Task scheduling optimization and feedback control method and system based on causal graph structure

Abstract

The invention discloses a task scheduling optimization and feedback control method and system based on a causal graph structure, and relates to the technical field of task scheduling, and the method comprises the steps: integrating multi-source information, identifying an entity, constructing a directed edge connection entity, calculating an edge weight, and forming a task causal graph; performing deep analysis on the causal graph, and identifying risk nodes and propagation paths thereof in the current project; combining the edge weight and the node attribute, quantifying the risk propagation possibility and influence range, and generating a risk thermodynamic diagram; on the basis of the identified risk nodes and propagation paths thereof, mining controllable variables in a task scheduling process, constructing a scheduling optimization objective function, searching a scheduling solution space under constraint conditions by using a search method, and evaluating the influence of different strategies on a task network through multiple rounds of simulation, so as to improve the task scheduling efficiency. A current optimal task adjustment strategy is screened out and practically applied; and collecting data generated in a project execution process, and updating the causal diagram. The method is suitable for intelligent scheduling and fine management scenes of projects.

Description

Technical Field

[0001] The present invention relates to the technical field of task scheduling, and in particular to a task scheduling optimization and feedback control method and system based on a causal graph structure. Background Art

[0002] As enterprise projects continue to expand and cross-departmental collaboration becomes increasingly frequent, task scheduling and progress control in project management face unprecedented complexity. Traditional project management practices often rely on static planning tools such as Gantt charts, critical path methods (CPM), and burndown charts for task scheduling. These tools employ linear scheduling based on a pre-set task sequence and estimated timelines. These tools struggle to address the complex and dynamic dependencies between tasks and are unable to effectively handle unexpected situations that arise during collaboration.

[0003] Furthermore, most current task scheduling systems lack the ability to model the causal impact paths between tasks, making them unable to accurately identify the propagation and cumulative effects of factors such as task delays, resource conflicts, and personnel changes within project networks. This directly leads to problems such as inaccurate schedule forecasts, delayed risk warnings, and slow scheduling responses. Furthermore, most existing systems rely on manual analysis and adjustments, and scheduling optimization relies heavily on the personal experience of project managers. They lack systematic reasoning mechanisms and intelligent feedback capabilities, making them unable to meet the needs of intelligent management in dynamic and complex environments.

[0004] In recent years, the development of artificial intelligence technologies such as graph neural networks, causal reasoning, and reinforcement learning has provided new solutions for the structural modeling and control optimization of complex systems. In the field of project management, the introduction of modeling methods based on causal graphs can more accurately express the causal dependencies and impact paths between tasks. Combining graph computing with optimized control mechanisms is expected to achieve intelligent closed-loop management of risk prediction, strategy adjustment, and execution feedback during task scheduling.

[0005] However, existing methods have yet to systematically integrate causal modeling, scheduling optimization, and real-time feedback control mechanisms, making it difficult to simultaneously meet the comprehensive requirements of project management for strong interpretability, high adaptability, and automated decision-making. Therefore, a new task scheduling optimization and feedback control method based on a causal graph structure is urgently needed to improve task scheduling efficiency and project execution resilience in complex project scenarios. Summary of the Invention

[0006] In response to the needs and shortcomings of current technological development, the present invention provides a task scheduling optimization and feedback control method and system based on a causal graph structure to solve technical problems existing in existing project management and task scheduling processes, such as rough dependency modeling, weak risk identification capabilities, lack of intelligent reasoning in scheduling optimization, and difficulty in closed-loop control of the execution process.

[0007] In the first aspect, the present invention provides a task scheduling optimization and feedback control method based on a causal graph structure, and the technical solutions adopted to solve the above technical problems are as follows:

[0008] A task scheduling optimization and feedback control method based on a causal graph structure comprises the following steps:

[0009] S1. By integrating multi-source information, using natural language processing and knowledge extraction technologies to identify entities, we construct directed edges connecting entities based on the dependencies, influence, and collaboration relationships between entities. We also calculate edge weights using historical task data, quantify the strength of causal relationships, the direction of influence, and the likelihood of risk transmission, and form a structured task causal graph, providing a knowledge foundation for subsequent analysis.

[0010] S2. Use graph neural networks or causal reasoning algorithms to conduct in-depth analysis of task causal graphs, identifying potential risk nodes and their propagation paths in the current project through node embedding learning or causal effect estimation. Combined with edge weights and node attributes, this approach quantifies the likelihood of risk propagation and the scope of impact, generating a risk heat map that visually displays systemic risk concentration areas and potential bottlenecks, providing a decision-making basis for scheduling optimization.

[0011] S3. Based on the identified risk nodes and their propagation paths, we explore the controllable variables in the task scheduling process, construct a weighted scheduling optimization objective function, and use reinforcement learning algorithms, graph optimization algorithms, or heuristic search methods to search the scheduling solution space under constraints. Through multiple rounds of simulation, we evaluate the impact of different strategies on the task network and select the optimal task adjustment strategy.

[0012] S4. Output the optimal task scheduling strategy and apply it to actual project execution to guide task allocation, resource scheduling, and progress control;

[0013] S5. Collect data generated during project execution in real time as feedback information, and update the task cause-and-effect diagram accordingly to achieve closed-loop control of task scheduling and project status, and continuously optimize task scheduling strategies to adapt to dynamic changes in the project.

[0014] Optionally, execute step S1, use natural language processing and knowledge extraction technology to identify three types of entities: task nodes, event nodes, and personnel nodes, and construct directed edges connecting the nodes based on the dependency, influence, and collaboration relationships between the entities.

[0015] Further optionally, when executing step S1, the edge weight is determined by the following operations:

[0016] Comprehensively analyze historical task data to extract the correlation strength and impact frequency between past tasks; conduct in-depth research on project documents to obtain the logical relationships and constraints between tasks; carefully analyze task descriptions to explore task characteristics and potential influencing factors; track task progress in real time to capture dynamic correlations during the execution of current tasks;

[0017] Based on the above data, edge weights are calculated using preset algorithms or rules. Specifically, for dependent edges, the edge weight is calculated based on the probability of the previous task in the history of tasks affecting the on-time completion of the subsequent task. The higher the probability of influence, the greater the edge weight. For impact edges, edge weights are assigned based on the degree of impact on the task duration, cost, and quality when the event occurs. For collaborative edges, the edge weight is determined based on the closeness of collaboration between personnel and the historical performance of collaboration.

[0018] The calculated edge weight intuitively reflects the strength of the causal relationship between tasks, the direction of influence, and the possibility of risk transmission. The larger the value, the stronger the correlation or influence.

[0019] Optionally, executing step S3, the controllable variables include the order of task execution, the assignment of task leaders, the amount of resource input, and whether the tasks can be executed in parallel;

[0020] Construct a scheduling optimization objective function that includes three goals: minimizing the overall delay risk, maintaining the stability of the critical path, and optimizing resource load balancing, and set weights for each goal.

[0021] Optionally, step S5 is executed, and the task causal graph is updated including node status update, edge weight adjustment, dependency modification, or node / edge addition and deletion to reflect actual changes occurring during task execution.

[0022] In a second aspect, the present invention provides a task scheduling optimization and feedback control system based on a causal graph structure, and the technical solutions adopted to solve the above technical problems are as follows:

[0023] A task scheduling optimization and feedback control system based on a causal graph structure, comprising:

[0024] The graph construction module integrates multi-source information, uses natural language processing and knowledge extraction techniques to identify entities, constructs directed edges to connect entities based on dependencies, influence, and collaboration between entities, and calculates edge weights based on historical task data to quantify the strength of causal relationships, direction of influence, and likelihood of risk transmission. This creates a structured task causal graph, providing a knowledge foundation for subsequent analysis.

[0025] The analysis and processing module uses graph neural networks or causal reasoning algorithms to conduct in-depth analysis of task causal graphs, identifying potential risk nodes and their propagation paths in the current project through node embedding learning or causal effect estimation. It combines edge weights and node attributes to quantify the likelihood of risk propagation and the scope of impact, generating a risk heat map that visually displays the clustering areas and potential bottlenecks of systemic risks, providing a decision-making basis for scheduling optimization.

[0026] The construction and search module is used to mine controllable variables in the task scheduling process based on the identified risk nodes and their propagation paths, construct a weighted scheduling optimization objective function, and use reinforcement learning algorithms, graph optimization algorithms, or heuristic search methods to search the scheduling solution space under constraints. Through multiple rounds of simulation, the impact of different strategies on the task network is evaluated to screen out the current optimal task adjustment strategy.

[0027] The strategy output module is used to output the optimal task scheduling strategy and apply it to actual project execution to guide task allocation, resource scheduling, and progress control;

[0028] The feedback update module is used to collect data generated during project execution in real time as feedback information, and update the task cause-effect diagram accordingly, to achieve closed-loop control of task scheduling and project status, and continuously optimize the task scheduling strategy to adapt to dynamic changes in the project.

[0029] Optionally, the graph construction modules involved specifically include:

[0030] Information integration unit, used to integrate multi-source information;

[0031] The identification and construction unit is used to identify three types of entities: task nodes, event nodes, and personnel nodes based on integrated multi-source information using natural language processing and knowledge extraction technologies. Based on the dependency, influence, and collaboration relationships between entities, directed edges are constructed to connect the nodes.

[0032] The edge weight calculation unit is used to calculate edge weights through historical task data, quantify the strength of causal relationships, the direction of impact, and the possibility of risk transmission, form a structured task causal graph, and provide a knowledge basis for subsequent analysis.

[0033] Further optionally, the edge weight calculation unit involved determines the edge weight by the following operations:

[0034] Comprehensively analyze historical task data to extract the correlation strength and impact frequency between past tasks; conduct in-depth research on project documents to obtain the logical relationships and constraints between tasks; carefully analyze task descriptions to explore task characteristics and potential influencing factors; track task progress in real time to capture dynamic correlations during the execution of current tasks;

[0035] Based on the above data, edge weights are calculated using preset algorithms or rules. Specifically, for dependent edges, the edge weight is calculated based on the probability of the previous task in the history of tasks affecting the on-time completion of the subsequent task. The higher the probability of influence, the greater the edge weight. For impact edges, edge weights are assigned based on the degree of impact on the task duration, cost, and quality when the event occurs. For collaborative edges, the edge weight is determined based on the closeness of collaboration between personnel and the historical performance of collaboration.

[0036] The calculated edge weight intuitively reflects the strength of the causal relationship between tasks, the direction of influence, and the possibility of risk transmission. The larger the value, the stronger the correlation or influence.

[0037] Optionally, the construction and search module involves mining controllable variables in the task scheduling process based on the identified risk nodes and their propagation paths, wherein the controllable variables include the order of task execution, the assignment of task leaders, the amount of resource input, and whether the tasks can be executed in parallel;

[0038] The construction and search module constructs a scheduling optimization objective function including three objectives: minimizing the overall delay risk, maintaining the critical path stability, and optimizing the resource load balance, and sets a weight for each objective.

[0039] Optionally, the feedback update module involved collects data generated during the project execution in real time as feedback information. The update of the task causal graph includes node status update, edge weight adjustment, dependency modification or node / edge addition and deletion to reflect the actual changes that occur during the task execution process.

[0040] The present invention provides a method and system for task scheduling optimization and feedback control based on a causal graph structure, which has the following beneficial effects compared with the prior art:

[0041] 1. This invention uses graph-structured modeling to describe the complex dependencies and potential risk propagation paths between tasks. It predicts the delay risks of key tasks through graph neural networks or causal reasoning algorithms. Based on the identification of controllable variables, it constructs a scheduling optimization objective function and dynamically generates scheduling adjustment strategies. Simultaneously, it continuously collects task execution feedback and updates and optimizes the causal graph structure, thereby achieving continuous iteration of scheduling solutions and adaptive evolution of the system.

[0042] 2. This invention constructs a task causal graph consisting of multiple types of nodes, such as tasks, events, and personnel, and multiple causal edges. This graph can comprehensively and structuredly depict the logical relationships and influencing mechanisms between tasks in a project. This graph structure significantly outperforms traditional linear scheduling or static dependency description methods, enabling earlier identification of risk propagation paths and systemic bottlenecks, and improving the accuracy and interpretability of schedule forecasts.

[0043] 3. This invention identifies controllable variables in task scheduling and constructs a scheduling optimization objective function that includes three goals: minimizing overall delay risk, maintaining critical path stability, and optimizing resource load balancing. Weights are assigned to each goal. Subsequently, a reinforcement learning algorithm, graph optimization algorithm, or heuristic search method is used to search the scheduling solution space under constraints. Multiple rounds of simulation evaluate the impact of different strategies on the task network, and the optimal task adjustment strategy is selected. This reduces reliance on the project manager's experience and judgment, reduces the cost of manual intervention in scheduling, and improves the scientific nature and efficiency of scheduling decisions.

[0044] 4. The present invention has closed-loop control capabilities. By collecting task execution data and dynamically updating the causal graph structure and edge weight parameters, the scheduling method or system can adapt to actual changes and support multiple rounds of iterative optimization scheduling, thereby improving the robustness and flexibility of the task scheduling method or system in uncertain environments.

[0045] 5. The present invention has the advantages of strong interpretability of task risks, high degree of automation of scheduling adjustments, and strong dynamic adaptability. This method is widely applicable to multi-task complex system scenarios such as enterprise project management, software development, engineering construction, and supply chain collaboration. It can significantly improve scheduling efficiency, reduce collaboration risks, and enhance the system's resilience and autonomy in dynamic environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Attachment Figure 1 is a flow chart of a method according to embodiment 1 of the present invention;

[0047] Attachment Figure 2 This is a module connection block diagram of the second embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to make the technical solution, the technical problems solved and the technical effects of the present invention more clear, the technical solution of the present invention is clearly and completely described below in conjunction with specific embodiments.

[0049] Example 1:

[0050] Reference Attachment Figure 1 This embodiment proposes a task scheduling optimization and feedback control method based on a causal graph structure, which includes the following steps:

[0051] S1. By integrating multi-source information and utilizing natural language processing and knowledge extraction technologies to identify three types of entities: task nodes, event nodes, and personnel nodes, directed edges are constructed to connect the nodes based on the dependencies, influence, and collaboration relationships between entities. Edge weights are calculated using historical task data to quantify the strength of causal relationships, the direction of influence, and the possibility of risk transmission, thus forming a structured task causal graph and providing a knowledge basis for subsequent analysis.

[0052] This step determines the edge weights by performing the following operations:

[0053] Comprehensively analyze historical task data to extract the correlation strength and impact frequency between past tasks; conduct in-depth research on project documents to obtain the logical relationships and constraints between tasks; carefully analyze task descriptions to explore task characteristics and potential influencing factors; track task progress in real time to capture dynamic correlations during the execution of current tasks;

[0054] Based on the above data, edge weights are calculated using preset algorithms or rules. Specifically, for dependent edges, the edge weight is calculated based on the probability of the previous task in the history of tasks affecting the on-time completion of the subsequent task. The higher the probability of influence, the greater the edge weight. For impact edges, edge weights are assigned based on the degree of impact on the task duration, cost, and quality when the event occurs. For collaborative edges, the edge weight is determined based on the closeness of collaboration between personnel and the historical performance of collaboration.

[0055] The calculated edge weight intuitively reflects the strength of the causal relationship between tasks, the direction of influence, and the possibility of risk transmission. The larger the value, the stronger the correlation or influence.

[0056] S2. Use graph neural networks or causal reasoning algorithms to conduct in-depth analysis of task causal graphs, and identify potential risk nodes and their propagation paths in the current project through node embedding learning or causal effect estimation. Combine edge weights and node attributes to quantify the possibility and scope of risk propagation, generate risk heat maps, and intuitively display the clustering areas and potential bottlenecks of systemic risks, providing a decision-making basis for scheduling optimization.

[0057] S3. Based on the identified risk nodes and their propagation paths, we explore the controllable variables in the task scheduling process. These controllable variables include the order of task execution, the assignment of task leaders, the amount of resource input, and whether tasks can be executed in parallel.

[0058] Construct a scheduling optimization objective function that includes three goals: minimizing overall delay risk, maintaining critical path stability, and optimizing resource load balancing, and set weights for each goal;

[0059] Subsequently, reinforcement learning algorithms, graph optimization algorithms or heuristic search methods are used to search the scheduling solution space under constraints. Through multiple rounds of simulation, the impact of different strategies on the task network is evaluated to screen out the current optimal task adjustment strategy.

[0060] S4. Output the optimal task scheduling strategy and apply it to actual project execution to guide task allocation, resource scheduling and progress control.

[0061] S5. Collect data generated during project execution in real time as feedback information, and update the task cause-and-effect diagram accordingly to achieve closed-loop control of task scheduling and project status, and continuously optimize task scheduling strategies to adapt to dynamic changes in the project.

[0062] The update of the task causal graph in this step includes node status update, edge weight adjustment, dependency modification or node / edge addition and deletion to reflect the actual changes that occur during task execution.

[0063] Example 2:

[0064] Reference Attachment Figure 2 This embodiment proposes a task scheduling optimization and feedback control system based on a causal graph structure, which includes:

[0065] The graph construction module integrates multi-source information, uses natural language processing and knowledge extraction techniques to identify entities, constructs directed edges to connect entities based on dependencies, influence, and collaboration between entities, and calculates edge weights based on historical task data to quantify the strength of causal relationships, direction of influence, and likelihood of risk transmission. This creates a structured task causal graph, providing a knowledge foundation for subsequent analysis.

[0066] The analysis and processing module uses graph neural networks or causal reasoning algorithms to conduct in-depth analysis of task causal graphs, identifying potential risk nodes and their propagation paths in the current project through node embedding learning or causal effect estimation. It combines edge weights and node attributes to quantify the likelihood of risk propagation and the scope of impact, generating a risk heat map that visually displays the clustering areas and potential bottlenecks of systemic risks, providing a decision-making basis for scheduling optimization.

[0067] The construction and search module is used to mine controllable variables in the task scheduling process based on the identified risk nodes and their propagation paths, construct a weighted scheduling optimization objective function, and use reinforcement learning algorithms, graph optimization algorithms, or heuristic search methods to search the scheduling solution space under constraints. Through multiple rounds of simulation, the impact of different strategies on the task network is evaluated to screen out the current optimal task adjustment strategy.

[0068] The strategy output module is used to output the optimal task scheduling strategy and apply it to actual project execution to guide task allocation, resource scheduling, and progress control;

[0069] The feedback update module is used to collect data generated during project execution in real time as feedback information, and update the task cause-effect diagram accordingly, to achieve closed-loop control of task scheduling and project status, and continuously optimize the task scheduling strategy to adapt to dynamic changes in the project.

[0070] In this embodiment, the graph construction modules involved specifically include:

[0071] Information integration unit, used to integrate multi-source information;

[0072] The identification and construction unit is used to identify three types of entities: task nodes, event nodes, and personnel nodes based on integrated multi-source information using natural language processing and knowledge extraction technologies. Based on the dependency, influence, and collaboration relationships between entities, directed edges are constructed to connect the nodes.

[0073] The edge weight calculation unit is used to calculate edge weights through historical task data, quantify the strength of causal relationships, the direction of impact, and the possibility of risk transmission, form a structured task causal graph, and provide a knowledge basis for subsequent analysis.

[0074] The edge weight calculation units involved determine the edge weights through the following operations:

[0075] Comprehensively analyze historical task data to extract the correlation strength and impact frequency between past tasks; conduct in-depth research on project documents to obtain the logical relationships and constraints between tasks; carefully analyze task descriptions to explore task characteristics and potential influencing factors; track task progress in real time to capture dynamic correlations during the execution of current tasks;

[0076] Based on the above data, edge weights are calculated using preset algorithms or rules. Specifically, for dependent edges, the edge weight is calculated based on the probability of the previous task in the history of tasks affecting the on-time completion of the subsequent task. The higher the probability of influence, the greater the edge weight. For impact edges, edge weights are assigned based on the degree of impact on the task duration, cost, and quality when the event occurs. For collaborative edges, the edge weight is determined based on the closeness of collaboration between personnel and the historical performance of collaboration.

[0077] The calculated edge weight intuitively reflects the strength of the causal relationship between tasks, the direction of influence, and the possibility of risk transmission. The larger the value, the stronger the correlation or influence.

[0078] In this embodiment, the construction and search module involved mines the controllable variables in the task scheduling process based on the identified risk nodes and their propagation paths. The controllable variables include the order of task execution, the assignment of task leaders, the amount of resource input, and whether tasks can be executed in parallel.

[0079] The construction and search module constructs a scheduling optimization objective function including three objectives: minimizing the overall delay risk, maintaining the critical path stability, and optimizing the resource load balance, and sets a weight for each objective.

[0080] In this embodiment, the feedback update module involved collects data generated during the project execution as feedback information in real time. The update of the task causal graph includes node status update, edge weight adjustment, dependency modification or node / edge addition and deletion to reflect the actual changes that occur during the task execution process.

[0081] In summary, the task scheduling optimization and feedback control method and system based on the causal graph structure of the present invention can solve technical problems existing in the existing project management and task scheduling processes, such as rough dependency modeling, weak risk identification capabilities, lack of intelligent reasoning in scheduling optimization, and difficulty in closed-loop control of the execution process.

[0082] The above specific examples are used to illustrate the principles and implementation methods of the present invention in detail. These examples are only used to help understand the core technical content of the present invention. Based on the above specific embodiments of the present invention, any improvements and modifications made by those skilled in the art without departing from the principles of the present invention should fall within the scope of patent protection of the present invention.

Claims

1. A task scheduling optimization and feedback control method based on a causal graph structure, characterized in that: The steps include: S1. By integrating multi-source information, using natural language processing and knowledge extraction technologies to identify entities, we construct directed edges connecting entities based on the dependencies, influence, and collaboration relationships between entities. We also calculate edge weights using historical task data, quantify the strength of causal relationships, the direction of influence, and the likelihood of risk transmission, and form a structured task causal graph, providing a knowledge foundation for subsequent analysis. S2. Use graph neural networks or causal reasoning algorithms to conduct in-depth analysis of task causal graphs, identifying potential risk nodes and their propagation paths in the current project through node embedding learning or causal effect estimation. Combined with edge weights and node attributes, this approach quantifies the likelihood of risk propagation and the scope of impact, generating a risk heat map that visually displays systemic risk concentration areas and potential bottlenecks, providing a decision-making basis for scheduling optimization. S3. Based on the identified risk nodes and their propagation paths, we explore the controllable variables in the task scheduling process, construct a weighted scheduling optimization objective function, and use reinforcement learning algorithms, graph optimization algorithms, or heuristic search methods to search the scheduling solution space under constraints. Through multiple rounds of simulation, we evaluate the impact of different strategies on the task network and select the optimal task adjustment strategy. S4. Output the optimal task scheduling strategy and apply it to actual project execution to guide task allocation, resource scheduling, and progress control; S5. Collect data generated during project execution in real time as feedback information, and update the task cause-and-effect diagram accordingly to achieve closed-loop control of task scheduling and project status, and continuously optimize task scheduling strategies to adapt to dynamic changes in the project.

2. The task scheduling optimization and feedback control method based on the causal graph structure according to claim 1 is characterized in that: Execute step S1, use natural language processing and knowledge extraction technology to identify three types of entities: task nodes, event nodes, and personnel nodes, and construct directed edge connection nodes based on the dependency, influence, and collaboration relationships between entities.

3. The task scheduling optimization and feedback control method based on the causal graph structure according to claim 2 is characterized in that: When executing step S1, the edge weight is determined by the following operations: Comprehensively analyze historical task data to extract the correlation strength and impact frequency between past tasks; conduct in-depth research on project documents to obtain the logical relationships and constraints between tasks; carefully analyze task descriptions to explore task characteristics and potential influencing factors; Track task progress in real time and capture dynamic correlations during the execution of the current task; Based on the above data, edge weights are calculated using preset algorithms or rules. Specifically, for dependent edges, the edge weight is calculated based on the probability of the previous task in the history of tasks affecting the on-time completion of the subsequent task. The higher the probability of influence, the greater the edge weight. For impact edges, edge weights are assigned based on the degree of impact on task duration, cost, and quality when the event occurs. For collaboration edges, the degree of collaboration between personnel and their historical performance are used to determine edge weights. The calculated edge weight intuitively reflects the strength of the causal relationship between tasks, the direction of influence, and the possibility of risk transmission. The larger the value, the stronger the correlation or influence.

4. The task scheduling optimization and feedback control method based on the causal graph structure according to claim 2 is characterized in that: Executing step S3, the controllable variables include the order of task execution, the assignment of task leaders, the amount of resource input, and whether the tasks can be executed in parallel; Construct a scheduling optimization objective function that includes three goals: minimizing the overall delay risk, maintaining the stability of the critical path, and optimizing resource load balancing, and set weights for each goal.

5. The task scheduling optimization and feedback control method based on the causal graph structure according to claim 2 is characterized in that: Executing step S5, updating the task causal graph includes updating node status, adjusting edge weights, modifying dependencies, or adding or deleting nodes / edges to reflect actual changes that occur during task execution.

6. A task scheduling optimization and feedback control system based on a causal graph structure, characterized in that: It includes: The graph construction module integrates multi-source information, uses natural language processing and knowledge extraction techniques to identify entities, constructs directed edges to connect entities based on dependencies, influence, and collaboration between entities, and calculates edge weights based on historical task data to quantify the strength of causal relationships, direction of influence, and likelihood of risk transmission. This creates a structured task causal graph, providing a knowledge foundation for subsequent analysis. The analysis and processing module uses graph neural networks or causal reasoning algorithms to conduct in-depth analysis of task causal graphs, identifying potential risk nodes and their propagation paths in the current project through node embedding learning or causal effect estimation. It combines edge weights and node attributes to quantify the likelihood of risk propagation and the scope of impact, generating a risk heat map that visually displays the clustering areas and potential bottlenecks of systemic risks, providing a decision-making basis for scheduling optimization. The construction and search module is used to mine controllable variables in the task scheduling process based on the identified risk nodes and their propagation paths, construct a weighted scheduling optimization objective function, and use reinforcement learning algorithms, graph optimization algorithms, or heuristic search methods to search the scheduling solution space under constraints. Through multiple rounds of simulation, the impact of different strategies on the task network is evaluated to screen out the current optimal task adjustment strategy. The strategy output module is used to output the optimal task scheduling strategy and apply it to actual project execution to guide task allocation, resource scheduling, and progress control; The feedback update module is used to collect data generated during project execution in real time as feedback information, and update the task cause-effect diagram accordingly, to achieve closed-loop control of task scheduling and project status, and continuously optimize the task scheduling strategy to adapt to dynamic changes in the project.

7. The task scheduling optimization and feedback control system based on the causal graph structure according to claim 6 is characterized in that: The graph construction module specifically includes: Information integration unit, used to integrate multi-source information; The identification and construction unit is used to identify three types of entities: task nodes, event nodes, and personnel nodes based on integrated multi-source information using natural language processing and knowledge extraction technologies. Based on the dependency, influence, and collaboration relationships between entities, directed edges are constructed to connect the nodes. The edge weight calculation unit is used to calculate edge weights through historical task data, quantify the strength of causal relationships, the direction of impact, and the possibility of risk transmission, form a structured task causal graph, and provide a knowledge basis for subsequent analysis.

8. The task scheduling optimization and feedback control system based on the causal graph structure according to claim 7 is characterized in that: The edge weight calculation unit determines the edge weight by the following operations: Comprehensively analyze historical task data to extract the correlation strength and impact frequency between past tasks; conduct in-depth research on project documents to obtain the logical relationships and constraints between tasks; carefully analyze task descriptions to explore task characteristics and potential influencing factors; Track task progress in real time and capture dynamic correlations during the execution of the current task; Based on the above data, edge weights are calculated using preset algorithms or rules. Specifically, for dependent edges, the edge weight is calculated based on the probability of the previous task in the history of tasks affecting the on-time completion of the subsequent task. The higher the probability of influence, the greater the edge weight. For impact edges, edge weights are assigned based on the degree of impact on task duration, cost, and quality when the event occurs. For collaboration edges, the degree of collaboration between personnel and their historical performance are used to determine edge weights. The calculated edge weight intuitively reflects the strength of the causal relationship between tasks, the direction of influence, and the possibility of risk transmission. The larger the value, the stronger the correlation or influence.

9. The task scheduling optimization and feedback control system based on the causal graph structure according to claim 7 is characterized in that: The construction and search module mines the controllable variables in the task scheduling process based on the identified risk nodes and their propagation paths. The controllable variables include the order of task execution, the assignment of task leaders, the amount of resource input, and whether tasks can be executed in parallel. The construction and search module constructs a scheduling optimization objective function including three objectives: minimizing the overall delay risk, maintaining the critical path stability, and optimizing the resource load balance, and sets a weight for each objective.

10. The task scheduling optimization and feedback control system based on the causal graph structure according to claim 7, characterized in that: The feedback update module collects data generated during the project execution as feedback information in real time. The update of the task causal graph includes node status update, edge weight adjustment, dependency modification or node / edge addition and deletion to reflect the actual changes that occur during the task execution.

Images