Photovoltaic project report generation method, system and equipment based on agent task planning and storage medium

Abstract

The invention provides a photovoltaic project report generation method, system and device based on agent task planning, and a storage medium. The method comprises the following steps: S1, inputting basic information, technical parameters and financial data of a photovoltaic project; s2, the intelligent agent initializes the current state according to the input data; s3, multi-level task planning architecture; s4, task dependency relationship management and priority adjustment; s5, task optimization and adaptive adjustment based on reinforcement learning; s6, multi-modal content generation and tool calling are carried out; s7, performing global consistency verification and report generation; and S8, optimizing a feedback mechanism. A multi-level task planning architecture enables the report generation process to be more modularized and ordered, and the task execution speed and efficiency are greatly improved; the intelligent agent dynamically adjusts resource allocation according to the task execution condition, and the resource utilization efficiency is improved; a real-time feedback and task re-planning mechanism ensures that timely adjustment can be performed when a problem occurs, and the stability and reliability of report generation are ensured.

Description

Technical Field

[0001] The present invention relates to the technical fields of artificial intelligence and photovoltaic project management, and particularly relates to a method, system, device and storage medium for generating a photovoltaic project report based on agent task planning. Background Art

[0002] In the process of photovoltaic project management, the compilation of project reports is crucial for project planning, execution, and evaluation. These reports usually involve a large amount of text description, technical parameters, charts, and financial analysis. The traditional manual writing and templated generation methods are inefficient and error-prone. Especially when dealing with complex multimodal content (such as text, pictures, and tables), it is difficult to ensure the consistency and professionalism of the reports.

[0003] With the rise of large models, report automatic generation methods based on artificial intelligence have gradually become possible. However, existing large models still face challenges when dealing with complex project reports, such as the management of task dependencies and the coordinated generation of multimodal content. The report generation process often needs to handle highly specialized content, complex data dependencies, and make real-time adjustments during task execution. Therefore, there is an urgent need for an intelligent task planning method that can automatically complete the generation of photovoltaic project reports. Summary of the Invention

[0004] The first object of the present invention is to provide a method for generating a photovoltaic project report based on agent task planning in view of the above-mentioned problems.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] A method for generating a photovoltaic project report based on agent task planning includes the following steps:

[0007] S1: Input the basic information, technical parameters, and financial data of the photovoltaic project;

[0008] S2: The agent initializes the current state according to the input data;

[0009] S3: Multilevel task planning architecture;

[0010] S4: Task dependency management and priority adjustment;

[0011] S5: Task optimization and adaptive adjustment based on reinforcement learning;

[0012] S6: Multimodal content generation and tool invocation;

[0013] S7: Global consistency verification and report generation;

[0014] S8: Feedback mechanism optimization.

[0015] While adopting the above technical solutions, the present invention may also adopt or combine the following technical solutions:

[0016] As a preferred technical solution of the present invention: In step S3, the multi-level task planning includes a global planning layer, a module planning layer, and a task execution layer.

[0017] As a preferred technical solution of the present invention: Step S4 includes the following sub-steps: Analyze the dependency relationships between tasks, and through a priority adjustment mechanism, ensure that important tasks or tasks with strong dependencies are completed first, and dynamically adjust the priorities and execution orders of tasks according to the actual situation.

[0018] As a preferred technical solution of the present invention: In step S5, task optimization and adaptive adjustment include dynamic task priority adjustment and task replanning mechanism.

[0019] As a preferred technical solution of the present invention: In step S6, multi-modal content generation and tool invocation include text generation, image generation, and table generation.

[0020] The second object of the present invention is to provide a photovoltaic project report generation system based on agent task planning.

[0021] To achieve this, the above object of the present invention is realized through the following technical solutions:

[0022] An input module, which is used for the basic information, technical parameters, and financial data of the photovoltaic project;

[0023] An initialization module, which is used to initialize the current state according to the input data;

[0024] A multi-level task planning architecture module, which is used to gradually refine the task planning and perform appropriate resource allocation and execution according to the task type;

[0025] A task dependency relationship management and priority adjustment module, which is used to analyze the dependency relationships between tasks, and through a priority adjustment mechanism, ensure that important tasks or tasks with strong dependencies are completed first, and dynamically adjust the priorities and execution orders of tasks according to the actual situation;

[0026] A task optimization and adaptive adjustment module, which is used to continuously optimize the task planning according to user feedback and task execution data;

[0027] A content generation and tool invocation module, which is used in task planning, where the agent intelligently allocates tasks and selects suitable tools for generation according to different types of content requirements;

[0028] A global consistency verification and report generation module, which is used to perform global consistency verification on the generated report content by the agent after all tasks are completed to ensure the consistency of data and logic among text, pictures, and tables.

[0029] A feedback mechanism optimization module, which is used to continuously learn and optimize the task planning strategy to improve the quality of subsequent report generation.

[0030] The third object of the present invention is to provide an electronic device.

[0031] For this reason, the above object of the present invention is achieved by the following technical solutions:

[0032] An electronic device includes a memory and a processor. An executable program is stored in the memory, and the processor is configured to run the executable program to execute the steps of the photovoltaic project report generation method based on agent task planning as described above.

[0033] The fourth object of the present invention is to provide a non-volatile storage medium.

[0034] For this reason, the above object of the present invention is achieved by the following technical solutions:

[0035] A non-volatile storage medium stores an executable program, and when the executable program is executed by a processor, it realizes the steps of the photovoltaic project report generation method based on agent task planning as described above.

[0036] The present invention provides a photovoltaic project report generation method, system, device, and storage medium based on agent task planning, which has the following beneficial effects: The multi-level task planning architecture makes the report generation process more modular and orderly, greatly improving the speed and efficiency of task execution; The task dependency relationship, priority management, and global consistency verification mechanism effectively avoid data dependency errors and content inconsistency problems, improving the quality of the report; The multi-objective optimization of reinforcement learning-driven adaptive task optimization and dynamic weight adjustment enables the report to better meet the needs of different photovoltaic projects; The multi-modal content generation and intelligent tool invocation ensure that the generation of text, pictures, tables, etc. in the report is more professional and accurate; The agent dynamically adjusts resource allocation according to the task execution situation, improving resource utilization efficiency; The real-time feedback and task replanning mechanism ensure timely adjustment when problems are encountered, guaranteeing the stability and reliability of report generation. Description of the Drawings

[0037] Figure 1The flowchart of the method for generating a photovoltaic project report based on agent task planning provided by the present invention. Detailed implementation manners

[0038] The present invention will be further described in detail with reference to the accompanying drawings and specific embodiments.

[0039] A method for generating a photovoltaic project report based on agent task planning includes the following steps:

[0040] S1: Input the basic information of the photovoltaic project (such as project name, scale, address), technical parameters (such as component type, inverter configuration), and financial data (such as investment cost, return period);

[0041] S2: The agent initializes the current state according to the input data, laying a foundation for task planning;

[0042] S3: Multilevel task planning architecture;

[0043] The agent adopts a multilevel task planning architecture, including a global planning layer, a module planning layer, and a task execution layer, gradually refining the task planning, and making appropriate resource allocation and execution according to the task type.

[0044] Global planning layer: The agent automatically selects a report template according to the project input, determines the overall structure of the report, and plans the layout of each chapter and its content type (text, picture, table, etc.).

[0045] Module planning layer: Further subdivide the specific content generation tasks in each chapter, such as text generation tasks, picture generation tasks, table generation tasks, etc.

[0046] Task execution layer: The agent assigns the tasks to the execution layer, calls the corresponding tools (such as large language models, APIs, data calculation tools) to generate specific content, and executes the tasks according to the priority and dependency relationship.

[0047] S4: Task dependency relationship management and priority adjustment;

[0048] The agent analyzes the dependency relationship between tasks to ensure that the tasks are executed in the correct order. For example, the generation of some tables in the technical chapter depends on the calculation results of meteorological data, and the content generation in the financial analysis chapter is based on the data input of technical parameters. The agent ensures that important tasks or tasks with strong dependencies are completed first through a priority adjustment mechanism, and dynamically adjusts the priority and execution order of tasks according to the actual situation.

[0049] S5: Task optimization and adaptive adjustment based on reinforcement learning;

[0050] The agent adopts a reinforcement learning algorithm and can continuously optimize the task plan based on user feedback and task execution data. The task planning process will be adaptively adjusted according to the task execution efficiency, the accuracy of the generated content, and user feedback to ensure the continuous improvement of the quality and efficiency of the generated report.

[0051] Dynamic task priority adjustment: During the task execution process, the agent automatically adjusts the task priorities according to the dependencies and execution situations of the tasks. For example, if the generated results of some tasks do not meet the requirements, the agent will reallocate resources and adjust the task order to ensure that key content is processed first.

[0052] Task replanning mechanism: During the task execution process, if a task fails due to tool call failure or the generated content does not meet expectations, the agent will automatically replan the task to avoid the impact of task failure on the overall report generation.

[0053] S6: Multi-modal content generation and tool call;

[0054] In task planning, the agent intelligently allocates tasks and selects suitable tools for generation according to different types of content requirements (text, pictures, tables, etc.).

[0055] Text generation: Call the large language model fine-tuned with photovoltaic field expertise to generate text for chapters such as project descriptions and technical solutions.

[0056] Picture generation: Call the map API to generate a site schematic diagram according to the project location, or generate relevant images such as meteorological maps according to project parameters.

[0057] Table generation: Call calculation tools according to technical and financial data to automatically generate table content such as project financial analysis tables and technical parameter tables.

[0058] The agent can dynamically coordinate various content generation tasks to ensure the reasonable distribution and consistency of text, pictures, and tables in the report.

[0059] S7: Global consistency verification and report generation;

[0060] After all tasks are completed, the agent conducts global consistency verification on the generated report content to ensure the data and logic consistency among text, pictures, and tables. For example, whether the project name, technical parameters, financial data, etc. are consistent in different chapters of the report to avoid data inconsistency caused by decentralized task execution.

[0061] S8: Feedback mechanism optimization.

[0062] Through the feedback mechanism, the agent can continuously learn and optimize the task planning strategy to improve the quality of subsequent report generation.

[0063] Specifically, the above-mentioned method for generating a photovoltaic project report based on agent task planning is implemented as follows:

[0064] Example:

[0065] Suppose the user needs to generate a project report for a 33.806 MW photovoltaic power station in Ningbo, Zhejiang Province. The specific steps are as follows:

[0066] S1: The user inputs the basic project information (such as project name, address, scale) and technical parameters (such as component type, inverter selection);

[0067] S2: The agent initializes the current state according to the input data;

[0068] S3: Multi-level task planning architecture:

[0069] In the global planning layer, the agent automatically selects a suitable report template and generates the chapter structure of the report;

[0070] In the module planning layer, the agent assigns specific content generation tasks to each chapter, including text generation, picture generation, and table generation tasks;

[0071] S4: Task dependency management and priority adjustment: The agent executes tasks according to the priority, first generates the content related to technical parameters, and then generates the meteorological analysis and financial tables to ensure that the dependencies are correctly processed;

[0072] S5: Task optimization and adaptive adjustment based on reinforcement learning;

[0073] S6: Multi-modal content generation and tool invocation;

[0074] S7: Global consistency verification and report generation: The agent performs global consistency verification on the generated report to ensure that the content of all chapters matches each other and exports the report;

[0075] S8: Feedback mechanism optimization: Optimize the task planning strategy according to the feedback information.

[0076] A photovoltaic project report generation system based on agent task planning, the system includes the following modules:

[0077] Input module, the input module is used for the basic information, technical parameters, and financial data of the photovoltaic project;

[0078] Initialization module, the initialization module is used to initialize the current state according to the input data;

[0079] Multi-level task planning architecture module, the multi-level task planning architecture module is used to gradually refine the task planning and perform appropriate resource allocation and execution according to the task type;

[0080] Task Dependency Management and Priority Adjustment Module. The task dependency management and priority adjustment module is used to analyze the dependencies between tasks, and through the priority adjustment mechanism, ensure that important tasks or tasks with strong dependencies are completed first, and dynamically adjust the priorities and execution orders of tasks according to the actual situation;

[0081] Task Optimization and Adaptive Adjustment Module. The task optimization and adaptive adjustment module is used to continuously optimize the task plan according to user feedback and task execution data;

[0082] Content Generation and Tool Invocation Module. The content generation and tool invocation module is used in task planning. The intelligent agent intelligently allocates tasks and selects suitable tools for generation according to different types of content requirements;

[0083] Global Consistency Verification and Report Generation Module. The global consistency verification and report generation module is used to, after all tasks are completed, the intelligent agent perform global consistency verification on the generated report content to ensure the data and logic between texts, pictures, and tables are mutually consistent;

[0084] Feedback Mechanism Optimization Module. The feedback mechanism optimization module is used to continuously learn and optimize the task planning strategy to improve the quality of subsequent report generation.

[0085] The present invention also provides an electronic device, including a processor and a memory for storing processor-executable instructions. Among them, when the processor is set to execute the executable instructions, it is to implement the steps of the photovoltaic project report generation method based on intelligent agent task planning described above.

[0086] The present invention also provides a non-volatile storage medium. The non-volatile storage medium stores an executable program, and when the executable program is executed by the processor, it is to implement the steps of the photovoltaic project report generation method based on intelligent agent task planning as described above.

[0087] The above specific embodiments are used to explain the present invention, and are only the preferred embodiments of the present invention, rather than limiting the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and scope of the claims of the present invention fall within the protection scope of the present invention.

Claims

1. A photovoltaic project report generation method based on agent task planning, characterized in that: The steps include: S1: Input basic information, technical parameters and financial data of the PV project; S2: The agent initializes the current state based on the input data; S3: Multi-level mission planning architecture; S4: Task dependency management and priority adjustment; S5: Task optimization and adaptive adjustment based on reinforcement learning; S6: Multimodal content generation and tool invocation; S7: Global consistency check and report generation; S8: Feedback mechanism optimization.

2. The method for generating a photovoltaic project report based on agent task planning according to claim 1, characterized in that: In step S3, the multi-level task planning includes a global planning layer, a module planning layer and a task execution layer.

3. The method for generating a photovoltaic project report based on agent task planning according to claim 1, characterized in that: Step S4 includes the following sub-steps: analyzing the dependencies between tasks, ensuring that important tasks or tasks with strong dependencies are completed first through a priority adjustment mechanism, and dynamically adjusting the priority and execution order of tasks according to actual conditions.

4. The method for generating a photovoltaic project report based on agent task planning according to claim 1, characterized in that: In step S5, task optimization and adaptive adjustment include dynamic task priority adjustment and task re-planning mechanism.

5. The method for generating a photovoltaic project report based on agent task planning according to claim 1, characterized in that: In step S6, multimodal content generation and tool calling include text generation, image generation, and table generation.

6. A photovoltaic project report generation system based on agent task planning, characterized in that: The system includes the following modules: An input module, which is used for basic information, technical parameters, and financial data of a photovoltaic project; An initialization module, the initialization module is used to initialize the current state according to input data; A multi-level task planning architecture module, which is used to gradually refine the task planning and perform appropriate resource allocation and execution according to the task type; A task dependency management and priority adjustment module, which is used to analyze the dependencies between tasks, ensure that important tasks or tasks with strong dependencies are completed first through a priority adjustment mechanism, and dynamically adjust the priority and execution order of tasks according to actual conditions; A task optimization and adaptive adjustment module, which is used to continuously optimize task planning based on user feedback and task execution data; A content generation and tool calling module, which is used in task planning for the agent to intelligently allocate tasks and select appropriate tools for generation according to different types of content requirements; A global consistency check and report generation module, which is used to perform a global consistency check on the generated report content after all tasks are completed to ensure that the data and logic between texts, pictures and tables are consistent with each other; A feedback mechanism optimization module is used to continuously learn and optimize task planning strategies to improve the quality of subsequent report generation.

7. An electronic device, comprising a memory and a processor, characterized in that: An executable program is stored in the memory, and the processor is configured to run the executable program to execute the steps of the method for generating a photovoltaic project report based on agent task planning as claimed in any one of claims 1 to 5.

8. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores an executable program, and when the executable program is executed by the processor, the steps of the method for generating a photovoltaic project report based on agent task planning as claimed in any one of claims 1 to 5 are implemented.

Images