Optimized communication network health degree prediction and collaboration method

Abstract

The invention provides an optimized communication network health degree prediction and collaboration method, and the method comprises the steps: building a high-precision virtual simulation model of a target communication network in a complex environment, calibrating the simulation model according to real-time operation data, and generating virtual-real fusion data; converting the fusion data into space-time diagram structure data based on a communication node and link connection relationship of the target communication network, and constructing a communication network performance prediction model based on a space-time diagram neural network; based on the constructed communication network performance prediction model, carrying out fault mode and influence analysis on the communication network in a complex environment, and predicting a health indication factor of the communication network in a typical working condition; calculating the health state of the target communication network according to the multi-dimensional health indication factor of the target communication network; and establishing a health state optimization model in a complex environment, and carrying out collaborative optimization on the health state of the target communication network from multiple dimensions.

Description

Technical Field

[0001] This application relates to the field of communication computing, and more specifically, to an optimized method for predicting and coordinating the health of communication networks. Background Technology

[0002] In the construction and operation of communication networks, network health status assessment and optimization are crucial for ensuring service quality. However, in special geographical environments (such as mountains, canyons, and tunnels), communication networks face unique physical challenges: multipath effects and signal shielding complicate wireless propagation characteristics, extreme temperature and humidity conditions accelerate equipment aging, and geographical isolation makes infrastructure maintenance difficult. Existing assessment systems typically use standardized indicator templates, failing to fully consider the differentiated impact of terrain on signal propagation. Furthermore, they lack response mechanisms for real-time environmental changes. Summary of the Invention

[0003] The purpose of this application is to provide an optimized method for predicting and coordinating the health of communication networks, so as to improve the accuracy of network health assessment in complex environments.

[0004] In a first aspect, the present invention provides a method for optimizing communication network health prediction and coordination, the method comprising: A high-precision virtual simulation model of the target communication network in a complex environment is built, and the simulation model is calibrated based on real-time running data to generate virtual-real fusion data; Based on the communication nodes and link connections of the target communication network, the fused data is converted into spatiotemporal graph structure data, and a communication network performance prediction model based on spatiotemporal graph neural network is constructed. Based on the constructed communication network performance prediction model, we conduct fault mode and impact analysis on communication networks under complex environments and predict the health indicator factors of communication networks under typical operating conditions. The health status of the target communication network is calculated based on the multi-dimensional health indicator factors of the target communication network. Establish a health status optimization model under complex environments and conduct collaborative optimization of the health status of the target communication network from multiple dimensions.

[0005] In an optional implementation, the communication network health indicator prediction model includes a graph convolutional layer, a fusion layer, and an attention mechanism layer, wherein, The spatiotemporal graph structures corresponding to multiple time points are input into the graph convolutional layer for spatial feature extraction, so as to output multiple spatial feature vectors. Multiple spatial feature vectors are input into a fusion layer for feature fusion to output a fused spatial feature vector; The fused spatial feature vector is input into the attention mechanism layer for temporal feature extraction, and the spatiotemporal feature vector is output to output the health indicator factor of the target communication network.

[0006] In an optional implementation, for each round of predictions by the communication network health indicator prediction model, a loss value of the communication network prediction model is calculated to adjust the hyperparameters of the communication network prediction model.

[0007] In an optional implementation, the loss value of the communication network prediction model is calculated in the following manner. : ; in, The current hyperparameters of the communication network prediction model. The loss value is calculated based on the current hyperparameters and the health indicator parameters predicted in this round. For the diagonal elements of the information matrix, The regularization coefficient is . The optimal hyperparameters for the historical prediction model of the communication network are... This represents the number of hyperparameters.

[0008] In an optional implementation, the scoring interval includes a first scoring interval [0, ...]. ], second scoring interval ( , ), third scoring interval[ ,1],wherein, < < The health status of the target communication network is determined in the following ways: When the health of the target communication network is in the first scoring range, the health status of the target communication network is determined to be faulty or high-risk. When the health of the target communication network is in the second scoring range, the health status of the target communication network is determined to be sub-healthy. When the health of the target communication network is within the first scoring range, the health status of the target communication network is determined to be healthy.

[0009] In an optional implementation, when the target communication network is in a sub-healthy state, the method further includes: Trigger a sub-health state warning and implement adjustments to optimize resource allocation for the target communication network; When the health status of the target communication network is faulty or high-risk, it also includes: Trigger a network anomaly alert and switch to another communication network to perform communication services.

[0010] In an optional implementation, the resource allocation vector is calculated in the following manner to perform adjustments to optimize resource allocation on the target communication network: ; ; ; in, For communication nodes The resource allocation vector, For communication nodes The health of the nodes, For link utilization, For business The actual delay For business Demand latency, , These are the weighting coefficients.

[0011] Secondly, the present invention provides an optimized communication network health prediction and coordination device, the device comprising: The acquisition module is used to acquire high-precision simulation data and real-time operation data of the target communication network in complex environments, and generate fused data; The data analysis module, based on virtual and real fusion data, performs fault mode and impact analysis on communication networks in complex environments, and determines the health indicator factors of communication networks in complex environments.

[0012] The module is used to convert fused data into spatiotemporal graph structure data based on the communication nodes and link connections of the target communication network, and to build a communication network performance prediction model based on spatiotemporal graph neural network. The prediction module is based on a spatiotemporal graph neural network-based communication network performance prediction model. It performs fault mode analysis and its impact analysis on communication networks under complex environments and predicts the health indicator factors of communication networks under typical operating conditions.

[0013] The calculation module is used to calculate the health of the target communication network based on the multi-dimensional health indicator factors of the target communication network. The evaluation module is used to determine the health status of the target communication network based on the relationship between the health of the target communication network and a preset scoring range.

[0014] The optimization module is used for collaborative optimization of the health status of the target communication network from multiple dimensions.

[0015] Thirdly, the present invention provides an electronic device, including: a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus, and the processor executes the machine-readable instructions to perform the steps of any of the optimized communication network health prediction and coordination methods described in the foregoing embodiments.

[0016] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of any of the methods for optimizing communication network health prediction and coordination as described in the foregoing embodiments.

[0017] This application provides a method for optimizing the prediction and coordination of communication network health. The method involves: constructing a high-precision virtual simulation model of the target communication network under complex environments; calibrating the simulation model based on real-time operational data to generate fused virtual-real data; converting the fused data into spatiotemporal graph structure data based on the communication nodes and link connections of the target communication network; constructing a communication network performance prediction model based on a spatiotemporal graph neural network; performing fault mode and impact analysis on the communication network under complex environments based on the constructed communication network performance prediction model; predicting the health indicator factors of the communication network under typical operating conditions; calculating the health status of the target communication network based on its multi-dimensional health indicator factors; and establishing a health status optimization model under complex environments to collaboratively optimize the health status of the target communication network from multiple dimensions. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating an optimized communication network health prediction and coordination method provided in this application embodiment; Figure 2 A link diagram of a target communication network provided in this application; Figure 3 This is a schematic diagram of the structure of a communication network performance prediction model provided in this application; Figure 4 A schematic diagram of the structure of an optimized communication network health prediction and coordination device provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0020] The technical solutions in the embodiments of this application will now be described with reference to the accompanying drawings.

[0021] Figure 1 This is a flowchart illustrating an optimized communication network health prediction and coordination method provided in an embodiment of this application. Figure 1As shown, in one embodiment of this application, an optimized communication network health prediction and coordination method is provided, the method comprising: S1. Build a high-precision virtual simulation model of the target communication network in a complex environment, and calibrate the simulation model based on real-time running data to generate virtual-real fusion data.

[0022] Figure 2 This application provides a link diagram of a target communication network. For example... Figure 2 As shown, the target communication network can consist of multiple communication nodes, such as terminal devices (monitors, servers, etc.), switches, routers, base stations, core networks, DTUs, MECs, etc. These communication nodes can be distributed in a three-dimensional, non-uniform manner in complex environments such as steep mountains and canyons.

[0023] Various types of sensors are deployed at each communication node of the communication network, including signal strength sensors, network latency sensors, packet loss rate sensors, equipment temperature sensors, humidity sensors, and geographic environment sensors (such as satellite positioning devices and weather stations). The real-time operational data collected by these sensors can be transmitted to the data processing center via communication links.

[0024] Simultaneously, target communication networks can be modeled to obtain real-time model prediction data. Specifically, numerical simulation models can be built based on MATLAB, using various functions and tools provided by MATLAB to generate signals, simulate transmission processes, and perform modulation, demodulation, and coding to evaluate the performance of communication equipment.

[0025] After collecting real-time operational data and model prediction data, preprocessing can be performed. Specifically, clustering algorithms can be used to identify and remove outliers for key parameters such as signal strength, ensuring reliable data quality. Simultaneously, data can be aligned using timestamps, associating temperature and humidity sensor data with communication performance parameters (such as signal strength and latency) and mapping them uniformly to a geographic coordinate system, establishing a complete spatiotemporally consistent dataset. For high-frequency sampled sensor data (such as IMU inertial measurement units), Kalman filtering algorithms can be used for real-time noise reduction and compression, significantly reducing data volume while preserving key features. No restrictions are placed on the data fusion algorithm used here.

[0026] S2. Based on the communication nodes and link connections of the target communication network, the fused data is converted into spatiotemporal graph structure data, and a communication network performance prediction model based on spatiotemporal graph neural network (ST-GNN) is constructed.

[0027] In step S2, a dynamic spatiotemporal graph structure can be constructed using communication nodes (such as base stations and terminal devices) as vertices and link connections (such as channel states and transmission delays) as edges. ,in A node set contains all the key communication devices in the communication network. Each node can be represented by a unique identifier (ID). For example, These correspond to terminal equipment, switches, 5G DTUs, 5G base stations, MECs, customer backbone routers, and service servers in the communication network, respectively.

[0028] This is a time-varying set of edges used to define the core architecture that constitutes the entire dynamic spatiotemporal graph model. Each edge is essentially a triple containing complete connectivity information and state attributes. , and These are two adjacent nodes. It is a dynamic feature vector that comprehensively represents the link quality at a specific time step through a function. At that time, from arrive This refers to the real-time quality status of the network link. For example, key performance indicators such as transmission latency, packet loss rate, and available bandwidth can be integrated into a comprehensive score or a multi-dimensional feature vector to more accurately characterize the dynamic performance of the network link.

[0029] It is a two-dimensional matrix, where each row corresponds to a node (in order and node set). (The order is consistent), each column represents a feature dimension, such as CPU utilization (%), memory utilization (%), transmission rate (Mbps), reception rate (Mbps), etc. The dynamics of the spatiotemporal graph are achieved through the adjacency matrix. Depiction, its elements This can be expressed as a function of link quality: ; In the formula, For the Sigmoid function, Link quality comprehensive evaluation function, used to quantify communication nodes With nodes At time step Connection performance during the process. The temperature parameter allows adjustment of the slope of the Sigmoid function, controlling the weight distribution shape of the adjacency matrix. The dynamic average of link quality is a standardized process that integrates multiple indicators. First, the raw data of key performance indicators of the link are collected in real time. Then, each indicator is normalized according to the preset ideal and unacceptable value ranges and mapped to the unified dimension [0, 1]. Finally, different weights are assigned to each indicator according to business needs, and the comprehensive score value is calculated by weighted summation to represent the quality average of the overall health of the link at the current moment.

[0030] ; In the formula, For signal-to-noise ratio, For bandwidth, For link stability, This refers to link latency.

[0031] S3. Based on the constructed communication network performance prediction model, conduct fault mode and impact analysis on communication networks under complex environments, and predict the health indicator factors of communication networks under typical operating conditions. Based on fused multimodal data, a spatiotemporal graph neural network (ST-GNN) model is constructed. This model can handle the correlation between spatiotemporal data and capture dynamic changes in network state. The model is trained and optimized using historical data, simulation data, and real-time acquired data, enabling it to accurately predict the health of communication networks.

[0032] Figure 3 This is a schematic diagram of the structure of a communication network performance prediction model provided in this application. The communication network prediction model may include a graph convolutional layer, a fusion layer, and an attention mechanism layer. Specifically, the spatiotemporal graph structure corresponding to multiple time points is input into the graph convolutional layer for spatial feature extraction, outputting multiple spatial feature vectors; these multiple spatial feature vectors are input into the fusion layer for feature fusion, outputting a fused spatial feature vector; and the fused spatial feature vector is input into the attention mechanism layer for temporal feature extraction, outputting a spatiotemporal feature vector to output the health index parameters of the target communication network.

[0033] The health indicators output by the predictive model of a communication network can include signal-to-noise ratio, bit error rate, power, CPU / memory utilization, link utilization, and latency. Spatiotemporal Graph Neural Networks (ST-GNNs) combine graph neural networks (GNNs) and time series models (such as RNNs and LSTMs) to simultaneously capture the temporal and spatial dependencies of data, extending traditional graph neural networks to a dynamic spatiotemporal dimension to meet the complex data modeling needs of communication networks.

[0034] Specifically, graph neural networks can use hierarchical graph convolution to aggregate neighbor information, totaling... Layer, for a single time step The picture Its node characteristics and adjacency matrix The input is fed into the first graph convolutional layer. The output of the first graph convolutional layer... and the same Continue inputting into the second convolutional layer, and then... The layer output is: ; In the formula, For activation functions, such as , It is the identity matrix. It is an adjacency matrix. For degree matrix, For the first Layer trainable weight matrix.

[0035] To predict time steps Taking the health indicator parameters as an example, we can respectively... and The spatiotemporal graph structure data corresponding to the time step is input into the trained communication network health indicator prediction model, and finally fused into a fusion spatial feature vector.

[0036] Fusion spatial feature vectors The query vector, key vector, and value vector can be generated by linearly projecting three independent learnable weight matrices respectively. Then, temporal features are extracted using the following method: ; In the formula, , is the query vector, and the query node is in time. Is the state related to other nodes? , where is the key vector, representing the node at time . Historical traffic status , which is a value vector that stores the information actually used for aggregation. , and These represent the feature mapping matrices, respectively. The projection dimension of the key and query vectors.

[0037] S4. Calculate the health status of the target communication network based on the multi-dimensional health indicator factors of the target communication network.

[0038] In step S4, the health of the communication network can be calculated by weighted averaging of multidimensional health indicator factors. Specifically, dynamic health indicator factors of the communication network can be extracted and weighted according to node-level and edge-level attributes, including parameters such as signal-to-noise ratio, bit error rate, power, CPU / memory utilization, link utilization, and latency.

[0039] S5. Establish a health status optimization model under complex environments and conduct collaborative optimization of the health status of the target communication network from multiple dimensions.

[0040] The rating range here can include the first rating range [0, ], second scoring interval ( , ), third scoring interval[ ,1],wherein, < < The health status of the target communication network is determined in the following ways: When the health of the target communication network is in the first scoring range, the health status of the target communication network is determined to be faulty or high-risk. When the health of the target communication network is in the second scoring range, the health status of the target communication network is determined to be sub-healthy. When the health of the target communication network is within the first scoring range, the health status of the target communication network is determined to be healthy.

[0041] When the target communication network is in a sub-healthy state, the system also triggers a sub-healthy state warning and implements adjustments to optimize resource allocation for the target communication network. When the target communication network is in a faulty or high-risk state, the system also triggers a network abnormal state alarm and switches to another communication network to perform communication services.

[0042] In one specific embodiment, the scoring range can be found in Table 1 below.

[0043] Table 1 Health Evaluation Form

[0044] Furthermore, when resource allocation needs to be optimized, network parameters can be dynamically adjusted using multi-objective optimization algorithms based on the collected real-time operational data and model prediction data to improve signal coverage, reduce latency and packet loss rate, and adapt to environmental changes (such as temperature, humidity, and moving obstacles).

[0045] This application provides an optimized method for predicting and coordinating the health status of communication networks. Through multimodal data acquisition and fusion, as well as spatiotemporal graph neural network modeling, it can comprehensively and accurately predict the health status of communication networks in complex environments. This overcomes the limitations of existing evaluation methods that rely solely on single-modal data, thereby improving the accuracy and reliability of health status prediction.

[0046] This application can also promptly identify problems in communication networks under complex environments and take corresponding optimization measures, thereby improving the reliability and stability of communication networks, providing strong support for communication needs in complex environments, avoiding large-scale maintenance and repair work caused by communication network failures, and reducing maintenance costs.

[0047] In one embodiment of this application, to further ensure the model's compatibility with the environment, online incremental learning can also be performed on the model. When the communication network topology changes (such as adding a new base station) or the traffic pattern changes abruptly (such as peak holiday periods), the model does not need to be fully retrained. Instead, it can maintain prediction accuracy by jointly utilizing offline stored historical data and real-time streaming data through incremental updates.

[0048] For each round of predictions by the communication network prediction model, the loss value of the communication network prediction model is calculated in order to adjust the hyperparameters of the communication network prediction model.

[0049] In an optional implementation, the loss value of the communication network prediction model is calculated in the following manner. : ; in, The current hyperparameters of the communication network prediction model. The loss value is calculated based on the current hyperparameters and the health indicator parameters predicted in this round. For the diagonal elements of the information matrix, The regularization coefficient is . The optimal hyperparameters for the historical prediction model of the communication network are... This represents the number of hyperparameters.

[0050] In one embodiment of this application, a collaborative optimization strategy for communication quality of a communication network in a complex environment is provided.

[0051] Based on numerical simulation and real-time sensor data, a multi-objective optimization algorithm can be used to dynamically adjust network parameters to improve signal coverage, reduce latency and packet loss rate, and adapt to environmental changes (such as temperature, humidity, and moving obstacles).

[0052] Specifically, the resource allocation vector can be calculated in the following way to perform optimization adjustments on the target communication network: The resource allocation vector here can include transmit power, operating frequency band, antenna azimuth, routing path, data compression ratio, etc. A multi-objective optimization model is established to maximize signal strength and health, minimize end-to-end delay, and minimize packet loss rate. ; ; ; in, For communication nodes The resource allocation vector, For communication nodes The health of the nodes, For link utilization, For business The actual delay For business Demand latency, , These are the weighting coefficients.

[0053] In one feasible implementation, the communication network in a complex environment can be divided into multiple autonomous sub-regions (such as cellular cells, industrial IoT clusters, or drone relay groups). Nodes (base stations, terminals, or edge servers) in each sub-region dynamically adjust their transmit power, routing paths, and spectrum allocation based on local sensor data (such as RSSI, latency, temperature, and humidity) and collaborative information exchanged with neighboring nodes (through a lightweight Gossip protocol or federated learning framework). This is achieved by using distributed algorithms (such as Alternating Directional Multiplier Method (ADMM) or distributed reinforcement learning) to meet global objectives (such as maximizing network energy efficiency and achieving latency balance). This approach relies only on local computation and limited communication to achieve adaptive optimization in complex environments, effectively avoiding the high latency and single-point failure risks of centralized control.

[0054] A real-time feedback mechanism can also be established here to feed back the state information of the communication network after implementing the dynamic optimization strategy to the spatiotemporal graph neural network model and the dynamic optimization strategy formulation module in real time. By monitoring the real-time performance indicators of the communication network (such as signal strength, network latency, packet loss rate, etc.), the dynamic optimization strategy can be adjusted in a timely manner to improve the performance and reliability of the communication network.

[0055] Based on feedback, the parameters of the spatiotemporal graph neural network model are adjusted to more accurately predict the health of the communication network. Simultaneously, the dynamic optimization strategy is optimized and adjusted based on feedback to adapt to changes in the communication network state and the dynamic changes in the complex environment.

[0056] The dynamic optimization strategy based on the prediction results of the spatiotemporal graph neural network model can adaptively adjust according to the real-time changes in the state of the communication network and the dynamic changes in the complex environment, effectively improving the resilience and reliability of the communication network and reducing the time and frequency of communication interruptions.

[0057] Figure 4 A schematic diagram of a device for optimizing communication network health prediction and coordination provided in an embodiment of this application. Based on the same inventive concept, such as Figure 4As shown in the embodiments of this application, an optimized communication network health prediction and coordination device 40 is also provided. The device includes: The acquisition module 410 is used to acquire real-time operating data of the target communication network under complex environments, obtain simulation analysis data based on the performance simulation analysis of the constructed target communication network, and generate fused data. Module 420 is used to convert fused data into spatiotemporal graph structure data based on the communication node and link connection relationship of the target communication network, and to build a communication network performance prediction model based on spatiotemporal graph neural network (ST-GNN). The prediction module 430, based on the constructed communication network performance prediction model, performs fault mode and impact analysis on the communication network under complex environments and predicts the health indicator factors of the communication network under typical operating conditions. The calculation module 440 is used to calculate the health of the target communication network based on the multi-dimensional indicator factors of the target communication network. The evaluation module 450 is used to determine the health status of the target communication network based on the relationship between the health of the target communication network and a preset scoring range.

[0058] In a preferred embodiment, the communication network prediction model includes a graph convolutional layer, a fusion layer, and an attention mechanism layer, wherein, The spatiotemporal graph structures corresponding to multiple time points are input into the graph convolutional layer for spatial feature extraction, so as to output multiple spatial feature vectors. Multiple spatial feature vectors are input into a fusion layer for feature fusion to output a fused spatial feature vector; The fused spatial feature vector is input into the attention mechanism layer for temporal feature extraction, and the spatiotemporal feature vector is output to output the health index parameters of the target communication network.

[0059] In a preferred embodiment, the system further includes an optimization module (not shown in the figure) for calculating the loss value of the communication network prediction model for each round of prediction, so as to adjust the hyperparameters of the communication network prediction model.

[0060] In a preferred embodiment, the optimization module calculates the loss value of the communication network prediction model in the following manner. : ; in, The current hyperparameters of the communication network prediction model. The loss value is calculated based on the current hyperparameters and the health indicator parameters predicted in this round. For the diagonal elements of the information matrix, The regularization coefficient is . The optimal hyperparameters for the historical prediction model of the communication network are... This represents the number of hyperparameters.

[0061] In a preferred embodiment, the scoring interval includes a first scoring interval [0, ...]. ], second scoring interval ( , ), third scoring interval[ ,1],wherein, < < The evaluation module 450 determines the health status of the target communication network in the following ways: When the health of the target communication network is in the first scoring range, the health status of the target communication network is determined to be faulty or high-risk. When the health of the target communication network is in the second scoring range, the health status of the target communication network is determined to be sub-healthy. When the health of the target communication network is within the first scoring range, the health status of the target communication network is determined to be healthy.

[0062] In a preferred embodiment, when the target communication network is in a sub-healthy state, the optimization module is further configured to: Trigger a sub-health state warning and implement adjustments to optimize resource allocation for the target communication network; When the target communication network is in a faulty or high-risk state, the optimization module is also used for: Trigger a network anomaly alert and switch to another communication network to perform communication services.

[0063] In a preferred embodiment, the optimization module calculates the resource allocation vector in the following manner to perform optimized resource allocation adjustments on the target communication network: ; ; ; in, For communication nodes The resource allocation vector, For communication nodes The health of the nodes, For link utilization, For business The actual delay For business Demand latency, , These are the weighting coefficients.

[0064] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 5 As shown, the electronic device 500 includes a processor 510, a memory 520, and a bus 530.

[0065] The memory 520 stores machine-readable instructions executable by the processor 510. When the electronic device 500 is running, the processor 510 and the memory 520 communicate via the bus 530. When the machine-readable instructions are executed by the processor 510, they can perform the steps of an optimized communication network health prediction and coordination method as described in the above method embodiment. For specific implementation details, please refer to the method embodiment, which will not be repeated here.

[0066] This application also provides a computer-readable storage medium storing a computer program. When the computer program is run by a processor, it can execute the steps of an optimized communication network health prediction and coordination method as described in the above method embodiments. For specific implementation details, please refer to the method embodiments, which will not be repeated here.

[0067] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0068] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0069] Furthermore, the units described as separate components may or may not be physically separate. 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 units can be selected to achieve the purpose of this embodiment according to actual needs.

[0070] Furthermore, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0071] It should be noted that if the function is implemented as a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0072] In this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between these entities or operations.

[0073] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for optimizing communication network health prediction and coordination, characterized in that, The method includes: A high-precision virtual simulation model of the target communication network in a complex environment is built, and the simulation model is calibrated based on real-time running data to generate virtual-real fusion data; Based on the communication nodes and link connections of the target communication network, the fused data is converted into spatiotemporal graph structure data, and a communication network performance prediction model based on spatiotemporal graph neural network is constructed. Based on the constructed communication network performance prediction model, we conduct fault mode and impact analysis on communication networks under complex environments and predict the health indicator factors of communication networks under typical operating conditions. The health status of the target communication network is calculated based on the multi-dimensional health indicator factors of the target communication network. Establish a health status optimization model under complex environments and conduct collaborative optimization of the health status of the target communication network from multiple dimensions.

2. The method according to claim 1, characterized in that, The communication network prediction model includes graph convolutional layers, fusion layers, and attention mechanism layers, among which... The spatiotemporal graph structures corresponding to multiple time points are input into the graph convolutional layer for spatial feature extraction, so as to output multiple spatial feature vectors. Multiple spatial feature vectors are input into a fusion layer for feature fusion to output a fused spatial feature vector; The fused spatial feature vector is input into the attention mechanism layer for temporal feature extraction, and the spatiotemporal feature vector is output to output the health indicator factor of the target communication network.

3. The method according to claim 2, characterized in that, For each round of predictions by the communication network prediction model, the loss value of the communication network prediction model is calculated in order to adjust the hyperparameters of the communication network prediction model.

4. The method according to claim 3, characterized in that, The loss value of the communication network prediction model is calculated using the following method. : ； in, The current hyperparameters of the communication network prediction model. The loss value is calculated based on the current hyperparameters and the health indicator parameters predicted in this round. For the diagonal elements of the information matrix, The regularization coefficient is . The optimal hyperparameters for the historical prediction model of the communication network are... This represents the number of hyperparameters.

5. The method according to claim 1, characterized in that, The rating interval includes the first rating interval [0, ... ], second scoring interval ( , ), third scoring interval[ ,1],wherein, < < The health status of the target communication network is determined in the following ways: When the health of the target communication network is in the first scoring range, the health status of the target communication network is determined to be faulty or high-risk. When the health of the target communication network is in the second scoring range, the health status of the target communication network is determined to be sub-healthy. When the health of the target communication network is within the first scoring range, the health status of the target communication network is determined to be healthy.

6. The method according to claim 5, characterized in that, When the target communication network is in a sub-healthy state, it also includes: Trigger a sub-health state warning and implement adjustments to optimize resource allocation for the target communication network; When the health status of the target communication network is faulty or high-risk, it also includes: Trigger a network anomaly alert and switch to another communication network to perform communication services.

7. The method according to claim 6, characterized in that, The resource allocation vector is calculated in the following way to perform adjustments to optimize resource allocation in the target communication network: ； ； ； in, For communication nodes The resource allocation vector, For communication nodes The health of the nodes, For link utilization, For business The actual delay For business Demand latency, , These are the weighting coefficients.

8. An optimized communication network health prediction and coordination device, characterized in that, The device includes: The acquisition module is used to acquire high-precision simulation data and real-time operation data of the target communication network in complex environments, and generate fused data; The module is used to convert fused data into spatiotemporal graph structure data based on the communication nodes and link connections of the target communication network, and to build a communication network performance prediction model based on spatiotemporal graph neural network. The prediction module, based on the constructed communication network health indicator factor prediction model, obtains the health indicator factors of the target communication network under typical operating conditions. The calculation module is used to calculate the health of the target communication network based on the multi-dimensional health indicator factors of the target communication network. The evaluation module is used to determine the health status of the target communication network based on the relationship between the health of the target communication network and a preset scoring range. The optimization module is used for collaborative optimization of the health status of the target communication network from multiple dimensions.

9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus, and the processor executes the machine-readable instructions to perform the steps of the optimized communication network health prediction and coordination method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the method for optimizing communication network health prediction and coordination as described in any one of claims 1 to 7.

Images