Intelligent modeling system and method for project task dependency relationship based on knowledge graph

Abstract

The application discloses a project task dependency relationship intelligent modeling system and method based on a knowledge graph, and belongs to the technical field of project task modeling. The method specifically comprises the following steps: acquiring project task image data, and pre-processing the acquired project task image data; constructing a preliminary project task knowledge graph according to the pre-processed project task image data; analyzing direct and indirect dependency relationships between project tasks; updating the preliminary project task knowledge graph to obtain a project task knowledge graph; based on the project task knowledge graph, combining attribute information of the project tasks, optimizing project task scheduling, determining the priority of project task execution and resource allocation, and automatically updating project task nodes and dependency relationships in the knowledge graph based on real-time data and project progress. The priority and resource allocation between project tasks can be processed in real time, so that a project manager can obtain the latest dependency relationships and progress of tasks in real time.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of project task modeling, and specifically relates to an intelligent project task dependency relationship modeling system and method based on a knowledge graph. BACKGROUND

[0002] In traditional software development and project management, modeling and management of task dependency relationships are one of the key factors for project success. With the continuous expansion of project scale and the increase of complexity, how to effectively identify, manage and optimize the dependency relationships between various tasks has become a challenge. Traditional task dependency relationship modeling methods are often based on static task lists or diagrams, such as Gantt charts or network diagrams. These methods, although intuitive, have the following limitations: static, lack of context awareness, and data silos.

[0003] As a graph-based data representation method, a knowledge graph is widely used in natural language processing, semantic analysis, recommendation systems, and other fields. It constructs a directed graph by connecting entities (such as tasks, personnel, resources, etc.) and their relationships (such as dependencies, constraints, priorities, etc.), making complex knowledge more structured and easier to store and reason.

[0004] Although the knowledge graph has the ability to update dynamically, in actual project management, the dependency relationships of tasks may change frequently, such as delays in certain tasks, resource reallocation, priority adjustment, etc., which need to be reflected in the knowledge graph in real time. If the update is not timely or accurate, it may lead to project managers being unable to obtain the latest task dependency information. SUMMARY

[0005] To address the deficiencies of the prior art, the present application proposes an intelligent project task dependency relationship modeling system and method based on a knowledge graph, which realizes automatic identification and modeling of task dependency relationships by constructing and using a task knowledge graph model, and adaptively adjusts the priority and resource allocation of tasks, thereby improving the overall efficiency of project management.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0007] The intelligent project task dependency relationship modeling method based on a knowledge graph comprises:

[0008] Obtain project task image data and preprocess the obtained project task image data;

[0009] Construct a preliminary project task knowledge graph according to the preprocessed project task image data;

[0010] Analyze the direct and indirect dependency relationships between project tasks, update the preliminary project task knowledge graph, and obtain a project task knowledge graph;

[0011] Based on the project task knowledge graph, combined with the attribute information of the project task, the project task scheduling is optimized to determine the priority and resource allocation of the project task execution;

[0012] Based on real-time data and project progress, the project task nodes and dependency relationships in the project task knowledge graph are automatically updated, and when there is a delay or change, the dependency relationships, priority and resource allocation of the project task are automatically adjusted.

[0013] Specifically, the construction of the preliminary project task knowledge graph includes:

[0014] The preprocessed project task image data is analyzed by using natural language processing technology to identify entities in the project task, extract project task dependency relationships and logical relationships between project tasks;

[0015] Based on the analysis results of the preprocessed project task image data, a preliminary project task knowledge graph is constructed, and the specific formula is: , represents the preliminary project task knowledge graph, V represents all nodes in the preliminary project task knowledge graph, including project task nodes and attribute nodes, , n represents the nth project task node, A m represents the mth attribute node, n represents the number of project task nodes, m represents the number of attribute nodes, E represents all edges in the preliminary project task knowledge graph, including relationship edges between project task nodes and attribute nodes and dependency relationship edges between project tasks.

[0016] Specifically, the analysis of direct and indirect dependency relationships between project tasks includes:

[0017] The direct dependency relationship between project tasks is analyzed, and it is set that there is a dependency relationship between project tasks T i and T j If the project task T i must be completed before the project task T j starts, then there is a direct dependency relationship between the project tasks T i and T j , and the project task T i directly depends on the project task T j , and the direct dependency relationship is represented as:

[0018] ,

[0019] wherein, represents the project task T i and Tj a direct dependency relationship between, denotes the end time of the project task T i , denotes the start time of the project task T j , denotes a dependency;

[0020] analyzing the indirect dependency relationship between project tasks, project task T i is indirectly affected by the start of the intermediate project task T j , there is an indirect dependency relationship between project tasks T i and T j , project task T i indirectly depends on project task T j , the indirect dependency relationship between project tasks is represented as:

[0021] ,

[0022] wherein, denotes an indirect dependency relationship between project tasks T i and T j , T k denotes the intermediate task between project tasks T i and T j ;

[0023] analyzing the complex dependency relationship between project tasks, including conditional dependency relationship, parallel dependency relationship and time window dependency relationship;

[0024] using a graph convolution network to embed and reason the nodes in the preliminary project task knowledge graph, learning the dependency relationship between nodes, updating the edges in the preliminary project task knowledge graph according to the learned dependency relationship between nodes, obtaining the project task knowledge graph G.

[0025] Specifically, the conditional dependency relationship, parallel dependency relationship and time window dependency relationship are specifically represented as:

[0026] ,

[0027] ,

[0028] ,

[0029] wherein, denotes that the project task T i is a conditional dependency relationship, C i denotes a condition, denotes the state of the condition C i , True represents true, represents that the project task T i and T j are parallel dependency relationships, represents that the project task T i and T j are executed in parallel, represents that the project task T i and T s are direct dependency relationships and the project task T j and T s are direct dependency relationships, represents that the project task T i and T j are time window dependency relationships, represents a time window.

[0030] Specifically, the project task scheduling is optimized based on the project task knowledge graph and in combination with attribute information of the project task, and the method comprises the following steps:

[0031] Based on the project task knowledge graph, attribute information of each project task is extracted, including task type, execution period, resource category, and a set of predecessor tasks;

[0032] According to the attribute information and dependency relationships of the project task, an ordered linked list is generated according to the predecessor constraints of the project task;

[0033] The ordered linked list is divided into multiple levels, and a task conflict detection subset is established within the same level. The project tasks in each level are cross-compared according to resource occupation, and a scheduling matrix is formed;

[0034] On the basis of the scheduling matrix, the conflicting project tasks are iteratively replaced until the dependency linked list constraints and the resource non-mutual exclusion conditions are met in each level, and a task scheduling sequence is obtained.

[0035] Specifically, on the basis of the scheduling matrix, the conflicting project tasks are iteratively replaced until the dependency linked list constraints and the resource non-mutual exclusion conditions are met in each level, and a task scheduling sequence is obtained, which comprises the following steps:

[0036] In the scheduling matrix, a set of project tasks with resource occupation overlap or time interval intersection within the same level is detected and marked as a conflict group;

[0037] Based on the project tasks in the conflict group, a plurality of candidate task sequences are generated according to the predecessor linked list order and the resource use priority, and each candidate task sequence meets the basic dependency constraints;

[0038] Different combinations of project tasks are selected in turn between the candidate task sequences, and are replaced to the conflict positions in the scheduling matrix to form a new scheduling matrix.

[0039] The new scheduling matrix is subjected to re-conflict detection, and if there are still project tasks that do not meet the constraint of the dependency list or the non-mutual exclusion condition of resources, the replacement operation is continued to iterate until all levels of project tasks meet the constraint, and finally a task scheduling sequence is generated.

[0040] Specifically, the project task image data includes project task basic information, explicit dependencies between project tasks, and project task attributes.

[0041] The preprocessing includes time format unification, converting the time information of the project tasks into a unified format.

[0042] Task classification and labeling, based on the content of the project tasks, classifying the project tasks by domain and task type; redundant information removal, removing irrelevant or redundant data.

[0043] Specifically, the project task dependency relationship intelligent modeling system based on the knowledge graph is used to implement the project task dependency relationship intelligent modeling method based on the knowledge graph, and includes a task acquisition module, a preliminary construction module, a dependency analysis module, a task scheduling optimization module, and a real-time updating module.

[0044] The task acquisition module is configured to acquire project task image data and preprocess the acquired project task image data.

[0045] The preliminary construction module is configured to construct a preliminary project task knowledge graph based on the preprocessed project task image data.

[0046] The dependency analysis module is configured to analyze the direct and indirect dependency relationships between project tasks, update the preliminary project task knowledge graph, and obtain a project task knowledge graph.

[0047] The task scheduling optimization module is configured to optimize project task scheduling based on the project task knowledge graph and in combination with attribute information of the project tasks, and determine the priority of project task execution and resource allocation.

[0048] The real-time updating module is configured to automatically update project task nodes and dependency relationships in the project task knowledge graph based on real-time data and project progress, and automatically adjust the dependency relationships, priority, and resource allocation of the project tasks when there is a delay or change.

[0049] Specifically, the dependency analysis module includes a dependency analysis unit and an updating unit.

[0050] The dependency analysis unit is configured to analyze the direct, indirect, and complex dependency relationships between project tasks.

[0051] The updating unit is configured to learn the dependency relationship between nodes, and update the edges in the preliminary project task knowledge graph according to the learned dependency relationship between nodes.

[0052] Specifically, the task scheduling optimization module comprises a linked list generation unit, a scheduling matrix generation unit and a task scheduling sequence unit.

[0053] The linked list generation unit is configured to generate an ordered linked list according to the attribute information and dependency relationship of the project task, and according to the predecessor constraint.

[0054] The scheduling matrix generation unit is configured to divide the ordered linked list into multiple levels, establish a task conflict detection subset within the same level, and cross-compare the project tasks in each level according to the resource occupation condition to form a scheduling matrix.

[0055] The task scheduling sequence unit is configured to iteratively replace the conflicting project tasks until the dependency linked list constraint and the resource non-mutual exclusion condition are met in each level, and obtain a task scheduling sequence.

[0056] Compared with the prior art, the present application has the following advantages:

[0057] The present application proposes an intelligent modeling method for project task dependency relationship based on a knowledge graph, constructs a preliminary knowledge graph, identifies and updates direct, indirect and complex dependency relationships between tasks, further generates a predecessor constraint linked list in combination with the task attribute information, constructs a hierarchical scheduling matrix, generates a task scheduling sequence meeting the dependency constraint and resource non-mutual exclusion condition by iteratively replacing conflicting tasks, dynamically adjusts the knowledge graph and task allocation in combination with real-time data, and unifies task modeling, dependency relationship reasoning and scheduling optimization in the knowledge graph framework, realizes intelligent analysis of complex dependency relationships and dynamic generation of executable sequences, and thus provides an expandable and efficient scheduling method in a large-scale and complex project environment, and ensures the orderliness of project task execution and the rationality of resource allocation. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 The present application provides an intelligent modeling method for project task dependency relationship based on a knowledge graph;

[0059] Figure 2 The present application provides a project task knowledge graph schematic diagram;

[0060] Figure 3 The present application provides a task scheduling sequence generation flowchart;

[0061] Figure 4 The present application provides an intelligent modeling system architecture diagram for project task dependency relationship based on a knowledge graph. Detailed Implementation

[0062] The present application will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present application, but do not limit the present application in any way. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present application. These all fall within the protection scope of the present application.

[0063] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0064] It should be noted that, unless there is a conflict, the various features in the embodiments of this application can be combined with each other, all of which are within the protection scope of this application. Furthermore, although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than the module division in the device or the order in the flowchart. In addition, the terms "first," "second," and "third" used in this application do not limit the data or execution order, but only distinguish identical or similar items with essentially the same function and effect.

[0065] Unless otherwise defined, all technical and scientific terms used in this specification have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The term "and / or" as used in this specification includes any and all combinations of one or more of the associated listed items.

[0066] Example 1:

[0067] Please see Figures 1-3 The present invention provides an embodiment of an intelligent modeling method for project task dependencies based on knowledge graphs, comprising the following specific steps:

[0068] Step S1: Acquire project task image data and preprocess the acquired project task image data;

[0069] Collecting image data of project tasks from a project management system, a task management tool or a related database, the project task image data including but not limited to the following: project task basic information, explicit dependencies between project tasks and project task attributes; the project task basic information including: task name, description, responsible person, time node, including start time, end time, required resources, including manpower, equipment, materials, etc.; the explicit dependencies between project tasks, for example: the start of a task requires the completion of the previous task (sequential dependency), or the task execution requires that certain conditions be met first (conditional dependency); the project task attributes including: the priority, importance, critical path information, etc. of each task;

[0070] The preprocessing includes: time format unification, converting the time information of project tasks into a unified format, for example, the ISO8601 date standard; task classification and labeling, classifying project tasks according to their content, such as research and development, testing, deployment, etc., and task type, such as requirement analysis, resource scheduling, etc.; and also including redundant information removal, removing irrelevant or redundant data.

[0071] Step S2: constructing a preliminary project task knowledge graph according to the preprocessed project task image data;

[0072] The specific steps of step S2 are:

[0073] Step S201: using natural language processing technology to analyze the preprocessed project task image data, identifying entities in the project tasks, extracting project task dependency relationships and logical relationships between project tasks;

[0074] In this embodiment, entities in the project tasks are identified, such as "project task A", "project manager", "research and development personnel", etc., and corresponding entity nodes are created; project task dependency relationships are extracted, i.e. project task dependency relationships in the project task description, for example, the relationship of "project task B depends on project task A", and dependency edges are automatically generated through relationship extraction algorithms; logical relationships between project tasks, such as sequential relationship, parallel relationship, etc.

[0075] Step S202: based on the analysis results of the preprocessed project task image data, constructing a preliminary project task knowledge graph, the specific formula being: , denotes the preliminary project task knowledge graph, V denotes all nodes in the preliminary project task knowledge graph, including project task nodes and attribute nodes, , T n denotes the nth project task node, A mrepresents the mth attribute node, n represents the number of project task nodes, m represents the number of attribute nodes, E represents all edges in the preliminary project task knowledge graph, including the relationship edges between the project task nodes and the attribute nodes and the dependency edges between the project tasks.

[0076] Step S3: analyzing the direct and indirect dependency relationships between the project tasks, updating the preliminary project task knowledge graph to obtain the project task knowledge graph;

[0077] The specific steps of step S3 are as follows:

[0078] Step S301: analyzing the direct dependency relationship between the project tasks, setting the project task T i and T j as the direct dependency relationship, if the project task T i must be completed before the project task T j starts, the project task T i and T j have a direct dependency relationship, the project task T i directly depends on the project task T j , and the direct dependency relationship is represented as:

[0079] ,

[0080] wherein, represents that the project task T i and T j have a direct dependency relationship, represents the end time of the project task T i , represents the start time of the project task T j , represents dependency;

[0081] This dependency relationship indicates that the project task T i must be completed before the project task T j starts, which is also called a sequential dependency.

[0082] Step S302: analyzing the indirect dependency relationship between the project tasks, the project task T i indirectly affects the start of the project task T j through intermediate project tasks, the project task T i and T j have an indirect dependency relationship, the project task T i indirectly depends on the project task T j , and the indirect dependency relationship between the project tasks is represented as:

[0083] ,

[0084] in, Project task T i and T j The relationship between them is indirect, T k Project task T i and T j Intermediate tasks between them;

[0085] Step S303: Analyze the complex dependencies between project tasks, including conditional dependencies, parallel dependencies, and time window dependencies, specifically represented as follows:

[0086] ,

[0087] ,

[0088] ,

[0089] in, Project task T i For conditional dependency, C i Indicates condition, Indicates condition C i The state is True, where True represents the condition C. i If true, then project task T i It can be executed. Project task T i and T j There is a parallel dependency between them. Project task T i and T j They can be executed in parallel, provided that their preceding tasks T are... s Completed. Project task T i and T s There is a direct dependency relationship and the project task T j and T s It is a direct dependency relationship. Project task T i and T j The relationship between them is a time window dependency. This indicates a time window, describing the project task T. i Must be in project task T j Completed Start within the specified time.

[0090] Step S304: Use a graph convolutional network to embed and reason about the nodes in the preliminary project task knowledge graph, learn the dependencies between nodes, and update the edges in the preliminary project task knowledge graph based on the learned dependencies between nodes to obtain the project task knowledge graph G.

[0091] Figure 2 The diagram illustrates the structure of a project task knowledge graph, showcasing the semantic relationships and dependencies between task nodes and attribute nodes. Specifically, rectangular nodes represent project task nodes, elliptical nodes represent task attribute nodes, and connecting lines represent different types of relationship edges. Attribute A1 is connected to both project tasks T2 and T4 through a relation, indicating that this attribute is associated with two different task entities simultaneously. A direct dependency exists between project tasks T2 and T3, meaning that the initiation of project task T3 requires the completion of project task T2. Project tasks T3 and T4 are connected through a parallel dependency, indicating that they can be executed in parallel under conditions where resources do not conflict. A conditional dependency, connected by a dashed line, between project tasks T1 and T4 indicates that the triggering of project task T4 depends on project task T1 satisfying specific conditional constraints. This constructed project task knowledge graph, through the links of attributes, tasks, and various types of dependencies, forms a semantic network structure that supports dependency analysis and task scheduling optimization.

[0092] Step S4: Based on the project task knowledge graph and combined with the attribute information of the project tasks, optimize the project task scheduling and determine the priority and resource allocation of project task execution;

[0093] The specific steps of step S4 are as follows:

[0094] Step S401: Based on the project task knowledge graph, extract the attribute information of each project task, including task type, execution time period, resource category, and predecessor task set.

[0095] In the embodiment, the attribute sets associated with each project task node are called from the constructed project task knowledge graph, and the attribute sets include the category definition of the task, the time segment for planned execution, the required resource category, and the reference relationship of the predecessor task; then, the task type is used to distinguish tasks of different fields or different properties, and subsequent scheduling is performed according to the task characteristics; the execution period is used as a limiting condition in the time dimension to determine the feasible start time of the task and the potential conflict point; the required resource category is mapped to the occupation factor in the resource pool as the basis for subsequent resource allocation; and the predecessor task set is determined by referring to the dependency edge in the project task knowledge graph to determine the logical order and dependency chain of the task; by extracting and organizing each attribute, a task attribute vector can be formed at the node level of the project task knowledge graph, which simultaneously reflects the three types of constraints of time, resource, and dependency.

[0096] Step S402: According to the attribute information and dependency relationship of the project task, the project task is generated as an ordered linked list according to the predecessor constraint.

[0097] In the embodiment, after the attribute information of the task and the dependency edge in the project task knowledge graph are obtained, the predecessor task set corresponding to each task, i.e., the task set that must be completed before the task is executed, needs to be identified first; then, all task nodes are classified according to whether there is an unfinished predecessor task, and the task with an empty predecessor task is preferentially placed in the starting position of the linked list, and then the tasks whose predecessor conditions have been met are sequentially added to the linked list, thereby forming a topological order; specifically, the system iterates each layer of arranged tasks, retrieves tasks that have not been added to the linked list, and determines whether the predecessor tasks of the tasks have all existed in the linked list; if the condition is met, the task is inserted into the subsequent position of the linked list; in this way, an ordered linked list can be generated without breaking the dependency logic.

[0098] Step S403: The ordered linked list is divided into multiple levels, and a task conflict detection subset is established within the same level; the project tasks in each layer are cross-compared according to the resource occupation, and a scheduling matrix is formed.

[0099] In the embodiment, on the basis of the generated ordered linked list, first, the tasks are divided into several layers according to the hierarchical relationship of the predecessor dependency, each layer only contains tasks whose dependency conditions are satisfied and can be considered in parallel; then, a task conflict detection subset is set in the same layer, tasks involving the same resource category or possibly having execution time overlap are classified into the same detection range; then, by comparing the task attributes in these subsets one by one, the resource occupation is cross-referenced, and the task combinations with resource mutual exclusion or time overlap are marked; finally, the comparison results in each layer are integrated to form a multi-dimensional scheduling matrix, which takes the layers as rows and the resource categories and time sections as columns, and records the conflict or compatibility state between tasks in the corresponding positions.

[0100] Step S404: On the basis of the scheduling matrix, the conflicting project tasks are iteratively replaced until each layer meets the dependency linked list constraint and the resource non-mutual exclusion condition, and a task scheduling sequence is obtained.

[0101] The specific steps of step S404 are:

[0102] Step S4041: In the scheduling matrix, detect a set of project tasks in the same layer that have resource occupation overlap or time interval overlap, and mark them as a conflict group.

[0103] In the embodiment, in the constructed scheduling matrix, first, the resource occupation information and execution time interval of each task unit in the same layer are extracted, the resource occupation information includes the required resource type and quantity, and the execution time interval is the planned start and end period; then, in a pairwise comparison manner, it is detected whether there is repeated demand for the same type of resource or overlapping intersection in the time period between tasks, if such overlap is detected, the related tasks are merged into the same set; then, the tasks in the set are assigned a unified conflict mark to indicate that the group of tasks cannot be executed in parallel under the current conditions; finally, all detected sets are defined as conflict groups, and the formation result of the conflict groups provides clear boundary conditions for subsequent replacement and scheduling optimization.

[0104] Step S4042: Based on the project tasks in the conflict group, a plurality of candidate task sequences are generated according to the predecessor linked list order and the resource use priority, each candidate task sequence meets the basic dependency constraint.

[0105] In the embodiment, in the obtained conflict group, the topological order information in the predecessor chain table is called first, the task dependency relationship is taken as a basic constraint for sorting, and it is ensured that any task cannot be arranged before its predecessor task; then, in combination with the resource use attribute of each task, the resource category, the required number and the scarcity are classified to form a resource use priority list; then, under the limitation of the predecessor order, the tasks are arranged in turn according to the resource use priority from high to low, and different arrangement branches are generated when there is a same-level resource competition, and the execution order is covered. In this way, a plurality of candidate task sequences can be derived in the same conflict group, each candidate task sequence meets the basic dependency condition, and a differentiated arrangement strategy is reflected at the resource level.

[0106] Step S4043: different item task combinations are selected in turn among the candidate task sequences, and are replaced into the conflict position in the scheduling matrix to form a new scheduling matrix.

[0107] In the embodiment, in the plurality of candidate task sequences generated, different sequences are selected in turn as scheduling replacement schemes, and the task combinations therein are mapped to the positions originally marked as conflicts in the scheduling matrix; in this process, the arrangement of non-conflict tasks is kept unchanged, and only the units corresponding to the conflict groups are replaced to ensure the stability of the overall scheduling framework; then, the scheduling matrix after replacement is subjected to consistency checking to confirm whether it still meets the dependency order and hierarchical constraint defined by the predecessor chain table; after a plurality of replacements and combination selections, a series of new scheduling matrices are gradually generated, each scheduling matrix being a possible conflict-free scheduling layout; finally, the operation evolves the original scheduling matrix containing conflicts into an adjusted candidate matrix.

[0108] Step S4044: the new scheduling matrix is subjected to conflict detection again, if there are still item tasks that do not meet the dependency chain table constraint or the resource non-mutual exclusion condition, the replacement operation is continued to iterate until all hierarchical item tasks meet the constraint, and finally a task scheduling sequence is generated.

[0109] In this embodiment, after obtaining the new scheduling matrix, first, the conflict detection operation is re-performed on each level task, and whether the predecessor constraint in the dependency chain table is satisfied and whether there is still overlapping use of tasks in the same time period or in the same resource pool is compared one by one; if the detection result shows that there is a dependency contradiction or resource mutual exclusion that has not been eliminated, the scheduling matrix is immediately determined as not meeting the requirements, and the replacement operation of the candidate task sequence is continued in the previous step; through the continuous iteration process, each replacement generates a new scheduling matrix and accepts the same conflict checking, until all levels of tasks pass the dual test of dependency consistency and resource exclusivity; when the scheduling matrix no longer has conflicts after multiple iterations, the final task scheduling sequence is determined, which is the executable order under the global constraint conditions.

[0110] Step S5: Based on real-time data and project progress, automatically update the project task nodes and dependency relationships in the project task knowledge graph, and automatically adjust the dependency relationships, priorities and resource allocation of project tasks when there are delays or changes.

[0111] The modeling and optimization results are displayed in a graphical manner, and interactive visualization tools are used to help project managers understand the task progress, dependency relationships and resource allocation in real time, support Gantt chart, network diagram, heat map and other display modes, help decision makers quickly analyze the project status, find potential problems and make adjustments.

[0112] Embodiment 2:

[0113] Please refer to Figure 4 The application provides another embodiment: a project task dependency relationship intelligent modeling system based on a knowledge graph, which comprises a task acquisition module, a preliminary construction module, a dependency analysis module, a task scheduling optimization module and a real-time updating module.

[0114] The task acquisition module is used for acquiring project task image data and pre-processing the acquired project task image data.

[0115] The preliminary construction module is used for constructing a preliminary project task knowledge graph according to the pre-processed project task image data.

[0116] The dependency analysis module is used for analyzing the direct and indirect dependency relationships between project tasks, updating the preliminary project task knowledge graph and obtaining the project task knowledge graph.

[0117] The task scheduling optimization module is used for optimizing project task scheduling based on the project task knowledge graph and in combination with the attribute information of the project tasks, determining the priority of project task execution and resource allocation.

[0118] The real-time updating module is configured to automatically update the project task nodes and the dependency relationships in the project task knowledge graph based on real-time data and project progress, and automatically adjust the dependency relationships, priorities and resource allocation of the project tasks when there is a delay or a change.

[0119] The dependency analysis module comprises a dependency analysis unit and an updating unit.

[0120] The dependency analysis unit is configured to analyze the direct, indirect and complex dependency relationships between the project tasks.

[0121] The updating unit is configured to learn the dependency relationships between the nodes, and update the edges in the preliminary project task knowledge graph according to the learned dependency relationships between the nodes.

[0122] The task scheduling optimization module comprises a linked list generation unit, a scheduling matrix generation unit and a task scheduling sequence unit.

[0123] The linked list generation unit is configured to generate an ordered linked list according to the attribute information and the dependency relationships of the project tasks based on the predecessor constraint.

[0124] The scheduling matrix generation unit is configured to divide the ordered linked list into multiple levels, establish a task conflict detection subset in the same level, and cross-compare the project tasks in each level according to the resource occupation to form a scheduling matrix.

[0125] The task scheduling sequence unit is configured to iteratively replace the conflicting project tasks until each level satisfies the dependency linked list constraint and the resource non-mutual exclusion condition, and obtain a task scheduling sequence.

[0126] In addition, the parts of the above technical solutions in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0127] The specific embodiments described above further explain the purposes, technical solutions and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A knowledge graph-based intelligent modeling method for project task dependencies, characterized in that, include: Acquire project task image data and preprocess the acquired project task image data; Based on the preprocessed project task image data, a preliminary project task knowledge graph is constructed. Analyze the direct and indirect dependencies between project tasks, update the preliminary project task knowledge graph, and obtain the final project task knowledge graph. Based on the project task knowledge graph and combined with the attribute information of the project tasks, the scheduling of project tasks is optimized to determine the priority of project task execution and resource allocation. Based on real-time data and project progress, the project task nodes and dependencies in the project task knowledge graph are automatically updated. When delays or changes occur, the dependencies, priorities, and resource allocation of project tasks are automatically adjusted. The optimization of project task scheduling based on project task knowledge graph and combined with project task attribute information includes: Based on the project task knowledge graph, the attribute information of each project task is extracted, including task type, execution time period, resource category and predecessor task set; Based on the attribute information and dependencies of the project tasks, generate an ordered linked list of project tasks according to the predecessor constraints; The ordered linked list is divided into multiple levels, and a task conflict detection subset is established within the same level. Within each level, the project tasks are cross-compared according to their resource usage to form a scheduling matrix. Based on the scheduling matrix, conflicting project tasks are iteratively replaced until all levels satisfy the dependency list constraints and resource non-exclusivity conditions, thus obtaining the task scheduling sequence. Based on the aforementioned scheduling matrix, conflicting project tasks are iteratively replaced until all levels satisfy dependency list constraints and resource non-exclusivity conditions, resulting in a task scheduling sequence, including: In the scheduling matrix, sets of project tasks that have overlapping resource usage or overlapping time intervals within the same level are detected and marked as conflict groups; Based on the project tasks in the conflict group, multiple candidate task sequences are generated according to the order of the predecessor linked list and the resource usage priority. Each candidate task sequence satisfies the basic dependency constraints. Different combinations of project tasks are selected sequentially from the candidate task sequences and replaced at conflicting positions in the scheduling matrix to form a new scheduling matrix; The new scheduling matrix is ​​subjected to conflict detection again. If there are still project tasks that do not meet the dependency list constraints or the resource non-mutual exclusion condition, the replacement operation is returned and the iteration continues until all project tasks at all levels meet the constraints, and finally a task scheduling sequence is generated.

2. The intelligent modeling method for project task dependencies based on knowledge graphs as described in claim 1, characterized in that, The construction of the preliminary project task knowledge graph includes: Natural language processing technology is used to analyze preprocessed project task image data to identify entities in project tasks, extract project task dependencies, and extract logical relationships between project tasks. Based on the analysis results of the preprocessed project task image data, a preliminary project task knowledge graph is constructed, with the following formula: , This represents a preliminary project task knowledge graph, where V represents all nodes in the preliminary project task knowledge graph, including project task nodes and attribute nodes. T n Let A represent the nth project task node. m Let represent the m-th attribute node, n represent the number of project task nodes, m represent the number of attribute nodes, and E represent all edges in the initial project task knowledge graph, including the relationship edges between project task nodes and attribute nodes and the dependency edges between project tasks.

3. The intelligent modeling method for project task dependencies based on knowledge graphs as described in claim 2, characterized in that, The analysis examines the direct and indirect dependencies between project tasks, updating the preliminary project task knowledge graph, including: Analyze the direct dependencies between project tasks. If project task T i Must be in project task T j If completed before starting, then project task T i and T j There is a direct dependency between them, project task T i Directly dependent on project task T j ; Analyze the indirect dependencies between project tasks, project task T i The project task T is indirectly affected by intermediate project tasks. j At the beginning, project task T i and T j There is an indirect dependency relationship between project task T and other tasks. i Indirectly dependent on project task T j ; Analyze the complex dependencies between project tasks, including conditional dependencies, parallel dependencies, and time window dependencies; Graph convolutional networks are used to embed and reason about nodes in the preliminary project task knowledge graph, learn the dependencies between nodes, and update the edges in the preliminary project task knowledge graph based on the learned dependencies, thus obtaining the project task knowledge graph G.

4. The intelligent modeling method for project task dependencies based on knowledge graphs as described in claim 3, characterized in that, The conditional dependency, parallel dependency, and time window dependency are specifically represented as follows: If and only if the project task T i Dependency C i When true, project task T i With condition C i There is a conditional dependency relationship, when project task T i and T j It can be executed in parallel, and the project task T i and T s For direct dependencies and project tasks T j and T s When there is a direct dependency, project task T i and T j There is a parallel dependency relationship between them, when project task T i Depends on project task T j And it satisfies the time interval constraint, project task T i and T j The relationship between them is a time window dependency.

5. The intelligent modeling method for project task dependencies based on knowledge graphs as described in claim 4, characterized in that, The project task image data includes: basic information of the project tasks, explicit dependencies between project tasks, and project task attributes; The preprocessing includes: unifying the time format, converting the time information of project tasks into a unified format; Task classification and tagging: Based on the content of project tasks, project tasks are classified by domain and task type; Redundant information removal: Remove irrelevant or redundant data.

6. A knowledge graph-based intelligent modeling system for project task dependencies, used to implement the knowledge graph-based intelligent modeling method for project task dependencies as described in any one of claims 1-5, characterized in that, include: The module includes a task acquisition module, a preliminary construction module, a dependency analysis module, a task scheduling optimization module, and a real-time update module. The task acquisition module is used to acquire project task image data and preprocess the acquired project task image data. The preliminary construction module is used to construct a preliminary project task knowledge graph based on the preprocessed project task image data. The dependency analysis module is used to analyze the direct and indirect dependencies between project tasks, update the preliminary project task knowledge graph, and obtain the project task knowledge graph. The task scheduling optimization module is used to optimize project task scheduling based on the project task knowledge graph and the attribute information of the project tasks, and to determine the priority and resource allocation of project task execution. The real-time update module is used to automatically update the project task nodes and dependencies in the project task knowledge graph based on real-time data and project progress. When delays or changes occur, it automatically adjusts the dependencies, priorities, and resource allocation of project tasks.

7. The intelligent modeling system for project task dependencies based on knowledge graphs as described in claim 6, characterized in that, The dependency analysis module includes: a dependency analysis unit and an update unit; The dependency analysis unit is used to analyze the direct, indirect, and complex dependencies between project tasks; The update unit is used to learn the dependencies between nodes and update the edges in the preliminary project task knowledge graph based on the learned dependencies between nodes.

8. The intelligent modeling system for project task dependencies based on knowledge graphs as described in claim 7, characterized in that, The task scheduling optimization module includes: a linked list generation unit, a scheduling matrix generation unit, and a task scheduling sequence unit; The linked list generation unit is used to generate an ordered linked list of project tasks according to the attribute information and dependencies of the project tasks. The scheduling matrix generation unit is used to divide the ordered linked list into multiple levels, establish a task conflict detection subset within the same level, and cross-compare the project tasks according to resource usage within each level to form a scheduling matrix. The task scheduling sequence unit is used to iteratively replace conflicting project tasks until all levels satisfy the dependency list constraints and resource non-exclusivity conditions, thus obtaining the task scheduling sequence.

Images