Network slice management method based on large language model and related device

Abstract

The invention discloses a network slicing management method based on a large language model and a related device, and relates to the technical field of network slicing, and the method comprises the steps: calling a global agent to receive a network slicing demand in a natural language format, and obtaining a demand intention; calling a global agent to interact with the large language model based on the demand intention to obtain a slicing request; calling a global agent to obtain a resource state in a current network; calling a large language model to determine a global resource allocation strategy based on the slicing request and a resource state in a current network, obtaining a resource allocation strategy of each sub-region, and obtaining a region slicing request of each sub-region; calling a local agent to interact with the large language model based on the region slicing request of each sub-region to obtain an operation instruction of network slicing operation performed by each sub-region; and executing the operation instruction to perform network slicing operation on each sub-region. The purpose of improving the management efficiency of the network slices is achieved based on the large language model.

Description

Technical Field

[0001] This application relates to the field of network slicing technology, and in particular to a network slice management method and related apparatus based on a large language model. Background Technology

[0002] Network slicing technology virtualizes physical network resources into multiple independent logical networks, thereby providing customized network services for different business needs. It can provide differentiated network service guarantees for diverse business scenarios such as the Industrial Internet, the Internet of Vehicles, and smart cities. However, due to the rapid development of networks, existing network slicing management technologies are mainly centralized. This centralized management solution mainly manages network slices based on preset rules, lacking intelligent understanding of complex scenarios. Furthermore, all resource allocation and adjustment rely on centralized management nodes for processing, resulting in a heavy operational burden on these nodes. Summary of the Invention

[0003] In view of the above problems, this application provides a network slice management method and related apparatus based on a large language model to improve the management efficiency of network slices. The specific solution is as follows:

[0004] The first aspect of this application provides a network slice management method based on a large language model, including:

[0005] The global intelligent agent is invoked to receive network slice requests in natural language format sent by the user, and the intent of the network slice requests is analyzed to obtain the request intent.

[0006] The global intelligent agent is invoked to interact with the large language model based on the demand intent, and a slice request containing multiple switching information is obtained;

[0007] Call the global agent to obtain the current resource status of the network;

[0008] The large language model is invoked to determine a global resource allocation strategy based on slice requests and the current resource status in the network.

[0009] Each sub-region agent is invoked to interact with the large language model based on the global resource allocation strategy to obtain the resource allocation strategy for each sub-region. Each sub-region agent is invoked to interact with the large language model based on the resource allocation strategy for each sub-region to obtain the region slice request for each sub-region. The region slice request contains various switching information.

[0010] The local agent interacts with the large language model based on the region slicing request for each sub-region to obtain the operation instructions for the network slicing operation to be performed separately for each sub-region.

[0011] Execute operation instructions to perform network slicing operations on each sub-region separately.

[0012] Optionally, the global agent is invoked to receive the network slice request in natural language format sent by the user, and the intent of the network slice request is analyzed to obtain the request intent, including:

[0013] Based on a global intelligent agent, semantic parsing is performed on network slicing requirements in natural language format to obtain semantic information in the network slicing requirements;

[0014] Generate demand intents to represent network slicing requirements based on semantic information.

[0015] Optionally, the global agent is invoked to interact with the large language model based on the demand intent, obtaining a slice request containing various switching information, including:

[0016] The global agent is invoked to construct interactive information for interacting with the large language model based on the demand intent;

[0017] The interactive information is input into the large language model to obtain a slice request containing multiple switching information.

[0018] Optionally, the global agent can be invoked to obtain the current resource status in the network, including:

[0019] The global agent is invoked to send a resource status acquisition command to the network management node to obtain the current usage of network computing resources, storage resources, and network bandwidth resources.

[0020] Based on usage, the resource status in the current network is obtained, reflecting the overall resource occupancy of the current network.

[0021] Optionally, the local agent interacts with the large language model based on the region slicing request for each sub-region to obtain the operation instructions for the network slicing operation to be performed separately for each sub-region, including:

[0022] The local agent is invoked to receive the region slice request for each sub-region and parse the resource allocation information and switching information contained in the region slice request for each sub-region.

[0023] The local intelligent agent is invoked to construct interactive content for interaction with the large language model based on resource allocation and switching information;

[0024] The local agent is invoked to send the interactive content to the large language model, which then obtains the operation instructions for network slicing operations to be performed on each sub-region separately.

[0025] Optionally, execute operation instructions to perform network slicing operations on each sub-region separately, including:

[0026] The operation instructions are sent to the network management nodes in the corresponding sub-regions;

[0027] The network management node performs network slicing operations on network resources within a sub-region based on operation commands.

[0028] Optionally, the method further includes:

[0029] During the network slicing operation, monitor the configuration status and execution results of network resources within the sub-region;

[0030] If an execution anomaly or resource conflict is detected, the operation instructions will be adjusted or the network slicing operation will be retried.

[0031] A second aspect of this application provides a computer program product including computer-readable instructions that, when executed on an electronic device, cause the electronic device to implement the network slice management method based on a large language model as described in the first aspect or any implementation thereof.

[0032] A third aspect of this application provides an electronic device, comprising at least one processor and a memory connected to the processor, wherein:

[0033] The memory is used to store computer programs;

[0034] The processor is used to execute the computer program so that the electronic device can implement the network slice management method based on a large language model as described in the first aspect or any implementation thereof.

[0035] A fourth aspect of this application provides a computer storage medium carrying one or more computer programs, which, when executed by an electronic device, enable the electronic device to implement the network slice management method based on a large language model as described in the first aspect or any implementation thereof.

[0036] By employing the above technical solution, compared to existing technical solutions that mainly rely on central control nodes and preset rules for resource allocation and are difficult to adapt to complex and ever-changing business needs, resulting in low efficiency in network slice management, this application receives network slice requests in natural language format sent by users and uses a global intelligent agent to parse the network slice requests to obtain the request intent, thereby generating slice requests containing various switching information. Based on this, a global resource allocation strategy is determined in combination with the current resource status in the network, and further, a resource allocation strategy for each sub-region is obtained based on the global resource allocation strategy. This enables collaborative management and optimized configuration of network resources in different regions, thereby reducing the operational burden of centralized management nodes and improving the management efficiency of network slices. Attached Figure Description

[0037] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0038] Figure 1 A flowchart illustrating a network slice management method based on a large language model provided in this application;

[0039] Figure 2 A flowchart illustrating another network slice management method based on a large language model provided in this application;

[0040] Figure 3 A schematic block diagram of an electronic device provided in this application. Detailed Implementation

[0041] The embodiments of this application are described below with reference to the accompanying drawings. The terminology used in the implementation section of this application is for explaining specific embodiments only and is not intended to limit the scope of this application.

[0042] The embodiments of this application will now be described with reference to the accompanying drawings. Those skilled in the art will recognize that, with technological advancements and the emergence of new scenarios, the technical solutions provided in the embodiments of this application are equally applicable to similar technical problems.

[0043] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such terms are interchangeable where appropriate; this is merely a way of distinguishing objects with the same attributes in the embodiments of this application. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, so that a process, method, system, product, or apparatus that comprises a series of elements is not necessarily limited to those elements, but may include other elements not explicitly listed or inherent to those processes, methods, products, or apparatuses.

[0044] This application provides a network slice management method based on a large language model, such as Figure 1 The method may include the following steps:

[0045] S101: Call the global intelligent agent to receive the network slice request in natural language format sent by the user, and perform intent analysis on the network slice request to obtain the request intent;

[0046] Optionally, in this embodiment, users can submit network slicing requests in natural language via terminal devices, such as "provide low-latency, high-reliability dedicated network slices for smart factory services" or "allocate high-bandwidth slice resources for high-definition video live streaming services." After receiving the network slicing request in natural language format, the global agent performs semantic understanding and parsing to identify key information related to the network slice, including but not limited to service type, quality of service requirements, latency constraints, bandwidth requirements, reliability level, and applicable area. After completing the semantic parsing, the global agent summarizes and abstracts the parsed semantic information, converting the user's natural language description into a structured demand expression, and generating a demand intent to represent the network slicing request. The demand intent accurately reflects the user's core needs for network slicing, providing a unified and clear semantic foundation for subsequent slice request generation, resource status analysis, and global resource allocation strategy formulation.

[0047] S102: Invoke the global intelligent agent to interact with the large language model based on the demand intent and obtain a slice request containing multiple switching information;

[0048] Optionally, in this embodiment, after obtaining the demand intent, the global agent constructs interaction information for interacting with the large language model based on the demand intent. The interaction information may include key elements represented by the demand intent, such as the service type, service level requirements, latency and bandwidth constraints, reliability indicators, and the network coverage that may be involved, and is organized in a preset prompt structure before being sent to the large language model.

[0049] The large language model infers and analyzes network slicing requirements based on interaction information, generating slice requests that match the intended needs. These slice requests contain various switching information, including slice type switching, resource allocation adjustment, service level parameter switching, and dynamic switching strategies under different network loads or service states. This process transforms the intended needs into structured slice requests that can be used for subsequent resource scheduling and slice execution. By introducing the large language model to infer and generate slice requests from the intended needs, a high-quality slice description containing various switching information can be automatically generated, enhancing the expressiveness and flexibility of network slice requests. This facilitates dynamic slice management in complex business scenarios and improves the intelligence level of network resource allocation.

[0050] S103: Call the global agent to obtain the current network resource status;

[0051] In this embodiment, after generating a slice request, the global agent sends a resource status acquisition instruction to the network management node to obtain the overall resource usage of the current network. The resource status can include the usage of computing resources, storage resource occupancy, and network bandwidth availability. The computing resource usage characterizes the computing load of each network node, the storage resource occupancy characterizes the data caching and storage capacity available for network slices, and the network bandwidth availability characterizes link bandwidth and its utilization rate. The global agent summarizes and analyzes the resource usage information returned from multiple network management nodes to generate a resource status description that reflects the overall resource occupancy level of the current network. This resource status description serves as an important input for subsequently determining the global resource allocation strategy. This application can grasp the real-time usage of network resources, providing a reliable basis for subsequent slice resource allocation, thereby avoiding resource conflicts and overload risks, and improving the stability and resource utilization efficiency of network slice management.

[0052] S104: Invoke the large language model to determine the global resource allocation strategy based on the slice request and the current resource status in the network;

[0053] Optionally, in this embodiment, the global agent can send a slice request containing various switching information and the acquired current network resource status as input to the large language model. The current network resource status reflects the availability and load of various network resources. Based on the slice request and the current network resource status, the large language model performs comprehensive reasoning and analysis on network resources to determine a global resource allocation scheme that meets business requirements and resource constraints, generating a global resource allocation strategy. The global resource allocation strategy characterizes the allocation ratio, priority adjustment method, and resource switching strategy of computing resources, storage resources, and network bandwidth resources globally, providing a unified guideline for subsequent resource allocation in each sub-region. By introducing the large language model to jointly analyze slice requests and network resource status, intelligent trade-offs between complex business requirements and dynamic network resource conditions can be achieved, thereby generating a more reasonable and flexible global resource allocation strategy, improving the adaptive capability of network slice management and the overall resource configuration efficiency.

[0054] S105: Call each sub-region agent to interact with the large language model based on the global resource allocation strategy to obtain the resource allocation strategy of each sub-region. Call each sub-region agent to interact with the large language model based on the resource allocation strategy of each sub-region to obtain the region slice request of each sub-region. The region slice request contains a variety of switching information.

[0055] Optionally, in this embodiment, after the global resource allocation strategy is distributed to each sub-region agent, each sub-region agent can combine the network topology, local resource status, and service load of its own sub-region to construct input information for interaction with the large language model, using the global resource allocation strategy as a constraint. Subsequently, each sub-region agent calls the large language model to perform a region-level refined analysis of the global resource allocation strategy, obtaining a resource allocation strategy that matches its own sub-region. After obtaining the resource allocation strategy for the sub-region, each sub-region agent can further interact with the large language model again based on the resource allocation strategy to generate a corresponding regional slice request. The regional slice request is used to characterize the network slice operation to be performed in the sub-region. The switching information may include resource adjustment methods, changes in the number of slice instances, bandwidth scaling information, and slice start / stop or reconfiguration instructions, thereby providing a clear basis for the subsequent execution of specific network slice operations by the local agent. By introducing a collaborative decision-making mechanism between sub-regional agents and large language models on the basis of the global resource allocation strategy, regional-level fine-grained adjustments to the resource allocation strategy can be achieved, making the network slicing scheme more closely aligned with the actual operating status of each sub-region, thereby improving the flexibility of network slice management and the overall resource utilization efficiency.

[0056] S106: The local agent interacts with the large language model based on the region slicing request for each sub-region to obtain the operation instructions for the network slicing operation to be performed on each sub-region separately.

[0057] Optionally, in this embodiment, the local agent receives region slice requests from the corresponding sub-region agents and parses the region slice requests. Based on this, the local agent combines the parsed region slice requests with the actual configuration capabilities of the network devices in the current sub-region to construct interactive content for interacting with the large language model.

[0058] Subsequently, the local agent invokes the large language model to perform reasoning analysis on the interaction content, generating network slicing operation instructions adapted to the current sub-region network environment. These instructions instruct the execution of specific network slicing operations within the corresponding sub-region. These instructions can include slice creation, deletion, parameter adjustment, and resource reallocation commands, ensuring that network slicing operations in each sub-region are executed in an orderly manner according to the expected strategy. By introducing a collaborative decision-making mechanism between the local agent and the large language model, regional-level slice requests can be transformed into directly executable network slicing operation instructions, improving the automation and accuracy of network slice configuration, reducing manual intervention costs, and enhancing the real-time performance and reliability of network slice management.

[0059] S107: Execute the operation instructions to perform network slicing operations on each sub-region separately;

[0060] Optionally, in this embodiment, the generated network slicing operation instructions can be sent to the network management node or network control unit within the corresponding sub-region. The network management node, based on the received operation instructions, configures and adjusts the computing resources, storage resources, and network bandwidth resources within the sub-region to complete the corresponding network slicing operation.

[0061] Specifically, network slicing operations include, but are not limited to, creating new network slices for sub-regions, adjusting parameters of existing network slices, and releasing or reconstructing network slices that no longer meet requirements. During execution, network management nodes dynamically schedule network resources within sub-regions according to the resource allocation ratios, quality of service parameters, and switching methods specified in the operation instructions, thereby ensuring that the network slice status of each sub-region remains consistent with the regional slicing request. By directly mapping network slicing operation instructions to sub-region-level network resource configuration behaviors, rapid implementation and precise execution of network slicing strategies can be achieved, enhancing the collaborative management capabilities of multi-region network resources and ensuring the stability and consistency of network slice service quality.

[0062] In one embodiment, semantic parsing of network slicing requirements in natural language format is performed based on a global intelligent agent to obtain semantic information in the network slicing requirements.

[0063] Generate demand intents to represent network slicing requirements based on semantic information.

[0064] Specifically, firstly, the global agent can perform semantic parsing on the network slicing requests sent by users in natural language format. Semantic parsing involves word segmentation, syntactic analysis, and semantic understanding of the text content in the network slicing requests, thereby identifying key semantic information contained within them. This semantic information includes at least the service type, quality of service requirements, latency requirements, bandwidth requirements, reliability level, applicable area, and slice activation time.

[0065] Subsequently, based on the parsed semantic information, the global agent abstracts and summarizes the network slicing requirements, generating requirement intents to represent these requirements. These requirement intents, in a structured or semi-structured form, uniformly describe the user's network slicing needs, clarifying the target type, resource requirement characteristics, and service constraints of the network slice. This provides a standardized input basis for subsequent interaction with the large language model and the formulation of global resource allocation strategies. Through this method, the network slicing requirements in natural language form are transformed into explicit and processable requirement intents, laying the foundation for the automated and intelligent execution of subsequent network slice management processes.

[0066] In one embodiment, a global agent is invoked to construct interactive information for interacting with a large language model based on demand intent;

[0067] The interactive information is input into the large language model to obtain a slice request containing multiple switching information.

[0068] Specifically, a global intelligent agent can be invoked to interact with a large language model based on the demand intent to obtain a slice request containing various switching information, including the following process.

[0069] First, the global agent can construct interaction information based on the generated demand intent to interact with the large language model. This interaction information is used to standardize the expression of the demand intent and organize key constraints related to network slicing in the form of natural language or structured prompts. The interaction information includes at least the target service type, service quality level requirements, bandwidth and latency constraints, reliability level, coverage area, and desired slice adjustment method, to guide the large language model in targeted reasoning and generation.

[0070] Subsequently, the global agent can input the interactive information into the large language model, enabling the large language model to comprehensively analyze network slicing requirements based on the intended intent. Leveraging its semantic understanding and reasoning capabilities, the large language model outputs slice requests containing various switching information. These slice requests describe specific adjustments to the network slice, and the switching information can include slice type switching information, resource allocation adjustment information, service level parameter switching information, and dynamic switching strategy information under different network loads or service states. Through this method, abstract intended intents are transformed into executable slice requests, providing a clear basis for subsequent network resource allocation and slicing operations, thereby improving the intelligence level of the network slice management process.

[0071] In one embodiment, a global intelligent agent can be invoked to send a resource status acquisition command to the network management node to obtain the current usage of network computing resources, storage resources, and network bandwidth resources.

[0072] Based on usage, the resource status in the current network is obtained, reflecting the overall resource occupancy of the current network.

[0073] Specifically, first, the global agent sends a resource status retrieval command to the network management node. The network management node can be a core network control node, a network orchestration and management node, or a centralized resource management node, used to uniformly maintain the operational status information of network resources. The resource status retrieval command is used to request the real-time usage status of various resources in the current network.

[0074] Subsequently, the network management node queries the current resource usage in the network based on the resource status acquisition command and feeds back the query results to the global agent. Resource usage can include computing resource usage, storage resource usage, and network bandwidth resource usage. Specifically, computing resource usage characterizes the load level of processing units, storage resource usage characterizes available storage capacity and its utilization rate, and network bandwidth resource usage characterizes the degree of link bandwidth utilization and remaining bandwidth capacity.

[0075] Finally, the global agent performs a comprehensive analysis based on the usage of computing resources, storage resources, and network bandwidth resources to generate resource status information reflecting the current overall resource occupancy of the network. The current resource status of the network characterizes its capacity to handle requests for new or adjusted network slices at the current moment and provides a basis for determining subsequent global resource allocation strategies.

[0076] In one embodiment, the local agent is invoked to receive the region slice request for each sub-region and parse the resource allocation information and switching information contained in the region slice request for each sub-region.

[0077] The local intelligent agent is invoked to construct interactive content for interaction with the large language model based on resource allocation and switching information;

[0078] The local agent is invoked to send the interactive content to the large language model, which then obtains the operation instructions for network slicing operations to be performed on each sub-region separately.

[0079] Specifically, firstly, the local agent can receive region slice requests from the corresponding sub-region. These requests can be generated by the sub-region agent and contain resource allocation and switching information for that sub-region. The resource allocation information characterizes the sub-region's allocation needs in terms of computing, storage, and network bandwidth resources, while the switching information characterizes operational needs such as creating, adjusting, or releasing network slices. The local agent then parses the region slice request, extracting the resource allocation parameters and switching strategy information.

[0080] Subsequently, based on the parsed resource allocation and switching information, the local agent constructs interaction content for interacting with the large language model. This interaction content describes the current network slicing demand status, resource constraints, and desired slicing operation objectives of the sub-region, guiding the large language model to generate operational decisions that conform to the actual network environment of the sub-region. The local agent then sends the interaction content to the large language model and invokes it for inference analysis. Based on the interaction content, the large language model generates network slicing operation instructions for that sub-region, specifying the detailed network slicing configuration parameters and execution methods.

[0081] Finally, the local agent receives the operation instructions returned by the large language model. These instructions serve as the basis for subsequent network slicing operations in the corresponding sub-regions, thereby enabling fine-grained control and flexible scheduling of network slicing operations in each sub-region.

[0082] In one embodiment, the operation instruction is sent to the network management node in the corresponding sub-region;

[0083] The network management node performs network slicing operations on network resources within a sub-region based on operation commands.

[0084] Specifically, firstly, operation instructions generated and confirmed by the local agent can be sent to the network management nodes within the corresponding sub-region. These instructions can include information such as network slicing operation parameters, resource allocation strategies, and slice switching methods for that sub-region, guiding the network management nodes to perform specific network configuration operations.

[0085] Subsequently, after receiving the operation instructions, the network management node parses them and schedules and configures the network resources within the sub-region based on the parsing results. Network resources include computing resources, storage resources, and network bandwidth resources. The network management node creates, adjusts, or releases network slices according to the operation instructions, thereby forming network slices within the sub-region that meet business requirements. Through this method, network slice operations within each sub-region can be executed in an orderly manner according to a unified scheduling strategy, achieving fine-grained management and dynamic adjustment of network resources in the sub-region, and ensuring the consistency and real-time nature of network slice configuration with business needs.

[0086] In one embodiment, during the network slicing operation, the configuration status and execution results of network resources within the sub-region are monitored;

[0087] If an execution anomaly or resource conflict is detected, the operation instructions will be adjusted or the network slicing operation will be retried.

[0088] Specifically, during the network slicing operation, the configuration status and execution results of network resources within the sub-region are continuously monitored. Specifically, after the network slicing operation command is issued and executed, the actual configuration of computing resources, storage resources, and network bandwidth resources within the sub-region is collected in real time, along with execution feedback information during the creation, adjustment, or switching of network slices, to determine whether the current network slicing operation has been completed as expected.

[0089] When monitoring results indicate a discrepancy between the network resource configuration status within a sub-region and the operational instructions, or when execution anomalies, resource contention, resource over-provisioning, or resource conflicts are detected during network slicing operations, an anomaly handling mechanism is triggered. In this case, the original operational instructions are dynamically adjusted, such as reallocating resource quotas, adjusting slice parameters, or modifying switching strategies; or the network slicing operation process is retried to ensure that network slicing can meet business needs while maintaining the stability and security of network resource usage. Through the above monitoring and adjustment mechanism, closed-loop control of the entire network slicing operation process is achieved, improving the reliability and fault tolerance of network slicing management and avoiding business interruptions or resource waste caused by anomalies.

[0090] like Figure 2 In another embodiment, the user can also send network slicing requests in natural language format to the input layer. The input layer then feeds these requests to the global LLM agent, which parses the network slicing requests to obtain prompts that the large language model can accept. These prompts are then fed to the large language model (LLM service), which returns the parsing results to the global agent. Based on the parsing results, the global agent generates structured slicing requests and obtains the overall network resource status (i.e., the current network resource status). This current network resource status is then fed into the large language model for analysis to obtain a global resource allocation strategy (i.e., the global strategy). Next, the regional LLM agent (i.e., the sub-region agent) optimizes the region based on the global resource allocation strategy, generating the prompts required by the large language model (i.e., the regional resource allocation scheme). These prompts are then fed into the large language model to obtain the regional slicing requests (i.e., the regional optimization schemes and slicing requirements) for each sub-region. Finally, the local LLM agent is invoked, which obtains the network slicing operation instructions based on the region slicing request of each sub-region, performs network slicing operation on each sub-region based on the operation instructions, and inputs the execution results to the result aggregator, which returns the execution results to the user.

[0091] This application also provides an electronic device in its embodiments. (See reference...) Figure 3 The diagram illustrates a structural schematic of an electronic device suitable for implementing the network slicing management method based on a large language model as described in this application. The electronic device in this application may include, but is not limited to, fixed terminals such as mobile phones, laptops, PDAs (personal digital assistants), PADs (tablet computers), desktop computers, etc. Figure 3 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0092] like Figure 3As shown, the electronic device may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. When the electronic device is powered on, the RAM 603 also stores various programs and data required for the operation of the electronic device. The processing unit 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0093] Typically, the following devices can be connected to I / O interface 605: input devices 606 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 607 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 608 including, for example, memory cards, hard drives, etc.; and communication devices 609. Communication device 609 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown. More or fewer devices may be implemented or have alternatively.

[0094] This application also provides a computer program product including computer-readable instructions, which, when executed on an electronic device, cause the electronic device to implement any of the network slice management methods based on a large language model provided in this application.

[0095] This application also provides a computer-readable storage medium that carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement any of the network slice management methods based on a large language model provided in this application.

[0096] It should also be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. In addition, in the device embodiment drawings provided in this application, the connection relationship between modules indicates that they have a communication connection, which can be implemented as one or more communication buses or signal lines.

[0097] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware, or it can be implemented by special-purpose hardware including application-specific integrated circuits, special-purpose CPUs, special-purpose memory, special-purpose components, etc. Generally, any function performed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be diverse, such as analog circuits, digital circuits, or special-purpose circuits. However, for this application, software program implementation is more often the preferred implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium, such as a computer floppy disk, USB flash drive, mobile hard disk, ROM, RAM, magnetic disk, or optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, training equipment, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0098] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product.

[0099] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, training device, or data center to another website, computer, training device, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).

Claims

1. A network slice management method based on a large language model, characterized in that, include: The global intelligent agent is invoked to receive network slice requests in natural language format sent by the user, and the intent of the network slice requests is analyzed to obtain the request intent; The global intelligent agent is invoked to interact with the large language model based on the demand intent, and a slice request containing multiple switching information is obtained; Invoke the global agent to obtain the current resource status of the network; The large language model is invoked to determine a global resource allocation strategy based on the slice request and the current resource status in the network. Each sub-region agent is invoked to interact with the large language model based on the global resource allocation strategy to obtain the resource allocation strategy for each sub-region. Each sub-region agent is invoked to interact with the large language model based on the resource allocation strategy for each sub-region to obtain the region slice request for each sub-region. The region slice request contains various switching information. The local intelligent agent interacts with the large language model based on the region slicing request of each sub-region to obtain the operation instructions for the network slicing operation to be performed on each sub-region separately; The operation instructions are executed to perform network slicing operations on each of the sub-regions respectively.

2. The network slice management method based on a large language model according to claim 1, characterized in that, The process involves invoking the global intelligent agent to receive network slice requests in natural language format sent by the user, and performing intent analysis on the network slice requests to obtain the request intent, including: Based on a global intelligent agent, semantic parsing is performed on the network slicing requirements in the natural language format to obtain semantic information in the network slicing requirements. Based on the semantic information, a demand intent is generated to characterize the network slice requirement.

3. The network slice management method based on a large language model according to claim 1, characterized in that, The invocation of the global agent interacts with the large language model based on the demand intent to obtain a slice request containing various switching information, including: The global intelligent agent is invoked to construct interactive information for interacting with the large language model based on the stated intent. The interactive information is input into the large language model to obtain a slice request containing multiple switching information.

4. The network slice management method based on a large language model according to claim 1, characterized in that, The step of invoking the global agent to obtain the current resource status in the network includes: The global intelligent agent is invoked to send a resource status acquisition command to the network management node to obtain the current usage of network computing resources, storage resources, and network bandwidth resources; Based on the usage, the resource status in the current network, reflecting the overall resource occupancy of the current network, is obtained.

5. The network slice management method based on a large language model according to claim 1, characterized in that, The process involves the local agent interacting with the large language model based on the region slicing request for each sub-region to obtain operation instructions for network slicing operations to be performed on each sub-region, including: The local agent is invoked to receive the region slice request for each sub-region and to parse the resource allocation information and switching information contained in the region slice request for each sub-region. The local intelligent agent is invoked to construct interactive content for interaction with the large language model based on the resource allocation information and switching information; The local agent is invoked to send the interactive content to the large language model, thereby obtaining operation instructions for network slicing operations to be performed on each sub-region separately.

6. The network slice management method based on a large language model according to claim 1, characterized in that, The execution of the operation instructions to perform network slicing operations on each of the sub-regions includes: The operation command is sent to the network management node in the corresponding sub-region; The network management node performs network slicing operations on network resources within the sub-region based on the operation instructions.

7. The network slice management method based on a large language model according to claim 6, characterized in that, Also includes: During the network slicing operation, monitor the configuration status and execution results of network resources within the sub-region; If an execution anomaly or resource conflict is detected, the operation instruction will be adjusted or the network slicing operation will be retried.

8. A computer program product, characterized in that, It includes computer-readable instructions that, when executed on an electronic device, cause the electronic device to implement the network slice management method based on a large language model as described in any one of claims 1 to 7.

9. An electronic device, characterized in that, It includes at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program to enable the electronic device to implement the network slice management method based on a large language model as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that, The storage medium carries one or more computer programs, which, when executed by an electronic device, enable the electronic device to implement the network slice management method based on a large language model as described in any one of claims 1 to 7.

Images