Digital twin system construction method and device, electronic equipment and storage medium

Abstract

The application discloses a digital twin system construction method and device, electronic equipment and a storage medium, and relates to the technical field of electronic information. The method comprises the following steps: acquiring system description data corresponding to a target system; the target system comprises a plurality of nodes which are connected to each other in communication; performing analysis and processing on the node behaviors based on the system description data by using a twin system analysis model to obtain a node behavior model of each node; the node behavior model is used for indicating the node configuration and the node behavior of the corresponding node; instantiating a virtual node corresponding to each node in a virtual digital system; and configuring the virtual node corresponding to each node based on the node behavior model of each node to obtain a target digital twin system corresponding to the target system. According to the method, the cost and difficulty of constructing the target digital twin system are reduced.

Description

Technical Field

[0001] This application relates to the field of electronic information technology, and more specifically, to a method, apparatus, electronic device, and storage medium for constructing a digital twin system. Background Technology

[0002] A digital twin is a virtual digital system created in a one-to-one, high-fidelity manner that corresponds to a physical entity system. However, related technologies face challenges in constructing virtual digital systems due to their high difficulty and cost. Summary of the Invention

[0003] In view of this, embodiments of this application propose a method, apparatus, electronic device, and storage medium for constructing a digital twin system.

[0004] In a first aspect, embodiments of this application provide a method for constructing a digital twin system. The method includes: acquiring system description data corresponding to a target system; the target system includes multiple nodes that are interconnected; analyzing and processing node behavior based on the system description data using a twin system analysis model to obtain node behavior models for each of the multiple nodes; the node behavior models are used to indicate the node configuration and node behavior of the corresponding nodes; instantiating a virtual node corresponding to each node in a virtual digital system; and configuring the virtual node corresponding to each node based on its respective node behavior model to obtain a target digital twin system corresponding to the target system.

[0005] Secondly, embodiments of this application provide a digital twin system construction apparatus, comprising: an acquisition module for acquiring system description data corresponding to a target system; the target system includes multiple nodes interconnected with each other; an analysis module for analyzing and processing node behavior based on the system description data using a twin system analysis model to obtain node behavior models for each of the multiple nodes; the node behavior models are used to indicate the node configuration and node behavior of the corresponding nodes; an instantiation module for instantiating a virtual node corresponding to each node in a virtual digital system; and a configuration module for configuring the virtual node corresponding to each node based on its respective node behavior model to obtain a target digital twin system corresponding to the target system.

[0006] Optionally, the analysis module is also used to infer topological relationships based on system description data through a twin system analysis model to obtain the node topological relationships corresponding to the target system; the node topological relationships indicate the logical connection relationships and / or physical connection relationships between multiple nodes; and to analyze and process node behavior based on the node topological relationships and the node information of each of the multiple nodes through the twin system analysis model to obtain the node behavior models of each of the multiple nodes.

[0007] Optionally, the configuration module is further configured to compare the state snapshots of the target digital twin system and the target system; if the difference between the state snapshots of the target digital twin system and the target system is greater than a preset difference, the twin system analysis model is used to perform difference analysis on the state snapshots of the target digital twin system and the target system to obtain calibration instructions for the target digital twin system; the target digital twin system is calibrated based on the calibration instructions to obtain a candidate digital twin system; the candidate digital twin system is obtained as a new target digital twin system, and the process of comparing the state snapshots of the target digital twin system and the target system is returned until the difference between the state snapshots of the target digital twin system and the target system is no greater than a preset difference, and the latest target digital twin system is obtained as the first digital twin system for the target system.

[0008] Optionally, the device further includes a change module, which, in response to a system change event for the target system, analyzes the system change event through a twin system analysis model to obtain a twin system change plan corresponding to the target digital twin system; and, based on the twin system change plan, performs change processing on the target digital twin system to obtain a changed digital twin system.

[0009] Optionally, the change module is also used to perform topology change analysis processing based on system change events and node topology relationships through a twin system analysis model to obtain node topology change information for the target digital twin system; the node topology relationship indicates the logical connection relationship and / or physical connection relationship between multiple nodes; the node topology relationship is obtained by the twin system analysis model based on the system description data through topology relationship analysis processing; and the twin system analysis model generates a change plan based on the node topology change information to obtain the twin system change plan corresponding to the target digital twin system.

[0010] Optionally, the change module is also used to determine whether the target digital twin system meets preset conditions based on the state snapshot of the target digital twin system; the preset conditions include that the target digital twin system is in a healthy state and can accept changes; if the target digital twin system meets the preset conditions, the target digital twin system is modified based on the twin system change plan to obtain the modified digital twin system.

[0011] Optionally, the modification module is also used to generate system health test cases for the modified digital twin system through the twin system analysis model; perform health tests on the modified digital twin system based on the system health test cases; and if the modified digital twin system passes the health test, obtain the modified digital twin system as a second digital twin system for the target system.

[0012] Optionally, the device further includes a dialogue processing module, used to respond to receiving dialogue information for the target digital twin system through a dialogue processing model, and to generate an operation plan based on the dialogue information through the dialogue processing model to obtain an operation plan for the target digital twin system; the operation plan includes at least the execution operation plan; controlling the operation of the target digital twin system based on the execution operation plan, and acquiring the operation log of the target digital twin system during the operation of the target digital twin system according to the execution operation plan, as the twin system operation log; and determining the system analysis results of the target digital twin system based on the twin system operation log and the real-time operation status of the target digital twin system.

[0013] Optionally, the processing module is further configured to, in response to receiving dialogue information for the target digital twin system through the dialogue processing model, generate an operation plan for the target digital twin system by performing operation planning based on the dialogue information and node topology relationships through the dialogue processing model; the node topology relationships indicate the logical connection relationships and / or physical connection relationships between multiple nodes; the node topology relationships are obtained by the twin system analysis model based on the system description data through topology relationship analysis and processing.

[0014] Optionally, the operation plan also includes a verification operation plan; the processing module is further used to perform verification processing on the target digital twin system based on the verification operation plan after the target digital twin system has finished running according to the execution operation plan, so as to obtain system verification data of the target digital twin system; and to determine the system analysis results of the target digital twin system based on the system verification data, the twin system operation log and the real-time operation status of the target digital twin system.

[0015] Optionally, the processing module is further configured to, if the system analysis results indicate that the target digital twin system has abnormal behavior, determine the natural language analysis report of the target digital twin system based on the system analysis results through a dialogue processing model; the natural language analysis report includes at least the abnormal behavior and the impact of the abnormal behavior; and display the natural language analysis report.

[0016] Optionally, the node behavior model includes node configuration information indicating the node configuration of the corresponding node and node behavior information constraining the node behavior of the corresponding node. The node configuration information indicates the forwarding logic and control logic between nodes, and the node behavior information indicates the node tasks executed by the node and the node functions implemented. The configuration module is also used to configure the node tasks and node functions of the virtual nodes corresponding to each node based on the node behavior information, and to configure the forwarding logic and control logic between the virtual nodes corresponding to each node based on the node configuration information, so as to obtain the target digital twin system corresponding to the target system.

[0017] Thirdly, embodiments of this application provide an electronic device, including: a processor and a memory; the memory stores computer-readable instructions, which, when executed by the processor, implement the above-described method.

[0018] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-readable instructions that, when executed by a processor, implement the above-described method.

[0019] Fifthly, embodiments of this application provide a computer program product including computer-readable instructions that, when executed by a processor, implement the method described above.

[0020] This application provides a method, apparatus, electronic device, and storage medium for constructing a digital twin system. In this application, a twin system analysis model is used to analyze and process node behavior based on the system description data of the target system to obtain the node configuration of the indicator node and the node behavior model of the node. Then, the virtual nodes instantiated in the virtual digital system are directly configured through the node behavior model to obtain the target digital twin system corresponding to the target system. It is no longer necessary to run the real image of the node in the simulation system to obtain the twin system, thus avoiding the difficulty of obtaining the real image of the node and the high computational resource consumption caused by running the real image. This greatly reduces the construction cost and difficulty of the digital twin system. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A schematic diagram illustrating an application scenario applicable to embodiments of this application is shown; Figure 2 A flowchart illustrating a method for constructing a digital twin system according to an embodiment of this application is shown; Figure 3 A flowchart of a digital twin system construction method according to another embodiment of this application is shown; Figure 4 A flowchart of a digital twin system construction method according to another embodiment of this application is shown; Figure 5 A schematic diagram illustrating the creation process of a digital twin in one embodiment of this application is shown; Figure 6A schematic diagram illustrating the modification process of a digital twin in one embodiment of this application is shown; Figure 7 A schematic diagram of the control process of a digital twin in one embodiment of this application is shown; Figure 8 A block diagram of a digital twin system construction apparatus according to an embodiment of this application is shown; Figure 9 A structural block diagram of an electronic device for performing the methods provided in this application is shown. Detailed Implementation

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0024] In the following description, the terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first" and "second" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit the application. It should be noted that "multiple" as used herein refers to two or more. "And / or" describes the association relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship.

[0026] The following is an explanation of the technical terms used in the embodiments of this application: LLM: Large Language Model, is the core brain of the AI ​​Agent (dialogue processing agent) and LLA Agent (twin system analysis agent) in this application, responsible for instruction parsing, reasoning planning and intelligent feedback.

[0027] SDN: Software-Defined Networking, refers to a network architecture that achieves network programmability through the separation of the control plane and forwarding plane and centralized control. A software-defined networking system is one type of system within the target system discussed in this application.

[0028] Digital Twin: A digital twin is a virtual model system created in virtual space that corresponds to a physical entity system in a one-to-one, high-fidelity manner.

[0029] AI Agent: AI Agent, or dialogue processing intelligent agent, in this solution refers to an intelligent agent with core LLM capabilities, responsible for receiving natural language instructions and driving the digital twin system to perform operations and analyses.

[0030] NLI: Natural Language Interface, refers to the interface that allows users to interact with the system in everyday language, and is the main user interaction method for AI Agents.

[0031] P4: Programming Protocol-Independent Packet Processors is a language specifically designed for programming the network data forwarding plane. It defines the processing logic of data packets and is one of the technological foundations for building high-fidelity forwarding behavior models.

[0032] KPIs: Key Performance Indicators, in this solution, refer to the core metrics used to measure network performance and service experience, such as latency, packet loss rate, and routing convergence time.

[0033] CMDB: Configuration Management Database, is a data system used to store network assets, configuration information, and change records. It is the data source used by the LLM Agent in this application for synchronization and inference.

[0034] Please refer to Figure 1 The diagram illustrates an application scenario applicable to the embodiments of this application. This application scenario includes a terminal 110 and a server 120.

[0035] Terminal 110 includes devices such as smartphones, tablets, e-book readers, music players, wearable devices, smart home devices, and in-vehicle terminals.

[0036] The server 120 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms.

[0037] In some embodiments, taking the electronic device as the server 120 as an example, the digital twin system construction method provided in this application embodiment can be implemented by the server 120, that is, the twin system analysis model is in the server 120. The terminal 110 can respond to the user's construction operation by generating and sending a digital twin system construction request to the server 120. The server 120 responds to the digital twin system construction request by obtaining the system description data corresponding to the target system. Then, based on the system description data, it performs node behavior analysis processing through the twin system analysis model to obtain the node behavior models of multiple nodes. Next, the server 120 continues to instantiate the virtual node corresponding to each node in the virtual digital system. Finally, based on the node behavior model of each node, the server 120 configures the virtual node corresponding to each node to obtain the target digital twin system corresponding to the target system. Furthermore, the server 120 can also return a prompt message to the terminal 110 to notify the user that the construction of the target digital twin system corresponding to the target system is complete.

[0038] In some embodiments, taking the electronic device as terminal 110 as an example, the digital twin system construction method provided in this application can be implemented by terminal 110, that is, the twin system analysis model is in terminal 110. Terminal 110 can respond to the user's construction operation, obtain the system description data corresponding to the target system, and then perform node behavior analysis processing based on the system description data through the twin system analysis model to obtain the node behavior models of multiple nodes respectively; then, terminal 110 continues to instantiate the virtual node corresponding to each node in the virtual digital system; finally, terminal 110 configures the virtual node corresponding to each node based on the node behavior model of each node to obtain the target digital twin system corresponding to the target system. In addition, terminal 110 can also output prompt information to notify the user that the construction of the target digital twin system corresponding to the target system is complete.

[0039] It is worth mentioning that the dialogue model upon which the twin system analysis model and dialogue processing model in this application are based can be trained by the server 120. For example, the server 120 can obtain various sample data from a remote database, and then train the dialogue model upon which the twin system analysis model and dialogue processing model are based based on the sample data. The dialogue model is then attached to the initialized intelligent agent to obtain the dialogue processing model. Thus, the server 120 can use the twin system analysis model and dialogue processing model in the digital twin system construction method of this application.

[0040] Of course, when the terminal 110 acts as the executing entity of the digital twin system construction method of this application, after the server 120 trains the dialogue model on which the twin system analysis model and the dialogue processing model are based, it can also send the dialogue model on which the twin system analysis model and the dialogue processing model are based to the terminal 110, so that the terminal 110 can store the dialogue model on which the twin system analysis model and the dialogue processing model are based, and attach the dialogue model to the initialized intelligent agent to obtain the dialogue processing model. Thus, the terminal 110 can use the twin system analysis model and the dialogue processing model in the digital twin system construction method of this application.

[0041] Alternatively, terminal 110 can obtain various sample data from a remote database through server 120, and train the twin system analysis model and the dialogue model on which the dialogue processing model is based based on the sample data. Then, terminal 110 stores the twin system analysis model and the dialogue model on which the dialogue processing model is based, and attaches the dialogue model to the initialized intelligent agent to obtain the dialogue processing model. Thus, terminal 110 can use the twin system analysis model and the dialogue processing model in the digital twin system construction method of this application.

[0042] For ease of description, this embodiment uses electronic devices as the executing entity of the digital twin system construction method for explanation.

[0043] Please see Figure 2 , Figure 2 This application illustrates a flowchart of a method for constructing a digital twin system according to an embodiment of the present application. The method can be applied to electronic devices, which may be... Figure 1 The terminal 110 or server 120 in the middle, the method may include: S110. Obtain the system description data corresponding to the target system.

[0044] A target system can refer to a system in a real-world environment. A target system includes multiple interconnected nodes, which can be physical or software nodes. For example, a target system can be a computer system composed of multiple computers, with each physical computer serving as a physical node. Alternatively, a target system can be a software-defined network system, where general-purpose network devices, switches, and smart network interface cards (NICs) serve as physical nodes, while data plane software, core controllers, and management and orchestration software—the software used to implement various functions—serve as software nodes.

[0045] System description data refers to data used to describe the target system. This can include data describing the target system's configuration and its operational status. For example, if the target system is a software-defined network (SDN) system, the system description data could include the target system's log files, CLI (Command-Line Interface) configuration information, performance monitoring data, SNMP (Simple Network Management Protocol) data, and node vendor documentation. The log files and performance monitoring data describe the target system's operational status, while the CLI configuration information, SNMP data, and node vendor documentation describe the target system's configuration.

[0046] S120. Analyze and process the node behavior based on the system description data using the twin system analysis model to obtain the node behavior models of multiple nodes.

[0047] The node behavior model is used to indicate the node configuration and node behavior of the corresponding node. That is, the node behavior model can include node configuration information for configuring the node and node behavior information for constraining the node behavior. The node configuration information can include forwarding logic and control logic between nodes, and the node behavior information can include the node tasks performed by the node and the node functions implemented.

[0048] Forwarding logic can include the next-hop node to which the data sent (or forwarded) by a node flows, or all nodes to which the data sent by a node flows. For example, if the next-hop nodes of the data sent by node 11 are nodes 12 and 13, then the forwarding logic of node 11 is "data sent to nodes 12 and 13". Or, if the next-hop nodes of the data sent by node 11 are nodes 12 and 13, and the next-hop node of the data sent by node 12 is node 14, then the forwarding logic of node 11 is "data sent to nodes 12 and 13, and node 12 sends data to node 14".

[0049] The control logic can include the nodes controlled by the node. For example, if node 21 controls nodes 22 and 23, then the control logic of node 21 is "node 21 controls nodes 22 and 23 respectively".

[0050] A node task can refer to the task performed by a node, such as a data acquisition task, a data routing task, or a data storage task; a node function refers to the function that a node has, such as a verification function, a validation function, or a rate limiting function.

[0051] A twin system analysis model can be a large language model with a certain degree of natural language understanding capability, enabling it to parse system description data. Of course, to enable this analytical capability, the parameters of the large language model can be adjusted using sample system description data from a sample system and sample node behavior models generated from the nodes in the sample system. The resulting parameter-adjusted large language model can then be used as the twin system analysis model.

[0052] In some implementations, pre-defined configuration information for a twin system analysis agent can be attached to a twin system analysis model to obtain a twin system analysis agent. The twin system analysis agent then performs the functions of the twin system analysis model—analyzing and processing node behaviors based on system description data to obtain node behavior models for multiple nodes.

[0053] The configuration information of the twin system analysis agent can include the name, description, thought process, task instructions, dialogue examples, recommended prompts, and advanced configuration information of the twin system analysis agent. The description summarizes the task to be performed by the twin system analysis agent. The task instructions are the output of the twin system analysis agent and the rules that the operation process must conform to. The dialogue examples are examples of dialogues made by the twin system analysis agent to understand its work tasks (which may include input questions and corresponding answers). The thought process is the thought process of the twin system analysis agent in creating the node behavior model. The recommended prompts may include prompts to prompt users to input system description data for the target system for node behavior analysis. The advanced configuration information may include some decoding parameters of the twin system analysis model on which the twin system analysis agent depends.

[0054] In other words, the system description data of the target system can be input into the twin system analysis model (or twin system analysis agent). Based on the system description data, the twin system analysis model (or twin system analysis agent) analyzes each node in the target system and the relationships between each node to obtain a node behavior model that indicates the node configuration and node behavior.

[0055] Typically, system description data for different target systems are diverse in format and form. To achieve accurate analysis of system description data, a twin system analysis model (or twin system analysis agent) can process the system description data. This leverages the powerful natural language understanding and reasoning capabilities of the twin system analysis model (or twin system analysis agent) to identify key information in data from different vendors and in different formats, and transform it into preprocessed system description data in a unified, structured data format (JSON or XML). Subsequently, the twin system analysis model (or twin system analysis agent) analyzes and processes node behavior based on the preprocessed system description data.

[0056] In some implementations, S120 may include: A1. Using a twin system analysis model, reason about topological relationships based on system description data to obtain the node topological relationships corresponding to the target system; the node topological relationships indicate the logical and / or physical connection relationships between multiple nodes; A2. By using a twin system analysis model, the node behavior is analyzed and processed based on the node topology relationship and the node information of multiple nodes, resulting in a node behavior model for each of the multiple nodes.

[0057] Among them, node information refers to the configuration information of the node, which may include the node type, operating system version, and complex configurations (such as QoS policies).

[0058] In other words, the system description data of the target system can be input into the twin system analysis model (or twin system analysis agent). The twin system analysis model (or twin system analysis agent) then uses the system description data to infer the topological relationships of each node in the target system, thereby obtaining the node topological relationships that indicate the logical and / or physical connections between multiple nodes. Subsequently, the twin system analysis model (or twin system analysis agent) continues to analyze and process the node behavior based on the obtained node topological relationships and the node information of each node, thereby obtaining the node behavior models of each node. The node behavior models can be in the form of P4 scripts or high-level language models, etc.

[0059] Of course, as mentioned above, the twin system analysis model (or twin system analysis agent) can first preprocess the system description data to obtain preprocessed system description data, and then continue to input the obtained preprocessed system description data into the twin system analysis model (or twin system analysis agent) to reason about the topological relationship in order to obtain the node topological relationship corresponding to the target system.

[0060] S130. Instantiate the virtual node corresponding to each node in the virtual digital system.

[0061] A virtual digital system refers to a virtual environment that can be used to configure virtual nodes corresponding to each node in a target system. This virtual digital system can be implemented through a virtual environment configuration tool, which is a tool that can create a virtual digital system, instantiate virtual nodes in the virtual digital system, and configure the created virtual nodes to obtain the target digital twin system corresponding to the target system.

[0062] In this application, instantiation refers to creating a virtual node corresponding to a given node in a virtual digital system. Typically, this creation can be based on the node information of a given node, where the node information of the created virtual node is usually the same as that of the corresponding node. For example, if mobile terminal sl1 is a node, a virtual mobile terminal xsl1 is created in the virtual digital system based on the node information sls1 of mobile terminal sl1. The node information of this virtual mobile terminal xsl1 can be sls1.

[0063] Electronic devices can use APIs (Application Programming Interfaces) to call virtual environment configuration tools to create virtual digital systems, and instantiate virtual nodes corresponding to each node in the virtual digital system, thereby creating an initial digital twin system for the target system.

[0064] S140. Based on the node behavior model of each node, configure the virtual node corresponding to each node to obtain the target digital twin system corresponding to the target system.

[0065] After creating the digital twin system that completes the initialization of the target system, the virtual nodes corresponding to each node are configured directly based on the node behavior model of each node to achieve the configuration of the initial digital twin system, and the configured digital twin system is obtained as the target digital twin system corresponding to the target system.

[0066] In some implementations, as described above, the node behavior model indicates the node configuration and node behavior of the corresponding node. The node behavior model may include node configuration information indicating the node configuration of the corresponding node and node behavior information constraining the node behavior of the corresponding node. The node configuration information indicates the forwarding logic and control logic between nodes, and the node behavior information indicates the node tasks performed by the node and the node functions implemented. Accordingly, S140 may include: configuring the node tasks and node functions of the virtual nodes corresponding to each node based on the node behavior information, and configuring the forwarding logic and control logic between the virtual nodes corresponding to each node based on the node configuration information, so as to obtain the target digital twin system corresponding to the target system.

[0067] In other words, based on node behavior information, the node tasks and functions of the virtual nodes corresponding to each node can be configured to configure the virtual nodes themselves. Furthermore, based on the node configuration information, the forwarding logic and control logic between the virtual nodes corresponding to each node can be configured to configure the relationships between the virtual nodes corresponding to each node. In this way, the configuration of the virtual nodes themselves and the relationships between different virtual nodes are completed, resulting in the target digital twin system corresponding to the target system.

[0068] In some implementations, after S140, the method may further include: B1. Compare the differences between the state snapshot of the target digital twin system and the state snapshot of the target system; B2. If the difference between the state snapshot of the target digital twin system and the state snapshot of the target system is greater than the preset difference, the difference analysis of the state snapshot of the target digital twin system and the state snapshot of the target system is performed through the twin system analysis model to obtain the calibration instructions for the target digital twin system. B3. Based on the calibration instructions, calibrate the target digital twin system to obtain the candidate digital twin system; B4. Obtain a candidate digital twin system as a new target digital twin system, and return to perform the step of comparing the state snapshots of the target digital twin system and the target system until the difference between the state snapshots of the target digital twin system and the target system is no greater than a preset difference. Obtain the latest target digital twin system as the first digital twin system for the target system.

[0069] In other words, in order to verify whether the configured target digital twin system is compatible with the target system, the differences between the state snapshots of the target digital twin system and the target system can be compared to determine whether the target digital twin system is compatible with the target system based on the differences in the comparison.

[0070] It is worth mentioning that if the target system is a software-defined network system, based on the node behavior model of each node, the virtual nodes corresponding to each node are configured. After obtaining the target digital twin system, we can continue to wait for the network protocol of the target digital twin system to converge, so as to ensure the stable operation of the target digital twin system. This will ensure that the state snapshot of the target digital twin system can accurately indicate the operation status of the target digital twin system.

[0071] The state snapshot of the target digital twin system can include at least the routing table and flow table data of the target digital twin system, and correspondingly, the state snapshot of the target system can include at least the routing table and flow table data of the target system. Of course, to ensure their temporal consistency, after the network protocol of the target digital twin system converges, the state snapshots of the target digital twin system and the target system are collected at the same time point.

[0072] After collecting state snapshots of the target digital twin system and the target system, the state snapshots of the target digital twin system and the target system are directly compared. If the difference between the state snapshots of the target digital twin system and the target system is no greater than a preset difference, it means that the operating state of the target digital twin system and the target system are highly consistent, the operating state of the target digital twin system can accurately indicate the operating state of the target system, and the target digital twin system is a high-fidelity twin system. Therefore, the target digital twin system is directly retained as the twin system for testing. The preset difference can be set based on requirements and is not limited in this application.

[0073] If the difference between the state snapshot of the target digital twin system and the state snapshot of the target system is greater than the preset difference, it means that the operating state of the target digital twin system and the target system is poorly matched, the operating state of the target digital twin system cannot accurately indicate the operating state of the target system, and the target digital twin system is not a high-fidelity twin system. In this case, the difference analysis between the state snapshot of the target digital twin system and the state snapshot of the target system can be further performed by the twin system analysis model (or twin system analysis agent) to locate the cause of the difference between the state snapshot of the target digital twin system and the state snapshot of the target system, and continue to analyze based on the cause of the difference to obtain calibration instructions for the target digital twin system.

[0074] Then, the target digital twin system is calibrated again based on the calibration instructions to obtain a candidate digital twin system. This candidate digital twin system is then used as the new target digital twin system. The process then returns to the previous step of comparing the state snapshots of the target digital twin system and the target system until the difference between their state snapshots is no greater than a preset difference. Finally, the latest target digital twin system is obtained and used as the first digital twin system for the target system. This achieves the verification and high-fidelity calibration of the target digital twin system.

[0075] In some implementations, after obtaining the target digital twin system, the method may further include: generating a first system health test case for the target digital twin system through a twin system analysis model; performing health tests on the target digital twin system based on the first system health test case; and if the target digital twin system passes the health test, obtaining the target digital twin system as a third digital twin system for the target system.

[0076] This can be achieved by inputting a state snapshot of the target digital twin system into a digital twin analysis model (or a digital twin analysis agent), which then analyzes the snapshot to obtain first system health test cases for conducting health tests on the target digital twin system. For example, if the target system is a software-defined network system, the first system health test cases created for the target digital twin system could include tests such as Ping connectivity testing, BGP (Border Gateway Protocol) neighbor status checks, and critical service traffic testing.

[0077] After obtaining the first system health test cases, the target digital twin system can be controlled to run these test cases to perform health testing and obtain test results. The test results can include test data from multiple test items. If the test data from multiple test items are all within the normal range, the target digital twin system is deemed to have passed the health test. If the test data from at least one of the multiple test items is outside the normal range, the target digital twin system is deemed to have failed the health test. The health range is set differently for different test items. For example, for the Ping connectivity test, the health range for the test data is a Ping value between 0-1000ms.

[0078] If the target digital twin system passes the health test, it is acquired as a third digital twin system for subsequent simulation testing, etc. If the target digital twin system fails the health test, a health test failure message can be output to inform the user that the target digital twin system has failed the health test.

[0079] Of course, if the target digital twin system fails the health test, the twin system analysis model (or twin system analysis agent) can analyze the test results of the target digital twin system to obtain a first health repair strategy. Then, the target digital twin system is repaired using the first health repair strategy to obtain a repaired target digital twin system. After that, the repaired target digital twin system is used as a new target digital twin system, and a health test is performed on the new target digital twin system. This process is repeated until a target digital twin system that passes the health test is obtained. The target digital twin system that passes the health test is then used as the third digital twin system.

[0080] In some implementations, after obtaining the first digital twin system, the method may further include: generating a second system health test case for the first digital twin system through a twin system analysis model; performing health tests on the first digital twin system based on the second system health test case; and if the first digital twin system passes the health test, obtaining the first digital twin system as a fourth digital twin system for the target system.

[0081] One approach is to input a state snapshot of the first digital twin system into the twin system analysis model (or twin system analysis agent), and then have the twin system analysis model (or twin system analysis agent) analyze the state snapshot of the first digital twin system to obtain second system health test cases for health testing of the first digital twin system.

[0082] After obtaining the second system health test cases, the first digital twin system can be controlled to run the second system health test cases to perform health testing on the first digital twin system and obtain the test results. The test results may include project test data from multiple test items. If the project test data from multiple test items are all within the normal range, the first digital twin system is determined to have passed the health test. If the project test data from at least one of the multiple test items is outside the normal range, the first digital twin system is determined to have failed the health test.

[0083] If the first digital twin system passes the health test, it is acquired as the fourth digital twin system for the target system and used for subsequent simulation tests, etc. If the first digital twin system fails the health test, a health test failure message can be output to inform the user that the first digital twin system has failed the health test.

[0084] Of course, if the first digital twin system fails the health test, the twin system analysis model (or twin system analysis agent) can analyze the test results of the first digital twin system to obtain a second health repair strategy. Then, the first digital twin system is repaired using the second health repair strategy to obtain a repaired first digital twin system. After that, the repaired first digital twin system is used as a new first digital twin system, and a health test is performed on the new first digital twin system. This process is repeated until a first digital twin system that passes the health test is obtained, and the first digital twin system that passes the health test is used as the fourth digital twin system.

[0085] In this embodiment, the twin system analysis model analyzes and processes node behavior based on the system description data of the target system to obtain the node configuration of the indicator node and the node behavior model of the node. Then, the virtual nodes instantiated in the virtual digital system are directly configured through the node behavior model to obtain the target digital twin system corresponding to the target system. It is no longer necessary to run the real image of the node in the simulation system to obtain the twin system, thus avoiding the difficulty of obtaining the real image of the node and the high computational resource consumption caused by running the real image. This greatly reduces the construction cost and difficulty of the digital twin system.

[0086] Furthermore, after obtaining the target digital twin system, a high-fidelity verification is performed on the target digital twin system based on the differences between the state snapshot of the target digital twin system and the state snapshot of the target system. If the target digital twin system fails the high-fidelity verification, it is further calibrated to obtain a first digital twin system that passes the high-fidelity verification. This makes the obtained first digital twin system more consistent with the target system, and the first digital twin system has high fidelity, thus improving the accuracy and authenticity of the constructed first digital twin system.

[0087] Furthermore, a health test is conducted on the target digital twin system to ensure that the target digital twin system used for subsequent simulation tests is healthy. This avoids situations where unhealthy target digital twin systems are used for subsequent simulation tests, which could lead to test failures or inaccuracies, and ensures the smooth execution of subsequent simulation tests and the accuracy of test results.

[0088] In some implementations, such as Figure 3 As shown, after S140, the method further includes: S210. In response to a system change event targeting the target system, analyze the system change event using a twin system analysis model to obtain a twin system change plan corresponding to the target digital twin system.

[0089] After obtaining the target digital twin system through initial modeling, the target system is generally not static. In order to ensure that the target digital twin system continues to serve as an effective pre-testing platform, it can be updated synchronously when the target system changes.

[0090] The target system typically experiences configuration changes, policy adjustments, or software patch updates daily. If these incremental changes are not synchronized to the target digital twin system in a timely and accurate manner, the validity of test results based on the target digital twin system will rapidly decline.

[0091] System change events refer to events that modify the configuration of the target system, adjust the policies of the target system, update the software of the target system, and add or delete nodes of the target system.

[0092] After detecting a system change event targeting the target system, the system change event is directly acquired. Then, the system change event of the target system is input into the twin system analysis model (or twin system analysis agent). The twin system analysis model (or twin system analysis agent) analyzes the system change event of the target system to obtain a twin system change plan for making changes to the target digital twin system.

[0093] Generally, system change events in the target system are usually unstructured events (e.g., Git commit messages or ticket content: "R1 BGP configuration update"). Therefore, the system change events of the target system can be structured by using a twin system analysis model (or twin system analysis agent) to convert the system change events into a structured, executable set of change instructions, which serves as a preprocessed system change event.

[0094] Subsequently, the preprocessed system change events are input into the twin system analysis model (or twin system analysis agent) for analysis and processing to obtain a twin system change plan for the target digital twin system.

[0095] In some implementations, S210 may include: B1. By using a twin system analysis model to perform topology change analysis based on system change events and node topology relationships, the node topology change information for the target digital twin system is obtained; the node topology relationship indicates the logical connection relationship and / or physical connection relationship between multiple nodes; the node topology relationship is obtained by the twin system analysis model based on the system description data through topology relationship analysis. B2. By using the twin system analysis model to generate a change plan based on the node topology relationship change information, the twin system change plan corresponding to the target digital twin system is obtained.

[0096] In other words, in this implementation, the twin system analysis model (or twin system analysis agent) first performs topology change analysis based on the system change event and the node topology relationship (obtained by the method in S120 of the aforementioned embodiment, which will not be repeated here), to obtain node topology relationship change information. Then, the twin system analysis model (or twin system analysis agent) continues to generate a change plan based on the node topology relationship change information obtained in the previous step, analyzes the impact range of this system change event and the change of node dependency relationship, and obtains a conflict-free, sequentially executed detailed change execution plan, which serves as the twin system change plan corresponding to the target digital twin system.

[0097] S220. Based on the twin system change plan, perform change processing on the target digital twin system to obtain the changed digital twin system.

[0098] After obtaining the digital twin system change plan, the change plan can be converted into a series of API call scripts or configuration scripts. The converted scripts are then sent to the target digital twin system created by the virtual environment configuration tool. The target digital twin system then applies the scripts to implement the change processing, resulting in the changed digital twin system.

[0099] In some implementations, before S220, the method may further include: determining whether the target digital twin system meets preset conditions based on a state snapshot of the target digital twin system; the preset conditions include that the target digital twin system is in a healthy state and is open to change; accordingly, S220 may include: if the target digital twin system meets the preset conditions, performing change processing on the target digital twin system based on the twin system change plan to obtain a changed digital twin system.

[0100] In other words, based on the state snapshot of the target digital twin system, it is first determined whether the digital twin system meets the preset conditions. If it does not meet the conditions, it means that the target digital twin system is not in a healthy state or cannot be changed, and changes to the target digital twin system cannot be made. A prompt message can be output to inform the user that the target digital twin system is not in a healthy state or cannot be changed.

[0101] If the conditions are met, it means that the target digital twin system is in a healthy state and is open to change. In this case, the target digital twin system can be directly modified based on the twin system change plan to obtain the modified digital twin system.

[0102] In some implementations, after S220, the method further includes: C1. Generate system health test cases for the changed digital twin system using the twin system analysis model; C2. Conduct health testing on the changed digital twin system based on system health test cases; C3. If the modified digital twin system passes the health test, obtain the modified digital twin system as a second digital twin system for the target system.

[0103] One approach is to input a snapshot of the changed digital twin system's state into the digital twin system analysis model (or digital twin system analysis agent), which then analyzes the snapshot to obtain system health test cases for conducting health tests on the changed digital twin system.

[0104] After obtaining the system health test cases, the system can be controlled to run these test cases to perform health tests on the modified digital twin system and obtain the test results. The test results can include test data from multiple test items. If the test data from multiple test items are all within the normal range, the modified digital twin system is determined to have passed the health test. If the test data from at least one of the multiple test items is outside the normal range, the modified digital twin system is determined to have failed the health test.

[0105] If the modified digital twin system passes the health test, the modified digital twin system is obtained as a second digital twin system for the target system. If the modified digital twin system fails the health test, a message indicating that the health test failed can be output to inform the user that the modified digital twin system obtained after the modification has failed the health test.

[0106] Of course, if the modified digital twin system fails the health test, the test results of the modified digital twin system can be analyzed by the twin system analysis model (or twin system analysis agent) to obtain a third system health repair strategy. Then, the modified digital twin system is repaired by the third system health repair strategy to obtain a repaired modified digital twin system. After that, the repaired modified digital twin system is used as a new modified digital twin system, and a health test is performed on the new modified digital twin system. This cycle is repeated until a modified digital twin system that passes the health test is obtained, and the modified digital twin system that passes the health test is used as the second digital twin system.

[0107] Generally, after completing the changes to the target digital twin system and obtaining the modified digital twin system, although the change to the target digital twin system has been achieved, since the change is based on the already created target digital twin system, some unknown faults may occur during the change process, which may result in the modified digital twin system being unhealthy. Therefore, system health test cases can be used to perform health tests on the modified digital twin system.

[0108] After obtaining the second digital twin system, it means that changes can be made to the target digital twin system. A difference comparison can be performed between the state snapshots of the second digital twin system and the target system. If the difference between the state snapshots of the second digital twin system and the target system is greater than a preset difference, a difference analysis is performed on the state snapshots of the second digital twin system and the target system using a twin system analysis model to obtain a change calibration instruction for the second digital twin system. Based on the change calibration instruction, the second digital twin system is calibrated to obtain the fifth digital twin system. The fifth digital twin system is then acquired as the new second digital twin system, and the process of comparing the state snapshots of the second digital twin system and the target system is repeated until the difference between the state snapshots of the second digital twin system and the target system is no greater than a preset difference. Finally, the latest second digital twin system is acquired as the sixth digital twin system for the target system.

[0109] That is, the second digital twin system is also subjected to high-fidelity verification, and if the second digital twin system fails the high-fidelity verification, the second digital twin system is calibrated.

[0110] Similarly, after obtaining the first digital twin system, in response to system change events targeting the target system, the system change events can be analyzed through the twin system analysis model to obtain the calibration twin system change plan corresponding to the first digital twin system; based on the calibration twin system change plan, the first digital twin system is modified to obtain the modified digital twin system.

[0111] In other words, the change processing also applies to the resulting first digital twin system. Accordingly, the process of determining the calibration twin system change plan may include: performing topology change analysis based on system change events and node topology relationships using a twin system analysis model to obtain node topology change information for the target digital twin system; node topology relationships indicate the logical and / or physical connection relationships between multiple nodes; node topology relationships are obtained by the twin system analysis model based on system description data; and generating a change plan based on the node topology change information using the twin system analysis model to obtain the calibration twin system change plan corresponding to the target digital twin system.

[0112] Similarly, before modifying the first digital twin system based on the calibration twin system change plan to obtain the modified digital twin system, it can be determined whether the first digital twin system meets the corresponding calibration preset conditions based on the state snapshot of the first digital twin system. The calibration preset conditions include that the first digital twin system is in a healthy state and is acceptable for modification. Accordingly, modifying the first digital twin system based on the calibration twin system change plan to obtain the modified digital twin system includes: if the first digital twin system meets the calibration preset conditions, modifying the first digital twin system based on the calibration twin system change plan to obtain the sixth digital twin system.

[0113] Of course, after obtaining the sixth digital twin system, health testing can continue to be conducted on the sixth digital twin system. Specifically, this can include: generating third system health test cases for the sixth digital twin system through the twin system analysis model; conducting health testing on the sixth digital twin system based on the third system health test cases; if the sixth digital twin system passes the health test, the sixth digital twin system is obtained as the seventh digital twin system for the target system.

[0114] Correspondingly, if the output change digital twin system fails the health test, the test results of the sixth digital twin system can be analyzed using the twin system analysis model (or twin system analysis agent) to obtain the third system health repair strategy. Then, the sixth digital twin system is repaired using the third system health repair strategy to obtain the repaired sixth digital twin system. After that, the repaired sixth digital twin system is used as the new sixth digital twin system, and the new sixth digital twin system is subjected to health testing. This process is repeated until a sixth digital twin system that passes the health test is obtained, and the sixth digital twin system that passes the health test is used as the seventh digital twin system.

[0115] After obtaining the seventh digital twin system, it means that changes to the target digital twin system can be implemented. A difference comparison can be performed between the state snapshots of the seventh digital twin system and the target system. If the difference between the state snapshots of the seventh digital twin system and the target system is greater than a preset difference, a difference analysis is performed on the state snapshots of the seventh digital twin system and the target system using a twin system analysis model to obtain a change calibration instruction for the seventh digital twin system. Based on the change calibration instruction, the seventh digital twin system is calibrated to obtain the eighth digital twin system. The eighth digital twin system is then acquired as the new seventh digital twin system, and the process of comparing the state snapshots of the seventh digital twin system and the target system is repeated until the difference between the state snapshots of the seventh digital twin system and the target system is no greater than a preset difference. Finally, the latest eighth digital twin system is acquired as the ninth digital twin system for the target system.

[0116] In this embodiment, after creating the target digital twin system, it can continue to respond to system change events of the target system and update the target digital twin system corresponding to the target system in real time based on the system change events of the target system. This realizes the real-time update of the target digital twin system, making the obtained modified digital twin system more consistent with the modified target system and the modified digital twin system more accurate.

[0117] Furthermore, after modifying the target digital twin system to obtain the modified digital twin system, a health test is conducted on the modified digital twin system to obtain a second digital twin system that passes the health test, ensuring that the second digital twin system is healthier and more stable.

[0118] In some implementations, such as Figure 4 As shown, after S140, the method further includes: S310. In response to receiving dialogue information for the target digital twin system through the dialogue processing model, the dialogue processing model generates an operation plan based on the dialogue information to obtain an operation plan for the target digital twin system.

[0119] Operation planning includes at least execution operation planning, which refers to the operation plan performed by the target digital twin system. For example, execution operation planning may include upgrade operation planning, etc.

[0120] Dialogue processing models can be large language models with a certain degree of natural language understanding capabilities, enabling them to analyze dialogue information.

[0121] In some implementations, pre-defined dialogue processing agent configuration information can be attached to the dialogue processing model to obtain a dialogue processing agent, which can then be used to implement the functions of the dialogue processing model.

[0122] The configuration information of a dialogue processing agent can include the agent's name, description, thought process, task instructions, dialogue examples, recommended prompts, and advanced configuration information. The description summarizes the task the agent must perform; the task instructions define the agent's output and the rules it must follow; the dialogue examples (which may include the input question and its corresponding answer) are examples of dialogues the agent uses to understand its task; the thought process describes the agent's analysis of the dialogue information; recommended prompts may include prompts for user input; and the advanced configuration information may include decoding parameters of the dialogue processing model on which the agent relies.

[0123] Users can send instructions in natural language format as dialogue information to the dialogue processing model (or dialogue processing model agent). The dialogue processing model (or dialogue processing model agent) analyzes the dialogue information to obtain an operation plan that can be executed by the target digital twin system.

[0124] In some implementations, S310 may further include: in response to receiving dialogue information for the target digital twin system through a dialogue processing model, generating an operation plan based on the dialogue information and node topology relationships through the dialogue processing model to obtain an operation plan for the target digital twin system; the node topology relationships indicate the logical connection relationships and / or physical connection relationships between multiple nodes; the node topology relationships are obtained by the twin system analysis model based on the system description data through topology relationship analysis processing.

[0125] Users can send commands in natural language format as dialogue information to the dialogue processing model (or dialogue processing model agent). The dialogue processing model (or dialogue processing model agent) analyzes the dialogue information to obtain clear and structured operation intentions and parameters. Then, the dialogue processing model (or dialogue processing model agent) continues to analyze the obtained operation intentions, parameters, and node topology relationships to automatically reason and generate a detailed, sequentially executed operation sequence (Action Plan) as an operation plan by combining the node topology relationships as accurate context dependencies.

[0126] S320. Control the operation of the target digital twin system based on the execution operation plan, and obtain the operation log of the target digital twin system during the operation process according to the execution operation plan, as the twin system operation log.

[0127] After obtaining the operation plan, the target digital twin system is directly controlled to run based on the operation plan, and the operation log of the target digital twin system during the execution of the operation plan is collected as the twin system operation log.

[0128] S330. Based on the twin system's operation log and the real-time operation status of the target digital twin system, determine the system analysis results of the target digital twin system.

[0129] After obtaining the twin system's operation log, the target digital twin system is analyzed based on the twin system's operation log and the target digital twin system's real-time operation status to obtain system analysis results indicating the target digital twin system's operation status under normal operating conditions.

[0130] Of course, the system analysis results of the target digital twin system can be determined by using a dialogue processing model (or a dialogue processing model agent) based on the twin system's operation logs and the real-time operation status of the target digital twin system.

[0131] In some implementations, the operation plan further includes a verification operation plan; the verification operation plan refers to an operation plan used to verify or test the target digital twin system, for example, the verification operation plan may include an operation plan for conducting connectivity tests; accordingly, S330 includes: after the target digital twin system has finished running according to the execution operation plan, performing verification processing on the target digital twin system based on the verification operation plan to obtain system verification data of the target digital twin system; and determining the system analysis results of the target digital twin system based on the system verification data, the twin system operation log, and the real-time operating status of the target digital twin system.

[0132] In other words, after obtaining the operation plan, the target digital twin system can be controlled to operate based on the verification operation plan included in the operation plan, and data from the target digital twin system during the operation of the verification operation plan can be collected as system verification data.

[0133] Finally, by combining the system verification data, the twin system operation logs, and the real-time operating status of the target digital twin system, the system analysis results of the target digital twin system are determined. At this point, the system analysis results can indicate not only the operating status of the target digital twin system under normal operating conditions, but also its operating status during the verification process.

[0134] Of course, here we can also use a dialogue processing model (or a dialogue processing model agent) to determine the system analysis results of the target digital twin system based on system verification data, the twin system operation logs, and the real-time operation status of the target digital twin system.

[0135] In some implementations, after S330, the method further includes: if the system analysis results indicate that the target digital twin system has abnormal behavior, determining the natural language analysis report of the target digital twin system based on the system analysis results through a dialogue processing model; the natural language analysis report includes at least the abnormal behavior and the impact of the abnormal behavior; and displaying the natural language analysis report.

[0136] In other words, the dialogue processing model (or dialogue processing model agent) analyzes and processes the system analysis results of the target digital twin system to obtain a visualized natural language analysis report, and then displays the visualized natural language analysis report to the user so that the user can quickly view the natural language analysis report.

[0137] It is worth mentioning that the modified digital twin system, the first digital twin system, and the second digital twin system obtained in the aforementioned embodiments can all be used as target digital twin systems and processed according to steps S310-S330 to achieve control and interaction with the modified digital twin system, the first digital twin system, and the second digital twin system through a dialogue processing model, thereby obtaining a natural language analysis report for the modified digital twin system, the first digital twin system, and the second digital twin system.

[0138] In this embodiment, the target digital twin system is controlled and tested through dialogue information in natural language by using a dialogue processing model. This eliminates the need for complex command-line interfaces or programming scripts, thereby improving the efficiency of controlling and testing the target digital twin system and reducing the difficulty and cost of doing so.

[0139] Furthermore, the dialogue processing model is used to further analyze the system analysis results of the target digital twin system to obtain a natural language analysis report, which allows users to quickly view the natural language analysis report without having to manually analyze logs and test data. This improves the efficiency of analyzing the operational and testing status of the target digital twin system and reduces the analysis cost of the operational and testing status of the target digital twin system.

[0140] Furthermore, it allows all execution and verification operations to be performed risk-free in a completely isolated and precisely replicated sandbox environment (i.e., the target digital twin system). This ensures that all potential faults, performance degradation, and business impacts have been identified and resolved before changes are deployed to the production environment, thereby guaranteeing network stability and business continuity.

[0141] To facilitate a better understanding of the solution presented in this application, an example is provided below to explain the digital twin system construction method of this application. In this example, the LLM Agent is a twin system analysis agent configured through a twin system analysis model, and the AI ​​Agent is a dialogue analysis agent configured through a dialogue analysis model.

[0142] The construction process of the target digital twin system is as follows Figure 5 As shown, the construction process of a digital twin system consists of four main stages: data acquisition, intelligent modeling, system instantiation, and fidelity calibration. Phase 1: Data Acquisition and Processing Data Import (1.1): Import a large amount of heterogeneous data from the production environment data source (that is, the target system in the production environment) in batches. The system description data includes unstructured CLI configuration output, logs, and performance monitoring data.

[0143] Parsing and Standardization (2.1): The LLM Agent intervenes as the core processor. It uses its powerful natural language understanding and reasoning capabilities to identify key information in the system description data from different manufacturers and in different formats, and transforms it into a unified, structured data format to obtain preprocessed system description data.

[0144] Topology Reasoning and Storage (2.2, 2.3): Based on the parsed preprocessed system description data, the LLM Agent intelligently derives the logical and physical connection relationships between devices—node topology relationships. These node topology relationships can be constructed into a network knowledge graph and stored in a knowledge base or CMDB, laying the foundation for subsequent dependency analysis.

[0145] Phase Two: Model Generation and System Instantiation Behavioral Model Generation (2.4): The LLM Agent generates or selects behavioral model code that accurately reflects the control and forwarding logic of a node based on the node information (including device type, operating system version and complex configuration) and the node topology, as the node behavioral model.

[0146] System instantiation and configuration distribution (3.1, 3.2): The LLM Agent calls the API of the virtual digital system to instantiate the virtual nodes corresponding to each node and distribute the node behavior model obtained above.

[0147] Phase 3: Fidelity Verification and Calibration State Acquisition and Comparison (4.1, 4.2, 4.3): The target digital twin system starts up and waits for the network protocol to converge. Then, the LL Agent acquires a state snapshot of the target digital twin system and compares it with the ground truth snapshot acquired from the target system.

[0148] Difference Analysis and Iterative Calibration (4.4, 4.5, 4.6): The LLM Agent runs a fidelity verification algorithm to analyze the difference between the state snapshot of the target digital twin system and the real state snapshot collected by the target system.

[0149] If the difference exceeds the preset threshold, the LLM Agent automatically infers and sends a calibration command to the target digital twin system (e.g., adjusting the latency or packet loss model parameters of a certain link), enters iterative optimization, until the difference meets the standard, and returns the calibrated state.

[0150] Modeling complete (4.7): Once the fidelity requirement is met, the LLM Agent reports successful modeling to the developers, at which point a high-fidelity first digital twin system is achieved.

[0151] The synchronization process of the target digital twin system is driven by the LLM Agent as the core intelligent coordinator and execution planner, ensuring full automation from change capture to application verification, involving two phases: change capture and intelligent planning, and automated application and behavior simulation.

[0152] The change process of the target digital twin system is as follows Figure 6 As shown, the first phase: change capture and intelligent planning Triggering a synchronization event (1): Any configuration or state change that occurs in the target system triggers a system change event, such as a system change event being submitted to the version control system (CMDB) or the controller's policy update, which triggers a policy update event.

[0153] Intent recognition and structuring (2): After receiving a system change event, the LLM Agent first uses its powerful natural language understanding ability to parse unstructured system change events (such as Git commit information or ticket content: "R1 BGP configuration update") and transforms them into a structured, executable set of change instructions.

[0154] Dependency Analysis and Planning (3, 6): Based on the previously constructed network knowledge graph, the LLM Agent analyzes the scope of impact and dependencies of this change, infers the downstream devices or logical components that may be affected by the change, and formulates a conflict-free, sequential twin system change plan accordingly.

[0155] Pre-synchronization checks (4, 5): Before actual synchronization, the LLM Agent requests a snapshot of the current state from the target digital twin system so that the LLM Agent can analyze the snapshot to ensure that the target digital twin system is in a healthy state that is open to change.

[0156] Phase Two: Automated Applications and Behavioral Simulation Execute synchronization instructions (7): The LLM Agent converts the twin system change plan into a series of API calls or configuration scripts and sends them to the target digital twin system. The target digital twin system applies the configuration based on the received scripts and simulates the behavior of network devices and controllers to realize the change processing of the target digital twin system.

[0157] Synchronous health checks (8, 9): After the target digital twin system change is completed, the LLM Agent immediately generates system health test cases for the changed digital twin system, and the changed digital twin system executes the system health test cases to realize a series of automated health checks to verify whether the changed digital twin system is successful and stable.

[0158] Fidelity verification (10): The LLM Agent compares the state snapshot of the changed digital twin system with the state snapshot of the target system to achieve high fidelity verification. This process is similar to the “third stage” of the aforementioned construction process and will not be described in detail here.

[0159] Interaction with R&D personnel (11): After the synchronization process is completed, the target digital twin system enters standby mode. R&D personnel can interact directly with the LLM Agen using natural language, for example, to inquire about the specific status after the changes.

[0160] Intelligent Feedback (12): The LLM Agent performs advanced analysis and refinement based on the massive logs and state data within the twin system, and provides real-time feedback to developers in a concise and highly readable natural language format (e.g., "R1 BGP has successfully converged, and the new strategy has taken effect"). This allows developers to understand the synchronization results without delving into complex system logs and complete the final verification before testing.

[0161] The AI ​​Agent that interacts with the digital twin system is built on a Large Model (LLM) and sits in the middle layer between the developers and the digital twin system's API / model. It transforms unstructured human intentions into structured system operations and converts complex system outputs into human-understandable insights.

[0162] The various capabilities of this AI Agent are shown in the table below: The process of controlling a target digital twin system through an AI Agent can include three stages: instruction input and planning (NLU & Reasoning), operation execution and data collection, and intelligent analysis and feedback (Synthesis & Pre-Judgment).

[0163] The control process of the target digital twin system is as follows Figure 7 As shown.

[0164] Phase 1: Instruction Input and Planning (NLU & Reasoning) R&D personnel → AI Agent (1. Natural language instructions): The process is initiated by the R&D personnel through dialogue information. This dialogue information can be a complex natural language instruction, which includes the operation goal, behavior and verification requirements.

[0165] AI Agent → AI Agent (2. Instruction Parsing): The AI ​​Agent uses its LLM core to parse the semantics of dialogue information, transforming unstructured text into clear, structured operational intentions and parameters.

[0166] AI Agent → Knowledge Base / Graph (3, 4: Knowledge Retrieval): The AI ​​Agent accesses the knowledge base graph to obtain node topological relationships, providing accurate context for planning.

[0167] AI Agent → AI Agent (5. Generate Operation Plan): Based on dialogue information and node topology relationships, the AI ​​Agent automatically infers and generates a detailed, sequentially executed operation sequence (Action Plan) as the operation plan.

[0168] Phase Two: Operation Execution and Data Collection AI Agent → Target Digital Twin System (6, 7. Execution Operations and State Updates): The AI ​​Agent translates the execution operation plan (such as upgrade) in the operation plan into API calls to the target digital twin system and executes them. The target digital twin system returns real-time logs of the execution process -- the twin system operation log.

[0169] AI Agent → Target Digital Twin System (8, 9. Validation and Data Collection): The AI ​​Agent executes the validation operation plan (such as connectivity testing) in the operation plan. The target digital twin system returns the key performance indicators (KPIs) and final state data of the test, which serve as system validation data.

[0170] Phase Three: Intelligent Analysis and Feedback (Synthesis & Pre-Judgment) AI Agent → AI Agent (10. Comprehensive Analysis): The AI ​​Agent integrates all collected twin system operation logs and system verification data. The LLM compares the actual results with the R&D personnel's expected goals to identify any abnormal behaviors and obtain system analysis results.

[0171] AI Agent → AI Agent (11. Risk and Impact Reasoning): The AI ​​Agent uses the reasoning capabilities of LLM to derive anomalous behavior and its impact based on system analysis results (e.g., determining the expected business impact if implemented in a production environment).

[0172] AI Agent → R&D Personnel (12. Structured Feedback Report): Finally, the AI ​​Agent extracts the complex results into a concise and easy-to-understand natural language report—the Natural Language Analysis Report—and feeds it directly back to the R&D personnel, completing the complete closed loop from instructions to intelligent prediction.

[0173] The core benefit of this solution is that it completely reduces the risks associated with network changes and significantly improves the intelligence level of network operation and maintenance and R&D. By creating a high-fidelity digital twin system and integrating an AI Agent, all device upgrades, configuration changes, or critical operations can be pre-simulated in a completely isolated and accurately replicated sandbox environment without risk. This ensures that all potential faults, performance degradation, and business impacts have been identified and resolved before changes are deployed to the production environment, thereby guaranteeing network stability and business continuity.

[0174] Furthermore, the AI ​​Agent enables seamless and efficient interaction between researchers and complex network systems. Researchers no longer need to master complex command-line interfaces or programming scripts; they can complete complex test scenario design, fault injection, and result analysis simply through natural language.

[0175] Furthermore, the Agent can extract massive and obscure system data into concise and accurate risk prediction and impact analysis reports, elevating the operation and maintenance work from tedious configuration execution and log investigation to decision-making based on intelligent insights, realizing a revolutionary transformation of network operation and maintenance mode from "passive response" to "proactive prediction".

[0176] Please see Figure 8 , Figure 8 This illustration shows a block diagram of a digital twin system construction apparatus according to an embodiment of this application. The apparatus 900 includes: The acquisition module 910 is used to acquire system description data corresponding to the target system; the target system includes multiple nodes that are interconnected. Analysis module 920 is used to analyze and process node behavior based on system description data through twin system analysis model to obtain node behavior models for multiple nodes; the node behavior model is used to indicate the node configuration and node behavior of the corresponding node; Instantiation module 930 is used to instantiate the virtual node corresponding to each node in the virtual digital system; Configuration module 940 is used to configure the virtual nodes corresponding to each node based on the node behavior model of each node, so as to obtain the target digital twin system corresponding to the target system.

[0177] Optionally, the analysis module is also used to infer topological relationships based on system description data through a twin system analysis model to obtain the node topological relationships corresponding to the target system; the node topological relationships indicate the logical connection relationships and / or physical connection relationships between multiple nodes; and to analyze and process node behavior based on the node topological relationships and the node information of each of the multiple nodes through the twin system analysis model to obtain the node behavior models of each of the multiple nodes.

[0178] Optionally, the configuration module is further configured to compare the state snapshots of the target digital twin system and the target system; if the difference between the state snapshots of the target digital twin system and the target system is greater than a preset difference, the twin system analysis model is used to perform difference analysis on the state snapshots of the target digital twin system and the target system to obtain calibration instructions for the target digital twin system; the target digital twin system is calibrated based on the calibration instructions to obtain a candidate digital twin system; the candidate digital twin system is obtained as a new target digital twin system, and the process of comparing the state snapshots of the target digital twin system and the target system is returned until the difference between the state snapshots of the target digital twin system and the target system is no greater than a preset difference, and the latest target digital twin system is obtained as the first digital twin system for the target system.

[0179] Optionally, the device further includes a change module, which, in response to a system change event for the target system, analyzes the system change event through a twin system analysis model to obtain a twin system change plan corresponding to the target digital twin system; and, based on the twin system change plan, performs change processing on the target digital twin system to obtain a changed digital twin system.

[0180] Optionally, the change module is also used to perform topology change analysis processing based on system change events and node topology relationships through a twin system analysis model to obtain node topology change information for the target digital twin system; the node topology relationship indicates the logical connection relationship and / or physical connection relationship between multiple nodes; the node topology relationship is obtained by the twin system analysis model based on the system description data through topology relationship analysis processing; and the twin system analysis model generates a change plan based on the node topology change information to obtain the twin system change plan corresponding to the target digital twin system.

[0181] Optionally, the change module is also used to determine whether the target digital twin system meets preset conditions based on the state snapshot of the target digital twin system; the preset conditions include that the target digital twin system is in a healthy state and can accept changes; if the target digital twin system meets the preset conditions, the target digital twin system is modified based on the twin system change plan to obtain the modified digital twin system.

[0182] Optionally, the modification module is also used to generate system health test cases for the modified digital twin system through the twin system analysis model; perform health tests on the modified digital twin system based on the system health test cases; and if the modified digital twin system passes the health test, obtain the modified digital twin system as a second digital twin system for the target system.

[0183] Optionally, the device further includes a dialogue processing module, used to respond to receiving dialogue information for the target digital twin system through a dialogue processing model, and to generate an operation plan based on the dialogue information through the dialogue processing model to obtain an operation plan for the target digital twin system; the operation plan includes at least the execution operation plan; controlling the operation of the target digital twin system based on the execution operation plan, and acquiring the operation log of the target digital twin system during the operation of the target digital twin system according to the execution operation plan, as the twin system operation log; and determining the system analysis results of the target digital twin system based on the twin system operation log and the real-time operation status of the target digital twin system.

[0184] Optionally, the processing module is further configured to, in response to receiving dialogue information for the target digital twin system through the dialogue processing model, generate an operation plan for the target digital twin system by performing operation planning based on the dialogue information and node topology relationships through the dialogue processing model; the node topology relationships indicate the logical connection relationships and / or physical connection relationships between multiple nodes; the node topology relationships are obtained by the twin system analysis model based on the system description data through topology relationship analysis and processing.

[0185] Optionally, the operation plan also includes a verification operation plan; the processing module is further used to perform verification processing on the target digital twin system based on the verification operation plan after the target digital twin system has finished running according to the execution operation plan, so as to obtain system verification data of the target digital twin system; and to determine the system analysis results of the target digital twin system based on the system verification data, the twin system operation log and the real-time operation status of the target digital twin system.

[0186] Optionally, the processing module is further configured to, if the system analysis results indicate that the target digital twin system has abnormal behavior, determine the natural language analysis report of the target digital twin system based on the system analysis results through a dialogue processing model; the natural language analysis report includes at least the abnormal behavior and the impact of the abnormal behavior; and display the natural language analysis report.

[0187] Optionally, the node behavior model includes node configuration information indicating the node configuration of the corresponding node and node behavior information constraining the node behavior of the corresponding node. The node configuration information indicates the forwarding logic and control logic between nodes, and the node behavior information indicates the node tasks executed by the node and the node functions implemented. The configuration module 940 is also used to configure the node tasks and node functions of the virtual nodes corresponding to each node based on the node behavior information, and to configure the forwarding logic and control logic between the virtual nodes corresponding to each node based on the node configuration information, so as to obtain the target digital twin system corresponding to the target system.

[0188] It should be noted that the device embodiments in this application correspond to the aforementioned method embodiments. The specific principles in the device embodiments can be found in the content of the aforementioned method embodiments, and will not be repeated here.

[0189] Please see Figure 9 This diagram illustrates a structural block diagram of an electronic device used to perform the methods provided in this application. The electronic device may be... Figure 1 The terminal is 110 or the server is 120. It should be noted that... Figure 9 The electronic device 1200 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0190] like Figure 9As shown, the electronic device 1200 includes a Central Processing Unit (CPU) 1201, which can perform various appropriate actions and processes, such as executing the methods described in the above embodiments, based on a program stored in Read-Only Memory (ROM) 1202 or a program loaded from storage portion 1208 into Random Access Memory (RAM) 1203. The RAM 1203 also stores various programs and data required for system operation. The CPU 1201, ROM 1202, and RAM 1203 are interconnected via a bus 1204. An Input / Output (I / O) interface 1205 is also connected to the bus 1204.

[0191] The following components are connected to I / O interface 1205: an input section 1206 including a keyboard, mouse, etc.; an output section 1207 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 1208 including a hard disk, etc.; and a communication section 1209 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 1209 performs communication processing via a network such as the Internet. A drive 1210 is also connected to I / O interface 1205 as needed. Removable media 1211, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 1210 as needed so that computer program products read from them can be installed into storage section 1208 as needed.

[0192] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program product can be downloaded and installed from a network via communication section 1209, and / or installed from removable media 1211. When the computer program product is executed by the central processing unit (CPU) 1201, it performs various functions defined in the system of this application.

[0193] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any group thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any group thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable group thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such transmitted data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0194] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and groups of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a group of dedicated hardware and computer-readable instructions.

[0195] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0196] In another aspect, this application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable storage medium carries computer-readable instructions that, when executed by a processor, implement the methods in any of the above embodiments.

[0197] According to one aspect of the embodiments of this application, a computer program product is provided, the computer program product including computer-readable instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer-readable instructions from the computer-readable storage medium, and the processor executes the computer-readable instructions, causing the electronic device to perform the methods of any of the above embodiments.

[0198] In the embodiments of this application, the terms "module" or "unit" refer to a computer program product or a part of a computer program product that has a predetermined function and works together with other related parts to achieve a predetermined goal. It can be implemented wholly or partially using software, hardware (e.g., processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that functions as a whole.

[0199] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0200] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause an electronic device (such as a personal computer, server, touch terminal, or network device, etc.) to execute the methods according to the embodiments of this application.

[0201] Other embodiments of this application will readily conceive of by those skilled in the art upon consideration of the specification and practice of the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. It should be understood that this application is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

[0202] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for constructing a digital twin system, characterized in that, The method includes: Obtain system description data corresponding to the target system; the target system includes multiple nodes that are interconnected. The twin system analysis model analyzes and processes the node behavior based on the system description data to obtain the node behavior model of each of the multiple nodes; the node behavior model is used to indicate the node configuration and node behavior of the corresponding node. Instantiate a virtual node corresponding to each of the aforementioned nodes in the virtual digital system; Based on the node behavior model of each node, the virtual node corresponding to each node is configured to obtain the target digital twin system corresponding to the target system.

2. The method according to claim 1, characterized in that, The step of analyzing and processing node behavior based on the system description data using a twin system analysis model to obtain the node behavior model for each of the multiple nodes includes: The twin system analysis model infers topological relationships based on the system description data to obtain the node topological relationships corresponding to the target system; the node topological relationships indicate the logical connection relationships and / or physical connection relationships between the multiple nodes. The twin system analysis model analyzes and processes node behavior based on the node topology and the individual node information of the multiple nodes to obtain the node behavior model of each of the multiple nodes.

3. The method according to claim 1, characterized in that, After configuring the virtual nodes corresponding to each node based on the node behavior model of each node to obtain the target digital twin system corresponding to the target system, the method further includes: A difference comparison is performed between the state snapshot of the target digital twin system and the state snapshot of the target system; If the difference between the state snapshot of the target digital twin system and the state snapshot of the target system is greater than a preset difference, the difference analysis of the state snapshot of the target digital twin system and the state snapshot of the target system is performed by the twin system analysis model to obtain a calibration instruction for the target digital twin system. The target digital twin system is calibrated based on the calibration instructions to obtain a candidate digital twin system; The candidate digital twin system is obtained as a new target digital twin system, and the process of comparing the state snapshots of the target digital twin system and the target system is returned until the difference between the state snapshots of the target digital twin system and the target system is no greater than the preset difference. The latest target digital twin system is then obtained as the first digital twin system for the target system.

4. The method according to claim 1, characterized in that, After configuring the virtual nodes corresponding to each node based on the node behavior model of each node to obtain the target digital twin system corresponding to the target system, the method further includes: In response to a system change event for the target system, the system change event is analyzed using the twin system analysis model to obtain a twin system change plan corresponding to the target digital twin system. Based on the twin system change plan, the target digital twin system is modified to obtain the modified digital twin system.

5. The method according to claim 4, characterized in that, In response to a system change event targeting the target system, the system change event is analyzed using the twin system analysis model to obtain a twin system change plan corresponding to the target digital twin system, including: The twin system analysis model performs topology change analysis based on system change events and node topology relationships to obtain node topology change information for the target digital twin system. The node topology relationships indicate the logical and / or physical connection relationships between the multiple nodes. The node topology relationships are obtained by the twin system analysis model based on the system description data. The twin system analysis model generates a change plan based on the node topology change information to obtain the twin system change plan corresponding to the target digital twin system.

6. The method according to claim 4, characterized in that, Before performing change processing on the target digital twin system based on the digital twin system change plan to obtain the changed digital twin system, the method further includes: Based on a snapshot of the target digital twin system's status, determine whether the target digital twin system meets preset conditions; the preset conditions include that the target digital twin system is in a healthy state and is open to change. The process of modifying the target digital twin system based on the digital twin system change plan to obtain a modified digital twin system includes: If the target digital twin system meets the preset conditions, the target digital twin system is modified based on the twin system change plan to obtain the modified digital twin system.

7. The method according to claim 4, characterized in that, After modifying the target digital twin system based on the digital twin system change plan to obtain the modified digital twin system, the method further includes: The twin system analysis model is used to generate system health test cases for the modified digital twin system; Health testing is performed on the modified digital twin system based on the system health test cases. If the modified digital twin system passes the health test, the modified digital twin system is obtained as a second digital twin system for the target system.

8. The method according to claim 1, characterized in that, After configuring the virtual nodes corresponding to each node based on the node behavior model of each node to obtain the target digital twin system corresponding to the target system, the method further includes: In response to receiving dialogue information for the target digital twin system through a dialogue processing model, the dialogue processing model performs operation planning generation processing based on the dialogue information to obtain an operation plan for the target digital twin system; the operation plan includes at least an execution operation plan. The target digital twin system is controlled to run based on the execution operation plan, and the operation log of the target digital twin system during the operation process according to the execution operation plan is obtained as the twin system operation log; Based on the twin system's operation logs and the real-time operating status of the target digital twin system, the system analysis results of the target digital twin system are determined.

9. The method according to claim 8, characterized in that, The step of responding to receiving dialogue information for the target digital twin system through a dialogue processing model, and generating an operation plan based on the dialogue information through the dialogue processing model to obtain an operation plan for the target digital twin system, includes: In response to receiving dialogue information for the target digital twin system through a dialogue processing model, the dialogue processing model generates an operation plan based on the dialogue information and node topology relationships to obtain the operation plan for the target digital twin system; the node topology relationships indicate the logical connection relationships and / or physical connection relationships between the multiple nodes; the node topology relationships are obtained by the twin system analysis model through topology relationship analysis based on the system description data.

10. The method according to claim 8, characterized in that, The operational plan also includes verification of the operational plan; The determination of the system analysis results of the target digital twin system based on the twin system's operation logs and the real-time operating status of the target digital twin system includes: After the target digital twin system finishes running according to the execution operation plan, the target digital twin system is verified based on the verification operation plan to obtain system verification data of the target digital twin system. Based on the system verification data, the twin system operation log, and the real-time operation status of the target digital twin system, the system analysis results of the target digital twin system are determined.

11. The method according to claim 8, characterized in that, After determining the system analysis results of the target digital twin system based on the twin system's operation logs and the real-time operating status of the target digital twin system, the method further includes: If the system analysis results indicate that the target digital twin system exhibits abnormal behavior, the dialogue processing model determines a natural language analysis report of the target digital twin system based on the system analysis results; the natural language analysis report includes at least the abnormal behavior and the impact of the abnormal behavior. Display the natural language analysis report.

12. The method according to claim 1, characterized in that, The node behavior model includes node configuration information indicating the node configuration of the corresponding node and node behavior information constraining the node behavior of the corresponding node. The node configuration information indicates the forwarding logic and control logic between nodes, and the node behavior information indicates the node tasks performed by the node and the node functions implemented. The step of configuring virtual nodes corresponding to each of the nodes based on their respective node behavior models to obtain the target digital twin system corresponding to the target system includes: Based on the node behavior information, the node tasks and node functions of the virtual nodes corresponding to each node are configured, and the forwarding logic and control logic between the virtual nodes corresponding to each node are configured based on the node configuration information, so as to obtain the target digital twin system corresponding to the target system.

13. A digital twin system construction apparatus, characterized in that, The device includes: The acquisition module is used to acquire system description data corresponding to the target system; the target system includes multiple nodes that are interconnected. The analysis module is used to analyze and process node behavior based on the system description data using a twin system analysis model to obtain the node behavior model of each of the multiple nodes; the node behavior model is used to indicate the node configuration and node behavior of the corresponding node. An instantiation module is used to instantiate the virtual node corresponding to each of the nodes in the virtual digital system; The configuration module is used to configure the virtual node corresponding to each node based on the node behavior model of each node, so as to obtain the target digital twin system corresponding to the target system.

14. An electronic device, characterized in that, include: processor; A memory storing computer-readable instructions that, when executed by the processor, implement the method as described in any one of claims 1-12.

15. A computer-readable storage medium, characterized in that, It stores computer-readable instructions that, when executed by a processor, implement the method as described in any one of claims 1-12.

16. A computer program product, characterized in that, It includes computer-readable instructions that, when executed by a processor, implement the method of any one of claims 1-12.

Images