An intelligent project management method and system based on enterprise digital information

Abstract

The application provides a project intelligent management method and system based on enterprise digital information. First, real-time process data and dynamic image data in the enterprise project are collected, and the latter covers operation records and an initial project framework. Then, personnel action trajectories and equipment states are extracted from the image data to form an operation feature set including operation continuity parameters and abnormal frequency, and the time sequence logical relationship of the process data is analyzed to generate a standard project task sheet containing a task decomposition structure. Then, these information is matched and converted into multi-dimensional management parameters containing dynamic task priority weights, and based on this, a dynamic process topology structure isosmorphic with the standard task sheet is constructed, and a real-time feedback path is embedded in the dynamic process topology structure to trigger version iteration of the standard project task sheet. The technical scheme provided by the application can improve the efficiency and flexibility of project intelligent management.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of project intelligent management, and particularly relates to a project intelligent management method and system based on enterprise digital information. BACKGROUND

[0002] In modern enterprise project management, as the business complexity and scale continue to grow, higher requirements are placed on real-time monitoring and dynamic adjustment capabilities during project execution. In projects involving cross-department collaboration, multi-task parallel processing, and high dependence on equipment operation and personnel coordination, effective decision support based on these data becomes particularly important. This not only requires management of static information such as operation records, task framework, etc. in the project process, but also real-time analysis of personnel action trajectories and equipment operating status to timely identify potential problems or bottlenecks and ensure efficient project advancement.

[0003] Currently, some enterprises have begun to adopt integrated enterprise resource planning systems to realize the digital transformation of project management. Through enterprise resource planning systems, various project-related data can be centrally managed and integrated, including but not limited to financial information, human resource allocation, material management, etc., thereby providing project managers with a comprehensive data view. In addition, some advanced enterprise resource planning systems also integrate data analysis tools that can deeply mine historical project data to assist in predicting future project trends, improving the overall execution and success rate of projects.

[0004] Although enterprise resource planning systems have made significant progress in improving project management efficiency, there are still certain limitations. First, most enterprise resource planning systems focus on post-analysis rather than real-time monitoring, and are slow to react to subtle changes in the project execution process, making it difficult to achieve real-time anomaly detection and rapid response. Second, the data analysis functions of such systems are generally generic, lacking customized analysis models for specific industries or project types, resulting in insufficient depth or accuracy of insight when dealing with complex and specific domain projects. Finally, due to the complexity and high cost of enterprise resource planning systems, small and medium-sized enterprises often find it difficult to bear the deployment and maintenance fees, limiting the popular application of this advanced technology. SUMMARY

[0005] The present application provides a project intelligent management method and system based on enterprise digital information to solve the problems of low efficiency and poor flexibility in project intelligent management in the prior art.

[0006] In a first aspect, the present application provides a project intelligent management method based on enterprise digital information, comprising:

[0007] Acquire real-time process data and dynamic image data generated in enterprise projects. The process data includes operation records of project execution nodes and the initial framework of the project task book.

[0008] The system extracts personnel movement trajectories and equipment operating status from the dynamic image data to generate a set of operation features that includes operation continuity parameters and abnormal action frequency.

[0009] Analyze the temporal logic relationships in the process data, and generate a standard project task book containing a task decomposition structure based on the initial framework of the project task book.

[0010] The temporal logic relationship is matched with the set of operational features, and a project execution feature vector is constructed based on the task decomposition structure. The project execution feature vector is then converted into a multi-dimensional management parameter containing dynamic task priority weights based on the resource allocation parameters and historical project evaluation indicators in the enterprise's digital information.

[0011] Based on the multidimensional management parameters, a dynamic process topology is generated that is isomorphic to the standard project task book. Real-time feedback paths are embedded in the dynamic process topology to trigger version iterations of the standard project task book.

[0012] Optionally, generating a dynamic process topology structure isomorphic to the standard project task book based on the multidimensional management parameters includes:

[0013] The dynamic task priority weight and resource adaptation coefficient in the multidimensional management parameters are respectively bound to the node connection rules and resource channel capacity of the topology node, wherein the node connection rules are associated with the task decomposition structure hierarchy in the standard project task book;

[0014] The initial connection path of the topology node is established based on the task decomposition structure hierarchy, and the path bandwidth is dynamically allocated according to the resource channel capacity to form a basic topology framework.

[0015] A state response mechanism is embedded in the basic topology framework. When the frequency of abnormal actions in the operation records and dynamic image data of the project execution node exceeds the preset tolerance, the connection rules of the topology node are recalculated.

[0016] Based on the temporal logical relationship constraints on the connection direction between nodes, and combined with the results of the recalculation of the connection rules, a dynamic process topology structure is generated that is isomorphic to the standard project task book.

[0017] Optionally, the step of activating the recalculation of the connection rules of the topology node when the frequency of abnormal actions in the operation records and dynamic image data of the project execution node exceeds a preset tolerance includes:

[0018] Real-time acquisition of operation records and dynamic image data of project execution nodes; extraction of event markers from the operation records and action trajectory features from the dynamic image data; alignment within a time window to generate a joint feature vector of events and actions.

[0019] Based on the joint feature vector of the event and action, the projection distance between the abnormal action matching degree index and the preset abnormal pattern library is calculated. When the average projection distance of the continuous window exceeds the preset tolerance threshold, a recalculation instruction is triggered.

[0020] Based on the recalculation instruction, the historical connection rules, resource channel capacity, and node level constraints of the topology nodes are integrated to construct a dynamic optimization target algorithm, and the weight allocation matrix of the node connection rules is iteratively adjusted until the convergence condition is met.

[0021] The adjusted weight allocation matrix is ​​updated to the connection rule engine of the topology node, activating the recalculation of the connection rules of the topology node.

[0022] Optionally, the step of calculating the abnormal action matching degree index based on the joint feature vector of the event and action and the projection distance of the index to the preset abnormal pattern library, and triggering a recalculation instruction when the average projection distance of consecutive windows exceeds a preset tolerance threshold, includes:

[0023] The event and action joint feature vector is decomposed into event marker and action trajectory subspaces, the principal component features of the action trajectory subspace are extracted, and a dynamic mapping matrix aligned with the dimensions of the preset abnormal pattern library is constructed.

[0024] Based on the dynamic mapping matrix, the joint feature vector of the event and action is projected onto the implicit space of the preset abnormal pattern library, the weighted Mahalanobis distance between the projected feature and the preset abnormal pattern library is calculated, and a multi-dimensional projection distance sequence is generated.

[0025] The multi-dimensional projection distance sequence is truncated by a sliding window, and the extreme value distribution and covariance parameter of the window projection distance are extracted. The sliding decay factor of the abnormal action matching degree index is obtained by aligning the window distribution characteristics through dynamic time warping.

[0026] When the weighted mean of the sliding decay factor of the continuous window exceeds the preset tolerance threshold, a recalculation instruction is triggered based on the extreme value distribution and the covariance parameter.

[0027] Optionally, embedding a real-time feedback path in the dynamic process topology to trigger version iteration of the standard project task book includes:

[0028] In the dynamic process topology, identify feedback trigger nodes that are associated with the dynamic task priority weights in the multi-dimensional management parameters, and extract the resource consumption rate and task delay rate of the feedback trigger nodes as real-time status parameters.

[0029] The real-time status parameters are compared with a preset threshold vector to generate the trigger signal strength of the feedback path. Based on the trigger signal strength, a feedback path priority queue is constructed from the feedback trigger node to the version control node of the standard project task book.

[0030] Based on the cumulative signal strength of the feedback path priority queue, the task level to be iterated is located in the task decomposition structure, and the logical dependency strength and historical version difference features are extracted to generate a task decomposition difference matrix.

[0031] Based on the task decomposition difference matrix, the node connection rules in the task decomposition structure are optimized and adjusted, triggering the version iteration of the standard project task book.

[0032] Optionally, the step of locating the task level to be iterated in the task decomposition structure based on the cumulative signal strength of the feedback path priority queue, and extracting the logical dependency strength and historical version difference features to generate a task decomposition difference matrix includes:

[0033] The cumulative signal strength of the feedback path priority queue is decomposed into a time decay factor and a spatial distribution weight, a hierarchical positioning weight allocation matrix is ​​generated, each task level in the task decomposition structure is traversed, and a logical dependency matrix is ​​constructed based on the logical dependency strength.

[0034] Extract the change operation sequence and parameter offset of the task level node from the task decomposition structure of the historical version, map the change operation sequence into an operation type encoding vector, and generate a level difference metric tensor by combining the parameter offset.

[0035] The logical dependency matrix is ​​modally multiplied with the hierarchical difference metric tensor to obtain cross-hierarchical difference propagation features, which are then correlated with the hierarchical localization weight allocation matrix using Hadamard to generate initial components of task decomposition differences.

[0036] Based on the sparsity of the initial components, reverse gradient pruning is performed on the task decomposition difference propagation features to retain the difference components associated with the current feedback signal intensity, thereby generating the task decomposition difference matrix.

[0037] Optionally, the step of parsing the temporal logical relationships in the process data and generating a standard project task book containing a task decomposition structure in conjunction with the initial framework of the project task book includes:

[0038] The project execution node operation records in the process data are parsed into discrete event tag sequences, the temporal density parameter is extracted, and a temporal directed graph structure with the event trigger timestamp interval as the edge weight is constructed.

[0039] Traverse the critical path in the time-series directed graph structure, calculate the overlapping execution probability and resource preemption conflict frequency of event nodes on the critical path, and generate the logical dependency strength matrix between the event nodes.

[0040] The task nodes and event nodes in the initial framework of the project task book are mapped and aligned. The task hierarchy is segmented into subgraphs based on the logical dependency strength matrix, and the task node embedding vectors are extracted.

[0041] Based on the cluster center distribution of the task node embedding vectors and the overlapping execution probability between the subgraphs, an initial hierarchical connection rule is generated. Combined with the dynamic offset in the temporal directed graph structure, the parallelism threshold in the initial hierarchical connection rule is adjusted to generate a standard project task book containing a task decomposition structure.

[0042] Secondly, this application provides a project intelligent management system based on enterprise digital information, including:

[0043] The acquisition module acquires real-time process data and dynamic image data generated in enterprise projects. The process data includes operation records of project execution nodes and the initial framework of the project task book.

[0044] The extraction module extracts personnel movement trajectories and equipment operating status from the dynamic image data, generating a set of operation features including operation continuity parameters and abnormal action frequencies.

[0045] The parsing module parses the temporal logical relationships in the process data and generates a standard project task book containing a task decomposition structure based on the initial framework of the project task book.

[0046] The construction module matches the temporal logical relationship with the set of operational features, constructs a project execution feature vector based on the task decomposition structure, and converts the project execution feature vector into a multi-dimensional management parameter containing dynamic task priority weights based on the resource allocation parameters and historical project evaluation indicators in the enterprise's digital information.

[0047] The generation module generates a dynamic process topology structure that is isomorphic to the standard project task book based on the multidimensional management parameters, and embeds a real-time feedback path in the dynamic process topology structure to trigger version iteration of the standard project task book.

[0048] Thirdly, this application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to realize a project intelligent management method based on enterprise digital information as described in the first aspect above.

[0049] Fourthly, this application provides a computer storage medium storing a computer program, which, when executed by a computer, implements a project intelligent management method based on enterprise digital information as described in the first aspect.

[0050] In this embodiment, real-time process data and dynamic image data generated in enterprise projects are acquired. The process data includes operation records of project execution nodes and an initial framework of the project task book. Personnel movement trajectories and equipment operating status are extracted from the dynamic image data to generate an operation feature set containing operation continuity parameters and abnormal action frequencies. The temporal logical relationships in the process data are parsed, and a standard project task book containing a task decomposition structure is generated in combination with the initial framework of the project task book. The temporal logical relationships are matched with the operation feature set, and a project execution feature vector is constructed based on the task decomposition structure. The project execution feature vector is converted into multi-dimensional management parameters containing dynamic task priority weights based on resource allocation parameters and historical project evaluation indicators in the enterprise's digital information. A dynamic process topology structure isomorphic to the standard project task book is generated based on the multi-dimensional management parameters, and a real-time feedback path is embedded in the dynamic process topology structure to trigger version iteration of the standard project task book.

[0051] The technical solution of this application has the following beneficial effects:

[0052] This application acquires real-time process data and dynamic image data generated in enterprise projects, ensuring comprehensive monitoring of project execution and capturing every detailed change during the project process. The collection of operation records and an initial framework provides foundational data support for subsequent analysis. Personnel movement trajectories and equipment operating status are extracted from the dynamic image data to generate an operational feature set. This step enables detailed tracking of personnel behavior and equipment status, helping to identify potential operational problems or efficiency bottlenecks and adjust workflows based on this information. A standard project task book containing a task decomposition structure is generated by combining the initial project task book framework with the data. This step clarifies the logical relationships between tasks through time sequence analysis, thereby optimizing task allocation and resource planning, and improving the overall project execution. Real-time feedback paths are embedded in the dynamic process topology to trigger version iterations of the standard project task book. The final step is to create a flexible project management model that can automatically adjust and update according to actual execution, ensuring the project always progresses along the optimal path.

[0053] Furthermore, the method for generating a dynamic process topology isomorphic to the standard project task book based on multi-dimensional management parameters includes binding dynamic task priority weights and resource adaptation coefficients to the rules and capacity of topology nodes, establishing preliminary connection paths, and forming a basic framework through dynamic allocation of path bandwidth. On this basis, a status response mechanism is implanted to handle abnormal operations, and the connection direction between nodes is constrained according to temporal logical relationships, thereby generating the dynamic process topology. This method makes project management more intelligent and flexible, not only adjusting resource configuration and task priorities based on real-time data, but also quickly responding to abnormal situations during project execution, effectively improving project execution efficiency and success rate, while promoting continuous optimization and updating of project documentation.

[0054] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 A flowchart of a project intelligent management method based on enterprise digital information provided in this application is shown;

[0057] Figure 2This application provides a schematic diagram of the structure of a project intelligent management system based on enterprise digital information.

[0058] Figure 3 A schematic diagram of the structure of a computing device provided in this application is shown. Detailed Implementation

[0059] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0060] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0061] This solution aims to achieve comprehensive monitoring and analysis of project execution by integrating real-time process data and dynamic image data from enterprise projects. By collecting operation records and initial framework information, it lays the foundation for subsequent analysis of task logic relationships, ensuring accurate capture of project progress.

[0062] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0063] Figure 1 A flowchart of a project intelligent management method based on enterprise digital information is provided in this application embodiment, such as... Figure 1 As shown, the method includes:

[0064] 101. Obtain real-time process data and dynamic image data generated in enterprise projects. The process data includes operation records of project execution nodes and the initial framework of the project task book.

[0065] In this step, the real-time process data generated in the enterprise project refers to the operation records of each node during the project execution process, including but not limited to information such as start time, end time, and operation content.

[0066] Dynamic image data refers to video streams or image sequences captured by cameras or other monitoring equipment, used to reflect the actual situation on site.

[0067] The initial framework of the project brief is a task outline defined at the beginning of the project, which includes the overall goals, expected outcomes and preliminary task breakdown of the project.

[0068] In this embodiment, firstly, real-time data streams are collected using sensors and cameras deployed in various work areas. This data, after preprocessing, is divided into two categories: process data and dynamic image data. Process data is captured from various management systems using log collection tools; while dynamic image data is analyzed using image recognition technology to extract keyframes and scene descriptions. Finally, these two types of data are integrated with the initial framework of the project task book, providing foundational data support for subsequent steps.

[0069] In a manufacturing company, every machine on the production line is equipped with sensors to monitor its operating status and upload data to a central server. Meanwhile, cameras installed in the workshop record employees' work, and video analytics software extracts their movement trajectories. The project manager developed a preliminary project brief based on project requirements, clearly defining the various performance indicators for producing the new product. As the production line operates, real-time data continuously flows into the system, providing rich material for subsequent analysis.

[0070] 102. Extract personnel movement trajectories and equipment operating status from the dynamic image data to generate a set of operation features including operation continuity parameters and abnormal action frequencies;

[0071] In this step, the set of operational features includes operational coherence parameters and the frequency of abnormal actions derived from dynamic image data analysis.

[0072] Operational coherence parameters measure the continuity and smoothness of the operational process.

[0073] Abnormal action frequency refers to the frequency of actions that deviate from pre-defined standard operating procedures within the operational process. These actions may include incorrect operating steps, non-standard work behaviors, or any behavior that deviates from the normal operating procedure. By statistically analyzing the frequency of abnormal actions, we can assess the efficiency and quality of operators' work, and also identify operational aspects that need improvement.

[0074] In this embodiment, computer vision algorithms are first used to perform deep analysis on dynamic image data to identify the specific actions and trajectories of personnel. Next, the actual operation is compared with a pre-defined standard operating procedure template to calculate the operational consistency parameters. Simultaneously, any behavior deviating from the normal behavioral pattern is marked as an abnormal action, and the frequency of its occurrence is counted. Finally, a comprehensive set of operational features is generated.

[0075] In the aforementioned manufacturing company case, by analyzing worker operation videos, the system can accurately track every action of each worker on the assembly line and assess whether these actions conform to predetermined operating procedures. For example, if a worker frequently needs to adjust the position of parts, this behavior will be recorded as an abnormal action, thus affecting the overall operational consistency score.

[0076] 103. Analyze the temporal logic relationships in the process data, and generate a standard project task book containing a task decomposition structure based on the initial framework of the project task book;

[0077] In this step, the temporal logic describes the temporal order and dependencies between tasks during project execution. This includes which tasks must be completed before other tasks can begin, as well as the expected start and end times for each task.

[0078] Task breakdown structure (TBS) is a method of breaking down a project into smaller, more manageable parts in a hierarchical structure. It not only clarifies the overall goals of the project but also details all the work tasks required to achieve those goals.

[0079] A standard project task statement is a document developed based on the temporal logical relationships in the analyzed process data and combined with the initial framework of the project task statement. It details the task decomposition structure of the project, that is, the result of dividing all the work tasks of the project according to a hierarchical structure.

[0080] In this embodiment, firstly, natural language processing technology is used to parse the text information in the process data to determine the sequence and dependencies between different tasks. Then, based on the initial framework of the project task book, a clear task decomposition structure is constructed to ensure that each subtask has a clear objective and deliverables. The final document not only covers all task details of the project but also reflects the temporal sequence and interrelationships between tasks.

[0081] Continuing with the example of a manufacturing company, after clarifying the operational procedures on the production line, the system automatically generates a detailed project task sheet. This document not only lists all the necessary production steps but also indicates which steps must be completed before proceeding to the next stage, thus helping the team better plan resources and time.

[0082] 104. Match the temporal logical relationship with the set of operational features, construct a project execution feature vector based on the task decomposition structure, and convert the project execution feature vector into a multi-dimensional management parameter containing dynamic task priority weights according to the resource allocation parameters and historical project evaluation indicators in the enterprise digital information.

[0083] In this step, the project execution feature vector is a mathematical model used to represent key features in the project execution process, including but not limited to operational consistency parameters and the frequency of abnormal actions.

[0084] Historical project evaluation indicators refer to various evaluation criteria and results of projects that have been completed in the past, such as cost control, adherence to schedules, and quality compliance rates.

[0085] Multidimensional management parameters are a set of values ​​derived from the project execution feature vector. They take into account factors such as resource allocation parameters and historical project evaluation indicators, and are used to quantify the dynamic priority weight of each task during project execution.

[0086] In this embodiment, firstly, the previously obtained set of operational features is combined with temporal logical relationships to form a project execution feature vector. Then, using mathematical models and optimization algorithms, based on the company's internal resource allocation and evaluation criteria for similar past projects, the dimensions of the feature vector are adjusted to transform it into multi-dimensional management parameters with practical guiding significance.

[0087] In the context of manufacturing enterprises, based on the data accumulated in previous stages, the system can calculate the importance and urgency of each production task. For example, when it is found that a certain process frequently causes production line stoppages, the system will automatically increase the weight of that process in the multi-dimensional management parameters, reminding management to pay attention to and resolve the issue.

[0088] 105. Generate a dynamic process topology structure isomorphic to the standard project task book based on the multidimensional management parameters, and embed a real-time feedback path in the dynamic process topology structure to trigger version iteration of the standard project task book.

[0089] In this step, the dynamic process topology is a visual representation method used to show the various activities and their interrelationships during project execution. It embeds real-time feedback paths, allowing the standard project task book to be updated according to the actual situation, thus promoting version iteration.

[0090] Real-time feedback paths refer to mechanisms embedded in dynamic process topologies that allow the system to automatically update project status based on the latest data input and provide immediate feedback.

[0091] In this embodiment, firstly, using a graphical modeling tool, multi-dimensional management parameters are transformed into a visual flowchart. Based on this, a real-time feedback mechanism is added, allowing the flowchart content to be adjusted immediately whenever new data is input or external conditions change, triggering updates to the corresponding parts and ensuring the flexibility and adaptability of the entire project management system.

[0092] For manufacturing companies, once the system detects equipment malfunctions or raw material shortages, it can quickly identify these issues in the dynamic process topology and suggest corresponding solutions. This not only helps resolve problems quickly but also prompts timely updates to project briefs, ensuring projects proceed as planned.

[0093] In summary, steps 101 to 105, by integrating multiple advanced technologies, achieve full-process coverage of enterprise projects, from data collection, feature extraction, task decomposition, priority setting to dynamic management. This method improves the efficiency and accuracy of project management, enhances the ability to respond to emergencies, promotes the effective allocation and utilization of resources, and provides a solid guarantee for the efficient operation of enterprises.

[0094] To further improve the flexibility and responsiveness of project management, this solution focuses on constructing a dynamic process topology isomorphic to the standard project task book using multi-dimensional management parameters. By binding priority weights and resource adaptation coefficients to node rules and capacities, an initial connection path is established, and path bandwidth is dynamically adjusted based on resource channel capacity, forming a highly adaptable basic framework. In some embodiments, step 105, which involves generating a dynamic process topology isomorphic to the standard project task book based on the multi-dimensional management parameters, includes:

[0095] 201. Bind the dynamic task priority weight and resource adaptation coefficient in the multidimensional management parameters to the node connection rules and resource channel capacity of the topology node, respectively, wherein the node connection rules are associated with the task decomposition structure hierarchy in the standard project task book;

[0096] In step 201, the dynamic task priority weight in the multi-dimensional management parameters is used to quantify the importance and urgency of each task during project execution, while the resource adaptation coefficient reflects the adjustment ratio of task allocation based on the current resource status. Node connection rules are associated with the task decomposition structure hierarchy in the standard project task book, defining the logical relationships and dependencies between different tasks. Resource channel capacity represents the amount of data that can be transmitted or the number of tasks that can be completed on a specific path, ensuring the effective utilization of resources.

[0097] In this embodiment, firstly, the dynamic task priority weight of each task is determined by analyzing multi-dimensional management parameters, and the resource adaptation coefficient is adjusted accordingly. Next, these weights and coefficients are bound to the node connection rules and resource channel capacity of the topology nodes. Then, graph theory algorithms are used to establish logical connections between tasks, ensuring that the task decomposition structure hierarchy is accurately reflected. Finally, by integrating the above information, a preliminary topology node configuration is formed, providing a foundation for subsequent steps.

[0098] 202. Based on the task decomposition structure hierarchy, establish the initial connection path of the topology node, and dynamically allocate path bandwidth according to the resource channel capacity to form a basic topology framework;

[0099] In step 202, the task decomposition structure hierarchy is the result of dividing project tasks according to their importance and execution order, helping to identify which tasks need to be completed first. The initial connection path is established based on this hierarchical relationship, while path bandwidth refers to the amount of data or tasks that can pass through a certain path within a certain time. Reasonable allocation of path bandwidth ensures efficient resource flow and smooth task execution.

[0100] In this embodiment, firstly, initial connection paths between topology nodes are established according to the task decomposition structure hierarchy. Next, a flow control algorithm is used to dynamically allocate path bandwidth based on resource channel capacity, ensuring that each node has sufficient resources to support its operation. Then, the entire network is checked for bottlenecks or redundancy, and the path design is optimized. Finally, a stable basic topology framework is built, providing a platform for real-time adjustments.

[0101] 203. An embedded state response mechanism is incorporated into the basic topology framework. When the frequency of abnormal actions in the operation records and dynamic image data of the project execution node exceeds the preset tolerance, the connection rules of the topology node are recalculated.

[0102] In step 203, the status response mechanism is a mechanism that monitors the system's operational status and automatically triggers corresponding measures. The preset tolerance refers to the allowable range of operational errors; exceeding this range is considered an anomaly. Connection rule recalculation refers to reassessing and adjusting the relationships between nodes to adapt to the new situation when an anomaly is detected. The preset tolerance refers to the maximum range of allowed operational errors or abnormal behavior; exceeding this range is considered an anomaly, requiring intervention and adjustment. Resource channel capacity plays a crucial role in this process, defining the amount of data that can be transmitted or the number of tasks that can be completed on a specific path, ensuring effective resource utilization and smooth task execution.

[0103] In this embodiment, firstly, a state response mechanism is embedded in the basic topology framework, setting a preset tolerance as the judgment criterion. Next, the frequency of abnormal actions in the operation records and dynamic image data of the project execution nodes is monitored in real time. Then, once a situation exceeding the preset tolerance is detected, such as frequent erroneous operations in a task leading to inefficiency, the connection rule recalculation process is immediately activated. At this time, the system re-evaluates the relationships between the affected nodes and their associated nodes based on the latest data, and adjusts the connection rules between nodes according to the current resource channel capacity. Finally, the dynamic process topology is updated to reflect these changes, ensuring the system's flexibility and adaptability while maintaining the overall project's stability and efficiency.

[0104] 204. Based on the temporal logical relationship constraints on the connection direction between nodes, and combined with the result of the recalculation of the connection rules, a dynamic process topology structure isomorphic to the standard project task book is generated.

[0105] In step 204, the temporal logic relationship is used to describe the temporal order and dependencies between tasks during project execution, ensuring that tasks are performed in the correct order. Connection rule recalculation refers to the process of re-evaluating and adjusting the connection rules between nodes when abnormal operations or exceeding preset tolerances are detected. The dynamic process topology is a visual representation used to show various activities and their interrelationships during project execution, allowing for rapid adjustments to the project plan based on actual conditions.

[0106] In this embodiment, firstly, based on the established basic topology framework, time series analysis technology is used to determine the temporal logical relationship of each task, ensuring that the order of tasks is accurately reflected. Next, combining the results of recalculating connection rules, graph theory algorithms are used to adjust the connection directions between nodes to conform to the actual operational flow and logical dependencies. Then, optimization algorithms are used to further optimize the entire network, ensuring reasonable allocation of resource channel capacity and path bandwidth to avoid bottlenecks. Finally, all information is integrated to generate a dynamic process topology structure isomorphic to the standard project task book, making project management more flexible and efficient, enabling real-time response to changes and rapid adjustments.

[0107] Here is a specific example:

[0108] On a manufacturing production line, real-time data is collected through integrated sensors and cameras. First, based on preliminary data analysis, the dynamic task priority weights and resource adaptation coefficients for each process are determined and applied to each workstation (node) on the production line. Next, connection paths between workstations are planned according to the task decomposition structure hierarchy, and bandwidth is dynamically adjusted based on resource requirements. Then, a status response mechanism is set up to monitor the production line's operating status. When a process experiences frequent failures, the system automatically triggers a recalculation of connection rules, rearranging the relationship between that process and other processes, optimizing resource allocation, and ensuring efficient production line operation.

[0109] In summary, steps 201 to 204, by dynamically adjusting task priorities and resource allocation during project execution, achieve precise management and rapid response to complex projects. This not only effectively avoids resource waste but also allows for swift handling of unforeseen circumstances, significantly improving project success rates and efficiency, and providing strong support for enterprises to achieve more efficient operational management.

[0110] To further improve the response speed and accuracy to anomalies during project execution, this solution further refines the recalculation mechanism triggered when the frequency of abnormal actions exceeds the tolerance. By extracting event markers from operation records and motion trajectory features from dynamic image data in real time, a joint feature vector is generated within a time window. Based on this, an anomaly matching degree index is calculated. Once it exceeds a preset threshold, a recalculation instruction is initiated to optimize node connection rules. In some embodiments, step 203, which involves activating the recalculation of the connection rules of the topology node when the frequency of abnormal actions in the operation records and dynamic image data of the project execution node exceeds the preset tolerance, includes:

[0111] 301. Real-time acquisition of operation records and dynamic image data of project execution nodes, extraction of event markers in the operation records and action trajectory features in the dynamic image data, and alignment within the time window to generate a joint feature vector of events and actions;

[0112] In step 301, event markers are key information points extracted from operation records to identify changes in specific operations or states. Motion trajectory features refer to the movement paths and patterns of personnel or equipment extracted from dynamic image data. A time-series window is a division of a series of consecutive time periods to align different data streams within the same time frame. The joint event and motion feature vector is a comprehensive description generated by combining event markers and motion trajectory features within the same time-series window.

[0113] In this embodiment, firstly, operation records and dynamic image data of project execution nodes are acquired through a real-time monitoring system. Next, data mining techniques are used to extract event markers from the operation records, and computer vision algorithms are used to extract motion trajectory features from the dynamic image data. Then, these features are aligned within the same time window to form a joint feature vector of events and actions. Finally, these vectors are used as input to provide basic data support for subsequent steps.

[0114] 302. Calculate the abnormal action matching degree index and the projection distance between the event and action joint feature vector and the preset abnormal pattern library. When the average projection distance of the continuous windows exceeds the preset tolerance threshold, trigger the recalculation instruction.

[0115] In step 302, the abnormal action matching index is a numerical value used to measure the similarity between the current operation and a preset abnormal pattern. The preset abnormal pattern library is a set of predefined abnormal behavior patterns used to compare and identify abnormal situations in actual operations. The projection distance refers to the difference between the joint feature vector of the event and action and the preset abnormal pattern library. When the average projection distance of a continuous window exceeds a preset tolerance threshold, it indicates that the current operation has a significant abnormality and a recalculation instruction needs to be triggered.

[0116] In this embodiment, firstly, the projection distance between the event and action joint feature vector and a preset abnormal pattern library is calculated, and machine learning algorithms such as support vector machines or neural networks are used to evaluate the abnormal action matching degree index. Next, the average projection distance within multiple consecutive time windows is continuously monitored. Once the average value exceeds a preset tolerance threshold, a recalculation instruction is immediately triggered. This process ensures timely detection and response to abnormal situations.

[0117] 303. Based on the recalculation instruction, integrate the historical connection rules, resource channel capacity, and node level constraints of the topology nodes to construct a dynamic optimization target algorithm, and iteratively adjust the weight allocation matrix of the node connection rules until the convergence condition is met;

[0118] In step 303, historical connection rules record the connection patterns and dependencies between topology nodes over a past period. Resource channel capacity reflects the amount of data that can be transmitted or the number of tasks that can be completed on a specific path. Node hierarchy constraints refer to the hierarchical relationships and logical order between nodes determined according to the task decomposition structure. The dynamic optimization objective algorithm is a mathematical model used to adjust the weight allocation matrix of node connection rules to meet a specific optimization objective. Convergence condition refers to the state where the algorithm no longer changes significantly after iterative updates to a certain extent.

[0119] In this embodiment, firstly, upon receiving a recalculation instruction, the historical connection rules of the topology nodes, resource channel capacity, and node hierarchical constraints are integrated to construct a dynamic optimization objective algorithm. Next, optimization algorithms such as genetic algorithms or gradient descent are used to iteratively adjust the weight allocation matrix of the node connection rules. Then, the adjusted results are continuously checked to ensure they meet the convergence conditions until the optimal solution is reached. Finally, a new set of weight allocation matrices is obtained, providing parameter support for subsequent steps.

[0120] 304. Update the adjusted weight allocation matrix to the connection rule engine of the topology node and activate the recalculation of the connection rules of the topology node.

[0121] In step 304, the weight allocation matrix is ​​a set of values ​​representing the importance or priority of connections between nodes. The connection rule engine is a processing module responsible for updating the connection rules of the topology nodes based on the weight allocation matrix. By updating the connection rule engine with the adjusted weight allocation matrix, the recalculation of the connection rules of the topology nodes can be activated, enabling dynamic adjustments to the project execution flow.

[0122] In this embodiment, the adjusted weight allocation matrix is ​​first imported into the connection rule engine. Then, the connection rule engine recalculates the connection rules between nodes based on the new weight allocation matrix, ensuring that all nodes operate according to the latest resource configuration and logical order. Next, the updated topology is applied to the actual production environment, enabling the entire system to quickly adapt to changes and maintain efficient operation. Ultimately, this achieves effective management and rapid response to anomalies.

[0123] Here is a specific example:

[0124] On a manufacturing production line, a real-time monitoring system collects operation records and video footage of employee actions. During a certain process, frequent downtime due to equipment malfunction exceeded the preset tolerance. The system first extracts event markers and action trajectory features, generating a joint feature vector within the same time window. Next, it calculates the projection distance between these features and a preset abnormal pattern library, triggering a recalculation instruction. Subsequently, it integrates historical connection rules, resource channel capacity, and node hierarchy constraints, and uses a genetic algorithm to iteratively adjust the weight allocation matrix until convergence. Finally, the adjusted matrix is ​​updated to the connection rule engine, recalculating the connection rules between nodes, thus resolving the problems caused by equipment malfunction and improving the overall efficiency of the production line.

[0125] In summary, steps 301 to 304, through real-time monitoring, data analysis, and dynamic adjustments, enable rapid response and precise management of anomalies during the execution of complex projects. This not only improves project success rates and efficiency but also enhances the system's flexibility and adaptability, providing strong support for enterprises to achieve more efficient operational management. In particular, the accurate identification and timely intervention of abnormal actions significantly reduce potential risks and ensure the smooth progress of projects.

[0126] To further improve the accuracy and response speed of abnormal action identification, the solution delves into specific methods for accurately calculating the abnormal matching degree index based on the joint feature vector of events and actions. This includes decomposing the feature vector, extracting principal component features, constructing a dynamic mapping matrix, evaluating abnormal behavior through a projection distance sequence, and triggering recalculation when the average decay factor of a continuous window exceeds a preset threshold, ensuring project execution quality. In some embodiments, step 302, which involves calculating the projection distance between the abnormal action matching degree index and a preset abnormal pattern library based on the joint feature vector of events and actions, and triggering a recalculation instruction when the average projection distance of a continuous window exceeds a preset tolerance threshold, includes:

[0127] 401. Decompose the event and action joint feature vector into event marker and action trajectory subspaces, extract the principal component features of the action trajectory subspace, and construct a dynamic mapping matrix aligned with the dimensions of the preset abnormal pattern library;

[0128] In step 401, event markers are key information points extracted from operation records to identify changes in specific operations or states. The action trajectory subspace refers to the feature space composed of action paths and patterns extracted from dynamic image data. Principal component features are the main feature vectors obtained through dimensionality reduction techniques, representing the main directions of change in the original data. The dynamic mapping matrix is ​​a transformation matrix used to map the joint feature vector of events and actions to a space with dimensions consistent with a preset abnormal pattern library for comparison.

[0129] In this embodiment, firstly, the joint feature vector of events and actions is decomposed into event markers and action trajectory subspaces. Next, principal component analysis (PCA) is used to extract principal component features from the action trajectory subspace. Then, a dynamic mapping matrix aligned with the dimensions of a pre-defined anomaly pattern library is constructed based on these principal component features. Finally, this matrix is ​​used to transform the original feature vectors into a form suitable for subsequent analysis, providing prepared data for later steps.

[0130] 402. Based on the dynamic mapping matrix, project the joint feature vector of the event and action onto the implicit space of the preset abnormal pattern library, calculate the weighted Mahalanobis distance between the projected features and the preset abnormal pattern library, and generate a multi-dimensional projection distance sequence.

[0131] In step 402, the dynamic mapping matrix is ​​used to transform the feature vectors to a space consistent with a preset anomaly pattern library. The implicit space refers to the feature representation space obtained through some mathematical transformation, typically used to capture the essential features of the data. Weighted Mahalanobis distance is a method for measuring the difference between two distributions, considering the correlation between different dimensions. A multidimensional projected distance sequence is a data sequence composed of projected distances at multiple time points, used to describe feature differences that change over time.

[0132] In this embodiment, firstly, based on a dynamic mapping matrix, the joint feature vector of events and actions is projected onto the implicit space of a preset anomaly pattern library. Next, a weighted Mahalanobis distance algorithm is used to calculate the distance between the projected features and the preset anomaly pattern library, generating a multi-dimensional projected distance sequence. Then, by analyzing the values ​​in these sequences, the similarity between the current operation and the preset anomaly pattern is evaluated. Finally, a set of projected distance sequences reflecting feature differences is obtained, providing basic data support for subsequent steps.

[0133] 403. The multi-dimensional projection distance sequence is truncated by a sliding window, and the extreme value distribution and covariance parameters of the window projection distance are extracted. The sliding decay factor of the abnormal action matching degree index is obtained by aligning the window distribution characteristics through dynamic time warping.

[0134] In step 403, sliding window truncation extracts fixed-length time periods from continuous time series data for local feature analysis. Extreme value distribution refers to statistical characteristics such as the maximum and minimum projected distances within the window. The covariance parameter describes the linear dependence between different variables. The sliding decay factor is a method of adjusting weights to balance the influence of new and old data, enabling the system to adapt to changes more quickly. Dynamic time warping is a technique for time series comparison, capable of aligning time series of different lengths.

[0135] In this embodiment, firstly, a sliding window technique is used to extract the multi-dimensional projected distance sequence, and the extreme value distribution and covariance parameters within each window are extracted. Next, a dynamic time warping method is used to align the window distribution characteristics, ensuring the comparability of data from different time periods. Then, a sliding decay factor is calculated, and its weight is adjusted according to the feature changes within the window. Finally, by integrating these parameters, a sliding decay factor for the abnormal action matching degree index is generated, providing a basis for determining whether a recalculation instruction should be triggered.

[0136] 404. When the weighted mean of the sliding decay factor of the continuous window exceeds the preset tolerance threshold, a recalculation instruction is triggered based on the extreme value distribution and the covariance parameter.

[0137] In step 404, the weighted average of the sliding attenuation factor is the average value of the sliding attenuation factor over a period of time, reflecting the cumulative impact of abnormal actions. The preset tolerance threshold is the allowable range of operational errors; anything exceeding this range is considered abnormal. The recalculation instruction is a process that triggers a re-evaluation of the topology node connection rules when a significant anomaly is detected.

[0138] In this embodiment, firstly, the weighted mean of the sliding decay factor of the continuous window is calculated. Next, this mean is compared with a preset tolerance threshold. If the weighted mean exceeds the threshold, a recalculation instruction is triggered based on the extreme value distribution and covariance parameter. Then, by integrating historical connection rules, resource channel capacity, and node hierarchy constraints, the weight allocation matrix of the node connection rules is iteratively adjusted until the convergence condition is met. Finally, the updated topology is applied to the actual production environment, achieving effective management and rapid response to abnormal situations.

[0139] Here is a specific example:

[0140] On a manufacturing production line, a real-time monitoring system collects operation records and video footage of employee actions. A certain process experienced frequent downtime due to equipment malfunction, exceeding a preset tolerance. The system first decomposes the joint feature vector of events and actions, extracts the principal component features of the action trajectory subspace, and constructs a dynamic mapping matrix. Next, it uses this matrix to project the feature vectors onto the implicit space of a preset abnormal pattern library, calculating the projection distance. These distances are analyzed using a sliding window technique to calculate the sliding attenuation factor. When the weighted average of the sliding attenuation factors across consecutive windows exceeds a preset tolerance threshold, the connection rules between nodes are adjusted, resolving the problems caused by equipment malfunctions and improving the overall efficiency of the production line.

[0141] In summary, steps 401 to 404, by accurately identifying and promptly intervening in abnormal actions, improved the stability and efficiency of project execution. This not only enhanced the system's flexibility and adaptability but also reduced potential risks, ensuring the smooth progress of the project. In particular, the precise analysis and dynamic adjustment mechanism for abnormal actions significantly improved the project's success rate and management level, providing strong support for enterprises to achieve more efficient operational management.

[0142] To further improve the adaptability and flexibility of standard project task specifications to actual project progress, this solution describes a method for embedding real-time feedback paths in a dynamic process topology. This involves identifying feedback trigger nodes associated with dynamic task priorities, comparing real-time status parameters with preset thresholds, constructing a feedback path priority queue, locating the task level requiring iteration based on accumulated signal strength, optimizing node connection rules in the task decomposition structure, and promoting version iteration. In some embodiments, step 105, embedding real-time feedback paths in the dynamic process topology to trigger version iteration of the standard project task specification, includes:

[0143] 501. Identify feedback triggering nodes in the dynamic process topology that are associated with the dynamic task priority weight in the multi-dimensional management parameters, and extract the resource consumption rate and task delay rate of the feedback triggering nodes as real-time status parameters.

[0144] In step 501, the feedback trigger node refers to a key node in the dynamic process topology that is associated with the dynamic task priority weight in the multi-dimensional management parameters, used to monitor and evaluate the execution status of the current project. Resource consumption rate represents the amount of resources required to complete a specific task and is an important indicator for measuring task execution efficiency. Task delay rate refers to the difference between the actual completion time and the planned completion time of a task, reflecting the time deviation during task execution. Real-time status parameters, including resource consumption rate and task delay rate, are used to reflect the current operating status of the feedback trigger node.

[0145] In this embodiment, firstly, feedback trigger nodes associated with the dynamic task priority weights in the multi-dimensional management parameters are identified within the dynamic process topology. Next, resource consumption rate and task latency rate data for these nodes are collected through sensors and the management system. Then, data analysis techniques are used to extract these real-time status parameters, providing foundational data support for subsequent steps. Finally, based on these parameters, the actual operational status of each key node can be accurately understood, providing a basis for adjusting the project plan.

[0146] 502. Compare the real-time status parameters with a preset threshold vector to generate the trigger signal strength of the feedback path, and construct a feedback path priority queue from the feedback trigger node to the version control node of the standard project task book based on the trigger signal strength.

[0147] In step 502, the threshold vector is a set of preset standard values ​​used to determine whether the real-time status parameters exceed the normal range. The trigger signal strength represents the signal strength generated based on the comparison between the real-time status parameters and the threshold vector, used to determine whether the feedback path needs to be triggered. The feedback path priority queue is a task list sorted by trigger signal strength, used to guide the data flow direction from the feedback trigger node to the version control node of the standard project task book.

[0148] In this embodiment, firstly, the real-time status parameters are compared with a preset threshold vector, and the trigger signal strength is calculated using statistical analysis methods. Next, a feedback path priority queue is constructed based on the trigger signal strength, from the feedback trigger node to the version control node of the standard project task book. Then, by comparing the trigger signal strengths of different nodes, it is determined which nodes require priority processing. Finally, an ordered task list is formed to ensure that high-priority issues are resolved promptly, thereby optimizing the entire project execution process.

[0149] 503. Based on the cumulative signal strength of the feedback path priority queue, locate the task level to be iterated in the task decomposition structure, extract the logical dependency strength and historical version difference features, and generate a task decomposition difference matrix.

[0150] In step 503, the cumulative signal strength is the cumulative value of the trigger signal strength over a period of time, reflecting the severity and duration of the problem. Logical dependency strength describes the interdependencies between tasks at different levels of the task decomposition structure. Historical version difference characteristics refer to the main changes or improvements between new and old versions. The task decomposition difference matrix is ​​a matrix dataset used to record and analyze changes in the task decomposition structure across different versions.

[0151] In this embodiment, firstly, the task level requiring iteration is located in the task decomposition structure based on the cumulative signal strength of the feedback path priority queue. Next, the logical dependency strength and historical version difference features within that level are extracted to generate a task decomposition difference matrix. Then, data analysis techniques are used to compare the differences between the old and new versions to identify the parts that need adjustment. Finally, by integrating this information, the specific content that needs to be updated is clarified, providing detailed operational guidelines for subsequent steps.

[0152] 504. Based on the task decomposition difference matrix, optimize and adjust the node connection rules in the task decomposition structure to trigger the version iteration of the standard project task book.

[0153] In step 504, the task decomposition difference matrix provides specific information about changes in the task decomposition structure, guiding optimization and adjustments. Node connection rules refer to the set of rules defining the logical relationships and dependencies between tasks. By optimizing and adjusting the node connection rules in the task decomposition structure, version iterations of the standard project task book can be achieved, ensuring that it always remains consistent with the actual project progress.

[0154] In this embodiment, firstly, based on the task decomposition difference matrix, the specific node connection rules that need optimization and adjustment are analyzed and determined. Next, a graph theory algorithm is used to re-plan the connection methods between tasks, ensuring that the new connection rules both meet current needs and maintain the stability of the overall structure. Then, the relevant parts of the standard project task book are updated, triggering a version iteration. Finally, the effectiveness of the new version is verified through the system to ensure that all adjustments meet the expected goals and effectively improve the project's management level.

[0155] Here is a specific example:

[0156] On a manufacturing production line, real-time data is collected through integrated sensors and cameras. A certain process experienced frequent downtime due to equipment malfunction, exceeding a preset tolerance. The system first identifies the feedback trigger nodes related to this process within the dynamic process topology, extracting resource consumption rate and task latency rate as real-time status parameters. Next, these parameters are compared with a preset threshold vector, and a feedback path priority queue is constructed accordingly. The system locates the task level requiring iteration within the task decomposition structure, generating a task decomposition difference matrix. Finally, based on this matrix, the node connection rules in the task decomposition structure are optimized and adjusted, triggering a version iteration of the standard project task book, resolving the problems caused by equipment malfunction and improving the overall efficiency of the production line.

[0157] In summary, steps 501 to 504, through real-time monitoring, data analysis, and dynamic adjustments, enable rapid response and precise management of anomalies during the execution of complex projects. This not only improves project success rates and efficiency but also enhances the system's flexibility and adaptability, providing strong support for enterprises to achieve more efficient operational management. In particular, the precise management of feedback paths and version iteration mechanisms significantly reduce potential risks and ensure the smooth progress of projects.

[0158] To further improve the dynamic adjustment capability of the task decomposition structure and the accuracy of version iteration, the solution details the process of locating the task level to be iterated based on the cumulative signal strength of the feedback path priority queue. This includes decomposing the cumulative signal strength into a time decay factor and a spatial distribution weight, traversing each level in the task decomposition structure, and generating a task decomposition difference matrix based on logical dependency strength and historical version differences, thereby achieving accurate task level updates and optimization. In some embodiments, step 503, which involves locating the task level to be iterated in the task decomposition structure based on the cumulative signal strength of the feedback path priority queue and extracting logical dependency strength and historical version difference features to generate a task decomposition difference matrix, includes:

[0159] 601. Decompose the cumulative signal strength of the feedback path priority queue into time decay factor and spatial distribution weight, generate a hierarchical positioning weight allocation matrix, traverse each task level in the task decomposition structure, and construct a logical dependency matrix based on logical dependency strength.

[0160] In step 601, the time decay factor is used to measure the importance of the accumulated signal strength over different time periods, gradually decreasing over time. Spatial distribution weights represent the importance and scope of influence of each task level within the overall project. The hierarchical location weight allocation matrix is ​​a matrix-like dataset used to quantify the importance of each task level, while the logical dependency matrix, constructed based on logical dependency strength, describes the interdependencies between tasks.

[0161] In this embodiment, firstly, the cumulative signal strength of the feedback path priority queue is decomposed into a time decay factor and a spatial distribution weight. Next, a weighted average method is used to generate a hierarchical positioning weight allocation matrix. Then, each task level in the task decomposition structure is traversed, and a logical dependency matrix is ​​constructed based on the logical dependency strength. Finally, by integrating this information, it is determined which task levels need to be prioritized for adjustment, providing basic data support for subsequent steps.

[0162] 602. Extract the change operation sequence and parameter offset of the task level node from the task decomposition structure of the historical version, map the change operation sequence into an operation type encoding vector, and generate a level difference metric tensor by combining the parameter offset.

[0163] In step 602, the change operation sequence refers to the change operations on task-level nodes recorded in the historical version of the task decomposition structure, such as adding, deleting, or modifying. The parameter offset is a numerical value describing the change in parameter values ​​before and after the change. The operation type encoding vector converts various change operations into a numerical encoding form for easy computer processing. The hierarchy difference metric tensor is a multidimensional array used to quantify the degree of difference between task levels.

[0164] In this embodiment, firstly, the change operation sequences and parameter offsets of task-level nodes are extracted from the task decomposition structure of historical versions. Next, these change operation sequences are mapped to operation type encoding vectors, and combined with the parameter offsets to generate a hierarchical difference metric tensor. Then, data analysis techniques are used to compare the differences between the old and new versions to identify the parts that need adjustment. Finally, a detailed hierarchical difference metric tensor is generated to provide specific operational guidelines for subsequent steps.

[0165] 603. Perform modal product operation on the logical dependency matrix and the hierarchical difference metric tensor to obtain cross-hierarchical difference propagation features, and perform Hadamard association with the hierarchical localization weight allocation matrix to generate the initial components of task decomposition differences.

[0166] In step 603, the modal product operation refers to the multiplication operation between two higher-order tensors, used to calculate the cross-level difference propagation features. Hadamard association is an element-level multiplication operation used to multiply two matrices of the same dimension. The initial components of the task decomposition differences are preliminary results obtained through the above operations, reflecting the differences in the task decomposition structure.

[0167] In this embodiment, firstly, the logical dependency matrix and the hierarchical difference metric tensor are modally multiplied to obtain cross-hierarchical difference propagation features. Next, these features are correlated with the hierarchical localization weight allocation matrix using Hadamard correlation to generate initial components of task decomposition differences. Then, these initial components are analyzed using a mathematical model to identify key difference points. Finally, by integrating this information, the specific content requiring updates is clarified, providing detailed operational preparation for subsequent steps.

[0168] 604. Based on the sparsity of the initial components, reverse gradient pruning is performed on the task decomposition difference propagation features to retain the difference components associated with the current feedback signal intensity, thereby generating a task decomposition difference matrix.

[0169] In step 604, sparsity refers to the characteristic that most elements in the dataset are zero. Backward gradient pruning is an optimization algorithm used to remove unimportant difference components and retain those related to the current feedback signal strength. The task decomposition difference matrix is ​​a final generated matrix containing all the task decomposition difference information that needs to be adjusted.

[0170] In this embodiment, firstly, based on the sparsity of the initial components, reverse gradient pruning is performed on the task decomposition difference propagation features to remove unimportant difference components. Next, difference components associated with the current feedback signal strength are retained. Then, the remaining difference components are adjusted using an optimization algorithm to ensure they meet actual requirements. Finally, a task decomposition difference matrix is ​​generated as the final adjustment basis, triggering version iterations of the standard project task book.

[0171] Here is a specific example:

[0172] On a manufacturing production line, a certain process experienced frequent downtime due to equipment failure, exceeding the preset tolerance. The system first decomposes the accumulated signal strength of the feedback path priority queue into a time decay factor and spatial distribution weights, constructing a logical dependency matrix. Next, it extracts the change operation sequences and parameter offsets of task-level nodes, generating a hierarchical difference metric tensor. Then, it performs a modal product operation between the logical dependency matrix and the hierarchical difference metric tensor to generate the initial components of the task decomposition difference. Finally, it performs backward gradient pruning on the task decomposition difference propagation characteristics, generating a task decomposition difference matrix, triggering a version iteration of the standard project task specification, thus resolving the problem caused by the equipment failure and improving the overall efficiency of the production line.

[0173] In summary, steps 601 to 604, by accurately identifying and promptly adjusting discrepancies in the task decomposition structure, improved the project's success rate and efficiency. This not only enhanced the system's flexibility and adaptability but also reduced potential risks, ensuring the project's smooth progress. In particular, the refined management and dynamic adjustment mechanism for task decomposition discrepancies significantly improved project management, providing strong support for enterprises to achieve more efficient operational management.

[0174] To further improve the accuracy and flexibility of project management, this solution focuses on parsing the temporal logical relationships in process data. It generates a logical dependency strength matrix by constructing a temporal directed graph structure and calculating the probability of overlapping execution and the frequency of resource conflicts on the critical path. Combined with the task node mapping alignment in the initial project task book framework, the subgraph is segmented and the parallelism threshold is adjusted, ultimately generating a standard project task book containing a task decomposition structure, thus improving the rationality of project planning. In some embodiments, step 103, parsing the temporal logical relationships in the process data and generating a standard project task book containing a task decomposition structure in conjunction with the initial project task book framework, includes:

[0175] 701. Parse the project execution node operation records in the process data into discrete event tag sequences, extract the temporal density parameter, and construct a temporal directed graph structure with the event trigger timestamp interval as the edge weight;

[0176] In step 701, the discrete event tag sequence is a time series of key operation records extracted from the process data, used to identify changes in specific operations or states. The temporal density parameter measures the time interval between different events, reflecting the compactness of event occurrence. The temporal directed graph structure is a graphical representation method where nodes represent events, edges represent the temporal relationships between events, and edge weights are calculated based on the event trigger timestamp interval.

[0177] In this embodiment, firstly, the project execution node operation records in the process data are parsed into a discrete event tag sequence. Next, time series analysis techniques are used to extract temporal density parameters, and a directed temporal graph structure with event trigger timestamp intervals as edge weights is constructed. Then, by analyzing these edge weights, the relative temporal order between events is determined. Finally, a directed graph structure that accurately reflects the temporal relationship between events is formed, providing basic data support for subsequent steps.

[0178] 702. Traverse the critical path in the time-series directed graph structure, calculate the overlapping execution probability and resource preemption conflict frequency of event nodes on the critical path, and generate the logical dependency strength matrix between the event nodes.

[0179] In step 702, the critical path refers to the series of consecutive events that take the longest time from start to finish, determining the shortest completion time of the project. Overlapping execution probability represents the likelihood of multiple events occurring simultaneously within the same time period. Resource preemption conflict frequency describes the competition for the same resource among different events. The logical dependency strength matrix is ​​a matrix dataset used to quantify the logical dependencies between event nodes.

[0180] In this embodiment, the critical path in the temporal directed graph structure is first traversed to calculate the overlapping execution probability and resource preemption conflict frequency of event nodes on the critical path. Next, a logical dependency strength matrix is ​​generated using statistical analysis methods. Then, by comparing the logical dependency strengths between different event nodes, task nodes requiring special attention are identified. Finally, a detailed logical dependency strength matrix is ​​obtained, providing specific operational guidelines for subsequent steps.

[0181] 703. Map and align the task nodes and event nodes in the initial framework of the project task book, perform subgraph segmentation on the task level based on the logical dependency strength matrix, and extract the task node embedding vector.

[0182] In step 703, mapping alignment refers to matching task nodes and event nodes in the initial framework of the project task book to ensure accurate correspondence between them. Subgraph segmentation is a process of dividing the task hierarchy based on the logical dependency strength matrix, aiming to identify sets of tasks that are independent but closely related. Task node embedding vectors are generated by converting task nodes into vector representations in a high-dimensional space using an algorithm (such as deep learning) to facilitate further analysis.

[0183] In this embodiment, firstly, the task nodes and event nodes in the initial framework of the project task book are mapped and aligned. Next, the task hierarchy is subgraphed based on the logical dependency strength matrix, and task node embedding vectors are extracted. Then, graph neural networks and other techniques are used to embed the task nodes, capturing their intrinsic features. Finally, a set of embedding vectors that accurately reflect the characteristics of the task nodes is formed, providing detailed information support for subsequent steps.

[0184] 704. Based on the cluster center distribution of the task node embedding vectors and the overlapping execution probability between the subgraphs, generate initial hierarchical connection rules. Combined with the dynamic offset in the temporal directed graph structure, adjust the parallelism threshold in the initial hierarchical connection rules to generate a standard project task book containing the task decomposition structure.

[0185] In step 704, the cluster center distribution describes the aggregation of task node embedding vectors in high-dimensional space, reflecting the similarity and differences between task nodes. The overlap execution probability refers to the likelihood that different task nodes will execute simultaneously within the same time period. The parallelism threshold is a set value used to control the degree of parallel execution among task nodes. The standard project task book is the final generated document, containing information about all task decomposition structures.

[0186] In this embodiment, firstly, initial hierarchical connection rules are generated based on the cluster center distribution of task node embedding vectors and the overlapping execution probability between subgraphs. Next, the parallelism threshold in the initial hierarchical connection rules is adjusted by incorporating the dynamic offsets in the temporal directed graph structure. Then, an optimization algorithm ensures a reasonable degree of parallel execution among the task nodes. Finally, a standard project task book containing a task decomposition structure is generated to ensure it always aligns with the actual project progress.

[0187] Here is a specific example:

[0188] On a manufacturing production line, the system first parses the project execution node operation records in the process data into discrete event tag sequences and constructs a temporal directed graph structure with event trigger timestamp intervals as edge weights. Next, it calculates the overlapping execution probability and resource preemption conflict frequency of event nodes on the critical path, generating a logical dependency strength matrix. Then, based on the logical dependency strength matrix, it performs subgraph segmentation on the task hierarchy and extracts task node embedding vectors. Finally, it generates initial hierarchical connection rules and, combined with dynamic offsets in the temporal directed graph structure, generates a standard project task book containing a task decomposition structure, improving the overall efficiency and management level of the production line.

[0189] In summary, steps 701 to 704, by accurately analyzing the temporal logic relationships during project execution and generating a detailed task breakdown structure based on the initial framework of the project task book, improve the project's success rate and efficiency. This not only enhances the system's flexibility and adaptability but also reduces potential risks, ensuring the project's smooth progress. In particular, the refined management and dynamic adjustment mechanism for task nodes significantly improves project management, providing strong support for enterprises to achieve more efficient operational management.

[0190] Figure 2 This application provides a schematic diagram of the structure of a project intelligent management system based on enterprise digital information, as shown in the embodiment of the present application. Figure 2 As shown, the device includes:

[0191] Module 21 acquires real-time process data and dynamic image data generated in the enterprise project. The process data includes operation records of project execution nodes and the initial framework of the project task book.

[0192] Extraction module 22 extracts personnel movement trajectories and equipment operating status from the dynamic image data, and generates a set of operation features including operation continuity parameters and abnormal action frequency;

[0193] Parsing module 23 parses the temporal logic relationship in the process data and generates a standard project task book containing a task decomposition structure in combination with the initial framework of the project task book.

[0194] The construction module 24 matches the temporal logical relationship with the set of operational features, constructs a project execution feature vector based on the task decomposition structure, and converts the project execution feature vector into a multi-dimensional management parameter containing dynamic task priority weights according to the resource allocation parameters and historical project evaluation indicators in the enterprise digital information.

[0195] The generation module 25 generates a dynamic process topology structure that is isomorphic to the standard project task book based on the multi-dimensional management parameters, and embeds a real-time feedback path in the dynamic process topology structure to trigger version iteration of the standard project task book.

[0196] Figure 2 The aforementioned intelligent project management system based on enterprise digital information can execute... Figure 1 The implementation principle and technical effects of the intelligent project management method based on enterprise digital information described in the illustrated embodiment will not be repeated here. The specific methods by which each module and unit of the intelligent project management system based on enterprise digital information in the above embodiments are performed have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0197] In one possible design, Figure 2 The illustrated embodiment of a project intelligent management system based on enterprise digital information can be implemented as a computing device, such as... Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0198] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 32.

[0199] The processing component 32 is used for the above Figure 1 The embodiment describes a project intelligent management method based on enterprise digital information.

[0200] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described method.

[0201] Storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0202] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0203] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0204] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0205] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0206] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The illustrated embodiment presents a project intelligent management method based on enterprise digital information.

[0207] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0208] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0209] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0210] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions 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 this application.

Claims

1. A project intelligent management method based on enterprise digital information, characterized in that, include: Acquire real-time process data and dynamic image data generated in enterprise projects. The process data includes operation records of project execution nodes and the initial framework of the project task book. The system extracts personnel movement trajectories and equipment operating status from the dynamic image data to generate a set of operation features that includes operation continuity parameters and abnormal action frequency. Analyze the temporal logic relationships in the process data, and generate a standard project task book containing a task decomposition structure based on the initial framework of the project task book. The temporal logic relationship is matched with the set of operational features, and a project execution feature vector is constructed based on the task decomposition structure. The project execution feature vector is then converted into a multi-dimensional management parameter containing dynamic task priority weights based on the resource allocation parameters and historical project evaluation indicators in the enterprise's digital information. Based on the multidimensional management parameters, a dynamic process topology is generated that is isomorphic to the standard project task book. Real-time feedback paths are embedded in the dynamic process topology to trigger version iterations of the standard project task book. The generation of a dynamic process topology structure isomorphic to the standard project task book based on the multi-dimensional management parameters includes: The dynamic task priority weight and resource adaptation coefficient in the multidimensional management parameters are respectively bound to the node connection rules and resource channel capacity of the topology node, wherein the node connection rules are associated with the task decomposition structure hierarchy in the standard project task book; The initial connection path of the topology node is established based on the task decomposition structure hierarchy, and the path bandwidth is dynamically allocated according to the resource channel capacity to form a basic topology framework. A state response mechanism is embedded in the basic topology framework. When the frequency of abnormal actions in the operation records and dynamic image data of the project execution node exceeds the preset tolerance, the connection rules of the topology node are recalculated. Based on the temporal logical relationship constraints on the connection direction between nodes, and combined with the results of the recalculation of the connection rules, a dynamic process topology structure is generated that is isomorphic to the standard project task book.

2. The method according to claim 1, characterized in that, When the frequency of abnormal actions in the operation records and dynamic image data of the project execution node exceeds a preset tolerance, the connection rules of the topology node are recalculated, including: Real-time acquisition of operation records and dynamic image data of project execution nodes; extraction of event markers from the operation records and action trajectory features from the dynamic image data; alignment within a time window to generate a joint feature vector of events and actions. Based on the joint feature vector of the event and action, the projection distance between the abnormal action matching degree index and the preset abnormal pattern library is calculated. When the average projection distance of the continuous window exceeds the preset tolerance threshold, a recalculation instruction is triggered. Based on the recalculation instruction, the historical connection rules, resource channel capacity, and node level constraints of the topology nodes are integrated to construct a dynamic optimization target algorithm, and the weight allocation matrix of the node connection rules is iteratively adjusted until the convergence condition is met. The adjusted weight allocation matrix is ​​updated to the connection rule engine of the topology node, activating the recalculation of the connection rules of the topology node.

3. The method according to claim 2, characterized in that, The step of calculating the abnormal action matching degree index based on the joint feature vector of the event and action and the projection distance of the index to the preset abnormal pattern library, and triggering a recalculation instruction when the average projection distance of consecutive windows exceeds a preset tolerance threshold, includes: The event and action joint feature vector is decomposed into event marker and action trajectory subspaces, the principal component features of the action trajectory subspace are extracted, and a dynamic mapping matrix aligned with the dimensions of the preset abnormal pattern library is constructed. Based on the dynamic mapping matrix, the joint feature vector of the event and action is projected onto the implicit space of the preset abnormal pattern library, the weighted Mahalanobis distance between the projected feature and the preset abnormal pattern library is calculated, and a multi-dimensional projection distance sequence is generated. The multi-dimensional projection distance sequence is truncated by a sliding window, and the extreme value distribution and covariance parameter of the window projection distance are extracted. The sliding decay factor of the abnormal action matching degree index is obtained by aligning the window distribution characteristics through dynamic time warping. When the weighted mean of the sliding decay factor of the continuous window exceeds the preset tolerance threshold, a recalculation instruction is triggered based on the extreme value distribution and the covariance parameter.

4. The method according to claim 1, characterized in that, The step of embedding a real-time feedback path in the dynamic process topology to trigger version iterations of the standard project task book includes: In the dynamic process topology, identify feedback trigger nodes that are associated with the dynamic task priority weights in the multi-dimensional management parameters, and extract the resource consumption rate and task delay rate of the feedback trigger nodes as real-time status parameters. The real-time status parameters are compared with a preset threshold vector to generate the trigger signal strength of the feedback path. Based on the trigger signal strength, a feedback path priority queue is constructed from the feedback trigger node to the version control node of the standard project task book. Based on the cumulative signal strength of the feedback path priority queue, the task level to be iterated is located in the task decomposition structure, and the logical dependency strength and historical version difference features are extracted to generate a task decomposition difference matrix. Based on the task decomposition difference matrix, the node connection rules in the task decomposition structure are optimized and adjusted, triggering the version iteration of the standard project task book.

5. The method according to claim 4, characterized in that, The step of locating the task level to be iterated in the task decomposition structure based on the cumulative signal strength of the feedback path priority queue, and extracting the logical dependency strength and historical version difference features to generate a task decomposition difference matrix includes: The cumulative signal strength of the feedback path priority queue is decomposed into a time decay factor and a spatial distribution weight, a hierarchical positioning weight allocation matrix is ​​generated, each task level in the task decomposition structure is traversed, and a logical dependency matrix is ​​constructed based on the logical dependency strength. Extract the change operation sequence and parameter offset of the task level node from the task decomposition structure of the historical version, map the change operation sequence into an operation type encoding vector, and generate a level difference metric tensor by combining the parameter offset. The logical dependency matrix is ​​modally multiplied with the hierarchical difference metric tensor to obtain cross-hierarchical difference propagation features, which are then correlated with the hierarchical localization weight allocation matrix using Hadamard to generate initial components of task decomposition differences. Based on the sparsity of the initial components, reverse gradient pruning is performed on the task decomposition difference propagation features to retain the difference components associated with the current feedback signal intensity, thereby generating the task decomposition difference matrix.

6. The method according to claim 1, characterized in that, The process involves parsing the temporal logic relationships within the process data and generating a standard project task book containing a task decomposition structure based on the initial framework of the project task book, including: The project execution node operation records in the process data are parsed into discrete event tag sequences, the temporal density parameter is extracted, and a temporal directed graph structure with the event trigger timestamp interval as the edge weight is constructed. Traverse the critical path in the time-series directed graph structure, calculate the overlapping execution probability and resource preemption conflict frequency of event nodes on the critical path, and generate the logical dependency strength matrix between the event nodes. The task nodes and event nodes in the initial framework of the project task book are mapped and aligned. The task hierarchy is segmented into subgraphs based on the logical dependency strength matrix, and the task node embedding vectors are extracted. Based on the cluster center distribution of the task node embedding vectors and the overlapping execution probability between the subgraphs, an initial hierarchical connection rule is generated. Combined with the dynamic offset in the temporal directed graph structure, the parallelism threshold in the initial hierarchical connection rule is adjusted to generate a standard project task book containing a task decomposition structure.

7. A project intelligent management system based on enterprise digital information, characterized in that, include: The acquisition module acquires real-time process data and dynamic image data generated in enterprise projects. The process data includes operation records of project execution nodes and the initial framework of the project task book. The extraction module extracts personnel movement trajectories and equipment operating status from the dynamic image data, generating a set of operation features including operation continuity parameters and abnormal action frequencies. The parsing module parses the temporal logical relationships in the process data and generates a standard project task book containing a task decomposition structure based on the initial framework of the project task book. The construction module matches the temporal logical relationship with the set of operational features, constructs a project execution feature vector based on the task decomposition structure, and converts the project execution feature vector into a multi-dimensional management parameter containing dynamic task priority weights based on the resource allocation parameters and historical project evaluation indicators in the enterprise's digital information. The generation module generates a dynamic process topology structure that is isomorphic to the standard project task book based on the multi-dimensional management parameters, and embeds a real-time feedback path in the dynamic process topology structure to trigger version iteration of the standard project task book. The generation of a dynamic process topology structure isomorphic to the standard project task book based on the multi-dimensional management parameters includes: The dynamic task priority weight and resource adaptation coefficient in the multidimensional management parameters are respectively bound to the node connection rules and resource channel capacity of the topology node, wherein the node connection rules are associated with the task decomposition structure hierarchy in the standard project task book; The initial connection path of the topology node is established based on the task decomposition structure hierarchy, and the path bandwidth is dynamically allocated according to the resource channel capacity to form a basic topology framework. A state response mechanism is embedded in the basic topology framework. When the frequency of abnormal actions in the operation records and dynamic image data of the project execution node exceeds the preset tolerance, the connection rules of the topology node are recalculated. Based on the temporal logical relationship constraints on the connection direction between nodes, and combined with the results of the recalculation of the connection rules, a dynamic process topology structure is generated that is isomorphic to the standard project task book.

8. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to realize a project intelligent management method based on enterprise digital information as described in any one of claims 1 to 6.

9. A computer storage medium, characterized in that, The system contains a computer program that, when executed by a computer, implements a project intelligent management method based on enterprise digital information as described in any one of claims 1 to 6.

Images