Intelligent resource configuration and organizational structure configuration system and method

Abstract

The invention discloses an intelligent resource configuration and organization structure configuration system and method, and belongs to the technical field of engineering project management. The system comprises an input and task structuring module, a multi-modal data perception and fusion module, an intelligent resource and organization knowledge retrieval module, a central scheduler, a professional resource configuration and organization reasoning module, a management decision generation module and an output and feedback learning module. According to the invention, the technical problems of low resource configuration efficiency, inaccurate man-post matching, frequent multi-project resource conflicts, rigid organization structure, fragmentation of a decision process and the like in traditional project management are solved.

Description

Technical Field

[0001] This invention relates to the field of engineering project management technology, and more specifically to an intelligent resource allocation and organizational structure configuration system and method. Background Technology

[0002] Current project management relies heavily on managers' personal experience in resource allocation and organizational structure design, resulting in the following technical shortcomings: Resource allocation relies on static quota standards, lacking the ability to dynamically respond to project-specific needs; the personnel-job matching mechanism is based on limited information and subjective judgment, making it difficult to achieve accurate matching of personnel capabilities with job requirements; resource conflict detection and coordination in multi-project environments rely on manual intervention, resulting in low efficiency and unstable decision-making quality; organizational structure design is mostly a one-time static setting, lacking the ability to dynamically adjust according to changes in project stages; decision-making links such as resource allocation, schedule planning, and organizational design are isolated from each other, lacking an integrated collaborative optimization mechanism.

[0003] While existing technologies attempt to improve these issues through information systems, most solutions remain at the level of business process automation, lacking deep intelligent decision-making capabilities. The few attempts to introduce artificial intelligence technology also focus primarily on optimizing single aspects, failing to build an intelligent decision-making system covering the entire project lifecycle.

[0004] Therefore, how to provide an intelligent resource allocation and organizational structure configuration system and method is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides an intelligent resource allocation and organizational structure configuration system and method to solve technical problems in traditional project management, such as low resource allocation efficiency, inaccurate matching of personnel and positions, frequent resource conflicts among multiple projects, rigid organizational structure, and fragmented decision-making process.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: An intelligent resource allocation and organizational structure configuration system, comprising: The input and task structuring module is used to receive project management goals input by the user and convert them into structured task instructions; The multimodal data perception and fusion module, connected to the input and task structuring module, is used to receive and process multi-source heterogeneous data related to project management, extract key information and construct a dynamically evolving project context through a preset model. The intelligent resource and organizational knowledge retrieval module is connected to the input and task structuring module and the multimodal data perception and fusion module, respectively, and is used to retrieve relevant information from the enterprise knowledge base based on the structured task instructions and the project context to form a decision support knowledge package; The central scheduler is connected to the input and task structuring module and the intelligent resource and organization knowledge retrieval module, respectively, and is used to receive the structured task instructions and the decision support knowledge package, and to perform task decomposition and scheduling. The professional resource allocation and organization reasoning module, connected to the central scheduler, includes multiple artificial intelligence sub-models for executing sub-tasks assigned by the central scheduler; The management decision generation module is connected to the central scheduler and the professional resource allocation and organization reasoning module, and is used to integrate the output results of each artificial intelligence sub-model and the decision support knowledge package to generate a comprehensive management decision document; The output and feedback learning module is connected to the management decision generation module and is used to output the management decision document and optimize each artificial intelligence sub-model and the enterprise knowledge base based on user feedback data.

[0007] Furthermore, the multimodal data sensing and fusion module includes: The document parsing unit, HR data interface unit, and historical data extraction unit are used to extract structured information from design documents, HR systems, and historical project databases, respectively.

[0008] Furthermore, the central scheduler includes: Task decomposition unit, dynamic scheduling unit, dependency management unit, and format standardization unit; It is used to decompose project management tasks into atomic sub-tasks and dynamically allocate them to the corresponding artificial intelligence sub-models. It is also used to manage the input and output data dependencies between artificial intelligence sub-models and standardize the data interaction format between modules.

[0009] Furthermore, the artificial intelligence sub-models include: a personnel-job matching sub-model, a multi-project resource conflict detection and resolution sub-model, a dynamic resource optimization sub-model, an organizational structure generation sub-model, a cost-schedule coupling analysis sub-model, and a construction safety risk identification sub-model.

[0010] Furthermore, the personnel-job matching sub-model is implemented using one or more methods among the Hungarian algorithm, deep recommendation networks, or knowledge graph reasoning.

[0011] Furthermore, the intelligent resource and organizational knowledge retrieval module employs vector retrieval technology for knowledge matching.

[0012] Furthermore, the management decision generation module includes: The solution document generation unit, the multi-format output unit, and the consistency verification unit are all included.

[0013] Furthermore, the output and feedback learning module includes: User feedback collection unit, model fine-tuning unit, and knowledge base update unit; It is used to fine-tune the models within the system, calibrate parameters, and update the knowledge base based on user feedback.

[0014] A method for intelligent resource allocation and organizational structure configuration, applied to an intelligent resource allocation and organizational structure configuration system, includes the following steps: S1. Receive user project management goals and convert them into structured task instructions; S2. Parse multi-source data to build dynamic project context; S3. Retrieve relevant knowledge bases based on task instructions and project context to form a decision support knowledge package; S4. Intelligently decompose and schedule the overall task to generate a sub-task queue; S5. Call the corresponding professional artificial intelligence sub-model to execute the sub-task; S6. Integrate the output results and knowledge packages of all sub-models to generate the final management decision document; S7. Output management decision documents and collect user feedback data to optimize the system model and knowledge base.

[0015] As can be seen from the above technical solution, compared with the prior art, the present invention provides an intelligent resource allocation and organizational structure configuration system and method, which has the following beneficial effects: It has enabled the intelligent and automated decision-making of project management, transforming the original management decision-making process that relied on human experience into a data-driven and model-enabled intelligent process, which has significantly improved the efficiency and scientific nature of decision-making. It achieves precise matching and dynamic optimization of resources and organization, realizes the optimal matching of people, positions, tasks and resources through advanced algorithm models, and can make dynamic adjustments according to the actual progress of the project to ensure maximum resource utilization efficiency. It has built an enterprise-level resource collaboration and conflict resolution capability. Through a multi-project resource conflict detection and resolution sub-model, it enables the visibility, management and control of resources at the enterprise-wide level, effectively solving the problem of resource conflict. This forms a closed-loop management system covering the entire project lifecycle, with system capabilities spanning the entire process from project initiation, planning, execution, monitoring, and closure, providing integrated intelligent decision support. By accumulating and revitalizing organizational knowledge assets, and solidifying continuously accumulated excellent management practices into corporate digital assets through feedback learning mechanisms, a virtuous cycle of knowledge accumulation and iterative optimization can be formed. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 The overall system architecture diagram provided by this invention; Figure 2 A flowchart illustrating the method provided by the present invention; Figure 3 This is a schematic diagram illustrating the working process of the personnel-job matching sub-model provided by the present invention; Figure 4 This is a schematic diagram of construction resource allocation provided by the present invention; Figure 5 An example diagram of the organizational structure provided by this invention is generated. Detailed Implementation

[0018] 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 only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1: See Figure 1 This invention discloses an intelligent resource allocation and organizational structure configuration system, comprising: The input and task structuring module is used to receive project management goals input by the user and convert them into structured task instructions; The multimodal data perception and fusion module, connected to the input and task structuring module, is used to receive and process multi-source heterogeneous data related to project management, extract key information and construct a dynamically evolving project context through a preset model. The intelligent resource and organizational knowledge retrieval module is connected to the input and task structuring module and the multimodal data perception and fusion module, respectively, and is used to retrieve relevant information from the enterprise knowledge base based on the structured task instructions and the project context to form a decision support knowledge package; The central scheduler is connected to the input and task structuring module and the intelligent resource and organization knowledge retrieval module, respectively, and is used to receive the structured task instructions and the decision support knowledge package, and to perform task decomposition and scheduling. The professional resource allocation and organization reasoning module, connected to the central scheduler, includes multiple artificial intelligence sub-models for executing sub-tasks assigned by the central scheduler; The management decision generation module is connected to the central scheduler and the professional resource allocation and organization reasoning module, and is used to integrate the output results of each artificial intelligence sub-model and the decision support knowledge package to generate a comprehensive management decision document; The output and feedback learning module is connected to the management decision generation module and is used to output the management decision document and optimize each artificial intelligence sub-model and the enterprise knowledge base based on user feedback data.

[0020] Specifically, the intelligent resource allocation and organizational structure configuration system of this invention includes seven core modules. Users input project management goals and parameters through the user interface, which are then converted into structured task instructions by the input and task structuring module. The multimodal data perception and fusion module processes multi-source information such as project design documents and real-time resource status data to construct a dynamic project context. The intelligent resource and organizational knowledge retrieval module retrieves relevant information from the enterprise knowledge base based on retrieval enhancement generation technology, forming a decision support knowledge package.

[0021] Specifically, the central scheduler performs the following functions: decomposes complex project management tasks into atomic subtasks that can be executed in parallel or serially; dynamically assigns each subtask to the corresponding professional sub-model; manages the input and output data dependencies between professional sub-models; and standardizes the data interaction format between all modules.

[0022] Specifically, the central scheduler receives task instructions and knowledge packages, performs task parsing and intelligent decomposition, and dynamically allocates sub-tasks to the corresponding sub-models in the professional resource allocation and organizational reasoning module. This module includes multiple professional artificial intelligence sub-models such as personnel-job matching, resource conflict detection, and organizational structure generation, with each sub-model executing professional computing and reasoning tasks in parallel.

[0023] The management decision generation module integrates the outputs of various sub-models to generate comprehensive management decision documents, including construction plans, schedules, resource plans, and organizational charts. The output and feedback learning module outputs these decision documents to users and continuously optimizes the system model and knowledge base based on user feedback.

[0024] Specifically, the multimodal data perception and fusion module includes a document parsing unit, an HR data interface unit, and a historical data extraction unit, which are used to extract structured information from design documents, HR systems, and historical project databases, respectively.

[0025] Specifically, the document parsing submodule (document parsing unit) is used to extract node details, text, charts and structured parameters from the design document through object detection, OCR technology and natural language processing models; The Human Resources Data Interface Submodule (HR Data Interface Unit) is used to obtain real-time data on personnel skills lists, work experience, performance evaluations, on-the-job status, and project preferences from the enterprise's human resources management system. The project historical data extraction submodule (historical data extraction unit) is used to extract organizational structure templates, resource allocation patterns, project execution records, and cost performance data of similar projects from the enterprise's historical project database.

[0026] Specifically, the central scheduler includes: Task decomposition unit, dynamic scheduling unit, dependency management unit, and format standardization unit; It is used to decompose project management tasks into atomic sub-tasks and dynamically allocate them to the corresponding artificial intelligence sub-models. It is also used to manage the input and output data dependencies between artificial intelligence sub-models and standardize the data interaction format between modules.

[0027] Specifically, the artificial intelligence sub-models include: personnel-job matching sub-model, multi-project resource conflict detection and resolution sub-model, dynamic resource optimization sub-model, organizational structure generation sub-model, cost-schedule coupling analysis sub-model, and construction safety risk identification sub-model.

[0028] Specifically, the personnel-job matching sub-model is used to make optimal assignments based on project requirements, job responsibilities and personnel capabilities in multiple dimensions. like Figure 3 As shown, the personnel-job matching sub-model generates dynamic job requirements based on the project overview. Combining personnel information and enterprise rule constraints, it outputs accurate job-person matching results through context-aware feature vectorization and multi-algorithm fusion matching, and continuously optimizes the matching model through a feedback mechanism.

[0029] Specifically, the personnel-job matching sub-models described herein employ one or more of the following methods to achieve accurate matching: Based on the precise matching of the Hungarian algorithm, a weight matrix is ​​constructed with job requirements and personnel capabilities as dimensions, and the optimal allocation is carried out with the goal of maximizing the overall matching score of the project. Semantic matching based on deep cross-networks involves high-quality text vectorization of project requirements, job descriptions, and personnel information, inputting them into a deep network to learn complex interaction relationships, and outputting matching probabilities. Based on knowledge graph-based associative reasoning, multi-hop reasoning is performed using a triplet relationship network consisting of job position, skill, personnel, and project history to recommend candidate personnel and generate explanatory reasons.

[0030] A multi-project resource conflict detection and resolution sub-model is used to identify key resource conflicts from an enterprise-wide perspective and provide intelligent solutions. The specific implementation of this sub-model includes the following technical steps: 1. Unified Resource Pool Modeling and Status Tracking: Establish an enterprise-level multi-dimensional resource pool data model, covering human resources, equipment resources, material resources, and financial resources. Continuously acquire and update resource planning and actual occupancy data for all projects under construction and planned.

[0031] 2. Conflict Detection (Temporal and Capacity Conflict Identification): Timing conflict detection: Based on the schedule of each project, the demand time windows of key resources are extracted. A timeline overlap detection algorithm is used to identify conflicts where the same resource is occupied by multiple projects within overlapping time periods.

[0032] Capacity conflict detection: For resources that can be shared but have a limited total amount, aggregate the total demand of each project within the same time period and compare it with the total available amount in the enterprise resource pool to identify over-limit conflicts.

[0033] 3. Conflict resolution and solution generation: Rule base and strategy engine: Built-in priority rules, optimization goals (such as minimizing overall delays and minimizing resource idle costs) and a solution strategy base (such as resource smoothing, resource balancing, and resource-constrained schedules).

[0034] Intelligent conflict resolution algorithm: For detected conflicts, the model calls a genetic algorithm to automatically solve them. The solution process aims to minimize overall project delays or total costs. Under the premise of satisfying the critical path constraints of each project, it rearranges the resource usage plan for non-critical tasks or adjusts the allocation ratio of resources among different projects, generating one or more feasible conflict resolution solutions.

[0035] Solution evaluation and recommendation: Simulate and analyze the generated solutions to assess their impact on the project schedule and cost, provide a comparison of solutions and reasons for recommendation, and finally output a conflict resolution report including adjustment suggestions.

[0036] The dynamic resource optimization sub-model is used to adjust and reallocate resource allocation schemes for future periods in real time based on actual project progress data. The specific implementation of this sub-model is based on the principles of rolling planning and feedback control, and the technical steps are as follows: 1. Schedule Variance Analysis and Impact Assessment: Input actual project schedule data (such as completed work and actual duration), perform earned value analysis against the baseline plan, and calculate schedule and cost variances. Use the critical path method to recalculate the earliest / latest start time of the remaining work, identify new critical paths and float time changes.

[0037] 2. Re-forecasting of future resource demands: Based on the revised Work Breakdown Structure (WBS) and schedule network diagram, combined with historical work efficiency data or work efficiency trends predicted by machine learning, the demand for various types of resources (labor, machinery, materials) for each time period in the future (such as day, week) is recalculated.

[0038] 3. Multi-objective optimization solution: Establish an optimization model: With the objective function of minimizing total project delay, minimizing resource adjustment costs, or maximizing resource utilization balance, and with constraints such as the process logic relationship of remaining work, resource availability, and resource intensity limits, construct an integer programming or mixed integer linear programming model.

[0039] Model Solving: The model is solved using an exact algorithm or a heuristic algorithm to obtain the optimal or near-optimal solution for resource allocation in the future time period. This may include: increasing / decreasing resource input, changing resource types, adjusting task execution order (within the logically permissible range), etc.

[0040] 4. Output dynamic adjustment plan: Generate resource load diagram, resource histogram and detailed resource adjustment instruction list from the solution results, which will serve as the basis for resource scheduling in the next stage.

[0041] An organizational structure generation sub-model is used to generate a flexible organizational structure that dynamically adapts to project type, scale, and phase characteristics. This sub-model adopts a layered construction strategy, and its specific technical implementation steps are as follows: 1. Layered architecture construction: Top-level management structure generation: Based on input project characteristics (such as type, scale, and investment amount), the system matches and calls the corresponding standard management structure template from the enterprise's predefined organizational structure template library. This template typically includes fixed management roles at the company / project department level (such as project manager, chief engineer, safety director, engineering management department, etc.) and their hierarchical reporting relationships. The system scales or reduces the number of staff in the departments within the template proportionally according to project size parameters (such as large, medium, and small).

[0042] Generation of underlying work team structure: This part is the core of dynamic generation.

[0043] Construction content analysis: This sub-model receives construction plans generated by other sub-modules coordinated by the central scheduler, and uses natural language processing technology to automatically identify all specific construction operations from the plans.

[0044] Work Team Knowledge Base Matching: The system maintains a work team knowledge base, which defines the typical personnel composition (e.g., team leader, skilled workers, general workers), skill requirements, and equipment configuration for various standard work teams (e.g., "reinforcing steel work team," "formwork support team," "concrete pouring team," "mechanical and electrical installation team"). Based on the parsed construction content, the model matches and instantiates the required work team from this knowledge base.

[0045] Subordination Establishment: Each instantiated work team is automatically assigned to the corresponding functional department or directly subordinate work team / work area in the top-level management structure according to its work nature and professional field, thus establishing a clear management and reporting link.

[0046] 2. Architecture integration and optimization: The dynamically generated underlying work team structure is seamlessly integrated with the fixed top-level management structure to form a complete project organizational structure tree.

[0047] Calculate the management span, management levels and other indicators of the integrated architecture, and make fine-tuning and optimization according to the preset organizational effectiveness rules (for example, if a department has too many subordinate teams, it is recommended to add a work area management team).

[0048] 3. Visual output: Data structuring: Transform the final organizational structure tree into a node-edge graph data structure. Each node contains attributes, and each edge represents a reporting or subordinate relationship.

[0049] Automatic layout and drawing: A breadth-first search algorithm is used to traverse the architecture tree to determine the level (y-coordinate) of each node on the canvas and its order within the same level (x-coordinate). The traversal starts from the root node and visits child nodes level by level, ensuring that the management hierarchy is visually clear and aligned.

[0050] Based on the calculated node positions, matplotlib or a similar graphics library is used to automatically plot the data. The plotting process includes: drawing node boxes, adding job description text, and drawing connecting lines.

[0051] Multi-format output: Finally, a high-resolution organizational chart (such as PNG and SVG formats) is generated, and a structured data file (such as JSON) describing the details of the architecture is output simultaneously for direct use by downstream systems or documents.

[0052] A cost-schedule coupling analysis sub-model is used for earned value management analysis and cost-to-completion prediction. The specific implementation of this sub-model is based on the earned value management system and a statistical prediction model, and the technical steps are as follows: 1. Real-time calculation of earned value indicators: Integrates actual project cost (AC) and actual progress (percentage of completion), compares them with planned value (PV) and earned value (EV), and automatically calculates and monitors core indicators such as cost variance (CV=EV-AC), schedule variance (SV=EV-PV), cost performance index (CPI=EV / AC), and schedule performance index (SPI=EV / PV).

[0053] 2. Estimate of Completion (EAC) Forecast: Model library: Integrates multiple EAC prediction formula models, including: Typical Deviation Model: EAC = BAC / CPI (assuming future performance is consistent with current CPI) Atypical Deviation Model: EAC = AC + (BAC - EV) (assuming the remaining work is completed on budget) Composite model: EAC = AC + (BAC - EV) / (CPI) SPI (While considering both cost and schedule performance) Machine learning prediction: In addition to the formula model, this sub-model is also trained with time series prediction models (such as ARIMA) or regression models, which take historical CPI, SPI series, project features, etc. as inputs to directly predict the final completion cost.

[0054] 3. Trend Analysis and Early Warning: Perform time series analysis on indicators such as CPI and SPI to identify their changing trends. Set thresholds (e.g., CPI < 0.9). When an indicator exceeds the threshold or shows a continuous deteriorating trend, an early warning signal is automatically triggered and linked to the dynamic resource optimization sub-model to provide a basis for adjustment decisions.

[0055] 4. Generate analysis reports: Automatically generate earned value analysis charts (such as S-curve charts) and reports, clearly showing the current cost / schedule health of the project, and providing completion cost forecast ranges and completion date forecasts based on multiple methods.

[0056] A construction safety risk identification sub-model is used to identify potential safety and quality risks by associating resource allocation schemes. The specific implementation of this sub-model is based on knowledge graphs and rule-based reasoning, and the technical steps are as follows: 1. Risk Knowledge Graph Construction: Construct a construction safety risk knowledge graph, whose nodes include: hazard sources (e.g., working at heights, lifting and hoisting), accident types (e.g., falls, impacts), unsafe conditions (e.g., lack of protection), unsafe behaviors (e.g., violations of operating procedures), causative factors (e.g., night work, fatigue work), and prevention and control measures (e.g., setting up safety nets, safety education). Nodes are connected by edges representing relationships such as "cause," "belongs to," and "need."

[0057] 2. Resource allocation scheme analysis and mapping: Analyze the resource allocation scheme and extract key elements: process / activity, resource type and quantity, environmental conditions, and process method.

[0058] 3. Risk Reasoning and Identification: Rule matching: Match the parsed elements with predefined "risk-configuration" association rules. Rules are in the IF-THEN format, for example: IF process includes "high-altitude operation" AND resources include "new workers" AND training time < specified value THEN risk level increases; IF environmental conditions include "nighttime" AND process includes "heavy equipment cross-operation" THEN risk level increases.

[0059] Graph Traversal Query: Starting with the parsed elements, a traversal query is performed within the risk knowledge graph. For example, starting from the "nighttime construction" node, the "visual impairment" node is found through the "aggravation" relationship, and then the "collision accident" node is found through the "cause" relationship, thereby identifying the potential risk chain that "nighttime construction may increase the risk of collision accidents".

[0060] 4. Risk Assessment and Output: For each identified potential risk, a risk level assessment (such as the LEC method or matrix method) is conducted, considering its probability of occurrence (based on historical statistical data or expert experience database) and the severity of its potential consequences. The final output is a risk list, specifying the identified risk points, risk levels, associated resource allocation items, and recommended preventative and control measures.

[0061] The central scheduler is the core coordination hub of the system, and its internal logic and workflow are implemented as follows: 1. Task analysis and atomization decomposition: Natural Language Understanding: Utilizes integrated large language models (LLM) to deeply understand structured task instructions, identifying core intentions, constraints, and implicit requirements.

[0062] Task Tree Generation: Based on a pre-defined project management task ontology, the overall task is decomposed into a task tree from top to bottom. The ontology defines task types, input and output data types, and logical relationships between tasks. For example, the overall task mentioned above may be decomposed into atomic or composite sub-tasks such as personnel-job matching, construction machinery configuration, schedule planning, cost estimation, and safety risk assessment.

[0063] 2. Dependency Analysis and Scheduling Graph Construction: Analyze the data dependencies (task B requires the output of task A as input) and logical dependencies (task C can only start after task D is completed) between the subtasks in the task tree. Based on this, construct a dynamic scheduling graph in the form of a directed acyclic graph (DAG).

[0064] 3. Dynamic scheduling and resource allocation: Sub-model registration and capability description: Each professional AI sub-model in the system registers with the central scheduler at startup and reports its capability description (such as the types of tasks it can handle, input format, output format, and current load status).

[0065] Task-model matching: For subtasks in the ready state (all their prerequisites have been satisfied) in the scheduling graph, the central scheduler matches a list of sub-models with corresponding processing capabilities from the registry based on the subtask type.

[0066] Load balancing and priority scheduling: Combining the current load of the sub-model (e.g., the number of tasks being processed) and the priority of the sub-tasks (determined by the urgency of the project or the criticality of the task), a load balancing algorithm (e.g., least-task-first) or a priority queue algorithm is used to dynamically allocate sub-tasks to the most suitable sub-model instances for execution. It supports assigning the same sub-task to multiple models for comparative analysis (e.g., using both the Hungarian algorithm and a deep learning model for personnel matching).

[0067] 4. Data flow management and format standardization: The central scheduler maintains a global data context to temporarily store the input and output data of each subtask. It is responsible for passing the output data of upstream tasks (after standardization) to downstream tasks as input according to the dependencies in the scheduling graph.

[0068] By using a predefined unified data exchange model, the accuracy and consistency of data interaction between different sub-models are ensured.

[0069] 5. Execution Monitoring and Exception Handling: Monitor the execution status of each subtask (waiting, executing, successful, failed). For subtasks that fail, retry according to predefined strategies or switch to a backup model, and log the results for system optimization.

[0070] like Figure 4 As shown, the construction resource allocation is based on the project overview and decomposes the work processes. Through mechanical combination configuration, quota matching and schedule resource calculation, the resource allocation scheme of each work process is output, and the configuration model is continuously improved through feedback optimization mechanism.

[0071] like Figure 5 As shown, the organizational structure generation sub-model dynamically generates a multi-level organizational structure, including management, functional departments, and operational teams, based on project characteristics, achieving the best match between organizational design and project requirements.

[0072] Specifically, the intelligent resource and organizational knowledge retrieval module uses vector retrieval technology to convert the key features and dynamic project context in the structured task instructions into high-dimensional vectors, and performs similarity matching in each vectorized knowledge base.

[0073] Specifically, the intelligent resource and organizational knowledge retrieval module performs vector retrieval, which converts the key features and dynamic project context in the structured task instructions into high-dimensional vectors, and performs similarity matching in various vectorized knowledge bases to retrieve the most relevant systems, cases, personnel and normative fragments.

[0074] Specifically, the management decision generation module receives the output results of each sub-model collected by the central scheduler and the decision support knowledge package. The large language model integrates the output results and the knowledge package content to generate a comprehensive management decision document, including construction-specific plans, project schedules, resource requirement plans, organizational charts, cost budget reports, and risk assessment reports. The solution document generation unit, the multi-format output unit, and the consistency verification unit are all included.

[0075] Specifically, the output and feedback learning module includes: User feedback collection unit, model fine-tuning unit, and knowledge base update unit; It is used to fine-tune the models within the system, calibrate parameters, and update the knowledge base based on user feedback.

[0076] When generating the final management decision document through the management decision generation module, the output supports multiple editable formats and automatically maintains the consistency of data, charts, and text descriptions in the document content.

[0077] Specifically, the output and feedback learning module outputs comprehensive management decision documents to users and receives feedback data from users in the actual management process. Based on this feedback data, it fine-tunes and calibrates the parameters of the large language model and professional sub-models in the system, and enriches and updates each knowledge base to achieve closed-loop learning and continuous evolution of the system. Specifically, the output and feedback learning module optimizes the system by: using user-adjusted organizational structure and resource plan data to supervise and fine-tune relevant sub-models; using user-confirmed calculation results and decision data to calibrate the parameters of analysis sub-models; and storing user-confirmed excellent management decision solutions into the knowledge base.

[0078] like Figure 2 As shown, the method of this invention includes seven core steps: input and task structuring, multimodal data perception and fusion, intelligent knowledge retrieval, task decomposition and scheduling, professional model reasoning, management decision generation, and output and feedback learning. Each step is executed sequentially to form a complete intelligent decision-making process.

[0079] Specifically, an intelligent resource allocation and organizational structure configuration method, applied to an intelligent resource allocation and organizational structure configuration system, includes the following steps: S1. Receive user project management goals and convert them into structured task instructions; S2. Parse multi-source data to build dynamic project context; S3. Retrieve relevant knowledge bases based on task instructions and project context to form a decision support knowledge package; S4. Intelligently decompose and schedule the overall task to generate a sub-task queue; S5. Call the corresponding professional artificial intelligence sub-model to execute the sub-task; S6. Integrate the output results and knowledge packages of all sub-models to generate the final management decision document; S7. Output management decision documents and collect user feedback data to optimize the system model and knowledge base.

[0080] Example 2: Application in railway bridge pier construction projects Users select the "Railway Bridge Abutment Construction" project type through the system interface and input parameters such as project scale, target construction period, and quality requirements. The system then generates a structured task instruction in JSON format containing complete technical parameter requirements through the input and task structuring module.

[0081] The multimodal data perception and fusion module receives user-uploaded foundation design drawings, geological survey reports, and design specifications. The document parsing submodule extracts text annotations using OCR technology, identifies structural elements using an object detection model, and parses process requirements using natural language processing. The human resources data interface submodule retrieves currently available personnel information from the company's HR system. The project historical data extraction submodule searches for historical data from similar projects. All information is fused using a multimodal big data model to generate a dynamic project context.

[0082] The intelligent resource and organizational knowledge retrieval module retrieves relevant information from construction specification knowledge bases, organizational template libraries, enterprise construction method libraries, and human resource libraries based on task instructions and project context. A large language model integrates the retrieval results to form a decision support knowledge package.

[0083] The central scheduler receives task instructions and knowledge packages, parses and decomposes them into multiple sub-tasks: personnel-job matching, organizational structure generation, resource conflict detection, and schedule-resource coupling optimization. The central scheduler distributes these sub-tasks in parallel to the corresponding sub-models in the professional resource allocation and organizational reasoning module.

[0084] The various professional sub-models are executed in parallel: the personnel-job matching sub-model uses the Hungarian algorithm to calculate the matching degree between each job and the candidate; the organizational structure generation sub-model outputs a three-level organizational structure chart; the resource conflict detection sub-model detects critical resource conflicts; and the schedule-resource coupling optimization sub-model calculates the optimal schedule and resource requirements based on the critical path method.

[0085] The large language model of the management decision generation module receives all intermediate results, generates a draft framework of the construction plan covering the standard chapters, integrates the outputs of each sub-model into the corresponding chapters, performs consistency verification and language polishing, and finally generates a standardized special construction plan and resource allocation plan document.

[0086] The output and feedback learning module presents the generated solution to the user. After review, the user can modify some parts, and the system records the user feedback, which is used for model fine-tuning and knowledge base updates.

[0087] Example 3: Dynamic resource adjustment application During the project execution phase, when the actual progress deviates from the target schedule, users input the updated progress data into the system and set the management goal as developing a plan to catch up on the project schedule.

[0088] The dynamic resource optimization sub-model is activated, recalculating the resource requirements of subsequent processes based on the latest schedule context. The model generates multiple catch-up scenarios, including aggressive, balanced, and conservative schemes, each corresponding to different degrees of schedule compression and cost increases.

[0089] The cost-schedule coupling analysis sub-model provides earned value analysis predictions for each option, while the safety risk identification sub-model assesses the safety impact of each option. The system comprehensively considers schedule, cost, and safety factors to recommend the optimal solution and provide detailed implementation suggestions.

[0090] The project team adopted the system's recommended solution and effectively controlled project delays and cost overruns through resource adjustments and process optimization.

[0091] Example 4: Multi-project resource collaborative optimization In a multi-project parallel environment within an enterprise, the system achieves global resource monitoring through an enterprise-level resource pool. The multi-project resource conflict detection and resolution sub-model tracks the resource requirements of all projects under construction in real time, predicts resource conflicts, and provides intelligent allocation solutions.

[0092] The system dynamically adjusts resource allocation priorities based on the strategic importance and urgency of projects. Through strategies such as resource smoothing, resource balancing, and activating backup resources, it effectively resolves resource conflicts and improves the overall resource utilization efficiency of the enterprise.

[0093] The system generates resource allocation suggestion reports to provide decision support for management and achieve optimal resource allocation in a multi-project environment.

[0094] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0095] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An intelligent resource allocation and organizational structure configuration system, characterized in that, include: The input and task structuring module is used to receive project management goals input by the user and convert them into structured task instructions; The multimodal data perception and fusion module, connected to the input and task structuring module, is used to receive and process multi-source heterogeneous data related to project management, extract key information and construct a dynamically evolving project context through a preset model. The intelligent resource and organizational knowledge retrieval module is connected to the input and task structuring module and the multimodal data perception and fusion module, respectively, and is used to retrieve relevant information from the enterprise knowledge base based on the structured task instructions and the project context to form a decision support knowledge package; The central scheduler is connected to the input and task structuring module and the intelligent resource and organization knowledge retrieval module, respectively, and is used to receive the structured task instructions and the decision support knowledge package, and to perform task decomposition and scheduling. The professional resource allocation and organization reasoning module, connected to the central scheduler, includes multiple artificial intelligence sub-models for executing sub-tasks assigned by the central scheduler; The management decision generation module is connected to the central scheduler and the professional resource allocation and organization reasoning module, and is used to integrate the output results of each artificial intelligence sub-model and the decision support knowledge package to generate a comprehensive management decision document; The output and feedback learning module is connected to the management decision generation module and is used to output the management decision document and optimize each artificial intelligence sub-model and the enterprise knowledge base based on user feedback data.

2. The intelligent resource allocation and organizational structure configuration system according to claim 1, characterized in that, The multimodal data sensing and fusion module includes: The document parsing unit, HR data interface unit, and historical data extraction unit are used to extract structured information from design documents, HR systems, and historical project databases, respectively.

3. The intelligent resource allocation and organizational structure configuration system according to claim 1, characterized in that, The central scheduler includes: Task decomposition unit, dynamic scheduling unit, dependency management unit, and format standardization unit; It is used to decompose project management tasks into atomic sub-tasks and dynamically allocate them to the corresponding artificial intelligence sub-models. It is also used to manage the input and output data dependencies between artificial intelligence sub-models and standardize the data interaction format between modules.

4. The intelligent resource allocation and organizational structure configuration system according to claim 1, characterized in that, The artificial intelligence sub-models include: personnel-job matching sub-model, multi-project resource conflict detection and resolution sub-model, dynamic resource optimization sub-model, organizational structure generation sub-model, cost-schedule coupling analysis sub-model, and construction safety risk identification sub-model.

5. The intelligent resource allocation and organizational structure configuration system according to claim 4, characterized in that, The personnel-job matching sub-model is implemented using one or more methods, such as the Hungarian algorithm, deep recommendation networks, or knowledge graph reasoning.

6. The intelligent resource allocation and organizational structure configuration system according to claim 1, characterized in that, The intelligent resource and organizational knowledge retrieval module uses vector retrieval technology for knowledge matching.

7. The intelligent resource allocation and organizational structure configuration system according to claim 1, characterized in that, The management decision generation module includes: The solution document generation unit, the multi-format output unit, and the consistency verification unit are all included.

8. The intelligent resource allocation and organizational structure configuration system according to claim 1, characterized in that, The output and feedback learning module includes: User feedback collection unit, model fine-tuning unit, and knowledge base update unit; It is used to fine-tune the models within the system, calibrate parameters, and update the knowledge base based on user feedback.

9. A method for intelligent resource allocation and organizational structure configuration, characterized in that, The intelligent resource allocation and organizational structure configuration system according to any one of claims 1-8 includes the following steps: S1. Receive user project management goals and convert them into structured task instructions; S2. Parse multi-source data to build dynamic project context; S3. Retrieve relevant knowledge bases based on task instructions and project context to form a decision support knowledge package; S4. Intelligently decompose and schedule the overall task to generate a sub-task queue; S5. Call the corresponding professional artificial intelligence sub-model to execute the sub-task; S6. Integrate the output results and knowledge packages of all sub-models to generate the final management decision document; S7. Output management decision documents and collect user feedback data to optimize the system model and knowledge base.

Images