A simulation processing method and device based on a large language model agent

Abstract

The embodiment of the application provides a simulation processing method and device based on a large language model agent, relates to the technical field of industrial simulation, and the method comprises the following steps: obtaining demand information input by a user, calling a large language model of a simulation process arrangement system through a model interface to perform intent recognition on the demand information, and obtaining an intent type corresponding to the demand information; if the intent type belongs to a simulation operation type, calling a simulation process management tool to perform simulation parameter setting according to demand information stored in a long short-term memory module, obtaining a corresponding simulation task, executing the simulation task, and generating a corresponding simulation result; obtaining a file type input by the user for the simulation result, determining a display template corresponding to the file type, inputting the simulation result into the display template, and obtaining a data file corresponding to the simulation result, so that the efficiency and accuracy of simulation process arrangement are improved, and the error rate is reduced.

Description

Technical Field

[0001] This invention relates to the field of industrial simulation technology, and in particular to a simulation processing method based on a large language model intelligent agent, a simulation processing device based on a large language model intelligent agent, an electronic device, and a computer-readable storage medium. Background Technology

[0002] In the field of industrial simulation, workflow orchestration is a core element in realizing complex simulation tasks. However, orchestrating corresponding simulation workflows often requires users to manually write scripts or configure settings through complex interfaces. This not only consumes a significant amount of user time and effort but also makes it difficult to cope with dynamically changing requirements. For example, when users need to frequently modify the relevant content of a simulation task, they often have to do so manually, increasing operational complexity and leading to low execution efficiency and a high risk of errors. Furthermore, the result display methods in simulation workflows are too simplistic and inflexible, unable to dynamically adjust the display method according to user needs, thus limiting the application value of simulation results. Summary of the Invention

[0003] This invention provides a simulation processing method, apparatus, electronic device, and computer-readable storage medium based on a large language model intelligent agent, to solve or partially solve the problems of low efficiency and error-proneness, and overly simplistic and inflexible result display in the simulation process.

[0004] This invention discloses a simulation processing method based on a large language model intelligent agent, applied to a simulation process orchestration system. The simulation process orchestration system includes at least a model interface, a long short-term memory module, and a simulation process management tool. The method includes:

[0005] Obtain the user's input requirement information, and call the large language model connected to the simulation process orchestration system through the model interface to perform intent recognition on the requirement information to obtain the intent type corresponding to the requirement information;

[0006] If the intent type belongs to the simulation operation type, the simulation process management tool is invoked to set simulation parameters according to the requirement information stored in the long and short memory module, obtain the corresponding simulation task, execute the simulation task, and generate the corresponding simulation result.

[0007] Obtain the file type of the simulation result input by the user, determine the display template corresponding to the file type, input the simulation result into the display template, and obtain the data file corresponding to the simulation result.

[0008] In some feasible implementations, the simulation process management tool includes at least tool nodes and flowcharts; wherein, the tool nodes encapsulate specific operations, and each tool node is used to perform one of the operations of setting parameters, running simulation, and generating a report; the flowchart is a tool for defining and managing state machines, and the flowchart consists of multiple working nodes and the transition logic between the working nodes.

[0009] In some feasible implementations, the step of calling the simulation process management tool to set simulation parameters and obtain the corresponding simulation task based on the requirement information stored in the long and short memory module includes:

[0010] Extract the demand information from the long and short memory module;

[0011] Extract target key information from the requirement information, where the target key information is information related to simulation testing;

[0012] The tool node is invoked to set simulation parameters that match the target key information;

[0013] The flowchart is invoked to set the simulation process corresponding to the simulation parameters, and a test task corresponding to the simulation process is generated.

[0014] In some feasible implementations, the simulation process orchestration system further includes simulation tool functions, which at least include parameter setting functions, simulation execution functions, and result generation functions. Executing the simulation task and generating the corresponding simulation results includes:

[0015] The simulation parameters are passed to the simulation software by calling the parameter setting function.

[0016] The simulation execution function is invoked to control the simulation software to execute the simulation process according to the simulation parameters;

[0017] In response to the completion of the simulation process, the result generation function is called to generate the simulation results corresponding to the simulation process.

[0018] In some feasible implementations, the simulation workflow orchestration system further includes a simulation manager; wherein the simulation manager is used for setting and managing simulation type, performance, architecture, and model, and supports setting simulation parameters in multi-turn dialogues; the method further includes:

[0019] Obtain the user's input adjustment instructions for the test parameters;

[0020] The large language model is called through the model interface to identify the adjustment command and obtain the simulation adjustment parameters corresponding to the adjustment command.

[0021] The simulation parameters are adjusted to match the target simulation parameters.

[0022] In some feasible implementations, the simulation parameters include at least a first simulation type, a first performance mode, and first model parameters; the simulation adjustment parameters include at least type adjustment parameters, performance adjustment parameters, and model adjustment parameters; and adjusting the simulation parameters to target simulation parameters that match the simulation adjustment parameters includes:

[0023] Adjust the first simulation type to a second simulation type that matches the type adjustment parameters;

[0024] Adjust the first performance mode to a second performance mode that matches the performance adjustment parameters;

[0025] The first model parameter is adjusted to a second model parameter that matches the model adjustment parameter.

[0026] Among some feasible implementation methods are:

[0027] If the intent type belongs to the knowledge acquisition category, then obtain the knowledge question and answer results corresponding to the requirement information.

[0028] In some feasible implementations, obtaining the knowledge-based question-and-answer results corresponding to the required information includes:

[0029] The model interface is used to call the large language model to generate knowledge-based question-and-answer results corresponding to the required information; and / or,

[0030] Call a preset knowledge extraction function to extract knowledge question-and-answer results corresponding to the required information from a preset model construction document; and / or,

[0031] The preset experience extraction function is invoked to extract the knowledge Q&A results corresponding to the required information from the preset experience document.

[0032] This invention also discloses a simulation processing device based on a large language model intelligent agent, applied to a simulation process orchestration system. The simulation process orchestration system includes at least a model interface, a long and short memory module, and a simulation process management tool. The device includes:

[0033] The intent recognition module is used to obtain the user's input requirement information, and to call the large language model connected to the simulation process orchestration system through the model interface to perform intent recognition on the requirement information and obtain the intent type corresponding to the requirement information.

[0034] The simulation processing module is used to call the simulation process management tool to set simulation parameters according to the requirement information stored in the long and short memory module if the intent type belongs to the simulation operation class, obtain the corresponding simulation task, execute the simulation task, and generate the corresponding simulation result.

[0035] The results presentation module is used to obtain the file type of the simulation results input by the user, determine the display template corresponding to the file type, input the simulation results into the display template, and obtain the data file corresponding to the simulation results.

[0036] In some feasible implementations, the simulation process management tool includes at least tool nodes and flowcharts; wherein, the tool nodes encapsulate specific operations, and each tool node is used to perform one of the operations of setting parameters, running simulation, and generating a report; the flowchart is a tool for defining and managing state machines, and the flowchart consists of multiple working nodes and the transition logic between the working nodes.

[0037] In some feasible implementations, the simulation processing module is specifically used for:

[0038] Extract the demand information from the long and short memory module;

[0039] Extract target key information from the requirement information, where the target key information is information related to simulation testing;

[0040] The tool node is invoked to set simulation parameters that match the target key information;

[0041] The flowchart is invoked to set the simulation process corresponding to the simulation parameters, and a test task corresponding to the simulation process is generated.

[0042] In some feasible implementations, the simulation process orchestration system further includes simulation tool functions, which at least include parameter setting functions, simulation execution functions, and result generation functions. The simulation processing module is specifically used for:

[0043] The simulation parameters are passed to the simulation software by calling the parameter setting function.

[0044] The simulation execution function is invoked to control the simulation software to execute the simulation process according to the simulation parameters;

[0045] In response to the completion of the simulation process, the result generation function is called to generate the simulation results corresponding to the simulation process.

[0046] In some feasible implementations, the simulation workflow orchestration system further includes a simulation manager; wherein the simulation manager is used for setting and managing simulation types, performance, architecture, and models, and supports setting simulation parameters in multi-turn dialogues; the device further includes:

[0047] The instruction acquisition module is used to acquire the adjustment instructions for the test parameters input by the user;

[0048] The recognition module is used to call the large language model through the model interface to recognize the adjustment command and obtain the simulation adjustment parameters corresponding to the adjustment command;

[0049] An adjustment module is used to adjust the simulation parameters to target simulation parameters that match the simulation adjustment parameters.

[0050] In some feasible implementations, the simulation parameters include at least a first simulation type, a first performance mode, and first model parameters; the simulation adjustment parameters include at least type adjustment parameters, performance adjustment parameters, and model adjustment parameters; and the adjustment module is specifically used for:

[0051] Adjust the first simulation type to a second simulation type that matches the type adjustment parameters;

[0052] Adjust the first performance mode to a second performance mode that matches the performance adjustment parameters;

[0053] The first model parameter is adjusted to a second model parameter that matches the model adjustment parameter.

[0054] Among some feasible implementation methods are:

[0055] The knowledge feedback module is used to obtain the knowledge Q&A results corresponding to the requirement information if the intent type belongs to the knowledge acquisition category.

[0056] In some feasible implementations, the knowledge feedback module is specifically used for:

[0057] The model interface is used to call the large language model to generate knowledge-based question-and-answer results corresponding to the required information; and / or,

[0058] Call a preset knowledge extraction function to extract knowledge question-and-answer results corresponding to the required information from a preset model construction document; and / or,

[0059] The preset experience extraction function is invoked to extract the knowledge Q&A results corresponding to the required information from the preset experience document.

[0060] This invention also discloses an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0061] The memory is used to store computer programs;

[0062] When the processor executes a program stored in the memory, it implements the method described in the embodiments of the present invention.

[0063] This invention also discloses a computer-readable storage medium storing instructions that, when executed by one or more processors, cause the processors to perform the methods described in this invention.

[0064] The embodiments of the present invention have the following advantages:

[0065] In this embodiment of the invention, a simulation process orchestration system for simulation processing is constructed. This system includes at least a model interface, a long short-term memory (LSM) module, and a simulation process management tool. By acquiring user-inputted requirements, the system uses the model interface to call a large language model connected to the simulation process orchestration system to perform intent recognition on the requirements, obtaining the intent type corresponding to the requirements. If the intent type belongs to the simulation operation category, the simulation process management tool is called to set simulation parameters based on the requirements stored in the LSM module, obtain the corresponding simulation task, execute the simulation task, and generate the corresponding simulation result. Then, the system acquires the file type of the simulation result input by the user, determines the display template corresponding to the file type, and inputs the simulation result into the display template to obtain the data file corresponding to the simulation result. Thus, by using the large language model as the "foundation" of the system, it deeply understands and automatically recognizes user intent, and completes the setting of simulation parameters, execution of simulation tasks, and output of simulation results according to the user intent. This helps users complete complex simulation tasks more efficiently, improves the efficiency and accuracy of simulation process orchestration, reduces the error rate, and enriches the diversity and flexibility of simulation result display based on the corresponding display template, effectively meeting user needs. Attached Figure Description

[0066] Figure 1 This is a flowchart of the steps of a simulation processing method based on a large language model intelligent agent provided in an embodiment of the present invention;

[0067] Figure 2 This is a structural block diagram of a simulation processing device based on a large language model intelligent agent provided in an embodiment of the present invention. Detailed Implementation

[0068] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0069] As an example, in the field of industrial simulation, process orchestration is a core element in realizing complex simulation tasks. It involves the automated management and optimization of multiple steps, including setting simulation parameters, the execution order of simulation tasks, and the processing and display of simulation results. However, with the increasing complexity and diversity of simulation tasks, related process orchestration methods have gradually revealed problems such as low efficiency, insufficient flexibility, and difficulty in adapting to diverse user needs. For instance, in the process of related process orchestration, relying on users to manually write scripts or configure through complex interfaces is not only time-consuming and labor-intensive but also difficult to cope with dynamically changing simulation requirements. Furthermore, the process orchestration process often requires manual completion by users, leading to low simulation efficiency and a high risk of errors. In addition, the result display methods in the simulation process are too simplistic and inflexible, unable to dynamically adjust the display method according to user needs, thus limiting the application value of the simulation results.

[0070] In this embodiment of the invention, a corresponding simulation process orchestration system is constructed to meet different user needs. This system includes a model interface, a long short-term memory (LSM) module, and simulation process management tools. The model interface is used to access a large language model, which serves as the "foundation" of the system, enhancing the deep understanding and automatic recognition capabilities of user-defined scenarios and improving the intelligence level of the simulation process orchestration. The LSM module further enhances the system's understanding of multi-turn dialogue content, allowing the system to dynamically plan the execution order of the simulation process by combining contextual information and historical data, ensuring efficient execution of simulation tasks. To meet diverse user needs, the system also provides various simulation result display templates, allowing users to choose the appropriate display method to present the simulation results. This effectively enhances the application value of the simulation results while satisfying user requirements.

[0071] Reference Figure 1 This diagram illustrates a flowchart of a simulation processing method based on a large language model intelligent agent provided in an embodiment of the present invention. The method is applied to a simulation process orchestration system, which includes at least a model interface, a long short-term memory module, and a simulation process management tool. Specifically, it may include the following steps:

[0072] Step 101: Obtain the user-inputted requirement information, and use the model interface to call the large language model connected to the simulation process orchestration system to perform intent recognition on the requirement information to obtain the intent type corresponding to the requirement information;

[0073] A simulation workflow orchestration system (hereinafter referred to as the system) may include a model interface, a long and short-term memory module, and simulation workflow management tools. The model interface is used to access a large language model, which serves as the "base" for constructing a corresponding "intelligent agent." This "intelligent agent" performs intent recognition on the user's natural language input, classifies the intent based on the recognition results, and then executes the corresponding workflow orchestration based on the classification results. The long and short-term memory module can manage the context information of the dialogue, and may include short-term memory and long-term memory, respectively storing temporary information of the current dialogue and persistent information across dialogues. For example... Short-term memory can be used to store contextual information of the current dialogue, including user input, system responses, and tool call results, to maintain dialogue coherence, support complex tasks, and reduce repetitive user input. Long-term memory is used to store persistent information across dialogues, including user preferences, historical simulation results, and commonly used configurations, to provide personalized services, historical records and backtracking, and cross-dialogue task support. Simulation process management tools are the core components of the simulation process orchestration system, responsible for coordinating and controlling the execution of simulation tasks. By integrating functions such as parameter setting, simulation execution, and result generation, they help users complete simulation tasks efficiently.

[0074] It should be noted that the system can be built around a one-dimensional simulation process from industrial simulation software, used for intelligent orchestration of simulation tasks within that scenario. Specifically, during the orchestration of the simulation process, the system can use the user's natural language input, through a model interface, to call the corresponding large language model to perform intent recognition on the user's input. Based on the intent recognition, it determines whether to execute the corresponding simulation process, and then performs the appropriate operation through the simulation process management tool. Simultaneously, during the above process, a long short-term memory module can manage the relevant contextual information to enhance the system's ability to understand multi-turn dialogue content.

[0075] In this embodiment of the invention, for the user-input requirement information, the system can call the large language model of the simulation process orchestration system through the model interface to perform intent recognition on the requirement information, so as to obtain the intent type corresponding to the requirement information, so as to further execute the corresponding process orchestration according to different intent types.

[0076] The requirement information can be natural language input by the user to express their needs in the system, such as "Set the temperature parameter to 300K and run the simulation", "Adjust the pressure parameter to 100Pa and rerun the simulation", "Generate a simulation report containing temperature and pressure distributions in Word format", "Set the performance mode to high precision mode", "I want to know what 100 km / h acceleration means? What factors is it related to in model A?" etc. Based on the user's input requirement information, the system can call the large language model to perform intent recognition and determine the intent type corresponding to the requirement information based on the intent recognition results.

[0077] Optionally, intent types can be used to characterize the purpose or intention expressed behind the user's input information. By identifying the user's intent type, the system can understand the user's needs and take corresponding actions. Intent types can include simulation operation types and knowledge acquisition types. Simulation operation types represent the user's desire to perform specific simulation tasks, such as setting parameters, running a simulation, and generating a report. Knowledge acquisition types represent the user's desire to obtain information or knowledge, such as querying simulation parameters, understanding simulation types, and obtaining help documentation. This invention does not impose limitations on these categories.

[0078] Furthermore, during the "dialogue" between the user and the system, the Long Short-Term Memory (LSTM) module can record contextual information from multiple rounds of dialogue. For example, short-term memory within the LTM module can be implemented using the `InputStoreMemory` class of `LangChain`, enabling real-time recording of the current dialogue's contextual information. STM saves user input and model output using the `save_context` method and loads contextual information using the `load_memory_variables` method. Long-term memory can be implemented using external vector storage (such as a vector database). By utilizing the embedding similarity calculation in the knowledge base, the agent can retrieve historical information relevant to the current task from external storage. The introduction of long-term memory allows the agent to access historical data or knowledge in complex tasks, enhancing its deep understanding and processing capabilities.

[0079] Optionally, the system can also combine prompt word engineering with the long and short-term memory module to achieve a deep understanding of the user's natural language input and dynamic scheduling of the simulation process. In some feasible implementations, the system can generate prompt information related to the user's query, combine the context information of short-term memory and the historical data of long-term memory, and dynamically plan the execution order of the simulation process. At the same time, combined with the simulation process management tool, the system can realize the dynamic scheduling of the simulation process. Thus, the system can call the corresponding tool functions according to the user's intention to complete the simulation parameter setting, simulation execution and result generation.

[0080] Step 102: If the intent type belongs to the simulation operation type, the simulation process management tool is called to set the simulation parameters according to the requirement information stored in the long and short memory module, obtain the corresponding simulation task, execute the simulation task, and generate the corresponding simulation result.

[0081] When the system identifies the intent type as a simulation operation, it can call the simulation process management tool to set simulation parameters based on the requirement information stored in the long and short memory module, obtain the corresponding simulation task, and then call the corresponding simulation software to execute the simulation task and generate the corresponding simulation results. This achieves full automation from intent recognition to simulation task execution, effectively helping users complete complex simulation tasks more efficiently, improving the efficiency and accuracy of simulation process orchestration, and reducing the error rate.

[0082] Optionally, the simulation process management tool includes at least tool nodes and flowcharts. Each tool node encapsulates specific operations, performing one of the following: setting parameters, running the simulation, or generating a report. The flowchart is a tool for defining and managing the state machine, consisting of multiple worker nodes and the transition logic between them. For example, a tool node can be a ToolNode, and the flowchart can be a StateGraph. Both enable tool invocation and process control, and support dynamic simulation process scheduling.

[0083] When the user intent is identified as belonging to the simulation operation category, the system can extract the requirement information from the long short-term memory module, and extract the target key information from the requirement information. The target key information is information related to the simulation test. Then, the tool node is called to set the simulation parameters that match the target key information. Then, the flowchart is called to set the simulation process corresponding to the simulation parameters, and the test task corresponding to the simulation process is generated.

[0084] The simulation process orchestration system also includes simulation tool functions, which include at least parameter setting functions, simulation execution functions, and result generation functions. Different simulation tool functions can be used to perform different operations. Optionally, after generating the corresponding simulation task, the system can first call the parameter setting function to pass the simulation parameters into the simulation software, then call the simulation execution function to control the simulation software to execute the simulation process according to the simulation parameters, and then, in response to the completion of the simulation process, call the result generation function to generate the simulation results corresponding to the simulation process.

[0085] In some examples, the system can extract relevant requirement information and key target information through the Long Short-Term Memory (LSTM) module. For instance, when the user intent is identified as a simulation operation, the system can extract requirement information related to the current task from the LSTM module. If the user previously inputs, "Set the temperature parameter to 300K," the system will extract the requirement information from the LSTM module: {"temperature": 300K}. Similarly, the system can extract key target information related to simulation testing from the requirement information, including simulation parameters, performance settings, and model configuration. For example, if the requirement information is {"temperature": 300K, "pressure": 100Pa}, the system will extract the key target information as {"temperature": 300K}. By using a large language model as the system's "foundation," the system can deeply understand and automatically recognize user intent, and based on the user intent, complete the setting of simulation parameters, the execution of simulation tasks, and the output of simulation results. This helps users complete complex simulation tasks more efficiently, improves the efficiency and accuracy of simulation process orchestration, and reduces the error rate.

[0086] Furthermore, tool nodes can serve as the basic units in simulation management tools, used to encapsulate specific operations. Each working node is responsible for executing a specific task. Flowcharts, on the other hand, are tools used to define and manage state machines. They consist of multiple working nodes and the transition logic between them. Flowcharts manage the execution order and conditional jumps of simulation tasks through state machines. For example, setting parameters, running simulations, and generating reports are all handled by the system based on the extracted key target information. The system can then call the corresponding working nodes to set simulation parameters, execute the simulation process, and generate simulation reports. The execution order of each working node is defined through the flowchart, allowing the system to execute the corresponding working nodes sequentially according to the flowchart.

[0087] In one example, assuming the user inputs the following requirements: set the temperature parameter to 300K and the pressure parameter to 100Pa, the system can extract the corresponding target key information from the requirements: {"temperature": 300K, "pressure": 100Pa}. Next, it calls the tool node to set the simulation parameters: calling the parameter setting tool node to set the temperature parameter to 300K and the pressure parameter to 100Pa; and calling the flowchart to set the simulation process, executing the tool nodes in sequence according to the flowchart: set parameters - run simulation - generate report. Then, the following test task is generated:

[0088] Verify that the temperature parameter is set to 300K;

[0089] Verify that the pressure parameter is set to 100 Pa;

[0090] Verify whether the simulation results contain the expected temperature and pressure distribution data;

[0091] Verify whether the simulation report includes charts of temperature and pressure distribution.

[0092] Finally, the generated test tasks are executed sequentially to ensure that each step of the simulation process is completed correctly. This allows for the setting of simulation parameters, the execution of simulation tasks, and the output of simulation results according to the user's intent. This helps users complete complex simulation tasks more efficiently, improves the efficiency and accuracy of simulation process orchestration, and reduces the error rate.

[0093] Optionally, the system can also generate the parameter settings and process generation during the simulation test through corresponding models. For example, the corresponding simulation parameters can be selected through a parameter selection model, or the corresponding simulation process can be generated through a corresponding simulation model. This invention does not limit this.

[0094] In some feasible implementations, the system may also include a simulation manager. The simulation manager is used for setting and managing simulation types, performance, architecture, and models, and supports simulation parameter settings in multi-turn dialogues. During the process of setting simulation parameters, executing simulation processes, and generating simulation reports through the simulation process management tool, if the user wants to adjust the simulation task, the user can input corresponding adjustment commands. The system can then obtain the user's input adjustment commands for the test parameters, and through the model interface, call a large language model to recognize the adjustment commands, obtain the simulation adjustment parameters corresponding to the adjustment commands, and then adjust the simulation parameters to the target simulation parameters that match the adjustment commands. Thus, the system can flexibly adjust the relevant content involved in the simulation orchestration process based on the user's dynamic adjustment commands, without requiring manual adjustments by the user. This not only improves the flexibility and efficiency of simulation process orchestration but also reduces the user's workload and minimizes errors caused by manual operation.

[0095] The simulation parameters include at least a first simulation type, a first performance mode, and a first model parameter. The simulation adjustment parameters include at least a type adjustment parameter, a performance adjustment parameter, and a model adjustment parameter. The parameter adjustment process may include at least one of the following: adjusting the first simulation type to a second simulation type that matches the type adjustment parameter; adjusting the first performance mode to a second performance mode that matches the performance adjustment parameter; and adjusting the first model parameter to a second model parameter that matches the model adjustment parameter.

[0096] For example, users can input commands to adjust the values ​​of simulation parameters (such as "adjust the temperature parameter to 350K"), commands to adjust the simulation process (such as "optimize the mesh resolution before running the simulation"), commands to adjust the performance mode (such as "set the performance mode to high precision mode"), and commands to adjust the simulation type (such as "set the simulation type to fluid dynamics"), etc. This invention does not limit these.

[0097] Step 103: Obtain the file type of the simulation result input by the user, determine the display template corresponding to the file type, and input the simulation result into the display template to obtain the data file corresponding to the simulation result.

[0098] After the system completes the corresponding simulation task through the simulation software, it can output the corresponding simulation results. During the output of simulation results, when the user wants to present the simulation results in a certain way, the user can input the file type of the simulation results into the system. The system can then determine the display template corresponding to the file type, input the simulation results into the display template, and obtain the data file corresponding to the simulation results. This enriches the diversity and flexibility of the simulation results display and can effectively meet the user's needs.

[0099] Optionally, the file type can include text, chart, image, etc. For example, users can choose to display simulation results in chart, PDF format, Word document, etc., according to their own needs, thereby enriching the diversity and flexibility of simulation result display and effectively meeting user needs.

[0100] In the above process, the simulation operation class as the intent type is used as an example for illustration. If the intent type belongs to the knowledge acquisition class, the system can obtain the knowledge question-and-answer results corresponding to the requirement information. In some feasible implementations, the system can call the large language model through the model interface to generate the knowledge question-and-answer results corresponding to the requirement information; and / or, call the preset knowledge extraction function to extract the knowledge question-and-answer results corresponding to the requirement information from the preset model construction document; and / or, call the preset experience extraction function to extract the knowledge question-and-answer results corresponding to the requirement information from the preset experience document.

[0101] Through the above process, the simulation workflow orchestration system can use a large language model as the foundation of an intelligent agent to achieve a deep understanding and automatic recognition of user intentions. Based on the user intentions, it can call the simulation software's API (Application Programming Interface) to complete the setting of simulation parameters, the execution of simulation tasks, and the display of simulation results in various templates. This significantly improves the intelligence level of the simulation software, helps users complete complex simulation tasks more efficiently, and promotes the further development of industrial simulation technology.

[0102] It should be noted that the embodiments of the present invention include, but are not limited to, the examples described above. It is understood that those skilled in the art can make further settings according to actual needs under the guidance of the ideas in the embodiments of the present invention, and the present invention does not limit such settings.

[0103] In this embodiment of the invention, a simulation process orchestration system for simulation processing is constructed. This system includes at least a model interface, a long short-term memory (LSM) module, and a simulation process management tool. By acquiring user-inputted requirements, the system uses the model interface to call a large language model connected to the simulation process orchestration system to perform intent recognition on the requirements, obtaining the intent type corresponding to the requirements. If the intent type belongs to the simulation operation category, the simulation process management tool is called to set simulation parameters based on the requirements stored in the LSM module, obtain the corresponding simulation task, execute the simulation task, and generate the corresponding simulation result. Then, the system acquires the file type of the simulation result input by the user, determines the display template corresponding to the file type, and inputs the simulation result into the display template to obtain the data file corresponding to the simulation result. Thus, by using the large language model as the "foundation" of the system, it deeply understands and automatically recognizes user intent, and completes the setting of simulation parameters, execution of simulation tasks, and output of simulation results according to the user intent. This helps users complete complex simulation tasks more efficiently, improves the efficiency and accuracy of simulation process orchestration, reduces the error rate, and enriches the diversity and flexibility of simulation result display based on the corresponding display template, effectively meeting user needs.

[0104] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the following examples are provided for illustrative purposes:

[0105] This invention provides an intelligent simulation process orchestration method based on a Large Language Model (LLM). It achieves deep understanding of user intent and dynamic simulation process scheduling through the Long Short-Term Memory module and Prompt Engineering within the LangChain framework. The specific implementation details of this invention are as follows:

[0106] 1. System Architecture and Module Design

[0107] This invention employs the LangChain framework as its core technology, combined with a large language model provided by corresponding AI (Artificial Intelligence), to construct an intelligent agent system for automating the orchestration of simulation processes. The system mainly consists of the following modules:

[0108] Large Language Model Interface: Uses a large language model provided by AI that supports natural language processing and intent recognition as the core inference engine.

[0109] Short-Term Memory Module: Implements short-term and long-term memory through LangChain's BaseMemory interface to store and manage dialogue context information.

[0110] Tool Nodes and Flowcharts: Tool calls and flow control are implemented through LangChain's ToolNode (a node used to call specific tools or operations) and StateGraph (a tool for defining and managing state machines), supporting dynamic simulation flow scheduling. The state machine allows for defining the calling order of tool nodes and supports unconditional jumps, enabling dynamic flow scheduling.

[0111] Simulation Manager: Responsible for setting and managing simulation type, performance, architecture, and model, and supports setting simulation parameters in multi-turn dialogues.

[0112] Simulation utility functions include SetParameterValue (sets simulation parameters), SimulateModel (executes simulation tasks), and generate_report (generates simulation result reports), which are used to call the simulation software API to set parameters, execute simulations, and generate results.

[0113] 2. Implementation of Long and Short Memory Module

[0114] To enhance the agent's ability to understand multi-turn dialogue content, this invention introduces a long short-term memory module, specifically implemented as follows:

[0115] Short-term memory: Implemented through LangChain's InputStoreMemory class, it can record the context information of the current dialogue in real time. Short-term memory saves user input and model output through the save_context method and loads context information through the load_memory_variables method.

[0116] Long-term memory: Implemented through external vector storage (such as a vector database), by utilizing the embedding similarity calculation of the knowledge base, the agent can retrieve historical information related to the current task from external storage. The introduction of long-term memory enables the agent to access historical data or knowledge in complex tasks, improving its ability to deeply understand and process tasks.

[0117] 3. Prompt word engineering and dynamic process scheduling

[0118] This invention utilizes LangChain's prompt engineering combined with a long short-term memory module to achieve deep understanding of user natural language input and dynamic simulation process scheduling. The specific implementation is as follows:

[0119] Prompt generation: LangChain's Prompt engine generates prompts related to the user's query. By combining contextual information from short-term memory and historical data from long-term memory, the agent can dynamically plan the execution order of the simulation process.

[0120] Flowchart and Tool Calls: Dynamic scheduling of the simulation process is achieved through LangChain's StateGraph and ToolNode. The flowchart defines multiple nodes (such as scheduling decision nodes and tool nodes), and transitions between nodes are achieved through conditional edges and ordinary edges. The agent calls the corresponding tool functions according to user intent to complete simulation parameter setting, simulation execution, and result generation.

[0121] 4. Implementation of the simulation manager

[0122] The SimulationManager is responsible for setting and managing simulation types, performance, architecture, and models. Its implementation is as follows:

[0123] Simulation type and performance settings: The agent can dynamically set the simulation type (such as dynamic simulation, economic simulation) and simulation performance (such as 0-100 km / h acceleration, top speed) according to user input.

[0124] Vehicle architecture and model settings: The intelligent agent can dynamically set the vehicle architecture and simulation model and other simulation types based on user input.

[0125] Simulation parameter settings: The agent can dynamically set simulation parameters based on user input and achieve parameter settings through LangChain's prompt words, projects, and tools.

[0126] 5. Implementation of simulation tool functions

[0127] This invention defines several simulation tool functions for calling simulation software APIs to set simulation parameters, execute simulations, and generate results.

[0128] 6. User Intent Recognition and Process Classification

[0129] Intent Recognition: Through a large language model, the agent can determine whether the user's intent is "to set simulation parameters" or "to acquire knowledge" based on the user's input. For example, if the user inputs "I want to perform dynamic simulation," the agent determines that the intent is "to set simulation parameters."

[0130] Process Classification: Based on the intent recognition results, the agent classifies the user's needs into "setting simulation parameters" or "acquiring knowledge" and executes the corresponding process. For example, if the user's intent is "setting simulation parameters," the agent calls the simulation manager to set the simulation parameters; if the user's intent is "acquiring knowledge," the agent calls the large language model to generate knowledge question-and-answer results.

[0131] In the above process, a large language model serves as the foundation for the intelligent agent, enabling deep understanding and automatic recognition of user intentions. Based on these intentions, the simulation software's APIs are invoked to set simulation parameters, execute simulation tasks, and display simulation results in various templates. The implementation of this invention significantly enhances the intelligence level of simulation software, helping users complete complex simulation tasks more efficiently and promoting the further development of industrial simulation technology.

[0132] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0133] Reference Figure 2 This diagram illustrates a structural block diagram of a simulation processing device based on a large language model intelligent agent provided in an embodiment of the present invention. It is applied to a simulation process orchestration system, which includes at least a model interface, a long and short memory module, and a simulation process management tool. Specifically, it may include the following modules:

[0134] The intent recognition module 201 is used to obtain the user's input requirement information, and to call the large language model connected to the simulation process orchestration system through the model interface to perform intent recognition on the requirement information and obtain the intent type corresponding to the requirement information.

[0135] The simulation processing module 202 is used to call the simulation process management tool to set simulation parameters according to the requirement information stored in the long and short memory module if the intent type belongs to the simulation operation class, obtain the corresponding simulation task, execute the simulation task, and generate the corresponding simulation result.

[0136] The result presentation module 203 is used to obtain the file type of the simulation result input by the user, determine the display template corresponding to the file type, input the simulation result into the display template, and obtain the data file corresponding to the simulation result.

[0137] In some feasible implementations, the simulation process management tool includes at least tool nodes and flowcharts; wherein, the tool nodes encapsulate specific operations, and each tool node is used to perform one of the operations of setting parameters, running simulation, and generating a report; the flowchart is a tool for defining and managing state machines, and the flowchart consists of multiple working nodes and the transition logic between the working nodes.

[0138] In some feasible implementations, the simulation processing module 202 is specifically used for:

[0139] Extract the demand information from the long and short memory module;

[0140] Extract target key information from the requirement information, where the target key information is information related to simulation testing;

[0141] The tool node is invoked to set simulation parameters that match the target key information;

[0142] The flowchart is invoked to set the simulation process corresponding to the simulation parameters, and a test task corresponding to the simulation process is generated.

[0143] In some feasible implementations, the simulation process orchestration system further includes simulation tool functions, which at least include parameter setting functions, simulation execution functions, and result generation functions. The simulation processing module 202 is specifically used for:

[0144] The simulation parameters are passed to the simulation software by calling the parameter setting function.

[0145] The simulation execution function is invoked to control the simulation software to execute the simulation process according to the simulation parameters;

[0146] In response to the completion of the simulation process, the result generation function is called to generate the simulation results corresponding to the simulation process.

[0147] In some feasible implementations, the simulation workflow orchestration system further includes a simulation manager; wherein the simulation manager is used for setting and managing simulation types, performance, architecture, and models, and supports setting simulation parameters in multi-turn dialogues; the device further includes:

[0148] The instruction acquisition module is used to acquire the adjustment instructions for the test parameters input by the user;

[0149] The recognition module is used to call the large language model through the model interface to recognize the adjustment command and obtain the simulation adjustment parameters corresponding to the adjustment command;

[0150] An adjustment module is used to adjust the simulation parameters to target simulation parameters that match the simulation adjustment parameters.

[0151] In some feasible implementations, the simulation parameters include at least a first simulation type, a first performance mode, and first model parameters; the simulation adjustment parameters include at least type adjustment parameters, performance adjustment parameters, and model adjustment parameters; and the adjustment module is specifically used for:

[0152] Adjust the first simulation type to a second simulation type that matches the type adjustment parameters;

[0153] Adjust the first performance mode to a second performance mode that matches the performance adjustment parameters;

[0154] The first model parameter is adjusted to a second model parameter that matches the model adjustment parameter.

[0155] Among some feasible implementation methods are:

[0156] The knowledge feedback module is used to obtain the knowledge Q&A results corresponding to the requirement information if the intent type belongs to the knowledge acquisition category.

[0157] In some feasible implementations, the knowledge feedback module is specifically used for:

[0158] The model interface is used to call the large language model to generate knowledge-based question-and-answer results corresponding to the required information; and / or,

[0159] Call a preset knowledge extraction function to extract knowledge question-and-answer results corresponding to the required information from a preset model construction document; and / or,

[0160] The preset experience extraction function is invoked to extract the knowledge Q&A results corresponding to the required information from the preset experience document.

[0161] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0162] In addition, this invention also provides an electronic device, including: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements the various processes of the above-described simulation processing method embodiment based on a large language model intelligent agent and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0163] This invention also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the various processes of the above-described simulation processing method embodiment based on a large language model intelligent agent, achieving the same technical effect. To avoid repetition, it will not be described again here. The computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0164] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0165] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, EEPROM, Flash, and eMMC, etc.) containing computer-usable program code.

[0166] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0167] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0168] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0169] Although preferred embodiments of the present invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0170] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0171] The above provides a detailed description of the simulation processing method and the simulation processing device based on a large language model intelligent agent provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A simulation processing method based on a large language model intelligent agent, characterized in that, The method is applied to a simulation workflow orchestration system, which at least includes a model interface, a short-term memory module, and simulation workflow management tools. The method includes: Obtain the user's input requirement information, and call the large language model connected to the simulation process orchestration system through the model interface to perform intent recognition on the requirement information to obtain the intent type corresponding to the requirement information; If the intent type belongs to the simulation operation type, the simulation process management tool is invoked to set simulation parameters according to the requirement information stored in the long and short memory module, obtain the corresponding simulation task, execute the simulation task, and generate the corresponding simulation result. Obtain the file type of the simulation result input by the user, determine the display template corresponding to the file type, and input the simulation result into the display template to obtain the data file corresponding to the simulation result; The simulation process management tool includes at least tool nodes and a flowchart. Each tool node encapsulates specific operations, and each tool node is used to perform one of the following operations: setting parameters, running the simulation, or generating a report. The flowchart is a tool for defining and managing state machines, and it consists of multiple work nodes and transition logic between those work nodes. Calling the simulation process management tool to set simulation parameters and obtain the corresponding simulation task based on the requirement information stored in the long short-term memory module includes: Extract the demand information from the long and short memory module; Extract target key information from the requirement information, where the target key information is information related to simulation testing; The tool node is invoked to set simulation parameters that match the target key information; The flowchart is invoked to set the simulation process corresponding to the simulation parameters, and a test task corresponding to the simulation process is generated.

2. The method according to claim 1, characterized in that, The simulation workflow orchestration system also includes simulation tool functions, which at least include parameter setting functions, simulation execution functions, and result generation functions. Executing the simulation task and generating the corresponding simulation results includes: The simulation parameters are passed to the simulation software by calling the parameter setting function. The simulation execution function is invoked to control the simulation software to execute the simulation process according to the simulation parameters; In response to the completion of the simulation process, the result generation function is called to generate the simulation results corresponding to the simulation process.

3. The method according to claim 1 or 2, characterized in that, The simulation workflow orchestration system also includes a simulation manager; wherein the simulation manager is used for setting and managing simulation type, performance, architecture, and model, and supports setting simulation parameters in multi-turn dialogues; the method further includes: Obtain the user's input adjustment instructions for the test parameters; The large language model is called through the model interface to identify the adjustment command and obtain the simulation adjustment parameters corresponding to the adjustment command. The simulation parameters are adjusted to match the target simulation parameters.

4. The method according to claim 3, characterized in that, The simulation parameters include at least a first simulation type, a first performance mode, and first model parameters; the simulation adjustment parameters include at least type adjustment parameters, performance adjustment parameters, and model adjustment parameters; adjusting the simulation parameters to target simulation parameters that match the simulation adjustment parameters includes: Adjust the first simulation type to a second simulation type that matches the type adjustment parameters; Adjust the first performance mode to a second performance mode that matches the performance adjustment parameters; The first model parameter is adjusted to a second model parameter that matches the model adjustment parameter.

5. The method according to claim 1, characterized in that, Also includes: If the intent type belongs to the knowledge acquisition category, then obtain the knowledge question and answer results corresponding to the requirement information.

6. The method according to claim 5, characterized in that, The process of obtaining the knowledge-based question-and-answer results corresponding to the required information includes: The model interface is used to call the large language model to generate knowledge-based question-and-answer results corresponding to the required information; and / or, Call a preset knowledge extraction function to extract knowledge question-and-answer results corresponding to the required information from a preset model construction document; and / or, The preset experience extraction function is invoked to extract the knowledge Q&A results corresponding to the required information from the preset experience document.

7. A simulation processing device based on a large language model intelligent agent, characterized in that, An apparatus for use in a simulation workflow orchestration system, the simulation workflow orchestration system including at least a model interface, a short-term memory module, and simulation workflow management tools, the apparatus comprising: The intent recognition module is used to obtain the user's input requirement information, and to call the large language model connected to the simulation process orchestration system through the model interface to perform intent recognition on the requirement information and obtain the intent type corresponding to the requirement information. The simulation processing module is used to call the simulation process management tool to set simulation parameters according to the requirement information stored in the long and short memory module if the intent type belongs to the simulation operation class, obtain the corresponding simulation task, execute the simulation task, and generate the corresponding simulation result. The results presentation module is used to obtain the file type of the simulation results input by the user, determine the display template corresponding to the file type, input the simulation results into the display template, and obtain the data file corresponding to the simulation results; The simulation process management tool includes at least tool nodes and a flowchart. Each tool node encapsulates specific operations, and each tool node is used to perform one of the following operations: setting parameters, running the simulation, or generating a report. The flowchart is a tool for defining and managing state machines, and it consists of multiple working nodes and transition logic between these nodes. The simulation processing module is specifically used for: Extract the demand information from the long and short memory module; Extract target key information from the requirement information, where the target key information is information related to simulation testing; The tool node is invoked to set simulation parameters that match the target key information; The flowchart is invoked to set the simulation process corresponding to the simulation parameters, and a test task corresponding to the simulation process is generated.

8. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; The memory is used to store computer programs; When the processor executes a program stored in the memory, it implements the method as described in any one of claims 1-6.

9. A computer-readable storage medium having instructions stored thereon that, when executed by one or more processors, cause the processors to perform the method as described in any one of claims 1-6.

Images