Network switching method, system, device and medium based on multi-round simulation optimization

Abstract

The application relates to the technical field of wireless communication, and discloses a network switching method, system, device and medium based on multi-round simulation optimization, which comprises the following steps: acquiring real-time network state information to perform flow prediction, obtaining an overall flow prediction result; acquiring a load value from the overall flow prediction result, and if the load value is higher than a load threshold value, determining a high-load access point and redirecting flow to a nearby access point to obtain an initial network configuration; performing multi-round switching simulation to generate a candidate sequence, and if a preset load balancing condition is met, obtaining a preliminary switching scheme and calculating a reward value, selecting a switching scheme with the highest reward value as an optimized switching scheme; executing the optimized switching scheme, recording access point states; performing abnormality detection, and if there is an abnormal access point, adjusting a prediction method and predicting overall flow, optimizing network configuration according to a prediction result, and obtaining a final network switching configuration. The method can efficiently and stably switch wireless networks in a complex network environment.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, and in particular to a network handover method, system, device, and medium based on multi-round simulation optimization. Background Technology

[0002] Currently, network communication is widely used in scenarios such as cloud computing, the Internet of Things and edge computing, and intelligent transportation. To ensure business continuity and service quality, terminals or service flows need to frequently switch networks between multiple access points, multiple links, and multiple operating domains. With the expansion of network scale and the intensification of heterogeneity, network states exhibit highly dynamic, highly uncertain, and strongly coupled characteristics. How to achieve global-level collaborative optimization switching under distributed conditions has become a key challenge.

[0003] In a wireless internet access (WiFi) scenario, network switching typically relies on centralized control and static threshold determination. Each access point collects link quality and traffic load data at fixed intervals and reports it to the central controller. The controller sorts candidate links by fixed weights and selects the target access point with higher priority. After a consistency check, the central controller issues a unified switching command.

[0004] However, fixed-period sampling struggles to capture instantaneous fluctuations in a timely manner, static thresholds lack adaptability to changes in service type and environment, fixed weights fail to reflect real-time service differences, and global state synchronization is delayed. In summary, existing technologies suffer from the inability to efficiently and stably switch wireless networks in complex network environments. Summary of the Invention

[0005] This invention provides a network handover method, system, device, and medium based on multi-round simulation optimization to achieve efficient and stable handover of wireless networks in complex network environments.

[0006] Firstly, in order to solve the above-mentioned technical problems, the present invention provides a network handover method based on multi-round simulation optimization, comprising:

[0007] Real-time network status information is acquired and features are extracted to obtain network status features;

[0008] Each distributed access point performs partial traffic prediction based on the network state characteristics to obtain partial traffic prediction results, and aggregates the partial traffic prediction results to obtain the overall traffic prediction result.

[0009] The load value of each access point is obtained from the overall traffic prediction result. If the load value is higher than the preset load threshold, it is determined to be a high-load access point, and the traffic of the high-load access point is redirected to a nearby access point to obtain the initial network configuration.

[0010] Based on the initial network configuration, perform multiple rounds of handover simulation to generate a candidate sequence containing handover actions. If the candidate sequence containing handover actions satisfies the preset load balancing conditions, a preliminary handover scheme is obtained. The reward value of the preliminary handover scheme is calculated according to the preset reward function. The handover scheme with the highest reward value is selected to obtain the optimized handover scheme.

[0011] The optimized switching scheme is distributed to each access point to obtain the optimized network configuration, and the adjusted access point status is recorded.

[0012] Anomaly detection is performed based on the adjusted access point status. If an abnormal access point is found, the prediction method is adjusted and the overall traffic is predicted to obtain an updated overall traffic prediction result.

[0013] Based on the updated overall traffic prediction results, the optimized network configuration is optimized to obtain the final network switching configuration.

[0014] In one optional implementation, the step of acquiring real-time network state information and extracting features to obtain network state features includes:

[0015] Obtain real-time network status information;

[0016] The real-time network status information is segmented using a time window to obtain segmented traffic data;

[0017] After filtering the segmented traffic data, feature vectors are calculated to obtain traffic pattern features;

[0018] The network state features are obtained by performing pattern recognition on the traffic pattern features using preset normal pattern features.

[0019] In one optional implementation, each distributed access point performs partial traffic prediction based on the network state characteristics to obtain partial traffic prediction results, and aggregates the partial traffic prediction results to obtain an overall traffic prediction result, including:

[0020] The network state features are encrypted to obtain encrypted network state features;

[0021] Each distributed access point makes predictions based on the encrypted network state characteristics to obtain partial traffic prediction results;

[0022] The overall traffic prediction result is obtained by weighting and averaging the partial traffic prediction results.

[0023] In one optional implementation, the step of obtaining the load value of each access point from the overall traffic prediction result, and determining that the load value is higher than a preset load threshold as a high-load access point, and redirecting the traffic of the high-load access point to a nearby access point to obtain the initial network configuration, includes:

[0024] Based on the overall traffic prediction results, the load values ​​of each access point are obtained to obtain the local load values;

[0025] If the local load value is higher than the preset load value threshold, it is determined to be a high-load access point, and a nearby access point is determined according to the preset switching rules. The traffic of the high-load access point is redirected to the nearby access point to obtain the initial network configuration.

[0026] In one optional implementation, the step of performing multiple rounds of handover simulation based on the initial network configuration to generate a candidate sequence containing handover actions, and if the candidate sequence containing handover actions satisfies a preset load balancing condition, obtaining a preliminary handover scheme, calculating the reward value of the preliminary handover scheme according to a preset reward function, and selecting the handover scheme with the highest reward value to obtain an optimized handover scheme, includes:

[0027] Simulate multiple rounds of traffic allocation scenarios based on the initial network configuration, and generate a candidate sequence containing switching actions;

[0028] Extract the load distribution of the candidate sequence and evaluate it. If the load distribution meets the preset load balancing conditions, then determine the candidate sequence as a preliminary switching scheme.

[0029] The reward value of the initial switching scheme is calculated according to the preset reward function, and the scheme with the highest reward value is selected as the optimized switching scheme.

[0030] In one optional implementation, the step of performing anomaly detection based on the adjusted access point status, and if an abnormal access point exists, adjusting the prediction method and predicting the overall traffic to obtain an updated overall traffic prediction result, includes:

[0031] The adjusted access point status is compared with the preset access point status template. If the access point status is abnormal, a list of abnormal access points is obtained.

[0032] Obtain real-time data of the abnormal access point list, adjust the prediction method to predict the overall traffic, and obtain the updated overall traffic prediction result.

[0033] In one optional implementation, optimizing the optimized network configuration based on the updated overall traffic prediction results to obtain the final network switching configuration includes:

[0034] Based on the updated overall traffic prediction results, the contribution value of each access point is calculated using a weighted average.

[0035] Based on the contribution value of each access point, the updated optimized network configuration is determined and distributed to each access point to obtain the final network switching configuration.

[0036] Secondly, the present invention provides a network handover system based on multi-round simulation optimization, comprising:

[0037] The data acquisition module is used to acquire real-time network status information and extract features to obtain network status features;

[0038] The traffic prediction module is used for each distributed access point to perform partial traffic prediction based on the network status characteristics, obtain partial traffic prediction results, and aggregate the partial traffic prediction results to obtain the overall traffic prediction result.

[0039] The load configuration module is used to obtain the load value of each access point from the overall traffic prediction result. If the load value is higher than the preset load threshold, it is determined to be a high-load access point, and the traffic of the high-load access point is redirected to a nearby access point to obtain the initial network configuration.

[0040] The handover strategy generation module is used to perform multiple rounds of handover simulation based on the initial network configuration, generate a candidate sequence containing handover actions, and if the candidate sequence containing handover actions meets the preset load balancing conditions, a preliminary handover scheme is obtained. The reward value of the preliminary handover scheme is calculated according to the preset reward function, and the handover scheme with the highest reward value is selected to obtain the optimized handover scheme.

[0041] The access point status recording module is used to distribute the optimized switching scheme to each access point, obtain the optimized network configuration, and record the adjusted access point status.

[0042] An anomaly detection module is used to perform anomaly detection based on the adjusted access point status. If an abnormal access point exists, the prediction method is adjusted and the overall traffic is predicted to obtain an updated overall traffic prediction result.

[0043] The network configuration module is used to optimize the optimized network configuration based on the updated overall traffic prediction results to obtain the final network switching configuration.

[0044] Thirdly, the present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the network switching method based on multi-round simulation optimization as described above.

[0045] Fourthly, the present invention also provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the network handover method based on multi-round simulation optimization described above.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] (1) This invention performs partial traffic prediction based on encrypted network state features at distributed access points, and generates an overall traffic prediction result by weighted aggregation of the prediction results of each access point. While ensuring data privacy and security, it avoids the defect of single access point prediction being susceptible to local anomalies, making the global prediction result more stable and comprehensive, improving the identification accuracy of high-load access points, providing a solid data foundation for subsequent traffic scheduling and switching optimization, and realizing accurate characterization and efficient prediction capability of network operation status from a global perspective.

[0048] (2) This invention obtains refined network state features by acquiring real-time network state information and performing time window segmentation, feature filtering, and pattern recognition. It aligns, smooths, and removes noise from multi-source heterogeneous raw data, making the input data continuous and stable, and effectively avoiding the failure problem of static thresholding methods in dynamic environments. The high-reliability feature vector formed on this basis provides a more accurate foundation for traffic prediction and load identification, realizing a fine perception of complex network states.

[0049] (3) This invention conducts multi-round handover simulations based on the initial network configuration and uses a multi-dimensional reward function to quantify and score candidate sequences, comprehensively evaluating the potential differences in topological constraints and resource consumption among different schemes. This results in the selected scheme having higher executability and load balancing when it is deployed and executed. At the same time, after the strategy is executed, by recording the status of access points in real time and detecting anomalies, deviation patterns are promptly fed back to the prediction model for adaptive adjustment, enabling the scheme to dynamically respond to fluctuations and anomalies during operation. Through this continuous derivation process from simulation evaluation to execution feedback and then to prediction correction, the handover strategy naturally maintains balance, stability and efficiency in actual operation, realizing the optimized utilization of network resources and the continuous robustness of the system. Attached Figure Description

[0050] Figure 1 This is a schematic diagram of the network handover method based on multi-round simulation optimization provided in the first embodiment of the present invention;

[0051] Figure 2 This is a schematic diagram of a network switching system structure based on multi-round simulation optimization provided in the second embodiment of the present invention. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Reference Figure 1 The first embodiment of the present invention provides a network handover method based on multi-round simulation optimization, including the following steps:

[0054] S11, Obtain real-time network status information and extract features to obtain network status features;

[0055] S12, each distributed access point performs partial traffic prediction based on the network state characteristics to obtain partial traffic prediction results, and aggregates the partial traffic prediction results to obtain the overall traffic prediction result;

[0056] S13, obtain the load value of each access point from the overall traffic prediction result. If the load value is higher than the preset load threshold, it is determined to be a high-load access point, and the traffic of the high-load access point is redirected to a nearby access point to obtain the initial network configuration;

[0057] S14. Perform multiple rounds of handover simulation based on the initial network configuration to generate a candidate sequence containing handover actions. If the candidate sequence containing handover actions satisfies the preset load balancing conditions, a preliminary handover scheme is obtained. Calculate the reward value of the preliminary handover scheme based on the preset reward function, and select the handover scheme with the highest reward value to obtain an optimized handover scheme.

[0058] S15, the optimized switching scheme is sent to each access point to obtain the optimized network configuration and the adjusted access point status is recorded.

[0059] S16, perform anomaly detection based on the adjusted access point status. If there is an abnormal access point, adjust the prediction method and predict the overall traffic to obtain the updated overall traffic prediction result.

[0060] S17. Based on the updated overall traffic prediction results, optimize the optimized network configuration to obtain the final network switching configuration.

[0061] In step S11, the process of acquiring real-time network state information and extracting features to obtain network state features includes:

[0062] Obtain real-time network status information;

[0063] The real-time network status information is segmented using a time window to obtain segmented traffic data;

[0064] After filtering the segmented traffic data, feature vectors are calculated to obtain traffic pattern features;

[0065] The network state features are obtained by performing pattern recognition on the traffic pattern features using preset normal pattern features.

[0066] It should be noted that network status information, including timestamps, source IPs, destination IPs, port numbers, and packet sizes, is collected in real time through sniffer cameras deployed on routers and switches. The network status information is divided into multiple time slices, each with a 10-second time window. Each slice contains all traffic metrics for that time period, including total access point traffic, number of packets, port receive and send traffic, port packet count, and packet size, resulting in segmented traffic data. The 10-second time window effectively reflects instantaneous fluctuations in network status, supporting timely traffic prediction and load balancing, while avoiding excessive data volume and noise interference from overly short windows. It also covers typical short-term traffic fluctuation cycles, facilitating the calculation of average load, peak values, and variance, forming stable and reliable segmented traffic data. This provides a highly reliable input foundation for subsequent traffic feature extraction, pattern recognition, and intelligent switching.

[0067] To ensure the effectiveness of segmented traffic data, network packets smaller than 50 bytes are filtered out. These packets are typically heartbeat packets, probe packets, or fragmented transmissions, which contribute little to network traffic analysis and are prone to generating high-frequency noise. Setting 50 bytes as the threshold is because the vast majority of real-world business traffic packets exceed this size, while packets smaller than the threshold are mainly for maintaining connections or are fragmented. Removing these packets significantly reduces the impact of abnormal fluctuations on statistical characteristics, improves the stability of traffic pattern extraction and the accuracy of prediction models, thus providing a reliable and highly dependable data foundation for subsequent traffic analysis and intelligent switching.

[0068] Meanwhile, there may still be some control packets with important business significance between 40 and 50 bytes, such as domain name resolution requests and network time synchronization messages. Therefore, an exception detection mechanism is introduced when performing threshold filtering. Specifically, the header information of the data packet is identified. When the protocol field of the small packet is detected to belong to DNS, NTP, DHCP, ICMP, IKE / ISAKMP or ARP, it is not directly discarded, but marked and retained.

[0069] Principal component analysis (PCA) is used to calculate the flow feature vectors for the filtered segmented flow information. Specifically, firstly, mean centering is performed on each flow indicator for each time slice. This involves calculating the average value of the indicator across all time slices and subtracting this average value from the indicator value in each time slice, centering the data around zero. Next, variance normalization is performed on the centered data. This involves dividing the value of each indicator by its standard deviation across all time slices, ensuring that indicators with different dimensions and numerical ranges have equal weight in subsequent analyses. After standardization, the data from each time slice are combined into a matrix, and the correlation between the indicators is calculated. Specifically, for each pair of indicators, their values ​​across all time slices are extracted, and their deviations from their respective averages are calculated. Then, the product of the deviations of the two indicators is averaged across all time slices to obtain the correlation of the indicator pair. Finally, the eigenvalues ​​and corresponding eigenvectors of the correlation matrix are calculated sequentially. Each eigenvalue reflects the magnitude of the variance in its corresponding direction. Among all eigenvalues, the direction in which a single eigenvalue contributes more than 5% of the total variance is selected as the principal component, and sorted from high to low according to the contribution rate; then the variance contribution of each principal component is calculated cumulatively, and the principal component with a cumulative variance contribution of 80% is selected to form the dimensionality-reduced feature vector. The projection values ​​of the data in each time slice in the directions of these principal components constitute the final feature vector, thus obtaining the flow pattern features.

[0070] In this process, directions whose individual eigenvalues ​​contribute more than 5% of the total variance are selected as principal components. This threshold is determined based on experience and actual network traffic analysis results: directions with less than 5% contribution have a small contribution to the overall data variation and limited significance for retention. Discarding these directions with small variance reduces noise interference and improves the stability and representativeness of the eigenvectors. Subsequently, the principal components are sorted from highest to lowest contribution rate, and the variance contribution is calculated cumulatively. Principal components with a cumulative variance contribution of 80% are selected as the dimensionality-reduced eigenvectors. This proportion can significantly reduce dimensionality while retaining most of the core variation characteristics of the data, ensuring that the eigenvectors can fully reflect network traffic patterns and provide a reliable foundation for subsequent traffic prediction and intelligent switching.

[0071] The absolute difference between the traffic pattern characteristics and the preset normal pattern characteristics is calculated for each dimension. Then, all absolute differences are divided by the corresponding value of the normal pattern characteristic, and finally multiplied by 100 to obtain the deviation percentage. The average deviation percentage is calculated for all obtained deviation percentages to obtain the deviation value between the traffic pattern characteristics and the normal pattern characteristics. The preset normal pattern characteristics are obtained based on historical network operation data and statistics from typical business scenarios. To comprehensively cover the network traffic characteristics under different load conditions and avoid bias caused by short-term observations, a week's worth of traffic feature vectors is collected when the network is stable and free from abnormal interference. The mean and variance of each indicator are calculated over time to form a feature vector representing the normal operating mode of the network. Network traffic typically exhibits obvious periodicity and daily regularity; for example, load patterns differ between weekdays and weekends, and between daytime and nighttime. Therefore, selecting a complete week's worth of data for calculation can simultaneously include peak periods, off-peak periods, and business fluctuations, ensuring that the statistically calculated normal pattern characteristics comprehensively reflect the overall behavior of the network within a typical operating cycle.

[0072] If the deviation exceeds 10%, it is judged as an abnormal state, and the current traffic pattern characteristics are recorded as network status characteristics. The deviation threshold of 10% is based on a comprehensive consideration of statistical analysis of historical network operation data and the patterns of business traffic characteristics. When the network is stable and free from abnormal interference, the percentage deviation between traffic characteristics and normal pattern characteristics for different time slots within a week is calculated. The results show that the deviation values ​​for the vast majority of time slots are below 10%, indicating that network load fluctuations and traffic patterns are within the normal range. When the deviation exceeds 10%, it means that the current traffic characteristics significantly deviate from the normal pattern in multiple indicators, exceeding the historical normal fluctuation range, and has a high probability of corresponding to abnormal traffic or sudden events. Therefore, using 10% as the threshold can effectively distinguish between normal fluctuations and abnormal states, ensuring that the deviation value judgment is sensitive to abnormal changes but not misjudged by slight fluctuations, thus providing a reliable basis for subsequent network status assessment, traffic prediction, and intelligent switching.

[0073] In step S12, each distributed access point performs partial traffic prediction based on the network state characteristics to obtain partial traffic prediction results, and aggregates the partial traffic prediction results to obtain an overall traffic prediction result, including:

[0074] The network state features are encrypted to obtain encrypted network state features;

[0075] Each distributed access point makes predictions based on the encrypted network state characteristics to obtain partial traffic prediction results;

[0076] The overall traffic prediction result is obtained by weighting and averaging the partial traffic prediction results.

[0077] It should be noted that, to prevent network state characteristics from being intercepted or tampered with during transmission, the TLS protocol is used to encrypt the network state characteristics, generating encrypted network state characteristics. Specifically, a TLS secure connection is established between distributed access points, and this TLS secure connection uses the RSA algorithm for key exchange and authentication. In the RSA algorithm implementation, firstly, an RSA public key and private key pair are generated, and the public key is sent to the distributed access points via a digital certificate. After receiving the public key, the access points use it to encrypt the generated symmetric encryption session key. Finally, the corresponding RSA private key is used to decrypt the ciphertext to recover the session key, thus completing the secure key exchange. Subsequently, the network state characteristics are encrypted using the key before transmission, converting the original feature vector into ciphertext, ensuring that even if the data is intercepted during transmission, it cannot be parsed or tampered with. After establishing a TLS connection, the receiving end uses the corresponding key to decrypt the ciphertext, restoring the original network state characteristics for subsequent traffic prediction and intelligent handover calculations, thereby achieving end-to-end secure transmission of network state data.

[0078] The encrypted network state features are distributed to each distributed access point. Each access point locally trains the model using a federated learning algorithm to generate the overall traffic prediction model. Specifically, initial weights are first generated using a uniform distribution to initialize the overall traffic prediction model, and the global model parameters are then distributed to each access point. Each access point then uses its own encrypted network state features to perform multiple rounds of gradient optimization to update the model weights locally, uploading only the encrypted weight updates to the center without transmitting the original data to ensure privacy and security. The weight updates uploaded by each access point are then weighted and aggregated to form new global model parameters, which are then distributed to the access points again. This process iterates until the mean absolute error is less than 0.01 or the maximum number of iterations (100 rounds) is reached. The reason for choosing 0.01 as the threshold for mean absolute error is that when the error is controlled within 0.01, the accuracy of the prediction model is sufficient to support traffic scheduling and intelligent switching. Further reducing the threshold will have limited performance improvement but will significantly increase the computational cost. The choice of 100 iterations as the upper limit takes into account the balance between convergence speed and resource consumption. At the same time, referring to mainstream federated learning practices, in most cases, convergence to the target accuracy can be achieved in 50 to 80 iterations. Setting 100 iterations can ensure that the model converges fully and avoid unnecessary computational waste.

[0079] The distributed access points employ a two-layer feedforward neural network model. The input is a dimensionality-reduced traffic feature vector, and the output is a traffic prediction value for a future period. Each distributed access point obtains its corresponding weight coefficient by dividing its local training data by the total training data. The partial traffic predictions from each access point are then weighted and averaged to obtain the global traffic prediction result. Specifically, based on the prediction results of each access point, the corresponding weight coefficient is obtained by dividing the amount of training data used by each access point by the total training data. For access points with missing prediction results, the average of the previous and subsequent predictions is used to fill in the gaps. For each access point's local prediction result, the predicted value is multiplied by the access point's weight and then summed sequentially to complete the weighted calculation and obtain the overall traffic prediction result.

[0080] In step S13, the process of obtaining the load value of each access point from the overall traffic prediction result, and determining that the load value is higher than a preset load threshold as a high-load access point, and redirecting the traffic of the high-load access point to a nearby access point to obtain the initial network configuration, includes:

[0081] Based on the overall traffic prediction results, the load values ​​of each access point are obtained to obtain the local load values;

[0082] If the local load value is higher than the preset load value threshold, it is determined to be a high-load access point, and a nearby access point is determined according to the preset switching rules. The traffic of the high-load access point is redirected to the nearby access point to obtain the initial network configuration.

[0083] It should be noted that after training and converging the overall traffic prediction model, the model is used to predict the overall traffic trend of the distributed network. Combined with the state characteristic data of each distributed access point, the global prediction result is decomposed and mapped as needed to calculate the predicted load value for each distributed access point. Specifically, the historical traffic of each point is first summed to obtain the total historical traffic. Then, the historical traffic of each access point is divided by the total historical traffic to obtain the historical traffic percentage as the allocation ratio. This allocation ratio is then multiplied by the overall traffic prediction result to obtain the local predicted load value for each access point. If the load value exceeds 20% of the normal load of the access point, the access point is determined to be a high-load access point. The 20% threshold was determined through a large number of historical operating samples and stress test experiments. In multiple experiments, when the access point load exceeded 20% of the normal baseline value, its average processing latency and packet loss rate increased significantly, and resource utilization efficiency decreased significantly. Therefore, 20% was selected as a reasonable threshold for high load determination to balance the sensitivity and tolerance of anomaly detection.

[0084] The non-high-load access points directly connected to the high-load access point are identified as nearby access points. Determining nearby access points according to preset switching rules refers to dynamically selecting access points that are directly connected to the high-load access point in the network, or logically adjacent in network topology and transmission quality, by acquiring real-time status characteristics (including bandwidth, latency, and load) through the monitoring function of distributed access points.

[0085] In one implementation, in a 5G base station network, neighboring access points are determined by calculating the current bandwidth utilization and end-to-end latency. The preset handover rule is that bandwidth utilization is less than 80% and end-to-end latency is less than 50ms, dynamically identifying temporary access points. The 80% threshold is based on historical data analysis, which found that when link bandwidth utilization consistently exceeds 80%, the risk of congestion and a sharp increase in queuing latency increases significantly. Therefore, a 20% bandwidth margin is reserved to provide a buffer for sudden traffic peaks. The 50ms threshold is set because the perception and decision-making requirements in real-world scenarios are extremely stringent; end-to-end latency must be controlled within tens of milliseconds. The 50ms threshold is also a commonly used and demanding QoS target.

[0086] The excess load is divided by the number of nearby access points and then redirected to each nearby access point sequentially to obtain the initial network configuration. The normal load value of each access point is based on historical monitoring data obtained under long-term stable operation. Specifically, under normal network conditions, traffic and processing volume data of each access point are collected over a complete one-week period. The mean and standard deviation of the access point's load are calculated, and the mean is used as the baseline value for the access point's normal load. If the real-time load value of an access point exceeds this baseline value by 20%, the access point is considered a high-load access point. Using a complete one-week period of data for calculation can simultaneously include peak periods, off-peak periods, and service fluctuations, ensuring that the statistically calculated normal pattern characteristics comprehensively reflect the overall network behavior within a typical operating cycle.

[0087] In step S14, the process involves performing multiple rounds of handover simulation based on the initial network configuration to generate a candidate sequence containing handover actions. If the candidate sequence containing handover actions satisfies a preset load balancing condition, a preliminary handover scheme is obtained. The reward value of the preliminary handover scheme is calculated based on a preset reward function, and the handover scheme with the highest reward value is selected to obtain an optimized handover scheme, including:

[0088] Simulate multiple rounds of traffic allocation scenarios based on the initial network configuration, and generate a candidate sequence containing switching actions;

[0089] Extract the load distribution of the candidate sequence and evaluate it. If the load distribution meets the preset load balancing conditions, then determine the candidate sequence as a preliminary switching scheme.

[0090] The reward value of the initial switching scheme is calculated according to the preset reward function, and the scheme with the highest reward value is selected as the optimized switching scheme.

[0091] It should be noted that the initial network configuration is used as the initial load distribution, and multiple rounds of traffic allocation simulation are conducted. Specifically, a genetic algorithm is used for multiple rounds of iterative optimization to search for the optimal traffic switching scheme. The population size is set at 50 individuals to strike a balance between computational efficiency and search space diversity; the maximum number of iterations is set at 100 generations, and if the optimal solution does not show significant improvement for 20 consecutive generations (the improvement threshold is set at 1%, i.e., when there is no significant improvement, the improvement is less than or equal to 1%), the process is terminated early to balance optimization effect and computational cost; a tournament selection method is used to retain high-fitness individuals and maintain population diversity, and the tournament size is set to 3; binary crossover is simulated to perform crossover operations, with the crossover probability set to 0.8 and the distribution exponent set to 5, to effectively explore the solution space; multinomial mutation is used to perform mutation operations, with the mutation probability set to 0.1 and the distribution exponent set to 10, to introduce random perturbation and avoid premature convergence; the top two individuals with the highest fitness in each generation are retained and directly enter the next generation to ensure that the optimal solution is not lost. During the optimization process, allocation simulation is performed based on the historical average traffic. Historical traffic averages are calculated using traffic data from a week, which can include peak periods, off-peak periods, and business fluctuations, thus ensuring that statistical calculations can comprehensively reflect the overall behavior of the network within a typical operating cycle.

[0092] The allocated traffic value for each access point is added to the initial load distribution to obtain the distributed load distribution. The load value of each access point is then checked to see if it exceeds or falls below 20% of the average load over the past week. The 20% threshold is chosen because when the actual load of an access point exceeds its normal load level by 20%, it can easily lead to overload risks such as processing delays and queue congestion; conversely, when the actual load is below 20% of the normal load, it indicates insufficient resource utilization, potentially resulting in idle computing power and a decrease in overall efficiency. Therefore, the 20% threshold effectively prevents performance degradation due to access point overload and also prevents waste caused by unbalanced resource allocation, ensuring a balance between stability and utilization. If the load exceeds or falls below 20% of the average load over the past week, the current candidate sequence is discarded; if it does not exceed or falls below 20% of the average load over the past week, the requirement is met, the current candidate sequence is recorded, and a preliminary switching plan is obtained.

[0093] The preset reward function calculates the reward value based on the load balancing of the handover scheme. Specifically, it calculates the average load value of each distributed access point, subtracts the average load value from the load value of each access point, and takes the absolute value to obtain the access point deviation. These access point deviations are accumulated to calculate the total global load deviation. The unit of the total global load deviation is the same as the access point load value, i.e., in Mbps of load traffic. The reciprocal of the total global load deviation is multiplied by 100 to convert it to a percentage, which is then used as the reward value. The handover scheme with the highest reward value is selected as the optimized handover scheme.

[0094] In step S15, the optimized switching scheme is sent to each access point to obtain the optimized network configuration and the adjusted access point status is recorded.

[0095] It should be noted that after the optimized switching plan is determined, it is distributed to each distributed access point. Each access point dynamically adjusts the received service traffic according to the proportional parameters allocated to it in the plan, i.e., receiving, forwarding, or processing traffic according to the allocated proportion, thereby achieving global load balancing. During the execution of the switching strategy, the access points continuously collect and store their status information to track and evaluate the performance. This status information includes processing latency, response time, throughput, and bandwidth utilization.

[0096] In step S16, the anomaly detection is performed based on the adjusted access point status. If an abnormal access point exists, the prediction method is adjusted and the overall traffic is predicted to obtain an updated overall traffic prediction result, including:

[0097] The adjusted access point status is compared with the preset access point status template. If the access point status is abnormal, a list of abnormal access points is obtained.

[0098] Obtain real-time data of the abnormal access point list, adjust the prediction method to predict the overall traffic, and obtain the updated overall traffic prediction result.

[0099] It should be noted that when calculating the adjusted access point status, it is compared item by item with the preset access point status template, and the deviation value is calculated. If the deviation value exceeds 10%, the access point is determined to have an abnormal status and is added to the list of abnormal access points. Specifically, the preset access point status template is obtained by averaging multi-dimensional status information collected during historical normal operation, which can objectively reflect the baseline operating level of the access point. The deviation value is calculated as follows: the absolute difference between the current access point status and the template value is calculated item by item, divided by the corresponding item in the access point status template, and the average value is taken as the relative deviation rate. If the relative deviation rate is greater than 10%, it indicates that the operating status of the access point has significantly deviated from the normal level.

[0100] It is worth noting that the 10% threshold was determined based on extensive operational experience and statistical analysis. Practice has shown that when the access point deviation exceeds this threshold, the probability of performance degradation or abnormal failure increases significantly. Therefore, this threshold can reduce the false positive rate while ensuring detection sensitivity.

[0101] When an access point is deemed abnormal, its real-time data is collected and incorporated into the corresponding historical dataset. The traffic prediction model for that access point is then retrained iteratively until the mean absolute error (MAE) on the validation set is less than 0.01 or the maximum number of iterations (100) is reached. The threshold of 0.01 is chosen because when the error is below 0.01, the accuracy of the prediction model is sufficient to support traffic scheduling and intelligent switching. Further reducing the threshold offers limited performance improvement but significantly increases computational overhead. The iteration limit of 100 rounds balances convergence speed and resource consumption. Referring to mainstream federated learning practices, convergence to the target accuracy is usually achieved within 50-80 rounds; setting 100 rounds ensures sufficient model convergence and avoids unnecessary computational waste. After the partial traffic prediction model converges, the prediction results of each access point are aggregated as a weighting factor based on the traffic allocation ratio of each access point in the scheduling scheme. Specifically, the allocation ratio of each access point is used as a weight to sum the local prediction results, thus obtaining the updated overall traffic prediction result.

[0102] In step S17, optimizing the optimized network configuration based on the updated overall traffic prediction results to obtain the final network switching configuration includes:

[0103] Based on the updated overall traffic prediction results, the contribution value of each access point is calculated using a weighted average.

[0104] Based on the contribution value of each access point, the updated optimized network configuration is determined and distributed to each access point to obtain the final network switching configuration.

[0105] It should be noted that after generating the updated overall traffic prediction result, the traffic prediction value of each access point is first extracted from the overall traffic prediction result. The ratio of the access point's traffic prediction value to the total overall traffic prediction value is then calculated to obtain the initial contribution value of each access point, which characterizes the relative role of that access point in the overall service traffic carrying capacity. Based on this, the initial contribution value is multiplied by the preset access point traffic allocation ratio in the optimized switching scheme to obtain the product factor combining the prediction result and the policy allocation. The preset access point traffic allocation ratio is derived from 100 rounds of traffic allocation simulation experiments. By evaluating the results of each round of simulation, the best-performing optimized switching scheme is selected. The access point traffic allocation ratio corresponding to this optimized switching scheme is the preset access point traffic allocation ratio. Using this ratio as the access point traffic allocation ratio achieves stable and relatively optimal results in all subsequent simulations and actual scheduling. Subsequently, the product factor corresponding to each access point is normalized to the sum of the product factors of all access points. That is, the product factor of a single access point is divided by the sum of the product factors of all access points, which is used as the updated access point traffic allocation ratio to obtain the final network handover configuration.

[0106] In summary, this invention conducts multiple rounds of handover simulations based on the initial network configuration, recording the load distribution, handover cost, and latency changes of each candidate scheme. A multi-dimensional reward function is used to quantify and score the candidate sequences, ensuring that the optimal scheme balances load balancing and resource consumption. After policy execution, access point status is continuously collected and compared with a preset template. Deviation patterns are identified through anomaly detection, and the results are fed back to the prediction model for adaptive adjustment, enabling candidate schemes to dynamically correct themselves to cope with network fluctuations and abnormal access points. Multiple rounds of simulation reveal topological constraints and cumulative costs, the reward function unifies multi-objective measurement standards, and execution feedback and anomaly correction form a closed-loop optimization, naturally harmonizing the handover strategy in terms of executability, load balancing, and long-term stability. Finally, the handover scheme, verified through multi-level simulation, quantitative evaluation, and closed-loop correction, is applied, achieving efficient and stable network handover in complex network environments.

[0107] Reference Figure 2 The second embodiment of the present invention provides a network handover system based on multi-round simulation optimization, comprising:

[0108] The data acquisition module is used to acquire real-time network status information and extract features to obtain network status features;

[0109] The traffic prediction module is used for each distributed access point to perform partial traffic prediction based on the network status characteristics, obtain partial traffic prediction results, and aggregate the partial traffic prediction results to obtain the overall traffic prediction result.

[0110] The load configuration module is used to obtain the load value of each access point from the overall traffic prediction result. If the load value is higher than the preset load threshold, it is determined to be a high-load access point, and the traffic of the high-load access point is redirected to a nearby access point to obtain the initial network configuration.

[0111] The handover strategy generation module is used to perform multiple rounds of handover simulation based on the initial network configuration, generate a candidate sequence containing handover actions, and if the candidate sequence containing handover actions meets the preset load balancing conditions, a preliminary handover scheme is obtained. The reward value of the preliminary handover scheme is calculated according to the preset reward function, and the handover scheme with the highest reward value is selected to obtain the optimized handover scheme.

[0112] The access point status recording module is used to distribute the optimized switching scheme to each access point, obtain the optimized network configuration, and record the adjusted access point status.

[0113] An anomaly detection module is used to perform anomaly detection based on the adjusted access point status. If an abnormal access point exists, the prediction method is adjusted and the overall traffic is predicted to obtain an updated overall traffic prediction result.

[0114] The network configuration module is used to optimize the optimized network configuration based on the updated overall traffic prediction results to obtain the final network switching configuration.

[0115] It should be noted that the network handover system based on multi-round simulation optimization provided in this embodiment of the invention is used to execute all the process steps of the network handover method based on multi-round simulation optimization in the above embodiments. The working principles and beneficial effects of the two are one-to-one, so they will not be described again.

[0116] This invention also provides an electronic device. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a data acquisition program. When the processor executes the computer program, it implements the steps in the various embodiments of the network handover method based on multi-round simulation optimization described above, for example... Figure 1 The step S11 shown. Alternatively, when the processor executes the computer program, it implements the functions of each module in the above-described device embodiments, such as the data acquisition module.

[0117] For example, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0118] The electronic device may be a desktop computer, laptop, handheld computer, or smart tablet, etc. The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. It may include more or fewer components than described above, or combine certain components, or different components. For example, the electronic device may also include input / output devices, network access devices, buses, etc.

[0119] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the electronic device, connecting all parts of the electronic device via various interfaces and lines.

[0120] The memory can be used to store the computer programs and modules. The processor implements various functions of the electronic device by running or executing the computer programs and modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0121] If the modules integrated into the electronic device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0122] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0123] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A network handover method based on multi-round simulation optimization, characterized in that, include: Real-time wireless network status information is acquired and features are extracted to obtain network status features; Each distributed access point performs partial traffic prediction based on the network state characteristics to obtain partial traffic prediction results, and aggregates the partial traffic prediction results to obtain the overall traffic prediction result. The load value of each access point is obtained from the overall traffic prediction result. If the load value is higher than the preset load threshold, it is determined to be a high-load access point, and the traffic of the high-load access point is redirected to a nearby access point to obtain the initial network configuration. Based on the initial network configuration, perform multiple rounds of handover simulation to generate a candidate sequence containing handover actions. If the candidate sequence containing handover actions satisfies the preset load balancing conditions, a preliminary handover scheme is obtained. The reward value of the preliminary handover scheme is calculated according to the preset reward function. The handover scheme with the highest reward value is selected to obtain the optimized handover scheme. The optimized switching scheme is distributed to each access point to obtain the optimized network configuration, and the adjusted access point status is recorded. Anomaly detection is performed based on the adjusted access point status. If an abnormal access point is found, the prediction method is adjusted and the overall traffic is predicted to obtain an updated overall traffic prediction result. Based on the updated overall traffic prediction results, the optimized network configuration is optimized to obtain the final network switching configuration; The step of acquiring real-time network status information and extracting features to obtain network status features includes: Obtain real-time network status information; The real-time network status information is segmented using a time window to obtain segmented traffic data; After filtering the segmented traffic data, feature vectors are calculated to obtain traffic pattern features; The network state features are obtained by performing pattern recognition on the traffic pattern features using preset normal pattern features; The process involves filtering the segmented traffic data and then calculating feature vectors to obtain traffic pattern features, including: The principal component analysis algorithm is used to calculate the flow feature vector of the filtered segmented flow information: First, the mean centering process of each flow indicator in each time slice is performed, and the average value of each indicator in each time slice is subtracted from the mean value to make the data centered at zero. The variance normalization process is performed on the centered data, the data of each time slice is combined into a matrix form, and the correlation between each indicator is calculated. The eigenvalues ​​and corresponding eigenvectors of the correlation matrix are calculated in turn. Among all eigenvalues, the directions in which a single eigenvalue contributes more than 5% to the total variance are selected as principal components and sorted from high to low according to their contribution rate. Then, the variance contribution of each principal component is calculated cumulatively, and the principal components with a cumulative variance contribution of 80% are selected to form the dimensionality-reduced feature vector. The projection values ​​of the data in each time slice in these principal component directions constitute the final feature vector, thus obtaining the flow pattern features. Each distributed access point performs partial traffic prediction based on the network state characteristics to obtain partial traffic prediction results, and aggregates these partial traffic prediction results to obtain an overall traffic prediction result, including: The network state features are encrypted to obtain encrypted network state features; Each distributed access point makes predictions based on the encrypted network state characteristics to obtain partial traffic prediction results; The overall traffic prediction result is obtained by weighted averaging the partial traffic prediction results. The model structure used by each distributed access point is a feedforward neural network with two hidden layers. The input is the dimensionality-reduced traffic feature vector, and the output is the traffic prediction value for a future period of time. Each distributed access point obtains the corresponding weight coefficient by dividing the amount of local training data used by the total amount of training data. The partial traffic prediction results of each access point are combined with the access point weights and weighted averaged to obtain the global traffic prediction result.

2. The network handover method based on multi-round simulation optimization according to claim 1, characterized in that, The process of obtaining the load value of each access point from the overall traffic prediction result, and determining a high-load access point if the load value is higher than a preset load threshold, and redirecting the traffic of the high-load access point to a nearby access point to obtain the initial network configuration, includes: Based on the overall traffic prediction results, the load values ​​of each access point are obtained to obtain the local load values; If the local load value is higher than the preset load value threshold, it is determined to be a high-load access point, and a nearby access point is determined according to the preset switching rules. The traffic of the high-load access point is redirected to the nearby access point to obtain the initial network configuration.

3. The network handover method based on multi-round simulation optimization according to claim 1, characterized in that, The process involves performing multiple rounds of handover simulation based on the initial network configuration to generate a candidate sequence containing handover actions. If the candidate sequence containing handover actions satisfies a preset load balancing condition, a preliminary handover scheme is obtained. The reward value of the preliminary handover scheme is calculated based on a preset reward function, and the handover scheme with the highest reward value is selected to obtain an optimized handover scheme, including: Simulate multiple rounds of traffic allocation scenarios based on the initial network configuration, and generate a candidate sequence containing switching actions; Extract the load distribution of the candidate sequence and evaluate it. If the load distribution meets the preset load balancing conditions, then determine the candidate sequence as a preliminary switching scheme. The reward value of the initial switching scheme is calculated according to the preset reward function, and the scheme with the highest reward value is selected as the optimized switching scheme.

4. The network handover method based on multi-round simulation optimization according to claim 1, characterized in that, The step involves performing anomaly detection based on the adjusted access point status. If an abnormal access point is found, the prediction method is adjusted and the overall traffic is predicted to obtain an updated overall traffic prediction result, including: The adjusted access point status is compared with the preset access point status template. If the access point status is abnormal, a list of abnormal access points is obtained. Obtain real-time data of the abnormal access point list, adjust the prediction method to predict the overall traffic, and obtain the updated overall traffic prediction result.

5. The network handover method based on multi-round simulation optimization according to claim 1, characterized in that, The step of optimizing the network configuration based on the updated overall traffic prediction results to obtain the final network switching configuration includes: Based on the updated overall traffic prediction results, the contribution value of each access point is calculated using a weighted average. Based on the contribution value of each access point, the updated optimized network configuration is determined and distributed to each access point to obtain the final network switching configuration.

6. A network handover system based on multi-round simulation optimization, characterized in that, A network handover method based on multi-round simulation optimization as described in any one of claims 1 to 5 includes: The data acquisition module is used to acquire real-time network status information and extract features to obtain network status features; The traffic prediction module is used for each distributed access point to perform partial traffic prediction based on the network status characteristics, obtain partial traffic prediction results, and aggregate the partial traffic prediction results to obtain the overall traffic prediction result. The load configuration module is used to obtain the load value of each access point from the overall traffic prediction result. If the load value is higher than the preset load threshold, it is determined to be a high-load access point, and the traffic of the high-load access point is redirected to a nearby access point to obtain the initial network configuration. The handover strategy generation module is used to perform multiple rounds of handover simulation based on the initial network configuration, generate a candidate sequence containing handover actions, and if the candidate sequence containing handover actions meets the preset load balancing conditions, a preliminary handover scheme is obtained. The reward value of the preliminary handover scheme is calculated according to the preset reward function, and the handover scheme with the highest reward value is selected to obtain the optimized handover scheme. The access point status recording module is used to distribute the optimized switching scheme to each access point, obtain the optimized network configuration, and record the adjusted access point status. An anomaly detection module is used to perform anomaly detection based on the adjusted access point status. If an abnormal access point exists, the prediction method is adjusted and the overall traffic is predicted to obtain an updated overall traffic prediction result. The network configuration module is used to optimize the optimized network configuration based on the updated overall traffic prediction results to obtain the final network switching configuration.

7. An electronic device, characterized in that, The method includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the network switching method based on multi-round simulation optimization as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the network handover method based on multi-round simulation optimization as described in any one of claims 1 to 5.

Images