Task splitting method, system and equipment based on multi-dimensional contract elements and medium

Abstract

The invention relates to the technical field of project management, in particular to a task splitting method, system and device based on multi-dimensional contract elements and a medium, and the method comprises the steps: obtaining an original contract, and determining multi-dimensional contract element data according to the original contract; based on a preset task splitting model, inputting the multi-dimensional contract element data into the task splitting model to obtain a modular task; obtaining a preset resource distribution condition, allocating the modular tasks according to the regional information and the resource distribution condition, and constructing a task allocation scheme; and monitoring a project execution state in real time, and optimizing the task allocation scheme based on the deviation between the project execution state and the task allocation scheme. The task splitting method and device have the effect of improving the task splitting efficiency and accuracy.

Description

Technical Field

[0001] This application relates to the field of project management technology, and in particular to a method, system, device and medium for task decomposition of multidimensional contract elements. Background Technology

[0002] Currently, in the field of project management and intelligent services, task decomposition is a key step in breaking down a complex project objective into a series of smaller and more manageable independent work units. It is a fundamental task to ensure that the project can be executed smoothly, resources can be allocated effectively, and costs can be controlled precisely.

[0003] Existing task breakdown methods mainly rely on the manual judgment and personal experience of project managers. While they are feasible for handling simple and clear tasks, their limitations are more apparent when faced with modern service contracts that are complex in terms, diverse in service content, geographically widespread, and have tight delivery cycles. Due to a lack of comprehensive analysis capabilities for multi-dimensional information in the contract, manual breakdown cannot fully and accurately consider the combined effects of factors such as subtle differences in contract terms, diversity of service content, complexity of geographical information, and urgency of delivery cycles. This can easily lead to problems such as unreasonable breakdown results, inaccurate resource matching, and untimely adjustments to project plans.

[0004] The existing technical solutions mentioned above have drawbacks such as reliance on human experience for task breakdown, difficulty in comprehensively processing multi-dimensional contract information, and low efficiency and accuracy in task breakdown. Therefore, there is room for improvement. Summary of the Invention

[0005] To improve the efficiency and accuracy of task decomposition, this application provides a task decomposition method, system, device, and medium based on multi-dimensional contract elements.

[0006] The above-mentioned objective of this application is achieved through the following technical solution: A task decomposition method based on multidimensional contract elements, the task decomposition method based on multidimensional contract elements includes: Obtain the original contract, and determine multi-dimensional contract element data based on the original contract. The multi-dimensional contract element data includes at least the contract terms, service content, geographical information, and delivery cycle. Based on a preset task splitting model, the multidimensional contract element data is input into the task splitting model to obtain modular tasks; Obtain the preset resource distribution information, and allocate the modular tasks according to the geographical information and the resource distribution information to construct a task allocation scheme; Monitor the project execution status in real time, and optimize the task allocation scheme based on the deviation between the project execution status and the task allocation scheme.

[0007] By adopting the above technical solutions, and by acquiring the original contract and identifying multi-dimensional contract element data, unstructured contract text can be transformed into structured data containing multiple logical dimensions, avoiding omissions and ambiguities caused by manual interpretation. Modular tasks are obtained based on a pre-set task decomposition model, which can decompose complex project objectives into clearly defined and specific task units, significantly improving the efficiency and scientific nature of task decomposition and reducing reliance on personal experience. By acquiring resource distribution information and allocating modular tasks, reasonable task-resource matching can be achieved based on objective conditions such as geographical location and resource availability, thereby improving the accuracy of resource scheduling and avoiding project delays and cost waste caused by resource mismatch. By monitoring the project execution status in real time and optimizing the task allocation scheme based on deviations, a responsive closed-loop optimization is established, allowing for real-time detection of differences between the plan and reality, greatly enhancing the flexibility and adaptability of the solution and ensuring the project proceeds smoothly as planned.

[0008] In a preferred embodiment, this application can be further configured such that the task decomposition method based on multidimensional contract elements also includes: Obtain historical project data, and from the historical project data, obtain historical multidimensional contract element data and the historical task breakdown results corresponding to the historical multidimensional contract element data; Based on the historical multidimensional contract element data and the historical task splitting results, a training dataset is constructed; The initial model is trained using the training dataset to obtain the task splitting model.

[0009] By adopting the above technical solution, a training dataset is constructed by acquiring historical project data, and the initial model is trained based on this dataset. This allows past project experience and best practices to be systematically and reusably incorporated into the machine learning model. As a result, task splitting decisions are no longer isolated and one-off judgments, but rather a sustainable optimization process that can inherit and utilize historical experience, thereby continuously improving the quality of task splitting.

[0010] In a preferred embodiment, this application can be further configured such that: determining the multidimensional contract element data based on the original contract specifically includes: Based on natural language processing technology, the original contract is parsed to obtain contract element data and corresponding direct constraint relationships; Based on the contract element data and the corresponding direct constraints, a knowledge graph is constructed. Through a rule engine and an external database, cross-clause dependencies and location-based environmental constraints are built in the knowledge graph to generate the multidimensional contract element data.

[0011] By adopting the above technical solution, and by using natural language processing technology to parse contracts and construct a knowledge graph that includes cross-clause dependencies and environmental constraints, a data structure that not only contains literal textual information but also integrates deep logical dependencies and external reality constraints can be generated. This enables subsequent models to make decisions based on a more comprehensive and profound understanding, rather than just on surface keywords, greatly improving the rationality and accuracy of subsequent task decomposition and resource allocation.

[0012] In a preferred embodiment, this application can be further configured such that: inputting the multidimensional contract element data into the task splitting model to obtain modular tasks specifically includes: The multidimensional contract element data is processed by feature vectorization to obtain vectorized information, and the vectorized information is fused by multimodal fusion technology to generate a joint feature matrix; Based on the clustering algorithm, cluster analysis is performed on the joint feature matrix to obtain task clusters; Based on the decision tree and the dependencies in the multidimensional contract element data, the joint feature matrix is ​​parsed to obtain a task priority list. Based on the task priority list and the delivery cycle of the multidimensional contract element data, the time nodes are obtained using the critical path method. The modular task is obtained by associating the task cluster, the task priority list, and the time node.

[0013] By adopting the above technical solution, through a series of precise processing steps such as feature vectorization, multimodal fusion, cluster analysis, dependency resolution, and time node planning of multidimensional contract element data, abstract multidimensional data can be systematically transformed into a modular task set containing task content, execution order, and time plan. This automates and standardizes the task decomposition process, completely replacing the tedious work of traditional manual planning, ensuring the consistency and traceability of the decomposition results, and greatly shortening the project preparation cycle.

[0014] In a preferred embodiment, this application can be further configured as follows: The allocation of the modular tasks based on the geographical information and the resource distribution, and the construction of a task allocation scheme, specifically includes: Based on the resource distribution, the modular tasks are matched using a geographic matching algorithm and the regional information to obtain a candidate resource list; A total cost calculation formula is constructed based on the candidate resource list, and constraints are determined according to the delivery cycle in the multidimensional contract element data and the upper limit of resource capacity in the resource distribution. Based on the planning algorithm, the optimal solution of the total cost calculation formula under the constraints is calculated. Based on the optimal solution, the corresponding resources for the modular task are selected from the candidate resource list. Based on the resources and the modular task, the task allocation scheme is generated.

[0015] By adopting the above technical solution, and using a two-stage optimization strategy that combines preliminary candidate set screening with global optimization solution to allocate modular tasks, it is possible to ensure that the allocation result is globally cost-optimal while effectively reducing the computational complexity of the solution. This makes it possible to find a high-quality allocation solution within a limited time, thus balancing solution quality and decision-making efficiency.

[0016] In a preferred embodiment, this application can be further configured such that: optimizing the task allocation scheme based on the deviation between the project execution status and the task allocation scheme specifically includes: Based on the rule engine, the project execution status is compared with the task allocation scheme to determine whether there is a deviation between the project execution status and the task allocation scheme; If there is a deviation between the project execution status and the task allocation scheme, the current project execution status is obtained, and the current project status is used as a new constraint condition for the total cost calculation formula to determine the optimized task allocation scheme. Based on the current project execution status and the optimized task allocation scheme, a visual display interface is generated for presentation on the visual monitoring interface.

[0017] By adopting the above technical solution, a monitoring-judgment-response mechanism is established by comparing the execution status based on the rule engine and replanning with the current project status as a new constraint when deviations occur. This mechanism can not only automatically detect the deviation between the plan and the actual situation, but also quantify the deviation into a new constraint and automatically call the optimization engine for replanning. This achieves closed-loop and proactive management of project risks, greatly improving the project's adaptability and self-correction capabilities in complex and ever-changing environments.

[0018] In a preferred embodiment, this application can be further configured such that the task decomposition method based on multidimensional contract elements also includes: The completed task data is used as incremental learning samples, and the task splitting model is updated based on the incremental learning samples.

[0019] By adopting the above technical solution, and by acquiring completed task data as incremental learning samples, and fine-tuning and updating the task splitting model, the task splitting model can be enabled to continuously learn from actual execution results and self-evolve. This allows the model's splitting strategy to iteratively approach the optimal practice in a specific organization or business scenario, significantly improving the intelligence level of task splitting.

[0020] The second objective of this invention is achieved through the following technical solution: A task decomposition system based on multidimensional contract elements, the task decomposition system based on multidimensional contract elements includes: The data determination module is used to obtain the original contract and determine multi-dimensional contract element data based on the original contract. The multi-dimensional contract element data includes at least contract terms, service content, geographical information, and delivery cycle. The task splitting module is used to input the multidimensional contract element data into the task splitting model based on a preset task splitting model to obtain modular tasks; The resource allocation module is used to obtain the preset resource distribution information, allocate the modular tasks according to the geographical information and the resource distribution information, and construct a task allocation scheme. The dynamic optimization module is used to monitor the project execution status in real time and optimize the task allocation scheme based on the deviation between the project execution status and the task allocation scheme.

[0021] By adopting the above technical solutions, and by acquiring the original contract and identifying multi-dimensional contract element data, unstructured contract text can be transformed into structured data containing multiple logical dimensions, avoiding omissions and ambiguities caused by manual interpretation. Modular tasks are obtained based on a pre-set task decomposition model, which can decompose complex project objectives into clearly defined and specific task units, significantly improving the efficiency and scientific nature of task decomposition and reducing reliance on personal experience. By acquiring resource distribution information and allocating modular tasks, reasonable task-resource matching can be achieved based on objective conditions such as geographical location and resource availability, thereby improving the accuracy of resource scheduling and avoiding project delays and cost waste caused by resource mismatch. By monitoring the project execution status in real time and optimizing the task allocation scheme based on deviations, a responsive closed-loop optimization is established, allowing for real-time detection of differences between the plan and reality, greatly enhancing the flexibility and adaptability of the solution and ensuring the project proceeds smoothly as planned.

[0022] The above-mentioned objective three of this application is achieved through the following technical solution: A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the task decomposition method based on multidimensional contract elements described above.

[0023] The fourth objective of this application is achieved through the following technical solution: A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the task decomposition method based on multidimensional contract elements described above.

[0024] In summary, this application includes at least one of the following beneficial technical effects: 1. By acquiring the original contract and identifying multi-dimensional contract element data, unstructured contract text can be transformed into structured data containing multiple logical dimensions, avoiding omissions and ambiguities from manual interpretation. Modular tasks are obtained based on a preset task decomposition model, which can decompose complex project goals into clearly defined and specific task units, thereby significantly improving the efficiency and scientific nature of task decomposition and reducing reliance on personal experience. By acquiring resource distribution information and allocating modular tasks, reasonable task-resource matching can be performed based on objective conditions such as geographical location and resource availability, thereby improving the accuracy of resource scheduling and avoiding project delays and cost waste caused by resource mismatch. By monitoring the project execution status in real time and optimizing the task allocation scheme based on deviations, a responsive closed-loop optimization is established, which can detect the difference between the plan and the actual situation in real time, thereby greatly enhancing the flexibility and adaptability of the scheme and ensuring that the project can proceed smoothly as planned. 2. By comparing the execution status based on the rule engine and replanning with the current project status as a new constraint when deviations occur, a monitoring-judgment-response mechanism is established. This mechanism can not only automatically detect the deviation between the plan and the actual situation, but also quantify the deviation into a new constraint and automatically call the optimization engine for replanning. This achieves closed-loop and proactive management of project risks, greatly improving the project's adaptability and self-correction capabilities in complex and ever-changing environments. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating the implementation of a task decomposition method based on multi-dimensional contract elements in one embodiment of this application; Figure 2 This is another implementation flowchart of the task decomposition method based on multi-dimensional contract elements in one embodiment of this application; Figure 3 This is a flowchart illustrating the implementation of step S10 in a task splitting method based on multidimensional contract elements in one embodiment of this application. Figure 4 This is a flowchart illustrating the implementation of step S20 in a task splitting method based on multidimensional contract elements in one embodiment of this application. Figure 5 This is a flowchart illustrating the implementation of step S30 in a task splitting method based on multidimensional contract elements in one embodiment of this application. Figure 6 This is a flowchart illustrating the implementation of step S40 in a task splitting method based on multidimensional contract elements in one embodiment of this application. Figure 7 This is a flowchart illustrating the implementation of step S40 and beyond in a task splitting method based on multidimensional contract elements according to an embodiment of this application. Figure 8 This is a principle block diagram of a task splitting system based on multi-dimensional contract elements in one embodiment of this application; Figure 9 This is a schematic diagram of the internal structure of a computer device according to an embodiment of this application. Detailed Implementation

[0026] The following embodiments will help those skilled in the art to further understand the function of this application, but do not limit this 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 this application. These all fall within the protection scope of this application.

[0027] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, circuits, and methods are omitted so as not to obscure the description of this application with unnecessary detail.

[0028] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0029] The present application will be further described in detail below with reference to the accompanying drawings.

[0030] In one embodiment, such as Figure 1 As shown, this application discloses a task decomposition method based on multi-dimensional contract elements, which specifically includes the following steps: S10: Obtain the original contract and determine the multi-dimensional contract element data based on the original contract. The multi-dimensional contract element data includes at least the contract terms, service content, geographical information, and delivery cycle.

[0031] Specifically, the original contract refers to an unprocessed commercial contract document, usually in PDF or Word format. It is the starting information source for the entire task breakdown process. The multidimensional contract element data is a structured data set formed after in-depth interpretation and information extraction of the original contract. The contract terms record legal and commercial constraints such as payment terms and liability for breach of contract. The service content defines the specific work items that need to be delivered. The geographical information marks the geographical scope of service execution. The delivery cycle clarifies the time nodes and final deadlines of each task. This whole set of data provides a preliminary classified and organized data foundation for subsequent automated processing.

[0032] S20: Based on the preset task splitting model, input the multi-dimensional contract element data into the task splitting model to obtain modular tasks.

[0033] Specifically, the obtained multidimensional contract element data is used as input to a pre-established and trained task decomposition model. The analytical capabilities of the task decomposition model are used to decompose the complex service project represented by the entire contract. The core function of this task decomposition model is to understand the internal logic between the elements of each dimension and, based on this, to transform the macro service objectives into a series of specific, independently manageable, and interrelated sub-task units. For example, a large-scale nationwide software deployment project may be decomposed by the model into multiple modular tasks such as software installation in North China, user training in South China, and data center construction. Each task contains clear work content and constraints with other tasks, providing clear objects for subsequent resource allocation.

[0034] S30: Obtain the preset resource distribution information, allocate modular tasks based on the geographical information and resource distribution information, and construct a task allocation scheme.

[0035] Specifically, the preset resource distribution refers to a database or list that records detailed information on all available execution resources, including information such as the engineer's technical level, city, and current workload. After obtaining the modular task list, in order to effectively assign these tasks to specific execution units, it is necessary to introduce data describing the availability of resources, namely the preset resource distribution. Through a matching decision-making process, this process examines the execution requirements of each modular task, focusing on the task's geographical attributes and the resource's geographical location, and finds the most suitable executor from the resource distribution. For example, a task that needs to be completed in Shenzhen may be assigned to a local engineering team in Shenzhen. Ultimately, a task allocation scheme that clearly defines which specific resource is responsible for executing each modular task and is operable is formed.

[0036] S40: Monitor project execution status in real time and optimize task allocation scheme based on the deviation between project execution status and task allocation scheme.

[0037] Specifically, to address various unforeseen circumstances during project execution, a continuous monitoring and adjustment mechanism needs to be established. Real-time monitoring of project execution status means that after the project begins, the monitoring and adjustment mechanism continuously collects actual progress data from the project site, such as the actual progress of tasks and the real-time status of resources. This actual progress data is then compared with the planned data in the constructed task allocation scheme. Once a significant deviation between the actual situation and the plan is detected, such as a task planned to be completed in 3 days actually taking 5 days, an automatic replanning action is triggered to optimize the current task allocation scheme. This optimization process will reallocate and plan the tasks that have not yet been completed based on the latest situation, generating an adjusted task allocation scheme that can adapt to the latest situation to ensure that the project goals can ultimately be achieved.

[0038] In one embodiment, such as Figure 2 As shown, before step S20, the task decomposition method based on multi-dimensional contract elements further includes: S201: Obtain historical project data, and extract historical multidimensional contract element data and the corresponding historical task breakdown results from the historical project data.

[0039] Specifically, historical project data refers to the complete archive of projects that the company has completed in the past. In order to train an intelligent task decomposition model, it is first necessary to prepare learning materials for training the model. These materials come from historical project data. Obtaining this historical project data includes not only the original contract documents, but more importantly, finding the historical task decomposition results of how these contracts were successfully decomposed and executed. For example, a contract for a national inspection project last year, and at the same time, finding the task decomposition table and execution sequence record that were formulated and verified to be effective at that time. These two parts of data together constitute a complete case that can be learned.

[0040] S202: Construct a training dataset based on historical multidimensional contract element data and historical task splitting results.

[0041] Specifically, the numerous historical project cases obtained in the previous step are organized and paired to construct a structured training dataset. In this process, each historical multidimensional contract element data is labeled as an input feature, while its corresponding historical task splitting result is labeled as the correct label. By combining hundreds or thousands of such (input feature, correct label) data pairs, a training dataset that can be learned by machine learning algorithms is constructed. The model will learn the mapping ability from contract information to reasonable task splitting by learning from this training dataset.

[0042] S203: Train the initial model based on the training dataset to obtain the task splitting model.

[0043] Specifically, the constructed training dataset is used to train an untrained initial model. The internal parameters are fine-tuned through repeated learning. During training, features from the training dataset are input into the untrained initial model in batches. The model will provide a predicted splitting result. Then, a preset loss function is used to calculate the difference between the predicted result and the correct label. This difference is then used to fine-tune the model's internal parameters through backpropagation. This process is repeated until the model's prediction error on the entire training dataset converges to a sufficiently low level, that is, the error between the model's predicted splitting result and the actual historical splitting result is reduced to an acceptable range. At this point, a task splitting model with the ability to learn from historical experience and perform intelligent task splitting is trained.

[0044] In one embodiment, such as Figure 3 As shown, in step S10, that is, determining the multidimensional contract element data based on the original contract, specifically includes: S11: Based on natural language processing technology, the original contract is parsed to obtain contract element data and corresponding direct constraint relationships.

[0045] Specifically, by using natural language processing technology to perform in-depth analysis and information extraction on the original contract, for example, by using named entity recognition technology, entities such as Party A and Party B can be identified, and by using semantic role labeling technology, logical relationships can be directly determined within a single sentence, thereby obtaining direct constraint relationships corresponding to the contract element data. For example, from the sentence "Party B shall respond within 7 days", the constraint relationship between the action of responding and the time within 7 days can be directly determined.

[0046] S12: Based on the contract element data and the corresponding direct constraints, construct a knowledge graph, and through a rule engine and an external database, construct cross-clause dependencies and location-based environmental constraints in the knowledge graph to generate multi-dimensional contract element data.

[0047] Specifically, the obtained contract element data and corresponding direct constraint relationships are integrated to construct a network-like knowledge graph, where contract elements are nodes, relationships are edges, and cross-clause dependencies refer to logical connections that require combining content from different chapters to deduce. For example, by analyzing the paths between the nodes of completed acceptance and final payment in the graph, the sequential dependency between the two is established. Then, by linking external databases such as national statutory holiday databases or local policy and regulation databases, and using a rule engine for reasoning, environmental constraints are automatically added to tasks in specific regions. For example, when the local information is Guangdong, after querying the external knowledge base, the rule engine automatically adds constraints requiring Cantonese language support to the relevant service content. The final result is a knowledge graph that integrates multiple sources and multiple levels, which is a multi-dimensional contract element data with extremely rich content.

[0048] In one embodiment, such as Figure 4 As shown, in step S20, the multidimensional contract element data is input into the task splitting model to obtain modular tasks, specifically including: S21: Perform feature vectorization processing on the multidimensional contract element data to obtain vectorized information, and then use multimodal fusion technology to fuse the vectorized information to generate a joint feature matrix.

[0049] Specifically, the determined multidimensional contract element data, such as contract terms, geographical location names, time series, and other different types of information, are uniformly converted into numerical vectors that can be mathematically processed by machines through technologies such as word embedding and geocoding. Then, multimodal fusion technologies, such as graph neural networks (GNN) or attention mechanisms (Transformer), are used to intelligently combine these numerical vectors from different sources and deeply understand their intrinsic relationships, rather than simply concatenating them, to finally generate a joint feature matrix. This joint feature matrix is ​​a high-dimensional numerical table in which each row represents a potential subtask and each column represents a feature dimension. Moreover, this matrix already contains the complex relationships between textual, geographical, and temporal information, preserving the complex correlations between information of different dimensions to the greatest extent.

[0050] S22: Based on the clustering algorithm, cluster analysis is performed on the joint feature matrix to obtain the task cluster.

[0051] Specifically, clustering algorithms refer to unsupervised learning algorithms, such as K-Means or DBSCAN. By employing one of these clustering algorithms, the algorithm performs unsupervised learning analysis on the joint feature matrix generated in the previous step. Since each point in the feature matrix represents a potential subtask, and similar tasks are close in the feature space, the clustering algorithm can automatically classify these points into different clusters. Each cluster corresponds to a task cluster. For example, all tasks that need to be executed in North China are clustered into one cluster, and all software development-related tasks are clustered into another cluster.

[0052] S23: Based on the dependency relationships in the decision tree and multidimensional contract element data, perform dependency parsing on the joint feature matrix to obtain a task priority list.

[0053] Specifically, a decision tree is a supervised learning algorithm that uses decision trees or similar rule-based learning algorithms to analyze the execution order of tasks. This algorithm takes the task clusters generated in the previous step as the analysis object and uses the dependencies defined in the multidimensional contract element data as strong guiding rules, such as the acceptance clauses obtained from the knowledge graph depending on the service completion clauses. By analyzing the characteristics of each task cluster in the joint feature matrix, a series of logical rules for determining the execution order of tasks are generated. For example, if task A is data preparation, then task B, such as data analysis, must be executed after A. This clarifies the pre- and post-constraints between different task clusters, constructs a logical execution order among all task clusters, and finally forms a task priority list with a clear order of tasks.

[0054] S24: Based on the delivery cycle of the task priority list and multidimensional contract element data, the time nodes are obtained using the critical path method.

[0055] Specifically, after determining the execution order of tasks, it is necessary to plan a specific time schedule for them. Critical Path Method (CPM) is a classic project management network diagram technique. By taking the task priority list generated in the previous step and the final delivery date specified in the contract as input, CPM calculates the earliest start time, latest start time, and latest end time of each task, thereby identifying the critical path that affects the overall project duration and assigning a specific planned execution window, i.e., a time node, to each task cluster.

[0056] S25: Associate the task cluster, task priority list, and time nodes to obtain modular tasks.

[0057] Specifically, the information about different aspects of the task produced by the preceding parallel or serial processing steps is integrated into one, that is, the three core pieces of information—the task cluster representing the task content, the task priority list representing the execution order, and the time node representing the execution time—are linked and encapsulated into a standardized data object with complete information and a unified structure. This data object, which ultimately forms and contains complete execution instructions, is a modular task that can be used for the next step of resource allocation, clearly describing what to do, when to do it, and after which task.

[0058] In one embodiment, such as Figure 5 As shown, in step S30, modular tasks are allocated based on geographical information and resource distribution to construct a task allocation scheme, specifically including: S31: Based on resource distribution, modular tasks are matched using a geographic matching algorithm and regional information to obtain a candidate resource list.

[0059] Specifically, the geographic matching algorithm is a fast heuristic filtering algorithm, such as the greedy algorithm or the Hungarian algorithm. To avoid the huge computational burden caused by global optimization in a large resource pool, one of the geographic matching algorithms is first adopted. This algorithm will, for each modular task generated in the previous step, prioritize matching and filtering several available resources that are geographically closest to the target based on the corresponding regional information from the global resource distribution. For example, for a task in Beijing, three local teams in Beijing and two teams in Tianjin will be filtered out. This generates a much smaller but higher-quality candidate resource list for each modular task, thus providing high-quality input for the next step of refined decision-making.

[0060] S32: Construct a total cost calculation formula based on the candidate resource list, and determine the constraints based on the upper limit of resource capacity according to the delivery cycle and resource distribution in the multidimensional contract element data.

[0061] Specifically, the total cost calculation formula is a mathematical objective function used to evaluate the merits of different allocation schemes. It quantifies and weights the labor costs, equipment costs, and penalty costs adjusted by the weighting coefficient α corresponding to the allocation of different candidate resources. That is, the total cost calculation formula = ∑(Cost 人力 +Cost 设备 +α×Delay Penalty The constraints are the red lines that must be followed when allocating resources. For example, as a hard constraint, the delivery cycle specifies the latest completion time of the task, while the resource capacity limit ensures that the workload of any allocated resource will not exceed its maximum capacity. This series of definitions together constitutes a complete and solvable optimization problem.

[0062] S33: Based on the planning algorithm, calculate the optimal solution of the total cost calculation formula under constraints. Based on the optimal solution, select the corresponding resources for the modular task from the candidate resource list, and generate a task allocation scheme according to the resources and the modular task.

[0063] Specifically, a planning algorithm is an optimization solver in operations research, such as linear programming (LP) or genetic algorithm (GA). By employing one of these planning algorithms, the algorithm is used to solve the optimization problem defined in the previous step. Through systematic calculation, under the premise of satisfying all constraints, it finds a specific combination of resource allocation that minimizes the total cost calculation formula. This combination is the optimal solution. Finally, based on this optimal solution, an optimal execution resource is deterministically selected from its corresponding candidate resource list for each modular task. Based on these determined "task-resource" pairings, the final task allocation scheme with the optimal global cost is generated.

[0064] In one embodiment, such as Figure 6 As shown, in step S40, the task allocation scheme is optimized based on the deviation between the project execution status and the task allocation scheme. This specifically includes: S41: Based on the rule engine, the project execution status is compared with the task allocation plan to determine whether there is a deviation between the project execution status and the task allocation plan.

[0065] Specifically, a rule engine refers to a logical judgment component, such as Drools. This rule engine has a set of pre-set rules for monitoring the health of a project. The project execution status is collected in real time through message queues such as Kafka, which collect data on the actual progress of tasks and resource utilization. The rule engine continuously collects this real-time project execution status data and compares it with the planned data in the original task allocation scheme. It then makes judgments based on pre-set rules. For example, a rule could be that if the actual start time of a task is later than its planned start time by more than a preset threshold, it is judged as a delay deviation, thereby identifying deviations in the project.

[0066] S42: If there is a deviation between the project execution status and the task allocation plan, obtain the current project execution status, and use the current project status as a new constraint condition for the total cost calculation formula to determine the optimized task allocation plan.

[0067] Specifically, once the rule engine in the previous step determines that there is a deviation, it will trigger a dynamic adjustment process. This process first captures and solidifies the current project execution status, for example, task A has been delayed for 3 days, resource B is temporarily unavailable, and adds these new facts as a set of unchangeable new constraints to the optimization problem defined in step S32. Then, the planning algorithm in step S33 is reused, and under the premise of satisfying this set of new constraints, a new optimization allocation calculation is performed on all unfinished tasks to determine a task allocation scheme that can adapt to the latest situation, that is, the optimized task allocation scheme.

[0068] S43: Based on the current project execution status and the optimized task allocation scheme, generate a visual display interface for display on the visual monitoring interface.

[0069] Specifically, for ease of understanding and management, the process and results of dynamic adjustments need to be presented intuitively. In particular, the current project execution status, such as which tasks have been completed and which are delayed, and the optimized task allocation plan, such as how it will be executed in the future, will be rendered together on a dynamic monitoring interface based on web technologies such as Vue and D3.js. This visual monitoring interface will include various charts such as real-time updated Gantt charts, resource heatmaps reflecting resource load, and dashboards displaying overall progress. For example, different colors and markers will be used to display these statuses on a real-time updated Gantt chart, allowing managers to have a clear understanding of the past, present, and future of the project and make efficient decisions.

[0070] In one embodiment, such as Figure 7 As shown, after step S40, the task decomposition method based on multi-dimensional contract elements further includes: S50: Obtain the completed task data as incremental learning samples, and update the task splitting model based on the incremental learning samples.

[0071] Specifically, completed task data refers to the complete data record generated after a project or modular task is completed, which includes the initial allocation plan and the final actual execution result. For example, it not only records the planned time and allocated resources of the task, but also its actual time and actual resources consumed. A new training case, namely the incremental learning sample, is built based on this verified real data. It takes the initial state of the task as the input feature and its actual execution result as the correct label. This new incremental learning sample is input into the deployed task splitting model, and its internal parameters are fine-tuned in a small and real-time manner. This allows the model to learn from the experience of this project and perform more accurately and intelligently in future task splitting.

[0072] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0073] In one embodiment, a task splitting system based on multi-dimensional contract elements is provided, which corresponds one-to-one with the task splitting method based on multi-dimensional contract elements described in the above embodiments. For example... Figure 8 As shown, this task decomposition system based on multi-dimensional contract elements includes a data determination module, a task decomposition module, a resource allocation module, and a dynamic optimization module. Detailed descriptions of each functional module are as follows: The data determination module is used to obtain the original contract and determine the multi-dimensional contract element data based on the original contract. The multi-dimensional contract element data includes at least the contract terms, service content, geographical information, and delivery cycle. The task splitting module is used to input multi-dimensional contract element data into the task splitting model based on a preset task splitting model to obtain modular tasks; The resource allocation module is used to obtain the preset resource distribution information, allocate modular tasks based on the geographical information and resource distribution information, and construct a task allocation scheme. The dynamic optimization module is used to monitor the project execution status in real time and optimize the task allocation scheme based on the deviation between the project execution status and the task allocation scheme.

[0074] Optionally, the task decomposition system based on multi-dimensional contract elements also includes: The training data acquisition module is used to acquire historical project data, and to obtain historical multidimensional contract element data and the corresponding historical task decomposition results from the historical project data. The dataset building module is used to construct a training dataset based on historical multidimensional contract element data and historical task splitting results; The model training module is used to train the initial model based on the training dataset to obtain the task splitting model.

[0075] Optional, the data determination module specifically includes: The text parsing submodule is used to parse the original contract based on natural language processing technology to obtain contract element data and corresponding direct constraint relationships; The knowledge graph construction submodule is used to construct a knowledge graph based on contract element data and corresponding direct constraints. Through a rule engine and an external database, it constructs cross-clause dependencies and location-based environmental constraints in the knowledge graph to generate multi-dimensional contract element data.

[0076] Optional, task splitting modules, specifically including: The feature engineering submodule is used to perform feature vectorization processing on multi-dimensional contract element data to obtain vectorized information, and then use multi-modal fusion technology to fuse the vectorized information to generate a joint feature matrix; The task clustering submodule is used to perform cluster analysis on the joint feature matrix based on clustering algorithms to obtain task clusters; The priority parsing submodule is used to parse the dependency relationships in the joint feature matrix based on the decision tree and multidimensional contract element data to obtain a task priority list. The time planning submodule is used to obtain time nodes based on the delivery cycle of the task priority list and multidimensional contract element data, using the critical path method. The task integration submodule is used to associate task clusters, task priority lists, and time nodes to obtain modular tasks.

[0077] Optional, resource allocation module, specifically including: The resource pre-matching submodule is used to match modular tasks based on resource distribution using geographic matching algorithms and regional information to obtain a candidate resource list; The cost modeling submodule is used to construct the total cost calculation formula based on the candidate resource list and determine the constraints based on the upper limit of resource capacity in the delivery cycle and resource distribution in the multidimensional contract element data. The optimization and solution submodule is used to calculate the optimal solution of the total cost calculation formula under constraints based on the planning algorithm. Based on the optimal solution, it selects the corresponding resources for the modular task from the candidate resource list and generates a task allocation scheme according to the resources and the modular task.

[0078] Optional, dynamic optimization module, specifically including: The deviation detection submodule is used to compare the project execution status with the task allocation scheme based on the rule engine in order to determine whether there is a deviation between the project execution status and the task allocation scheme. The task replanning submodule is used to obtain the current project execution status if there is a deviation between the project execution status and the task allocation plan, and to determine the optimized task allocation plan based on the current project status as a new constraint in the total cost calculation formula. The status visualization submodule is used to generate a visual display interface based on the current project execution status and the optimized task allocation scheme, so as to be displayed on the visual monitoring interface.

[0079] Optionally, the task decomposition system based on multi-dimensional contract elements also includes: The model update module is used to obtain completed task data as incremental learning samples and update the task splitting model based on the incremental learning samples.

[0080] For specific limitations regarding the task decomposition system based on multidimensional contract elements, please refer to the limitations of the task decomposition method based on multidimensional contract elements mentioned above, which will not be repeated here. Each module in the aforementioned task decomposition system based on multidimensional contract elements can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0081] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 9 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores data such as original contracts, task decomposition models, resource distribution, and task allocation schemes. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements a task decomposition method based on multi-dimensional contract elements.

[0082] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps: Obtain the original contract, and based on the original contract, determine the multi-dimensional contract element data, which includes at least the contract terms, service content, geographical information, and delivery cycle. Based on the preset task splitting model, multidimensional contract element data is input into the task splitting model to obtain modular tasks; Obtain the preset resource distribution information, allocate modular tasks based on the geographical information and resource distribution information, and construct a task allocation scheme; Monitor project execution status in real time, and optimize task allocation scheme based on the deviation between project execution status and task allocation scheme.

[0083] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor: Obtain the original contract, and based on the original contract, determine the multi-dimensional contract element data, which includes at least the contract terms, service content, geographical information, and delivery cycle. Based on the preset task splitting model, multidimensional contract element data is input into the task splitting model to obtain modular tasks; Obtain the preset resource distribution information, allocate modular tasks based on the geographical information and resource distribution information, and construct a task allocation scheme; Monitor project execution status in real time, and optimize task allocation scheme based on the deviation between project execution status and task allocation scheme.

[0084] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0085] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0086] The above-described 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, and should all be included within the protection scope of this application.

Claims

1. A task decomposition method based on multi-dimensional contract elements, characterized in that, The task decomposition method based on multidimensional contract elements includes: Obtain the original contract, and determine multi-dimensional contract element data based on the original contract. The multi-dimensional contract element data includes at least the contract terms, service content, geographical information, and delivery cycle. Based on a preset task splitting model, the multidimensional contract element data is input into the task splitting model to obtain modular tasks; Obtain the preset resource distribution information, and allocate the modular tasks according to the geographical information and the resource distribution information to construct a task allocation scheme; Monitor the project execution status in real time, and optimize the task allocation scheme based on the deviation between the project execution status and the task allocation scheme.

2. The task decomposition method based on multi-dimensional contract elements according to claim 1, characterized in that, The task decomposition method based on multidimensional contract elements also includes: Obtain historical project data, and from the historical project data, obtain historical multidimensional contract element data and the historical task breakdown results corresponding to the historical multidimensional contract element data; Based on the historical multidimensional contract element data and the historical task splitting results, a training dataset is constructed; The initial model is trained using the training dataset to obtain the task splitting model.

3. The task decomposition method based on multi-dimensional contract elements according to claim 1, characterized in that, The step of determining the multidimensional contract element data based on the original contract specifically includes: Based on natural language processing technology, the original contract is parsed to obtain contract element data and corresponding direct constraint relationships; Based on the contract element data and the corresponding direct constraints, a knowledge graph is constructed. Through a rule engine and an external database, cross-clause dependencies and location-based environmental constraints are built in the knowledge graph to generate the multidimensional contract element data.

4. The task decomposition method based on multi-dimensional contract elements according to any one of claims 1-3, characterized in that, The step of inputting the multidimensional contract element data into the task splitting model to obtain modular tasks specifically includes: The multidimensional contract element data is processed by feature vectorization to obtain vectorized information, and the vectorized information is fused by multimodal fusion technology to generate a joint feature matrix; Based on the clustering algorithm, cluster analysis is performed on the joint feature matrix to obtain task clusters; Based on the decision tree and the dependencies in the multidimensional contract element data, the joint feature matrix is ​​parsed to obtain a task priority list. Based on the task priority list and the delivery cycle of the multidimensional contract element data, the time nodes are obtained using the critical path method. The modular task is obtained by associating the task cluster, the task priority list, and the time node.

5. The task decomposition method based on multi-dimensional contract elements according to claim 4, characterized in that, The process of allocating the modular tasks based on the geographical information and resource distribution, and constructing a task allocation scheme, specifically includes: Based on the resource distribution, the modular tasks are matched using a geographic matching algorithm and the regional information to obtain a candidate resource list; A total cost calculation formula is constructed based on the candidate resource list, and constraints are determined according to the delivery cycle in the multidimensional contract element data and the upper limit of resource capacity in the resource distribution. Based on the planning algorithm, the optimal solution of the total cost calculation formula under the constraints is calculated. Based on the optimal solution, the corresponding resources for the modular task are selected from the candidate resource list. Based on the resources and the modular task, the task allocation scheme is generated.

6. The task decomposition method based on multi-dimensional contract elements according to claim 5, characterized in that, The step of optimizing the task allocation scheme based on the deviation between the project execution status and the task allocation scheme specifically includes: Based on the rule engine, the project execution status is compared with the task allocation scheme to determine whether there is a deviation between the project execution status and the task allocation scheme; If there is a deviation between the project execution status and the task allocation scheme, the current project execution status is obtained, and the current project status is used as a new constraint condition for the total cost calculation formula to determine the optimized task allocation scheme. Based on the current project execution status and the optimized task allocation scheme, a visual display interface is generated for presentation on the visual monitoring interface.

7. The task decomposition method based on multi-dimensional contract elements according to claim 2, characterized in that, The task decomposition method based on multidimensional contract elements also includes: The completed task data is used as incremental learning samples, and the task splitting model is updated based on the incremental learning samples.

8. A task decomposition system based on multi-dimensional contract elements, characterized in that, The task decomposition system based on multidimensional contract elements includes: The data determination module is used to obtain the original contract and determine multi-dimensional contract element data based on the original contract. The multi-dimensional contract element data includes at least contract terms, service content, geographical information, and delivery cycle. The task splitting module is used to input the multidimensional contract element data into the task splitting model based on a preset task splitting model to obtain modular tasks; The resource allocation module is used to obtain the preset resource distribution information, allocate the modular tasks according to the geographical information and the resource distribution information, and construct a task allocation scheme. The dynamic optimization module is used to monitor the project execution status in real time and optimize the task allocation scheme based on the deviation between the project execution status and the task allocation scheme.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the task splitting method based on multidimensional contract elements as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the task splitting method based on multidimensional contract elements as described in any one of claims 1 to 7.

Images