Optimization method for automated testing of wireless network high-speed switching based on edge computing

Abstract

This invention discloses an automated testing and optimization method for wireless network high-speed handover based on edge computing, which relates to the field of wireless network handover. The method uses an improved gravitational search algorithm combined with GIS to deploy nodes, collects multi-dimensional data using intelligent heterogeneous sensors and reinforcement learning, pre-processes data using a GAN-based denoising autoencoder, models the data using a hybrid model of spatiotemporal graph convolution and LSTM, generates test cases using reinforcement learning Monte Carlo tree search, executes the tests in a distributed manner at edge nodes, formulates optimization strategies using multi-agent deep reinforcement learning, and implements strategy implementation and feedback using SDN and blockchain. The method utilizes multi-dimensional data collection, intelligent modeling, automated testing, and deep reinforcement learning optimization to improve high-speed handover performance, reduce latency and failure rates, increase testing efficiency, optimize resource allocation, enhance network security, and comprehensively ensure stable and efficient operation of wireless networks.

Description

Technical Field

[0001] The present invention relates to the field of wireless network switching technology, and in particular to an edge computing-based wireless network high-speed switching automated testing optimization method. Background Art

[0002] With the rapid development of technologies such as mobile internet and the Internet of Things, the application scope of wireless networks continues to expand, and the requirements for network performance and stability are also increasing. In wireless network environments, the mobility of user devices makes network handover a common operation. For example, in scenarios such as high-speed trains and subways, user devices need to frequently switch between different base stations to ensure network continuity. However, existing wireless network handover mechanisms have many problems in high-speed mobility scenarios.

[0003] Traditional wireless network handovers are primarily based on a single metric, such as signal strength. When signal strength falls below a certain threshold, the device triggers a handover. However, this approach is insufficient in high-speed mobile scenarios. Firstly, due to the rapid movement speed, signal strength changes very quickly, and a single threshold judgment can easily lead to frequent handovers or untimely handovers. Frequent handovers increase handover latency, reduce user experience, and even cause connection interruptions. Untimely handovers, on the other hand, can leave devices in a low-quality network environment for extended periods, impacting data transmission stability. Secondly, existing handover mechanisms lack comprehensive awareness of the network environment and fail to comprehensively consider multiple factors, such as bandwidth, packet loss rate, and latency, resulting in poor network performance after handover.

[0004] Currently, automated wireless network testing primarily relies on traditional testing methods, which often require extensive manual intervention and result in low test efficiency. Furthermore, traditional testing methods struggle to simulate complex real-world network environments and cannot accurately assess wireless network performance in high-speed handover scenarios. Furthermore, existing test result analysis methods often only provide basic statistical data, making it difficult to deeply explore correlations between network parameters and provide effective guidance for network optimization.

[0005] The development of edge computing technology has provided new insights for optimizing high-speed handover in wireless networks. Edge computing can store computation and data closer to the data source, reducing data transmission latency and improving network responsiveness. However, effectively applying edge computing technology to automated testing and optimization of high-speed handover in wireless networks remains a pressing challenge. Summary of the Invention

[0006] The present invention proposes an edge computing-based wireless network high-speed switching automated test optimization method to solve the problems mentioned in the above-mentioned prior art.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: a method for optimizing automated testing of high-speed handover in wireless networks based on edge computing, comprising:

[0008] Edge computing node deployment steps: Use the improved gravitational search algorithm (IGSA) combined with geographic information system (GIS) data to optimize edge computing node deployment. Consider network topology, geographical environment, and expected traffic distribution to build a comprehensive objective function, and use GIS data to plan node locations.

[0009] Multi-dimensional data collection steps: Deploy an intelligent heterogeneous sensor network, integrate visible light communication (VLC) and traditional radio frequency (RF) sensing technology to collect data, dynamically adjust the sampling strategy based on reinforcement learning, and use formulas to adapt the sampling frequency and data type focus;

[0010] Data preprocessing steps: Use the denoising autoencoder DAGAN based on the generative adversarial network (GAN) to purify the raw data, use the adaptive normalization method to process the data, and use tensor decomposition technology to reduce the dimensionality of multi-source data and preserve high-order feature relationships;

[0011] Switching scene modeling steps: Build a hybrid model based on the spatiotemporal graph convolutional network (ST-GCN) and the long short-term memory (LSTM) network. ST-GCN captures spatial associations, LSTM learns the changing patterns of time series, and introduces the attention gating mechanism (AGC) to focus on key spatiotemporal features.

[0012] Automated test case generation steps: Generate test cases using Monte Carlo Tree Search (MCTS-RL) based on reinforcement learning, construct a reward function based on performance indicators, and dynamically adjust test priorities using a fuzzy clustering algorithm;

[0013] Test execution and result analysis: Deploy a distributed test execution framework on edge computing nodes, use a real-time streaming data analysis engine to analyze test results, and locate problem areas through a deep association rule mining algorithm combined with a self-organizing map (SOM).

[0014] Optimization strategy formulation: Multi-agent deep reinforcement learning (MA-DRL) is used to formulate optimization strategies. The agents are responsible for optimizing different network parameters and learning the optimal strategy through collaborative competition.

[0015] Strategy implementation and feedback: Use software-defined network (SDN) technology to implement optimization strategies, adjust network configurations through interfaces, establish a blockchain-based feedback mechanism, and use smart contracts to verify and analyze data.

[0016] Furthermore, it also includes:

[0017] Network security monitoring steps: Deploy an intrusion detection system based on federated learning and generative adversarial networks (FL-GAN). Use local data to train the generator at each edge node to generate normal traffic samples, and the discriminator to identify abnormal traffic. Through the federated learning mechanism, each node collaborates in training while protecting data privacy. When an anomaly is detected, the emergency response smart contract is activated to isolate the risk area and notify the administrator.

[0018] Furthermore, it also includes:

[0019] Steps for dynamic allocation of edge computing resources: Use the adaptive resource allocation algorithm to build a resource allocation agent, and dynamically adjust the computing and storage allocation network resources through the Q-learning algorithm based on the real-time network load, task priority and node resource status. The formula is: ,in Is the action taken in state s Q value, is the learning rate, r is the reward, γ is the discount factor, It is the next state.

[0020] Furthermore, the multi-dimensional data collection step uses the state-action-reward mechanism to adaptively change the sampling frequency and data type focus according to the formula: , is to take action in state s The probability of is the state-action-value function, is the temperature parameter;

[0021] Introduce data collection technology based on quantum key distribution (QKD), use the principle of quantum encryption to encrypt collected data, and combine distributed ledger technology to record data collection timestamps and hash values.

[0022] Furthermore, the switching scene modeling step introduces the attention gating mechanism AGC, which focuses on key spatiotemporal features through the formula: in is the gating signal, σ is the activation function, is the weight matrix, is the hidden state at the previous moment, is the current input;

[0023] A combination of variational autoencoder (VAE) and generative adversarial network (GAN) is used to expand scene data. VAE learns the real switching scene distribution and generates potential samples. GAN optimizes the generated samples and transfers knowledge from similar network scene models through transfer learning technology.

[0024] Furthermore, the automated test case generation step uses a Bayesian optimization-based test case generation algorithm to build a Bayesian model to evaluate the impact of different test cases on network performance, and selects the optimal test case by maximizing the expected improvement (EI) criterion. The formula is: ,in is the current optimal solution, is the objective function value.

[0025] Furthermore, the test execution and result analysis steps use an interpretable analysis method based on deep learning to interpret the test results. By generating explanations for the prediction results of the convolutional neural network (CNN) and recurrent neural network (RNN) models, the key factors affecting the switching performance are located.

[0026] Furthermore, the optimization strategy formulation step introduces a multi-objective optimization algorithm based on game theory, constructs a game model considering different needs, and obtains the optimal switching strategy by solving the Nash equilibrium.

[0027] Furthermore, the strategy implementation and feedback steps use digital twin technology to build a virtual network model, evaluate the strategy effect through simulation operation, utilize the virtual network to interact with the real network, and apply the optimized strategy in the virtual network to the real network.

[0028] Furthermore, an intelligent monitoring and early warning system will be established, using the Internet of Things and big data analysis technologies to monitor edge computing node hardware, network connections and system performance, using deep learning-based anomaly detection models to identify abnormal behaviors, combining natural language processing technology to generate easy-to-understand early warning information, and using knowledge graph technology to associate various types of monitoring data.

[0029] Compared with the existing technology, the beneficial effects of the present invention are:

[0030] In terms of network handover performance, multi-dimensional data collection and comprehensive analysis, taking into account factors such as signal strength, bandwidth, packet loss rate, and latency, allows for the development of more rational handover strategies, avoiding the limitations of traditional methods based solely on signal strength. This enables devices to perform handovers more promptly and accurately in high-speed mobile scenarios, reducing handover latency and handover failure rates, improving the continuity and stability of network connections, and significantly enhancing the user experience.

[0031] In terms of testing efficiency, automated test case generation and a distributed test execution framework enable automated high-speed handover testing of wireless networks. This reduces manual intervention, improves testing efficiency, and enables rapid and comprehensive evaluation of network performance in various scenarios. Furthermore, a real-time streaming data analysis engine and deep association rule mining algorithms enable in-depth analysis of test results and the discovery of correlations between network parameters, providing strong support for network optimization.

[0032] In terms of resource utilization, the adaptive resource allocation algorithm based on reinforcement learning can dynamically adjust the resource allocation of edge computing nodes according to the real-time network load and task priority, thereby improving resource utilization efficiency and reducing operating costs.

[0033] In terms of network security, the deployment of an intrusion detection system based on federated learning and generative adversarial networks can effectively identify and prevent abnormal network behavior, ensuring the security of the wireless network. Furthermore, a blockchain-based feedback mechanism and emergency response smart contracts ensure the authenticity and immutability of data, improving the reliability and traceability of the system.

[0034] In summary, the method of the present invention can effectively improve the performance and stability of high-speed switching in wireless networks, increase testing efficiency, optimize resource utilization, and ensure network security, and has important application value and market prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a schematic block diagram of an edge computing-based wireless network high-speed handover automated test optimization method proposed by the present invention;

[0036] Figure 2 A schematic diagram comparing the switching success rates of different methods;

[0037] Figure 3 This is a schematic diagram comparing switching delays in different scenarios;

[0038] Figure 4 A schematic diagram showing the comparison of test efficiency. DETAILED DESCRIPTION

[0039] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0040] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0041] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the said features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. In addition, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be a connection between the two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. The present invention will be further described in detail below with reference to the accompanying drawings.

[0042] Reference Figures 1 to 4 : A wireless network high-speed switching automated test optimization method based on edge computing, comprising:

[0043] Edge computing node deployment steps: When deploying edge computing nodes, first collect network topology information of the wireless network coverage area, such as base station distribution, signal transmission path, etc., as well as geographic information system (GIS) data, including topography, building distribution, etc. At the same time, predict the expected traffic distribution in different areas based on historical traffic data and business needs. Use the improved gravitational search algorithm (IGSA) to optimize node positions. Initialize a set of random node positions as the initial population, and each node position is regarded as an object with mass and position attributes. By calculating the gravitational force between objects, updating the position and velocity of the objects, and gradually iterating to find the optimal node deployment position. During the calculation process, the comprehensive objective function is used ,in is the distance from the node to the key business area, For node construction and operation and maintenance costs, is the signal strength, is the average signal strength, 、 、 It can be adjusted according to the actual situation. For example, in areas with high signal coverage requirements, the The optimized node locations are marked on the GIS map and deployed on the ground to ensure that the nodes can process network data efficiently.

[0044] Multi-dimensional data collection steps: Deploy an intelligent heterogeneous sensor network, rationally distributing visible light communication (VLC) sensors and traditional radio frequency (RF) sensors within the wireless network coverage area. VLC sensors can be installed on indoor lighting fixtures to collect network data in indoor environments, while RF sensors can be installed at base stations, utility poles, and other locations to collect network signals outdoors. Sensors use reinforcement learning to dynamically adjust their sampling strategies. Sensor states are defined as real-time fluctuations in the network, such as the rate of change of signal strength and the amplitude of bandwidth fluctuations. Actions include adjusting the sampling frequency and data type focus. For example, when signal strength fluctuates dramatically, the sampling frequency is increased and signal strength data is prioritized. The reward function is designed based on the accuracy and timeliness of collected data, providing positive rewards when the collected data accurately reflects the network status and is transmitted in a timely manner. Through continuous interaction with the network environment, sensors gradually learn the optimal sampling strategy, comprehensively collecting network data such as signal strength, bandwidth, packet loss rate, and latency, as well as information such as the location, speed, and direction of the mobile terminal.

[0045] Data preprocessing steps: The collected raw data is first denoised using a denoising autoencoder (DAGAN) based on a generative adversarial network (GAN). The generator is trained to generate simulated noise data, and the discriminator distinguishes between real noise and simulated noise. Through continuous adversarial training, the denoising effect is improved. For example, for signal strength data affected by electromagnetic interference, after DAGAN processing, the noise interference can be removed and the true signal strength value can be restored. Data normalization uses an adaptive normalization method, based on the dynamic mean of the data. and standard deviation , using the formula Normalize the data. This ensures that different types of data have the same scale, facilitating subsequent analysis and processing. Use tensor decomposition technology to reduce the dimensionality of multi-source data. Represent the collected multi-dimensional data as tensors. Using a tensor decomposition algorithm, decompose high-dimensional tensors into low-dimensional tensors, preserving high-order feature relationships in the data, reducing data redundancy, and improving data processing efficiency.

[0046] Switching scene modeling steps: Build a hybrid model based on the spatiotemporal graph convolutional network (ST-GCN) and the long short-term memory network (LSTM). The base stations and mobile terminals in the network are regarded as graph nodes, and the connections between the nodes represent the signal transmission relationship between them, and a graph structure is constructed. ST-GCN performs convolution operations on the graph structure to capture the correlation between network parameters in the spatial dimension, such as the signal interference relationship between different base stations. LSTM processes time series data and learns the temporal variation of network parameters, such as the fluctuation of signal strength over time. Introducing the attention gating mechanism (AGC), through the formula Calculate the gating signal, where σ is the activation function, is the weight matrix, is the hidden state at the previous moment, The gating signal controls the model's attention to different spatiotemporal features, focusing on features that have a significant impact on handover decisions, thereby improving the model's ability to capture complex handover scenarios. The model is trained using extensive historical data and its parameters are continuously adjusted to accurately predict network handovers in various scenarios.

[0047] Automated test case generation steps: Test cases are generated based on Monte Carlo Tree Search with Reinforcement Learning (MCTS-RL). The agent's state is defined as the current network environment, including network load, signal strength, and mobile terminal speed. Actions include selecting different combinations of test parameters, such as handover thresholds and signal strength assessment algorithms. The reward function is designed based on performance indicators of the test results, such as handover success rate and latency. A positive reward is given when the test results meet the expected goals. The agent explores the simulated network environment, constructing a search tree using the Monte Carlo Tree Search algorithm to select the optimal test action. During the search process, the nodes in the search tree are continuously expanded, the value of each node is evaluated, and the node with the highest value is selected for further exploration. A fuzzy clustering algorithm is used to classify the generated test cases, dynamically adjusting their priority based on different network scenarios and test objectives. For example, test cases in high-speed mobile scenarios are given higher priority to ensure test coverage of key scenarios.

[0048] Test execution and result analysis steps: Deploy a distributed test execution framework on edge computing nodes and distribute the generated test cases to different edge nodes for parallel execution. Use a real-time streaming data analysis engine (such as Apache Flink) to process and analyze test data in real time. During testing, continuously collect network performance metrics such as handover success rate, latency, and packet loss rate. Use deep association rule mining algorithms (such as an extended Eclat algorithm) to discover correlations between network parameters. For example, analyze the correlation between signal strength and packet loss rate to identify key factors affecting handover performance. Combined with a self-organizing map (SOM) neural network, map the high-dimensional test data onto a two-dimensional plane, visually presenting the distribution of handover performance. By observing clustering in the SOM map, areas of poor network performance can be quickly located, providing a basis for network optimization.

[0049] Optimization strategy formulation steps: Use multi-agent deep reinforcement learning (MA-DRL) to formulate the optimization strategy. Assign different parameter optimization tasks in the network to multiple agents, such as one agent responsible for optimizing the switching threshold and another responsible for optimizing the signal strength evaluation algorithm. Each agent explores and learns in the network environment based on its own tasks. Through collaboration and competition, the agents continuously adjust their strategies with the goal of optimizing the overall network performance. Use the Q-learning algorithm to update the Q value of the agent, the formula is , is the learning rate, r is the reward, γ is the discount factor, Through continuous iterative learning, the agent gradually finds the optimal optimization strategy and improves the performance of high-speed switching in wireless networks.

[0050] Policy Implementation and Feedback Steps: Software-Defined Networking (SDN) technology is used to implement the optimization strategy. The SDN controller's southbound interface allows real-time adjustments to network device configurations, such as modifying base station handover parameters and adjusting signal strength assessment algorithms. A blockchain-based feedback mechanism is also established, with each edge node encrypting and uploading network performance data after policy implementation to the blockchain. Smart contracts automatically verify and analyze this data to determine policy effectiveness. If network performance improves after policy implementation, the strategy is continued. If performance degrades, the new data is fed back to the handover scenario model and optimization strategy module, updating and optimizing the model and adjusting the optimization strategy, forming a closed-loop optimization process.

[0051] The present invention further comprises the following steps:

[0052] Network security monitoring steps: Deploy an intrusion detection system based on federated learning and generative adversarial networks (FL-GAN). Each edge node utilizes abundant local network traffic data to train the generator. The generator utilizes a deep neural network architecture, combining multi-layer convolutional and deconvolutional neural networks to learn and extract features from local normal network traffic data, thereby generating realistic normal traffic samples. The discriminator, also built on a deep neural network with a convolutional neural network as its backbone, is responsible for discriminating input traffic samples and accurately identifying anomalous traffic. Leveraging the federated learning mechanism, edge nodes achieve collaborative training while strictly protecting local data privacy. Specifically, each node only uploads updated model parameters, not raw data. These parameters are aggregated on a central server using secure aggregation technology, and the updated model is distributed to all nodes for continuous iterative optimization. This significantly improves the intrusion detection model's ability to identify diverse attacks, such as DDoS attacks and SQL injection attacks. Once the discriminator successfully detects anomalous traffic, the system immediately triggers a blockchain-based emergency response smart contract. The smart contract pre-sets detailed rules and execution logic, ensures the reliability of instruction execution through the tamper-proof nature of blockchain, automatically isolates risk areas, and promptly notifies network administrators through reliable communication methods, enabling rapid response and effective disposal of network security threats.

[0053] The present invention further comprises the following steps:

[0054] Steps for dynamic allocation of edge computing resources: Adopt an adaptive resource allocation algorithm based on reinforcement learning to build a resource allocation agent. The agent collects network load information based on real-time perception of the network environment, including but not limited to the traffic rate and number of data packets of each edge node, to measure the current network busyness; at the same time, it obtains task priorities. For example, tasks with high real-time requirements (such as high-definition video streaming, industrial control instruction transmission, etc.) need to prioritize resource supply; it also monitors node resource status, such as the CPU usage of computing nodes, the remaining memory, and the storage space occupancy of storage nodes. The agent uses the Q-learning algorithm (formula is ,in Is the action taken in state s Q value, is the learning rate, r is the reward, γ is the discount factor, is the next state) to dynamically adjust the allocation of computing, storage, and network resources to ensure efficient resource utilization and stable switching performance.

[0055] In this invention, the multi-dimensional data collection step incorporates a data collection technology based on quantum key distribution (QKD) to ensure data transmission security. Based on the fundamental principles of quantum mechanics, QKD utilizes the quantum states of photons to generate and distribute encryption keys. During the data collection process, the transmitter randomly generates a sequence of single photons in different quantum states (such as polarization states) to encode the key information. The receiver then measures the received photons using a corresponding measurement basis. Due to the non-cloning nature of quantum states and the collapse of quantum states upon measurement, any third-party eavesdropping will be detected by both the transmitter and receiver. Once the key distribution process is secure, the generated quantum key is used to encrypt the collected data, ensuring that even if intercepted during transmission to the edge node, the data cannot be eavesdropped or tampered with. To further enhance data credibility, distributed ledger technology is incorporated. Each sensor participates as a ledger node. Upon completion of data collection, the sensor records a precise data acquisition timestamp. This timestamp is calibrated using a time synchronization protocol (such as the Network Time Protocol (NTP)) to ensure time accuracy. The sensor also calculates a hash value for the collected data. This hash function uses a secure cryptographic hash algorithm, such as SHA-256, to map the data into a fixed-length hash value. These timestamps and hash values are recorded in a distributed ledger. The data in the ledger is interconnected through the blockchain's chain structure and cryptographic techniques, making it tamper-proof. This allows for reliable verification of the data's source and integrity, thus comprehensively ensuring the security and credibility of multi-dimensional data collection.

[0056] In the present invention, the switching scene modeling step utilizes a combination of a variational autoencoder (VAE) and a generative adversarial network (GAN) to expand the scene data. As a generative model, the variational autoencoder (VAE) uses an encoder to map real-world switching scene data into a low-dimensional latent space and learn its distribution characteristics. The encoder network, typically composed of a multi-layer neural network, extracts features from the input switching scene data (such as signal strength, node location, network topology, and other information) to obtain the mean and variance parameters of the latent space. Reparameterization techniques are used to sample from the latent distribution, and the decoder then converts the sampled latent vectors into switching scene data samples, thereby learning the distribution of real-world switching scenarios and generating latent samples. The generative adversarial network (GAN) further optimizes the samples generated by the VAE. The GAN consists of a generator and a discriminator. The generator receives the latent samples generated by the VAE and processes them to generate data that is closer to the real scene. The discriminator is responsible for determining whether the input data is from the real scene or samples generated by the generator. During the continuous adversarial training process, the generator strives to produce more realistic data to deceive the discriminator, while the discriminator continuously improves its discrimination ability, thereby increasing the diversity and authenticity of the generated samples. Furthermore, through transfer learning techniques, knowledge is transferred from models of similar network scenarios. For example, in network scenarios with similar topologies or communication patterns, the feature representations and parameters learned by the trained model are transferred to the current handover scenario model. This reduces the amount of data and computing resources required for model training, accelerates the current model training process, and improves the model's generalization ability, enabling it to better adapt to complex and changing wireless network handover scenarios.

[0057] In the present invention, the automated test case generation step adopts a test case generation algorithm based on Bayesian optimization. First, a Bayesian model is constructed. The model is based on probability theory and uses prior knowledge and continuously acquired new data to update the understanding of the relationship between network performance and test cases. In this model, different test cases are regarded as input variables, which cover a variety of factors that affect the high-speed switching of wireless networks, such as changes in network topology, node movement speed, signal interference intensity, etc. Network performance indicators, such as switching success rate, delay time, packet loss rate, etc., are used as objective function values. Based on the constructed Bayesian model, the optimal test case is screened by maximizing the expected improvement (EI) criterion. Formula is the core expression of this principle. The optimal solution currently determined is the test case parameter combination found during previous testing and analysis that can achieve the best network performance. For any test case The corresponding objective function value reflects the performance of the network under this test case. E represents the expected value. This means that only when The difference between the two, that is, the measurement from the current test case Improve to the optimal solution By continuously applying this criterion to evaluate and screen numerous potential test cases, the number of test cases can be significantly reduced while ensuring sufficient test coverage of various scenarios and influencing factors for high-speed wireless handovers. This avoids a large number of redundant tests, greatly improving test efficiency and enabling more accurate identification and optimization of network performance issues.

[0058] In the present invention, the test execution and result analysis steps employ deep learning-based interpretability analysis methods (such as the LIME algorithm) to interpret the test results. Taking the LIME algorithm (Local Interpretable Model-Independent Explanations) as an example, it can provide intuitive and easy-to-understand explanations for the prediction results of deep learning models. In wireless network high-speed handover testing, models such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs) are commonly used for performance prediction. CNNs excel at processing grid-structured data and excel at analyzing network characteristics related to spatial distribution (such as the distribution of signal strength at different geographic locations). RNNs, due to their excellent processing capabilities for time series data, are suitable for analyzing temporal changes in network performance (such as the fluctuation of handover latency at consecutive time points). The LIME algorithm builds on these models. It perturbs local data near the model's prediction results to generate a series of new data samples. Based on these new samples, a simple and interpretable surrogate model (such as a linear regression model) is constructed. The coefficients and feature importance of this surrogate model clearly demonstrate which input features (such as network topology, node mobility, and signal interference strength) have a key impact on the model's prediction results. This approach allows testers to gain a deeper understanding of the model's decision-making process. Rather than simply facing a prediction from a deep learning model, they can clearly understand why the model makes the judgment it does. This allows them to quickly identify key factors impacting wireless network high-speed handover performance, such as whether signal interference in a specific area is causing an increased handover failure rate, or whether excessive node movement is causing handover latency to exceed standards. This significantly improves the efficiency of troubleshooting and resolution, enabling more targeted optimization of wireless network high-speed handover performance.

[0059] In this invention, the optimization strategy formulation step incorporates a multi-objective optimization algorithm based on game theory, comprehensively considering the needs of different stakeholders in the network. Users pursue a high-quality network experience, focusing on low latency, high bandwidth, and high stability during handovers. They expect seamless network connection switching during mobility, ensuring smooth operation of various services (such as HD video streaming and real-time gaming). Operators prioritize efficient utilization of network resources, control of operating costs, and maintenance of service quality, striving to achieve profitability while meeting user needs. Service providers want their services to remain unaffected during network handovers, ensuring service reliability and availability. Based on these diverse needs, a game model is constructed. In this model, each stakeholder is considered a game participant, each possessing a distinct strategy space. For example, users can choose different device configurations and network access modes; operators can adjust base station layouts and resource allocation strategies; and service providers can optimize service deployment architectures and data transmission protocols. The strategy choices of each participant influence each other, generating corresponding benefits or costs. The optimal handover strategy is determined by solving a Nash equilibrium. A Nash equilibrium is a stable state in a game where all participants have chosen their respective strategies, and no single participant can improve their own benefits by changing their strategy. In wireless network handover scenarios, complex mathematical calculations and iterative problem-solving processes are used to find a strategy combination that balances the interests of all parties. This ensures network performance, improves user experience, and effectively controls operating costs, achieving a win-win situation for all parties and maximizing overall network benefits.

[0060] In this invention, the policy implementation and feedback steps utilize digital twin technology to construct a virtual network model. This technology creates a highly realistic virtual network model that comprehensively maps the topology of a real wireless network, accurately capturing the locations and connections of base stations, access points, and terminal devices. It also simulates the network's physical layer characteristics, including signal propagation, attenuation, and interference. The network's protocol and application layers are also meticulously modeled, including the operational mechanisms of various network protocols and the traffic characteristics of different service applications. After the virtual network model is constructed, the optimization policy is rehearsed within it. During the rehearsal, the policy's effectiveness is comprehensively evaluated by simulating various network operating scenarios, such as high-traffic periods and rapid user mobility. Advanced simulation tools and algorithms are used to accurately calculate and monitor network performance indicators (such as handover success rate, latency, and throughput) in real time. This process allows for the early detection of potential issues, such as localized network congestion or handover failures caused by excessive load on certain devices, which may occur after policy implementation. By leveraging real-time interaction between the virtual network and the real network, efficient policy migration is achieved. By continuously adjusting and optimizing policies within the virtual network, and once their effectiveness is confirmed, leveraging technologies like the Internet of Things and big data, the optimized policy parameters in the virtual network are quickly and accurately transferred to the various control units and devices in the real network. Simultaneously, the real network feeds real-time operational data back to the virtual network, further calibrating and optimizing the virtual model. This creates a closed-loop system of continuous iteration and bidirectional optimization, significantly improving the accuracy and efficiency of policy implementation and ensuring the stable and efficient operation of high-speed handovers in wireless networks.

[0061] The present invention establishes an intelligent monitoring and early warning system. Leveraging Internet of Things (IoT) technology, various sensors are deployed at edge computing nodes to collect real-time operating parameters of hardware devices (such as CPUs, GPUs, and memory modules), including temperature, usage, and voltage. The system also monitors network connectivity, such as signal strength, packet loss rate, and network latency, as well as system performance indicators such as task processing speed and resource allocation efficiency. The collected data is aggregated via high-speed communication links to a big data analysis platform. For anomaly detection, an autoencoder model based on deep learning is employed. The autoencoder consists of an encoder and a decoder. The encoder compresses high-dimensional monitoring data into low-dimensional feature representations, which the decoder then attempts to restore. Under normal operating conditions, reconstruction error is minimal. However, once a network or device anomaly occurs, the data features change, and reconstruction error increases significantly, thereby identifying abnormal behavior. Incorporating natural language processing technology, the system can convert detected anomaly information into easily understandable early warning messages. For example, using a text generation algorithm, complex technical indicator anomalies can be translated into intuitive statements such as "The edge node CPU temperature is too high, which may affect network switching performance. Please check the cooling system immediately," making it easier for operations and maintenance personnel to understand. Furthermore, knowledge graph technology is used to perform correlation analysis on various monitoring data. Knowledge graphs use a graph structure to store and represent semantic relationships between data, for example, linking hardware failures with network performance degradation, system error logs, and other information. Based on this, the system can intelligently diagnose the root cause of the problem and provide targeted solutions, such as "Detecting a disconnection between a base station and an edge node, this may be a fiber optic failure; it is recommended to check the corresponding line." This series of technical measures comprehensively ensures the stable operation of the wireless network, promptly identifies and resolves potential problems, and ensures the reliability of high-speed wireless network switching in edge computing environments.

[0062] The above are only preferred specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solutions and inventive concepts of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A wireless network high-speed switching automated test optimization method based on edge computing, characterized in that: The following steps are involved: Edge computing node deployment steps: Use the improved gravitational search algorithm (IGSA) combined with geographic information system (GIS) data to optimize edge computing node deployment. Consider network topology, geographical environment, and expected traffic distribution to build a comprehensive objective function, and use GIS data to plan node locations. Multi-dimensional data collection steps: Deploy an intelligent heterogeneous sensor network, integrate visible light communication (VLC) and traditional radio frequency (RF) sensing technology to collect data, dynamically adjust the sampling strategy based on reinforcement learning, and use formulas to adapt the sampling frequency and data type focus; Data preprocessing steps: Use the denoising autoencoder DAGAN based on the generative adversarial network (GAN) to purify the raw data, use the adaptive normalization method to process the data, and use tensor decomposition technology to reduce the dimensionality of multi-source data and preserve high-order feature relationships; Switching scene modeling steps: Build a hybrid model based on the spatiotemporal graph convolutional network (ST-GCN) and the long short-term memory (LSTM) network. ST-GCN captures spatial associations, LSTM learns the changing patterns of time series, and introduces the attention gating mechanism (AGC) to focus on key spatiotemporal features. Automated test case generation steps: Generate test cases using Monte Carlo Tree Search (MCTS-RL) based on reinforcement learning, construct a reward function based on performance indicators, and dynamically adjust test priorities using a fuzzy clustering algorithm; Test execution and result analysis: Deploy a distributed test execution framework on edge computing nodes, use a real-time streaming data analysis engine to analyze test results, and locate problem areas through a deep association rule mining algorithm combined with a self-organizing map (SOM). Optimization strategy formulation: Multi-agent deep reinforcement learning (MA-DRL) is used to formulate optimization strategies. The agents are responsible for optimizing different network parameters and learning the optimal strategy through collaborative competition. Strategy implementation and feedback: Use software-defined network (SDN) technology to implement optimization strategies, adjust network configurations through interfaces, establish a blockchain-based feedback mechanism, and use smart contracts to verify and analyze data.

2. The method for optimizing the automated test of wireless network high-speed switching based on edge computing according to claim 1 is characterized in that: Also includes: Network security monitoring steps: Deploy an intrusion detection system based on federated learning and generative adversarial networks (FL-GAN). Use local data to train the generator at each edge node to generate normal traffic samples, and the discriminator to identify abnormal traffic. Through the federated learning mechanism, each node collaborates in training while protecting data privacy. When an anomaly is detected, the emergency response smart contract is activated to isolate the risk area and notify the administrator.

3. The method for optimizing the automated test of wireless network high-speed switching based on edge computing according to claim 1 is characterized in that: Also includes: Steps for dynamic allocation of edge computing resources: Use the adaptive resource allocation algorithm to build a resource allocation agent, and dynamically adjust the computing and storage allocation network resources through the Q-learning algorithm based on the real-time network load, task priority and node resource status. The formula is: ,in Is the action taken in state s Q value, is the learning rate, r is the reward, γ is the discount factor, It is the next state.

4. The method for optimizing the automated test of wireless network high-speed switching based on edge computing according to claim 1, characterized in that: The multi-dimensional data collection step uses the state-action-reward mechanism to adaptively change the sampling frequency and data type focus according to the formula: , is to take action in state s The probability of is the state-action-value function, is the temperature parameter; Introduce data collection technology based on quantum key distribution (QKD), use the principle of quantum encryption to encrypt collected data, and combine distributed ledger technology to record data collection timestamps and hash values.

5. The method for optimizing the automated test of wireless network high-speed switching based on edge computing according to claim 1 is characterized in that: The switching scene modeling step introduces the attention gating mechanism AGC, which focuses on key spatiotemporal features through the formula: in is the gating signal, σ is the activation function, is the weight matrix, is the hidden state at the previous moment, is the current input; A combination of variational autoencoder (VAE) and generative adversarial network (GAN) is used to expand scene data. VAE learns the real switching scene distribution and generates potential samples. GAN optimizes the generated samples and transfers knowledge from similar network scene models through transfer learning technology.

6. The method for optimizing the automated test of wireless network high-speed switching based on edge computing according to claim 1, characterized in that: The automated test case generation step uses a Bayesian optimization-based test case generation algorithm to build a Bayesian model to evaluate the impact of different test cases on network performance, and selects the optimal test case by maximizing the expected improvement (EI) criterion. The formula is: ,in is the current optimal solution, is the objective function value.

7. The edge computing-based wireless network high-speed switching automated test optimization method according to claim 1, characterized in that: The test execution and result analysis steps use an interpretable analysis method based on deep learning to interpret the test results. By generating explanations for the prediction results of the convolutional neural network (CNN) and recurrent neural network (RNN) models, the key factors affecting the switching performance are located.

8. The method for optimizing the automated test of wireless network high-speed switching based on edge computing according to claim 1, characterized in that: The optimization strategy formulation step introduces a multi-objective optimization algorithm based on game theory, constructs a game model considering different needs, and obtains the optimal switching strategy by solving the Nash equilibrium.

9. The method for optimizing the automated test of wireless network high-speed switching based on edge computing according to claim 1, characterized in that: The strategy implementation and feedback steps use digital twin technology to build a virtual network model, evaluate the strategy effect through simulation operation, utilize the virtual network to interact with the real network, and apply the optimized strategy in the virtual network to the real network.

10. The method for optimizing the automated test of wireless network high-speed switching based on edge computing according to claim 1, characterized in that: Establish an intelligent monitoring and early warning system, use the Internet of Things and big data analysis technology to monitor edge computing node hardware, network connections and system performance, use deep learning-based anomaly detection models to identify abnormal behavior, combine natural language processing technology to generate easy-to-understand early warning information, and associate various types of monitoring data through knowledge graph technology.

Images