Railway engineering implementation construction organization generation method and related assembly

Abstract

The invention discloses a railway engineering implementation construction organization generation method and related components, and the method comprises the steps: building a railway engineering knowledge base which comprises a railway engineering implementation construction specification file and a historical railway engineering implementation construction organization, and carrying out the extraction of key elements through a key element extraction agent; the method comprises the steps that a directory structure is generated through a directory generation agent, and then a directory text generation agent calls related texts from a railway engineering knowledge base based on keywords in the directories to conduct text filling on all the directories in the directory structure. Related contents are extracted from a railway engineering knowledge base by calling a key element extraction agent, a directory generation agent and a directory text generation agent to perform directory generation and text filling of the railway engineering implementation construction organization, so that the compiling efficiency of the railway engineering implementation construction organization is improved; and the accuracy of semantic understanding of the large language model is relatively high, so that the accuracy of the content of the implementation construction organization of the railway engineering is ensured.

Description

Technical Field

[0001] The invention relates to the field of railway engineering, and in particular to a method for generating an implementable construction organization for railway engineering and related components. Background Art

[0002] The implementation construction organization of railway projects is a detailed and quantitative organization and management for specific construction stages, focusing on the specific operations of on-site construction, including construction plans, schedules, resource allocation, on-site management, etc., to ensure that the project is implemented efficiently and safely as planned. The implementation construction organization emphasizes the dynamic monitoring and adjustment of the construction process to adapt to changes in on-site conditions and ensure the quality and progress of the project. When compiling the implementation construction organization of railway projects, first of all, it is necessary to ensure that the construction organization design is completed before the start of the project to avoid the situation of compiling while construction or supplementing after construction. This requires the construction organization to be highly forward-looking and accurately planned; secondly, the construction organization design is a programmatic document guiding project construction. Once approved, it becomes an important basis for project construction management. Therefore, its accuracy and feasibility are extremely high.

[0003] At present, the design and preparation of construction organization mainly adopts manual paper preparation. Since there are many normative documents that need to be consulted during the preparation process, the preparation is difficult and the efficiency is low. In addition, it is difficult to adjust the construction organization after the preparation is completed. Summary of the invention

[0004] The purpose of the present invention is to provide a method for generating an implementable construction organization for railway engineering and related components. By calling an intelligent agent for extracting key elements, an intelligent agent for generating a catalogue and an intelligent agent for generating a catalogue text, relevant contents are extracted from a railway engineering knowledge base to generate a catalogue and fill in the text of the implementable construction organization for railway engineering, thereby improving the generation efficiency of the implementable construction organization for railway engineering. In addition, due to the high accuracy of semantic understanding of the large language model, the accuracy of the content of the implementable construction organization for railway engineering is guaranteed.

[0005] In order to solve the above technical problems, the present invention provides a method for generating a railway engineering implementation construction organization, comprising:

[0006] Establishing a railway engineering knowledge base, wherein the railway engineering knowledge base includes railway engineering implementation construction specification documents and historical railway engineering implementation construction organizations;

[0007] The key elements extraction agent is called to extract the key elements from the pre-implementation data of the railway project using the large language model;

[0008] Based on the key elements, calling a directory generation agent to generate a directory structure of the railway engineering implementation construction organization;

[0009] The directory text generating intelligent body is called to extract the keywords of each directory in the directory structure, and based on the keywords, relevant texts are retrieved from the railway engineering knowledge base to fill the text of each directory in the directory structure to obtain the construction organization.

[0010] Preferably, before calling the key element extraction agent to extract the key elements from the railway engineering implementation pre-data using the large language model, the method further includes:

[0011] Acquire the pre-condition data of the implementation construction organization of the historical railway project, and divide the pre-condition data of the implementation construction organization of the historical railway project into a training set of the key element extraction intelligent agent and a test set of the key element extraction intelligent agent;

[0012] Presetting the key elements of the pre-construction organization data of each historical railway project implementation in the key element extraction agent training set and the key element extraction agent test set;

[0013] Controlling the key element extraction agent to learn the key elements pre-set in the key element extraction agent training set to perform pre-training of the key element extraction agent;

[0014] Controlling the pre-trained key element extraction agent to extract key elements in the key element extraction agent test set using the large language model, and verifying the key elements extracted by the key element extraction agent based on the key elements pre-set in the key element extraction agent test set;

[0015] If after verification, the accuracy of the key element extraction by the key element extraction agent is greater than the preset key element extraction accuracy, then the step of calling the key element extraction agent to use the large language model to extract the key elements from the railway project implementation prerequisite data is entered.

[0016] Preferably, based on the key elements, before calling the directory generating agent to generate the directory structure of the railway engineering implementation construction organization, the method further includes:

[0017] Predetermine the correspondence between the preset key elements of the implementation construction organization of each historical railway project and the existing directory structure, and determine the generated directory agent training set and the generated directory agent test set;

[0018] Controlling the directory generation agent to learn the correspondence between the preset key elements in the directory generation agent training set and the existing directory structure, so as to perform pre-training of the directory generation agent;

[0019] Controlling the pre-trained directory generation agent to use the large language model to generate a corresponding directory structure based on preset key elements in the directory generation agent test set, and verifying the directory structure generated by the directory generation agent based on the correspondence between the preset key elements in the directory generation agent test set and the existing directory structure;

[0020] If after verification the accuracy of the directory structure generated by the directory generating agent based on the preset key elements is greater than the preset directory structure generation accuracy, then the step of calling the directory generating agent to generate the directory structure of the railway engineering implementation construction organization based on the key elements is entered.

[0021] Preferably, before calling the directory text generation agent, extracting keywords of each directory in the directory structure, and retrieving relevant texts from the railway engineering knowledge base based on the keywords to fill the text of each directory in the directory structure to obtain the construction organization, the method further includes:

[0022] Predetermine the correspondence between the keywords of each directory and the text under each directory in the implementation construction organization of each historical railway project, and determine the training set of the directory text agent and the test set of the directory text agent;

[0023] Controlling the catalog text generation agent to learn the correspondence between the keywords of each catalog in each historical railway engineering implementation construction organization in the catalog text generation agent training set, so as to perform pre-training of the catalog text generation agent;

[0024] Control the pre-trained catalog text generation agent to use the large language model to extract test keywords for each catalog in each historical railway engineering implementation construction organization in the catalog text generation agent test set, and retrieve relevant texts from the established railway engineering knowledge base based on the test keywords to fill the catalog structure in each historical railway engineering implementation construction organization in the catalog text generation agent test set;

[0025] Verifying the text filling result of the generated directory text agent based on the correspondence between the texts under the keywords of each directory in the generated directory text agent test set;

[0026] If the accuracy of the text filling result of the directory text generation intelligent agent after verification is greater than the preset text generation accuracy, the step of calling the directory text generation intelligent agent is entered to extract the keywords of each directory in the directory structure, and based on the keywords, relevant texts are retrieved from the railway engineering knowledge base to fill the text of each directory in the directory structure.

[0027] Preferably, a railway engineering knowledge base is established, including:

[0028] Collect railway project implementation construction specification documents and historical railway project implementation construction organization;

[0029] Classify the railway project implementation construction specification documents according to the content of the documents;

[0030] Classifying the text content of each historical railway project implementation construction organization based on the directory structure in each of the historical railway project implementation construction organizations;

[0031] The large language model is used to extract key elements of the knowledge base from the classified text contents of the railway engineering implementation construction specification documents and the classified historical railway engineering implementation construction organizations; so that the directory text generation intelligent agent extracts the keywords of each directory in the directory structure and compares them with the key elements of the knowledge base to retrieve relevant texts from the railway engineering knowledge base to fill the text of each directory in the directory structure.

[0032] Preferably, the large language model is used to extract key elements of the knowledge base from the classified text contents of the railway engineering implementation construction specification documents and the classified historical railway engineering implementation construction organization, including:

[0033] The large language model is used to extract key elements of the knowledge base from the classified text contents of the railway engineering implementation construction specification documents and the classified historical railway engineering implementation construction organizations, and generate knowledge base key element word vectors, so that the directory text generation agent can extract the keywords of each directory in the directory structure and compare them with the knowledge base key element word vectors to retrieve relevant texts from the railway engineering knowledge base to fill the text of each directory in the directory structure.

[0034] Preferably, after calling the directory text generating agent, extracting keywords of each directory in the directory structure, and retrieving relevant texts from the railway engineering knowledge base based on the keywords to fill the text of each directory in the directory structure to obtain the construction organization, the method further includes:

[0035] Traversing the text corresponding to each directory in the directory structure to determine whether each directory includes the corresponding text content;

[0036] If yes, then generate the implementation construction organization of the railway project;

[0037] If not, the intelligent agent for generating the directory text is called to extract keywords from the directory that does not include the corresponding text content, and based on the keywords, relevant texts are retrieved from the railway engineering knowledge base to fill the directory that does not include the corresponding text content with text.

[0038] In order to solve the above technical problems, the present invention provides a railway engineering implementation construction organization generation system, comprising:

[0039] A knowledge base establishment unit, used to establish a railway engineering knowledge base, wherein the railway engineering knowledge base includes railway engineering implementation construction specification documents and historical railway engineering implementation construction organizations;

[0040] The first calling unit is used to call the key element extraction intelligent agent to extract the key elements from the pre-placed data of the railway project implementation by using the large language model;

[0041] The second calling unit is used to call the directory generation intelligent body to generate the directory structure of the railway engineering implementation construction organization based on the key elements;

[0042] The third calling unit is used to call the directory text generation intelligent body, extract the keywords of each directory in the directory structure, and retrieve relevant texts from the railway engineering knowledge base based on the keywords to fill the text of each directory in the directory structure to obtain the construction organization.

[0043] In order to solve the above technical problems, the present invention provides a device for generating an implementable construction organization for a railway project, comprising:

[0044] Memory for storing computer programs;

[0045] The processor is used to implement the steps of the method for generating a feasible construction organization for railway engineering as described above when executing a computer program.

[0046] In order to solve the above technical problems, the present invention provides a non-volatile storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for generating an implementable construction organization for a railway project as described above are implemented.

[0047] The present application provides a method for generating an implementable construction organization for railway engineering and related components. By establishing a railway engineering knowledge base including railway engineering implementable construction specification documents and historical railway engineering implementable construction organizations, a key element extraction intelligent agent is used to extract key elements, so that a directory generation intelligent agent generates a directory structure, and then a directory body generation intelligent agent retrieves relevant texts from the railway engineering knowledge base based on keywords in the directory to fill in the text of each directory in the directory structure. By calling the key element extraction intelligent agent, the directory generation intelligent agent, and the directory body generation intelligent agent to extract relevant content from the railway engineering knowledge base to generate the directory and fill in the text of the implementable construction organization for railway engineering, the efficiency of compiling the implementable construction organization for railway engineering is improved, and because the semantic understanding accuracy of the large language model is high, the accuracy of the content of the implementable construction organization for railway engineering is guaranteed. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0049] Figure 1 A schematic diagram of a flow chart of a method for generating an implementable construction organization for a railway project provided by the present invention;

[0050] Figure 2 A schematic diagram of the structure of a railway engineering implementation construction organization generation system provided by the present invention;

[0051] Figure 3 A schematic diagram of the structure of a railway engineering implementation construction organization generation device provided by the present invention;

[0052] Figure 4 A schematic structural diagram of a non-volatile storage medium provided by the present invention. DETAILED DESCRIPTION

[0053] The core of the present invention is to provide a method for generating an implementable construction organization for railway engineering and related components. By calling an intelligent body for extracting key elements, an intelligent body for generating a catalogue and an intelligent body for generating a catalogue text, relevant contents are extracted from a railway engineering knowledge base to generate a catalogue and fill in the text of the implementable construction organization for railway engineering, thereby improving the generation efficiency of the implementable construction organization for railway engineering. In addition, due to the high accuracy of the semantic understanding of the large language model, the accuracy of the content of the implementable construction organization for railway engineering is guaranteed.

[0054] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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.

[0055] Please refer to Figure 1 , Figure 1 A schematic flow chart of a method for generating a railway engineering implementation construction organization provided by the present invention, the method comprising:

[0056] S11: Establish a railway engineering knowledge base, which includes railway engineering implementation construction specification documents and historical railway engineering implementation construction organizations;

[0057] In the existing technology, when generating the implementation construction organization of railway projects, detailed and quantitative organization management is carried out for specific construction stages based on the construction organization design. It focuses on the specific operations of on-site construction, including construction plans, schedules, resource allocation, on-site management, etc., to ensure that the project is implemented efficiently and safely as planned. The implementation construction organization emphasizes the dynamic monitoring and adjustment of the construction process to adapt to changes in on-site conditions and ensure the quality and progress of the project.

[0058] When compiling the implementation construction organization of railway projects, it is first necessary to ensure that the implementation construction organization of railway projects is completed before the start of the project to avoid the situation where the implementation construction organization of railway projects is generated while the construction is in progress or supplemented after the construction. This requires the implementation construction organization of railway projects to be highly forward-looking and accurately planned. Secondly, the implementation construction organization of railway projects is a programmatic document guiding project construction. Once approved, it becomes an important basis for management. Therefore, the requirements for its accuracy and feasibility are extremely high. In addition, the classification and division of labor of the implementation construction organization of railway projects are complex, involving the coordination of multiple units and departments, and the approval process is cumbersome. These factors increase the difficulty of generating the implementation construction organization of railway projects.

[0059] Since the implementation construction organization of railway engineering is a detailed and quantitative organization and management for specific construction stages. The content of the implementation construction organization is jointly affected by the railway knowledge base data, project data and the compiled production capacity data. Therefore, in order to solve the above technical problems, this application systematically collects various design data and specifications covering the entire process of construction organization design. Based on this, a railway engineering knowledge base is first established, in which the railway engineering knowledge base includes railway engineering implementation construction specification documents and historical railway engineering implementation construction organizations, that is, including the national standard documents stipulated when the railway engineering is implemented, which can specifically include the relevant national and national railway group specifications, acceptance standards and technical regulations, railway engineering construction organization design compilation guidelines and other materials; it also includes the historical railway engineering implementation construction organizations that have been implemented for reference.

[0060] S12: calling the key element extraction agent to extract key elements from the pre-implementation data of the railway project using the large language model;

[0061] Before generating the implementation construction organization of the railway project, the implementation pre-data of the railway project are first obtained, which may include but are not limited to project-related documents approved by the state and the China Railway Group, construction drawing design documents, survey and design contracts, effective contract components and geological survey reports, construction organization surveys and on-site survey related documents and materials, and related bidding and tendering documents and document attachments, etc. Based on this, the specific geographical location, terrain and related requirements of the railway project can be determined, so that the key element extraction agent can extract key elements from the implementation pre-data of the railway project, such as extracting key elements including longitude, latitude, terrain, regional characteristics and transportation conditions. In addition, because in this application, when calling the key element extraction agent to extract key elements, it specifically uses a large language model for extraction. The large language model can not only recognize individual words and sentences, but also understand semantics, and continuously optimize itself to ensure that the key elements extracted by the key element extraction agent meet the generation requirements of the implementation construction organization of the railway project.

[0062] Specifically, for the pre-data, that is, the project data, including the project-related documents approved by the state and the China Railway Corporation, construction drawing design documents, survey and design contracts, effective component documents of the contract and geological survey reports, construction organization investigations and on-site survey related documents and materials and related bidding and tendering documents and document attachments, etc., the basis for generating the implementation construction organization of the railway project is directly derived from these materials, such as the line scope, engineering quantity, geological survey data in the design, etc. of the implementation construction organization of the railway project. This solution uses the key element extraction intelligent agent generated by the large language model to extract key elements, in preparation for generating the directory structure of the implementation construction organization of the railway project.

[0063] It should be noted that by utilizing a large language model to extract key elements, it is ensured that the extracted key elements are closely centered around the professional field of construction organization design.

[0064] S13: Based on the key elements, the catalog generation agent is called to generate the catalog structure of the railway project implementation construction organization;

[0065] After determining the key elements, the directory generation intelligent body first generates the directory structure of the railway project implementation construction organization based on the key elements, that is, first establishes the overall structure of the railway project implementation construction organization, and summarizes the content covered by the railway project implementation construction organization, such as the generation basis of the railway project implementation construction organization, the geographical location of the railway project, specific construction requirements, the classification of projects to be carried out, the land acquisition information of the railway project construction site, environmental protection projects, construction arrangements, etc.

[0066] In addition, when generating the construction organization arrangement chapter of the railway project implementation construction organization, priority must be given to production capacity data. The production capacity data is determined by the production capacity of the construction unit that generates the railway project implementation construction organization, including human resources data, machinery and equipment, construction progress indicators and other factors.

[0067] S14: Calling the directory text generation intelligent body, extracting keywords of each directory in the directory structure, and retrieving relevant texts from the railway engineering knowledge base based on the keywords to fill the text of each directory in the directory structure to obtain the construction organization.

[0068] After calling the Generate Directory Intelligent Agent to generate the directory structure of the railway engineering implementation construction organization, the Generate Directory Text Intelligent Agent is called to fill in the contents of each directory in the directory structure. Specifically, the Generate Directory Text Intelligent Agent extracts the keywords of each directory. For example, if the keywords of the directory are transportation conditions, weather conditions, and construction organization arrangements, then the Generate Directory Text Intelligent Agent will retrieve the relevant text corresponding to the keywords from the railway engineering knowledge base based on each keyword, and fill it in the corresponding directory text. For example, if the keyword of the directory to be filled is weather conditions, then the Generate Directory Text Intelligent Agent will retrieve the weather conditions of the area where the railway project is located from the railway engineering knowledge base, and fill it in the text of the directory. After each directory has been filled with the text, a complete railway engineering implementation construction organization is generated.

[0069] Of course, both the catalog generation agent and the catalog body generation agent can use the large language model to extract keywords and search text to ensure the accuracy of the generated railway project implementation construction organization. Based on this, the large language model is introduced, the algorithm is continuously optimized, and the ability to capture contextual information is enhanced. By introducing a deeper semantic network and attention mechanism, the large language model can more accurately parse complex sentences and implicit meanings, and can improve understanding in specific areas. In addition, continuous training and feedback adjustments are also helping the large language model gradually adapt to the diversity and complexity of human language, thereby improving its semantic understanding ability.

[0070] In addition, the directory text generating agent can first generate an outline of the text based on the keywords of each directory in the directory structure, and then retrieve relevant texts from the railway engineering knowledge base according to the outline to fill the text of each directory in the directory structure to obtain the construction organization.

[0071] Aiming at the time-consuming and error-prone problems in the traditional railway engineering implementation construction organization generation process, this application has realized the rapid extraction of key elements, automatic generation of directory structure and automatic filling of text by establishing a railway engineering knowledge base and extracting key elements intelligent body, generating directory intelligent body and generating directory text intelligent body. It has significantly improved the efficiency of railway engineering implementation construction organization generation, reduced human errors, and ensured the quality and accuracy of railway engineering implementation construction organization plan.

[0072] Secondly, in view of the problem that the implementation organization content of railway engineering is difficult to be closely integrated with the actual needs of the project, the intelligent agent of this application can extract key elements based on the pre-conditioning data of the implementation organization of railway engineering, and generate a targeted implementation organization directory structure of railway engineering based on these key elements and fill in the text. It can better meet the personalized needs of the project, improve the practical application value of the implementation organization plan of railway engineering, and thus ensure the smooth progress of the construction process.

[0073] In summary, in this application, the relevant content is extracted from the railway engineering knowledge base by calling the key element extraction intelligent agent, the directory generation intelligent agent and the directory text generation intelligent agent to generate the directory and fill the text of the railway engineering implementation construction organization, thereby improving the efficiency of compiling the railway engineering implementation construction organization. In addition, due to the high accuracy of the semantic understanding of the large language model, the accuracy of the content of the railway engineering implementation construction organization is guaranteed.

[0074] Based on the above embodiments:

[0075] As a preferred embodiment, before calling the key element extraction agent to extract the key elements from the railway engineering implementation pre-data using the large language model, the method further includes:

[0076] Acquire the pre-condition data of the implementation construction organization of the historical railway project, and divide the pre-condition data of the implementation construction organization of the historical railway project into a training set of the key element extraction agent and a test set of the key element extraction agent;

[0077] Presetting the key elements of the pre-construction organization data of each historical railway project implementation in the key element extraction agent training set and the key element extraction agent test set;

[0078] Controlling the key element extraction intelligent agent to learn the key elements pre-set in the key element extraction intelligent agent training set to perform pre-training of the key element extraction intelligent agent;

[0079] Control the pre-trained key factor extraction agent to extract key factors in the key factor extraction agent test set using the large language model, and verify the key factors extracted by the key factor extraction agent based on the key factors pre-set in the key factor extraction agent test set;

[0080] If the accuracy of the key element extraction by the key element extraction agent after verification is greater than the preset key element extraction accuracy, the step of calling the key element extraction agent to use the large language model to extract the key elements from the railway project implementation prerequisite data is entered.

[0081] In this embodiment, the key element extraction agent is first pre-trained. The pre-training process is to first obtain the prior information of multiple historical railway engineering implementation construction organizations, that is, obtain the prior information of multiple railway engineering implementation construction organizations that have been put into implementation, and divide the prior information of each historical railway engineering implementation construction organization into a training set of the key element extraction agent and a test set of the key element extraction agent. The key elements of the prior information of each historical railway engineering implementation construction organization have been set in advance. For example, relying on the rich experience of industry experts, a list of key elements is screened and determined. The list of key elements covers important data such as the scope of the bid section, the amount of engineering work, the cost of the project, the process and method of the railway engineering, etc. The key element extraction agent is first called to learn the key elements pre-set in the training set of the key element extraction agent, that is, based on the corresponding relationship between the prior information of the historical railway engineering implementation construction organization and the pre-set key elements in the training set of the key element extraction agent, the type of key elements that need to be extracted from the prior information is learned, so as to realize the pre-training of the key element extraction agent. After pre-training, the key element extraction agent is called to extract key elements from the prior data of the implementation construction organization of historical railway projects in the test set of the key element extraction agent, and the key elements extracted by the pre-trained key element extraction agent from the prior data of the implementation construction organization of historical railway projects in the test set of the key element extraction agent are compared with the key elements pre-set in the prior data of the implementation construction organization of historical railway projects in the test set of the key element extraction agent. If the key elements extracted by the pre-trained key element extraction agent are consistent with the pre-set key elements, then the accuracy of the key elements extracted by the pre-trained key element extraction agent is The accuracy is high. If the accuracy of key element extraction by the pre-trained key element extraction agent is greater than the preset key element extraction accuracy, then the pre-trained key element extraction agent can extract the key elements of the prerequisite data of the actual railway engineering implementation construction organization. Otherwise, pre-training is required to continue, such as re-obtaining the prerequisite data of different historical railway engineering implementation construction organizations, and dividing the prerequisite data of the historical railway engineering implementation construction organizations into a training set of the key element extraction agent and a test set of the key element extraction agent for training and testing. The present application does not limit this, and it is sufficient to improve the accuracy of key element extraction by the key element extraction agent.

[0082] As a preferred embodiment, based on the key elements, before calling the directory generating agent to generate the directory structure of the railway engineering implementation construction organization, the method further includes:

[0083] Predetermine the correspondence between the preset key elements of the implementation construction organization of each historical railway project and the existing directory structure, and determine the generated directory agent training set and the generated directory agent test set;

[0084] Controlling the directory generation agent to learn the correspondence between the preset key elements in the directory generation agent training set and the existing directory structure, so as to perform pre-training of the directory generation agent;

[0085] Controlling the pre-trained directory generation agent to use the large language model to generate a corresponding directory structure based on preset key elements in the directory generation agent test set, and verifying the directory structure generated by the directory generation agent based on the correspondence between the preset key elements in the directory generation agent test set and the existing directory structure;

[0086] If the accuracy of the directory structure generated by the generated directory intelligent agent based on the preset key elements after verification is greater than the preset directory structure generation accuracy, then the step of calling the generated directory intelligent agent to generate the directory structure of the railway project implementation construction organization based on the key elements is entered.

[0087] In this embodiment, the directory generation agent is pre-trained to determine the correspondence between the preset key elements of the implementation construction organization of each historical railway project and the existing directory structure. The preset key elements of the implementation construction organization of each historical railway project here can be the key elements pre-determined when the key element extraction agent is trained, and this application does not limit this. Then determine the directory generation agent training set and the directory generation agent test set, wherein the directory generation agent training set and the directory generation agent test set both include the preset key elements of the implementation construction organization of each historical railway project that have been determined and the corresponding relationship between the existing directory structure, call the directory generation agent to learn the corresponding relationship between the preset key elements in the directory generation agent training set and the existing directory structure, so that the directory generation agent learns the corresponding relationship between the preset key elements and the existing directory structure, and learns which directory to generate for different preset key elements, so as to achieve pre-training of the directory generation agent. Then call the pre-trained generated directory agent to generate a directory structure based on the preset key elements of each historical railway engineering implementation construction organization in the generated directory agent test set, by comparing the directory structure generated by the pre-trained generated directory agent with the existing directory structure corresponding to the preset key elements of each historical railway engineering implementation construction organization in the directory agent test set. If the directory structure generated by the pre-trained generated directory agent is consistent with the existing directory structure corresponding to the preset key elements of each historical railway engineering implementation construction organization in the directory agent test set, then the accuracy of the directory structure generated by the pre-trained generated directory agent is higher. If the accuracy of the directory structure generated by the generated directory agent based on the preset key elements is greater than the accuracy of the preset directory structure generation, then the pre-trained generated directory agent can generate a directory structure based on the key elements of the prerequisite data of the actual railway engineering implementation construction organization. Otherwise, pre-training needs to be continued, such as re-acquiring different historical railway engineering implementation construction organizations, and determining their preset key elements and existing directory structures, and re-training and testing the generated directory agent. This application is not limited to this, and it is sufficient to improve the accuracy of the directory structure generated by the generated directory agent.

[0088] As a preferred embodiment, before calling the directory text generation agent, extracting keywords of each directory in the directory structure, and retrieving relevant texts from the railway engineering knowledge base based on the keywords to fill the text of each directory in the directory structure to obtain the construction organization, it also includes:

[0089] Predetermine the correspondence between the keywords of each directory and the text under each directory in the implementation construction organization of each historical railway project, and determine the training set of the directory text agent and the test set of the directory text agent;

[0090] Controlling the catalog text generation agent to learn the correspondence between the keywords of each catalog in each historical railway project implementation construction organization in the catalog text generation agent training set, so as to perform pre-training of the catalog text generation agent;

[0091] Control the pre-trained catalog text generation agent to use the large language model to extract the test keywords of each catalog in the implementation construction organization of each historical railway project in the catalog text generation agent test set, and retrieve relevant texts from the established railway engineering knowledge base based on the test keywords to fill the catalog structure in the implementation construction organization of each historical railway project in the catalog text generation agent test set;

[0092] Based on the correspondence between the keywords of each directory in the directory text agent test set, the text filling result of the generated directory text agent is verified;

[0093] If the accuracy of the text filling result of the generated directory text intelligent body after verification is greater than the preset text generation accuracy, the step of calling the generated directory text intelligent body is entered to extract the keywords of each directory in the directory structure, and based on the keywords, relevant texts are retrieved from the railway engineering knowledge base to fill the text of each directory in the directory structure.

[0094] In this embodiment, the directory text generating agent is trained. Similarly, the implementation construction organizations of various historical railway projects are first determined in advance, and the correspondence between the texts under the keywords of each directory in each historical railway project implementation construction organization is determined. The directory text generating agent training set and the directory text generating agent test set respectively include the correspondence between the texts under the keywords of each directory in different historical railway project implementation construction organizations. The directory text generating agent is first called to learn the correspondence between the texts under the keywords of each directory in the implementation construction organizations of various historical railway projects in the directory text generating agent training set. That is, the directory text generating agent determines which text contents the keywords of different directories need to correspond to. After learning, the pre-trained directory text generating agent is called to retrieve the text contents from the railway engineering knowledge base based on the keywords of each directory in the implementation construction organizations of various historical railway projects in the directory text generating agent test set. Take the relevant content and fill it into each directory, and compare the text content filled by the pre-trained generated directory text intelligent agent with the text content of each directory in the historical railway engineering implementation construction organization in the generated directory text intelligent agent test set. If they are consistent, the accuracy of the text filling result of the generated directory text intelligent agent is relatively large. If the accuracy of the text filling result of the generated directory text intelligent agent is greater than the preset text generation accuracy, the pre-trained generated directory text intelligent agent can extract content from the railway engineering knowledge base based on the various keywords of the directory structure of the actual railway engineering implementation construction organization and fill the text of each directory. Otherwise, pre-training needs to be continued, such as re-acquiring different historical railway engineering implementation construction organizations, and determining the correspondence between each keyword in its directory structure and the text under each directory, and re-training and testing the generated directory text intelligent agent. This application does not limit this, and it is sufficient to improve the accuracy of the text filling of the generated directory text intelligent agent.

[0095] It should be noted that when pre-training the directory text generating agent, the directory text generating agent can also be trained to extract keywords from each directory. Please refer to the training process of the key element extraction agent for details, which will not be repeated in this embodiment.

[0096] As a preferred embodiment, a railway engineering knowledge base is established, including:

[0097] Collect railway project implementation construction specification documents and historical railway project implementation construction organization;

[0098] Classify the railway project implementation construction specification documents according to the content of the documents;

[0099] Classify the text contents of each historical railway project implementation construction organization based on the directory structure in each historical railway project implementation construction organization;

[0100] The large language model is used to extract key elements of the knowledge base from the classified railway engineering implementation construction specification documents and the classified historical railway engineering implementation construction organization texts.

[0101] After extracting the keywords of each directory in the directory structure, the intelligent agent for generating the directory text compares them with the key elements of the knowledge base, and retrieves relevant texts from the railway engineering knowledge base to fill the text of each directory in the directory structure.

[0102] In this embodiment, since there are many contents in the railway engineering knowledge base, each file in the railway engineering knowledge base is also classified, such as classifying the railway engineering implementation construction specification documents according to the content of the documents, and dividing them into relevant national and national railway group specifications, acceptance standards and technical regulations, and railway engineering construction organization design compilation guidelines; and classifying the text content of each historical railway engineering implementation construction organization, such as classifying the text content about the natural environment in each historical railway engineering implementation construction organization into one category, and further classifying it based on the natural environment of different geographical regions in the natural environment, or according to the directory structure of each historical railway engineering implementation construction organization, the text content in each historical railway engineering implementation construction organization is classified and stored and the text is sliced, and the key elements of the knowledge base are extracted from the classified railway engineering knowledge base using a large language model, so as to generate a directory body. The intelligent agent compares the keywords of the directory with the key elements of the knowledge base, and specifically and accurately finds the corresponding text content to fill. Based on this, in this embodiment, a railway engineering knowledge base for survey, design, construction and other stages is constructed based on a large language model, and the railway knowledge base tends to be accurate as the large language model continues to learn.

[0103] It should be noted that the large language model used in this embodiment has memorized countless language rules and usages through a lot of reading and learning. It can split a sentence into individual words or phrases, and then analyze the relationship between each word and phrase, and can understand the relationship between words or phrases from different angles, so as to grasp the meaning of the sentence more comprehensively. In addition, the large language model can also remember the position of the word in the sentence. Because the same word may have completely different meanings in different positions. Through position encoding, the model can distinguish these subtle differences, so as to understand the semantics more accurately. Finally, the large language model will continue to optimize itself. When it encounters a sentence that is difficult to understand, it will adjust the internal parameters so that it can better deal with similar problems next time. This self-learning ability makes the model more and more accurate in understanding semantics. In short, the large language model achieves accurate understanding of semantics through self-attention mechanism, multi-head attention, position encoding and continuous learning. It performs well in tasks such as reading comprehension, text classification, machine translation, and writing summaries, and has become a powerful assistant for writing construction organizations. Therefore, the use of a large language model can accurately understand and transform the design data and various specifications in the railway engineering knowledge base.

[0104] As a preferred embodiment, the large language model is used to extract key elements of the knowledge base from the classified railway engineering implementation construction specification documents and the classified historical railway engineering implementation construction organization texts, including:

[0105] A large language model is used to extract key elements of the knowledge base from the classified railway engineering implementation construction specification documents and the classified historical railway engineering implementation construction organization text contents, and generate knowledge base key element word vectors, so that the directory text generating agent can extract the keywords of each directory in the directory structure and compare them with the knowledge base key element word vectors to retrieve relevant texts from the railway engineering knowledge base to fill the text of each directory in the directory structure.

[0106] Since the prerequisite data of different railway engineering implementation construction organizations are different, the key elements extracted by the key element extraction agent from different prerequisite data are different, and the directory structure generated by the directory generation agent based on the key elements is also different. The keywords of each directory in the directory structure extracted by the directory text generation agent are also different. However, since the text content corresponding to the semantically similar keywords is the same, therefore, in this embodiment, when the key elements of the knowledge base in the classified railway engineering knowledge base are extracted using a large language model, a knowledge base key element word vector is also generated. The word vector is a technology in natural language processing, which is used to represent vocabulary as a real number vector of fixed dimension. This method maps each word to a continuous vector space so that the similarity between words can be quantified by the mathematical distance between the vectors. Therefore, after extracting the keywords of each directory in the directory structure, the directory text generation agent compares them with the knowledge base key element word vector, thereby improving the accuracy of text filling for each directory in the directory structure and the completeness of the text content.

[0107] As a preferred embodiment, after calling the directory text generation agent, extracting keywords of each directory in the directory structure, and retrieving relevant texts from the railway engineering knowledge base based on the keywords to fill the text of each directory in the directory structure to obtain the construction organization, it also includes:

[0108] Traverse the text corresponding to each directory in the directory structure to determine whether each directory includes the corresponding text content;

[0109] If yes, then generate the implementation construction organization of the railway project;

[0110] If not, the directory text generation intelligent agent is called to extract keywords from the directory that does not include the corresponding text content, and based on the keywords, relevant texts are retrieved from the railway engineering knowledge base to fill the text of the directory that does not include the corresponding text content.

[0111] In this embodiment, after obtaining the construction organization, each directory needs to be traversed to determine whether each directory has been filled with text, that is, whether each directory includes the corresponding text content. If so, the real-time construction organization of the railway project can be directly generated. However, if there is a directory that does not include the corresponding text content, it is necessary to call the directory text generation agent to fill the directory with text again to ensure that each directory includes the corresponding text content.

[0112] After the railway engineering implementation construction organization is generated, adaptive modification can be performed manually, that is, professional engineers conduct a detailed review of the railway engineering implementation construction organization output by the large language model, and feedback the verification results to make fine adjustments to ensure the accuracy of the generated railway engineering implementation construction organization. In this way, the present invention not only improves the accuracy of the railway engineering implementation construction organization, but also significantly reduces the need for manual intervention, and improves the automation level of the entire railway engineering implementation construction organization generation process. This technical solution provides an efficient and accurate tool for railway engineering, ensuring the quality and efficiency of the generation of railway engineering implementation construction organization.

[0113] Please refer to Figure 2 , Figure 2 A schematic diagram of a railway engineering implementation construction organization generation system provided by the present invention, the system comprising:

[0114] A knowledge base building unit 21 is used to build a railway engineering knowledge base, which includes railway engineering implementation construction specification documents and historical railway engineering implementation construction organizations;

[0115] The first calling unit 22 is used to call the key element extraction agent to extract the key elements from the pre-positioned data of the railway project implementation by using the large language model;

[0116] The second calling unit 23 is used to call the directory generation intelligent body to generate the directory structure of the railway engineering implementation construction organization based on the key elements;

[0117] The third calling unit 24 is used to call the directory text generation intelligent body, extract the keywords of each directory in the directory structure, and retrieve relevant texts from the railway engineering knowledge base based on the keywords to fill the text of each directory in the directory structure to obtain the construction organization.

[0118] For an introduction to the railway engineering implementation construction organization generation system provided by the present invention, please refer to the above-mentioned method embodiment, and the present invention will not be repeated here.

[0119] Please refer to Figure 3 , Figure 3 A schematic diagram of a railway engineering implementation construction organization generation device provided by the present invention, the device comprising:

[0120] A memory 31, used for storing computer programs;

[0121] The processor 32 is used to implement the steps of the above-mentioned method for generating a feasible construction organization for railway engineering when executing a computer program.

[0122] For an introduction to the railway engineering implementation construction organization generation device provided by the present invention, please refer to the above-mentioned method embodiment, and the present invention will not be repeated here.

[0123] Please refer to Figure 4 , Figure 4 A schematic diagram of the structure of a non-volatile storage medium provided by the present invention, wherein a computer program 42 is stored on the non-volatile storage medium 41, and when the computer program is executed by a processor, the steps of the method for generating an implementable construction organization for a railway project as described above are implemented.

[0124] For an introduction to the non-volatile storage medium 41 provided by the present invention, please refer to the above method embodiment, and the present invention will not be described in detail here.

[0125] It should also be noted that, in this specification, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises", "comprising" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the statement "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0126] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for generating a feasible construction organization for a railway project, characterized in that: include: Establishing a railway engineering knowledge base, wherein the railway engineering knowledge base includes railway engineering implementation construction specification documents and historical railway engineering implementation construction organizations; The key elements extraction agent is called to extract the key elements from the pre-implementation data of the railway project using the large language model; Based on the key elements, calling a directory generation agent to generate a directory structure of the railway engineering implementation construction organization; The directory text generating intelligent body is called to extract the keywords of each directory in the directory structure, and based on the keywords, relevant texts are retrieved from the railway engineering knowledge base to fill the text of each directory in the directory structure to obtain the construction organization.

2. The method for generating a feasible construction organization for a railway project according to claim 1, characterized in that: Before calling the key element extraction agent to extract the key elements from the pre-implementation data of the railway project using the large language model, it also includes: Acquire the pre-condition data of the implementation construction organization of the historical railway project, and divide the pre-condition data of the implementation construction organization of the historical railway project into a training set of the key element extraction intelligent agent and a test set of the key element extraction intelligent agent; Presetting the key elements of the pre-construction organization data of each historical railway project implementation in the key element extraction agent training set and the key element extraction agent test set; Controlling the key element extraction agent to learn the key elements pre-set in the key element extraction agent training set to perform pre-training of the key element extraction agent; Controlling the pre-trained key element extraction agent to extract key elements in the key element extraction agent test set using the large language model, and verifying the key elements extracted by the key element extraction agent based on the key elements pre-set in the key element extraction agent test set; If after verification, the accuracy of the key element extraction by the key element extraction agent is greater than the preset key element extraction accuracy, then the step of calling the key element extraction agent to use the large language model to extract the key elements from the railway project implementation prerequisite data is entered.

3. The method for generating a feasible construction organization for a railway project according to claim 1, characterized in that: Based on the key elements, before calling the directory generation intelligent agent to generate the directory structure of the railway project implementation construction organization, it also includes: Predetermine the correspondence between the preset key elements of the implementation construction organization of each historical railway project and the existing directory structure, and determine the generated directory agent training set and the generated directory agent test set; Controlling the directory generation agent to learn the correspondence between the preset key elements in the directory generation agent training set and the existing directory structure, so as to perform pre-training of the directory generation agent; Controlling the pre-trained directory generation agent to use the large language model to generate a corresponding directory structure based on preset key elements in the directory generation agent test set, and verifying the directory structure generated by the directory generation agent based on the correspondence between the preset key elements in the directory generation agent test set and the existing directory structure; If after verification the accuracy of the directory structure generated by the directory generating agent based on the preset key elements is greater than the preset directory structure generation accuracy, then the step of calling the directory generating agent to generate the directory structure of the railway engineering implementation construction organization based on the key elements is entered.

4. The method for generating a feasible construction organization for a railway project according to claim 1, characterized in that: Before calling the directory text generating agent, extracting keywords of each directory in the directory structure, and retrieving relevant texts from the railway engineering knowledge base based on the keywords to fill the text of each directory in the directory structure to obtain the construction organization, the method further includes: Predetermine the correspondence between the keywords of each directory and the text under each directory in the implementation construction organization of each historical railway project, and determine the training set of the directory text agent and the test set of the directory text agent; Controlling the catalog text generation agent to learn the correspondence between the keywords of each catalog in each historical railway engineering implementation construction organization in the catalog text generation agent training set, so as to perform pre-training of the catalog text generation agent; Control the pre-trained catalog text generation agent to use the large language model to extract test keywords for each catalog in each historical railway engineering implementation construction organization in the catalog text generation agent test set, and retrieve relevant texts from the established railway engineering knowledge base based on the test keywords to fill the catalog structure in each historical railway engineering implementation construction organization in the catalog text generation agent test set; Verifying the text filling result of the generated directory text agent based on the correspondence between the texts under the keywords of each directory in the generated directory text agent test set; If the accuracy of the text filling result of the directory text generation intelligent agent after verification is greater than the preset text generation accuracy, the step of calling the directory text generation intelligent agent is entered to extract the keywords of each directory in the directory structure, and based on the keywords, relevant texts are retrieved from the railway engineering knowledge base to fill the text of each directory in the directory structure.

5. The method for generating a feasible construction organization for a railway project according to claim 1, characterized in that: Establish a railway engineering knowledge base, including: Collect railway project implementation construction specification documents and historical railway project implementation construction organization; Classify the railway project implementation construction specification documents according to the content of the documents; Classifying the text content of each historical railway project implementation construction organization based on the directory structure in each of the historical railway project implementation construction organizations; The large language model is used to extract key elements of the knowledge base from the classified text contents of the railway engineering implementation construction specification documents and the classified historical railway engineering implementation construction organizations; so that the directory text generation intelligent agent extracts the keywords of each directory in the directory structure and compares them with the key elements of the knowledge base to retrieve relevant texts from the railway engineering knowledge base to fill the text of each directory in the directory structure.

6. The method for generating a feasible construction organization for a railway project according to claim 5, characterized in that: The large language model is used to extract key elements of the knowledge base from the classified text contents of the railway engineering implementation construction specification documents and the classified historical railway engineering implementation construction organization, including: The large language model is used to extract key elements of the knowledge base from the classified text contents of the railway engineering implementation construction specification documents and the classified historical railway engineering implementation construction organizations, and generate knowledge base key element word vectors, so that the directory text generation agent can extract the keywords of each directory in the directory structure and compare them with the knowledge base key element word vectors to retrieve relevant texts from the railway engineering knowledge base to fill the text of each directory in the directory structure.

7. The method for generating a feasible construction organization for a railway project according to any one of claims 1 to 6, characterized in that: After calling the directory text generation agent, extracting keywords of each directory in the directory structure, and retrieving relevant texts from the railway engineering knowledge base based on the keywords to fill the text of each directory in the directory structure to obtain the construction organization, the method further includes: Traversing the text corresponding to each directory in the directory structure to determine whether each directory includes the corresponding text content; If yes, then generate the implementation construction organization of the railway project; If not, the intelligent agent for generating the directory text is called to extract keywords from the directory that does not include the corresponding text content, and based on the keywords, relevant texts are retrieved from the railway engineering knowledge base to fill the directory that does not include the corresponding text content with text.

8. A railway engineering implementation construction organization generation system, characterized in that: include: A knowledge base establishment unit, used to establish a railway engineering knowledge base, wherein the railway engineering knowledge base includes railway engineering implementation construction specification documents and historical railway engineering implementation construction organizations; The first calling unit is used to call the key element extraction intelligent agent to extract the key elements from the pre-placed data of the railway project implementation by using the large language model; The second calling unit is used to call the directory generation intelligent body to generate the directory structure of the railway engineering implementation construction organization based on the key elements; The third calling unit is used to call the directory text generation intelligent body, extract the keywords of each directory in the directory structure, and retrieve relevant texts from the railway engineering knowledge base based on the keywords to fill the text of each directory in the directory structure to obtain the construction organization.

9. A device for generating an implementable construction organization for a railway project, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the method for generating a feasible construction organization for a railway project as described in any one of claims 1 to 7 when executing a computer program.

10. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for generating an implementable construction organization for a railway project as described in any one of claims 1 to 7.

Images