions are authorized in a chain; tag changes are traced with trusted information; and approval processes are stored on the blockchain for evidence preservation. Table 1 Its main innovation lies in: (1) The first "twin mapping" model of organizational structure adapted to the complex governance structure of universities: Unlike the existing campus digital twins that only map physical space, this invention proposes a digital twin mapping of organizational structure for the first time, targeting the complex scenario where the three-element governance structure of universities coexists and special organizational relationships such as subordinate / joint office / affiliation are intertwined. It constructs abstract management elements such as secondary institutions (including special relationship types), positions, and personnel (including complex identities such as dual-role and one person holding multiple positions) into calculable, simulable, and predictable digital entities, realizing a new management model of "organizational adjustment is first rehearsed in the twin, and then implemented in the physical world".

[0014] (2) Construct a three-layer dynamic relationship network of “organization-position-personnel”: Innovatively establish a three-layer coupling model of inter-organization relationship graph (including administrative level, business dependence, resource allocation, collaborative cooperation and organizational correspondence), inter-position reporting and business contact network, and personnel tag application chain (supporting flexible combination of job-title-rank three-element identity), realize the global impact quantitative analysis of organizational structure adjustment, solve the limitation of “seeing the trees but not the forest” in traditional management, and support precise governance under complex organizational forms.

[0015] (3) Trusted organizational data governance mechanism integrating blockchain: Introduce consortium blockchain technology into organizational structure management, realize immutable evidence storage and multi-party consensus verification of key operations such as organizational adjustment, job change, and personnel appointment and dismissal through distributed ledger, build a trust infrastructure for cross-departmental data sharing, and solve the problems of data silos and lack of trust.

[0016] (4) Intelligent document parsing and “AI-assisted + manual confirmation” dual-track decision-making model: Based on the AI ​​big model, an intelligent parsing engine for university documents is built to realize the intelligent processing of the entire process of document uploading, importing, recognition, text extraction, organization / position / personnel action recognition, rule conflict detection, and draft adjustment generation; strictly follow the “AI-assisted + manual confirmation” mechanism, all parsing results must be manually reviewed and confirmed before they can take effect, and direct automatic storage is strictly prohibited, realizing the leap from “manual review” to “intelligent assisted decision-making”, and ensuring a balance between efficiency and security.

[0017] (5) Propose a complete digital twin system development framework: Based on the above research, form a feasible development framework including data standards, model specifications, interface protocols, and blockchain integration solutions, and provide a systematic solution for the digital transformation of university organizational structure management.

[0018] This invention uses digital twin technology to create a high-fidelity virtual mapping of the complex organizational structure of universities, uses blockchain technology to build a cross-departmental trust infrastructure, and uses AI big data models to provide intelligent decision support. Ultimately, it forms a new paradigm for university organizational structure governance characterized by "virtual-real mapping, trusted flow, intelligent decision-making, and human controllability," which significantly improves data consistency, scientific decision-making, traceability credibility, and the level of collaborative automation. Attached Figure Description

[0019] The accompanying drawings described herein are for illustrative purposes only and are not intended to limit the scope of the invention in any way; the shapes and proportions of the components in the drawings are merely illustrative and are intended to aid in understanding the invention, and are not intended to specifically limit the shapes and proportions of the components of the invention; those skilled in the art, under the guidance of this invention, can select various possible shapes and proportions to implement the invention according to specific circumstances.

[0020] Figure 1 This is a schematic diagram of the overall architecture of the blockchain-based digital twin system for university organizational structures, as described in this embodiment of the invention. Figure 2 This is a seven-step closed-loop flowchart of the mechanism adjustment digital twin path model in this embodiment of the invention; Figure 3 This is a flowchart of the document intelligent analysis and adjustment model according to an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can be arranged and designed in various different configurations.

[0022] ( Example 1 )refer to Figure 1 This embodiment provides a blockchain-based digital twin system for university organizational structures, which adopts a "four-layer architecture + blockchain foundation" design, covering a data acquisition layer, a data resource layer, a digital twin layer, and an application service layer; the specific composition and interaction of each layer will be described in detail below.

[0023] The data acquisition layer serves as the system's data entry point, responsible for interfacing with the university's dispersed and heterogeneous source business systems. These systems typically include, but are not limited to, Human Resource Management Systems (HRMS), Office Automation Systems (OA), Academic Affairs Management Systems, Research Management Systems, Financial Management Systems, and Asset Management Systems. This layer utilizes data extraction, transformation, and loading (ETL) tools or API-based data interfaces to achieve real-time aggregation or periodic incremental collection of organizational structure-related data from these systems. The collected data covers: basic organizational information (code, name, level), basic personnel information (employee ID, name, department), job information, personnel tags (title, position, rank), and business-related data (budget, project). To ensure data quality, this layer also includes preliminary data cleaning and formatting rules to convert non-standard data into a unified format that can be processed downstream.

[0024] The data resource layer is the data hub of the system. It performs standardized processing, fusion, and persistent storage of the collected raw data, and constructs the following core databases to form a unified data platform: 1) Organizational master database: Stores verified and authoritative core entity information of institutions, positions, and personnel, serving as the "single trusted data source" for the entire system; 2) Relationship Graph Library: Constructed using a graph database (such as Neo4j); with organizations, positions, and personnel as nodes, and their administrative affiliations, business collaborations, resource allocations, and organizational correspondences as edges, a large-scale, multi-dimensional organizational relationship knowledge graph is built; each edge can be assigned attributes such as weight, type, and time. 3) Historical version repository: Using a time-series database or a versioned data storage solution, it records the status change history of any entity (organization, position, personnel employment relationship) in the organizational structure, supports backtracking queries by time point, and meets auditing and traceability needs; 4) Tag Mapping Library: Stores the definition and classification system of personnel tags (identity, position, job title, job level, business, skills), and most importantly, the dynamic mapping rules between tags and downstream business system permissions / resources; each rule defines what operation should be automatically triggered on the target system (such as access control, finance, and academic affairs systems) when a specific tag combination appears or changes.

[0025] The digital twin layer is the core intelligent engine of the system. Based on the data from the data resource layer, it constructs and runs a dynamic digital twin model, and establishes a secondary organizational twin model, an organizational adjustment digital twin path model (i.e., a job relationship twin model), and a personnel tag (digitalization) twin model to achieve virtual mapping of all elements of the organizational structure.

[0026] To address the unique complexity of university organizational structures—the coexistence of a three-tiered governance structure and the intertwining of special organizational relationships such as subordinate / joint / affiliated entities—a digital twin system for secondary institutions is constructed. This system achieves digital mapping of all organizational elements through an institutional digital twin (ODT), encompassing five dimensions: basic information, spatial resources, staffing, business functions, and related relationships, forming a dynamically updatable virtual organizational profile. Furthermore, a graph database is used to construct a multi-dimensional relationship network, including four types of relationships: administrative hierarchy, business dependencies, resource allocation, and collaborative cooperation. This transforms the static organizational structure into a computable and analyzable relationship graph, supporting precise governance under complex organizational structures.

[0027] This model aims to create a corresponding, computable, and simulable "Organizational Digital Twin" (ODT) for each secondary organization (such as a college or department) in the physical world; each ODT contains the following digitally defined attributes, as shown in Table 2: Basic attributes Organization code, name, abbreviation, English name, date of establishment, nature of the organization (teaching / research / management / teaching support), administrative level, type of organizational relationship (subordinate / joint / affiliated) Structured data model Spatial attributes Office location (multiple campuses), building area, number of laboratories, signage for shared office spaces GIS Spatial Database Personnel attributes Number of authorized positions, actual number of employees, job structure distribution, identification of employees with dual roles, and pool of part-time employees. Data extraction from human resources system Business attributes Main functions, core business indicators, annual budget, and extended management / territorial management attributes Business system data extraction Relationship attributes Superior supervisory authority (administrative subordinate), peer cooperating institutions, intensity of business dealings, and corresponding party organization relationships. Graph database association Table 2 1) Basic attributes: organization code, name, establishment date, organization nature (teaching / research / management), administrative level, organizational relationship type (subordinate / joint office / affiliated); among them, special relationship types such as "subordinate", "joint office", and "affiliated" have clear digital logical definitions. For example, the "joint office" relationship will trigger the creation of virtual shared space and joint approval workflow; 2) Spatial attributes: Link to the GIS system to record its office location (supports multiple campuses), building area, number of laboratories, etc.; 3) Personnel attributes: Synchronize from HRMS, including the number of authorized positions, actual number of employees, distribution of various positions (teaching, research, management), list of personnel with dual roles, and pool of available part-time personnel, etc. 4) Business attributes: Extracted from the business system, including main function descriptions, core business indicators (KPIs), annual budget amount, and whether it has extended management or local management attributes, etc. 5) Relationship attributes: This is the dynamic part of the ODT, obtained by querying the relationship graph library; it defines the relationship and strength between the organization and its superiors, peers, and subordinates in multiple dimensions such as administration, business, resources, and collaboration.

[0028] Based on the ODTs and their relational attributes of all institutions, this model utilizes graph database technology to construct a global, multidimensional secondary institutional relationship network, as shown in Table 3: Hierarchical relationship Administrative affiliation (tree structure), dual-line correspondence between the Party Committee and the administration Tree diagram model Business Relationship Cross-departmental business process dependencies (directed graph), collaborative office processes Directed graphical model Resource relationships The allocation of shared equipment, venues, and funding; resource coordination across multiple campuses. Resource allocation network Collaborative relationships Based on the strength of collaboration (weighted edges) of historical project cooperation and the participation of deliberative bodies Weighted network model Table 3 This network transforms the static "department tree" into a complex relationship graph capable of network analysis (such as centrality analysis, community discovery, and influence propagation simulation), providing a data foundation for quantifying the mutual influence between organizations. For specific organizational relationships, the digital twin employs the following technical means: Subordinate relationships: Strongly correlated structure, vertical leadership, with superiors having full management authority over subordinates, including personnel appointments and dismissals, financial approvals, and performance evaluations; digital twins are represented through composite relationships, enabling automatic inheritance of permissions and data penetration queries; Joint operation relationship: Two or more independent agencies share administrative teams and office space, retain their respective names externally, and operate in an integrated manner internally; digital twins are represented by independent categories + shared related categories, establishing virtual shared spaces and joint approval processes; Affiliation relationship: Entities without independent legal person status are attached to the main responsible entity, maintaining business autonomy but relying on resource support; digital twins represent the dependency relationship through combination relationships, and adopt a data interaction mode of logical isolation + limited interfaces.

[0029] refer to Figure 2 This model provides a full-lifecycle digital management channel for institutional adjustments (establishment, merger, split, abolition, and renaming). Through a "seven-step closed-loop" process, it achieves scientific and controllable management of major organizational restructuring in universities. Its core value lies in transforming traditional "experience-based decision-making and post-event recording" into "data-driven, pre-event simulation"—simulating scenarios such as institutional mergers, splits, and abolitions in the digital space through an adjustment simulation engine, quantitatively assessing the chain reaction impact on personnel placement, budget transfers, and business continuity; automatically identifying contradictions with existing systems using conflict detection algorithms, and intelligently generating optimized solutions; and utilizing blockchain notarization to ensure the entire process of proposal, approval, and execution is tamper-proof and fully traceable, significantly reducing the trial-and-error costs and decision-making risks of institutional adjustments, providing key technical support for the modernization of university governance systems and capabilities. It includes the following seven sequentially executed steps forming a feedback loop: (1) Adjustment Proposal: As a digital entry point for organizational adjustment needs, this step transforms unstructured adjustment intentions into standardized digital models; the initiator fills in the adjustment type (establishment, merger, split, revocation, renaming), the system automatically extracts existing organizational data to generate a baseline twin before adjustment, and preliminarily verifies the completeness and legality of the proposal, outputting structured proposal data (pre-adjusted organizational structure information, personnel change information), baseline status before adjustment, and a unique proposal identifier; the user initiates the adjustment intention through the front-end interface; the system guides the user to select the adjustment type, involved organizations, reasons for adjustment, etc., and automatically extracts the baseline status (ODT snapshot) of relevant organizations from the current digital twin system to generate a structured digital proposal.

[0030] (2) Adjustment plan: The adjustment plan module is the core processing unit of the system. It is responsible for converting the unstructured adjustment intentions input by users into standardized digital plans (i.e., twin simulations). This module generates the complete adjusted plan and the adjusted organizational status directly by automatically matching the database related data, providing a structured basis for subsequent approval decisions.

[0031] 1) Input information: a. Adjusted organizational structure information: Adjustment types: establish, merge, split, cancel, rename; different types of adjustments will trigger differentiated data processing rules and state transition logic.

[0032] Baseline status before adjustment: organization code, organization name, organization level, affiliation, functional positioning, authorized number of employees, actual number of employees, and list of internal departments.

[0033] b. Personnel change information: Personnel list: Name, Employee ID, Original position, Original organization, Personnel category (permanent / non-permanent / rehired); Types of changes: retention, transfer in, transfer out, job change, retirement, dismissal; Target positions: new organization, new department, new job title, new rank; Change time: Proposed effective date and transitional arrangements.

[0034] c. Associated resource configuration (Associated resources refer to all business and asset data under the name of the personnel being transferred, which are automatically extracted by the system from each business sub-database without manual entry): Asset information: office space, equipment assets, budget, and research projects; Business information: Courses undertaken, ongoing research projects, external cooperation agreements, and management system documents.

[0035] 2) System processing: Based on the input adjustment information, the system automatically matches all associated data in the database and generates the adjusted mechanism state according to predetermined rules. The processing logic is divided into four parallel stages: a. Personnel mobility processing The system updates the personnel's affiliated organization, department, position, and rank information based on the type of change. For retained personnel, the original position association is maintained, but the organizational affiliation is updated; for transferred personnel, the position association is added after verifying the staffing quota; for transferred-out or reassigned personnel, the original position association is removed and a new association is established; for retired or dismissed personnel, the personnel status is marked as "off-duty" and the staffing quota is released. The system automatically calculates the availability of key positions (responsible persons, discipline leaders, technical backbones) and personnel structure distribution (professional title, education, age) after the processing.

[0036] b. Asset transfer processing Based on personnel changes, the system automatically links their asset data and reallocates assets according to the target organization. Office space requirements are recalculated based on the new organization's staffing and departmental setup; equipment assets are transferred to a different organization along with changes in custodians; budget funds are divided or transferred entirely according to the personnel transfer ratio; and research projects are updated with the implementing entity based on changes in project leaders.

[0037] c. Business handover process Based on personnel changes, the system automatically links the business data to the new personnel and reassigns them according to the new institution. Courses are offered by a different institution based on the new instructor's affiliation; ongoing projects are managed according to the new project leader's affiliation; and external cooperation agreements are transferred to a new responsible party.

[0038] d. Handling of institutional transitions The system identifies the scope of application of the original institutional regulations and marks them into three categories: In Use (adjustments do not affect applicability), Requires Revision (changes in institutional name or affiliation render the wording invalid), and Requires New Formulation (newly established institutions or functional integration after mergers create gaps in regulations). For regulations requiring revision or new formulation, the system generates suggested responsible departments and completion deadlines.

[0039] Based on the input adjustment information, the system generates a revised organizational twin plan (including new personnel, assets, and business configurations), as shown in Table 4: New organizational structure New organization code, new organization name, new organization abbreviation, new organization type, new organization level, new affiliation, new establishment date (at the time of establishment / merger), new dissolution date (at the time of dissolution), new main responsibilities, new business list. New internal architecture List of newly established internal departments (department code, department name, new functions, new head, staffing), and table of new job positions (job code, job name, job category, new job level, current staff). New personnel configuration Staffing Summary Table (Name, Employee ID, New Organization, New Department, New Position, New Rank, Type of Change, Effective Date), Key Position Availability (Person in Charge / Discipline Leader / Technical Backbone), Staff Structure Statistics (New Distribution of Professional Titles / Education Levels / Ages) New asset ownership New allocation of office space (adjustments to area, location, and purpose); list of transferred equipment and assets (asset code, asset name, transferring agency, receiving agency, transfer date, new custodian); new allocation of budget (annual budget, source composition, new implementing entity). New business undertaking Course Transfer Form (Course Code, Course Name, New Responsible Institution, New Instructor, Semester Offered), Ongoing Project Transfer Form (Project Code, Project Name, New Responsible Institution, New Person in Charge, Transfer Status), Cooperation Agreement Amendment Form (Agreement Number, Partner, New Implementing Entity, Amendment Description) New institutional system List of existing regulations (regulation name, document number, explanation of continued application), list of regulations requiring revision (regulation name, key revision points, responsible department, completion deadline), and suggestions for newly formulated regulations (regulation name, basis for formulation, drafting department). New relationship New list of collaborating departments (new collaborating institutions and collaborating matters), new external affiliations, and historical background (background, basis for adjustment, and approval document number). Table 4 The twin simulation model calculates a professional matching score S1 by measuring the cosine similarity between the personnel skill tag vector and the job requirement vector, and combines this with a job grade weight table to obtain a job grade suitability score S2. The final placement decision is based on the comprehensive score S = α. S1 + β S2, and is subject to the number of staff.

[0040] (3) Impact assessment: Based on the results of the plan, this step relies on AI big models (such as DeepSeek-V4-Flash, Qwen3.6) to create an intelligent agent for analyzing organizational structure adjustment proposals, conduct multi-dimensional and multi-scenario intelligent analysis of the plan results, analyze the chain impact on the overall operation of the university, identify potential risk points, and output impact assessment reports, risk ratings and high-risk point lists.

[0041] The following is an example of the prompt words for the AI ​​agent proposing organizational restructuring: "Role setting: You are an intelligent pre-simulation assistant for institutional adjustments in higher education institutions, specializing in organizational structure simulation and multi-dimensional impact assessment; based on the input adjustment plan, you simulate the entire execution process and output a structured analysis report."

[0042] Input data: Baseline status information before adjustment, and twin contingency plan after adjustment.

[0043] Analysis dimensions: Business continuity Identify potential disruptions to ongoing projects and gaps in course offerings after the implementation of contingency plans; output a list of business interruption risks and project handover safeguards. Financial impact Calculate budget transfer deviations and variable costs of salaries; output detailed cost increases and decreases, budget execution forecasts, financial health scores, and early warnings of funding gaps; Institutional compliance By comparing the organizational structure with the superior's regulations and current systems, personnel policies, and financial systems, identify the points of conflict with existing rules and regulations (such as organizational structure regulations and budget management methods), and output the system conflict identification, compliance review opinions, and a list of systems that need to be revised; Multi-scene comparison Construct parameters for three sets of solutions: aggressive, conservative, and compromise; output a comparison table of the execution results of the three solutions, a recommendation of the optimal solution, key assumptions, and sensitivity analysis; Initial risk screening Aggregate the above analysis results and assess them according to the transmission path; output a risk classification list (green / yellow / red), special explanations for high-risk points, risk mitigation suggestions, and emergency response plans.

[0044] Output: Adjusted mechanism twin (JSON format), multi-scenario simulation report (with conflict point annotations), risk assessment report (explaining the risk points of each solution).

[0045] (4) Solution Optimization: For high-risk points identified in the impact assessment, the AI ​​big model generates optimization suggestions or alternative solutions based on the local predefined optimization strategy knowledge base; for the identified risk points, the optimization strategy knowledge base is matched to generate multiple alternative solutions and conduct multi-dimensional comparisons, providing a "hypothesis analysis" tool to support decision-makers in manually adjusting parameters and observing, and outputting optimization solution suggestions, solution comparison analysis and recommended implementation strategies; for example, if "conflicts in the placement of core business backbones" are identified, the strategy knowledge base matches and suggests "establishing a dual-responsibility system during the transition period" or "providing special training for job transfer"; the system can generate multiple optimized alternative solutions and provide comparative analysis to help decision-makers weigh the pros and cons.

[0046] (5) On-chain evidence storage: Key information (or its hash value) of the final plan (including proposals, preliminary reports, evaluation reports, optimized plans, and approval records) confirmed by the decision-maker will be written to the blockchain by calling the smart contract of the blockchain base layer; this step ensures that the entire decision-making process is tamper-proof and traceable; at the same time, the relevant execution instructions (such as "create a new organization A" or "transfer personnel X from organization B to organization A") are also encoded as the triggering conditions of the smart contract; this method of writing the key data of the entire process to the blockchain ensures credible traceability and automatic execution; and the hash of proposals, preliminary results, approval records, optimized plans, etc. are on the chain, and the smart contract can automatically trigger the execution instructions when the preset conditions are met, reach a consensus across departments and be endorsed by multiple nodes, and output the on-chain transaction hash, smart contract execution instructions and an immutable evidence chain.

[0047] (6) Physical execution: The digital twin adjustment plan confirmed by the decision is transformed into task decomposition and execution instructions for functional departments. After the departments execute the plan, the specific operations are completed through the business system, the progress is written back in real time, and an early warning is issued when the deviation exceeds the threshold. The execution instruction package, execution status feedback and physical world adjustment implementation results are output.

[0048] (7) Effect monitoring: Evaluate the actual effect of the adjustment, compare it with the pre-adjustment prediction to form a closed-loop optimization, and output the effect evaluation report. If the plan is adjusted during the implementation process, the basis should be provided and the corresponding adjustment information should be added to the strategy knowledge base mentioned in the plan optimization steps.

[0049] To achieve the aforementioned twin simulation, the following provides a specific feasible solution including inputs, calculation rules, outputs, and constraints. For example, a personnel assignment model based on a weighted scoring algorithm, where the system calls and executes the core simulation model.

[0050] 1) Personnel reassignment simulation model: a. Input: The list of personnel involved in the adjustment L_person (including personnel ID, original position, skill tag set TS_skill, job level LR) and the list of target positions L_position (including position ID, required skill tag set TS_required, job level requirement LR_req).

[0051] b. Processing Logic: A multi-rule weighted scoring algorithm is used to calculate the comprehensive matching score Score_ij for each person P_i with each vacant position POS_j. The calculation formula is as follows: Score_ij = α * S_similarity(TS_skill_i, TS_required_j) + β * F_level(LR_i, LR_req_j); Wherein, S_similarity is the skill tag matching function, which calculates the Jaccard similarity coefficient between TS_skill_i and TS_required_j, with a value range of [0, 1]; F_level is the job level fit function, which defines a mapping table, such as "1.0 for a perfect job level match, 0.8 for a difference of one level, 0.5 for a difference of two levels, and 0 otherwise"; α and β are configurable weight coefficients (default settings α=0.7, β=0.3), used to balance professional matching and job level qualifications.

[0052] c. Output and Constraints: The model sorts Score_ij from high to low. Under the premise of satisfying the "one person, one position" and position staffing constraints, it performs optimal or suboptimal matching through the Hungarian algorithm or greedy algorithm, outputs a list of personnel-position mapping relationships, and marks matching risk points where Score_ij is lower than the threshold (e.g., 0.6).

[0053] 2) Asset transfer model: a. Model basis: Based on the resource dependency graph, personnel, equipment, and budget are regarded as nodes, and their affiliation relationships are regarded as edges.

[0054] b. Processing Logic: Based on the above personnel assignment results, the resource dependency graph is automatically traversed; when a person P_i is transferred from institution A to institution B, the model automatically redirects the ownership edges of the non-shared asset nodes (such as dedicated equipment, personal research funding accounts) under their name from institution A to institution B; for shared assets (such as public laboratories), the model checks the changes in the set of associated personnel. If more than a preset proportion (such as 70%) of the associated personnel of an asset are transferred to institution B, it is recommended to transfer the main management authority of the asset to institution B.

[0055] c. Output: Generate an asset transfer list, including the ID, name, original owner, new owner, transfer type (assigned to person / redistributed), and basis for transfer for each asset.

[0056] 3) Business process integration model: a. Model basis: Based on business process diagrams, where nodes are business links and edges represent the logical and data dependencies between links.

[0057] b. Processing Logic: Identify the business process sub-graphs affected by organizational mergers / splitting; for mergers, the model attempts to merge and normalize parallel or similar business processes originally belonging to different organizations, and checks the consistency of input and output interfaces; for splits, the model cuts the original complete process into independently runnable sub-processes based on the binding relationship between business processes and personnel, and automatically inserts standardized data handover interfaces at the cutting points.

[0058] c. Output: Output the updated business process diagram, and list the changed steps, interfaces, and potential business interruption risks.

[0059] This module addresses the complexity of university personnel identities—characterized by parallel roles, titles, and ranks, separation of positions from personnel, and individuals holding multiple positions or dual roles—by constructing a personnel tagging application model. Through a multi-dimensional tagging system and dynamic permission mapping mechanism, it achieves refined definition and intelligent cross-system management of university personnel identity attributes. Given the complexity of university personnel identities (multiple positions or dual roles), this model utilizes a multi-dimensional tagging system to achieve refined and dynamic management.

[0060] Multi-dimensional Tag Forest (PTF): This constructs a dynamic set of tags for each individual, covering five major categories: identity, position, job title, professional title, job level, business responsibilities, and skills. See Table 5 for details. Identity tags Faculty / Students / Visitors / Retirees Access control permissions, application system account permissions, etc. Job tags Professor / Associate Professor / Lecturer / Teaching Assistant (Professional Title Position); Director / Section Chief / Clerk (Administrative Position) Teaching workload calculation, evaluation qualifications, and salary levels Job title tags Professional and technical positions (levels 1-13), management positions (levels 1-10), and skilled worker positions (levels 1-5) Human Resources System, Salary Grading, Retirement Benefit Calculation Professional title tags Senior Professional (Professor / Researcher), Associate Senior Professional (Associate Professor / Associate Researcher), Intermediate Professional (Lecturer / Assistant Researcher), Junior Professional (Teaching Assistant / Research Intern) Human resources system, review of qualifications, and certification of mentor qualifications Job Tag Dean / Vice Dean / Department Head; Director / Deputy Director of Academic Affairs Approval authority, scope of meeting notification, decision-making weight Business Tags Project Manager / Lab Technician / Safety Officer / Purchasing Officer Business system permissions, pending tasks, and specific responsibilities Skill Tags Large-scale models / Blockchain / Knowledge graphs / Higher education management Expert database retrieval and intelligent task allocation Table 5 It supports flexible combinations of three identities: position, title, and rank; identity tags distinguish personnel types such as faculty, students, visitors, and retirees, supporting access control and application system account permission control; position tags identify professional titles such as professor, associate professor, lecturer, and teaching assistant, serving the calculation of teaching workload and review of qualifications; job title tags reflect administrative levels such as dean, department head, section chief, and staff member, determining approval authority and the scope of meeting notifications; business tags mark specific roles such as project leader and laboratory technician, associating them with business system permissions and push notifications for pending tasks; skill tags record professional capabilities such as large-scale models, blockchain, and knowledge graphs, supporting expert database retrieval and intelligent task allocation; all kinds of tags together constitute a digital profile of personnel, providing a data foundation for refined permission management; for example, a professor may simultaneously possess multiple tags such as identity: faculty, position: professor, title: vice dean, professional title: senior professor, business: project leader, and skill: blockchain.

[0061] Special identity scenarios are supported: a. Dual-role personnel: Simultaneously link administrative position tags and professional technical title tags, and set conflict detection rules (e.g., allowed by scientific research management, prohibited by financial approval). b. One person, multiple positions: Dynamically bind multiple positions through the employment relationship table and switch identity contexts according to the scenario; c. Separation of evaluation and appointment: The title label and the position appointment status are managed separately, supporting "high-level appointment to low-level position" or "low-level appointment to high-level position".

[0062] Tag-Based Access Control (TBAC) mechanism: A tag-based access control model is established, and automatic mapping between tags and downstream system permissions is achieved through a dynamic rule base; the rule engine supports condition combinations and state triggers, which is key to achieving automation; a series of rules are predefined in the tag mapping library, see Table 6: The tag includes "laboratory safety officer" Automatically assign "data entry permissions" Laboratory Management System Static authorization The tag contains "Project Leader" AND Project Status="In Progress" Dynamically assign "approval authority" Financial System Conditional authorization The tag "Professor" has been changed to "Associate Professor". Triggering "Mentor Qualification Review" Graduate System State change trigger Table 6 Example of a rule: The IF tag containing "Laboratory Safety Officer" will automatically grant "Safety Inspection Reporting" permissions in the laboratory management system. If the title is changed from "Associate Professor" to "Professor", then the "Supervisor Qualification Review" process in the graduate student system will be automatically triggered, and the Human Resources Department will be notified to update the expert database. If the IF tag contains "Project Leader" AND the associated project status is "In Progress" THEN, the funding approval authority for the corresponding project will be dynamically opened in the financial system; when the project status changes to "Completed", the authority will be automatically revoked.

[0063] When personnel tags change due to appointment, dismissal, or evaluation, the mapping engine automatically detects and executes the corresponding permission operations, realizing "tags are permissions, and changes take effect immediately." This reduces the original manual coordination work across systems from several days to seconds. This mechanism breaks through the traditional role-based static permission management model, significantly improves the real-time performance and accuracy of cross-system permission configuration, and reduces manual maintenance costs and the risk of permission lag.

[0064] This model applies a large-scale AI model to the processing of documents related to organizational restructuring, automating document classification, element extraction, and compliance verification, thereby improving efficiency and reducing human error; see also Figure 3We have built an intelligent document parsing engine, which is based on an AI big data model intelligent agent to realize intelligent processing of the entire process, including document uploading, importing, recognition, text extraction, organizational action recognition, job action recognition, personnel appointment and removal action recognition, rule conflict detection and draft adjustment generation. We strictly follow the dual-track mechanism of "AI assistance + manual confirmation" to ensure the human controllability of key decisions.

[0065] Before using the AI ​​model, it was supervised fine-tuned using a dataset consisting of historical university documents and corresponding labeled information (institutions, actions, and personnel entities). The standard cross-entropy loss function was used to iteratively train it for N epochs, enabling it to have the specific recognition ability of university document terminology and structure.

[0066] Intelligent document parsing engine: Based on the construction of a domain-adapted AI large model, to adapt a general large model (such as Qwen3.6) to the vertical domain task of university document parsing, any or a combination of the following technical approaches can be adopted: (Path 1) Supervised Fine-Tuning (SFT) specifically includes: 1) Training data construction: Collect historical university documents (such as notices of establishment of institutions, appointment and dismissal, etc.) to form the original text set; domain experts perform sequence labeling on each document, labeling the entities in the text (such as institution name, person name, position) and their types (ORG, PERSON, TITLE), as well as keywords (ACTION_MERGE, ACTION_APPOINT) and attributes (effective date DATE) describing the adjustment action.

[0067] 2) Model training: Using a general large model as a base, the above-mentioned labeled data is used, and the standard cross-entropy loss function is adopted. Supervised full parameter fine-tuning or LoRA / LoRA+ parameter efficient fine-tuning is carried out at a low learning rate (such as 2e-5) for 3-5 epochs. This process enables the model to internalize the language style, entity composition and action logic of university documents.

[0068] 3) Application: The fine-tuned model has enhanced end-to-end information extraction capabilities from input documents, and can directly replace or assist the multiple serial agents in the original document to complete the mapping from raw text to structured information.

[0069] (Path Two) Retrieval-Augmented Generation (RAG) combined with hint engineering, specifically including: 1) Knowledge base construction: The texts of the university's regulations and systems, such as the "Regulations on Institutional Establishment Management", "Post Setting Plan" and "Regulations on the Management of Leading Cadres", are segmented and vectorized, and stored in a vector database to build a compliance knowledge base.

[0070] 2) System workflow: a. Analysis and Extraction: First, the zero-shot / few-shot learning capability of the general large model is utilized, combined with carefully designed prompt words (such as the example provided later), to initially extract the adjustment action elements from the official document; b. Compliance verification: Using the initially extracted elements (such as "proposed establishment of an Artificial Intelligence College with a staff of 50") as query conditions, relevant institutional provisions (such as "the maximum staffing of school-level teaching institutions is 40") are retrieved from the vector knowledge base. c. Conflict Analysis and Report Generation: Input the document elements and relevant policy clauses retrieved into the large model, instruct it to perform comparative analysis and generate a compliance report; for example, the prompt could be: "Please compare the following 'adjustment items' with 'related policies' and list all conflict points: Adjustment items: [elements extracted above]; Related policies: [retrieved clauses]".

[0071] Through deep adaptation training via Path One or external knowledge enhancement via Path Two, the AI ​​large model system acquires reliable intelligent parsing and compliance verification capabilities for university official document scenarios, thereby supporting the actual operation of the "official document intelligent analysis and adjustment model".

[0072] The document intelligent parsing engine consists of a series of interconnected, function-specific AI agents, as shown in Table 7: Document Upload, Import, and Recognition Document access to intelligent agents Supports uploading and format recognition of official documents in multiple formats. PDF / Word / scanned document parsing, official document template matching, document number extraction Document Extraction Text Extraction Agent Automatically extract document title, addressee, body text, and attachment list. Title, issuing authority, document number, date of issuance, and structured body paragraphs Organizational Action Recognition Organize action recognition intelligent agents Identify actions such as the establishment, merger, split, cancellation, and renaming of an organization. Organization Name, Type of Adjustment, Change of Affiliation, Effective Date Job motion recognition Job Action Recognition Intelligent Agent Identify actions such as job creation, adjustment, cancellation, and changes in the number of positions. Job title, job category, changes in number of positions, and department. Personnel appointment and removal action recognition Intelligent agent for personnel appointment and removal Identify personnel appointments, removals, concurrent appointments, and dismissals. Personnel name, ID number, original position, new position, appointment / removal date Rule Conflict Detection Compliance verification intelligent agent Points of conflict between testing and existing systems Organizational establishment: Does it exceed the authorized staffing level? Does it conflict with the functions of existing organizations? Job adjustments: Do they comply with the job setting plan? Do they exceed the quota limit? Personnel appointments and removals: Do they meet the qualifications for the position? Are the procedures compliant? Generate draft adjustment Draft generates intelligent agents A structured adjustment draft is generated based on the identification results. Target of adjustment, content of adjustment, scope of impact, implementation steps, and risk warnings Table 7 This AI agent uses current mainstream AI models (such as DeepSeek-V4-Flash, Qwen3.6): 1. Document Information Extraction Intelligent Agent: Parses text, uploaded PDF / Word / scanned documents, etc., and identifies document type, document number, and template; extracts structured information such as document title, issuing authority, date of issuance, core paragraphs of the main text, and list of attachments.

[0073] The following are examples of prompt words for the document information extraction intelligent agent: "Role setting: Document parsing assistant, automatically identifies and extracts key information from official documents."

[0074] Input data: Official documents (PDF / Word / images), official document text.

[0075] Task: Identify document type and document number; extract structured information such as document title, issuing authority, date of issuance, and abstract.

[0076] Output example: Title: xxx, Document Type: xxx, Document Number: xxx, Abstract: xxx, Keywords: xxx, Contact Person: xxx, Contact Information: xxx, Issuing Unit: xxx, Issuance Date: xxx.

[0077] 2. Organizational Structure Action Recognition Agent: Identifies whether an action is an organizational action, a job action, or a personnel appointment or dismissal. If it is an organizational action, it identifies actions related to "establishment," "merger," "splitting," "cancellation," or "renaming" of the organization in the main text, and extracts the organization name, adjustment type, change of affiliation, and effective date. If it is a job action, it identifies actions related to "addition," "adjustment," "cancellation," or "change in the number of positions." If it is a personnel appointment or dismissal, it identifies actions such as "appointment," "removal," or "part-time employment," and extracts the personnel's name, ID number, original / new position, and appointment / dismissal date.

[0078] Examples of prompts for organizational action recognition agents are as follows: "Role setting: You are an expert in identifying organizational restructuring actions, responsible for accurately identifying three types of adjustment behaviors from official documents: organizational actions, job actions, and personnel appointments and removals, and extracting key information in a structured manner."

[0079] Input data: Official documents (PDF / Word / images), official document text.

[0080] Tasks: 1) Determine the type of adjustment: If the text involves the establishment, integration, splitting, cancellation, or renaming of an organization, classify it as an organizational action; if the text involves the addition, adjustment, cancellation, or change of the number of positions, classify it as a position action; if the text involves the appointment, dismissal, or concurrent employment of personnel, classify it as personnel appointment and dismissal. 2) Extract elements by type: If it is an organizational action, identify actions related to the "establishment," "merger," "split," "cancellation," or "renaming" of the organization in the text, and extract the organization name, adjustment type, change of affiliation, and effective date; if it is a position action, identify actions related to the "addition," "adjustment," "cancellation," or "change of the number of positions" of positions; if it is a personnel appointment and dismissal, identify actions such as "appointment," "dismissal," or "concurrent employment" of personnel, and extract the personnel name, ID number, original / new position, and appointment / dismissal date. 3) Analyze affiliation: "Managed by XX," "Leaded by XX," "Belonging to XX," "Affiliated to XX" → extract as affiliation; when multiple organizations are merged, list all the names of the merged organizations and the name of the new organization after the merger. 4) Date recognition: Prioritize extracting dates with explicit effective dates such as "from X year X month X day", "effective from X year X month X day", and "decided on X year X month X day"; if there is no explicit date, mark it as "not specified".

[0081] Output example: Output as a JSON array, each record contains: { "adjustment_type": "Organizational actions / Job actions / Personnel appointments and removals", "action": "Specific action keywords", "details": { / / Organization Action Field "institution_name": "Institution Name", "adjustment_category": "Establish / Merge / Split / Revoke / Rename", "original_affiliation": "original affiliation", "current_affiliation": "Current affiliation", / / or job action field "position_name": "Job Title", "belonging_institution": "belonging institution", "original_quota": "original number of positions", "current_quota": "current number of positions", / / or personnel appointment and removal field "person_name": "Person's name", "id_number": "ID card number", "original_position": "original position", "new_position": "new position", "appointment_type": "appointment / removal / part-time" }, "effective_date": "Effective date" }".

[0082] 3. Compliance Analysis and Draft Generation Intelligent Agent: The system automatically compares the identified adjustment actions with a pre-set database of rules and regulations (such as organizational structure regulations, job setting plans, and qualification regulations) to identify conflict points (such as "the proposed establishment of an organization exceeds the total number of authorized positions" or "the proposed personnel do not meet the minimum service period"). It also summarizes the information from the document information extraction intelligent agent and the organizational structure action identification intelligent agent to automatically generate a structured "Implementation Plan for Organizational Adjustment (Draft)", which includes the adjustment content, scope, implementation steps, and risk warnings.

[0083] Examples of prompts for compliance analysis and draft generation agents are as follows: "Role setting: You are a senior compliance review expert and official document drafting specialist in the field of institutional staffing management in higher education institutions. You are proficient in institutional documents such as the 'Regulations on the Management of Institutional Staffing of Public Institutions,' 'Implementation Plan for Post Setting Management,' and 'Regulations on the Qualification of Leading Cadres.' You are skilled at transforming unstructured adjustment intentions into structured implementation plan drafts."

[0084] Input data: Organizational structure action recognition intelligent agent output data, and a pre-set library of rules and regulations.

[0085] Task: Output the following in tabular form: rule number, review item, review result (compliant / conflicting / requires review), conflict explanation, risk level (high / medium / low), and recommended measures; generate the "Institutional Adjustment Implementation Plan (Draft)", including the adjustment content, scope, implementation steps, and risk warnings.

[0086] Output Example: Draft Implementation Plan for Institutional Restructuring of XX University I. Basis for Adjustment 1.1 Policy Basis (Reference of superior document number and clauses) 1.2 Practical Basis (Business Needs, Development Requirements, etc.) II. Adjustments 2.1 Organizational Structure Adjustment (List of Established / Revoked / Merged / Renamed Organizations) 2.2 Job Posting Adjustments (List of New / Removed / Changed Job Positions and Number of Positions) 2.3 Personnel allocation adjustments (appointment, removal, transfer, and reassignment plans) III. Scope of Involvement 3.1 Scope of Organizations (List of Departments / Units Affected by the Adjustment) 3.2 Scope of Personnel (Statistics on the Number of Personnel and Their Identity Categories) 3.3 Assets and Business Scope (List of Asset Transfers and Business Handovers) IV. Implementation Steps 4.1 Preparation Phase (Timeline, Responsible Departments, Prerequisite Procedures) 4.2 Implementation Phase (Phase-by-Phase Implementation Plan, Key Milestones) 4.3 Acceptance Phase (Evaluation Criteria, Feedback Mechanism) V. Risk Warnings and Countermeasures 5.1 Risk Assessment (Overstaffing Risk and Mitigation Plan) 5.2 Personnel risks (disputes over relocation and resettlement, labor relations risks) 5.3 Business Risks (Service Interruption, Data Migration Risks) 5.4 Compliance Risks (Conflicts with Existing Regulations and Rectification Path)

[0087] The "AI-assisted + human confirmation" dual-track mechanism is the core constraint to ensure safety and controllability, as shown in Table 8. Document Analysis Automatically extract elements and identify questionable content. Verify the accuracy of key elements and supplement missing information. Structured document data (pending confirmation) Action recognition Categorize and label organizational / position / personnel actions Confirm the action type and adjustment range Action recognition report (pending confirmation) Conflict detection Automatically compare rules and regulations, and highlight conflict points. Review the reasonableness of conflict determination and approve exceptions. Compliance verification report (pending confirmation) Draft generation Automatically generate draft adjustments and implementation recommendations Revise the draft content and confirm the implementation strategy Draft amendments (to be effective upon manual confirmation) Result writing Submit the confirmed draft to the event center. Final review and approval are required for the document to take effect. Official adjustment order (confirmed) Table 8 All the analysis or output results of the aforementioned intelligent agents are marked as "pending confirmation" in the system. They must be manually reviewed and confirmed before entering the event center. Administrators with appropriate permissions (such as staff from the organization department or personnel department) must review, correct, and finally confirm each element extracted by AI, the identified actions, the detected conflicts, and the generated drafts. It is strictly forbidden for AI analysis results to be directly and automatically written into the business database to avoid automatic database entry and effectiveness. Only results that have been manually confirmed to be correct will be submitted to the event center, thereby triggering subsequent processes such as twin simulation and on-chain evidence storage. In this process, AI plays an "advanced auxiliary" role, which greatly reduces the burden of manual review, extraction, and drafting, but the final decision-making and control rights are firmly in the hands of humans.

[0088] Pending Confirmation Mechanism: All output results are marked as "Pending Confirmation" in the system and must be manually reviewed and confirmed before entering the event center. The technical implementation of this mechanism is as follows: Database status field: A unified status field, ai_confirm_status, is set in the main table of the proposal (t_adjustment_proposal) and each sub-table (personnel, assets, business, system). The value range is {pending confirmation, approved, rejected, corrected}, and the default value is "pending confirmation". This field is decoupled from the business data field, only controlling the workflow permissions, and does not affect data integrity.

[0089] Workflow Engine Driven: The system employs an event-driven workflow engine (based on a state machine pattern), with the "Pending Confirmation" status serving as the starting node of the workflow. Only users with ROLE_ORG_ADMIN (Organization Department) or ROLE_HR_ADMIN (Personnel Department Officer) permissions can advance the status from "Pending Confirmation" to "Approved" or "Rejected." Status change triggers an event log written to the t_workflow_event table, recording the operator, operation time, status before and after the change, and a summary of the correction content.

[0090] Data isolation and prevention of accidental writes: AI output results are written to a separate temporary cache table (t_ai_draft_*), physically isolated from the formal business database. During manual confirmation, the system performs a difference comparison (diff), and the corrected data is written to the formal business table only after being electronically signed by the reviewer. It is strictly prohibited for AI parsing results to bypass the review process and be automatically stored in the database for effective implementation.

[0091] Event center access control: Only plans with `ai_confirm_status = 'approved'` can generate an event ID and be submitted to the event center (`t_event_center`), thereby triggering subsequent data changes, on-chain notarization, and other processes. The event center subscribes to the status change messages of `t_workflow_event` to achieve asynchronous decoupling.

[0092] Intelligent adjustment suggestion generation: This feature utilizes a large AI model to provide intelligent decision support for document processing. Key functions include impact range prediction, risk warning, and optimization suggestions, assisting decision-makers in formulating scientifically feasible implementation strategies. (See Table 9 for details.) Impact range prediction Quantitative adjustment of the impact area Number of departments affected, number of personnel involved, number of business process changes, and scope of impact across campuses. Confirm the completeness of the forecast range and adjust special impact points. Risk warning Identify potential hazards Risks include: personnel placement conflicts, business interruption, budget overruns, systemic conflicts, and legacy issues. Review risk level assessment and supplement hidden risk points Optimization suggestions Recommended execution strategy Step-by-step implementation plan, transition period setup, contingency plan, communication strategy, and timeline. Choose to adopt the suggestion, revise the implementation details, and confirm the final plan. Table 9 This model transforms the traditional document review process that relies on human experience into a dual-track decision-making model of "AI intelligent analysis + human review and confirmation," which improves the efficiency of document processing while ensuring the human controllability of key decisions, thus achieving a balance between efficiency and security.

[0093] This layer provides a visual application interface for different users, such as school leaders, functional departments, and college administrators: Twin simulation platform: Provides a drag-and-drop interface, allowing users to intuitively design organizational adjustment plans and visually view simulation animations, impact assessment maps, and plan comparison dashboards; Intelligent document processing platform: Provides a one-stop interface for document uploading, AI analysis result display, manual review and confirmation, and process tracking; Tag Application Center: Allows administrators to configure the tag system, define mapping rules, and view permission operation logs triggered by tag changes; Decision-Making Dashboard: Displays a comprehensive overview of the organizational structure, real-time status, early warning information, and key performance indicators in the form of a large data dashboard.

[0094] This layer provides the trust infrastructure for the entire system, including the consortium blockchain network architecture, six types of core smart contracts, and data on-chain strategies: Consortium Blockchain Network Architecture: A university organizational structure consortium blockchain is constructed using Hyperledger Fabric or FISCO BCOS frameworks. Key functional departments such as the university office, organization department, personnel department, academic affairs office, and finance department are invited as peer nodes in the network to jointly maintain the ledger, forming a multi-center co-governance network structure. The consensus mechanism adopts PBFT, adapting to the relatively trusted environment within the university and achieving efficient consensus. Fabric (Practical Byzantine Fault Tolerance) Channel technology is utilized to meet the efficient consensus needs of multi-departmental collaboration scenarios within the university. Privacy protection employs Channel technology to create independent privacy channels for data of different sensitivities (such as personnel appointments and dismissals, financial budgets), ensuring that sensitive data is only visible to relevant departments, achieving secure sharing of "data usable but not visible."

[0095] Core Smart Contracts: Deploy a set of core business logic contracts on the blockchain, see Table 10: OrganizationRegistry Record keeping and verification of registration, alteration and cancellation of organizations PositionRegistry Management of authority for creating, adjusting, and canceling positions PersonnelTagRegistry Issuance, renewal, and cancellation of personnel tags DocumentNotarization Document Preservation and Hash Verification AccessControl Tag-based cross-system permission auto-configuration AuditTrail Historical Inquiry into Organizational Structure Adjustments Table 10 All contracts operate collaboratively to ensure the coded execution of business rules and full traceability; six types of smart contracts are deployed to support business scenarios: OrganizationRegistry.sol: This contract is used to store and verify the registration, modification, and cancellation of organizations; when a new organization is established or its information changes, this contract is called to register / update and generate a unique on-chain identity. PositionRegistry.sol: This contract is used to manage the transfer of authority for the creation, adjustment, and cancellation of positions; it manages the creation, adjustment, and cancellation of positions and records the affiliation between positions and organizations; PersonnelTagRegistry.sol: This contract is responsible for the issuance, updating, and cancellation of personnel tags; when a personnel tag undergoes an authoritative change (such as through a formal appointment or dismissal document), this contract is invoked to issue or update the on-chain tag certificate as a trusted source; DocumentNotarization.sol: This contract is used to complete the entire process of document notarization and hash verification; it puts the document (or its main hash), the hash of key nodes in the approval flow, and the hash of the final effective adjustment draft on the chain for notarization, forming an immutable chain of evidence. AccessControl.sol: This contract is used to implement tag-based cross-system automatic permission configuration; it receives tag change events from the PersonnelTagRegistry contract and automatically generates standardized permission configuration instructions for downstream business systems (such as access control and OA) according to pre-defined on-chain rules; the downstream systems listen for and execute the instructions. AuditTrail.sol: This contract is used to provide historical traceability query services for organizational restructuring; it provides rich query interfaces, allowing authorized users (such as audit departments) to quickly retrieve and verify the full-process on-chain records of any organizational restructuring in history based on conditions such as organization ID, time range, and transaction type.

[0096] Data on-chain strategy: A hybrid model combining on-chain notarization and off-chain storage is adopted to balance trustworthiness and efficiency; on-chain storage includes data hashes, operation logs, approval records, and timestamps to ensure the immutability of critical information; off-chain storage includes complete organizational structure data and personnel information to ensure high-frequency query efficiency; and Merkle root hashes of off-chain data are periodically written to the blockchain through an anchoring mechanism to achieve batch data verifiability, reducing on-chain storage costs while maintaining the overall data's trustworthy traceability capabilities.

[0097] ( Example 2 This embodiment provides a construction method based on the system described in Embodiment 1, and the specific steps are as follows: S1. Construct digital twins of secondary institutions and their relationship networks. First, extract existing data from HRMS, OA, and other systems through the data acquisition layer interface. Then, based on the predefined ODT data model (containing five major categories of attributes: basic, spatial, personnel, business, and relationship), instantiate a digital twin for each institution and store it in the organization's master database at the data resource layer. Next, based on the actual relationships between institutions, construct four types of relationship edges using a graph database: 1) administrative hierarchy edges (superior-subordinate relationships); 2) business dependency edges (based on business process analysis); 3) resource allocation edges (based on budget and asset sharing data); 4) collaborative edges (based on historical joint projects and meeting data weighting). Finally, a complete institutional relationship network graph is formed.

[0098] S2. Construct and execute a digital twin path for institutional restructuring; this path is initiated upon receiving a restructuring request; users submit restructuring proposals through the application interface, and the system automatically generates a baseline; entering the twin pre-simulation phase, the system loads the relevant institutional ODT and relationship network according to the proposal type (e.g., "merging College A and Department B"), runs a personnel reassignment simulation model (considering professional matching and job level), an asset transfer model, and a business process integration model to simulate the new ODT status after the merger and a list of possible problems; subsequently, the impact assessment module calculates the impact of this merger on six dimensions of indicators, including the intensity of cooperation in relevant research projects and cross-campus management costs, based on the new ODT and the global relationship network. The system assesses the impact and outputs a risk assessment report. For high-risk scenarios such as "conflicts in core course scheduling," the optimization module recommends suggestions from the strategy library, such as "establishing a transitional teaching guidance committee," generating 2-3 optimized solutions for decision-makers to choose from. After selecting a solution, the decision-maker invokes a blockchain smart contract to store the core information (hash) and execution logic of the solution on the blockchain. Upon reaching the execution node, the smart contract automatically triggers, issuing instructions to the academic affairs system (adjusting course affiliations) and the financial system (merging budget items). Finally, at the end of the semester, actual teaching operation and financial data are collected, compared with pre-simulation predictions, and an effectiveness report is generated and used to optimize simulation model parameters.

[0099] S3. Construct a personnel tagging application model; First, synchronize the basic information and tags (titles, positions, etc.) of all personnel from the HR system to build a preliminary tag forest; then, business departments (such as the research institute and international office) add business tags (such as "major project leader" and "international exchange specialist") and skill tags (such as "proficient in Python") to personnel as needed; next, in the tag mapping library, the system administrator, together with the administrators of each business system, jointly define specific mapping rules; for example, define the rule: "When a person's tag 'graduate supervisor' is added, automatically create a supervisor account for him in the graduate management system and assign supervisor selection permissions"; when the HR department officially appoints an associate professor as a graduate supervisor and updates his tag in HRMS, this event will trigger the PersonnelTagRegistry smart contract to be uploaded to the chain through the interface, which will then trigger the AccessControl contract, generate permission instructions, and notify the graduate management system to execute.

[0100] S4. Construct an intelligent document analysis and adjustment model; When the Organization Department uploads an official document (PDF format) regarding the "establishment of an Artificial Intelligence Academy," the document intelligent parsing engine is activated; First, the document access agent identifies the document as a "notice of institutional establishment"; The text extraction agent extracts key fields such as the issuing authority, document number, and basis for establishment; The organizational action recognition agent accurately identifies the action of "establishment" and "Artificial Intelligence Academy"; The compliance verification agent compares this action with the institutional staffing regulations, and if it does not exceed the authorized staffing, it marks it as compliant; At the same time, it may find doubts about the qualifications of the proposed dean and highlight them in red; The draft generation agent integrates the above information and automatically generates a draft establishment plan including the academy's name, affiliation, preliminary staffing, and preparatory group suggestions; All these AI outputs are presented on the document processing platform's review interface, where Organization Department staff make final confirmations, modifications, and approvals; Only after a manual click of "confirm effective" will this draft be used as formal input and enter the institutional adjustment twin path described in S2 to begin subsequent processes such as rehearsal, evaluation, and on-chain processing.

[0101] S5. Construct a trusted blockchain foundation; deploy a Hyperledger Fabric consortium blockchain network in the university's information center; designate the university office, organization department, personnel department, academic affairs department, finance department, and audit department as founding nodes to jointly deploy chaincode; based on the six types of smart contracts designed in Example 1, developers write chaincode (such as in Go language) and install and instantiate it on all nodes; create independent channels for highly sensitive matters such as "cadre appointment and removal" and "organizational adjustments," with only the organization department, personnel department, and audit department nodes joining these channels to ensure data privacy; configure event listening services to enable applications in the digital twin layer to easily call contract interfaces for evidence storage and enable downstream business systems to listen to on-chain events to execute automatic instructions; ultimately, the entire system operates on the trusted, traceable, and automatically executed foundation provided by the blockchain.

[0102] The specific beneficial effects and advantages of this invention are as follows: 1) Enhanced Data Consistency and Trustworthiness: This invention addresses data silos through a unified digital twin model and ensures data immutability and multi-party consensus through blockchain notarization. Existing solutions rely on each business system independently maintaining organizational structure data, leading to severe data silos of "multiple versions of the same organization." This invention constructs a unified data source through an organizational digital twin (ODT). Addressing the unique complexity of universities—a coexistence of a three-tiered governance structure and intertwined organizational relationships such as subordinate / joint / affiliated entities—this invention standardizes and models all elements of the organization's information (basic attributes, spatial attributes, personnel attributes, business attributes, and relationship attributes) to form a dynamically updatable virtual organizational profile. Simultaneously, leveraging the distributed ledger and multi-party consensus mechanism of the blockchain platform, it ensures the trustworthiness of cross-system data synchronization, eliminating information silos. This advantage stems from the digital modeling of secondary organizations in Module 1 (including digital mapping of special organizational relationships) and the blockchain platform design in Module 5.

[0103] 2) Reduced Risk of Organizational Adjustment Decisions: Through twin simulation and impact assessment, the "simulate first, then execute" approach is achieved, quantifying decision risks. Existing solutions lack simulation capabilities, and direct physical execution of organizational adjustments incurs high trial-and-error costs. This invention utilizes a seven-step closed-loop process of digital twin path for organizational adjustments, particularly the twin simulation engine and impact assessment model, to simulate scenarios such as organizational mergers, splits, and cancellations in the digital space, quantifying the chain reaction impacts on personnel placement, budget transfers, and business continuity. Combined with AI-powered large-scale models to automatically identify contradictions with existing systems, it exposes personnel placement conflicts and business interruption risks in advance, reducing trial-and-error costs and improving the scientific nature of decision-making. This advantage stems from the adjustment simulation engine, impact assessment, and solution optimization technologies in Module Two.

[0104] 3) Optimization of Accountability and Audit Efficiency: End-to-end on-chain evidence storage provides a complete and reliable chain of evidence, enabling one-click traceability during audits. Existing solutions rely on paper documents or centralized databases, making data easily tampered with and difficult to trace. This invention, through an on-chain evidence storage mechanism, hashes key data such as adjustment proposals, pre-performance results, approval records, and execution instructions onto the chain, combining this with smart contracts to achieve automatic execution and multi-party endorsement. The AuditTrail contract provides a historical traceability query service for organizational structure adjustments. This achieves end-to-end traceability and improves audit efficiency, meeting the compliance requirements of inspection audits. This advantage stems from the blockchain evidence storage strategy and core smart contract design in Module Five.

[0105] 4) Improved cross-departmental collaboration efficiency: Through dynamic tag-permission mapping and smart contracts, cross-system coordination is transformed from "manual errands" to "automatic, second-level synchronization." Existing solutions rely on manual coordination, resulting in long implementation cycles for organizational adjustments. This invention achieves automated management and control through a tag-permission mapping mechanism (TBAC) that ensures "tags equal permissions, and changes take effect immediately." Addressing the complexity of personnel identities in universities—with parallel roles, titles, and ranks, separation of positions and personnel, and multiple roles or dual responsibilities—it transforms personnel tag changes from manual, system-by-system maintenance (which takes several days) to automatic, real-time synchronization (in seconds). By automatically executing smart contracts to replace manual coordination, routine documents (such as minor job adjustments and tag updates) are processed automatically, improving response speed and collaboration efficiency. This advantage stems from the tag-permission mapping mechanism in Module 3 (including support for special identity scenarios such as dual responsibilities) and the automatic execution technology of smart contracts in Module 5.

[0106] 5) Enhanced Data Security and Privacy Protection: The consortium blockchain's multi-center governance and channel technology reduce the risk of single-point failures and data leaks in centralized systems. Existing solutions use centralized storage, which carries the risk of single-point failures and data leaks. This invention adopts a hybrid model of on-chain evidence storage and off-chain storage. Key information is hashed and stored on-chain to ensure immutability, while complete data is stored off-chain to ensure query efficiency. The PBFT consensus mechanism adapts to collaborative scenarios involving multiple departments in universities. The risk of data tampering is significantly reduced, achieving secure sharing of data that is usable but not visible. This advantage stems from the consortium blockchain network architecture and data on-chain strategy in Module 5.

[0107] 6) Enhanced Intelligence and Scalability: The AI-powered large-scale model enhances the intelligence of document processing. The standardized digital twin model and interface support system are easily expandable and adaptable. Existing solutions rely on human experience and lack predictive analysis capabilities. This invention achieves intelligent processing of the entire process—uploading, importing, recognizing, extracting text, identifying organizational / position / personnel actions, detecting rule conflicts, and generating revised drafts—through an intelligent document parsing engine (based on an AI-powered large-scale model for classification, extraction, verification, and recognition). It strictly adheres to the "AI-assisted + human confirmation" mechanism to ensure the human controllability of key decisions. A closed loop is formed through effect monitoring and feedback optimization to continuously calibrate and simulate model parameters. Through standardized interface design and a universal digital twin system model, it supports flexible adaptation to universities of different sizes and types and lays the foundation for cross-university alliances.

[0108] In summary, the technical advantages of this invention stem from the deep integration of multiple technologies, including digital twin modeling (adapting to the complex organizational structure of universities), blockchain trusted storage, automatic execution of smart contracts, and AI intelligent analysis ("AI assistance + manual confirmation"), forming a new paradigm for university organizational structure governance characterized by "virtual-real mapping, trusted circulation, intelligent decision-making, and human controllability." This advantage originates from the document intelligent analysis and adjustment model (including human-machine collaboration mechanism) in Module 4 and the general modeling method in Module 1.

[0109] It should be understood that the above description is only a preferred embodiment of the present invention and is not sufficient to limit the technical solution of the present invention. For those skilled in the art, within the spirit and principles of the present invention, additions, subtractions, substitutions, transformations or improvements can be made based on the above description, and all such additions, subtractions, substitutions or improvements should fall within the protection scope of the appended claims of the present invention.

Claims

1. A blockchain-based digital twin system for university organizational structure, characterized in that, It includes, in sequence, a data acquisition layer, a data resource layer, a digital twin layer, an application service layer, and a blockchain foundation layer serving as a trusted base; among them, The data acquisition layer is used to connect with multiple heterogeneous business systems scattered within the university and to aggregate organizational structure-related data in real time. The data resource layer is used to standardize and store the aggregated data, and to build a unified data platform including an organizational master database, a relational graph library, a historical version library, and a tag mapping library. The digital twin layer is used to construct and maintain a virtual mapping model of all elements of the organizational structure based on the data from the data resource layer, and includes: The secondary organization twin model is used to create a digital twin of each secondary organization, which includes basic attributes, spatial attributes, personnel attributes, business attributes and relational attributes. Based on a graph database, a multi-dimensional relationship network is constructed to represent the administrative hierarchy, business dependencies, resource allocation and collaborative relationships between organizations. The digital twin path model for institutional restructuring is used to digitally simulate and manage the entire lifecycle of institutional restructuring, including seven closed-loop steps executed sequentially: restructuring proposal, twin simulation, impact assessment, scheme optimization, on-chain notarization, physical execution, and effect monitoring. The personnel tagging application model is used to build a multi-dimensional tag forest for personnel, including identity, position, job title, business, and skills, and to establish dynamic mapping rules between tags and downstream business system permissions; The document intelligent analysis and adjustment model is used to intelligently analyze, detect conflicts, and generate adjustment drafts for organizational adjustment documents based on a large AI model, and follows a dual-track decision-making mechanism that combines AI assistance and human confirmation. The application service layer is used to provide users with visual and interactive intelligent applications based on the model of the digital twin layer. The blockchain foundation layer is used to build a trusted execution environment based on consortium blockchain technology. By deploying and running a set of smart contracts, it provides trusted services to the digital twin layer and the application service layer, including institutional registration and notarization, document hash notarization, automatic permission configuration, and historical traceability auditing.

2. The blockchain-based digital twin system for university organizational structure as described in claim 1, characterized in that, In the aforementioned secondary-level organizational twin model, the organizational digital twin digitally defines and maps the three-tiered governance structure of the Party Committee system, the administrative system, and the deliberative body, as well as the special organizational relationships of subordinates, joint offices, and affiliations. The multi-dimensional relationship network transforms the static organizational structure into a computable relationship graph. Among these, subordinate relationships use a combination relationship to represent automatic inheritance of permissions, joint office relationships use independent classes and shared association classes to represent virtual shared spaces, and affiliation relationships use a combination relationship to represent dependency relationships and adopt a logically isolated data interaction mode.

3. The blockchain-based digital twin system for university organizational structure as described in claim 1, characterized in that, The organization adjusts the digital twin path model as follows: The proposed adjustment steps are used to transform the organization's intention to adjust its structure into a structured digital proposal. The twin simulation step is used to call a simulation model in the digital space to simulate the chain effects of different adjustment plans on personnel, assets, and business, and generate the predicted state after adjustment; The impact assessment step is used to calculate and adjust the six-dimensional impact indicators on personnel, finance, business, collaboration, system, and reputation based on the AI ​​big data model, and to classify the risks. The optimization step is used to match the identified risk points with an optimization strategy library, generate and compare multiple alternative optimization solutions. The on-chain evidence storage step is used to write the hash value of the key data of the entire process into the blockchain base layer; The physical execution step is used to decompose the confirmed optimization scheme into instructions and send them to each physical business system for execution. The effect monitoring step is used to collect the adjusted actual data and compare it with the pre-simulation predicted values ​​to calibrate the model parameters.

4. The blockchain-based digital twin system for university organizational structure as described in claim 1, characterized in that, The personnel tagging application model includes a tag-permission dynamic mapping mechanism. This mechanism predefines a set of mapping rules, and each rule includes triggering conditions, triggering actions, target systems, and rule types. When a person's tag status meets the triggering conditions, the corresponding triggering action is automatically executed to grant, modify, or revoke permissions in the target system.

5. The blockchain-based digital twin system for university organizational structure according to claim 1, characterized in that, The document intelligent analysis and adjustment model includes: The document intelligent parsing engine includes a document access intelligent agent, a text extraction intelligent agent, an organizational action recognition intelligent agent, a job action recognition intelligent agent, a personnel appointment and removal intelligent agent, a compliance verification intelligent agent, and a draft generation intelligent agent, which are used to sequentially complete document format recognition, text structuring, organizational / job / personnel adjustment action recognition, rule conflict detection, and draft generation; The human-machine collaboration mechanism stipulates that the analysis results of all AI agents are marked as pending confirmation and can only be converted into formal adjustment instructions after being reviewed and confirmed by humans.

6. The blockchain-based digital twin system for university organizational structure according to claim 1, characterized in that, The blockchain base layer uses Hyperledger Fabric or FISCO BCOS framework to build a consortium blockchain network, with participating nodes including key management departments within the university; the consensus mechanism adopts a practical Byzantine fault-tolerant algorithm; and data privacy is protected through channel technology.

7. The blockchain-based digital twin system for university organizational structure as described in claim 1 or 6, characterized in that, The smart contracts deployed in the blockchain foundation layer include at least: The OrganizationRegistry contract is used for the storage and verification of organization registration, changes, and cancellation. The PositionRegistry contract is used for the management of permissions for creating, adjusting, and revoking job positions. The PersonnelTagRegistry contract is used for the issuance, updating, and cancellation of personnel tags; The DocumentNotarization contract is used for hash-based notarization and verification of official documents throughout the entire process. AccessControl contracts are used for automatic cross-system permission configuration based on user tags; The AuditTrail contract is used to provide historical traceability and query of the entire process of organizational restructuring.

8. A method for constructing a digital twin of a university organizational structure based on blockchain, applied to the system described in any one of claims 1-7, characterized in that, The method includes: S1. Constructing digital twins of secondary organizations and their relationship networks: Aggregate data from various business systems, create a digital twin of each secondary organization containing multi-dimensional attributes, and use a graph database to construct an organizational relationship network graph representing different relationship types; S2. Construct and execute a digital twin path for organizational restructuring: In response to organizational restructuring needs, execute the following closed-loop process in sequence: generating restructuring proposals, conducting digital space simulation and pre-rehearsal, conducting multi-dimensional impact quantification assessment, optimizing intelligent solutions based on assessment results, hashing and storing key process data on the blockchain, issuing execution instructions to physical business systems, and monitoring actual effects and providing feedback to the optimization model. S3. Construct a personnel tagging application model: Create a multi-dimensional tag set for each person and configure dynamic mapping rules between tag status and downstream business system permissions; S4. Construct an intelligent document analysis and adjustment model: Use an AI big data model to automatically analyze and identify the input organizational adjustment documents, generate a preliminary adjustment draft and indicate conflict risks, and after manual review and confirmation, form a formal adjustment instruction; S5. Construct a trusted blockchain foundation: Deploy a consortium blockchain network and core smart contracts to provide an immutable and automated trusted execution environment for the registration of institutional entities, adjustment operation records, official documents, and permission changes generated in steps S1 to S4.

9. The method for constructing a digital twin of a university organizational structure based on blockchain according to claim 8, characterized in that, In step S2, the digital space simulation pre-play specifically includes: based on the adjustment type, calling the corresponding personnel flow simulation model, asset transfer simulation model and business continuity simulation model, simulating the execution process of the adjustment plan in the digital space, and outputting a multi-scenario simulation report containing the status of the adjusted organizational twin.

10. The method for constructing a digital twin of a university organizational structure based on blockchain according to claim 8, characterized in that, In step S4, the automated parsing and recognition of the input organizational adjustment documents using a large AI model is specifically executed sequentially by multiple interconnected AI agents, including: a document access agent for format parsing, a text extraction agent for text structuring, an organizational action recognition agent, a job action recognition agent, and a personnel appointment and removal agent for recognizing the corresponding entity adjustment actions, a compliance verification agent for rule conflict detection, and a draft generation agent for summarizing and generating a structured adjustment draft; and all the outputs of the agents are in a pending-effectiveness state before being manually confirmed.

Images