Multi-agent-based PDM system data modeler generation method and system

Abstract

The invention belongs to the technical field of ships, and particularly relates to a multi-agent-based PDM system data modeler generation method, which comprises the steps of acquiring original data, constructing a multi-agent template library, inputting the acquired original data, mapping different types of data into a uniform data structure, analyzing an association relationship among the data, and establishing a data modeler model of a PDM system. The method comprises the following steps of: establishing a communication and cooperation framework among intelligent agents, automatically generating a database table structure document meeting business and performance requirements, converting a table structure into an SQL (Structured Query Language) code, automatically generating a table establishment statement and connecting a database execution code, establishing a database table into a target database, and storing the database table into the database table according to a table requirement. And testing the generated database table and the filling data, and displaying the modeled data form in a front-end system in a table form. According to the method, the modeling efficiency and accuracy can be improved in the data modeling process in the field of automatic ship design.

Description

Technical Field

[0001] The invention belongs to the technical field of ships, and in particular is a data modeler generation method and system of a PDM system based on multi-agents. Background Art

[0002] With the increasing complexity of ship design, traditional data management systems have been unable to meet the complex multi-dimensional data management needs in modern ship design projects. Ship design involves multiple data types such as two-dimensional drawings, three-dimensional models, structural data, equipment selection, etc., and must comply with industry standards and design specifications. In addition, the data associations between different systems are complex, and data modifications are frequent during the design process, resulting in high management and data consistency maintenance costs for existing systems. Traditional PDM data modelers rely on manual data analysis, abstraction, and model building, and are difficult to automatically adapt to dynamic changes and complex associations in the design process. In order to solve these problems, the present invention proposes a method for generating a PDM system data modeler based on multi-agents. Through the adaptive and collaborative capabilities of the multi-agent system, dynamic management and intelligent modeling of ship design data are realized, the efficiency and quality of data modeling are improved, and the complex business needs in the field of ship design are met. Summary of the invention

[0003] In order to solve the problems in the above background technology, the present invention provides a data modeler generation method of a PDM system based on a multi-agent.

[0004] To achieve the above object, the present invention proposes a method for generating a data modeler of a multi-agent PDM system in a first aspect, comprising:

[0005] Step S100, obtaining original data, including original data related to ship design, standards and specifications in the field of ship design, data templates, and functional requirements and performance requirements of the PDM system;

[0006] Step S200, constructing a multi-agent template library, including an analysis agent, a central agent, a model generation agent, a code generation agent, a data collection agent, a data processing agent, a test agent, and a visualization agent, and inputting the acquired raw data;

[0007] Step S300, the analytical agent automatically abstracts and classifies the original data, maps different types of data into a unified data structure, and analyzes the associations between the data;

[0008] Step S400, the central agent automatically calls the specific required agent from the agent template library and builds a communication and collaboration framework between the agents;

[0009] Step S500, the model generation agent automatically generates a database table structure document that meets business and performance requirements according to the requirements and data template;

[0010] Step S600, the code generation agent converts the table structure into SQL code, automatically generates table creation statements and connects to the database execution code, and creates the database table in the target database;

[0011] Step S700, the data collection agent extracts and fills corresponding data from the input raw data according to the generated table structure and requirements, cleans and annotates the collected data, and saves the processed data into the database table according to the form requirements;

[0012] Step S800, the test agent tests the generated database table and the filled data;

[0013] Step S900: the visualization agent displays the modeled data form in a table format on the front-end system so that the user can operate on the data.

[0014] As a technical solution of the present invention, in step S300, a source data model and a target data model are determined, data elements that need to be mapped are identified in the source data model and the target data model, data mapping rules are defined based on the correspondence between the data elements, a programming language or a data integration tool is used to implement the defined mapping rules, and the results of the data mapping are tested and verified to ensure that the data is not lost or damaged during the conversion process and meets the requirements of the target data model.

[0015] As a technical solution of the present invention, in step S100, obtaining the original data includes:

[0016] Collect existing industry data;

[0017] Label target data from existing data;

[0018] The target data is cleaned and classified, and the missing data is interpolated and supplemented.

[0019] As a technical solution of the present invention, in step S600, a code generation model is used to automatically parse the requirements to obtain modeling statements expressed in SQL form, and the language structure tree expressed in SQL form is converted into the language of the target database and the SQL text is output to obtain the corresponding SQL code.

[0020] The second aspect of the present invention proposes a data modeler generation system for a multi-agent PDM system, characterized in that it includes:

[0021] The acquisition module is used to acquire original data, including original data related to ship design, standards and specifications in the field of ship design, data templates, and functional requirements and performance requirements of the PDM system;

[0022] The multi-agent module is used to build a multi-agent template library, including an analytical agent, a central agent, a model generation agent, a code generation agent, a data collection agent, a data processing agent, a test agent, and a visualization agent, and input the acquired raw data;

[0023] The first data processing module is used for the analytical agent to automatically abstract and classify the original data, map different types of data into a unified data structure, and analyze the association relationship between the data;

[0024] A calling module is used for the central agent to automatically call the specific required agent from the agent template library and build a communication and collaboration framework between agents;

[0025] Table structure module, used for the model generation agent to automatically generate database table structure documents that meet business and performance requirements according to requirements and data templates;

[0026] A conversion module, used for the code generation agent to convert the table structure into SQL code, automatically generate table creation statements and connect to the database execution code, and create the database table in the target database;

[0027] The second data processing module is used for the data collection agent to extract and fill corresponding data from the input raw data according to the generated table structure and requirements, clean and annotate the collected data, and save the processed data into the database table according to the form requirements;

[0028] A test module, used for the test agent to test the generated database table and filled data;

[0029] The visualization module is used for the visualization agent to display the modeled data form in a table form on the front-end system so that users can operate on the data.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The present invention can improve the modeling efficiency and accuracy in the data modeling process in the field of automated ship design. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] In order to better understand the present invention, in the PDM system, it is necessary to convert the design BOM tree into the production BOM tree. However, traditional BOM conversion mostly relies on rules, requires a lot of manual adjustments, and is difficult to cover all design changes or specific production requirements. The multi-agent system can dynamically adapt to complex rules and changes, and significantly reduce the workload and improve accuracy through intelligent automatic recognition and conversion. The present invention is further described below.

[0035] The data modeler generation method of the multi-agent PDM system proposed in the first aspect of the present invention includes step S100, obtaining original data, including original data related to ship design, standards and specifications in the field of ship design, data templates, and functional requirements and performance requirements of the PDM system; step S200, building a multi-agent template library, including parsing agents, central agents, model generation agents, code generation agents, data acquisition agents, data processing agents, test agents, and visualization agents, and inputting the acquired original data. Specifically, the original data of the design BOM, the standards in the field of ship design and the structural requirements of the production BOM, as well as the functional requirements and performance standards of the PDM system are input to provide unified target specifications and data conversion requirements for the agents.

[0036] Step S300, the analytical agent automatically abstracts and classifies the original data, maps different types of data into a unified data structure, and analyzes the associations between the data;

[0037] Specifically, the parsing agent calls the large model to automatically abstract and classify the original data of the design BOM, parse the part relationships and hierarchical information, and use the graph neural network model to analyze the correlation between different data items, laying the foundation for subsequent conversion.

[0038] First, determine the source data model and the target data model, identify the data elements that need to be mapped in the source data model and the target data model, define the data mapping rules based on the correspondence between the data elements, use programming languages ​​​​or data integration tools to implement the defined mapping rules, and test and verify the results of the data mapping to ensure that the data is not lost or damaged during the conversion process and meets the requirements of the target data model.

[0039] Step S400: the central agent automatically calls the specific required agent from the agent template library and builds a communication and collaboration framework between the agents.

[0040] Specifically, the object factory pattern is used to automatically create an agent instance for each module, and the code model is called to build the communication system framework.

[0041] Step S500: the model generation agent automatically generates a database table structure document that meets business and performance requirements based on demand and data template.

[0042] Specifically, the model generation agent uses a large language model to re-analyze and organize the design BOM structure. It also uses a large language model to analyze the historical design knowledge graph to fill in the missing tree structure and automatically generate a database table structure document.

[0043] For example, interpolation can be used to fill in the missing values, and linear interpolation or other forms of interpolation can be performed based on the values ​​of adjacent nodes to estimate the missing values. Or the KNN interpolation method can be used to fill in the missing values ​​by finding the K nodes that are most similar to the missing value node and then calculating the mean or weighted mean of these nodes.

[0044] Step S600, the code generation agent converts the table structure into SQL code, automatically generates table creation statements and connects to the database execution code, and creates the database table in the target database;

[0045] The code generation model automatically generates table creation SQL code that adapts to the target database based on the predefined table structure. The code generation agent uses the code generation model to convert the processed BOM structure into a database structure, convert the table structure into SQL code, automatically generate table creation statements and create a BOM table in the database to provide support for the PDM system.

[0046] In step S700, the data acquisition agent extracts corresponding data from the input raw data and fills it into the generated BOM table according to the generated table structure and requirements, cleans and annotates the collected data, and saves the processed data into the database table according to the form requirements to ensure that all data meets the production consistency and specification requirements.

[0047] Step S800: the test agent tests the generated database table and the filled data. The test agent performs function test, performance test and consistency test on the generated BOM database to ensure the availability of the generated BOM data in the PDM system.

[0048] Step S900: The visual agent displays the modeled data form in a table format on the front-end system for users to operate on the data. Specifically, the visual agent displays the final generated BOM in a table format on the front-end of the PDM system, supports user query, screening and modification operations, and provides intuitive data feedback for the design and production teams.

[0049] In step S300, the source data model and the target data model are determined, the data elements that need to be mapped are identified in the source data model and the target data model, the data mapping rules are defined according to the correspondence between the data elements, a programming language or a data integration tool is used to implement the defined mapping rules, and the data mapping results are tested and verified to ensure that the data is not lost or damaged during the conversion process and meets the requirements of the target data model.

[0050] In step S100, the acquisition of raw data includes: collecting existing industry data; marking target data from the existing data; cleaning and classifying the target data, and interpolating and completing missing data.

[0051] In step S600, a conversion table structure identifier is preset, the identifier corresponds to a table structure conversion template, and a table structure conversion function is performed to obtain a language structure number expressed in SQL form, convert the language structure tree expressed in SQL form into the language of the target database and output the SQL text to obtain the corresponding SQL code.

[0052] The second aspect of the present invention proposes a data modeler generation system for a multi-agent PDM system, including: an acquisition module, used to acquire original data, including original data related to ship design, standards and specifications in the field of ship design, data templates, and functional requirements and performance requirements of the PDM system; a multi-agent module, used to build a multi-agent template library, including an analytical agent, a central agent, a model generation agent, a code generation agent, a data acquisition agent, a data processing agent, a test agent, and a visualization agent, and input the acquired original data; a first data processing module, used for the analytical agent to automatically abstract and classify the original data, and at the same time map different types of data into a unified data structure, and analyze the association relationship between the data; a calling module, used for the central agent to automatically call the specific required agent from the agent template library, and match A communication and collaboration framework is established between intelligent agents; a table structure module is used for the model generation agent to automatically generate a database table structure document that meets the business and performance requirements according to the requirements and data templates; a conversion module is used for the code generation agent to convert the table structure into SQL code, automatically generate table creation statements and connect to the database execution code, and create the database table in the target database; a second data processing module is used for the data acquisition agent to extract and fill in the corresponding data from the input raw data according to the generated table structure and requirements, and clean and annotate the collected data, and save the processed data into the database table according to the form requirements; a testing module is used for the test agent to test the generated database table and the filled data; a visualization module is used for the visualization agent to display the modeled data form in a tabular form on the front-end system for users to process the data.

[0053] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A data modeler generation method for a multi-agent based PDM system, characterized in that: include: Step S100, obtaining original data, including original data related to ship design, standards and specifications in the field of ship design, data templates, and functional requirements and performance requirements of the PDM system; Step S200, constructing a multi-agent template library, including an analysis agent, a central agent, a model generation agent, a code generation agent, a data collection agent, a data processing agent, a test agent, and a visualization agent, and inputting the acquired raw data; Step S300, the analytical agent automatically abstracts and classifies the original data, maps different types of data into a unified data structure, and analyzes the associations between the data; Step S400, the central agent automatically calls the specific required agent from the agent template library and builds a communication and collaboration framework between the agents; Step S500, the model generation agent automatically generates a database table structure document that meets business and performance requirements according to the requirements and data template; Step S600, the code generation agent converts the table structure into SQL code, automatically generates table creation statements and connects to the database execution code, and creates the database table in the target database; Step S700, the data collection agent extracts and fills corresponding data from the input raw data according to the generated table structure and requirements, cleans and annotates the collected data, and saves the processed data into the database table according to the form requirements; Step S800, the test agent tests the generated database table and the filled data; Step S900: the visualization agent displays the modeled data form in a table format on the front-end system so that the user can operate on the data.

2. The method for generating a data modeler for a multi-agent PDM system according to claim 1, characterized in that: In step S300, the source data model and the target data model are determined, the data elements that need to be mapped are identified in the source data model and the target data model, the data mapping rules are defined according to the correspondence between the data elements, a programming language or a data integration tool is used to implement the defined mapping rules, and the data mapping results are tested and verified to ensure that the data is not lost or damaged during the conversion process and meets the requirements of the target data model.

3. The method for generating a data modeler for a multi-agent PDM system according to claim 1, characterized in that: In step S100, obtaining the original data includes: Collect existing industry data; Label target data from existing data; The target data is cleaned and classified, and the missing data is interpolated and supplemented.

4. The method for generating a data modeler for a multi-agent PDM system according to claim 1, characterized in that: In step S600, a code generation model is used to automatically parse the table structure, generate modeling statements expressed in SQL form, convert the language structure tree expressed in SQL form into the language of the target database and output the SQL text to obtain the corresponding SQL code.

5. A data modeler generation system for a multi-agent based PDM system, characterized in that: include: The acquisition module is used to acquire original data, including original data related to ship design, standards and specifications in the field of ship design, data templates, and functional requirements and performance requirements of the PDM system; The multi-agent module is used to build a multi-agent template library, including an analytical agent, a central agent, a model generation agent, a code generation agent, a data collection agent, a data processing agent, a test agent, and a visualization agent, and input the acquired raw data; The first data processing module is used for the analytical agent to automatically abstract and classify the original data, map different types of data into a unified data structure, and analyze the association relationship between the data; A calling module is used for the central agent to automatically call the specific required agent from the agent template library and build a communication and collaboration framework between agents; Table structure module, used for the model generation agent to automatically generate database table structure documents that meet business and performance requirements according to requirements and data templates; A conversion module, used for the code generation agent to convert the table structure into SQL code, automatically generate table creation statements and connect to the database execution code, and create the database table in the target database; The second data processing module is used for the data collection agent to extract and fill corresponding data from the input raw data according to the generated table structure and requirements, clean and annotate the collected data, and save the processed data into the database table according to the form requirements; A test module, used for the test agent to test the generated database table and filled data; The visualization module is used for the visualization agent to display the modeled data form in a table form on the front-end system so that users can operate on the data.

Images