Enterprise management optimization method and system of big data enabling ERP

Abstract

The invention discloses an enterprise management optimization method and system of big data enabling ERP, and relates to the technical field of enterprise management, the method comprises the following steps: creating an enterprise research and development task, obtaining research and development resource information, and constructing a task-resource-technology associated dynamic knowledge graph; the task priority is adjusted based on a multi-task game, and a task-resource-technology task line is reconstructed; carrying out adaptive granularity disassembly on the reconstructed task line, generating task key points and establishing parallel sub-task lines; a task line is further optimized, a cross-department cooperation scheme is formulated, and enterprise management optimization is promoted. Therefore, real-time perception, dynamic priority response and efficient resource allocation are achieved, and the cross-department cooperation level is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of enterprise management, and particularly to an enterprise management optimization method and system empowered by big data for ERP. Background Art

[0002] In enterprise R & D management, it is usually necessary to handle multiple R & D project tasks simultaneously and reasonably allocate limited R & D resources (such as manpower, equipment, funds, etc.). In the prior art, enterprises usually rely on traditional project management tools (such as Gantt charts, resource allocation tables, etc.) to plan tasks and resource allocation. These tools define task priorities and resource allocation strategies in a static manner to ensure the timely completion of projects. However, with the increasing complexity and dynamics of enterprise R & D projects, traditional methods are gradually showing deficiencies in handling multi-task collaboration and resource optimization. For example, in the product development management of trading enterprises, there are often risks and problems such as mismatches between R & D and requirements, disconnection between production and sales, and delays caused by slow supply chain response.

[0003] Traditional methods mostly set task priorities based on fixed rules or historical data, and cannot dynamically perceive the real-time association between tasks and resources. At the same time, in the scenario of multi-task parallelism, there is a lack of an efficient cross-departmental collaboration mechanism, resulting in a serious information island phenomenon. To sum up, the prior art has technical problems such as poor management responsiveness and real-time performance, unbalanced and lagging resource allocation, and low cross-departmental collaboration efficiency in enterprise R & D project management. Summary of the Invention

[0004] The present invention provides an enterprise management optimization method and system empowered by big data for ERP to solve the technical problems of poor management responsiveness and real-time performance, unbalanced and lagging resource allocation, and low cross-departmental collaboration efficiency in the prior art, and to achieve the technical effects of real-time perception and dynamic priority response, efficient resource allocation, and improved cross-departmental collaboration level.

[0005] In a first aspect, the present invention provides an enterprise management optimization method empowered by big data for ERP, wherein the enterprise management optimization method empowered by big data for ERP includes: Create multiple enterprise R & D project tasks and obtain the R & D resource information of the enterprise.

[0006] Construct a dynamic knowledge graph of task-resource-technology association according to the multiple enterprise R & D project tasks and the R & D resource information.

[0007] Perform multi-task game on the multiple enterprise R & D project tasks, and after adjusting the task priorities of the enterprise R & D project tasks according to the multi-task game results, reconstruct the task-resource-technology task line.

[0008] Perform adaptive task granularity decomposition using the reconstructed task-resource-technology task line, create task key points based on the results of the adaptive task decomposition, and establish parallel sub-task lines.

[0009] After optimizing the task-resource-technology task line according to the parallel sub-task line, create a decision-making plan for cross-departmental collaboration, and optimize enterprise management according to the decision-making plan.

[0010] In a feasible implementation, the multi-task game for the multiple enterprise R & D project tasks includes: Model the multiple enterprise R & D project tasks as task agents.

[0011] Establish a state space, which includes the current resource occupancy state, task state, task dependency graph, and executed records.

[0012] After the task agent selects an execution action in the dynamic knowledge graph associated with task-resource-technology, perform game analysis under the game objective depending on the state space to complete the multi-task game.

[0013] In a feasible implementation, the game analysis under the game objective depending on the state space includes: Create a game objective, which includes the objective of maximizing the overall project benefit, the objective of minimizing conflicts, and the objective of local optimal matching, where the local optimal matching objective is the alternative abandonment objective.

[0014] After all task agents access the state space, perform a self-evaluation of the current situation.

[0015] Based on the results of the self-evaluation of the current situation, perform an action search for maximizing the self-benefit of each task agent, establish agent actions, and the evaluation indicators of the self-benefit include task value, delay penalty, and requested resource cost.

[0016] Establish update actions, which include self-scheduling actions, delayed resource release actions, conflict priority promotion actions, and shared cooperation actions. Establish a resource conflict degree according to the agent actions, select an update action using the resource conflict degree, then perform task agent update, and perform game evaluation through the game objective.

[0017] Complete the game analysis according to the game evaluation results.

[0018] In a feasible implementation, the selection of update actions using the resource conflict degree includes: Perform task association analysis on the multiple enterprise R & D project tasks, establish the association analysis results, which include sequential association, collaborative association, and conflict association.

[0019] Establish action update constraints using the correlation analysis results, and select update actions using the action update constraints and the resource conflict degree.

[0020] In a feasible implementation manner, the completing game analysis according to the game evaluation results includes: Set the game round period. In each game round, the task agent updates the task body according to the selected update action, and optimizes the action according to the updated state space and the reward function.

[0021] Simulate the action interactions of all task agents, and detect resource conflicts and dependency contradictions.

[0022] Establish game feedback, optimize the update actions in the next game round, and complete the game when the game round period is satisfied or a game stable state appears.

[0023] In a feasible implementation manner, the adaptively disassembling the task granularity using the reconstructed task-resource-technology task line and creating task key points according to the adaptive task disassembling results includes: Adaptively disassemble the task according to the complexity, resource dependency, and technical connectivity of the task, and establish a subtask set.

[0024] Configure key point identification features, where the key point identification features include task transition bottleneck features, connectivity hub features, and technical breakthrough features.

[0025] Use the key point identification features to identify the key task points of the subtask set and create task key points.

[0026] In a feasible implementation manner, the establishing parallel subtask lines includes: Call the expert knowledge base, use the expert knowledge base to identify alternative technical paths for task key points, and establish branch technical paths.

[0027] Use the branch technical paths to establish parallel subtask lines.

[0028] In a feasible implementation manner, the multi-task game for the multiple enterprise R & D project tasks further includes: Establish an expected revenue model, which is used for revenue prediction of enterprise R & D project tasks.

[0029] Create a customer strategic weight factor, and the evaluation indicators of the customer strategic weight factor include stability indicators, profit contribution indicators, cooperation history indicators, and strategic fit indicators.

[0030] Establish a revenue function based on the expected revenue model and the customer strategic weight factor, and use the revenue function to conduct a game evaluation of multi-task games.

[0031] In a feasible implementation manner, the enterprise management optimization according to the solution decision includes: Create a monitoring time window using the solution decision.

[0032] Conduct execution monitoring of the solution decision within the monitoring time window and establish monitoring feedback.

[0033] Use the monitoring feedback to predict the execution impact, generate a decision optimization plan, and perform optimization management of the solution decision through the decision optimization plan.

[0034] In a second aspect, the present invention also provides an enterprise management optimization system empowered by big data for ERP. Among them, the enterprise management optimization system empowered by big data for ERP includes: A task and resource information acquisition module, which is used to create multiple enterprise R & D project tasks and acquire the R & D resource information of the enterprise.

[0035] A dynamic graph module, which is used to construct a dynamic knowledge graph related to tasks, resources, and technologies based on the multiple enterprise R & D project tasks and the R & D resource information.

[0036] A priority reconstruction module, which is used to conduct multi-task games on the multiple enterprise R & D project tasks, adjust the task priorities of the enterprise R & D project tasks according to the results of the multi-task games, and then reconstruct the task-resource-technology task line.

[0037] A key point definition module, which is used to perform adaptive task granularity decomposition using the reconstructed task-resource-technology task line, create task key points according to the results of the adaptive task decomposition, and establish parallel sub-task lines.

[0038] A collaboration decision module, which is used to optimize the task-resource-technology task line according to the parallel sub-task lines, create a solution decision for cross-departmental collaboration, and perform enterprise management optimization according to the solution decision.

[0039] The present invention discloses an enterprise management optimization method and system empowered by big data for ERP, including: creating multiple enterprise R & D project tasks and obtaining relevant R & D resource information; based on these R & D project tasks and R & D resource information, constructing a dynamic knowledge graph associated with tasks, resources, and technologies. Conducting multi-task games for these R & D project tasks, adjusting the priorities of tasks according to the game results, and then reconstructing the task-resource-technology task line. Using the reconstructed task-resource-technology task line for adaptive task granularity decomposition, creating task key points according to the decomposition results, and then establishing parallel sub-task lines. Optimizing the task-resource-technology task line according to the parallel sub-task lines, creating a decision-making plan for cross-departmental collaboration, and optimizing enterprise management according to this decision. The enterprise management optimization method and system empowered by big data for ERP disclosed by the present invention solve the technical problems of poor management responsiveness and real-time performance, unbalanced and lagging resource allocation, and low cross-departmental collaboration efficiency, and achieve the technical effects of real-time perception and dynamic priority response, efficient resource allocation, and improved cross-departmental collaboration level. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a flowchart of the enterprise management optimization method empowered by big data for ERP of the present invention.

[0041] Figure 2 It is a structural diagram of the enterprise management optimization system empowered by big data for ERP of the present invention.

[0042] Description of the reference numerals: Task and resource information acquisition module 11, dynamic graph module 12, priority reconstruction module 13, key point definition module 14, collaboration decision-making module 15. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] The following will combine the specification drawings and specific embodiments to elaborate on the above technical solutions in detail to better understand the above technical solutions. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments for explaining the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention. In addition, it should be noted that for the sake of description, only the parts related to the present invention are shown in the drawings rather than all of them.

[0044] Embodiment 1, as Figure 1 It is a flowchart of the enterprise management optimization method empowered by big data for ERP of the present invention. Among them, the enterprise management optimization method empowered by big data for ERP includes: S100: Create multiple enterprise R & D project tasks and obtain the R & D resource information of the enterprise.

[0045] Specifically, first, break down complex enterprise R & D goals into multiple executable independent tasks (multiple enterprise R & D project tasks) in a systematic and structured manner. Among them, the task information of each enterprise R & D project task includes clear goals, time limits, resource requirements, etc.

[0046] Specifically, according to the multiple enterprise R & D project tasks created, collect all R & D-related resource information within the enterprise through data collection and integration, including human resources (such as R & D personnel skills and available time), equipment (such as laboratory instruments, testing equipment), funds (such as budget allocation), and other technical resources (such as patents, technical documents, etc.).

[0047] S200: Construct a dynamic knowledge graph that associates tasks, resources, and technologies based on the multiple enterprise R & D project tasks and the R & D resource information.

[0048] Specifically, the dynamic knowledge graph of task-resource-technology association is a knowledge network constructed with tasks, resources, and technologies as core nodes through dynamic association relationships. This dynamic graph can reflect the dynamic relationships between tasks, resources, and technologies in real time and support real-time updates and reasoning. Among them, the process of constructing the dynamic knowledge graph includes steps such as data collection, preprocessing, entity and relationship extraction, graph construction and storage, real-time update and reasoning.

[0049] Specifically, first, obtain structured and unstructured data from multiple data sources within the enterprise (such as project management systems, resource management systems, technical document libraries, etc.) and perform cleaning, format standardization, and event extraction to ensure data consistency and availability; then, extract task, resource, and technology entities from the text through named entity recognition technology or large language models and identify the relationships between them. For example, through the LLM (large language model), relationships such as "Task A requires the support of Technology B, and Technology B depends on Resource C" can be extracted from technical documents; then, map the extracted entities and relationships to the graph structure and store them using a temporal graph database (such as TigerGraph). Among them, each entity and relationship is attached with a timestamp attribute to support time window queries and historical backtracking.

[0050] Optionally, through streaming computing and incremental learning technologies, the graph structure is updated in real time, and real-time reasoning is performed using graph neural networks (GNNs) or rule engines. By constructing a dynamic knowledge graph, the dynamic relationships between tasks, resources, and technologies can be perceived in real time, providing data support and intelligent decision-making basis for subsequent multi-task game optimization and cross-departmental collaboration.

[0051] S300: Perform multi-task game on the multiple enterprise R & D project tasks. After adjusting the task priorities of the enterprise R & D project tasks according to the multi-task game results, reconstruct the task-resource-technology task line.

[0052] Specifically, in the case of limited resources among multiple tasks and dependencies and competition relationships among tasks, introduce the game theory concept to optimize task scheduling, so as to determine the task priority allocation strategy under the maximization of overall benefits or local balance. The task-resource-technology task line is the mapping relationship chain among tasks, allocable resources, and adapted technologies in the R & D process, and is the basic structure for task execution planning and resource allocation.

[0053] In some embodiments, the performing multi-task game on the multiple enterprise R & D project tasks includes: Model the multiple enterprise R & D project tasks as task agents; establish a state space, where the state space includes the current resource occupancy state, task state, task dependency graph, and executed records; after the task agents select execution actions in the dynamic knowledge graph associated with task-resource-technology, perform game analysis under the game objective depending on the state space to complete the multi-task game.

[0054] Specifically, a task agent is an intelligent individual with independent decision-making ability abstracted from each independent enterprise R & D project task, and is used to simulate its behavior choices under different states and resource conditions. The state space represents the multi-dimensional information set in the process of task scheduling and execution, including the resource occupancy state (such as the occupancy of developers, equipment, and budget), task state (such as completed, in progress, not started), task dependency graph (indicating the sequential and dependency relationships among tasks), executed records (including historical task decisions and execution feedback), etc.

[0055] Specifically, first, according to information such as task requirements, task return rate, and task cost, abstractly model the multiple enterprise R & D project tasks as independent task agents respectively, so that they have the ability of state judgment and strategy game. In other words, each agent has clear goals (such as completion time, resource requirements) and decision-making ability (such as selecting execution actions); subsequently, construct a multi-dimensional state space including resources, tasks, dependencies, and execution history to completely describe the operation situation of the current R & D system. Then, model the connections between tasks and their required resources and adapted technologies through a dynamic knowledge graph, so that task agents can select the optimal execution actions (such as applying for resources, adjusting the execution order, requesting cooperation, etc.) in the graph structure, and perform game analysis according to the state space and the determined execution actions, so as to evaluate the task completion effects (such as completion time, resource utilization rate) under different strategies and select the optimal strategy.

[0056] Specifically, after the agent selects an action, multi-task game analysis is carried out based on preset game objectives (such as the shortest total execution time, the maximum output benefit, the minimum conflict rate, etc.), the optimal decision-making strategy of each task in the current state is calculated, the task priorities are dynamically adjusted to ensure that critical tasks can be completed in a timely manner, and the task-resource-technology task line is reconstructed according to the adjustment results to ensure the efficiency of task execution and the rationality of resource utilization, so as to achieve dynamic and optimal R & D resource scheduling and execution path configuration.

[0057] In the above process, by introducing a multi-task game mechanism, the originally independent and static tasks of enterprise R & D projects are abstractly transformed into task agents with dynamic game behaviors, thereby improving the flexibility of task scheduling and the operating efficiency of the overall system. At the same time, with the help of the task-resource-technology association knowledge graph and state space modeling, the system can fully explore the dependency relationships and cooperation potential among tasks and achieve efficient reconstruction of the task line.

[0058] In some implementation manners, the game analysis under the game objective depending on the state space includes: Create a game objective, where the game objective includes the objective of maximizing the overall project benefit, the objective of minimizing conflicts, and the objective of local optimal matching. Among them, the local optimal matching objective is an alternative abandonment objective; after all task agents visit the state space, perform a self-evaluation of the current situation; based on the self-evaluation result of the current situation, conduct an action search for maximizing the self-benefit of each task agent, establish agent actions, and the evaluation indicators of the self-benefit include task value, delay penalty, and requested resource cost; establish update actions, where the update actions include self-scheduling actions, delayed resource release actions, conflict priority promotion actions, and shared cooperation actions, establish a resource conflict degree according to the agent actions, select an update action using the resource conflict degree, then perform task agent update, and conduct game evaluation through the game objective; complete the game analysis according to the game evaluation result.

[0059] Specifically, the game objective is the optimization objective in the set multi-task game, including the objective of maximizing the overall project benefit (that is, the highest overall task value and resource utilization efficiency), the objective of minimizing conflicts (that is, the least resource requests or task dependency conflicts), and the objective of local optimal matching (referring to the scheme of allocating tasks and resources nearby within a specific time window, but it is a sub-optimal strategy and can be abandoned when conflicts are severe).

[0060] Specifically, the current situation self-assessment is a predictive analysis of the current task-resource matching pattern by the task agent after reading the state space information. According to the results of the current situation self-assessment, an action search for maximizing its own benefits can be carried out. That is, the task agent tries to select the strategy action with the optimal benefits under the current resource conditions according to the situation assessment results. Among them, its own benefits are composed of task value (the contribution degree of the task to the overall goal), delay penalty (the penalty coefficient generated when the task is not completed according to the plan), and requested resource cost (the cost or scarcity corresponding to the requested resources).

[0061] Specifically, the update action refers to the action selected by the task agent according to the game analysis results, including its own scheduling action (such as advancing or delaying task execution), delaying resource release action (that is, delaying the release of resources that should have been released to hedge risks), conflict priority promotion action (promoting the game priority of urgent tasks), and sharing cooperation action (sharing resources with other tasks to achieve win-win results). Among them, the resource conflict degree refers to the conflict degree generated among task agents due to resource competition, and is used to select the optimal update action.

[0062] Specifically, first, create game goals: set the goal of maximizing the overall project benefits (such as improving the project completion rate and resource utilization rate), the goal of minimizing conflicts (such as reducing resource conflicts between tasks), and the goal of local optimal matching (such as optimizing the resource allocation of a single task). The goal of local optimal matching can be discarded when necessary to give priority to achieving the first two goals.

[0063] Then, conduct a current situation self-assessment: each task agent accesses the state space, evaluates the situation it is in according to the current resource occupancy status, task status, task dependency graph, and executed records, and searches for possible actions according to the evaluation indicators of its own benefits (task value, delay penalty, requested resource cost), and selects the action that can maximize its own benefits.

[0064] Next, the task agent selects update actions according to the search results, including its own scheduling action (such as adjusting the task execution order), delaying resource release action (such as delaying the release of currently occupied resources), conflict priority promotion action (such as promoting the priority of high-conflict tasks), and sharing cooperation action (such as sharing resources with other task agents).

[0065] Furthermore, based on the agent actions after action update, calculate the resource conflict degree among task agents to select the update action with the lowest conflict degree for execution. Thus, after executing the update action, evaluate the entire game process according to the game goals (maximizing the overall project benefits, minimizing conflicts) to ensure the optimization of task priorities and resource allocations.

[0066] Through the above multi-task game analysis, the task priorities and resource allocations can be dynamically adjusted, the task execution order can be optimized, resource conflicts can be reduced, and the overall project benefits can be improved.

[0067] In some implementation manners, the step of selecting an update action by using the resource conflict degree includes: Performing task association analysis on the multiple enterprise R & D project tasks to establish an association analysis result, where the association analysis result includes sequential association, collaborative association, and conflict association; establishing action update constraints by using the association analysis result, and selecting an update action by using the action update constraints and the resource conflict degree.

[0068] Specifically, first, perform systematic modeling and classification recognition on the mutual relationships between multiple enterprise R & D project tasks to obtain an association analysis result including sequential association (task A must be completed before task B), collaborative association (multiple tasks need to be executed in parallel or cooperatively within a certain time window), and conflict association (multiple tasks compete for the same type of resources or there is logical exclusion), that is, the structured output of the above associations, as the basic data for subsequent task scheduling and behavior strategy optimization.

[0069] Specifically, the action update constraint is a restrictive condition set based on the association relationship between tasks. For example, if tasks have a sequential association, subsequent tasks cannot be advanced; collaborative tasks should be synchronized in time, etc. The resource conflict degree is used to quantify the degree of resource contention among multiple tasks and is an important factor in task scheduling optimization to determine whether an update action needs to be executed.

[0070] Furthermore, in combination with the current resource conflict degree, that is, the competition intensity of multiple tasks for key resources, comprehensively consider the action update constraint and the conflict degree, select the most suitable update action type, such as delayed release, collaborative sharing, or priority adjustment, etc., and update the behavior plan of the task agent accordingly, so as to avoid resource competition as much as possible on the basis of conforming to the task logical dependencies and optimize the overall scheduling strategy.

[0071] In some implementation manners, the step of completing the game analysis according to the game evaluation result includes: Setting a game round period. In each game round, the task agent updates the task body according to the selected update action and optimizes the action according to the updated state space and reward function; simulating the action interactions of all task agents, detecting resource conflicts and dependency contradictions; establishing game feedback, optimizing the update action of the next game round, and completing the game when the game round period is satisfied or a game stable state appears.

[0072] Specifically, the game round cycle refers to a discrete time unit in the game process, which is used to synchronize the strategy updates and interaction behaviors of each intelligent agent. In other words, within the game round cycle, the task intelligent agent adjusts its own state and strategy according to the selected update actions (such as its own scheduling action, delaying the release of resources action, etc.). The game stable state refers to the state where the strategies and resource allocations of the task intelligent agents reach a relatively stable state after multiple rounds of games. At this time, it can be considered that the game analysis is completed.

[0073] Specifically, by simulating the actions of all task intelligent agents (such as resource application, task migration, priority adjustment, etc.), resource conflicts (such as two tasks simultaneously applying for the same GPU) and dependency contradictions (such as task B depending on task A but task A not being completed) are detected, and reference information for strategy adjustment in the next round (such as resource conflict degree, task completion rate, revenue function value) is provided according to the game results of the current round, guiding it to adjust strategies in the next round (such as avoiding highly competitive resources, adjusting the execution order, etc.). This process continues until the preset game round upper limit is reached, or the system enters a certain stable state (such as Nash equilibrium or resource conflict-free state).

[0074] Through this step, dynamic collaboration and resource optimization scheduling among multiple task intelligent agents can be achieved, avoiding local optimum problems caused by static allocation or single-agent optimization. Among them, the game round mechanism ensures the orderliness and traceability of strategy updates, and the feedback mechanism enhances the adaptive ability of the system; at the same time, by detecting resource conflicts and dependency contradictions, the operation risks such as system deadlocks and task blockages are effectively reduced.

[0075] In some embodiments, the multi-task game for the multiple enterprise R & D project tasks further includes: Establishing an expected revenue model, which is used to predict the revenue of enterprise R & D project tasks; creating a customer strategic weight factor, and the evaluation indicators of the customer strategic weight factor include a stability indicator, a profit contribution indicator, a cooperation history indicator, and a strategic fit indicator; establishing a revenue function according to the expected revenue model and the customer strategic weight factor, and using the revenue function to conduct game evaluation of the multi-task game.

[0076] Specifically, the expected revenue model is a mathematical model used to predict the potential revenue of enterprise R & D project tasks, and this expected revenue model can be constructed based on historical data, market trends, and project characteristics. The customer strategic weight factor is a weight indicator used to evaluate the importance of the customer to the enterprise strategy, and its evaluation indicators include a stability indicator (the stability of the customer's business), a profit contribution indicator (the profit brought by the customer to the enterprise), a cooperation history indicator (the cooperation duration and frequency with the customer), and a strategic fit indicator (the matching degree between the customer and the enterprise's strategic goals).

[0077] Specifically, the revenue function refers to a function that combines the expected revenue model and the customer strategic weight factor, and is used to quantify the comprehensive revenue of tasks, providing an evaluation basis for multi-task games. Exemplarily, this function can be expressed as: Revenue = Expected Revenue × Customer Strategic Weight Factor, where the expected revenue comes from the expected revenue model, and the customer strategic weight factor is calculated based on the above four indicators.

[0078] By establishing the expected revenue model and the customer strategic weight factor, and using the revenue function for multi-task game evaluation, it is possible to more comprehensively and accurately evaluate the comprehensive revenue of tasks, avoiding decision-making biases caused by single-indicator evaluation; at the same time, the introduction of the customer strategic weight factor ensures that customer tasks crucial to the enterprise's development can be prioritized; the comprehensively constructed revenue function provides a quantitative evaluation basis for multi-task games, making task priority adjustment and resource allocation more scientific and reasonable.

[0079] S400: Use the reconstructed task-resource-technology task line to perform adaptive task granularity decomposition, create task key points according to the adaptive task decomposition results, and establish parallel sub-task lines.

[0080] Specifically, adaptive task granularity decomposition is a process of dynamically decomposing a task into multiple sub-tasks according to the task's complexity, resource dependence, and technical connectivity, in order to improve the flexibility and efficiency of task execution. Task key points refer to task nodes that have a decisive impact on the overall performance or structural stability in the task network. The sub-task set is a collection of multiple sub-tasks formed after the task is decomposed, and each sub-task has clear goals and execution steps.

[0081] In some embodiments, the use of the reconstructed task-resource-technology task line to perform adaptive task granularity decomposition and create task key points according to the adaptive task decomposition results includes: Perform adaptive decomposition of the task according to the task's complexity, resource dependence, and technical connectivity to establish a sub-task set; configure key point identification features, where the key point identification features include task conversion bottleneck features, connectivity hub features, and technical breakthrough features; use the key point identification features to identify the key task points of the sub-task set and create task key points.

[0082] Specifically, the key point identification features are features used to identify key nodes in a task, including task conversion bottleneck features (such as long conversion time between tasks), connectivity hub features (such as connection points between multiple sub-tasks), and technical breakthrough features (such as nodes that require key technical breakthroughs in the task).

[0083] Specifically, first, based on the reconstructed task-resource-technology task line, the original task is adaptively decomposed into granularity. Exemplarily, it includes analyzing the complexity of each task (such as the number of dependencies, the depth of execution logic), resource dependencies (such as whether it depends on specific hardware or shared resources), and technical connectivity (such as whether it connects multiple technical paths or modules), and dynamically decomposing the task into multiple subtasks. For example, a complex product R & D task can be decomposed into multiple subtasks such as design, testing, and optimization. Then, the decomposed subtasks are organized into a subtask set, where each subtask stores clear goals, resource requirements, and execution steps. Preferably, the subtask set is managed through a tree structure for easy tracking and adjustment.

[0084] Furthermore, then, configure a set of characteristic indicators for identifying key task points, including: task transition bottleneck characteristics (such as resource switching delay, task waiting time), connectivity hub characteristics (such as the degree centrality of the task in the task network), technical breakthrough characteristics (such as whether new algorithms or key processes are introduced), and analyze the subtask set based on these characteristics to identify the subtasks that have the greatest impact on the task process, which are marked as task key points.

[0085] The function of the above process is that through adaptive task granularity decomposition and key task point identification, complex tasks can be managed more flexibly, adapting to the complexity and changes of tasks, and improving task execution efficiency. Among them, by identifying key task points, it helps to preferentially allocate resources to key tasks and improve resource utilization efficiency.

[0086] In some embodiments, the establishment of parallel subtask lines includes: Call the expert knowledge base, use the expert knowledge base to identify alternative technical paths for task key points, and establish branch technical paths; use the branch technical paths to establish parallel subtask lines.

[0087] Specifically, expand the original linear task process into multiple sub-processes (subtask lines) that can be executed in parallel. These sub-processes can run simultaneously on the premise of meeting the dependency relationship to further improve the execution efficiency and enhance the stability of task execution; among them, the expert knowledge base is a data system containing knowledge such as industry experience, process rules, and technical alternative solutions, and is used to provide the general and domain information required for intelligent reasoning and complementation in complex decision-making.

[0088] Specifically, the alternative technical path is an alternative technical solution or process flow provided by the expert knowledge base at the task key point in addition to the default technical implementation path. In other words, the alternative technical path has functional equivalence or compatibility with the existing technical path of the task key point. The branch technical path refers to an alternative execution path formed outside the original task main line based on the above alternative technical path, that is, it constitutes a branch structure of the task process.

[0089] Specifically, outside the original task main line, one or more functionally equivalent branch paths are introduced for key task nodes and organized into a structured parallel sub-task line. This parallel sub-task line forms a concurrent execution relationship with the main task line, providing the possibility of dynamically selecting the optimal path according to resource status, performance feedback, fault conditions, etc. during task execution, which enhances the flexibility of the task process, improves the task parallelism and the system fault tolerance ability.

[0090] S500: After optimizing the task-resource-technology task line according to the parallel sub-task line, create a solution decision for cross-departmental collaboration, and optimize enterprise management according to the solution decision.

[0091] Specifically, after optimizing the task line, based on the task execution model, resource distribution status, and technical ability boundary, make a solution decision for cross-departmental collaboration. Among them, the solution decision includes key elements such as task assignment, resource scheduling, responsibility division, and time nodes.

[0092] Specifically, create a solution decision for cross-departmental collaboration to break down departmental barriers, promote information sharing and collaborative work. Exemplarily, through goal alignment, task responsibility assignment, information sharing platform (such as building a digital dashboard to enable the R & D department to instantly retrieve market feedback data and shorten the product iteration decision-making time), and hierarchical decision-making mechanism (such as setting up a joint decision-making group composed of technology, market, and production leaders, and shortening the approval cycle through a pre-empowerment mechanism).

[0093] Specifically, after introducing the cross-departmental collaboration solution decision mechanism, the enterprise can achieve a deep integration between the task execution process and organizational management. Exemplarily, in an intelligent manufacturing enterprise, the R & D task of a new type of denitration catalyst is decomposed into multiple parallel sub-tasks: ① Material ratio optimization (responsible by the R & D department), ② Sample trial production and performance testing (completed jointly by the trial production workshop and the quality inspection department), ③ Process parameter modeling (completed by the data analysis department), ④ Production line deployment (responsible by the engineering and operation and maintenance departments). Then, the collaboration solution automatically generated according to the task-resource-technology model can clarify the person in charge of each task node, cross-departmental contacts, required resources, time nodes, and data interaction interfaces.

[0094] Preferably, through the above cross-departmental collaboration solution decision mechanism, creating a solution decision for cross-departmental collaboration to optimize enterprise management further includes: A knowledge R & D assistant for synchronizing knowledge content and standardizing terms among multiple business departments; supporting the automatic extraction of key terms and knowledge points from internal department documents, meeting minutes, and project materials; achieving term alignment, concept unification, and knowledge sharing through semantic analysis and knowledge graph technology; helping to improve cross - department communication efficiency and reduce understanding deviations caused by inconsistent terms.

[0095] An intelligent recommendation module for providing multi - dimensional intelligent recommendation services based on the current project or task requirements, including but not limited to: personnel recommendation: recommending the most suitable personnel to participate according to the ability profile, historical project experience, and current workload; tool recommendation: recommending appropriate collaboration tools or platforms according to the task type and department - preferred tools; methodology recommendation: recommending standard processes, analysis methods, or industry best practices applicable to the current task; scheduling recommendation: intelligently generating feasible project scheduling suggestions by combining the resource occupancy of each department.

[0096] A dynamic monitoring module for real - time monitoring and progress synchronization of the execution process of cross - department collaboration tasks; supporting the dynamic display and warning of information such as task status, key nodes, and resource occupancy; setting an automatic coordination mechanism, when the progress of a certain department lags behind, the system automatically notifies the relevant responsible person or adjusts the schedule; ensuring the completion of cross - department collaboration tasks on schedule and reducing the risk of delays.

[0097] A technical document assistant for intelligently parsing unstructured data such as technical documents and project materials; automatically extracting key entity content, such as functional modules, technical parameters, interface information, and dependency relationships; supporting the generation of structured knowledge cards for quick cross - department understanding and reuse; at the same time, it can be linked with the knowledge R & D assistant module to achieve knowledge precipitation and accumulation.

[0098] The above process, through the optimization of the task - resource - technology task line and the establishment of a cross - department collaboration mechanism, can improve task execution efficiency and resource utilization rate, while promoting collaborative work among departments and enhancing the overall operation efficiency and competitiveness of the enterprise.

[0099] In some embodiments, the enterprise management optimization according to the solution decision includes: Creating a monitoring time window using the solution decision; performing execution monitoring of the solution decision within the monitoring time window and establishing monitoring feedback; using the monitoring feedback for execution impact prediction, generating a decision optimization plan, and performing optimization management of the solution decision through the decision optimization plan.

[0100] Specifically, the monitoring time window refers to a specific time period or time length set according to the task execution plan, resource scheduling cycle, and key node time points for real-time monitoring and evaluation of the implementation of the solution. This window has attributes such as start and end times, monitoring frequency, and monitoring indicators.

[0101] Specifically, within the monitoring time window, real-time or periodic data collection and status analysis are carried out on the actual implementation of the solution decision, including task progress, resource usage, cross-departmental collaboration status, etc., so as to obtain data results and status evaluation information, including the identification of problems such as task delays, resource conflicts, and collaboration bottlenecks, and the output is the monitoring feedback.

[0102] Furthermore, based on the monitoring feedback, combined with a preset impact prediction model (such as a rule-based, time series model trained based on historical data, causal inference model, etc.), trend analysis and result deduction are carried out to determine the problems such as subsequent task delays, resource bottlenecks, and collaboration failures that may be caused by the current execution status for execution impact prediction; furthermore, according to the prediction results, a decision optimization plan is automatically generated, for example, advancing a certain task node, changing the resource allocation order, adjusting the collaboration interface, etc.; finally, this optimization plan is fed back into the enterprise decision-making process to dynamically adjust the original plan and achieve the closed-loop control of optimized management.

[0103] Exemplarily, taking an environmental protection equipment manufacturing enterprise as an example, when carrying out a modular design task of an SCR denitration system, the system has generated parallel subtask lines and a cross-departmental collaboration plan, and set a monitoring time window of 48 hours, collecting the task execution status every 30 minutes. During the monitoring process, it is found that the "module interface standardization" task responsible by the design department is delayed by more than 4 hours, resulting in the subsequent structural simulation task unable to start on time. At this time, a monitoring feedback is immediately generated, and through the execution impact prediction model, it is judged that if not adjusted, this delay will cause the overall delivery time to be postponed by 1.5 days. At this time, a corresponding optimization plan is generated, suggesting that some simulation tasks be changed to use historical data for pre-simulation, and temporarily call external resources to assist in completing the interface standardization task, so that the project is completed and delivered on time.

[0104] Through the above mechanism, the enterprise can achieve real-time monitoring and dynamic optimization of the solution decision during the task execution process, significantly improving the agility and execution efficiency of management.

[0105] In summary, the enterprise management optimization method empowered by big data for ERP provided by the present invention has the following technical effects: By creating multiple enterprise R & D project tasks and obtaining relevant R & D resource information; based on these R & D project tasks and R & D resource information, constructing a dynamic knowledge graph that associates tasks, resources, and technologies. Conduct multi-task games for these R & D project tasks, adjust the priorities of the tasks according to the game results, and then reconstruct the task-resource-technology task line. Use the reconstructed task-resource-technology task line for adaptive task granularity decomposition, create task key points according to the decomposition results, and then establish parallel sub-task lines. Optimize the task-resource-technology task line according to the parallel sub-task lines, create a decision-making plan for cross-departmental collaboration, and optimize enterprise management according to this decision, so as to achieve the technical effects of real-time perception and dynamic priority response, efficient resource allocation, and improved cross-departmental collaboration level.

[0106] Embodiment 2, such as Figure 2 is a schematic structural diagram of the enterprise management optimization system for big data empowered ERP of the present invention. For example, Figure 1 In this, the flow schematic diagram of the enterprise management optimization method for big data empowered ERP of the present invention can be implemented through a structure such as Figure 2 shown.

[0107] Based on the same concept as the enterprise management optimization method for big data empowered ERP in the above embodiment, the enterprise management optimization system for big data empowered ERP provided by the present invention further includes: A task and resource information acquisition module 11, configured to create multiple enterprise R & D project tasks and obtain the R & D resource information of the enterprise.

[0108] A dynamic graph module 12, configured to construct a dynamic knowledge graph that associates tasks, resources, and technologies according to the multiple enterprise R & D project tasks and the R & D resource information.

[0109] A priority reconstruction module 13, configured to conduct multi-task games for the multiple enterprise R & D project tasks, reconstruct the task-resource-technology task line after adjusting the task priorities of the enterprise R & D project tasks according to the multi-task game results.

[0110] A key point definition module 14, configured to perform adaptive task granularity decomposition using the reconstructed task-resource-technology task line, create task key points according to the adaptive task decomposition results, and establish parallel sub-task lines.

[0111] A collaboration decision-making module 15, configured to optimize the task-resource-technology task line according to the parallel sub-task lines, create a decision-making plan for cross-departmental collaboration, and optimize enterprise management according to the plan decision.

[0112] In some embodiments, the priority reconstruction module 13 includes: A task agent modeling unit for modeling multiple enterprise R & D project tasks as task agents.

[0113] A state space establishment unit for establishing a state space, where the state space includes the current resource occupancy state, task state, task dependency graph, and executed records.

[0114] A multi - task game execution unit for, after the task agent selects an execution action in the dynamic knowledge graph of task - resource - technology association, performing game analysis under the game goal depending on the state space to complete the multi - task game.

[0115] In some implementation manners, the multi - task game execution unit in the priority reconstruction module 13 includes: A game goal creation sub - unit for creating game goals, where the game goals include the goal of maximizing the overall project benefit, the goal of minimizing conflicts, and the goal of local optimal matching, and among them, the local optimal matching goal is an alternative abandonment goal.

[0116] A situation self - assessment sub - unit for performing current situation self - assessment after all task agents access the state space.

[0117] An agent action search sub - unit for performing action search for each task agent to maximize its own benefit based on the current situation self - assessment result, establishing agent actions, and the evaluation indicators of its own benefit include task value, delay penalty, and requested resource cost.

[0118] An updated action establishment sub - unit for establishing updated actions, where the updated actions include its own scheduling actions, delayed resource release actions, conflict priority promotion actions, and shared cooperation actions, establishing a resource conflict degree according to the agent actions, selecting updated actions using the resource conflict degree, then performing task agent update, and performing game evaluation through the game goal.

[0119] A game analysis completion sub - unit for completing game analysis according to the game evaluation result.

[0120] Further, the execution steps of the updated action establishment sub - unit further include: Performing task association analysis on the multiple enterprise R & D project tasks, establishing an association analysis result, where the association analysis result includes sequential association, collaborative association, and conflict association.

[0121] Establishing action update constraints using the association analysis result, and selecting updated actions using the action update constraints and the resource conflict degree.

[0122] Further, the execution steps of the game analysis completion sub - unit further include: Set the game round period. In each game round, the task agent updates the task body according to the selected update action, and optimizes the action based on the updated state space and reward function.

[0123] Simulate the action interactions of all task agents, and detect resource conflicts and dependency contradictions.

[0124] Establish game feedback, optimize the update actions of the next game round, and complete the game when the game round period is satisfied or a game steady state appears.

[0125] In some embodiments, the key point definition module 14 includes: A task adaptive decomposition unit, which is used to adaptively decompose the task according to the complexity, resource dependency, and technical connectivity of the task, and establish a subtask set.

[0126] A key point recognition feature configuration unit, which is used to configure key point recognition features, and the key point recognition features include task conversion bottleneck features, connectivity hub features, and technical breakthrough features.

[0127] A task key point creation unit, which is used to identify the key task points of the subtask set by using the key point recognition features and create task key points.

[0128] In some embodiments, the key point definition module 14 further includes: A backup technical path identification unit, which is used to call the expert knowledge base and use the expert knowledge base to identify the backup technical paths of task key points and establish branch technical paths.

[0129] A parallel subtask line establishment unit, which is used to establish parallel subtask lines by using the branch technical paths.

[0130] In some embodiments, the priority reconstruction module 13 further includes: An expected revenue model establishment unit, which is used to establish an expected revenue model for predicting the revenue of enterprise R & D project tasks.

[0131] A customer strategy weight factor creation unit, which is used to create customer strategy weight factors, and the evaluation indicators of the customer strategy weight factors include stability indicators, profit contribution indicators, cooperation history indicators, and strategic fit indicators.

[0132] A reward function establishment unit, which is used to establish a reward function according to the expected revenue model and the customer strategy weight factor, and use the reward function to evaluate the multi-task game.

[0133] In some embodiments, the collaboration decision module 15 includes: A monitoring time window creation unit, configured to create a monitoring time window by using the solution decision.

[0134] An execution monitoring and feedback unit, configured to perform execution monitoring of the solution decision in the monitoring time window and establish monitoring feedback.

[0135] A decision optimization solution generation unit, configured to perform execution impact prediction by using the monitoring feedback, generate a decision optimization solution, and perform optimization management of the solution decision through the decision optimization solution.

[0136] It should be understood that the key point of the embodiments mentioned in this specification lies in their differences from other embodiments. The specific embodiments in the foregoing Embodiment 1 are equally applicable to the enterprise management optimization system for big data empowered ERP described in Embodiment 2. For the sake of brevity of the specification, no further elaboration will be made here.

[0137] It should be understood that the disclosed embodiments of the present invention and the above descriptions enable those skilled in the art to implement the present invention by using the present invention. At the same time, the present invention is not limited to the above-mentioned part of the embodiments. It should be understood that those of ordinary skill in the art can still modify the technical solutions recorded in the foregoing embodiments or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. The enterprise management optimization method of big data empowering ERP is characterized by: include: Create multiple enterprise R&D project tasks and obtain the enterprise's R&D resource information; Constructing a dynamic knowledge graph of task-resource-technology association according to the multiple enterprise R&D project tasks and the R&D resource information; Conducting a multi-task game on the multiple enterprise R&D project tasks, adjusting the task priorities of the enterprise R&D project tasks according to the multi-task game results, and reconstructing the task-resource-technology task line; Use the reconstructed task-resource-technology task line to perform adaptive task granularity decomposition, create task key points based on the adaptive task decomposition results, and establish parallel sub-task lines; After optimizing the task-resource-technology task line according to the parallel sub-task lines, a cross-departmental collaborative solution decision is created, and enterprise management optimization is performed according to the solution decision.

2. The enterprise management optimization method of big data-enabled ERP as claimed in claim 1, characterized in that: The multi-task game for the multiple enterprise R&D project tasks includes: Model multiple enterprise R&D project tasks as task agents; Establishing a state space, wherein the state space includes current resource occupancy status, task status, task dependency graph, and executed records; After the task agent selects an execution action in the dynamic knowledge graph of task-resource-technology association, the game analysis under the game goal is performed relying on the state space to complete the multi-task game.

3. The enterprise management optimization method of big data-enabled ERP as claimed in claim 2, characterized in that: The game analysis under the game objective based on the state space includes: Creating a game objective, wherein the game objective includes an overall project benefit maximization objective, a conflict minimization objective, and a local optimal matching objective, wherein the local optimal matching objective is an alternative discarded objective; After all task agents have visited the state space, they perform a self-assessment of the current situation; Based on the self-evaluation results of the current situation, each task agent searches for actions that maximize its own benefits and establishes agent actions. The evaluation indicators of its own benefits include task value, delay penalty, and request resource cost; Establish update actions, the update actions include self-scheduling actions, delayed resource release actions, conflict priority enhancement actions, and shared cooperation actions, establish resource conflict degrees according to the agent actions, use the resource conflict degrees to select update actions, execute task agent updates, and conduct game evaluation through the game objectives; Complete the game analysis based on the game evaluation results.

4. The enterprise management optimization method of big data-enabled ERP as claimed in claim 3, characterized in that: The step of selecting an update action by utilizing the resource conflict degree comprises: Performing task correlation analysis on the multiple enterprise R&D project tasks and establishing correlation analysis results, wherein the correlation analysis results include sequential correlation, collaborative correlation, and conflict correlation; An action update constraint is established using the association analysis result, and an update action is selected using the action update constraint and the resource conflict degree.

5. The enterprise management optimization method of big data-enabled ERP as claimed in claim 3, characterized in that: The game analysis is completed according to the game evaluation result, including: Set the game round cycle. In each game round, the task agent performs task body updates according to the selected update actions and optimizes the actions according to the updated state space and benefit function. Simulate the action interactions of all task agents and detect resource conflicts and dependency contradictions; Establish game feedback, optimize the update action of the next game round, and complete the game when the game round cycle is met or the game stable state appears.

6. The enterprise management optimization method of big data-enabled ERP according to claim 1, characterized in that: The adaptive task granularity decomposition is performed using the reconstructed task-resource-technology task line, and task key points are created according to the adaptive task decomposition results, including: Adaptively decompose tasks according to their complexity, resource dependency, and technical connectivity to create subtask sets; Configuring key point identification features, wherein the key point identification features include task conversion bottleneck features, connectivity hub features, and technology breakthrough features; The key point recognition features are used to identify the key task points of the subtask set and create task key points.

7. The enterprise management optimization method of big data-enabled ERP as claimed in claim 6, characterized in that: The establishment of parallel sub-task lines includes: Calling an expert knowledge base, using the expert knowledge base to identify alternate technical paths for key mission points, and establishing branch technical paths; The branching technical paths are used to establish parallel sub-task lines.

8. The enterprise management optimization method of big data-enabled ERP as claimed in claim 1, characterized in that: The multi-task game for the multiple enterprise R&D project tasks also includes: Establishing an expected revenue model, wherein the expected revenue model is used to predict the revenue of enterprise R&D project tasks; Creating a customer strategy weight factor, wherein the evaluation indicators of the customer strategy weight factor include a stability indicator, a profit contribution indicator, a cooperation history indicator, and a strategic fit indicator; A profit function is established according to the expected profit model and the customer strategy weight factor, and the profit function is used to perform game evaluation of the multi-task game.

9. The enterprise management optimization method of big data-enabled ERP according to claim 1, characterized in that: The enterprise management optimization according to the solution decision includes: Creating a monitoring time window using the solution decision; Performing execution monitoring of the solution decision within the monitoring time window and establishing monitoring feedback; The monitoring feedback is used to predict the execution impact, generate a decision optimization plan, and optimize the management of plan decisions through the decision optimization plan.

10. Big data empowers ERP's enterprise management optimization system, which is characterized by: The enterprise management optimization method for implementing the big data-enabled ERP of any one of claims 1 to 9 comprises: The task and resource information acquisition module is used to create multiple enterprise R&D project tasks and obtain the enterprise's R&D resource information; A dynamic graph module, used to construct a dynamic knowledge graph of task-resource-technology association according to the multiple enterprise R&D project tasks and the R&D resource information; A priority reconstruction module, used to conduct a multi-task game on the multiple enterprise R&D project tasks, adjust the task priorities of the enterprise R&D project tasks according to the multi-task game results, and reconstruct the task-resource-technology task line; The key point definition module is used to use the reconstructed task-resource-technology task line to perform adaptive task granularity decomposition, create task key points based on the adaptive task decomposition results, and establish parallel sub-task lines; The collaborative decision-making module is used to optimize the task-resource-technology task line according to the parallel sub-task lines, create a cross-departmental collaborative solution decision, and optimize enterprise management according to the solution decision.

Images