Affinity propagation flow clustering method and device based on parameter adaptive sparrow search

Abstract

The invention discloses an affinity propagation traffic clustering method and device based on parameter adaptive sparrow search, and relates to the field of 5G network traffic clustering, and the method comprises the steps: obtaining network traffic data of a 5G base station, and carrying out the preprocessing of the network traffic data; determining an optimal deviation parameter and a damping coefficient by using a parameter adaptive sparrow search algorithm; according to the parameter adaptive sparrow search algorithm, an initial population is optimized by using a parameter adaptive ICMIC-Logistic mapping method on the basis of a traditional sparrow search algorithm, and positions of individuals in the population are updated by using a parameter adaptive spiral exploration strategy; and according to the optimal deviation parameter and the damping coefficient, clustering the preprocessed network traffic by using an affinity propagation algorithm to obtain a clustering result so as to determine a dormancy timing strategy of the 5G base station. The network traffic data of the 5G base station is clustered, so that the problems that the no-load energy consumption of the 5G base station is high and manual timing energy saving is influenced by human factors are solved.

Description

Technical Field

[0001] The present application relates to the field of 5G network traffic clustering, and particularly to a method and device for near neighbor propagation traffic clustering based on parameter adaptive sparrow search. Background Art

[0002] The 5G base station industry is currently in a stage of rapid development. According to the data released by the Ministry of Industry and Information Technology, as of October 15, 2024, the total number of 5G base stations in China has reached 4.042 million, accounting for 32.1% of the total number of mobile base stations, ranking first in the world in terms of network scale. Related applications have covered 74 major categories of the national economy. Compared with 4G technology, 5G technology has significantly improved in terms of rate, latency, and data throughput. However, high performance brings high energy consumption. The DC load power consumption of a single 5G base station is about 3 to 4 times that of a 4G base station, with an average power consumption of 60 kWh. Even the no-load power consumption of the main equipment of a 5G base station is about 2.2 - 2.3 kilowatts, much higher than the few hundred watts of 4G no-load consumption, bringing huge pressure and challenges to operators. Therefore, it is urgent to effectively solve the problems of high no-load energy consumption of 5G base stations and the influence of manual timing energy saving by human factors. Summary of the Invention

[0003] The purpose of the present application is to provide a method and device for near neighbor propagation traffic clustering based on parameter adaptive sparrow search, which solves the problems of high no-load energy consumption of 5G base stations and the influence of manual timing energy saving by human factors by clustering the network traffic data of 5G base stations.

[0004] To achieve the above purpose, the present application provides the following solutions:

[0005] In the first aspect, the present application provides a method for near neighbor propagation traffic clustering based on parameter adaptive sparrow search, including:

[0006] Obtain the network traffic data of 5G base stations, preprocess the network traffic data, and obtain the preprocessed network traffic;

[0007] Use the parameter adaptive sparrow search algorithm to determine the optimal preference parameter and the optimal damping coefficient of the near neighbor propagation algorithm; the parameter adaptive sparrow search algorithm optimizes the initial population using the parameter adaptive ICMIC-Logistic mapping method on the basis of the traditional sparrow search algorithm, and updates the positions of individuals in the population using the parameter adaptive spiral exploration strategy;

[0008] According to the optimal preference parameter and the optimal damping coefficient, apply the near neighbor propagation algorithm to cluster the preprocessed network traffic, and obtain a clustering result; the clustering result is used to determine the sleep timing strategy of 5G base stations.

[0009] In a second aspect, the present application provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the above-mentioned nearest neighbor propagation traffic clustering method based on parameter adaptive sparrow search.

[0010] In a third aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the above-mentioned nearest neighbor propagation traffic clustering method based on parameter adaptive sparrow search.

[0011] In a fourth aspect, the present application provides a computer program product, including a computer program, and when the computer program is executed by a processor, it implements the above-mentioned nearest neighbor propagation traffic clustering method based on parameter adaptive sparrow search.

[0012] According to the specific embodiments provided by the present application, the following technical effects are disclosed in the present application:

[0013] The present application provides a nearest neighbor propagation traffic clustering method and device based on parameter adaptive sparrow search, which uses the parameter adaptive sparrow search algorithm to determine the optimal bias parameter and damping coefficient; the parameter adaptive sparrow search algorithm optimizes the initial population using the parameter adaptive ICMIC-Logistic mapping method on the basis of the traditional sparrow search algorithm, and updates the positions of individuals in the population using the parameter adaptive spiral exploration strategy; according to the optimal bias parameter and damping coefficient, the nearest neighbor propagation algorithm is applied to cluster the preprocessed network traffic, and the clustering result is obtained to determine the sleep timing strategy of the 5G base station. Among them, the present application uses the parameter adaptive sparrow search algorithm to improve the search efficiency and accuracy of the optimal bias parameter and optimal damping coefficient of the nearest neighbor propagation algorithm, thereby improving the clustering efficiency and clustering accuracy of 5G network traffic data, and finally solving the problems of high no-load energy consumption of 5G base stations and manual timing energy saving being affected by human factors according to the clustering result. Description of the Drawings

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained according to these drawings.

[0015] Figure 1 It is an application environment diagram of a nearest neighbor propagation traffic clustering method based on parameter adaptive sparrow search in an embodiment of the present application;

[0016] Figure 2Schematic flowchart of a nearest neighbor propagation traffic clustering method based on parameter adaptive sparrow search provided by an embodiment of the present application;

[0017] Figure 3 Schematic diagram of the clustering result of nearest neighbor propagation traffic clustering based on parameter adaptive sparrow search provided by an embodiment of the present application;

[0018] Figure 4 Schematic diagram of the functional modules of a nearest neighbor propagation traffic clustering device based on parameter adaptive sparrow search provided by an embodiment of the present application;

[0019] Figure 5 Schematic diagram of the structure of a computer device provided by an embodiment of the present application. Detailed implementation manners

[0020] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts shall fall within the protection scope of the present application.

[0021] To make the above objects, features, and advantages of the present application more obvious and understandable, the present application will be further described in detail below with reference to the accompanying drawings and specific implementation manners.

[0022] The nearest neighbor propagation traffic clustering method based on parameter adaptive sparrow search provided by the embodiments of the present application can be applied to, for example Figure 1In the application environment shown. Among them, the terminal communicates with the server through the network. The data storage system can store the data that the server needs to process. The data storage system can be set separately, integrated on the server, or placed on the cloud or other servers. The terminal can send the network traffic data of the 5G base station to the server. After the server receives the network traffic data of the 5G base station, the server preprocesses the network traffic data to obtain the preprocessed network traffic; uses the parameter adaptive sparrow search algorithm to determine the optimal preference parameter and the optimal damping coefficient of the affinity propagation algorithm; the parameter adaptive sparrow search algorithm optimizes the initial population using the parameter adaptive ICMIC-Logistic mapping method on the basis of the traditional sparrow search algorithm, and updates the positions of the individuals in the population using the parameter adaptive spiral exploration strategy; according to the optimal preference parameter and the optimal damping coefficient, applies the affinity propagation algorithm to cluster the preprocessed network traffic to obtain a clustering result; the clustering result is used to determine the sleep timing strategy of the 5G base station. The server can feedback the obtained optimal preference parameter, optimal damping coefficient and clustering result to the terminal. In addition, in some embodiments, the affinity propagation traffic clustering method based on parameter adaptive sparrow search can also be implemented by the server or the terminal alone. For example, the terminal can directly perform affinity propagation traffic clustering based on parameter adaptive sparrow search on the network traffic data of the 5G base station, or the server can obtain the network traffic data of the 5G base station from the data storage system and perform affinity propagation traffic clustering based on parameter adaptive sparrow search.

[0023] Among them, the terminal can be, but is not limited to, various desktop computers, laptop computers, smartphones, tablets, Internet of Things devices, and portable wearable devices. The server can be implemented by an independent server or a server cluster composed of multiple servers, and can also be a cloud server.

[0024] To effectively solve the problems of high no-load energy consumption in 5G base stations and the influence of manual timing energy saving by human factors, this application clusters according to the characteristics of the 5G base station traffic dataset to determine the sleep timing strategy of the base station. Clustering is a statistical method in the field of machine learning. According to the internal characteristics between different object data, data samples with high similarity are divided into the same cluster. There are many traditional clustering methods, each with limitations; the fuzzy C-means clustering has a high dependence on the initial value and is prone to falling into local minima; the density-based DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm has a high time complexity and is not applicable to large datasets; in the prior art, a clustering algorithm is used to process large optical network traffic data; in the prior art, a two-step clustering method is used to perform clustering analysis on a 5G base station cell of a certain operator, and the BIRCH algorithm is used to pre-cluster the data points in the dense area, and then the preprocessed results are clustered according to the principle of the smallest distance; in the prior art, a traffic clustering method based on semi-supervised learning is used to cluster Internet user traffic to shorten the traffic scheduling time; in the prior art, a k-means algorithm based on particle swarm optimization is proposed to perform clustering analysis on the traffic data prediction results, effectively solving the problem of the k-means algorithm being affected by the clustering center; however, k-means clustering requires manual specification of the initial clustering center and number, and is relatively sensitive to noise and outliers; in the prior art, an affinity propagation clustering algorithm (AP) is proposed, which does not require prior determination of the clustering number. Compared with the above clustering methods, it has a faster processing speed and has been applied to many fields, such as face recognition, construction engineering, fault diagnosis and processing, etc. In the AP algorithm, the preference parameter and the damping coefficient need to be set manually. Among them, the preference parameter determines the clustering effect. The larger the preference parameter, the more clustering numbers, and vice versa; the damping coefficient affects the oscillation degree of the clustering process. Increasing the damping coefficient will reduce the clustering speed, and decreasing the damping coefficient will increase the oscillation. In recent years, using swarm intelligence optimization algorithms to optimize AP clustering and determine appropriate preference parameters and damping coefficients has become a research hotspot. In the prior art, the particle swarm algorithm is combined with AP, and the robustness and stability are improved; in the prior art, an improved fruit fly optimization algorithm is used to optimize the selection of the preference parameters of AP; in the prior art, the idea of fireworks explosion is introduced to balance the global and local search capabilities of AP. The above existing algorithms have made some improvements in avoiding local optima and improving the global search ability, but there is still room for improvement in the search efficiency and accuracy. The sparrow search algorithm (SSA) is a new swarm intelligence optimization algorithm proposed by XUE in 2020. It simulates the process of sparrows foraging, has the characteristics of fast convergence speed, strong adaptability, and easy modification of the model, but is prone to falling into local optimal solutions.

[0025] In an exemplary embodiment, Figure 2 As shown, a method for clustering neighbor propagation traffic based on parameter adaptive sparrow search is provided. The method is executed by a computer device, and can be executed by a computer device such as a terminal or a server alone, or by a terminal and a server together. In the embodiment of the present application, the method is applied to Figure 1 The server in the example is used to illustrate, including the following steps 101 to 103. Among them:

[0026] Step 101, obtain network traffic data of the 5G base station, preprocess the network traffic data, and obtain the preprocessed network traffic.

[0027] Read the network traffic data of 5G base stations, perform data cleaning and normalization to eliminate invalid data caused by equipment failure, network interruption, etc. Perform subsequent cluster analysis based on the preprocessed data. Network traffic data is time series data, which contains traffic data corresponding to different time points.

[0028] Step 102, using a parameter-adaptive sparrow search algorithm to determine the optimal bias parameter and optimal damping coefficient of the nearest neighbor propagation algorithm; the parameter-adaptive sparrow search algorithm uses a parameter-adaptive ICMIC-Logistic mapping method to optimize the initial population on the basis of a traditional sparrow search algorithm, and uses a parameter-adaptive spiral exploration strategy to update the position of individuals in the population. ICMIC-Logistic hybrid mapping is to substitute the formula of ICMIC chaotic mapping into the Logistic hybrid mapping, so the parameter that changes with iteration is the control parameter of the Logistic hybrid mapping, that is, 1 / z in the sin function i The coefficient of .

[0029] The neighbor propagation clustering algorithm does not need to determine the number of clusters in advance. It can divide data samples with high similarity into the same cluster according to the intrinsic characteristics between different object data. The two important parameters of the neighbor propagation clustering algorithm are the bias parameter and the damping coefficient. Among them, the bias parameter determines the clustering effect. The larger the bias parameter, the more clusters there are, and vice versa. The damping coefficient affects the oscillation degree of the clustering process. Increasing the damping coefficient will reduce the clustering speed, and reducing the damping coefficient will increase the oscillation. Therefore, it is necessary to determine the optimal bias parameter and optimal damping parameter of the neighbor propagation distance algorithm so as to perform clustering analysis based on the optimal value and obtain the optimal clustering result.

[0030] Step 103: According to the optimal deflection parameter and the optimal damping coefficient, the neighbor propagation algorithm is applied to cluster the preprocessed network traffic to obtain a clustering result; the clustering result is used to determine the sleep timing strategy of the 5G base station.

[0031] By implementing the above steps 101 to 103, based on the traditional sparrow search algorithm, this application optimizes the initial population using the parameter - adaptive ICMIC - Logistic mapping method, and updates the positions of individuals in the population using the parameter - adaptive spiral exploration strategy, introducing an improved sparrow search algorithm to solve the problems of existing clustering algorithms being prone to falling into local optima, slow convergence speed, and low search efficiency. Therefore, this application uses the parameter - adaptive sparrow search algorithm to improve the search efficiency and accuracy of the optimal preference parameter and the optimal damping coefficient of the affinity propagation algorithm, thereby improving the clustering efficiency and clustering accuracy of 5G network traffic data, and finally solving the problems of high no - load energy consumption of 5G base stations and the influence of manual timing energy - saving by human factors according to the clustering results.

[0032] In another exemplary embodiment of this application, in step 102, the parameter adaptability is reflected in using the parameter - adaptive ICMIC - Logistic mapping method to optimize the initial population of the sparrow search algorithm and adopting the parameter - adaptive spiral exploration strategy to avoid falling into local optima. Determining the optimal preference parameter and the optimal damping coefficient of the affinity propagation algorithm using the parameter - adaptive sparrow search algorithm specifically includes:

[0033] (1) Determine the parameter search space of the preference parameter and the parameter search space of the damping coefficient of the affinity propagation algorithm.

[0034] The sparrow search algorithm is an optimization algorithm based on swarm intelligence. By simulating the foraging behavior of sparrows, it searches for the optimal solution in the parameter space. The parameter dimension of the affinity propagation clustering algorithm is 2 - dimensional. Among them, the value range of the preference parameter is set from - 500 to 0, initially set to - 500, and the value range of the damping coefficient is 0.5 to 1, initially set to 0.5. As an example, the initial number of sparrow individuals in the sparrow search algorithm is 100, and the maximum number of iterations is 100. In the sparrow search algorithm, each sparrow individual represents a set of affinity propagation clustering parameters (preference parameter and damping coefficient). By continuously iteratively updating the positions of sparrow individuals, the parameter combination corresponding to the sparrow individual with the maximum fitness function value is found.

[0035] (2) According to the parameter search space of the preference parameter and the parameter search space of the damping coefficient, use the parameter - adaptive ICMIC - Logistic mapping method to construct the initial population; each sparrow individual in the initial population represents a numerical combination of the preference parameter and the damping coefficient.

[0036] (3) Calculate the silhouette coefficient corresponding to each sparrow individual in the current population; the silhouette coefficient corresponding to each sparrow individual is used as the fitness value corresponding to each sparrow individual.

[0037] The design of the fitness function in the sparrow search algorithm is the key to evaluating the clustering effect, and it is necessary to comprehensively consider the compactness and separation of the clustering results. In this application, the silhouette coefficient is used as the fitness function because it can simultaneously reflect the similarity of data points with other points within their respective clusters and the differences with points in other clusters. The silhouette coefficient is used as the fitness function to evaluate the quality of the clustering results of the affinity propagation clustering algorithm under different parameter combinations.

[0038] (4) Determine the discoverers and followers in the current population according to the fitness values corresponding to each sparrow individual in the current population.

[0039] (5) Update the positions of the discoverers and followers in the current population; the position update of the discoverers adopts a parameter adaptive spiral exploration strategy.

[0040] (6) Randomly determine the vigilant ones from the current population and update the positions of the vigilant ones to obtain the next generation population.

[0041] Select the top 10% of the solutions after sorting the fitness values, and use them as the sparrows that may perceive danger, that is, the vigilant ones. The position update of the vigilant ones is carried out after the position updates of the discoverers and followers, and it is an operation that will be carried out whether there is danger or not. Their positions are randomly generated in the population. The positions of these sparrows are random at the beginning. When they realize the danger, they need to move their positions. The sparrows on the periphery of the population move closer to the safe area, and the sparrows in the center of the population walk randomly to get closer to other sparrows, and then judge the relationship between the fitness values of each vigilant one and the global optimal fitness value to update the positions.

[0042] (7) Judge whether the current iteration number has reached the maximum iteration number.

[0043] The sparrow search algorithm is used to search for the optimal parameter combination of the affinity propagation clustering algorithm within the set parameter space. During the iteration process, the positions of sparrow individuals are updated according to the fitness values, the excellent individuals are retained, and the poor individuals are eliminated. When the iteration number is reached or the convergence condition is met, the optimal parameter combination is output.

[0044] (8) If so, use the numerical combination corresponding to the sparrow individual with the optimal fitness value in the next generation population as the optimal preference parameter and the optimal damping coefficient.

[0045] (9) If not, use the next generation population as the current population and return to the step of "calculating the silhouette coefficient corresponding to each sparrow individual in the current population".

[0046] In another exemplary embodiment of this application, the expression of the parameter adaptive ICMIC-Logistic mapping method is:

[0047]

[0048] Among them, z i is the chaotic value of the position of the i-th sparrow, and the chaotic value is a number chaotic sequence between -1 and 1; α is the control parameter of ICMIC, and the range of α is 0.5 to 1; α 0 is the initial iteration value of the initial position of the sparrow, α 0 = 0.7; α f is the final iteration value of the initial position of the sparrow, α f = 1; N is the maximum number of iterations of the sparrow search algorithm SSA. z i+1 is the chaotic value of the initial position of the (i + 1)-th sparrow; n represents the number of sparrow individuals;

[0049] Apply the adaptive ICMIC-Logistic mapping to the search space initialization process of the sparrow search algorithm SSA, and use the generated chaotic sequence to set the initial position of each sparrow individual. The specific formula is as follows:

[0050]

[0051] Among them, X i represents the initial position of the i-th sparrow; X l , X u represents the upper and lower bounds of the initial position of the sparrow. The initial position of each sparrow individual in each dimension is jointly determined by the upper bound, lower bound, and the generated chaotic sequence value of its corresponding dimension. It not only ensures the randomness and uniform distribution of population initialization, but also makes full use of the ergodicity of the chaotic sequence, so that the initial population can cover the solution space more comprehensively.

[0052] In another exemplary embodiment of this application, the expression for updating the position of the discoverer is:

[0053]

[0054] Among them,

[0055] In the formula, X i t represents the position of the i-th individual in the t-th generation of the population; γ is a uniform random number in (0, 1); N is the maximum number of iterations of the sparrow search algorithm SSA; Q is a standard normal distribution random number; L is a multi-dimensional all-one matrix; R 2 is a uniform random number in the interval [0, 1], θ is the warning threshold, and its value is 0.6; β is the spiral exploration factor applied with the parameter adaptive spiral exploration strategy, b i is the spiral shape constant, b f is the final spiral shape constant, b 0is the initial spiral shape constant, η represents the path coefficient, and η is a random number in [-1, 1].

[0056] In another exemplary embodiment of the present application, in step 103, using the optimized optimal bias parameter and damping coefficient as inputs, perform the affinity propagation clustering algorithm, and output the clustering result, including the cluster centers and the cluster labels to which each data point belongs. According to the clustering result, take the time corresponding to the label of the category with the lowest traffic as the timing strategy for the 5G base station to go to sleep, and give the sleep period of the base station, such as Figure 3 shown, Figure 3 In, the horizontal axis is time, the vertical axis is the corresponding traffic value, the unit is GB, the same clusters are divided into the same color, and the "×" in the figure is the cluster center. It can be seen that the traffic values at points 3, 4, and 5 are relatively low. Taking them as the time period for the base station to go to sleep will neither affect the user experience nor waste energy consumption. Specifically, according to the optimal bias parameter and the optimal damping coefficient, apply the affinity propagation algorithm to cluster the preprocessed network traffic, and obtain the clustering result, specifically including:

[0057] (1) Calculate the similarity between every two data points in the preprocessed network traffic to obtain a similarity matrix; the diagonal elements of the similarity matrix are the optimal bias parameters; the non-diagonal elements of the similarity matrix are the composite similarities of the time dimension and the traffic dimension between the corresponding two data points, and the similarities of the two dimensions are obtained by weighted summation to obtain the final similarity, that is, the non-diagonal element values in the similarity matrix.

[0058] According to the characteristics of the 5G base station traffic dataset, the present application constructs a similarity function for the affinity propagation clustering algorithm, which fuses the similarities of the traffic and time dimensions to more effectively identify the similarity between traffic.

[0059] (2) Update the attractiveness matrix and the responsibility matrix according to the similarity matrix and the optimal damping coefficient; at the initial moment, the attractiveness matrix and the responsibility matrix respectively adopt the preset initial attractiveness matrix and initial responsibility matrix.

[0060] (3) Determine whether the current iteration meets the termination condition; the termination condition is that the cluster centers no longer update or the current iteration number reaches the maximum iteration number.

[0061] (4) If so, obtain the finally divided cluster centers, and cluster the network traffic data according to the cluster centers to obtain the clustering result; the clustering result includes the cluster centers and the cluster labels to which each data point belongs.

[0062] (5) If not, return to the step of "updating the attractiveness matrix and the responsibility matrix according to the similarity matrix and the optimal damping coefficient".

[0063] In another exemplary embodiment of the present application, the clustering results are evaluated by the Silhouette Coefficient, the Calinski-Harabasz index, and the Davies-Bouldin index. Therefore, after performing step 103, "Apply the affinity propagation algorithm to cluster the preprocessed network traffic according to the optimal bias parameter and the optimal damping coefficient to obtain clustering results", the affinity propagation traffic clustering method based on parameter adaptive sparrow search further includes:

[0064] (1) Calculate the Silhouette Coefficient, the Calinski-Harabasz index, and the Davies-Bouldin index corresponding to the clustering results.

[0065] (2) Evaluate the clustering results according to the Silhouette Coefficient, the Calinski-Harabasz index, and the Davies-Bouldin index corresponding to the clustering results.

[0066] In another exemplary embodiment of the present application, after performing step 103, "Apply the affinity propagation algorithm to cluster the preprocessed network traffic according to the optimal bias parameter and the optimal damping coefficient to obtain clustering results", the affinity propagation traffic clustering method based on parameter adaptive sparrow search further includes: visualizing the clustering results. Visualizing the clustering results can represent different labels in different colors, which is convenient for users to intuitively understand the distribution characteristics of network traffic.

[0067] Compared with the prior art, the affinity propagation traffic clustering method based on parameter adaptive sparrow search of the present application does not require manual setting of the number of clusters, and has the advantages of high search efficiency, fast convergence speed, excellent sleep timing strategy, and the ability to avoid falling into local optimal solutions. Specifically:

[0068] (1) Do not require manual setting of the number of clusters

[0069] Traditional K-means and K-means++ algorithms require manual setting of the number of clusters, while the affinity propagation clustering algorithm adopted in the present application can automatically identify the internal characteristics between different object data, and divide data samples with high similarity into the same cluster without manual setting of the number of clusters.

[0070] (2) High search efficiency

[0071] Traditional optimization algorithms may require a large number of iterations to find the optimal solution during the optimization process. However, the present application uses a new swarm intelligence optimization algorithm, namely the sparrow search algorithm, which simulates the foraging process of sparrows and has the advantage of fast convergence speed. Moreover, a parameter adaptive mechanism is introduced, which can dynamically adjust the algorithm parameters according to the current clustering state and the number of iterations, thereby accelerating the convergence process of the algorithm.

[0072] (3) Fast convergence speed

[0073] Through parameter adaptation, the algorithm can identify potential clustering centers faster and reduce the search in unnecessary areas, thus improving the overall processing efficiency.

[0074] (4) Avoiding falling into local optimal solutions

[0075] By designing a parameter-adaptive ICMIC-Logistic mapping to optimize the initial sparrow population of the sparrow search algorithm and adopting a parameter-adaptive spiral exploration strategy to avoid local optima.

[0076] (5) Optimal sleep timing strategy

[0077] First, according to the characteristics of the 5G base station traffic dataset, construct a similarity function for the affinity propagation clustering algorithm, fuse traffic and time similarity, and more effectively identify the similarity between traffic, thus proposing a better sleep timing strategy; second, use the sparrow search method to find the optimal preference parameter and damping coefficient of the affinity propagation clustering algorithm, making the clustering effect more accurate. Therefore, the proposed sleep timing strategy is better.

[0078] To fully demonstrate the flexibility and innovation of this application, the following several alternative solutions are provided, which can also achieve the purpose of this application:

[0079] 1. Affinity propagation traffic clustering based on genetic algorithm:

[0080] By introducing selection, crossover, and mutation operations in the genetic algorithm, optimize the affinity propagation traffic clustering. While maintaining the clustering effect, this solution improves the convergence speed and global search ability of the algorithm.

[0081] 2. Affinity propagation traffic clustering based on particle swarm algorithm:

[0082] Utilize the particle position and velocity update mechanism in the particle swarm algorithm to iteratively optimize the affinity propagation traffic clustering. While maintaining the clustering accuracy, this solution reduces the computational complexity of the algorithm.

[0083] 3. Affinity propagation traffic clustering combined with deep learning:

[0084] By introducing deep learning technology, optimize the feature extraction and clustering process of the affinity propagation traffic clustering. While maintaining the clustering effect, this solution improves the generalization ability and robustness of the algorithm.

[0085] The above alternative solutions have all been experimentally verified and have their own advantages and disadvantages compared with this application. However, overall, this application has significant advantages in terms of clustering effect, convergence speed, and computational complexity.

[0086] The present application also provides an application scenario, which applies the above-mentioned nearest neighbor propagation traffic clustering method based on parameter adaptive sparrow search. Specifically: The nearest neighbor propagation traffic clustering method based on parameter adaptive sparrow search provided in this embodiment can be applied to the scenario of setting the sleep timing strategy of 5G base stations. This scenario includes a data collection link for collecting the network traffic data of 5G base stations; a data clustering link for obtaining the network traffic data of 5G base stations and performing nearest neighbor propagation traffic clustering based on parameter adaptive sparrow search to obtain a clustering result; a sleep timing strategy setting link for setting the sleep timing strategy of 5G base stations according to the clustering result. The nearest neighbor propagation traffic clustering method based on parameter adaptive sparrow search provided in this embodiment belongs to the data clustering link.

[0087] Based on the same inventive concept, an embodiment of the present application also provides a nearest neighbor propagation traffic clustering device based on parameter adaptive sparrow search for implementing the above-mentioned nearest neighbor propagation traffic clustering method based on parameter adaptive sparrow search. The solution provided by this device for solving problems is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the nearest neighbor propagation traffic clustering device based on parameter adaptive sparrow search provided below can refer to the limitations on the nearest neighbor propagation traffic clustering method based on parameter adaptive sparrow search in the above text, and will not be repeated here.

[0088] In an exemplary embodiment, as Figure 4 shown, a nearest neighbor propagation traffic clustering device based on parameter adaptive sparrow search is provided, aiming to efficiently and accurately complete the clustering analysis of 5G network traffic. The device includes: a data processing module - labeled as A; an algorithm execution module - labeled as B, which contains: a parameter adaptive sparrow search sub-module - B1 and a nearest neighbor propagation clustering sub-module - B2; a data storage module - labeled as C; a user interface - labeled as D; and a communication interface - labeled as E.

[0089] The connection relationship between each structure:

[0090] Data processing module A: Responsible for receiving the original 5G network traffic data input from the outside, and performing preprocessing, such as data cleaning and normalization operations, and then transmitting the processed data to the algorithm execution module B.

[0091] Algorithm execution module B:

[0092] Parameter adaptive sparrow search sub-module B1: Receives the data from the data processing unit A, and dynamically adjusts the parameters of the search algorithm according to the characteristics of the data to optimize the search efficiency. The optimized parameters will be transmitted to the nearest neighbor propagation clustering sub-module B2.

[0093] Nearest Neighbor Propagation Clustering Sub-module B2: Based on the parameters provided by B1, execute the nearest neighbor propagation algorithm to perform clustering analysis on network traffic data and output the clustering results.

[0094] Data Storage Module C: Store the original network traffic data, preprocessed network traffic data, intermediate results during algorithm execution, and final clustering results for subsequent analysis or query.

[0095] User Interface D: Provide an interactive interface between the user and the device, including functions such as data input, parameter setting, and clustering result display.

[0096] Communication Interface E: Support communication with other systems or devices to achieve data input / output and remote monitoring and management.

[0097] Each structure and its functions:

[0098] Data Processing Module A: As the starting point of the data stream, ensure the accuracy and availability of the input data, and provide a high-quality data basis for subsequent algorithm execution.

[0099] Parameter Adaptive Sparrow Search Sub-module B1: By intelligently adjusting search parameters, improve the adaptability and clustering efficiency of the algorithm on different datasets, which is the key to algorithm performance optimization.

[0100] Nearest Neighbor Propagation Clustering Sub-module B2: Utilize the advantages of the nearest neighbor propagation clustering algorithm to effectively identify and classify different patterns or categories in network traffic and output highly accurate clustering results.

[0101] Data Storage Module C: Ensure data integrity and traceability, and provide support for subsequent operations such as data analysis and report generation.

[0102] User Interface D: Enhance the usability and interactivity of the device, enabling users to intuitively understand the clustering process and results for easy decision-making.

[0103] Communication Interface E: Promote the integration and scalability of the device, enabling this product to be easily integrated into a wider network environment.

[0104] In an exemplary embodiment, a computer device is provided. The computer device can be a server or a terminal, and its internal structure diagram can be as Figure 5As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store relevant data for nearest neighbor propagation traffic clustering based on parameter adaptive sparrow search. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it implements a method for nearest neighbor propagation traffic clustering based on parameter adaptive sparrow search.

[0105] Those skilled in the art can understand that Figure 5 the structure shown in the figure is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0106] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0107] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0108] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0109] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random-access memories (ReRAM), magnetoresistive random-access memories (MRAM), ferroelectric random-access memories (FRAM), phase change memories (PCM), graphene memories, etc. Volatile memories can include random-access memories (RAM) or external caches, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random-access memory (SRAM) or dynamic random-access memory (DRAM), etc.

[0110] The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logics, data processing logics based on quantum computing, etc., without limitation.

[0111] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0112] In this article, specific examples are used to elaborate on the principles and implementation manners of the present application. The descriptions of the above embodiments are only used to help understand the method and its core idea of the present application; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. A neighbor propagation traffic clustering method based on parameter adaptive sparrow search, characterized in that: The neighbor propagation traffic clustering method based on parameter adaptive sparrow search includes: Obtain network traffic data of a 5G base station, preprocess the network traffic data, and obtain preprocessed network traffic; The optimal bias parameter and the optimal damping coefficient of the neighbor propagation algorithm are determined by using a parameter-adaptive sparrow search algorithm; the parameter-adaptive sparrow search algorithm uses a parameter-adaptive ICMIC-Logistic mapping method to optimize the initial population on the basis of a traditional sparrow search algorithm, and uses a parameter-adaptive spiral exploration strategy to update the position of individuals in the population; According to the optimal bias parameter and the optimal damping coefficient, the neighbor propagation algorithm is applied to cluster the preprocessed network traffic to obtain a clustering result; the clustering result is used to determine the sleep timing strategy of the 5G base station.

2. According to claim 1, the neighbor propagation traffic clustering method based on parameter adaptive sparrow search is characterized in that: The optimal bias parameter and optimal damping coefficient of the neighbor propagation algorithm are determined by using the parameter adaptive sparrow search algorithm, including: Determine the parameter search space of the bias parameter and the parameter search space of the damping coefficient of the neighbor propagation algorithm; According to the parameter search space of the deflection parameter and the parameter search space of the damping coefficient, an initial population is constructed using a parameter adaptive ICMIC-Logistic mapping method; each sparrow individual in the initial population represents a numerical combination of the deflection parameter and the damping coefficient; Calculate the silhouette coefficient corresponding to each individual sparrow in the current population; the silhouette coefficient corresponding to each individual sparrow is used as the fitness value corresponding to each individual sparrow; Determine the discoverer and follower in the current population according to the fitness value corresponding to each individual sparrow in the current population; Update the positions of the discoverers and followers in the current population; the discoverer position update adopts a parameter-adaptive spiral exploration strategy; Randomly determine a sentinel from the current population, and update the position of the sentinel to obtain the next generation population; Determine whether the current number of iterations has reached the maximum number of iterations; If so, the numerical combination corresponding to the sparrow individual with the best fitness value in the next generation population is used as the optimal bias parameter and the optimal damping coefficient; If not, the next generation population is taken as the current population, and the process returns to step "calculating the silhouette coefficient corresponding to each individual sparrow in the current population".

3. The method for clustering neighbor propagation traffic based on parameter adaptive sparrow search according to claim 1 or 2, characterized in that: The expression of the parameter adaptive ICMIC-Logistic mapping method is: Among them, z i is the chaotic value of the initial position of the i-th sparrow; α is the control parameter of ICMIC; α0 is the initial iteration value of the initial position of the sparrow; α f is the final iteration value of the initial position of the sparrow; N is the maximum number of iterations of the parameter adaptive sparrow search algorithm; z i+1 is the chaos value of the initial position of the i+1th sparrow.

4. The method for clustering neighbor propagation traffic based on parameter adaptive sparrow search according to claim 1 or 2, characterized in that: The expression for the discoverer's position update is: in, Where, X i t represents the position of the i-th individual in the t-th generation in the population, γ is a uniform random number in (0,1), N is the maximum number of iterations of SSA, Q is a standard normal distribution random number, and L is a multi-dimensional all-one matrix. R2 is a uniform random number in the interval [0,1], θ is the warning threshold, and its value is 0.

6. β is the spiral exploration factor, b i is the spiral shape constant, b f is the final spiral shape constant, b0 is the initial spiral shape constant, and η represents the path coefficient, which is a random number in [-1,1].

5. The method for clustering neighbor propagation traffic based on parameter adaptive sparrow search according to claim 1 is characterized in that: According to the optimal deflection parameter and the optimal damping coefficient, the neighbor propagation algorithm is applied to cluster the pre-processed network traffic to obtain a clustering result, which specifically includes: Calculate the similarity between every two data points in the preprocessed network traffic to obtain a similarity matrix; the diagonal elements of the similarity matrix are the optimal bias parameters; the non-diagonal elements of the similarity matrix are the composite similarities of the time dimension and the traffic dimension between the corresponding two data points; The attraction matrix and the attribution matrix are updated according to the similarity matrix and the optimal damping coefficient; at the initial moment, the attraction matrix and the attribution matrix respectively use the preset initial attraction matrix and the initial attribution matrix; Determine whether the current iteration meets the termination condition; the termination condition is that the cluster center is no longer updated or the current iteration number reaches the maximum iteration number; If yes, the final cluster center is obtained, and the network traffic data is clustered according to the cluster center to obtain a clustering result; the clustering result includes the cluster center and the cluster label to which each data point belongs; If not, return to step "update the attraction matrix and the attribution matrix according to the similarity matrix and the optimal damping coefficient".

6. The method for clustering neighbor propagation traffic based on parameter adaptive sparrow search according to claim 1 is characterized in that: After executing the step of "clustering the pre-processed network traffic using the nearest neighbor propagation algorithm according to the optimal deflection parameter and the optimal damping coefficient to obtain a clustering result", the nearest neighbor propagation traffic clustering method based on parameter adaptive sparrow search also includes: Calculate the silhouette coefficient, CH index and DB index corresponding to the clustering result; The clustering results are evaluated according to the silhouette coefficient, CH index and DB index corresponding to the clustering results.

7. The method for clustering neighbor propagation traffic based on parameter adaptive sparrow search according to claim 1 is characterized in that: After executing the step of "clustering the preprocessed network traffic using the nearest neighbor propagation algorithm according to the optimal bias parameter and the optimal damping coefficient to obtain a clustering result", the nearest neighbor propagation traffic clustering method based on parameter adaptive sparrow search also includes: visualizing the clustering results.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the neighbor propagation traffic clustering method based on parameter adaptive sparrow search as described in any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for clustering neighbor propagation traffic based on parameter adaptive sparrow search described in any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for clustering neighbor propagation traffic based on parameter adaptive sparrow search described in any one of claims 1 to 7 is implemented.

Images