An AI-based software project scheduling and supervision system

Abstract

The present invention discloses a software project schedule and supervision system based on AI, which relates to the field of artificial intelligence technology, including optimizing the initial schedule of the project by combining reinforcement learning with simulated annealing algorithm, comprehensively considering multi-dimensional factors between tasks, and generating a preliminary schedule plan; dynamically adjusting the initial schedule of the project based on genetic algorithm, reordering the priority and execution order of tasks, and minimizing project delays and resource waste; using multimodal data fusion to predict the risk of project progress, and identifying potential risks in the initial schedule by integrating various data sources; finally simulating and optimizing the initial schedule plan by generating adversarial networks, generating multiple candidate schedule plans and selecting the optimal schedule plan. The present invention uses generative adversarial networks to generate multiple candidate plans, combines non-dominated sorting with Monte Carlo simulation to screen the optimal schedule, and ensures that the project achieves a comprehensive optimal balance between construction period, cost and risk.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an AI-based software project schedule and supervision system. Background Art

[0002] In recent years, artificial intelligence technology has been gradually introduced into the field of software project management to optimize schedule and resource scheduling. Traditional methods rely on the critical path method (CPM) and the program evaluation and review technique (PERT) to generate scheduling plans through static task dependencies and fixed resource allocation. With the expansion of the scale of complex projects, dynamic scheduling algorithms based on machine learning (such as reinforcement learning and genetic algorithms) are applied to multi-objective optimization to deal with problems such as dynamic adjustment of task priorities and resource competition conflicts. In addition, graph neural networks and multimodal data fusion technologies are gradually used to model complex dependencies between tasks and risk prediction. However, existing technologies mostly focus on a single optimization dimension (such as time or resources), lack a dynamic collaborative optimization mechanism for multi-dimensional factors (time, resources, risks), and have limitations in real-time anomaly detection and global solution robustness.

[0003] The main shortcomings of current technology are reflected in two aspects: first, traditional reinforcement learning algorithms are prone to fall into local optimality in scheduling optimization, and they do not effectively combine heuristic algorithms (such as simulated annealing) to balance exploration and development capabilities, resulting in insufficient adaptability of scheduling solutions to dynamic environments; second, existing resource scheduling models are mostly based on static competition graph analysis, and fail to model dynamic resource sharing relationships between tasks in real time through adaptive network algorithms, resulting in low resource utilization and loss of execution efficiency. For example, although the adjustment strategy based on genetic algorithms can optimize priorities, it lacks closed-loop feedback with the anomaly detection module, making it difficult to respond to sudden delays or resource bottlenecks in a timely manner. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an AI-based software project scheduling and supervision system to solve the problem of insufficient efficiency in comprehensive optimization and real-time adjustment of multi-dimensional factors in software project scheduling under a dynamic environment.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: The present invention provides an AI-based software project schedule and supervision system, which includes a schedule optimization module, which optimizes the initial schedule of the project by combining reinforcement learning with simulated annealing algorithms, comprehensively considers multi-dimensional factors between tasks, and generates a preliminary schedule plan; The resource scheduling module uses an adaptive network algorithm to optimize resource scheduling, analyze resource competition between tasks in real time, and dynamically adjust resource allocation plans to maximize project execution efficiency; Anomaly detection module tracks project progress and detects anomalies through unsupervised learning algorithms to identify potential task delays and resource bottlenecks in advance; Priority adjustment module, which dynamically adjusts the initial schedule of the project based on genetic algorithms, reorders the priority and execution order of tasks, and minimizes project delays and resource waste; The risk prediction module uses multimodal data fusion to predict the risk of project progress and identifies potential risks in the initial schedule by integrating various data sources; The solution optimization module uses a generative adversarial network to perform final simulation and optimization on the initial scheduling solution, generate multiple candidate scheduling solutions and select the optimal scheduling solution.

[0007] As a preferred solution of the AI-based software project schedule and supervision system of the present invention, the specific steps of generating a preliminary schedule plan are as follows: Receive demand information and resource data related to software projects, clean and standardize them, and generate a structured project demand database; Perform multi-dimensional feature extraction on each task, normalize and standardize different types of features, and output multi-dimensional factors of project time, resources, and risks; Use the Q-learning reinforcement learning algorithm to take task scheduling as agent action, define the state space and action space of the project, and establish a scheduling decision model; Continuously adjust task scheduling through interaction with the environment, build reward functions, and select the optimal scheduling action; Based on reinforcement learning, the simulated annealing algorithm is used to further optimize the scheduling plan; Generate neighborhood solutions by adjusting task order and resource allocation, and use acceptance criteria to update the latest scheduling solution to avoid local optimality; Taking multi-dimensional factors into consideration, the updated scheduling plan is further evaluated and adjusted in real time to finally generate a preliminary scheduling plan.

[0008] As a preferred solution of the AI-based software project schedule and supervision system of the present invention, the state space of the project includes the start time, end time, resource usage and task dependency of the task, the action space of the project consists of the scheduling selection of the task, and for each state, the agent selects the execution order and resource allocation of the task; The specific steps of selecting the optimal scheduling action are as follows: Based on the state space of the project, a reward function is constructed by giving positive rewards for timeliness and resource efficiency, and negative rewards for delays and resource waste; The agent performs actions in the current state and adjusts the scheduling plan based on the reward value and state transition fed back by the environment; Through the Q-learning reinforcement learning algorithm, the Q value is updated in combination with the reward value of environmental feedback, the optimal scheduling action is selected, and the scheduling plan is continuously optimized.

[0009] As a preferred solution of the AI-based software project scheduling and monitoring system of the present invention, the specific steps of maximizing the execution efficiency of the project are as follows: Extract the task resource requirements in the preliminary scheduling plan, analyze the resource competition relationship between tasks, and construct a resource competition diagram; Use pre-trained graph neural networks to model resource sharing and dependencies between tasks, and combine deep reinforcement learning algorithms to optimize the resource allocation strategy of tasks in real time; Monitor task resource usage in real time, predict task progress based on pre-trained long short-term memory networks, and adjust resource allocation, dynamically adjusting resources based on progress feedback and prediction results.

[0010] As a preferred solution of the AI-based software project scheduling and monitoring system of the present invention, the specific steps of identifying potential task delays and resource bottlenecks in advance are as follows: In the process of dynamically adjusting resources, the task execution status data is recorded in real time, the task progress, resource consumption, deviation between the estimated and actual completion time are tracked, and a dynamically updated project progress database is established; Extract time series features, resource consumption features and task dependencies from the project progress database, perform normalization and standardization, and output feature data sets; Use the Isolation Forest algorithm to train the feature data set, identify potential patterns of abnormal progress and abnormal resource consumption, and build a detection model for task delays and resource bottlenecks; Each time the task progress is updated, the task abnormality score is evaluated in real time through the task delay and resource bottleneck detection model to identify the risks of delays and resource bottlenecks.

[0011] As a preferred solution of the AI-based software project scheduling and monitoring system of the present invention, wherein: minimizing project delays and resource waste, the specific steps are as follows: Create a genetic algorithm environment, determine the individual encoding method, and generate the initial population; Each individual expresses task sequencing and resource allocation through genetic coding; Evaluate the quality of each scheduling solution and define a fitness function based on project progress, resource utilization, delay penalties, and task priority; Use the tournament selection method to select individuals with higher fitness from the current initial population as parent individuals to participate in crossover and mutation operations; The genes of the parent individuals are exchanged through crossover operation to generate a new scheduling plan; Evaluate the fitness of each generation of individuals and select the best solution; Monitor task execution in real time based on project progress, dynamically adjust scheduling plans, and use genetic algorithms to further optimize scheduling if delays or resource bottlenecks occur.

[0012] As a preferred solution of the AI-based software project schedule and supervision system of the present invention, wherein: the quality of each schedule plan is evaluated and the fitness function is defined, the specific steps are as follows: Collect project task data, clean and standardize it; Extract numerical features from the processed task data as pre-input of the fitness function; According to the proportion of completed tasks and the total number of tasks, the project progress is calculated as the first weight coefficient of the fitness function; Calculate the ratio of actual resource consumption to total resources and evaluate resource utilization as the second weight coefficient of the fitness function; By calculating the difference between the actual completion time and the planned completion time of the task, a penalty value is assigned to each delayed part of the task as the third weight coefficient of the fitness function; According to the priority and actual completion of the task, the priority score is calculated as the fourth weight coefficient of the fitness function; All weight coefficients are weighted and summarized to generate the final fitness function.

[0013] As a preferred solution of the AI-based software project schedule and supervision system of the present invention, the specific steps of identifying potential risks in the initial schedule by integrating various data sources are as follows: The time series features of task execution status are extracted through pre-trained long short-term memory networks, resource dependencies are analyzed using pre-trained graph neural networks, and project progress features are extracted through self-attention mechanisms; Use a multimodal fusion neural network to weight and merge various features and their dependencies to generate a comprehensive feature representation; A multi-layer perceptron is used to receive comprehensive feature representations, perform nonlinear transformation and processing, and output risk scores for tasks and projects; The output process is trained using historical project data and the multi-layer perceptron layer is optimized using the cross entropy loss function; Track task progress in real time and detect delay risks and resource bottlenecks using predicted risk scores.

[0014] As a preferred solution of the AI-based software project scheduling and supervision system of the present invention, the specific steps of extracting the time series characteristics of the task execution status through the pre-trained long short-term memory network are as follows: Based on the time series data of historical task execution status, after normalization and sliding window segmentation, some time steps are randomly masked to construct a self-supervised pre-training task; Taking the masked time step prediction as a self-supervised task, a bidirectional long short-term memory network is constructed, which inputs the masked time series and outputs the numerical prediction of the masked position; Pre-training is performed through the MSE loss function and the Adam optimizer to learn the temporal rules of task states; Remove the regression layer of the pre-trained LSTM network, freeze the first layer parameters and fine-tune the last layer, and extract the hidden state of the last layer as the time series feature of the current task.

[0015] As a preferred solution of the AI-based software project schedule and supervision system of the present invention, the specific steps of generating multiple candidate schedule plans and selecting the optimal schedule plan are as follows: The generator generates a probability matrix of task order and resource allocation, and the discriminator evaluates the feasibility of the initial scheduling plan and predicts multi-objective indicators, and outputs multiple candidate scheduling plans; The candidate solutions are non-dominatedly sorted using the NSGA-II algorithm, retaining the Pareto frontier solution set with the best combination of duration, cost, resource utilization and risk score; Monte Carlo simulation is used to inject task time fluctuations and resource changes, and the robustness scores and resource elasticity indexes of candidate solutions are calculated; Combine mixed integer programming with a visual interface to select the final scheduling solution based on multi-objective weights and constraint penalty coefficients.

[0016] The beneficial effects of the present invention are as follows: by combining reinforcement learning with simulated annealing algorithms to optimize the initial schedule, a preliminary plan that takes into account both timeliness and resource efficiency is generated, providing a reliable basis for subsequent dynamic adjustments. An adaptive network algorithm is used to analyze the resource competition relationship between tasks in real time, dynamically adjust the resource allocation strategy, and maximize the project execution efficiency and resource utilization. An unsupervised learning algorithm is used to track project progress and detect anomalies, identify potential task delays and resource bottlenecks in advance, and reduce project risks. Task priorities and execution orders are dynamically adjusted based on genetic algorithms to minimize project delays and resource waste. Various data sources are integrated through multimodal data fusion technology to accurately identify potential risks in the initial schedule and enhance the robustness of the solution. Generative adversarial networks are used to generate multiple candidate solutions, and the optimal schedule is selected in combination with non-dominated sorting and Monte Carlo simulation to ensure that the project achieves a comprehensive optimal balance between construction period, cost and risk. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0018] Figure 1 It is a module diagram of the AI-based software project scheduling and supervision system in the embodiment.

[0019] Figure 2 Schematic diagram of the function of the abnormality detection module in the embodiment. DETAILED DESCRIPTION

[0020] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0021] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0022] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0023] Reference Figure 1 and Figure 2, is an embodiment of the present invention, which provides an AI-based software project schedule and supervision system, comprising the following steps: The scheduling optimization module uses a combination of reinforcement learning and simulated annealing algorithms to optimize the initial scheduling of the project, comprehensively considers multi-dimensional factors between tasks, and generates a preliminary scheduling plan.

[0024] Specifically, the following steps are included: Receive demand information and resource data related to software projects, clean and standardize them, and generate a structured project demand database; Specifically, collect project-related requirement information from project management tools, documents, and various interfaces, including detailed descriptions of tasks (including workload, priority, and dependencies); start and end dates of tasks; and resource data (including developers, hardware resources, software tools, etc.).

[0025] Data cleaning includes: cleaning the collected raw data (demand information and resource data related to the software project) to ensure that the data is complete and non-duplicate; pre-processing the text of the task description, including removing redundant information, normalizing the format, unifying the date format, etc.; checking whether the dependencies and constraints between tasks are consistent, and eliminating contradictions in the raw data.

[0026] Standardized processing and structuring include: using standardized methods to convert task descriptions, resource information, etc. into structured data formats to ensure that the original data can be quickly read and processed; digitizing information such as resource allocation, task priority, workload, etc., and converting them into structured data that can be directly processed by machine learning algorithms.

[0027] Perform multi-dimensional feature extraction on each task, normalize and standardize different types of features, and output multi-dimensional factors of project time, resources, and risks; It should be noted that for the time characteristics of tasks (such as start time and end time), the minimum-maximum normalization method is used to scale its value range to the interval [0,1] to eliminate the differences in the time scales of different tasks and ensure that the time characteristics of all tasks are compared under the same standard. For the resource requirements and risk factors of tasks, the Z-score standardization method is used to calculate the mean and standard deviation of each feature, and the feature values ​​are converted to a standard normal distribution with a mean of 0 and a standard deviation of 1 to ensure the consistency of the scale of the feature values.

[0028] Use the Q-learning reinforcement learning algorithm to take task scheduling as agent action, define the state space and action space of the project, and establish a scheduling decision model; The state space of the project includes the start time, end time, resource usage, and task dependencies of the task; the action space of the project consists of the scheduling selection of the task. For each state, the agent selects the execution order of the task and the resource allocation; It should be noted that in the reinforcement learning framework, "state" refers to a complete set of information about the project execution environment at a certain moment, which reflects the progress of the current task and the allocation of resources. The state consists of multiple dimensions, including the time schedule of the task, the use of resources, the dependencies between tasks, and potential risk indicators. When the project advances to a certain moment, this information will be captured and form a state snapshot, which not only contains the status of completed tasks, but also reflects the detailed information of ongoing tasks and tasks to be started. Based on this state, the agent will evaluate the current environmental conditions and select the best action according to the preset goals and constraints, such as adjusting the execution order of tasks or reallocating resources. As the project progresses, the state will be continuously updated, and the agent will make corresponding decisions based on the new state, thus forming a dynamic feedback loop to ensure that the project can proceed efficiently and orderly.

[0029] Example description: When "front-end development task resources are over-allocated (state dimension 1) and the dependent back-end interface is delayed (state dimension 3)", the current state will trigger the agent to generate a series of scheduling optimization actions, such as adjusting the order of tasks, giving priority to independent modules (action 1), and reallocating full-stack engineer resources (action 2). The generation of these actions is based on a comprehensive evaluation of the current state to ensure more reasonable resource utilization and more efficient task execution. Through this dynamic response mechanism, the agent can optimize the project schedule in real time according to environmental changes, thereby effectively dealing with problems such as uneven resource allocation and task dependency delays, and promoting the smooth progress of the project.

[0030] Furthermore, the task start time and end time can be represented by specific time points on the timeline.

[0031] Resource usage: It can be expressed by resource utilization rate (such as manpower, equipment, budget).

[0032] Task dependencies: The dependencies between tasks can be represented by a directed acyclic graph (DAG).

[0033] Continuously adjust task scheduling through interaction with the environment, build reward functions, and select the optimal scheduling action; It should be noted that in reinforcement learning, task scheduling as an action of the intelligent agent means that the intelligent agent affects the state of the project by selecting different scheduling plans (i.e., actions), and then evaluates the effect of the action through the reward function to optimize the overall scheduling plan.

[0034] Specifically, the reward function is constructed by comprehensively considering the multi-dimensional factors in the state space of the project (i.e. the start time, end time, resource usage and task dependencies of the task), giving positive rewards for completion timeliness and resource utilization efficiency, and negative rewards for delays and resource waste; the agent performs actions in the current state (selects a scheduling plan) and adjusts the scheduling plan based on the reward value and state transition (the process of the environment transferring from the current state to the next state after the agent performs a certain action) fed back by the environment; through the Q-learning reinforcement learning algorithm, the Q value is updated in combination with the reward value fed back by the environment, the optimal scheduling action is selected, and the scheduling plan is continuously optimized.

[0035] Based on reinforcement learning, the simulated annealing algorithm is used to further optimize the scheduling plan; The simulated annealing algorithm can effectively avoid local optimal solutions through global search capabilities and improve the global optimization of the scheduling solution. Through randomness, the simulated annealing algorithm allows a certain degree of non-optimal solutions during the search process, thereby jumping out of the trap of local optimality and gradually approaching the global optimal solution. This can not only improve the flexibility of task execution order and resource allocation, but also improve the efficiency of resource utilization, ensuring that the scheduling solution has higher adaptability when facing dynamic changes.

[0036] Generate neighborhood solutions by adjusting task order and resource allocation, and use acceptance criteria to update the latest scheduling solution to avoid local optimality; The acceptance criterion refers to the standard used to decide whether to accept a new solution when generating a neighborhood solution. In the simulated annealing algorithm, the role of the acceptance criterion is particularly important.

[0037] Specifically, the acceptance criterion is used to determine whether to accept the generated neighborhood solution (i.e., the adjusted scheduling solution), even if the objective function value of this new solution (such as project delay time, resource waste, etc.) is not better than the current solution. The core role of the acceptance criterion is to avoid falling into the local optimal solution during the neighborhood search process, increase the breadth of the search space, and improve the chance of global optimization.

[0038] Taking into account multi-dimensional factors, further comprehensively evaluate the updated scheduling plan and adjust the plan in real time to finally generate a preliminary scheduling plan; It should be noted that in order to comprehensively consider the multi-dimensional factors of project time, resources and risks, it is necessary to first define the evaluation indicators of each dimension, as follows: Time factor: By monitoring the gap between the actual completion time of the task and the scheduled time, the deviation of the task progress is evaluated; if the task is delayed, the overall progress of the project will also be affected, so the time factor has an important impact on scheduling optimization; Resource factor: including the efficiency of resource allocation to ensure that each task can obtain the required resources (such as manpower, equipment, etc.); resource utilization efficiency directly affects the execution speed of the task and the overall efficiency of the project; Risk factor: Assess potential risks that may cause deviations from the project schedule, such as resource bottlenecks, task dependency conflicts, emergencies, etc.; identify high-risk tasks through risk analysis and take priority measures.

[0039] Furthermore, by quantifying time factors, resource factors, and risk factors, a comprehensive evaluation model is established using the weighted average method or other optimization methods to output the evaluation results of the three dimensions of time, resources, and risk and combine them into a comprehensive score. This comprehensive score reflects the overall performance of the current scheduling plan and provides a basis for subsequent optimization.

[0040] The resource scheduling module uses an adaptive network algorithm to optimize resource scheduling, analyze resource competition between tasks in real time, dynamically adjust resource allocation plans, and maximize project execution efficiency.

[0041] Specifically, the following steps are included: Extract the task resource requirements in the preliminary scheduling plan, analyze the resource competition relationship between tasks, and construct a resource competition diagram; It should be noted that the preliminary scheduling plan is usually presented in Excel, CSV format or Gantt chart in project management tools. If it is an electronic table (such as Excel), the tasks and their resource requirements can be directly extracted from the table through data extraction tools (such as Python's Pandas library). If it is a Gantt chart or a scheduling plan in a project management tool, it is necessary to obtain the resource requirement information of the task through an API interface (for example, using the Microsoft Project API or Asana API).

[0042] For example, suppose that task A requires 3 developers, task B requires 2 test equipment, task C requires 1 test equipment and 2 engineers, and task D requires 1 test equipment. Next, analyze the resource competition relationship between tasks. For example, task A and task B overlap in time and both require test equipment, so there is resource competition between them. Then, further construct a resource competition graph, with tasks A, B, C, and D as nodes, and the edges between them represent their resource competition relationship. In this way, it is possible to clearly identify which tasks have resource conflicts in the same time period, providing a basis for subsequent resource scheduling and optimization.

[0043] It should also be noted that the concept of task resource requirements comes from the resource requirement information of the tasks clearly recorded in the preliminary scheduling plan. Whether it is through data extraction in electronic spreadsheets (such as Excel, CSV) or task resource allocation details obtained through project management tools (such as Gantt charts or API interfaces), specific definitions and quantitative information on the resources required for each task (such as manpower, equipment, time, etc.) are provided.

[0044] Use pre-trained graph neural networks to model resource sharing and dependencies between tasks, and combine deep reinforcement learning algorithms to optimize task resource allocation strategies in real time to improve resource utilization efficiency. Specifically, the specific steps of pre-training graph neural network (GNN) are as follows: First, use the historical data in the structured project requirements database to build a task relationship graph, where nodes represent tasks and edges represent resource sharing or dependencies between tasks. Next, multi-dimensional features are extracted for each task, including factors such as time, resources, and risks, and feature normalization and standardization are performed to ensure the uniformity of input data. Then, use the graph convolutional network (GCN) as the GNN model to learn the complex relationship patterns between tasks through multi-layer information transmission and aggregation. The training goal is to enable GNN to accurately model resource sharing and dependencies between tasks.

[0045] Furthermore, during the training process, the Adam optimizer is used, the learning rate is set to 0.001, and the mean square error (MSE) is used as the loss function to gradually optimize the GNN model parameters through back propagation and gradient descent. After the training is completed, the pre-trained GNN model is saved and used in the subsequent steps to model the relationship between tasks in real time.

[0046] Monitor task resource usage in real time, combine long short-term memory network (LSTM) to predict task progress, and adjust resource allocation, dynamically adjusting resources based on progress feedback and prediction results.

[0047] Specifically, by deploying resource monitoring agents, the CPU, memory, disk I / O and network bandwidth occupancy rates of task nodes are collected at a frequency of seconds, and input into the pre-trained long short-term memory network (LSTM) model (input features include historical resource sequences, task progress curves and external interference factors). The pre-trained long short-term memory network (LSTM) model outputs the probability of progress deviation and key resource bottleneck prediction values ​​within the next hour; based on the difference between the real-time progress and the prediction result, combined with the preset resource elasticity threshold (such as the CPU continuously exceeds the limit by 80% for more than 5 minutes), the dynamic priority scheduling algorithm (such as the improved strategy based on weighted shortest job priority) is used to automatically generate resource reallocation instructions, and the adjusted actual effect data is fed back to the LSTM model for online parameter fine-tuning to form a closed-loop optimization.

[0048] The anomaly detection module tracks project progress and performs anomaly detection through unsupervised learning algorithms to identify potential task delays and resource bottlenecks in advance.

[0049] Specifically, the following steps are included: In the process of dynamically adjusting resources, the task execution status data is recorded in real time, the task progress, resource consumption, deviation between the estimated and actual completion time are tracked, and a dynamically updated project progress database is established; Among them, the task execution status data includes basic task information (such as task identification and type), progress-related information (progress percentage, completion status of key milestones), resource consumption information (CPU, memory, disk I / O, network bandwidth usage), time information (start, expected and actual completion time), resource allocation details (allocation and usage status of various resources), task dependency status (completion status of dependent tasks) and exception information (errors, exceptions and resource competition conflicts).

[0050] Extract time series features, resource consumption features, and task dependencies from project progress data, and perform normalization (such as linear normalization) and standardization (Z-score standardization); Specifically, the following are the specific methods for extracting time series features, resource consumption features and task dependencies: for time series feature extraction, it will be extracted through time interval calculation method, sliding window analysis method and trend fitting method; for resource consumption feature extraction, it will be extracted through statistical analysis method and association analysis method; for task dependency extraction, it will be extracted through database query method, topological sorting method and weight assignment method.

[0051] Use the isolation forest algorithm to train the processed task progress data, identify potential patterns of abnormal progress and abnormal resource consumption, and build a detection model for task delays and resource bottlenecks; Specifically, the feature data set is divided into a training set and a test set, usually in a certain ratio (such as 8:2); then, the relevant parameters of the isolation forest algorithm are initialized, such as the number of trees, subsample size, etc., where the selection of relevant parameters can be adjusted and optimized according to the characteristics of the data and actual needs; then, the training set data is input into the isolation forest algorithm for detection model training, and the isolation forest algorithm will automatically learn the normal patterns in the training set data, build multiple decision trees by randomly selecting features and segmentation values, and calculate the anomaly score of each data point; after the training is completed, the test set data is used to verify and evaluate the trained detection model, and the accuracy, recall rate, F1 value and other evaluation indicators of the detection model are calculated. According to the evaluation results, the detection model parameters are fine-tuned until the detection model performance reaches a satisfactory effect, forming a detection model for task delays and resource bottlenecks.

[0052] Assess task anomaly scores in real time with every task progress update to identify risks of delays and resource bottlenecks.

[0053] The priority adjustment module dynamically adjusts the initial schedule of the project based on genetic algorithms, reorders the priority and execution order of tasks, and minimizes project delays and resource waste.

[0054] Specifically, the following steps are included: Create a genetic algorithm environment, determine the individual encoding method, and generate the initial population; It should be noted that when creating a genetic algorithm environment, you first need to clarify the specific problem to be solved, such as project scheduling optimization, determine the objective function (such as minimizing project delay time, minimizing resource waste, etc.) and various constraints (such as task dependencies, resource limitations, etc.). This is the basis for all subsequent operations. Only when the nature and requirements of the problem are clear can better coding and population generation be carried out.

[0055] Each individual expresses task sequencing and resource allocation through genetic coding; Based on project progress, resource utilization, delay penalty and task priority, evaluate the quality of each scheduling plan and define the fitness function. The specific steps are as follows: Collect project task data, clean and standardize it; Specifically, after collecting the task data of the project, the data is first cleaned, including deleting duplicate records (such as duplicate task IDs or exactly the same task descriptions), filling missing values ​​(such as using forward filling or interpolation when the task start time is missing), and correcting outliers (such as adjusting based on historical data or domain experience when the task duration exceeds the preset threshold); then the cleaned data is standardized, including unifying the time format (such as converting the task start time to the ISO 8601 standard), normalizing numeric fields (such as mapping the task priority to a value between 0 and 1), and encoding categorical variables (such as converting the task status 'in progress' or 'completed' to One-Hot encoding), and finally generating a structured task data set to provide high-quality input for subsequent analysis.

[0056] Extract numerical features from the processed task data as pre-input of the fitness function; According to the proportion of completed tasks and the total number of tasks, the project progress is calculated as the first weight coefficient of the fitness function; Calculate the ratio of actual resource consumption to total resources and evaluate resource utilization as the second weight coefficient of the fitness function; By calculating the difference between the actual completion time and the planned completion time of the task, a penalty value is assigned to each delayed part of the task as the third weight coefficient of the fitness function; Specifically, the process of assigning a penalty value to the delayed part of each task is as follows: first, calculate the difference between the actual completion time and the planned completion time of the task to obtain the delay time (for example, if Task A is planned to be completed in 5 days and actually takes 7 days, the delay time is 2 days); then, according to the preset penalty rule (for example, the penalty value for each day of delay is 0.1), multiply the delay time by the penalty coefficient to obtain the penalty value of each task (for example, if Task A is delayed by 2 days, the penalty value is 0.2); finally, use the penalty value as the third weight coefficient of the fitness function to evaluate the overall performance of the task scheduling plan. The higher the penalty value, the lower the plan score.

[0057] According to the priority and actual completion of the task, the priority score is calculated as the fourth weight coefficient of the fitness function; All weight coefficients are weighted and summarized to generate the final fitness function.

[0058] The fitness function takes into account four key dimensions: project progress, resource utilization, delay penalty, and task priority. This avoids the imbalance caused by focusing on a single factor in traditional methods, such as focusing only on progress while ignoring resource waste, or emphasizing resource utilization but causing project delays. By comprehensively considering these four dimensions, the quality of each scheduling plan can be evaluated more comprehensively and objectively, ensuring that the selected plan performs well in multiple aspects.

[0059] Use the tournament selection method to select individuals with higher fitness from the current initial population as parent individuals to participate in crossover and mutation operations; The genes of the parent individuals are exchanged through crossover operation to generate a new scheduling plan; Specifically, two parent individuals are randomly selected from the current population. These two parent individuals should meet certain selection strategies, such as selection based on fitness values ​​(individuals with higher fitness values ​​have a greater probability of being selected), to ensure that the selected parent individuals have better genetic characteristics. Assume that the two selected parent individuals are Parent1 and Parent2, which represent two different task scheduling schemes, and the genetic code of each individual represents the order of task arrangement or the encoding information of tasks and other related attributes.

[0060] Furthermore, the type of crossover operation and the crossover point position are determined. Common crossover operation types include single-point crossover, multi-point crossover, uniform crossover, etc. Taking single-point crossover as an example, a crossover point position is randomly selected in the gene sequence of Parent1 and Parent2, and this position divides the gene sequence into two parts.

[0061] For example, if the gene code of Parent1 is [1,2,3,4,5,6,7,8,9,10] and the gene code of Parent2 is [10,9,8,7,6,5,4,3,2,1], and the randomly selected crossover point is the 5th gene position, then Parent1 is split into two parts [1,2,3,4,5] and [6,7,8,9,10], and Parent2 is split into [10,9,8,7,6] and [5,4,3,2,1] accordingly.

[0062] Next, perform a gene exchange operation. Exchange the second half of Parent1's genes with the first half of Parent2's genes, and at the same time exchange the second half of Parent2's genes with the first half of Parent1's genes, thereby generating two new individuals, i.e., offspring individuals. According to the above example, the gene code of the generated offspring individual Child1 is [1,2,3,4,5,5,4,3,2,1], and the gene code of the offspring individual Child2 is [10,9,8,7,6,6,7,8,9,10].

[0063] It should be noted that after the generation of offspring individuals, the genes of the offspring individuals need to be checked for legitimacy and repaired. The crossover operation may cause the gene encoding to not meet the constraints of the problem, such as the destruction of the dependency relationship between tasks and unreasonable resource allocation. For the gene parts that do not meet the constraints, adjustments are made according to the specific constraints and repair strategies. For example, if there is a sequential dependency relationship between tasks, and the execution order of a task after the crossover violates the dependency relationship, then the violation of the constraint can be repaired by adjusting the order of the task and its subsequent related tasks to ensure that the generated offspring individuals are a legal and valid scheduling plan at the genetic level.

[0064] Finally, the newly generated legal offspring individuals are added to the population, replacing the individuals with lower fitness values ​​in the original population, so as to keep the population size unchanged and update the genetic diversity of the population. In this way, the genes of the parent individuals are successfully exchanged through the crossover operation, and a new scheduling plan is generated, which provides more possibilities for the subsequent iterative optimization of the genetic algorithm.

[0065] Evaluate the fitness of each generation of individuals and select the best solution; Monitor task execution in real time based on project progress, dynamically adjust scheduling plans, and use genetic algorithms to further optimize scheduling if delays or resource bottlenecks occur.

[0066] The risk prediction module uses multimodal data fusion to predict the risks of project progress and identifies potential risks in the initial scheduling by integrating various data sources.

[0067] Specifically, the following steps are included: The time series features of task execution status are extracted through pre-trained long short-term memory networks, resource dependencies are analyzed using pre-trained graph neural networks, and project progress features are extracted through self-attention mechanisms; Among them, the task execution status refers to the actual dynamic performance of the task in the project, including the progress of the time dimension (such as the completion time of each stage), the actual completion degree (compared with the plan), the dependency relationship with other tasks (such as whether the predecessor task is completed) and the resource consumption (whether the use of manpower and material resources is reasonable). These task execution status data are the basis for subsequent feature extraction (such as LSTM and GCN modeling), and directly affect the ability to predict task delay risks and resource bottlenecks.

[0068] The time series features of task execution status are extracted through the pre-trained long short-term memory network. The specific steps are as follows: Based on the time series data of historical task execution status, after normalization (Min-Max) and sliding window segmentation (window length 30 days, step length 1 day), a part of the time is randomly masked (e.g., masking 15% of the time steps) to construct a self-supervised pre-training task; Taking the masked time step prediction as a self-supervised task, a bidirectional long short-term memory network is constructed, which inputs the masked time series and outputs the numerical prediction of the masked position; Pre-training is performed using the MSE loss function and the Adam optimizer (learning rate 1e-3) to learn the temporal regularity of task states; Remove the regression layer of the pre-trained long short-term memory network, freeze the first layer parameters and fine-tune the last layer (learning rate 1e-4), and extract the last layer hidden state (dimension 256) as the time series feature of the current task.

[0069] Use a multimodal fusion neural network to weight and merge various features and their dependencies to generate a comprehensive feature representation; A multi-layer perceptron (MLP) is used to receive comprehensive feature representations and perform nonlinear transformation and processing to output risk scores for tasks and projects, indicating potential delays. Preferably, MLP can learn the complex relationship between input features by stacking multiple fully connected layers and nonlinear activation functions (such as ReLU and Sigmoid), and is suitable for processing fused high-dimensional features.

[0070] For example, by constructing a multilayer perceptron (MLP) with two hidden layers, the number of neurons in each layer is 128 and 64 respectively, and nonlinear mapping is achieved through the ReLU activation function.

[0071] The output process is trained using historical project data and the multi-layer perceptron layer is optimized using the cross entropy loss function; Track task progress in real time and detect delay risks and resource bottlenecks based on predicted risk scores; The solution optimization module uses a generative adversarial network to perform final simulation and optimization on the initial scheduling solution, generate multiple candidate scheduling solutions and select the optimal scheduling solution.

[0072] Specifically, the following steps are included: The generator generates a probability matrix of task order and resource allocation, and the discriminator evaluates the feasibility of the initial scheduling plan and predicts multi-objective indicators, and outputs multiple candidate scheduling plans; Specifically, the architecture design of the Generative Adversarial Network (GAN) is as follows: Generator: Input: task dependency graph embedding vector + resource constraint encoding; Network structure: Transformer decoder layer (self-attention mechanism generates task sequence probability distribution); Output: genetic encoding of candidate solutions (task order arrangement matrix + resource allocation probability table).

[0073] Discriminator: Input: Gene encoding of candidate solutions; Network structure: Graph convolution branch: parse resource dependencies, time series convolution branch: predict duration and cost; Output: feasibility probability of solution (0-1) + multi-objective prediction value (duration / cost / risk).

[0074] The candidate solutions are non-dominatedly sorted using the NSGA-II algorithm, retaining the Pareto frontier solution set with the best combination of duration, cost, resource utilization and risk score; Specifically, non-dominated sorting includes the first level of screening: eliminating solutions that violate hard constraints (resource excess, dependency break); the second level of sorting: calculating the dominance relationship based on the target vector and dividing the Pareto hierarchy.

[0075] Monte Carlo simulation is used to inject task time fluctuations and resource changes, and the robustness scores and resource elasticity indexes of candidate solutions are calculated; Specifically, the task timing, resource allocation and dependency data of candidate solutions are first loaded from the Pareto frontier solution set generated by the previous multi-objective optimization to build a simulation input template; then, an execution time fluctuation model is defined for each task (based on Beta distribution, parameters α=2, β=5), and availability change rules are configured for resource nodes (Poisson process trigger, λ=0.1 / hour), and a random scenario set containing task time offsets and resource interruption events is generated (1000 samplings by default); then, the candidate solutions are combined with random disturbance scenarios, and the task execution process is simulated through a parallel simulation engine to track progress deviations, resource conflicts and dynamic adjustment records in real time; based on the simulation results, the task delay ratio (number of delayed tasks / total number of tasks) of each solution in all disturbance scenarios is counted to calculate the robustness score, and the resource elasticity index is calculated according to the ratio of successful adjustments after resource conflicts (number of successful adjustments / total number of conflicts); finally, dynamic thresholds of robustness score ≥ 0.7 and resource elasticity index ≥ 0.6 are set to eliminate substandard solutions and retain high-stability candidate sets for subsequent decision-making processes.

[0076] Furthermore, the role of defining the execution time fluctuation model: In project scheduling, the execution time of tasks is often not fixed and may fluctuate due to various factors (such as insufficient resources, unexpected problems, etc.). In order to simulate the project execution process more realistically, it is necessary to define an execution time fluctuation model for each task. By defining the execution time fluctuation, it is possible to predict the possible time deviation of the task in actual execution, so as to more accurately evaluate the robustness of the scheduling plan. The specific steps are as follows: Select distribution model: Beta distribution is selected as the execution time fluctuation model. Beta distribution is a continuous probability distribution suitable for describing random variables in a finite interval. Parameters α=2 and β=5 represent the shape of the distribution. Specifically, this parameter setting means that the task execution time tends to be completed in a shorter time, but there is also a certain degree of volatility.

[0077] Generate random time offset: Based on Beta distribution, a random time offset is generated for each task. This offset represents the deviation between the actual execution time of the task and the planned time. For example, if the planned time is 10 days, the generated offset may be +2 days or -1 day, indicating that the task may be completed earlier or later.

[0078] Configure resource node availability change rules: The availability of resources (such as manpower and equipment) may also change. The document uses the Poisson process to simulate the availability changes of resource nodes. The Poisson process is a model that describes the occurrence of random events over time. The parameter λ=0.1 / hour represents the average number of resource interruptions per hour.

[0079] Generate a random scenario set: Combine the time offset of the task and the resource interruption event to generate multiple random scenarios (default 1000 samples). Each scenario represents a possible project execution situation, including the time deviation of the task and the resource interruption situation.

[0080] It should also be noted that random perturbation scenarios refer to specific scenarios generated after introducing random factors (such as task time fluctuations, resource interruptions, etc.) during the simulation of project execution. Each scenario represents a possible project execution situation, including the actual execution time of the task, resource availability, etc. By generating multiple random perturbation scenarios, the performance of the scheduling solution under different circumstances can be more comprehensively evaluated.

[0081] Combine mixed integer programming with a visualization interface (Gantt chart / heat map) to select the final scheduling solution based on multi-objective weights and constraint penalty coefficients.

[0082] In summary, the present invention optimizes the initial schedule by combining reinforcement learning with simulated annealing algorithms, generates a preliminary plan that takes into account both timeliness and resource efficiency, and provides a reliable basis for subsequent dynamic adjustments. Adaptive network algorithms are used to analyze the resource competition relationship between tasks in real time, dynamically adjust resource allocation strategies, and maximize project execution efficiency and resource utilization. Unsupervised learning algorithms are used to track project progress and detect anomalies, identify potential task delays and resource bottlenecks in advance, and reduce project risks. Task priorities and execution orders are dynamically adjusted based on genetic algorithms to minimize project delays and resource waste. Various data sources are integrated through multimodal data fusion technology to accurately identify potential risks in the initial schedule and enhance the robustness of the solution. Generative adversarial networks are used to generate multiple candidate solutions, and the optimal schedule is screened in combination with non-dominated sorting and Monte Carlo simulation to ensure that the project achieves a comprehensive optimal balance between schedule, cost, and risk.

[0083] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An AI-based software project scheduling and supervision system, characterized by: include, The scheduling optimization module uses a combination of reinforcement learning and simulated annealing algorithms to optimize the initial scheduling of projects, comprehensively considers multi-dimensional factors between tasks, and generates a preliminary scheduling plan; The resource scheduling module uses an adaptive network algorithm to optimize resource scheduling, analyze resource competition between tasks in real time, and dynamically adjust resource allocation plans to maximize project execution efficiency; Anomaly detection module tracks project progress and detects anomalies through unsupervised learning algorithms to identify potential task delays and resource bottlenecks in advance; Priority adjustment module, which dynamically adjusts the initial schedule of the project based on genetic algorithms, reorders the priority and execution order of tasks, and minimizes project delays and resource waste; The risk prediction module uses multimodal data fusion to predict the risk of project progress and identifies potential risks in the initial schedule by integrating various data sources; The solution optimization module uses a generative adversarial network to perform final simulation and optimization on the initial scheduling solution, generate multiple candidate scheduling solutions and select the optimal scheduling solution.

2. The AI-based software project scheduling and monitoring system according to claim 1, characterized in that: The specific steps of generating a preliminary scheduling plan are as follows: Receive demand information and resource data related to software projects, clean and standardize them, and generate a structured project demand database; Perform multi-dimensional feature extraction on each task, normalize and standardize different types of features, and output multi-dimensional factors of project time, resources, and risks; Use the Q-learning reinforcement learning algorithm to take task scheduling as agent action, define the state space and action space of the project, and establish a scheduling decision model; Continuously adjust task scheduling through interaction with the environment, build reward functions, and select the optimal scheduling action; Based on reinforcement learning, the simulated annealing algorithm is used to further optimize the scheduling plan; Generate neighborhood solutions by adjusting task order and resource allocation, and use acceptance criteria to update the latest scheduling solution to avoid local optimality; Taking multi-dimensional factors into consideration, the updated scheduling plan is further evaluated and adjusted in real time to finally generate a preliminary scheduling plan.

3. The AI-based software project scheduling and monitoring system according to claim 2, characterized in that: The state space of the project includes the start time, end time, resource usage and task dependencies of the task, and the action space of the project consists of the scheduling selection of the task. For each state, the agent selects the execution order of the task and the resource allocation; The specific steps of selecting the optimal scheduling action are as follows: Based on the state space of the project, a reward function is constructed by giving positive rewards for timeliness and resource efficiency, and negative rewards for delays and resource waste; The agent performs actions in the current state and adjusts the scheduling plan based on the reward value and state transition fed back by the environment; Through the Q-learning reinforcement learning algorithm, the Q value is updated in combination with the reward value of environmental feedback, the optimal scheduling action is selected, and the scheduling plan is continuously optimized.

4. The AI-based software project scheduling and monitoring system according to claim 3, characterized in that: The specific steps to maximize the efficiency of project execution are as follows: Extract the task resource requirements in the preliminary scheduling plan, analyze the resource competition relationship between tasks, and construct a resource competition diagram; Use pre-trained graph neural networks to model resource sharing and dependencies between tasks, and combine deep reinforcement learning algorithms to optimize the resource allocation strategy of tasks in real time; Monitor task resource usage in real time, predict task progress based on pre-trained long short-term memory networks, and adjust resource allocation, dynamically adjusting resources based on progress feedback and prediction results.

5. The AI-based software project scheduling and monitoring system according to claim 4, characterized in that: The specific steps for identifying potential task delays and resource bottlenecks in advance are as follows: In the process of dynamically adjusting resources, the task execution status data is recorded in real time, the task progress, resource consumption, deviation between the estimated and actual completion time are tracked, and a dynamically updated project progress database is established; Extract time series features, resource consumption features and task dependencies from the project progress database, perform normalization and standardization, and output feature data sets; Use the Isolation Forest algorithm to train the feature data set, identify potential patterns of abnormal progress and abnormal resource consumption, and build a detection model for task delays and resource bottlenecks; Each time the task progress is updated, the task abnormality score is evaluated in real time through the task delay and resource bottleneck detection model to identify the risks of delays and resource bottlenecks.

6. The AI-based software project scheduling and monitoring system according to claim 5, characterized in that: The specific steps to minimize project delays and resource waste are as follows: Create a genetic algorithm environment, determine the individual encoding method, and generate the initial population; Each individual expresses task sequencing and resource allocation through genetic coding; Evaluate the quality of each scheduling solution and define a fitness function based on project progress, resource utilization, delay penalties, and task priority; Use the tournament selection method to select individuals with higher fitness from the current initial population as parent individuals to participate in crossover and mutation operations; The genes of the parent individuals are exchanged through crossover operation to generate a new scheduling plan; Evaluate the fitness of each generation of individuals and select the best solution; Monitor task execution in real time based on project progress, dynamically adjust scheduling plans, and use genetic algorithms to further optimize scheduling if delays or resource bottlenecks occur.

7. The AI-based software project scheduling and monitoring system according to claim 6, characterized in that: The specific steps of evaluating the quality of each scheduling plan and defining the fitness function are as follows: Collect project task data, clean and standardize it; Extract numerical features from the processed task data as pre-input of the fitness function; According to the proportion of completed tasks and the total number of tasks, the project progress is calculated as the first weight coefficient of the fitness function; Calculate the ratio of actual resource consumption to total resources and evaluate resource utilization as the second weight coefficient of the fitness function; By calculating the difference between the actual completion time and the planned completion time of the task, a penalty value is assigned to each delayed part of the task as the third weight coefficient of the fitness function; According to the priority and actual completion of the task, the priority score is calculated as the fourth weight coefficient of the fitness function; All weight coefficients are weighted and summarized to generate the final fitness function.

8. The AI-based software project scheduling and monitoring system according to claim 7, characterized in that: The specific steps of identifying potential risks in initial scheduling by integrating various data sources are as follows: The time series features of task execution status are extracted through pre-trained long short-term memory networks, resource dependencies are analyzed using pre-trained graph neural networks, and project progress features are extracted through self-attention mechanisms; Use a multimodal fusion neural network to weight and merge various features and their dependencies to generate a comprehensive feature representation; A multi-layer perceptron is used to receive comprehensive feature representations, perform nonlinear transformation and processing, and output risk scores for tasks and projects; The output process is trained using historical project data and the multi-layer perceptron layer is optimized using the cross entropy loss function; Track task progress in real time and detect delay risks and resource bottlenecks using predicted risk scores.

9. The AI-based software project scheduling and monitoring system according to claim 8, characterized in that: The specific steps of extracting the time series features of the task execution status through the pre-trained long short-term memory network are as follows: Based on the time series data of historical task execution status, after normalization and sliding window segmentation, some time steps are randomly masked to construct a self-supervised pre-training task; Taking the masked time step prediction as a self-supervised task, a bidirectional long short-term memory network is constructed, which inputs the masked time series and outputs the numerical prediction of the masked position; Pre-training is performed through the MSE loss function and the Adam optimizer to learn the temporal rules of task states; Remove the regression layer of the pre-trained LSTM network, freeze the first layer parameters and fine-tune the last layer, and extract the hidden state of the last layer as the time series feature of the current task.

10. The AI-based software project scheduling and monitoring system according to claim 9, characterized in that: The specific steps of generating multiple candidate scheduling plans and selecting the optimal scheduling plan are as follows: The generator generates a probability matrix of task order and resource allocation, and the discriminator evaluates the feasibility of the initial scheduling plan and predicts multi-objective indicators, and outputs multiple candidate scheduling plans; The candidate solutions are non-dominatedly sorted using the NSGA-II algorithm, retaining the Pareto frontier solution set with the best combination of duration, cost, resource utilization and risk score; Monte Carlo simulation is used to inject task time fluctuations and resource changes, and the robustness scores and resource elasticity indexes of candidate solutions are calculated; Combine mixed integer programming with a visual interface to select the final scheduling solution based on multi-objective weights and constraint penalty coefficients.

Images