Business system simulation method and device based on large model

Abstract

The invention provides a business system simulation method and device based on a large model. The device comprises a knowledge engineering module, an interpretation engineering module and an interaction engineering module. The knowledge engineering module is used for providing content generation elements of the large model and integrating professional knowledge and technical methods; the interpretation engineering module is used for constructing knowledge reasoning of the professional field through a large model to obtain interpretation data and a simulation model which conform to knowledge rules of the professional field; the interaction engineering module is used for performing optimization through an abstract prompt model of dialogues, interaction and interaction contents in professional fields, outputting an abnormal value and a predicted value to the interpretation engineering module, and performing interaction verification and quality verification of a large model; the knowledge engineering module comprises a data information module, a knowledge graph module and a knowledge engineering library; the interpretation engineering module comprises an interpretation engineer and a large model. According to the invention, the automatic intelligentization of the simulation system is realized, a mathematical analysis basis is provided for the measurement of a simulation scene, the implementation efficiency is greatly improved, and the implementation cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of computer modeling and simulation, and in particular to a business system simulation method and device based on a large model. Background Art

[0002] In recent years, deception defense (commonly known as honeypot) technology has made great progress. Deception defense (commonly known as honeypot) technology is based on the idea of ​​attack and defense confrontation, and the defender can observe the behavior of the attacker. Attack and defense drills are becoming more and more normalized in various industries. Honeypot products are widely used in attack and defense confrontation to build a deception defense system and realize active defense against attacks, which enables the defender to achieve good results in attack and defense drills.

[0003] The simulation capability of the honeypot is the core capability of honeypot trapping. However, at present, due to the high complexity of the business system and the heavy workload, it is difficult to fully obtain the business network information, and thus it is difficult to identify the characteristics of the business network and to perform simulation protection on the business system. The degree of simulation of the business network relies on manual experience, and the efficiency of analyzing the business network is low. The simulation effect lacks measurement, making the simulated network easy to be identified by attackers. The construction of the simulation environment still restricts the further application of honeypot products.

[0004] At present, with the release and application of artificial intelligence big models represented by ChatGPT, all walks of life have begun to consider how to use big models to assist or replace human work. A variety of commercial and open source big models provide plug-in functions and interfaces for the application of big models. Using the logical reasoning ability of big models to simulate human understanding ability, perform knowledge processing and output to third-party programs has gradually become a hot technology. Summary of the invention

[0005] The present invention proposes the following technical solutions to address one or more technical deficiencies in the above-mentioned prior art.

[0006] Based on the first aspect of the present invention, a business system simulation device based on a large model is proposed, including a knowledge engineering module, an explanation engineering module and an interaction engineering module; wherein,

[0007] The knowledge engineering module is used to provide content generation elements for large models and integrate professional knowledge and technical methods;

[0008] The explanation engineering module is used to construct knowledge reasoning in a professional field through a large model to obtain explanation data and simulation models that conform to the knowledge rules of the professional field;

[0009] The interactive engineering module is used to optimize the summary prompt model of professional field dialogue, interaction and interactive content, output abnormal values ​​and predicted values ​​to the interpretation engineering module, and perform interactive verification and quality verification of the large model;

[0010] The knowledge engineering module includes a data information module, a knowledge graph module and a knowledge engineering library;

[0011] The interpretation engineering module includes an interpretation engineer and a large model.

[0012] Furthermore, the data information module is used to define knowledge information of the business system simulation, and the knowledge information includes topology data, traffic data and data content of the business network.

[0013] Furthermore, the knowledge graph module is used to store the knowledge characteristics of the knowledge information and convert the knowledge information into characteristic parameters of the business system;

[0014] The characteristic parameters of the business system include detailed information of network topology and assets in the network;

[0015] The detailed information in the network assets includes knowledge data vectors of operating system, software information, hardware information, database, protocol information of business system and port information and their associations.

[0016] Furthermore, the knowledge engineering library includes a knowledge base and a historical interpretation library generated using a business system simulation template;

[0017] The business system simulation template is used as the input of the interpretation engineering module, and the historical interpretation library stores the interpretation information in the large model for modifying the knowledge base.

[0018] Furthermore, the interpretation engineer is used to convert the knowledge information in the knowledge engineering module into prompt information and send it to the large model, as well as interact with the interactive engineering module, receive abnormal values ​​and predicted values, and send the predicted value results to the knowledge engineering library for correction.

[0019] Furthermore, the large model receives the prompt information in the interpretation engineer, performs model calculations and obtains interpretation information, and sends it to the historical interpretation library in the knowledge engineering library to output a simulation model.

[0020] Large models can learn more subtle patterns and regularities, have stronger generalization and expression capabilities, and can handle more complex and larger data sets or tasks.

[0021] Based on the second aspect of the present invention, a method for simulating a business system using the above-mentioned device is also proposed, comprising:

[0022] The knowledge engineering module provides content generation elements for large models and integrates professional knowledge and technical methods, converts knowledge information into business system characteristic parameters, forms a knowledge graph representing data knowledge characteristics, and uses business system simulation templates to form a knowledge base;

[0023] The explanation engineer converts the knowledge information in the knowledge engineering module into prompt information, the large model receives the prompt information in the explanation engineer to perform model calculation, and constructs knowledge reasoning in the professional field to obtain explanation information that conforms to the knowledge rules of the professional field, sends the explanation information to the historical explanation library in the knowledge engineering library, and outputs the simulation model;

[0024] The interactive engineering module interacts with the explanation engineer, optimizes the large model through dialogues, interactions and summary of interaction content in professional fields, and outputs abnormal values ​​and predicted values ​​to the explanation engineer;

[0025] The explanation engineer receives abnormal values ​​and predicted values, sends the predicted value results to the historical explanation library of the knowledge engineering library, and makes corrections based on the explanation information of the large model in the historical explanation library;

[0026] The interactive engineering module performs interactive verification and quality verification on the large model.

[0027] Furthermore, the large model includes an input-output layer, a fine-tuning layer and a large model layer, wherein:

[0028] The input and output layer uses the LangChain framework to design related parameters, including Model I / O, Prompts, Output parsers, Data Connection, Chain, Agent, and Callcacks;

[0029] The fine-tuning layer fine-tunes different large models and different scenarios, and the fine-tuning methods include LoRA method, Adapter method and P-tuning method;

[0030] The large model used in the large model layer includes ChatGPT4 or ChatGLM2-6B.

[0031] Based on the third aspect of the present invention, a computer storable medium is further proposed, on which one or more computer programs are stored, and when the computer programs are executed by a computer processor, the above-mentioned apparatus is implemented.

[0032] The technical effect of the present invention is that: the present invention performs business system simulation based on a large model, realizes the intelligence and automation of simulation, greatly improves the implementation efficiency, reduces the implementation cost, and draws a knowledge graph for simulating the business system, generates a simulation template, and realizes the normalization of simulation scenarios, which provides a mathematical analysis basis for simulation measurement and provides an important foundation for the construction of a deception defense system. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Other features, objects and advantages of the present application will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings.

[0034] Figure 1 It is an architecture diagram of a business system simulation device based on a large model provided according to an embodiment of the present invention.

[0035] Figure 2 It is a construction flow chart of a business system simulation method based on a large model provided according to an embodiment of the present invention.

[0036] Figure 3 It is a big model construction architecture diagram of a big model-based business system simulation method provided according to an embodiment of the present invention.

[0037] Figure 4 It is a LoRA fine-tuning flow chart of a business system simulation method based on a large model provided according to an embodiment of the present invention.

[0038] Figure 5 It is a structural diagram of a computer system suitable for implementing an electronic device of an embodiment of the present application. DETAILED DESCRIPTION

[0039] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It should also be noted that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.

[0040] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0041] Figure 1 A large-model-based business system simulation device of the present invention is shown, which includes a knowledge engineering module a, an explanation engineering module b and an interaction engineering module c.

[0042] In a specific embodiment, the knowledge engineering module a is used to provide content generation elements for a large model and integrate professional knowledge and technical methods.

[0043] In a specific embodiment, the explanation engineering module b is used to construct knowledge reasoning in a professional field through a large model to obtain explanation data and simulation models that conform to the knowledge rules of the professional field.

[0044] In a specific embodiment, the interactive engineering module c is used to optimize through the summary prompt model of professional field dialogue, interaction and interactive content, output abnormal values ​​and predicted values ​​to the interpretation engineering module, and perform interactive verification and quality verification of the large model;

[0045] The knowledge engineering module includes a data information module, a knowledge graph module and a knowledge engineering library;

[0046] The interpretation engineering module includes an interpretation engineer and a large model.

[0047] It should be noted that the big model, namely the large-scale language model (LLM), is an artificial intelligence model that is trained to understand and generate human language. It has a huge parameter scale and complexity. The design and training of the big model aims to provide more powerful and accurate model performance to cope with more complex and larger data sets or tasks. The big model can learn more subtle patterns and regularities and has stronger generalization and expression capabilities.

[0048] It should be noted that the data information module is used to define knowledge information of the business system simulation, and the knowledge information includes topology data, flow data and data content of the business network.

[0049] It should be noted that the knowledge graph module is used to store the knowledge characteristics of the knowledge information and convert the knowledge information into characteristic parameters of the business system;

[0050] The characteristic parameters of the business system include detailed information of network topology and assets in the network;

[0051] The detailed information in the network assets includes knowledge data vectors of operating system, software information, hardware information, database, protocol information of business system and port information and their associations.

[0052] It should be noted that the knowledge engineering database includes a knowledge database and a historical interpretation database generated using a business system simulation template;

[0053] The business system simulation template is used as the input of the interpretation engineering module, and the historical interpretation library stores the interpretation information in the large model for modifying the knowledge base.

[0054] It should be noted that the interpretation engineer is used to convert the knowledge information in the knowledge engineering module into prompt information and send it to the large model, as well as interact with the interactive engineering module, receive abnormal values ​​and predicted values, and send the predicted value results to the knowledge engineering library for correction.

[0055] It should be noted that the large model receives the prompt information in the interpretation engineering device, performs model calculations and obtains interpretation information, and sends it to the historical interpretation library in the knowledge engineering library to output a simulation model.

[0056] It should be noted that the method for performing business system simulation using the business system simulation device based on the large model is as follows: Figure 2 As shown, including:

[0057] The knowledge engineering module provides content generation elements for large models and integrates professional knowledge and technical methods, converts knowledge information into business system characteristic parameters, forms a knowledge graph representing data knowledge characteristics, and uses business system simulation templates to form a knowledge base;

[0058] The explanation engineer converts the knowledge information in the knowledge engineering module into prompt information, the large model receives the prompt information in the explanation engineer to perform model calculation, and constructs knowledge reasoning in the professional field to obtain explanation information that conforms to the knowledge rules of the professional field, sends the explanation information to the historical explanation library in the knowledge engineering library, and outputs the simulation model;

[0059] The interactive engineering module interacts with the explanation engineer, optimizes the large model through dialogues, interactions and summary of interaction content in professional fields, and outputs abnormal values ​​and predicted values ​​to the explanation engineer;

[0060] The explanation engineer receives abnormal values ​​and predicted values, sends the predicted value results to the historical explanation library of the knowledge engineering library, and makes corrections based on the explanation information of the large model in the historical explanation library;

[0061] The interactive engineering module performs interactive verification and quality verification on the large model.

[0062] It should be noted that the large model includes an input and output layer, a fine-tuning layer and a large model layer, wherein:

[0063] The input and output layer uses the LangChain framework to design related parameters, including Model I / O, Prompts, Output parsers, Data Connection, Chain, Agent, and Callcacks;

[0064] The fine-tuning layer fine-tunes different large models and different scenarios, and the fine-tuning methods include LoRA method, Adapter method and P-tuning method;

[0065] The large model used in the large model layer includes ChatGPT4 or ChatGLM2-6B.

[0066] It should be noted that the construction of the large model includes input and output layers, fine-tuning layers and large model layers, such as Figure 3 As shown, the input and output layers use the LangChain framework to set the main parameters, the fine-tuning layer uses the LoRA method, Adapter method and P-tuning method for fine-tuning, and the large model layer uses ChatGPT and other open source large models.

[0067] In a specific embodiment, the input and output layer uses the LangChain framework to set the main parameters, including: setting the Model I / O parameters by selecting the language model (LLMs) and giving the input and output as text format; setting the Prompts parameters by selecting the PromptTemplate mode of the prompter; setting the Output parsers parameters by using a custom string with a fixed structure for the simulation model; and,

[0068] Use the Data Connection parameter to implement data connection retrieval, use the Chains parameter to implement complex task models, multiple fine-tuning components and multi-task applications of large models, use the Agent parameter to implement role definition and enhance the reasoning ability of large models, and use the Callbacks parameter to implement log printing, monitoring and streaming.

[0069] It should be noted that in order to adapt to the fine-tuning effects of different large models and different scenarios, the fine-tuning methods used in the design of the fine-tuning layer include the LoRA method, the Adapter method and the P-tuning method.

[0070] In a specific embodiment, during the training of the large model, a plurality of typical business network scenarios are selected for knowledge engineering construction, and corresponding simulation templates are generated for each typical business network to form a knowledge engineering library;

[0071] The input and output interfaces of the ChatGLM2-6B large model are established through the LangChain framework, and the data interface format is defined. The large model is trained based on the information of the simulation template in the knowledge engineering library. The interactive engineering module trains and corrects the large model. The large model outputs interpretation data to correct the knowledge base algorithm and trains to obtain the initialization model W0.

[0072] When applied on-site, the on-site business network topology data, mirrored traffic data, and classified and graded data content are input into the knowledge graph data access device, and the large model is fine-tuned according to the simulation network parameters output by the large model and the LoRA method. After fine-tuning, the simulation model data output by the large model is input into the deception defense device.

[0073] In a specific embodiment, the fine-tuning process of the LoRA method is as follows Figure 4 As shown, specifically, the parameters W0 and variable parameters ΔW of the pre-trained model are set, and a small parameter module is added to learn the variable parameter matrix ΔW. The specific formula is:

[0074] h=W0x+ΔWx=W0x+BAx,

[0075] Where x represents the input of the model, A represents the dimension reduction matrix, B represents the dimension increase matrix, W0 represents the parameters initialized by the pre-trained model, and ΔW represents the variable parameter matrix of the training model;

[0076] A bypass is added to the original pre-trained language model PLM, and the bypass is first reduced in dimension and then increased in dimension to simulate the intrinsic rank. During training, the parameters of the language model PLM are fixed, and the reduced-dimensional matrix A and the increased-dimensional matrix B are trained separately. The calculation matrix BA and the parameters of the language model PLM are superimposed without changing the input and output dimensions of the model. The reduced-dimensional matrix A is initialized by random Gaussian distribution, and the increased-dimensional matrix B is initialized by a 0 matrix, and the model is fine-tuned.

[0077] It should be noted that the large model layer is mainly based on natural language human-computer interaction and has weak adaptability to transaction scenarios. The ChatGPT4 large model and the ChatGLM2-6B large model have better effects on business system simulation.

[0078] In a specific embodiment, the application method of ChatGLM2-6B includes: using the LangChain framework to establish a simulation input parameter text and set the Prompts parameter and the number of Embeddig embedding layers, setting GLMBLock to 28 cycles, and using the LangChain framework to establish a simulation output parameter text and set the next token parameter.

[0079] It should be noted that when the customer network node scale is 10,000 points, the mirror traffic of the network core interaction is generally 5Gbps. In the manual simulation method of the traditional honeypot, the business network simulation workload is no less than 3 man-days, and the simulation degree is <60%; in contrast, the business network simulation of the present invention is <60s, and the simulation degree is >95%, which greatly improves the simulation efficiency and simulation accuracy.

[0080] Reference below Figure 5, which shows a schematic diagram of the structure of a computer system suitable for implementing an electronic device of an embodiment of the present application. Figure 5 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0081] like Figure 5 As shown, the computer system includes a central processing unit (CPU) 501, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 502 or the program loaded from the storage part 508 to the random access memory (RAM) 503. In the RAM 503, various programs and data required for system operation are also stored. The CPU 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0082] The following components are connected to the I / O interface 505: an input section 506 including a keyboard, a mouse, etc.; an output section 507 including a liquid crystal display (LCD), etc. and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, a modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the I / O interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed, so that a computer program read therefrom is installed into the storage section 508 as needed.

[0083] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 509, and / or installed from the removable medium 511. When the computer program is executed by the central processing unit (CPU) 501, the above functions defined in the method of the present application are executed. It should be noted that the computer-readable storage medium of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, - but not limited to - an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more conductors, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable storage medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, device, or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wireless, wireline, optical cable, RF, etc., or any suitable combination of the foregoing.

[0084] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0085] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0086] The modules involved in the embodiments of the present application may be implemented by software or by hardware.

[0087] As another aspect, the present application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment; or it may exist independently without being assembled into the electronic device. The above computer-readable storage medium carries one or more programs. When the above one or more programs are executed by the electronic device, the electronic device: the knowledge engineering module is used to provide content generation elements of the large model and integrate professional knowledge and technical methods; the knowledge engineering module provides content generation elements of the large model and integrates professional knowledge and technical methods, converts knowledge information into business system characteristic parameters, forms a knowledge graph representing the knowledge characteristics of the data, and uses a business system simulation template to form a knowledge base; the interpretation engineer converts the knowledge information in the knowledge engineering module into prompt information, and the large model receives the prompt information in the interpretation engineer for The model is operated and professional knowledge reasoning is constructed to obtain explanation information that conforms to the professional knowledge rules, and the explanation information is sent to the historical explanation library in the knowledge engineering library, and the simulation model is output; the interactive engineering module interacts with the explanation engineer, and optimizes the large model through professional dialogue, interaction and summary of interaction content, and outputs abnormal values ​​and predicted values ​​to the explanation engineer; the explanation engineer receives abnormal values ​​and predicted values, and sends the predicted value results to the historical explanation library of the knowledge engineering library, and corrects according to the explanation information of the large model in the historical explanation library; the interactive engineering module performs interactive verification and quality verification on the large model.

[0088] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above invention concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.

[0089] Finally, it should be noted that the above embodiments are only intended to illustrate rather than limit the technical solutions of the present invention. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the present invention can still be modified or replaced by equivalents. Any modification or partial replacement that does not depart from the spirit and scope of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A business system simulation device based on a large model, characterized in that: It includes knowledge engineering module, interpretation engineering module and interactive engineering module; among them, The knowledge engineering module is used to provide content generation elements for large models and integrate professional knowledge and technical methods; The explanation engineering module is used to construct knowledge reasoning in a professional field through a large model to obtain explanation data and simulation models that conform to the knowledge rules of the professional field; The interactive engineering module is used to optimize the summary prompt model of professional field dialogue, interaction and interactive content, output abnormal values ​​and predicted values ​​to the interpretation engineering module, and perform interactive verification and quality verification of the large model; The knowledge engineering module includes a data information module, a knowledge graph module and a knowledge engineering library; The interpretation engineering module includes an interpretation engineer and a large model.

2. The device according to claim 1, characterized in that The data information module is used to define knowledge information of the business system simulation, and the knowledge information includes topology data, flow data and data content of the business network.

3. The device according to claim 1, characterized in that The knowledge graph module is used to store the knowledge characteristics of the knowledge information and convert the knowledge information into characteristic parameters of the business system; The characteristic parameters of the business system include detailed information of network topology and assets in the network; The detailed information in the network assets includes knowledge data vectors of operating system, software information, hardware information, database, protocol information of business system and port information and their associations.

4. The device according to claim 1, characterized in that The knowledge engineering database includes a knowledge database and a historical interpretation database generated using a business system simulation template; The business system simulation template is used as the input of the interpretation engineering module, and the historical interpretation library stores the interpretation information in the large model for modifying the knowledge base.

5. The device according to claim 1, characterized in that The explanation engineer is used to convert the knowledge information in the knowledge engineering module into prompt information and send it to the large model, and interact with the interactive engineering module, receive abnormal values ​​and predicted values, and send the predicted value results to the knowledge engineering library for correction.

6. The device according to claim 1, characterized in that The large model receives the prompt information in the interpretation engineering device, performs model calculations and obtains interpretation information, and sends it to the historical interpretation library in the knowledge engineering library to output a simulation model.

7. A method for simulating a business system using the device according to any one of claims 1 to 6, characterized in that: include: The knowledge engineering module provides content generation elements for large models and integrates professional knowledge and technical methods, converts knowledge information into business system characteristic parameters, forms a knowledge graph representing data knowledge characteristics, and uses business system simulation templates to form a knowledge base; The explanation engineer converts the knowledge information in the knowledge engineering module into prompt information, the large model receives the prompt information in the explanation engineer to perform model calculation, and constructs knowledge reasoning in the professional field to obtain explanation information that conforms to the knowledge rules of the professional field, sends the explanation information to the historical explanation library in the knowledge engineering library, and outputs the simulation model; The interactive engineering module interacts with the explanation engineer, optimizes the large model through dialogues, interactions and summary of interaction content in professional fields, and outputs abnormal values ​​and predicted values ​​to the explanation engineer; The explanation engineer receives abnormal values ​​and predicted values, sends the predicted value results to the historical explanation library of the knowledge engineering library, and makes corrections based on the explanation information of the large model in the historical explanation library; The interactive engineering module performs interactive verification and quality verification on the large model.

8. The method according to claim 7, characterized in that The large model includes an input and output layer, a fine-tuning layer and a large model layer, wherein: The input and output layer uses the LangChain framework to design related parameters, including Model I / O, Prompts, Output parsers, Data Connection, Chain, Agent and Call lcacks; The fine-tuning layer fine-tunes different large models and different scenarios, and the fine-tuning methods include LoRA method, Adapter method and P-tuning method; The large model used in the large model layer includes ChatGPT4 or ChatGLM2-6B.

9. A computer storable medium, characterized in that: One or more computer programs are stored thereon, and when the computer programs are executed by a computer processor, the apparatus according to any one of claims 1 to 6 is implemented.

Images