A project whole-cycle management platform

Abstract

The application provides a project whole cycle management platform, and belongs to the technical field of project management, and specifically comprises: an early warning rule adjustment module responsible for low performance early warning processing of a project based on evaluation analysis processing results, and adjusting processing of early warning rules by using the influence correlation characteristics; a conclusion processing module responsible for determining the generation mode of the project case of the project by using the project portrait of the project and the adjustment data of the early warning rules; and a knowledge base generation module responsible for constructing a knowledge base according to different project cases and performance evaluation analysis results, and outputting the performance analysis processing results according to user requirements by using the knowledge base, so that the whole cycle management processing of the project of different project portraits is realized.

Description

[0001] This application is a divisional application of the invention patent application with application number CN202511174378.8 and invention title "A Project Lifecycle Management Platform", the parent application being filed on August 21, 2025. Technical Field

[0002] This invention belongs to the field of project management technology, and in particular relates to a project lifecycle management platform. Background Technology

[0003] With the acceleration of enterprise informatization, project management has gradually shifted from offline to online. However, due to the discrepancies in performance evaluation standards and evaluation targets among different projects, and the fact that performance evaluation often involves multiple stages of a project, especially with a large number of projects, the performance evaluation of different projects is quite difficult to handle. Therefore, how to achieve collaborative management of the entire project lifecycle has become an urgent technical problem to be solved.

[0004] To address the aforementioned technical problems, this invention provides a project lifecycle management platform. Summary of the Invention

[0005] The purpose of this invention is to provide a project lifecycle management platform and a performance evaluation processing method.

[0006] To address the aforementioned technical problems, this invention provides a project lifecycle management platform, specifically comprising: Project profile generation module, performance evaluation module, early warning rule adjustment module, project completion processing module, knowledge base generation module; The project profile generation module is responsible for generating project profiles based on project information. The performance evaluation module is responsible for evaluating and analyzing the performance of the project and different personnel based on the project profile, calling performance evaluation rules to obtain evaluation and analysis results, and using future project data to determine the impact correlation characteristics of different personnel's performance. The early warning rule adjustment module is responsible for issuing low-performance early warnings for projects based on the evaluation and analysis results, and for adjusting the early warning rules using the influence correlation characteristics. The adjustment process for the early warning rules specifically includes: using the distribution data of the influence-related characteristics, identifying personnel with influence-related characteristics; designating these personnel as potential performance-influencing personnel, and determining performance-influencing personnel based on the composition data of the influence-related characteristics of these potential performance-influencing personnel; and determining whether the early warning rules need to be adjusted based on the performance-influencing personnel of the project. The project completion processing module is responsible for determining the method of generating project cases for the project based on the project profile and the adjustment data of the early warning rules. The method for determining the generation method of project cases for the project specifically includes: based on the project profile of the project, determining the associated historical projects of the project profile; based on the adjustment data of the early warning rules of the associated historical projects, determining the similarity of the impact of different adjustment times of the early warning rules of the project on the associated features; based on the similarity, determining the feature adjustment times of the early warning rules of the project; and determining the generation method of project cases for the project through the feature adjustment times of the early warning rules of the project. The knowledge base generation module is responsible for building a knowledge base based on different project cases and performance evaluation analysis results, and using the knowledge base to output the performance analysis results according to user needs.

[0007] Preferably, the project information includes project schedule, project hours, project documents, and customer information.

[0008] Preferably, the project profile is determined based on the parsing results of the project information, and specifically includes the project client, project type, and project cycle.

[0009] Preferably, the performance evaluation rules are determined based on preset performance evaluation rules corresponding to the project profile.

[0010] Preferably, the method for determining the impact correlation characteristics of the performance is as follows: The basic information of the personnel is used to determine the professional skills and work experience characteristics of different personnel; Based on the aforementioned professional skills and work experience characteristics, determine the matching degree with the content of the project to be completed; Based on the matching results, the impact-related characteristics of the personnel's performance are determined.

[0011] Preferably, the professional skills characteristics and work experience characteristics are determined based on the personnel information of the personnel.

[0012] Preferably, the low-performance early warning processing for the project includes low man-month performance early warning, over-investment early warning, and bad debt early warning.

[0013] Preferably, the construction of the knowledge base includes: Long Short-Term Memory (LSTM) networks are used to process performance data from different project cases. The processed performance data is then input into a convolutional neural network to perform in-depth mining of the impact correlation features in the performance data of different project profiles, resulting in a model between the impact correlation features and the performance evaluation results. We use stochastic gradient descent and its variants to optimize the model parameters of a long short-term memory network-convolutional neural network, minimizing the loss function of the convolutional neural network that affects the relationship between associated features and performance evaluation results. When the damage function reaches the target constraint, a knowledge base based on the model parameters after training is constructed.

[0014] Preferably, the output of performance analysis results is based on user needs, specifically including: The user's needs are input into the intelligent agent, which automatically generates prompts based on the keywords of the user's needs; Based on the prompt words, determine the influence association features of the project profile corresponding to the prompt words in the knowledge base; The system takes the influencing features, the basic information of the employees corresponding to the prompt words, and the project profile of the corresponding projects as inputs, and outputs the performance analysis results based on the output of the intelligent agent.

[0015] Preferably, the performance analysis results are presented in a user-friendly format, including text reports and charts.

[0016] Other features and advantages will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained through the structures particularly pointed out in the description and the drawings.

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0018] The above and other features and advantages of the present invention will become more apparent from a detailed description of exemplary embodiments thereof with reference to the accompanying drawings.

[0019] Figure 1 This is a framework diagram of a project lifecycle management platform.

[0020] Figure 2 It is a flowchart of data annotation using historical input data and the generation of prompt words.

[0021] Figure 3 It is a flowchart of the method for determining the data processing result. Detailed Implementation

[0022] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the embodiments set forth herein; rather, they are provided so that the invention will be thorough and complete, and the concept of the exemplary embodiments will be fully conveyed to those skilled in the art. The same reference numerals in the drawings denote the same or similar structures, and therefore their detailed description will be omitted.

[0023] The terms “a,” “one,” “the,” and “the” are used to indicate the existence of one or more elements / components / etc.; the terms “including” and “having” are used to indicate an open-ended meaning of inclusion and that there may be other elements / components / etc. in addition to the listed elements / components / etc.

[0024] Example To solve the above technical problems, such as Figure 1 As shown, this invention provides a project lifecycle management platform, specifically including: Project profile generation module, performance evaluation module, early warning rule adjustment module, project completion processing module, knowledge base generation module; The project profile generation module is responsible for generating project profiles based on project information. The performance evaluation module is responsible for evaluating and analyzing the performance of the project and different personnel based on the project profile, calling performance evaluation rules to obtain evaluation and analysis results, and using future project data to determine the impact correlation characteristics of different personnel's performance. The early warning rule adjustment module is responsible for issuing low-performance early warnings for projects based on the evaluation and analysis results, and for adjusting the early warning rules using the influence correlation characteristics. The project completion processing module is responsible for determining the method of generating project cases for the project based on the project profile and the adjustment data of the early warning rules. The knowledge base generation module is responsible for constructing the knowledge base based on different project cases and performance evaluation analysis results. It then utilizes the knowledge base for accounting and quarterly assessments (including task weighting evaluation, supervision performance, development performance, operational performance, and payment collection performance). Furthermore, the performance evaluation rules are determined based on the preset performance evaluation rules corresponding to the project profile.

[0025] Output of performance analysis results based on user needs.

[0026] Furthermore, the project information includes project schedule, project hours, project documents, and client information.

[0027] Specifically, the project profile is determined based on the analysis results of the project information, and includes project clients, project type, and project cycle.

[0028] Specifically, project types include public welfare projects, commercial projects, and innovative projects. Public welfare projects include park construction and public library building, which focus on social benefits. Commercial projects include commercial real estate development and factory construction, which are profit-oriented. Innovative projects include technology research and development or new product development, which have higher risks but greater potential for returns.

[0029] In another embodiment, project types include agricultural projects, industrial projects, service industry projects, infrastructure projects, public utilities, and livelihood projects.

[0030] Specifically, after approval, a project initiation meeting is held to generate a project profile. The project profile is updated according to different stages of the project, including basic project information such as project number, project name, project year, budget amount, client information, signing date, business attributes, and project manager, as well as information such as project performance, project schedule, project hours, project documents, client information, project payment, project cost, completion information, and results and benefits. Contract preparation and formal signing, initial payment (invoicing), etc. are carried out.

[0031] During project execution, the process involves project schedule, project plan (adjusted based on whether an early warning is issued), project supervisor input completion assessment values, monthly and quarterly performance evaluation of the project management team (including task weight assessment, supervisor performance, development performance, operational performance, and payment collection performance, etc.). Furthermore, the performance evaluation rules are determined based on the preset performance evaluation rules corresponding to the project profile. Table 1 lists the project early warning rules as follows: Specifically, the future project data includes the content of projects to be completed.

[0032] Employee performance is evaluated monthly based on their participation in various projects. The monthly salary payment plan is determined according to the employee's performance level. The steps for employee monthly performance evaluation are as follows: Step 1: Quantify the project by setting different calculation formulas according to different types; calculate the project initiation form separately, and determine the project contract accounting amount, the total planned working hours of the project and the proportion of planned completed tasks, and the actual total working hours of the project each month.

[0033] 1. Project contract accounting amount = contract amount * difficulty coefficient; 2. Total planned project man-hours = Calculated amount * 21 / Planned monthly performance per person; 3. The planned completion percentage (β) is confirmed by the project manager's revenue input; 4. Employee working hours (ti) are obtained from employee attendance hours by the system, and employees allocate their attendance hours according to the project schedule; 5. The project's actual total working hours per month = ∑ (all working hours of employees allocated to the project this month).

[0034] Step 2: Establish a project performance distribution mechanism and generate monthly project performance reports.

[0035] 1. The project supervisor evaluates the monthly completion rate (γ) of the project, manually inputting the task completion percentage. 2. The system calculates the actual operational performance achieved; Actual operational performance = Percentage of monthly completed tasks * Calculated amount; 3. The system calculates the actual monthly performance per person. Actual monthly performance = Monthly actual operating performance * 21 / Monthly total working hours; Step 3: The project manager evaluates the team members' task completion percentage and completion rate for the month and generates the employee monthly performance appraisal form.

[0036] 1. The project manager manually inputs the employee's task completion percentage θj and task completion rate δj for the month; 2. The system calculates employee performance (the actual business volume handled by the employee on the project) based on a formula.

[0037] Employee performance = Employee's participation in projects ∑ (Percentage of tasks completed this month * Actual performance completed * Completion rate); Step 4: Standardize employee performance ratings. Rank employee performance scores from highest to lowest and divide them into four categories (A, B, C, and D) according to a certain percentage range. Based on employee performance ratings, employee person-months, and performance targets, determine the monthly salary payment plan and distribute monthly performance-based wages.

[0038] 1. Performance targets are the performance targets stipulated in the annual performance agreement signed by the employee upon joining the company; 2. Person-month = (12 - month of employment) / 12 + (30 - date of employment) / 30; 3. Monthly performance target = performance target / person-month; 4. Performance rating = Monthly operational performance / Monthly performance benchmark value; Table 2 shows the performance rating rules as follows: Step 5: Performance Adjustment. The performance classification is automatically or manually adjusted using the rules engine. The adjusted level is the employee's final performance level.

[0039] Specifically, such as Figure 2 As shown, the method for determining the correlation characteristics of the performance impact is as follows: The basic information of the personnel is used to determine the professional skills and work experience characteristics of different personnel; Based on the aforementioned professional skills and work experience characteristics, determine the matching degree with the content of the project to be completed; Based on the matching results, the impact-related characteristics of the personnel's performance are determined.

[0040] Specifically, the professional skills characteristics and work experience characteristics are determined based on the personnel information.

[0041] It is understood that the matching results are determined based on the matching of the professional skills and work experience characteristics required by the project content to be completed with the user. To illustrate specifically, chip design requires fundamental academic skills, including mastery of semiconductor physics, mathematical modeling, and circuit theory, such as band structure theory, carrier transport mechanisms, MOSFET characteristics, and GAA transistor modeling. Hardware design capabilities are also crucial, encompassing RTL design (e.g., low-power design, asynchronous reset and synchronous release mechanisms), physical implementation (e.g., parasitic parameter extraction), power optimization, and verification (improving formal verification coverage). System-level capabilities are essential, requiring understanding of heterogeneous architecture design, hardware-software co-verification, and system-level thinking to ensure chip compatibility with other parts of the system. Mathematical analysis skills are also necessary, including 3D-IC thermodynamic analysis, GAA transistor modeling, and Chiplet technology. These are core technologies for advanced processes below 5nm. Proficiency in EDA tools (such as Cadence and Synopsys) and programming languages ​​(Verilog / VHDL) is indispensable. If a user lacks relevant chip design experience and has only previously used EDA tools, their academic skills, hardware design capabilities, system-level capabilities, mathematical analysis skills, programming language skills, and work experience are all mismatched. Therefore, these characteristics are identified as influencing factors.

[0042] Furthermore, the low-performance early warning processing for the project includes low man-month performance early warning, over-investment early warning, and bad debt early warning.

[0043] Specifically, such as Figure 3 As shown, the adjustment and processing of the early warning rules includes: Based on the distribution data of the influencing correlation characteristics, identify the personnel who influence the correlation characteristics; Individuals with influence-related characteristics are identified as potential performance influencers. Performance influencers are determined based on the composition data of the influence-related characteristics of potential performance influencers. By analyzing the personnel affected by the project's performance, it can be determined whether adjustments to the early warning rules are necessary.

[0044] Specifically, when the number of personnel affected by the project's performance does not meet the requirements, the risk of low performance per capita is relatively high. Therefore, it is determined that the early warning rules need to be adjusted to promptly and effectively identify the risk of low performance and improve the efficiency of identifying and handling low performance.

[0045] In one possible embodiment, the performance influencer is the person whose number of influence related features does not meet the requirement; in another possible embodiment, the person whose number of influence related features is more than two is the performance influencer.

[0046] It should be noted that if the number of performance-influencing personnel is 5 or more, or if their proportion among all personnel in the project is 0.2 or more, then the number of performance-influencing personnel in the project is determined to be insufficient.

[0047] Understandably, adjustments to the early warning rules will be made, specifically including: The performance warning cycle is adjusted according to a preset ratio. Specifically, in one possible embodiment, the new performance warning cycle is determined based on the product of the original performance warning cycle and the preset ratio. The warning rules are adjusted based on the new performance warning cycle.

[0048] Furthermore, the method for determining the generation method of the project cases is as follows: Based on the project profile of the project, identify the associated historical projects of the project profile; Based on the adjustment data of the warning rules of the associated historical projects, determine the similarity of the impact of different adjustment numbers of the warning rules of the project on the associated features, and determine the feature adjustment number of the warning rules of the project based on the similarity. The method for generating project cases is determined by the number of times the early warning rules of the project are adjusted.

[0049] It should be noted that the associated historical projects are the historical projects in the project cases of the project profile.

[0050] Furthermore, the similarities include the same number of factors affecting the associated features.

[0051] It is understood that the number of feature adjustments refers to the number of adjustments made to the same item that affects the associated feature, which is not found in any other set of adjustments made to the warning rules for related historical items.

[0052] It is understandable that the method for generating project cases is determined by the number of feature adjustments to the project's early warning rules, specifically including: When the number of feature adjustments to the project's early warning rules does not meet the requirements, in one possible embodiment, when the number of feature adjustments to the project's early warning rules is more than 4, project cases will be generated based on performance evaluation data of different personnel in the project, impact correlation features in different performance evaluation processes, and adjustment data of the early warning rules.

[0053] When the feature adjustment count of the project's warning rule meets the requirements, if none of the associated historical projects of the project's user profile have feature adjustment counts, then project cases will be generated based on the performance evaluation data of different personnel in the project, the influence-related features in different performance evaluation processes, and the adjustment data of the warning rule. In other cases, if the number of project cases for the project profile used for knowledge base construction is greater than the preset project case number threshold (more than 20 in one possible embodiment), then the project does not need to be used as a project case for knowledge base update construction. In other cases, project cases will be generated based on the performance evaluation data of different personnel in the project, the influence-related features in different performance evaluation processes, and the adjustment data of the warning rule.

[0054] It should be noted that if the project is not required to be used as a project case for updating and building the knowledge base, the project can be directly processed for completion. This includes project team self-evaluation, project supervisor evaluation, final confirmation of documents (legal, development, deliverables, etc.), completion meeting (status adjustment), internal training, final verification of project operational performance, and project completion archiving. When generating project cases, basic project information, project performance, project schedule, project hours, project documents, client information, project payment, project cost, completion information, and deliverables can also be considered.

[0055] Optionally, the method for determining the generation method of the project cases is as follows: Based on the project profile of the project, the associated historical projects of the project profile are determined. According to the adjustment data of the early warning rules of the associated historical projects, the similarity of the impact of different adjustment times of the early warning rules of the project on the associated features is determined. Based on the similarity, the feature adjustment times of the early warning rules of the project are determined. It should be noted that if the number of associated historical projects in the project profile of the project is less than the preset project number threshold in the above steps, the project case will be generated based on the performance evaluation data of different personnel in the project, the influence correlation characteristics in different performance evaluation processes, and the adjustment data of the early warning rules. Furthermore, in the above steps, if the number of associated historical projects in the project profile of the project is not less than the preset project number threshold, and if the number of feature adjustments to the project's warning rules does not meet the requirements, in one possible embodiment, when the number of feature adjustments to the project's warning rules is more than 4 times, then the project case will be generated based on the performance evaluation data of different personnel in the project, the influence association features in different performance evaluation processes, and the adjustment data of the warning rules.

[0056] Additionally, it is understandable that if there are no feature adjustment counts for the warning rules in the project, then the project does not need to be used as a project case for knowledge base update and construction. However, if there are feature adjustment counts for the warning rules in the project, and if none of the associated historical projects of the project's user profile have feature adjustment counts, then the project case will be generated based on the performance evaluation data of different personnel in the project, the impact-related features in different performance evaluation processes, and the adjustment data of the warning rules. Otherwise, proceed to the next step.

[0057] Based on the similarity of the influence correlation characteristics of different historical projects and the project at different adjustment times, the project correlation factors of different historical projects are determined; It is understood that the project correlation factor is determined based on the proportion of the number of adjustments with the same influence correlation characteristics as the project in all the number of adjustments to the project.

[0058] It should be noted that if the project correlation factors of different related historical projects are all less than the preset correlation factor threshold in the above steps, the project case will be generated based on the performance evaluation data of different personnel in the project, the influence correlation characteristics in different performance evaluation processes, and the adjustment data of the early warning rules. Furthermore, if there are historical projects with a project association factor greater than the preset value, then the project does not need to be used as a project case for knowledge base update and construction. Otherwise, proceed to the next step.

[0059] The method for generating project cases is determined by the number of feature adjustments to the project's early warning rules and the project association factors of different associated historical projects.

[0060] It is understandable that, in the above steps, the maximum value of the project association factor of different related historical projects is used as the basis to determine the preset adjustment number threshold for the number of feature adjustments. When the number of feature adjustments is greater than the preset adjustment number threshold, the project case will be generated based on the performance evaluation data of different personnel in the project, the impact association features in different performance evaluation processes, and the adjustment data of the early warning rules. In other cases, it is not necessary to use the project as a project case for the knowledge base update and construction process.

[0061] Specifically, the larger the maximum value of the project correlation factor, the larger the preset adjustment number threshold, which is determined by the product of the maximum value of the project correlation factor and the preset proportional factor.

[0062] Furthermore, the knowledge base will be constructed, specifically including: Long Short-Term Memory (LSTM) networks are used to process performance data from different project cases. The processed performance data is then input into a convolutional neural network to perform in-depth mining of the impact correlation features in the performance data of different project profiles, resulting in a model between the impact correlation features and the performance evaluation results. We use stochastic gradient descent and its variants to optimize the model parameters of a long short-term memory network-convolutional neural network, minimizing the loss function of the convolutional neural network that affects the relationship between associated features and performance evaluation results. When the damage function reaches the target constraint, a knowledge base based on the model parameters after training is constructed.

[0063] Specifically, the knowledge base is used to determine the correlation between different impact correlation features and the performance evaluation results of projects with different project profiles, which is the impact correlation features of users required for different project profiles.

[0064] Specifically, the variation algorithms include Adagrad and Adadelta.

[0065] It is understood that the target constraints include the loss function being less than a preset threshold or reaching the target number of iterations.

[0066] Furthermore, the output of performance analysis results is based on user needs, specifically including: The user's needs are input into the intelligent agent, which automatically generates prompts based on the keywords of the user's needs. Based on the prompt words, determine the influence association features of the project profile corresponding to the prompt words in the knowledge base; The system takes the influencing features, the basic information of the employees corresponding to the prompt words, and the project profile of the corresponding projects as inputs, and outputs the performance analysis results based on the output of the intelligent agent.

[0067] Understandably, the agent is constructed based on a large model.

[0068] It should be noted that the performance analysis results are presented in a user-friendly format, including text reports and charts.

[0069] In one possible specific embodiment: 3.1 Knowledge Perception Extensive collection of business-related information from both inside and outside the enterprise, i.e., project case studies.

[0070] 3.2 Knowledge Base Construction The raw data collected for knowledge perception is preprocessed by cleaning, transformation, and labeling, and then stored in a certain structure to form a knowledge base.

[0071] 3.2.1 The intelligent agent connects to internal enterprise systems (such as human resource management systems, project management systems, etc.) and external data sources (such as industry data platforms) through interfaces to collect multi-source data related to performance, including basic employee information (age, education, work experience, work skills), historical performance data (past performance scores, project completion status, etc.), work behavior data (working hours, timeliness of task completion, etc.), project-related data (project difficulty coefficient, resource input, etc.), and market environment data (industry dynamics, economic indicators, etc.).

[0072] 3.2.2 The collected data is cleaned to remove duplicates, errors, and missing values; the data is scaled using normalization and standardization to ensure a uniform scale for easier subsequent model processing; missing values ​​in employee performance data are filled using mean imputation or machine learning-based missing value prediction algorithms; numerical data such as employee age and working hours are normalized to scale their range to the [0, 1] interval; high-dimensional features are reduced using methods such as principal component analysis (PCA) to remove redundant features while retaining information that has a significant impact on performance prediction.

[0073] 3.2.3 Use feature engineering techniques to extract and construct features that have a significant impact on performance prediction, such as calculating the average performance growth rate of employees in different projects and the correlation features between work behavior and performance.

[0074] 3.3 Forming a knowledge base It involves dataset preprocessing, algorithm model library, and toolset integration, and supports the storage of structured, semi-structured, and unstructured data in relational databases, graph databases, and other databases.

[0075] 3.3.1 Use Long Short-Term Memory (LSTM) networks to process time-series related performance data, capture the trends and patterns of performance changes over time, such as analyzing the performance fluctuations of employees over multiple consecutive assessment periods, and use this as input to a convolutional neural network; 3.3.2 Combining Convolutional Neural Networks (CNNs) to deeply mine local features in the data, such as extracting professional skill features and work experience features from project-related data and their feature relationships with project profiles. Simultaneously, an attention mechanism is introduced, enabling the model to automatically focus on professional skill features and work experience features that are correlated with the performance evaluation results of project profiles, thereby improving the model's prediction accuracy.

[0076] 3.3.3 During model training, stochastic gradient descent (SGD) and its variants (such as Adagrad, Adadelta, etc.) are used to optimize model parameters, minimizing the loss function (such as mean squared error loss function MSE, cross-entropy loss function, etc.) between the predicted results and actual performance. The model is trained using a large amount of historical data, and its parameters are continuously adjusted to enable it to accurately learn the complex patterns and rules in the data.

[0077] 3.4 Obtaining User Input The platform receives user instructions, questions, or requests through intelligent question answering, and supports text.

[0078] 3.4.1 Interaction method: Receive user instructions through the platform's intelligent question and answer module (such as text input: "Analyze the matching of project profiles between employees and the projects they participate in" or "Predict the performance data of an employee in Q3 of 2025").

[0079] 3.4.2 Requirements Classification: Data forecasting: such as "predicting the performance data of an employee in Q3 of 2025".

[0080] Decision support: such as "analyzing the matching of an employee's project profile with the projects they participate in".

[0081] 3.5 Knowledge Base Invocation The agent retrieves and extracts data from the knowledge base based on prompts, locating the corresponding data table, document, or knowledge entry.

[0082] 3.6 Decision Reasoning The intelligent agent uses algorithms (such as statistical models and machine learning models) from the algorithm model library to analyze and process the called data knowledge (logical reasoning, pattern matching, prediction calculation, etc.), and arrives at solutions or conclusions. The decision reasoning results are presented in a form that is easy for users to understand (text reports, charts, etc.).

[0083] The employee's basic information, influencing features, and project profiles of the projects they participate in are input into the trained agent. The agent outputs a predicted score of the matching between the employee and the project profiles of the projects they participate in, such as 85 points (out of 100), which corresponds to the level of "good" with a probability of 80%.

[0084] It is understandable that the intelligent agent is built after training the large private domain model. Specifically, the basic information of the employee, the influence and correlation features, and the project profile of the projects participated in are used as input features. The output result is the prediction score of the matching between the employee and the project profile of the projects participated in. A training set is constructed, and the intelligent agent is built based on the large model after training the large model.

[0085] In the description of this specification, the terms "an embodiment," "a preferred embodiment," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0086] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. For those skilled in the art, the embodiments of the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments of the present invention should be included within the protection scope of the embodiments of the present invention.

Claims

1. A project lifecycle management platform, characterized in that, Specifically, it includes: Project profile generation module, performance evaluation module, early warning rule adjustment module, project completion processing module, knowledge base generation module; The project profile generation module is responsible for generating project profiles based on project information. The performance evaluation module is responsible for evaluating and analyzing the performance of the project and different personnel based on the project profile, calling performance evaluation rules to obtain evaluation and analysis results, and using future project data to determine the impact correlation characteristics of different personnel's performance. The early warning rule adjustment module is responsible for issuing low-performance early warnings for projects based on the evaluation and analysis results, and for adjusting the early warning rules using the influence correlation characteristics. The adjustment process for the early warning rules specifically includes: using the distribution data of the influence-related characteristics, identifying personnel with influence-related characteristics; designating these personnel as potential performance-influencing personnel, and determining performance-influencing personnel based on the composition data of the influence-related characteristics of these potential performance-influencing personnel; and determining whether the early warning rules need to be adjusted based on the performance-influencing personnel of the project. The project completion processing module is responsible for determining the method of generating project cases for the project based on the project profile and the adjustment data of the early warning rules. The method for determining the generation method of project cases for the project specifically includes: based on the project profile of the project, determining the associated historical projects of the project profile; based on the adjustment data of the early warning rules of the associated historical projects, determining the similarity of the impact of different adjustment times of the early warning rules of the project on the associated features; based on the similarity, determining the feature adjustment times of the early warning rules of the project; and determining the generation method of project cases for the project through the feature adjustment times of the early warning rules of the project. The knowledge base generation module is responsible for building a knowledge base based on different project cases and performance evaluation analysis results, and using the knowledge base to output the performance analysis results according to user needs.

2. The project lifecycle management platform as described in claim 1, characterized in that, The project information includes project schedule, project hours, project documents, and client information.

3. The project lifecycle management platform as described in claim 1, characterized in that, The project profile is determined based on the analysis results of the project information, and specifically includes the project client, project type, and project cycle.

4. The project lifecycle management platform as described in claim 1, characterized in that, The performance evaluation rules are determined based on the preset performance evaluation rules corresponding to the project profile.

5. The project lifecycle management platform as described in claim 1, characterized in that, The method for determining the correlation characteristics of the performance impact is as follows: The basic information of the personnel is used to determine the professional skills and work experience characteristics of different personnel; Based on the aforementioned professional skills and work experience characteristics, determine the matching degree with the content of the project to be completed; Based on the matching results, the impact-related characteristics of the personnel's performance are determined.

6. The project lifecycle management platform as described in claim 5, characterized in that, The professional skills and work experience characteristics are determined based on the personnel information.

7. The project lifecycle management platform as described in claim 1, characterized in that, The project's low-performance early warning system includes low man-month performance warning, over-investment warning, and bad debt warning.

8. The project lifecycle management platform as described in claim 1, characterized in that, The construction of a knowledge base specifically includes: Long Short-Term Memory (LSTM) networks are used to process performance data from different project cases. The processed performance data is then input into a convolutional neural network to perform in-depth mining of the impact correlation features in the performance data of different project profiles, resulting in a model between the impact correlation features and the performance evaluation results. We use stochastic gradient descent and its variants to optimize the model parameters of a long short-term memory network-convolutional neural network, minimizing the loss function of the convolutional neural network that affects the relationship between associated features and performance evaluation results. When the damage function reaches the target constraint, a knowledge base based on the model parameters after training is constructed.

9. The project lifecycle management platform as described in claim 1, characterized in that, The output of performance analysis results based on user needs includes: The user's needs are input into the intelligent agent, which automatically generates prompts based on the keywords of the user's needs; Based on the prompt words, determine the influence association features of the project profile corresponding to the prompt words in the knowledge base; The system takes the influencing features, the basic information of the employees corresponding to the prompt words, and the project profile of the corresponding projects as inputs, and outputs the performance analysis results based on the output of the intelligent agent.

10. The project lifecycle management platform as described in claim 9, characterized in that, The performance analysis results are presented in a user-friendly format, including text reports and charts.

Images