Communication network deployment method and device based on 5G, and electronic equipment

Abstract

The invention discloses a 5G-based communication network deployment method and device, and electronic equipment, and relates to the 5G field or other related technical fields, and the method comprises the steps: inputting a multi-dimensional traffic demand data set into a multi-class network demand prediction model, evaluating the network demands in a specified time period in the future, and determining the network demands in the specified time period; and fusing the network demand prediction results to generate a network state report, formulating a base station layout based on the network state report and various network traffic influence factors to obtain a base station layout optimization scheme, and formulating an edge computing node configuration strategy based on the network state report and the base station layout optimization scheme. And monitoring the network operation state after configuration of the regional network nodes, collecting user feedback data, analyzing the network operation state and the user feedback data by using a linear regression model to obtain a network performance index, and determining the network performance based on the network performance index. And dynamically adjusting the base station layout optimization scheme and the edge computing node configuration strategy to obtain a 5G communication network deployment scheme.

Description

Technical Field

[0001] The present invention relates to the field of 5G technology, communication technology or other related fields, and specifically to a 5G-based communication network deployment method and device, and electronic equipment. Background Art

[0002] The rapid development of fifth-generation mobile communication technology (5G), with its ultra-high speed, low latency, and massive connectivity capabilities, has provided a powerful impetus for society's digital transformation. 5G networks not only offer a quantum leap in speed and connection quality, but are also being widely adopted in areas such as the Internet of Things, smart cities, and autonomous driving, fostering deep integration and innovation across industries. However, the efficient deployment and optimization of 5G communication networks, particularly in complex urban environments and diverse user scenarios, while ensuring network service quality and stability, remain key challenges in their implementation.

[0003] 5G network planning and deployment methods often rely on empirical formulas and statistical models, particularly in network demand forecasting, base station layout optimization, and edge computing node configuration. These methods face numerous challenges in accuracy and flexibility. Network demand forecasting typically relies on simple linear extrapolation of historical data, making it difficult to accurately predict sudden traffic fluctuations or nonlinear changes in user demand. Base station layout optimization often overlooks the dynamic distribution of users and the diversity of business models, leading to uneven distribution of network resources. Edge computing node configuration strategies also fail to fully consider the matching of real-time network status with future demand, impacting the performance and efficiency of 5G networks.

[0004] To address the above-mentioned problems, no effective solutions have been proposed so far. Summary of the Invention

[0005] The embodiments of the present invention provide a 5G-based communication network deployment method, device, and electronic device to at least solve the technical problem in related technologies that, when deploying 5G networks, network demand forecasting is usually based on simple linear extrapolation of historical data, making it difficult to cope with sudden fluctuations in network traffic.

[0006] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a 5G-based communication network deployment method is provided, including: inputting a preprocessed multi-dimensional traffic demand data set into a pre-trained multi-category network demand prediction model, and using the network demand prediction model to evaluate the network demand within a specified time period in the future; fusing the network demand prediction results output by each type of network demand prediction model to generate a network status report, and formulating a base station layout based on the network status report and multiple network traffic influencing factors to obtain a base station layout optimization plan; formulating an edge computing node configuration strategy based on the network status report and the base station layout optimization plan, wherein the edge computing node configuration strategy is used to configure network node resources in each area; monitoring the network operation status after configuring regional network nodes, and collecting user feedback data, using a linear regression model to analyze the network operation status and the user feedback data to obtain network performance indicators; based on the network performance indicators, dynamically adjusting the base station layout optimization plan and the edge computing node configuration strategy to obtain a 5G communication network deployment plan.

[0007] According to another aspect of an embodiment of the present invention, a 5G-based communication network deployment device is also provided, including: a multi-type network demand prediction unit, which is used to input the preprocessed multi-dimensional traffic demand data set into a pre-trained multi-type network demand prediction model, and use the network demand prediction model to evaluate the network demand within a specified time period in the future; a demand fusion unit, which is used to fuse the network demand prediction results output by each type of network demand prediction model, generate a network status report, and formulate a base station layout based on the network status report and multiple network traffic influencing factors to obtain a base station layout optimization plan; a node configuration unit, which is used to formulate an edge computing node configuration strategy based on the network status report and the base station layout optimization plan, wherein the edge computing node configuration strategy is used to configure network node resources in each area; a network operation monitoring unit, which is used to monitor the network operation status after the regional network nodes are configured, and collect user feedback data, and use a linear regression model to analyze the network operation status and the user feedback data to obtain network performance indicators; a configuration adjustment unit, which is used to dynamically adjust the base station layout optimization plan and the edge computing node configuration strategy based on the network performance indicators to obtain a 5G communication network deployment plan.

[0008] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is also provided, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned 5G-based communication network deployment methods.

[0009] According to another aspect of an embodiment of the present invention, an electronic device is also provided, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the above-mentioned 5G-based communication network deployment methods.

[0010] According to another aspect of an embodiment of the present invention, a computer program product is also provided, including a computer program, which, when executed by a processor, implements the steps of any one of the above-mentioned 5G-based communication network deployment methods.

[0011] Based on the above disclosure, the present invention can evaluate network demand within a specified future time period through multiple types of network demand prediction models, thereby improving the accuracy of network demand prediction. Accurate network demand prediction is crucial for subsequent base station layout optimization and edge computing node configuration. It can avoid over-allocation or under-allocation of resources, significantly reduce network traffic fluctuations caused by emergencies, and improve the resource utilization efficiency and overall performance of 5G communication networks, thereby solving the technical problem in related technologies that when deploying 5G networks, network demand prediction is usually based on simple linear extrapolation of historical data, which makes it difficult to cope with sudden network traffic fluctuations. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0013] Figure 1 A hardware structure block diagram of a computer terminal (or mobile device) for implementing a 5G-based communication network deployment method is shown;

[0014] Figure 2 is a flowchart of an optional 5G-based communication network deployment method according to an embodiment of the present invention;

[0015] Figure 3 is a schematic diagram of an optional 5G-based communication network deployment device according to an embodiment of the present invention;

[0016] Figure 4 This is a structural block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0017] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0018] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0019] To facilitate those skilled in the art to understand the present invention, some of the terms or nouns involved in the embodiments of the present invention are explained below:

[0020] The fifth generation of mobile communication technology, 5th Generation Mobile Network, referred to as 5G, is designed to support massive machine-type communications, enhanced mobile broadband, and ultra-reliable low-latency communications, thus meeting the needs of future Internet, Internet of Things and other application scenarios.

[0021] The Autoregressive Integrated Moving Average Model (ARIMA) is a statistical model commonly used for time series data forecasting. It can handle data with non-stationary characteristics and fits the data through three components: autoregression, differencing, and moving average. In this paper, the ARIMA model is used to predict future network demand based on historical traffic demand data. In particular, for time series data with trend and seasonal changes, the ARIMA model can provide more accurate forecasts.

[0022] LSTM (Long Short-Term Memory Network) is a special type of recurrent neural network (RNN) designed to overcome long-term dependencies. It uses a gating mechanism to retain or forget information, enabling the model to remember information over long sequences. In this paper, the LSTM model is used to process nonlinear relationships in multidimensional historical datasets, capable of capturing long-term dependencies in time series, thereby providing complex network demand forecasting.

[0023] Akaike Information Criterion (AIC) is an information-theoretic metric used to evaluate statistical models. It is used to balance model complexity with data fit during model selection. When selecting ARIMA model parameters, AIC can help select the most appropriate combination of model parameters to reduce model complexity while maintaining a good data fit.

[0024] The Bayesian Information Criterion (BIC) is also used to evaluate models. Compared to the AIC, the BIC places a greater penalty on model complexity and tends to select simpler models. Similar to the AIC, the BIC is used to select ARIMA models, especially for large datasets, where it can better mitigate the risk of overfitting.

[0025] eMBB, Enhanced Mobile Broadband, is a scenario in 5G applications that pursues high-speed data transmission, aiming to meet users' high-bandwidth requirements for services such as high-quality video, virtual reality, and augmented reality.

[0026] Ultra Reliable Low Latency Communications (URLLC) is another key 5G application scenario, focused on providing ultra-low latency and highly reliable communication connections for applications such as autonomous driving and remote surgery. URLLC ensures that 5G networks can meet the stringent latency and reliability requirements of mission-critical services, enhancing communication security and responsiveness.

[0027] MSE (mean squared error) is a commonly used metric for predictive model performance evaluation, measuring the square of the average error between the predicted value and the true value. MSE is used during model training and evaluation to quantify prediction error and guide the adjustment of model parameters to ensure that they are as close to actual network requirements as possible.

[0028] MAE (Mean Absolute Error) is another metric used to evaluate the performance of a forecasting model. It calculates the average absolute difference between the predicted and actual values. MAE directly reflects the average level of forecast error and helps understand and improve the consistency of a forecasting model's performance across different data points.

[0029] R 2 , Coefficient of Determination, is used to indicate the proportion of variation explained by a model to the total variation, and its value ranges from 0 to 1. In model evaluation, R 2 It reflects the explanatory power of the model and the degree of fitting of the data. The closer the value is to 1, the better the model fits the data.

[0030] IoT, Internet of Things, is a conceptual collection of network devices, referring to physical objects that collect and exchange data by embedding electronic devices, software, sensors and network connections.

[0031] It should be noted that the 5G-based communication network deployment method and apparatus thereof in the present disclosure can be used in the field of 5G technology. In the case of 5G-based communication network deployment, it can also be used in any field other than the field of 5G technology. In the case of 5G-based communication network deployment, the present disclosure does not limit the application field of the 5G-based communication network deployment method and apparatus thereof.

[0032] It should be noted that in this disclosure, when collecting and analyzing customer information, the corresponding operation entrance is provided for users to choose to agree or reject the automated decision-making results; if the user chooses to reject, the expert decision-making process will be entered.

[0033] The following embodiments of the present invention can be applied to various systems / applications / devices deployed based on 5G communication networks. The present invention can be applied to the field of communication network optimization and deployment, especially for the intelligent planning, optimization and dynamic adjustment of 5G communication networks. Specific application scenarios may include but are not limited to: smart city and Internet of Things infrastructure construction, large-scale events and high-density crowd area network security scenarios, unmanned vehicle communication network optimization scenarios, etc. For example, for smart city and Internet of Things infrastructure construction scenarios, 5G network is a key technology for realizing the Internet of Everything. The embodiments of the present application can intelligently predict network demand through the collection and analysis of multi-dimensional data, optimize base station layout and edge computing resource configuration, ensure efficient and stable communication connections, and support the effective deployment and management of large-scale Internet of Things devices.

[0034] The present invention predicts network demand by integrating the ARIMA model and the LSTM model, combining real-time network performance monitoring and dynamic adjustment mechanisms. By leveraging the ability of the ARIMA model to process time series data and the advantage of the LSTM model in capturing complex relationships, the fusion prediction results significantly improve the accuracy of future demand forecasts for 5G networks, laying the foundation for the efficient allocation of network resources. The present invention can dynamically adjust the base station layout and the configuration strategy of edge computing nodes based on the predicted network demand and actual network performance indicators, realize network resource optimization in different regions and at different time points, and ensure the service quality and resource utilization efficiency of the 5G network.

[0035] At the same time, the present invention can also respond to changes in network status in a timely manner through real-time monitoring and feedback mechanisms, quickly adjust network deployment plans, enhance the 5G communication network's ability to respond to emergencies, and improve the adaptability and flexibility of the overall network. By optimizing network performance, it can enhance the user's network experience in different scenarios, while meeting business needs such as high-density data transmission and low-latency real-time interaction, and promote the in-depth application of 5G networks in multiple industries.

[0036] The present invention will be described in detail below with reference to various embodiments.

[0037] Example 1

[0038] According to an embodiment of the present invention, an embodiment of a 5G-based communication network deployment method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0039] The 5G-based communication network deployment method embodiment provided in Example 1 of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 The following is a hardware block diagram of a computer terminal (or mobile device) for implementing a 5G-based communication network deployment method. Figure 1 As shown, the computer terminal 10 (or mobile device) may include one or more ( Figure 1The computer system includes a processor 102 (shown as 102a, 102b, ..., 102n) (the processor 102 may include but is not limited to a microcontroller unit (MCU) or a programmable logic device (FPGA)), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, the computer system may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS), a network interface, a power supply, and / or a camera. It will be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components than shown, or with Figure 1 Different configurations shown.

[0040] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuitry". The data processing circuitry may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuitry may be a single independent processing module, or may be incorporated in whole or in part into any of the other components of the computer terminal 10 (or mobile device). As described in the embodiments of the present application, the data processing circuitry serves as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0041] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the 5G-based communication network deployment method in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, implementing the above-mentioned 5G-based communication network deployment method. The memory 104 may include a high-speed random access memory and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely located relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0042] The transmission device 106 is configured to receive or transmit data via a network. A specific example of the aforementioned network may include a wireless network provided by the communications provider of the computer terminal 10. In one embodiment, the transmission device 106 includes a network interface controller (NIC), which can be connected to other network devices via a base station to enable communication with the Internet. In another embodiment, the transmission device 106 may be a radio frequency (RF) module, which is configured to communicate with the Internet wirelessly.

[0043] The display may be, for example, a touch screen liquid crystal display (LCD), which enables a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0044] Under the above operating environment, this application provides Figure 2 The 5G-based communication network deployment method shown. Figure 2 is a flowchart of an optional 5G-based communication network deployment method according to an embodiment of the present invention, such as Figure 2 As shown, the method includes the following steps:

[0045] Before the deployment of 5G communication networks, it is necessary to build a multi-source information collection network through various types of sensors to obtain multi-dimensional data sets, and use the pre-processed data sets to train multiple types of network demand prediction models.

[0046] Optionally, before inputting the preprocessed multi-dimensional traffic demand data set into the pre-trained multi-category network demand prediction model, it also includes: configuring communication sensors and communication network protocols in each network coverage area based on the geographical environment factors of different network coverage areas, the density of user distribution and the network service demand characteristics of each area, and building a multi-source information collection network; using the multi-source information collection network to collect user location, traffic demand data, network load, environmental data and device type data to obtain a historical multi-dimensional traffic demand data set; using a group mean filling strategy to process missing values in the historical multi-dimensional traffic demand data set, and normalizing the processed historical multi-dimensional traffic demand data set through a minimum and maximum scaling strategy to obtain a preprocessed historical multi-dimensional traffic demand data set; organizing the preprocessed historical multi-dimensional traffic demand data set into a time series data format, and training the ARIMA model and LSTM model; and using the trained ARIMA model and LSTM model as network demand prediction models.

[0047] In this embodiment, taking into account the geographical environment of different network coverage areas in the 5G communication network, the density of user distribution, and the diversity of network service demand characteristics, a multi-source information collection network is provided to comprehensively obtain various dimensions of network operation status data. First, based on the geographical environment factors of each area, such as topography, building distribution, population density, etc., the configuration location of communication sensors is planned to ensure that the sensors can effectively cover and perceive the network conditions in a specific area; second, based on the density of user distribution, the sensor density is reasonably allocated, and more sensors are deployed in key areas (such as commercial centers and transportation hubs) to capture more user behavior data; finally, based on the network service demand characteristics of each area, appropriate communication network protocols are selected, such as using low-latency communication technology in areas with high real-time requirements, thereby building a multi-source information collection network that can comprehensively and real-timely reflect the network status.

[0048] It should be noted that when planning sensor locations, multiple factors need to be considered, such as the geographic environment of the network coverage area (including topography, building distribution, etc.), the density of user distribution, and the business needs of different regions. Appropriate communication technologies (such as ZigBee, Bluetooth, Wi-Fi, or other proprietary communication protocols) are selected, and a multi-source information collection network is constructed through carefully planned sensor layouts and reliable communication connections. This multi-source information collection network collects various key data, including user location information for analyzing user distribution and mobility patterns; traffic demand data for predicting future traffic trends; network load data for evaluating the carrying capacity of existing networks; environmental data for understanding the impact of external factors on network performance; and device type data for analyzing the differences in network resource requirements of different devices. This creates a multi-dimensional traffic demand dataset that comprehensively reflects the historical network status.

[0049] Next, this embodiment preprocesses the collected historical multi-dimensional traffic demand dataset to ensure data quality and improve the training effect of the prediction model. First, a group mean filling strategy is used to address missing values that may appear in the dataset. This strategy divides the dataset into several groups based on the intrinsic properties of the data, such as grouping by user type, geographic location, or business type. The average value of the non-missing data within each group is then calculated to fill the missing values. This maintains data integrity while minimizing prediction bias caused by missing data.

[0050] It should be noted that when processing historical multidimensional traffic demand datasets, missing values are handled using a group mean filling method. This involves first dividing the multidimensional dataset into different groups based on their characteristics or attributes, such as user type, region, or time period. Within each group, the mean of the non-missing values is calculated and used to fill in the missing values. This maintains data integrity while minimizing prediction bias caused by missing data. This method can, to a certain extent, preserve the overall distribution characteristics of the multidimensional data and avoid bias introduced by simple filling.

[0051] In addition, when preprocessing historical multi-dimensional traffic demand datasets, outliers must be addressed using statistical methods, such as the Z-score method or boxplot method based on the mean and standard deviation. Taking the Z-score method as an example, by calculating the distance between each data point and the mean (in units of standard deviation), outliers are identified and then processed according to the specific situation. For example, statistical measures such as the median and mode can be used for replacement, or corrections can be made based on the contextual logic of the multi-dimensional data to ensure the quality and reliability of the multi-dimensional data.

[0052] Subsequently, the dataset is normalized using a minimum-maximum scaling strategy, scaling the numerical range of all data to a uniform scale (usually between [0, 1]), eliminating the impact of different feature dimensions, thereby improving the efficiency of model training and the accuracy of predictions. Here, the normalization expression can be expressed as: x' = (x-x_min) / (x_max-x_min), where x' is the normalized historical multi-dimensional traffic demand data, x is the original value in the historical multi-dimensional traffic demand dataset, x_max is the maximum value in the historical multi-dimensional traffic demand dataset, and x_min is the minimum value in the historical multi-dimensional traffic demand dataset.

[0053] Optionally, the preprocessed historical multi-dimensional traffic demand data set is organized into a time series data format, and the steps of training the ARIMA model and the LSTM model include: arranging the preprocessed historical multi-dimensional traffic demand data set according to predetermined time intervals, establishing an association relationship between each historical traffic demand data point and the corresponding timestamp, and obtaining traffic demand time series data; dividing the traffic demand time series data into a training set and a validation set according to a predetermined ratio, and using the training set to train the initial ARIMA model of the selected model parameters, and estimating the parameter values of the ARIMA model through the maximum likelihood estimation strategy; using the validation set to evaluate the trained ARIMA model, and adjusting the parameter values of the ARIMA model according to the evaluation results to obtain a trained ARIMA model; converting the preprocessed historical multi-dimensional traffic demand data set into a three-dimensional tensor form, and training the LSTM model.

[0054] After completing the data preprocessing, this embodiment organizes the preprocessed historical multi-dimensional traffic demand data set into a time series data format for the training of the ARIMA model and the LSTM model. Specifically, the data set is first arranged according to a predetermined time interval (for example, hourly, daily or weekly) to ensure that each traffic demand data point is associated with its corresponding timestamp to form traffic demand time series data. Then, the generated traffic demand time series data is divided according to a predetermined ratio (for example, 70% for the training set and 30% for the validation set). When training the ARIMA model, based on the training set data and through the maximum likelihood estimation strategy, the optimal parameter combination of the ARIMA model is determined, including the autoregressive order, the difference order and the moving average order, to ensure that the model can capture the trend and seasonal changes of the time series data. Here, it is necessary to preliminarily determine the parameter range by analyzing the statistical characteristics of the time series data such as the autocorrelation function and the partial autocorrelation function, and using methods such as grid search and information criteria (such as AIC, BIC, etc.). During the training process, parameters are continuously adjusted to optimize the ARIMA model's fit. Convergence of the ARIMA model can be determined by observing changes in the loss function during training. The trained ARIMA model is evaluated using validation data, and relevant metrics such as mean squared error (MSE) and mean absolute error (MAE) are calculated. Based on the evaluation results, parameters are further adjusted or the ARIMA model is optimized until a high-performance ARIMA model is obtained that can accurately predict historical traffic demand data trends.

[0055] Furthermore, this embodiment also provides the latest traffic demand data as input to the trained ARIMA model in time series order. Future network demand is predicted through operations such as autoregression, differencing, and moving average. Specifically, this embodiment can use the ARIMA model to gradually extrapolate future traffic demand values based on the patterns of historical traffic demand data. For example, based on the traffic data of the past few hours and the autoregressive coefficients of the ARIMA model, the traffic demand for the next hour can be predicted. The prediction results are then adjusted and optimized by combining the differencing and moving average operations to obtain the ARIMA model prediction results, which include the future network demand forecast and prediction error.

[0056] This embodiment also converts the preprocessed historical multidimensional traffic demand dataset into a three-dimensional tensor format to facilitate LSTM model training. It should be noted that when converting the preprocessed historical multidimensional traffic demand dataset into a three-dimensional tensor format, if using NumPy, the data dimensions can be rearranged based on application requirements and data processing workflow using the reshape method. If using PyTorch, the historical multidimensional data can be converted into a tensor first, and then the dimensions can be adjusted using the view method. It is also important to note that the order of the historical multidimensional data should conform to the order of samples, then time steps, and finally the features of the historical multidimensional data. After conversion to a three-dimensional tensor format, the historical multidimensional data is more effectively organized and represented in spatial and temporal dimensions, where each of the three dimensions can correspond to different meanings. During training, the LSTM (Long Short-Term Memory) model continuously adjusts the connection weights between neurons to learn the patterns and regularities inherent in the historical multidimensional dataset, minimizing the error between predicted and true values. This process involves numerous computational and optimization operations, such as forward propagation to calculate predictions and backpropagation to update weights. Through multiple iterations of training, the LSTM model gradually converges to an optimal state, thus acquiring stronger predictive capabilities. New three-dimensional tensor data is input into the LSTM model. Based on previously learned knowledge and patterns, the LSTM model processes and analyzes the input three-dimensional tensor data, comprehensively considering the interactions of various factors and dynamic changes in the time series to produce LSTM model predictions. These predictions include future network demand forecasts and trend capture.

[0057] Ultimately, the trained ARIMA model and LSTM model will be integrated as a network demand forecasting model. By fusing the prediction results of the two models, it not only takes into account the linear trend of time series data but also captures the nonlinear dynamic changes of the data, thereby achieving more accurate network demand forecasting.

[0058] It should be noted that in the fusion process of the ARIMA model and the LSTM model, the initial weight distribution is set based on the historical performance of the two models. For example, if the LSTM model performs better in handling nonlinear relationships, a higher initial weight is given. After the initial weight distribution is completed, the weights are dynamically adjusted according to the performance of the ARIMA model and the LSTM model in different time periods. The weight ratio is optimized by regularly evaluating the prediction error (such as the mean square error (MSE) or the mean absolute error (MAE). Cross-validation and comparative analysis methods are used to verify the accuracy and robustness of the fused ARIMA model and LSTM model. The prediction results are compared with the single model and actual data, and the error indicators are evaluated to confirm the improvement effect.

[0059] The fusion model's predictions are compared with historical data and actual historical network demand data. This includes plotting time series graphs to visually demonstrate the changing trends between the predicted and actual values. For example, if the predicted values closely track actual traffic demand during specific time periods (such as holidays or major events), the fusion model's demand forecasts for that time period are relatively accurate. The fusion model's predictions are then compared with predictions generated using either the ARIMA model or the LSTM model alone. Various error metrics (such as mean squared error (MSE) and mean absolute error (MAE)) are calculated to quantify the advantages of the fusion model over the individual models. For example, if the MSE of the fusion model is significantly lower than that of the individual models, this indicates improved prediction accuracy. This comprehensive and in-depth comparative analysis clearly demonstrates network demand changes, potential issues, and development trends, thereby generating a detailed network status report.

[0060] In step S201 , the pre-processed multi-dimensional traffic demand data set is input into a pre-trained multi-class network demand prediction model, and the network demand within a specified time period in the future is evaluated using the network demand prediction model.

[0061] Before inputting the data into the prediction model, this embodiment preprocesses the collected traffic demand data, including data cleaning, missing value filling, outlier detection and processing, and data normalization, to ensure the quality and applicability of the model input data. The preprocessed multi-dimensional traffic demand dataset not only includes information such as user location, traffic demand, network load, environmental data, and device type, but also undergoes appropriate formatting and feature engineering to ensure that the data better meets the input requirements of the prediction model.

[0062] This embodiment uses the ARIMA and LSTM models as tools for network demand forecasting. ARIMA (autoregressive integrated moving average model) excels at processing time series data and can capture trends and cyclical changes in data; LSTM (long short-term memory network) is a deep learning model that is particularly good at processing long-term dependencies in sequences and has powerful analytical capabilities for complex multi-dimensional data. During the training process, the two models are fitted using historical traffic demand datasets, and the prediction error is minimized by continuously adjusting the parameters until the model converges, thereby obtaining a network demand forecasting model that can accurately predict future network demand.

[0063] After the model training is completed, this embodiment will evaluate the prediction effects of the ARIMA model and the LSTM model, and compare the performance differences of the two models in predicting future network demand. Subsequently, the prediction results of the two models are fused using a weighted average method, aiming to combine the advantages of both and improve the accuracy and stability of the prediction. During the weighted fusion process, this embodiment may dynamically adjust the weight distribution based on the performance history of the model in different scenarios to achieve the best prediction effect. After completing the model fusion, this embodiment inputs the preprocessed multi-dimensional traffic demand data set into the fused network demand prediction model, and outputs the future network demand forecast within the specified time length. The prediction results not only include the size of future traffic demand, but may also involve the changing trend of traffic demand and potential periodic fluctuations, providing a key reference basis for subsequent base station layout optimization and edge computing node configuration.

[0064] Based on the results of future network demand forecasts, this embodiment generates a detailed network status report. The report not only includes the predicted traffic demand data, but also may cover multi-dimensional information such as the geographical distribution, time distribution, and service type distribution of traffic demand, as well as comparative analysis of the forecast results, such as comparison with historical data and comparison of results from different forecast models, to comprehensively evaluate the expected status of the network and possible problems.

[0065] Step S202 , the network demand forecast results output by various types of network demand forecast models are integrated to generate a network status report, and a base station layout is formulated based on the network status report and various network traffic influencing factors to obtain a base station layout optimization plan.

[0066] Optionally, the steps of fusing the network demand forecast results output by various types of network demand forecast models to generate a network status report include: dynamically adjusting the corresponding model weight ratio according to the forecast error of each type of network demand forecast model in different time periods; based on the model weight ratio, adopting a weighted average strategy to fuse the network demand forecast results output by each type of network demand forecast model to obtain future network demand; and comparing the future network demand with historical network demand data to generate a network status report.

[0067] First, this embodiment can dynamically adjust the weight ratio of each type of network demand forecasting model (such as ARIMA model and LSTM model) according to its forecast performance in different time periods (such as weekdays, weekends, holidays, etc.). The adjustment process is based on the forecast error of the model in each time period, aiming to more fairly and accurately reflect the actual forecasting capabilities of the two models. By regularly back-testing and calculating indicators such as the mean square error (MSE) or mean absolute error (MAE) of different models on historical data, the forecast accuracy of the model can be quantified. For example, if the LSTM model shows a smaller forecast error in a certain time period, then when the forecast fusion is performed in that time period, the LSTM model will be given a higher weight. After determining the weight ratio of the model, this embodiment adopts a weighted average strategy to fuse the output results of each type of network demand forecasting model. The forecast results of each model will be multiplied by its corresponding weight and then summed to obtain a comprehensive forecast value, i.e., future network demand. It can effectively combine the advantages of the ARIMA model in capturing the trend and seasonality of time series data and the ability of the LSTM model to handle complex nonlinear relationships, so that the forecast results are more in line with the changes in actual network demand. It should be noted that during the fusion process of the ARIMA and LSTM models, the initial weight assignment is set based on the historical performance of the two models. For example, if the LSTM model performs better in handling nonlinear relationships, it is given a higher initial weight. Cross-validation and comparative analysis methods are used to verify the accuracy and robustness of the fused ARIMA and LSTM models. The forecast results are compared with those of the individual models and actual data, and error indicators are evaluated to confirm the effectiveness of the improvements. For seasonal fluctuations and sudden events, cyclical component analysis is added to help the fused model understand regular changes. The adaptability is tested by introducing sudden event datasets to ensure that efficient performance is maintained.

[0068] After obtaining the future network demand, the next step is to compare and analyze this forecast result with the historical network demand data. This embodiment evaluates the accuracy and potential deviation of the forecast by comparing the predicted value with the actual network demand data over a period of time. This comparative analysis is not limited to numerical comparison, but more importantly, it is to analyze the reasons for the differences, such as whether it is affected by special events (such as major events, changes in weather conditions), and the prediction capabilities of the model in different scenarios. Based on these analyses, a detailed network status report is generated, which includes: future network demand forecast (providing a comprehensive forecast of future network demand, covering traffic peaks, trend changes, regional demand distribution, etc.), forecast accuracy assessment (quantifying the difference between the forecast results and historical actual data, and evaluating the accuracy and stability of the forecast model), and optimization suggestions (combining the forecast error with network demand analysis to propose optimization suggestions for base station layout, edge computing node configuration, etc., to improve resource utilization efficiency and overall network performance).

[0069] In this embodiment, the prediction results are compared with historical data, and the prediction results of the fusion model are compared with the actual historical network demand data, including drawing a time series graph to intuitively show the changing trend of the predicted value and the true value. The prediction results of the fusion model are respectively compared with the prediction results made using only the ARIMA model or the LSTM model. The advantages of the fusion model over the single model are quantified by calculating various error indicators (such as mean square error MSE and mean absolute error MAE). For example, if the MSE of the fusion model is significantly lower than that of the single model, it indicates that the prediction accuracy is improved. Through this comprehensive and in-depth comparative analysis, the changes in network demand, potential problems and development trends can be clearly presented, thereby generating a detailed network status report.

[0070] Optionally, a base station layout is formulated based on the network status report and a variety of network traffic influencing factors to obtain a base station layout optimization solution, including: performing different dimensional analysis on the network status report to obtain high network demand areas, base station location areas to be optimized, and key business focus areas, wherein the high network demand area refers to an area where the traffic demand assessment value in the network status report is greater than a preset traffic demand threshold, and the traffic demand assessment value is determined based on user-side traffic demand parameters, population density, and current regional signal strength; the base station location area to be optimized refers to an area where the base station coverage assessment value in the network status report is lower than a preset base station coverage threshold, and the base station coverage assessment value is based on The scheduled survey factors and network planning objectives are determined. The scheduled survey factors include the current geographical location of the base station, surrounding environmental factors and equipment operating status. The key business focus area refers to the area where the specific business evaluation value in the network status report is higher than the preset business threshold. The specific business evaluation value is determined based on the traffic proportion of the specific business, business type distribution and business growth trend parameters; determine the new base station plan based on the high network demand area, formulate the base station migration plan based on the base station location area to be optimized, and formulate the base station parameter adjustment plan based on the demand characteristics of the key business focus area; integrate the new base station plan, base station migration plan and base station parameter adjustment plan to formulate the base station layout optimization plan.

[0071] In this embodiment, multi-dimensional analysis based on network status reports is a prerequisite for developing a base station layout optimization plan. Network status reports are a comprehensive dataset that includes multi-dimensional data collected by various sensors, such as user location, traffic demand, network load, environment, and device type data. These data are fused with the prediction results of ARIMA and LSTM models to reflect future network demand and current network performance.

[0072] When analyzing the report, this embodiment first identifies areas with high network demand, that is, areas where the traffic demand assessment value in the network status report exceeds the preset traffic demand threshold (that is, find out those areas where the traffic demand is continuously high, close to or exceeds the carrying capacity), which are usually places where users are concentrated and activities are frequent, such as commercial centers, large event venues, etc. By combining the signal strength and coverage range with the base station coverage data, check whether there are areas with weak signals but high traffic demand. These places may have a sudden increase in demand due to dense population or specific events, and the existing base stations cannot meet the demand. These areas are divided into high network demand areas, and the traffic demand assessment value is calculated based on the user-side traffic demand parameters (such as data transmission volume, multimedia service usage, etc.), population density and regional signal strength, and is used to measure the supply and demand balance of network resources in the region. The identification of high network demand areas helps to prioritize network resources and ensure smooth communication in important areas.

[0073] Based on indicators such as base station signal strength, coverage range and signal quality in the network status report, identify areas where base station coverage has blind spots, unstable signals or coverage does not match actual needs.

[0074] Organize a professional team to conduct on-site surveys, and record the current geographical location of the base station, the surrounding environment (including building height, obstruction, etc.) and the operating status of the equipment. Compare the survey data with the network planning goals to determine which base stations have poor network performance due to poor geographical location, and consider migration plans to divide these areas into areas where base station locations need to be optimized. In this embodiment, when determining the base station location area to be optimized, it determines the area where the base station coverage evaluation value is lower than the preset base station coverage threshold. The base station coverage evaluation value is calculated by comparing the current geographical location of the base station, the surrounding environmental factors (such as building obstruction, topography, etc.) and the operating status of the equipment (such as transmission power, receiving sensitivity) with the network planning goals. The identification of the base station location area to be optimized helps to discover the shortcomings of network coverage and take measures to strengthen signal coverage.

[0075] Finally, identify key business focus areas based on areas where specific business assessment values exceed preset business thresholds. Specific business assessment values take into account the traffic proportion, business type distribution, and business growth trends of specific businesses (such as high-definition video streaming and virtual reality applications), and are used to reveal areas in the network that are particularly sensitive to certain business types. The identification of key business focus areas helps to optimize network configuration to meet the high-quality communication requirements of specific businesses. This embodiment can determine areas with high demand and rapid growth for specific businesses based on information such as the traffic proportion, business type distribution, and growth trends of different businesses in the network status report. For example, the demand for video streaming services may be particularly high in certain areas. Collect and analyze the usage habits of users in the region, such as peak hours, popular applications, etc., to understand users' preferences for different types of services and demand change trends.

[0076] Based on the above analysis results, this embodiment develops a base station layout optimization plan, covering three aspects: a new base station plan, a base station migration plan, and a base station parameter adjustment plan. For areas with high network demand, this embodiment develops a new base station plan to increase network capacity and improve service quality by adding base stations or improving base station performance (such as increasing transmit power or introducing more advanced communication technologies), ensuring that the regional network needs are met. For areas with high network demand, this embodiment can collect multi-dimensional data on high-network demand areas, including population density, commercial activity distribution, user traffic hotspot distribution (such as peak network usage hours and traffic data for specific locations such as large shopping malls, office buildings, and university dormitories), and existing network coverage (through signal strength testing, network speed monitoring, etc.). A comprehensive analysis is then performed to identify areas with weak network coverage but strong demand. Based on geographical features, building layout, and user distribution patterns, the general type of new base stations is determined (such as macro base stations for larger coverage areas and micro base stations for smaller hotspots). Using network planning tools, combined with factors such as topography, network coverage and capacity improvement are simulated under different base station layout plans to select the optimal new base station plan.

[0077] A base station relocation plan is developed for areas where base station locations are being optimized. Based on the results of on-site surveys and assessments, base stations in areas with poor coverage or excessive load are relocated to locations with better signal transmission, thereby improving network coverage and stability. A professional on-site survey team is organized for areas where base station locations need to be optimized. The surveyor should conduct a detailed inspection of the base stations within the target area, recording the base station's current location, surrounding environment (including building height, obstruction, and electromagnetic interference sources), equipment operating status (parameters such as transmit power and receive sensitivity), and network performance indicators within the base station's coverage area (such as coverage, signal strength, call drop rate, and handover success rate). The data obtained from the on-site survey is compared and analyzed with network planning and optimization goals to identify base stations whose network performance fails to meet requirements due to poor location. Target relocation locations for these base stations are determined by further analyzing factors such as surrounding available space resources, topography, and synergy with other base stations. Once the target relocation locations are determined, a detailed base station relocation plan is developed.

[0078] For key service focus areas, this embodiment develops a base station parameter adjustment plan. Based on the varying network performance requirements (e.g., bandwidth and latency) of different service types, this plan dynamically adjusts relevant base station parameters, such as frequency band allocation, signal power control, and multiple access technology selection, to enhance the service capabilities of specific services. In-depth research and analysis is conducted on key service focus areas, comprehensively collecting service data. This includes the traffic distribution of different services (e.g., the volume of video, voice, and data download services at different times and locations), user behavior patterns (the frequency, duration, and time period of user usage of various services), and the specific network performance requirements of these services (e.g., the requirements of HD video services for low latency and high bandwidth, and the requirements of IoT services for stable connections and low power consumption). Based on the collected service data, the demand characteristics of the key service areas are carefully analyzed to identify the key network performance requirements of different services. For example, some areas may have extremely high bandwidth and real-time requirements due to the large number of HD video live broadcasts, while other areas may have special requirements for network connection stability and coverage due to the concentrated deployment of IoT devices. Based on these demand characteristics, a base station parameter adjustment plan is developed.

[0079] Integrating the above three solutions, this embodiment formulates a base station layout optimization plan to ensure that the addition of new base stations, base station migration and parameter adjustment can complement each other and jointly improve the overall performance of the network. For example, the location and number of new base stations may affect the migration decisions and parameter adjustment ranges of surrounding base stations; after the base station is migrated, changes in the coverage area may require corresponding adjustments to the parameters of the relevant base stations; and the adjustment of base station parameters may also have a feedback effect on the layout of new base stations and migrated base stations. By comprehensively weighing these factors, comprehensively coordinating the layout of new base stations, the migration path of base stations and the specific settings of parameter adjustments, a scientific, reasonable and feasible base station layout optimization plan is formulated to ensure that the implementation of the plan can maximize the efficiency of network resource utilization while minimizing the impact on the existing network architecture and achieving a smooth transition.

[0080] Step S203: Based on the network status report and the base station layout optimization plan, an edge computing node configuration strategy is formulated, wherein the edge computing node configuration strategy is used to configure network node resources in each area.

[0081] Optionally, step S203 includes: analyzing the network status report and the base station layout optimization plan to obtain the network demand characteristics, network change trends and base station layout of different regions affecting the edge computing node configuration of each region, and obtaining the first type of node influencing factors; obtaining the network demand forecast results, the current user terminal distribution status and the impact of each business characteristic on the edge computing node configuration of each region, and obtaining the second type of node influencing factors; based on the first type of node influencing factors and the second type of node influencing factors, dividing the overall area into different types of sub-regions, and formulating corresponding node deployment strategies for the regional characteristics of different types of sub-regions; allocating computing resources according to regional business demand and number of users to obtain the node resource allocation plan for each region, combining the node deployment strategies of different regions with the node resource allocation plans of each region to obtain the edge computing node configuration strategy.

[0082] Step S203 first involves an in-depth analysis of the network status report and the base station layout optimization plan, aiming to extract the key factors that affect the configuration of edge computing nodes in each region. The network status report contains a wealth of information, such as network traffic, latency, signal strength, and network coverage, reflecting the actual operating status of the network. The base station layout optimization plan clarifies the strategy for base station adjustment, including the addition, migration, and parameter modification of base stations, which is directly related to the network coverage and data transmission speed of each region. In this link, this embodiment focuses on understanding the characteristics of network requirements in different regions, such as the high dependence of commercial areas on real-time data processing and low-latency services, and the requirements of residential areas for stable connections and sufficient bandwidth. At the same time, through comparative analysis, the dynamic changes in network demand trends are identified, such as traffic peaks during holidays or major events, and nighttime troughs. More importantly, it is necessary to evaluate how the base station layout optimization plan affects the configuration of edge computing nodes, including the changes in data transmission paths caused by changes in base station locations, and the impact of the increase in the number of base stations on the computing load of edge nodes.

[0083] The first step is to analyze the network demand characteristics, network change trends, and the potential impact of base station layout optimization solutions on edge computing node configuration, as reflected in the network status report. These factors constitute the first category of node influencing factors. Network demand characteristics may include peak data traffic times, user preference for specific services, and sensitivity to network latency. Network change trends involve temporal and seasonal fluctuations in network traffic, as well as long-term growth trends. Base station layout optimization not only changes the physical network architecture but also indirectly affects the workload and data processing efficiency of edge computing nodes. Comprehensively and meticulously analyze the network status report, delving into detailed information on network traffic distribution (including traffic volume and fluctuation patterns by region and time period), signal strength and quality data (such as average signal strength and signal fluctuations in each region), network latency (latency levels in different service scenarios), and network congestion information (frequency and duration of congestion). In conjunction with the base station layout optimization solution, analyze the network demand characteristics of different regions from multiple dimensions. For example, commercial areas may have higher requirements for high bandwidth and low latency, while residential areas may prioritize network stability and coverage. Consider the impact of new base station networks. We analyze how factors such as base station location, number, and coverage interact with the configuration of edge computing nodes. We consider how to adjust the number, location, and resource allocation of edge computing nodes based on base station layout. Based on these findings, we analyze differences in service types, user density, and data transmission requirements across different regions (such as commercial, residential, and industrial areas), thereby clarifying the network demand characteristics of each region. By comparing network status data over different time periods and integrating regional development trends (such as urban construction planning and industrial upgrading), we explore the changing trends of network demand over time, including growth or decline in demand and demand fluctuation patterns. Furthermore, we conduct an in-depth analysis of the impact of base station layout optimization solutions on the configuration of edge computing nodes in each region, considering how factors such as base station location changes and coverage adjustments affect data transmission paths and processing methods. This analysis then analyzes the changes in demand for edge computing node computing power, storage capacity, and number of connections. Combined with these analytical results, we comprehensively understand the network demand characteristics, network change trends, and the impact of base station layout on the configuration of edge computing nodes in each region.

[0084] By comprehensively considering these first-category node influencing factors, this embodiment can preliminarily identify which areas may require additional computing resources, which areas' network change trends prompt adjustments to edge node configurations, and which areas have new demands for edge computing nodes due to base station layout optimization.

[0085] Next, this embodiment requires obtaining the second type of node influencing factors, namely, the impact of network demand forecasts, current user-end distribution, and various service characteristics on edge computing node configuration. Network demand forecasts provide an estimate of network traffic over the next period of time, which is crucial for determining the scale and capacity of edge nodes in each region. The user-end distribution reveals the geographic distribution and density of terminal devices, helping to determine the optimal layout of edge computing nodes and ensure that all users can enjoy efficient, low-latency services. Service characteristics cover the different network requirements of different types of services (such as video streaming, online gaming, and IoT data transmission), which helps to fine-tune the configuration of computing, storage, and network resources of edge nodes. In this embodiment, from the perspective of network demand, the traffic demand, network latency, and bandwidth requirements of different regions can be analyzed. For example, some areas have extremely high real-time requirements, such as intelligent transportation hubs, and require low-latency network support. From the perspective of user distribution, attention is paid to user density and distribution patterns. For example, users are highly concentrated in urban central business districts, while users are more sparsely distributed in remote suburbs. From the perspective of service characteristics, the main service types of different regions can be distinguished. For example, some areas are dominated by high-definition video services, while others are dominated by IoT data transmission. Based on these comprehensive analyses, the overall region is divided into different types of sub-regions, and then corresponding node deployment strategies are formulated according to the characteristics of different regions. After completing the regional classification and node deployment strategy formulation, computing resources are allocated according to regional business needs and the number of users, thus obtaining a node resource allocation plan for each region.

[0086] Based on a full understanding of the factors affecting the two types of nodes, this embodiment further subdivides the overall area into different types of sub-areas, such as high-density commercial areas, low-density residential areas, industrial areas, etc., and formulates a special node deployment strategy based on the characteristics of each sub-area. For example, high-density commercial areas may need to deploy more high-performance edge computing nodes to cope with sudden traffic peaks; while low-density residential areas may focus more on node coverage and stability. Through analysis of regional business needs and the number of users, this embodiment also formulates a node resource allocation plan for each area to ensure that the edge nodes in each area can obtain computing resources that match them, while avoiding resource waste and overload risks.

[0087] Integrating these two approaches requires in-depth analysis of the inherent connections and mutual influences between node deployment strategies in different regions and their node resource allocation schemes. On the one hand, node deployment strategies in different regions determine the basic framework and focus of resource allocation, and different deployment models directly impact the distribution and demand of computing resources. On the other hand, regional node resource allocation schemes provide resource guarantees for the implementation of node deployment strategies. Reasonable resource allocation maximizes the effectiveness of node deployment. Node deployment strategies and node resource allocation schemes are linked to establish a mapping between factors such as node location, number, and type, and resource allocation. For example, high-performance computing nodes deployed in central business districts are allocated more computing resources and network bandwidth, while storage nodes deployed in remote areas are allocated more storage resources. A joint optimization algorithm is used to collaboratively optimize and adjust the node deployment strategy and resource allocation scheme. Aiming to optimize overall network performance, the node deployment location, number, and resource allocation ratio are continuously adjusted, taking into account factors such as inter-node load balancing, data transmission efficiency, and user experience, until a preferred edge computing node configuration strategy is found.

[0088] When faced with sudden high-traffic events, such as live sports broadcasts, the computing power and storage capacity of edge computing nodes in the relevant areas are enhanced in advance. Load balancing strategies are used to distribute traffic pressure during the event, and the nearest edge computing node is intelligently dispatched to handle video streaming requests, ensuring service quality and user experience. After the event, further data collection and analysis are performed to optimize future response strategies, ensuring rapid response and efficient resource utilization even in the face of sudden increases in network demand.

[0089] Step S204: monitor the network operation status after configuring the regional network nodes, collect user feedback data, and use a linear regression model to analyze the network operation status and user feedback data to obtain network performance indicators.

[0090] Step S204 in this embodiment can monitor the network operating status in real time and collect user feedback to build and utilize a linear regression model to analyze network performance, providing a basis for dynamically adjusting the network deployment plan. That is, after configuring the edge computing nodes and base station layout, the system will continuously monitor the network's operating status, including but not limited to key performance indicators such as network latency, throughput, and packet loss rate, so as to provide real-time understanding of network service quality and performance. Through online surveys and built-in feedback mechanisms in user applications, this embodiment will collect user experience and satisfaction information for network services. This feedback will include but not be limited to perceived evaluations of network speed, stability, latency, and other aspects, which are important dimensions for measuring network service quality.

[0091] Optionally, the linear regression model is constructed in the following manner: extracting historical user feedback data and historical network performance data, and dividing the historical user feedback data and historical network performance data into a model training set and a model test set according to a time sequence or random sampling principle; initializing the parameters of the linear regression model based on pre-defined independent variables and dependent variables, wherein the independent variables include: network delay, throughput, packet loss rate, and the dependent variables include target network performance indicators; using the model training set to train the initialized linear regression model, and iteratively adjusting the parameters of the linear regression model by the gradient descent method until the loss function converges or reaches a preset number of iterations, completing the training of the linear regression model and obtaining a trained linear regression model, wherein, in each iteration process, the loss function value under the current parameters is calculated, and the parameters of the linear regression model are updated according to the loss function value; evaluating the trained linear regression model by the model test set, calculating the evaluation index value, adjusting the hyperparameters of the linear regression model shown, and obtaining a verified linear regression model.

[0092] First, based on historical user feedback and network performance data, we divide this data into a model training set and a model test set, either in chronological order or using random sampling. The model training set is used to train the linear regression model, while the model test set is used to evaluate the model's generalization capabilities. Before model training, we define independent variables, including network performance metrics such as network latency, throughput, and packet loss rate, and the dependent variable, the target network performance metric.

[0093] This embodiment can divide all historical user feedback and historical performance data into training sets and validation sets according to a certain ratio (for example, 70%-30%) based on the principle of chronological order or random sampling, to ensure that the training set contains sufficiently rich information for model learning, while the validation set is used to evaluate the generalization ability of the model. In the data preprocessing stage, the group mean filling method is used for missing values. The historical user feedback and historical performance data are divided into different groups according to their characteristics or attributes, and the mean of the non-missing values in each group is calculated to fill the missing values; for outliers, statistical methods such as the Z-score method are used to identify and correct them. In order to eliminate the impact of differences between data features of different magnitudes, the data must also be normalized through methods such as minimum and maximum scaling.

[0094] After initializing the parameters of the linear regression model, the model is trained using the model training set. During training, the model parameters are iteratively adjusted using gradient descent to minimize the loss function (such as mean squared error (MSE)) so that the model's predicted values are as close to the actual values as possible. During the iterative process, the loss function value is calculated for the current parameters, and the model parameters are updated accordingly until the loss function converges or the preset number of iterations is reached, completing the linear regression model training.

[0095] After training, the linear regression model's predictive performance is evaluated using the model test set. Evaluation metrics are calculated and, based on the results, hyperparameters of the linear regression model, such as the learning rate and regularization parameter, are adjusted to optimize the model's generalization performance. Ultimately, a validated linear regression model is obtained, serving as an important tool for analyzing network performance metrics.

[0096] It should be noted that the validation set data, as a sample of historical performance data independent of the training set, can objectively test the performance of the trained linear regression model on unseen performance data. In the evaluation process, it is necessary to select appropriate evaluation indicators, such as mean square error (MSE), mean absolute error (MAE), coefficient of determination (R 2 ) etc. For example, the smaller the MSE value, the smaller the prediction error of the trained linear regression model; R 2 The closer the value is to 1, the better the trained linear regression model fits the performance data. The gradient descent algorithm is used to continuously adjust the linear regression model parameters so that the model minimizes the sum of squared errors between the predicted and actual values. To prevent overfitting, regularization terms (such as L1 or L2 regularization) are often introduced. Hyperparameter tuning can employ strategies such as grid search, random search, or Bayesian optimization. Cross-validation is used to find the optimal parameter combination, such as the optimal learning rate and regularization coefficient, to improve the expressiveness and generalization ability of the linear regression model. After repeated adjustments and evaluations, the hyperparameters are continuously optimized until the evaluation metrics of the trained linear regression model reach a satisfactory level, resulting in a linear regression model with improved performance. K-fold cross-validation is used to evaluate the performance of the linear regression model on different data subsets to test its generalization ability and verify the validity of the linear regression model assumptions, such as homoscedasticity and lack of autocorrelation of the residuals. Outliers that may have a significant impact on the linear regression model are identified and analyzed, and confidence intervals or prediction intervals are provided for the predicted values to reflect the uncertainty of the results. Combining these methods, the accuracy and reliability of the model can be comprehensively evaluated.

[0097] Leveraging a trained and validated linear regression model, this embodiment conducts in-depth analysis of real-time collected network operating status and user feedback data to obtain network performance indicators such as average latency, peak throughput, and packet loss rate trends. These indicators directly reflect the quality and efficiency of network services, providing key data support for subsequent dynamic adjustments to network deployment plans.

[0098] This embodiment constructs and trains a linear regression model by combining real-time monitoring with user feedback, thereby achieving accurate quantification and evaluation of 5G communication network performance. The collected user feedback and performance data cover a wealth of information related to network operation, including user experience, network connection stability, performance data transmission speed and other aspects, reflecting the various states and characteristics of the network in actual operation. The pre-processed user feedback and performance data are input into the linear regression model for in-depth analysis and processing based on the internal learned rules and algorithms. Taking into account the mutual relationship and influence between the various performance data characteristics, the corresponding network performance indicators are output. The network performance indicators include: Delay: The time interval from the start of data transmission at the sender to the receipt of data at the receiver, usually in milliseconds (ms). Throughput: The amount of data successfully transmitted in a given time, usually expressed in Mbps (megabits per second). Packet loss rate: The proportion of data packets lost during network transmission to the total number of data packets sent, usually expressed as a percentage (%). Network performance metrics are converted into numerical values. Latency is measured by calculating the time difference between sending and receiving data packets using the Ping command or the timestamp method. Throughput is calculated by dividing the total amount of data transmitted within a specific time period by the duration. Packet loss rate is calculated by recording the total number of packets sent by the sender and the number of packets successfully received by the receiver, providing a quantitative assessment of the specific network status. This process not only enhances the intelligence level of network deployment and optimization, but also ensures that network service quality is always optimized to meet growing user needs and business challenges.

[0099] Step S205: Based on the network performance indicators, dynamically adjust the base station layout optimization plan and edge computing node configuration strategy to obtain a 5G communication network deployment plan.

[0100] Optionally, step S205 includes: dividing the 5G network coverage area into different performance areas according to network performance indicators, analyzing various network performance indicators of different performance areas, and obtaining performance characteristics, user needs and user distribution changes of different performance areas; dynamically adjusting the base station layout optimization plan based on the performance characteristics and user needs of different performance areas; dynamically adjusting the edge computing node configuration plan based on the user distribution changes of different performance areas; integrating the dynamically adjusted base station layout optimization plan and edge computing node configuration plan to obtain a 5G communication network deployment plan.

[0101] The entire 5G network coverage area is subdivided into different performance zones based on network performance indicators such as signal strength, data transmission rate, latency, and packet loss rate. This division takes into account the actual network operating status and can accurately locate performance bottlenecks or areas of advantage. By analyzing the network performance indicators of each performance zone, we can gain a deeper understanding of the performance characteristics of different regions. For example, some areas exhibit high performance due to dense base station distribution and low user traffic; while other areas may exhibit lower network performance due to factors such as terrain and building obstruction. In addition, by collecting user feedback and business needs, we can further understand the changing trends of user distribution and the specific needs of users in different regions.

[0102] Based on the above analysis results, this embodiment can dynamically adjust the base station layout optimization plan according to the characteristics and needs of each performance area. For example, in low-performance areas with weak signals and high user demand, the number of base stations can be appropriately increased or their locations adjusted to enhance signal coverage. In areas with high user density, base station capacity expansion should be considered to cope with peak traffic periods. In high-performance areas where network performance has already reached high standards, minor base station adjustments may be reduced to avoid resource redundancy. This dynamic adjustment takes into account the actual operating status of the network and future demand forecasts, aiming to achieve efficient resource allocation and improve overall network performance.

[0103] For example, the network performance threshold range is set through network standards and specifications, where a is the threshold boundary that distinguishes the high-performance area from the medium-performance area, and b is the threshold boundary that distinguishes the medium-performance area from the low-performance area (a is generally -80dBm, and b is generally -90dBm). When the network performance threshold is greater than a, it is a high-performance area; when the network performance threshold is less than a and greater than b, it is a medium-performance area; when the network performance threshold is less than b, it is a low-performance area. In high-performance areas, the reasons for the excellent performance of various indicators are analyzed. This may be due to a good base station layout, fewer interference sources, and a favorable geographical environment. This will further clarify the needs of users in this area for high-speed and stable networks, such as high demand for high-traffic, low-latency services such as HD video and virtual reality. At the same time, attention will be paid to whether the user distribution in this area is relatively concentrated and whether there are dynamic network change trends. In medium-performance areas, the key factors affecting performance improvement are explored, which may be problems such as signal obstruction or uneven base station load in some areas. The characteristics of users in this area who expect certain improvements in network performance while meeting basic business needs are understood, and how the number and distribution of users change over time and in different scenarios. In low-performance areas, the focus is on analyzing the root causes of poor performance, such as insufficient base station coverage and severe signal interference. The basic needs of users in this area for improved network connectivity are clarified, as well as the characteristics such as whether users are relatively dispersed or concentrated in specific scenarios. Through such comprehensive and in-depth analysis, the performance characteristics, user needs, and user distribution changes of different regions can be accurately grasped.

[0104] In response to changes in user distribution and business needs, this embodiment can also intelligently adjust the edge computing node configuration strategy. In areas with high user density and diverse business needs, edge computing nodes need to be equipped with more powerful computing capabilities and sufficient storage resources to support real-time processing and local caching; in areas where users are more evenly distributed, node configuration can be moderately optimized to ensure service response speed while avoiding resource waste. The dynamic adjustment mechanism also takes time factors into account. For example, during peak hours in the morning and evening, the configuration resources of edge computing nodes should be tilted towards high-density user areas to cope with traffic surges.

[0105] This embodiment integrates the dynamically adjusted base station layout optimization scheme with the edge computing node configuration scheme to obtain a complete 5G communication network deployment scheme. This integration takes into account the global optimization of the network architecture, ensuring that the base station layout and edge computing node configuration complement each other and jointly serve the goal of improving network performance and user experience. For example, after optimizing the base station layout, this embodiment will re-evaluate the location and resource allocation of edge computing nodes to ensure that the node deployment can effectively share the base station load and quickly respond to user needs. It not only realizes dynamic deployment optimization based on network performance indicators, but also ensures the foresight and adaptability of the scheme, providing solid support for the efficient operation and high-quality services of the 5G communication network.

[0106] Through the above steps, the preprocessed multi-dimensional traffic demand data set can be input into the pre-trained multi-category network demand prediction model, and the network demand within a specified time period in the future can be evaluated using the network demand prediction model; the network demand prediction results output by various types of network demand prediction models are integrated to generate a network status report, and a base station layout is formulated based on the network status report and various network traffic influencing factors to obtain a base station layout optimization plan; based on the network status report and the base station layout optimization plan, an edge computing node configuration strategy is formulated, wherein the edge computing node configuration strategy is used to configure network node resources in each area; the network operation status after the regional network nodes are configured is monitored, and user feedback data is collected, and the network operation status and user feedback data are analyzed using a linear regression model to obtain network performance indicators; based on the network performance indicators, the base station layout optimization plan and the edge computing node configuration strategy are dynamically adjusted to obtain a 5G communication network deployment plan. In this embodiment, the network demand within a specified time period in the future can be evaluated through multiple types of network demand prediction models, thereby improving the accuracy of network demand prediction. Accurate network demand prediction is crucial for subsequent base station layout optimization and edge computing node configuration. It can avoid over-allocation or under-allocation of resources, significantly reduce network traffic fluctuations caused by emergencies, and improve the resource utilization efficiency and overall performance of the 5G communication network, thereby solving the technical problem in related technologies that when deploying 5G networks, network demand prediction is usually based on simple linear extrapolation of historical data, which makes it difficult to cope with sudden network traffic fluctuations.

[0107] The following describes it in detail with reference to another embodiment.

[0108] Example 2

[0109] A 5G-based communication network deployment device provided in this embodiment includes multiple implementation units, each implementation unit corresponds to each implementation step in the above-mentioned embodiment one. Its specific implementation method and beneficial effects can refer to the above-mentioned method embodiment and will not be repeated here.

[0110] Figure 3 is a schematic diagram of an optional 5G-based communication network deployment device according to an embodiment of the present invention, such as Figure 3 As shown, the 5G-based communication network deployment device may include: a multi-type network demand prediction unit 31, a demand fusion unit 32, a node configuration unit 33, a network operation monitoring unit 34, and a configuration adjustment unit 35.

[0111] The multi-class network demand prediction unit 31 is used to input the pre-processed multi-dimensional traffic demand data set into a pre-trained multi-class network demand prediction model, and use the network demand prediction model to evaluate the network demand within a specified time period in the future.

[0112] The demand fusion unit 32 is used to fuse the network demand prediction results output by various types of network demand prediction models, generate a network status report, and formulate a base station layout based on the network status report and various network traffic influencing factors to obtain a base station layout optimization plan.

[0113] The node configuration unit 33 is used to formulate an edge computing node configuration strategy based on the network status report and the base station layout optimization plan, wherein the edge computing node configuration strategy is used to configure network node resources in each area.

[0114] The network operation monitoring unit 34 is used to monitor the network operation status after the regional network nodes are configured, collect user feedback data, and analyze the network operation status and user feedback data using a linear regression model to obtain network performance indicators.

[0115] The configuration adjustment unit 35 is used to dynamically adjust the base station layout optimization plan and edge computing node configuration strategy based on network performance indicators to obtain a 5G communication network deployment plan.

[0116] The above-mentioned 5G-based communication network deployment device can input the pre-processed multi-dimensional traffic demand data set into a pre-trained multi-class network demand prediction model through a multi-class network demand prediction unit 31, and use the network demand prediction model to evaluate the network demand within a specified time period in the future. The network demand prediction results output by each type of network demand prediction model are integrated through the demand fusion unit 32 to generate a network status report, and a base station layout is formulated based on the network status report and various network traffic influencing factors to obtain a base station layout optimization plan. The node configuration unit 33 formulates an edge computing node configuration strategy based on the network status report and the base station layout optimization plan, wherein the edge computing node configuration strategy is used to configure the network node resources of each area, and the network operation status after the regional network nodes are configured is monitored through the network operation monitoring unit 34, and user feedback data is collected. The network operation status and user feedback data are analyzed using a linear regression model to obtain network performance indicators. The configuration adjustment unit 35 dynamically adjusts the base station layout optimization plan and the edge computing node configuration strategy based on the network performance indicators to obtain a 5G communication network deployment plan. In this embodiment, the network demand within a specified time period in the future can be evaluated through multiple types of network demand prediction models, thereby improving the accuracy of network demand prediction. Accurate network demand prediction is crucial for subsequent base station layout optimization and edge computing node configuration. It can avoid over-allocation or under-allocation of resources, significantly reduce network traffic fluctuations caused by emergencies, and improve the resource utilization efficiency and overall performance of the 5G communication network, thereby solving the technical problem in related technologies that when deploying 5G networks, network demand prediction is usually based on simple linear extrapolation of historical data, which makes it difficult to cope with sudden network traffic fluctuations.

[0117] Optionally, the 5G-based communication network deployment device also includes: an acquisition network construction unit, which is used to configure communication sensors and communication network protocols in each network coverage area to build a multi-source information acquisition network based on the geographical environment factors of different network coverage areas, the density of user distribution and the network service demand characteristics of each area before inputting the pre-processed multi-dimensional traffic demand data set into the pre-trained multi-category network demand prediction model; a multi-source acquisition unit, which is used to use the multi-source information acquisition network to collect user location, traffic demand data, network load, environmental data and equipment type data to obtain a historical multi-dimensional traffic demand data set; a missing filling unit, which is used to process the missing values in the historical multi-dimensional traffic demand data set using a group mean filling strategy, and normalize the processed historical multi-dimensional traffic demand data set through a minimum and maximum scaling strategy to obtain a pre-processed historical multi-dimensional traffic demand data set; a data sorting unit, which is used to sort the pre-processed historical multi-dimensional traffic demand data set into a time series data format and train the ARIMA model and the LSTM model; and a model determination unit, which is used to use the trained ARIMA model and LSTM model as network demand prediction models.

[0118] Optionally, the data sorting unit includes: a data arrangement module, which is used to arrange the preprocessed historical multi-dimensional traffic demand data set according to a predetermined time interval, establish an association relationship between each historical traffic demand data point and the corresponding timestamp, and obtain traffic demand time series data; a data partitioning module, which is used to divide the traffic demand time series data into a training set and a validation set according to a predetermined ratio, and use the training set to train the initial ARIMA model of the selected model parameters, and estimate the parameter values of the ARIMA model through the maximum likelihood estimation strategy; a model evaluation module, which is used to evaluate the trained ARIMA model using the validation set, adjust the parameter values of the ARIMA model according to the evaluation results, and obtain a trained ARIMA model; a data conversion module, which is used to convert the preprocessed historical multi-dimensional traffic demand data set into a three-dimensional tensor form, and train the LSTM model.

[0119] Optionally, the demand fusion unit includes: a weight ratio adjustment module, which is used to dynamically adjust the corresponding model weight ratio according to the prediction error of each type of network demand prediction model in different time periods; a demand fusion module, which is used to fuse the network demand prediction results output by each type of network demand prediction model based on the model weight ratio and adopt a weighted average strategy to obtain future network demand; a demand comparison module, which is used to compare future network demand with historical network demand data to generate a network status report.

[0120] Optionally, the demand fusion unit also includes: a network status report multi-dimensional analysis module, which is used to analyze the network status report in different dimensions to obtain high network demand areas, base station location areas to be optimized and key business focus areas, wherein the high network demand area refers to the area where the traffic demand assessment value in the network status report is greater than the preset traffic demand threshold, and the traffic demand assessment value is determined based on the user-side traffic demand parameters, population density and the current regional signal strength; the base station location area to be optimized refers to the area where the base station coverage assessment value in the network status report is lower than the preset base station coverage threshold, and the base station coverage assessment value is determined based on predetermined survey factors and network planning goals, and the predetermined survey factors are used to determine the area. The factors include the current geographical location of the base station, surrounding environmental factors and equipment operating status. The key business focus area refers to the area where the specific business evaluation value in the network status report is higher than the preset business threshold. The specific business evaluation value is determined based on the traffic proportion of the specific business, business type distribution and business growth trend parameters; the base station migration plan formulation module is used to determine the new base station plan based on the high network demand area, formulate the base station migration plan based on the base station location area to be optimized, and formulate the base station parameter adjustment plan based on the demand characteristics of the key business focus area; the base station layout plan formulation module is used to integrate the new base station plan, base station migration plan and base station parameter adjustment plan to formulate the base station layout optimization plan.

[0121] Optionally, the node configuration unit includes: a solution analysis module, which is used to analyze the network status report and the base station layout optimization plan, obtain the network demand characteristics of different regions, the network change trend and the influence of the base station layout on the edge computing node configuration of each region, and obtain the first type of node influencing factors; a node impact acquisition module, which is used to obtain the network demand forecast results, the current user terminal distribution status and the influence of each business characteristic on the edge computing node configuration of each region, and obtain the second type of node influencing factors; a regional division module, which is used to divide the overall region into different types of sub-regions based on the first type of node influencing factors and the second type of node influencing factors, and formulate corresponding node deployment strategies for the regional characteristics of different types of sub-regions; a computing resource allocation module, which is used to allocate computing resources according to regional business demand and the number of users, obtain the node resource allocation plan for each region, combine the node deployment strategies of different regions with the node resource allocation plans of each region, and obtain the edge computing node configuration strategy.

[0122] Optionally, when constructing a linear regression model, the 5G-based communication network deployment device includes: a data extraction unit for extracting historical user feedback data and historical network performance data, and dividing the historical user feedback data and historical network performance data into a model training set and a model test set according to a time sequence or random extraction principle; a linear regression model parameter initialization unit for initializing the parameters of the linear regression model based on predefined independent variables and dependent variables, wherein the independent variables include: network delay, throughput, packet loss rate, and the dependent variables include target network performance indicators; a model training unit for training the initialized linear regression model using the model training set, and iteratively adjusting the parameters of the linear regression model through the gradient descent method until the loss function converges or reaches a preset number of iterations, completing the training of the linear regression model, and obtaining a trained linear regression model, wherein, in each iteration process, the loss function value under the current parameters is calculated, and the parameters of the linear regression model are updated according to the loss function value; a linear regression model evaluation unit for evaluating the trained linear regression model through the model test set, calculating the evaluation index value, adjusting the hyperparameters of the linear regression model shown, and obtaining a verified linear regression model.

[0123] Optionally, the configuration adjustment unit includes: a performance indicator analysis module, which is used to divide the 5G network coverage area into different performance areas according to network performance indicators, analyze various network performance indicators of different performance areas, and obtain performance characteristics, user needs and user distribution changes of different performance areas; a base station layout optimization adjustment module, which is used to dynamically adjust the base station layout optimization plan based on the performance characteristics and user needs of different performance areas; a node configuration adjustment module, which is used to dynamically adjust the edge computing node configuration plan based on the user distribution changes in different performance areas; a plan integration module, which is used to integrate the dynamically adjusted base station layout optimization plan and edge computing node configuration plan to obtain a 5G communication network deployment plan.

[0124] The above-mentioned 5G-based communication network deployment device may also include a processor and a memory. The above-mentioned multi-type network demand prediction unit 31, demand fusion unit 32, node configuration unit 33, network operation monitoring unit 34, configuration adjustment unit 35, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize the corresponding functions.

[0125] The processors mentioned above include a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured, and kernel parameters can be adjusted to enable 5G-based communication network deployment.

[0126] The above-mentioned memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0127] Example 3

[0128] An embodiment of the present application may provide an electronic device, Figure 4 This is a structural block diagram of an electronic device according to an embodiment of the present application. Figure 4 As shown, the electronic device may include: one or more ( Figure 4 Only one is shown) processor 402, memory 404, storage controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0129] Among them, the memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the 5G-based communication network deployment method and device in the embodiments of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned 5G-based communication network deployment method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely located relative to the processor, and these remote memories can be connected to the terminal via a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0130] The processor can call the information and application programs stored in the memory through the transmission device to perform the following steps: input the preprocessed multi-dimensional traffic demand data set into a pre-trained multi-category network demand prediction model, and use the network demand prediction model to evaluate the network demand within a specified time period in the future; integrate the network demand prediction results output by various types of network demand prediction models to generate a network status report, and formulate a base station layout based on the network status report and various network traffic influencing factors to obtain a base station layout optimization plan; formulate an edge computing node configuration strategy based on the network status report and the base station layout optimization plan, wherein the edge computing node configuration strategy is used to configure network node resources in each area; monitor the network operation status after configuring regional network nodes, and collect user feedback data, use a linear regression model to analyze the network operation status and user feedback data to obtain network performance indicators; based on the network performance indicators, dynamically adjust the base station layout optimization plan and the edge computing node configuration strategy to obtain a 5G communication network deployment plan.

[0131] It can be understood by those skilled in the art that Figure 4 The structure shown is for illustration only, and the electronic device may also be a smart phone, a tablet computer, a PDA, a mobile Internet device (MID), a PAD or other terminal device. Figure 4 It does not limit the structure of the above electronic device. For example, the electronic device may also include Figure 4 More or fewer components (such as network interfaces, display devices, etc.) shown in, or with Figure 4 Different configurations shown.

[0132] A person of ordinary skill in the art can understand that all or part of the steps in the various 5G-based communication network deployment methods in the above-mentioned embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0133] Example 4

[0134] The embodiment of the present application further provides a storage medium. Optionally, in this embodiment, the storage medium can be used to store the program code executed by the 5G-based communication network deployment method provided in the first embodiment.

[0135] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the 5G-based communication network deployment methods in the above-mentioned embodiment 1.

[0136] Optionally, in this embodiment, the storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0137] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the 5G-based communication network deployment method described in each embodiment of the present application.

[0138] The present application also provides a computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the 5G-based communication network deployment method described in each embodiment of the present application are implemented.

[0139] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0140] In the above embodiments of the present invention, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0141] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only exemplary. For example, the division of the units can be a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0142] 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 units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0143] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0144] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk, etc. Various media that can store program codes.

[0145] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A 5G-based communication network deployment method, characterized in that: include: Input the pre-processed multi-dimensional traffic demand dataset into a pre-trained multi-class network demand prediction model, and use the network demand prediction model to estimate the network demand within a specified time period in the future; The network demand forecast results output by each type of network demand forecast model are integrated to generate a network status report, and a base station layout is formulated based on the network status report and multiple network traffic influencing factors to obtain a base station layout optimization plan; Formulate an edge computing node configuration strategy based on the network status report and the base station layout optimization plan, wherein the edge computing node configuration strategy is used to configure network node resources in each area; Monitoring the network operation status after configuring regional network nodes, collecting user feedback data, and analyzing the network operation status and the user feedback data using a linear regression model to obtain network performance indicators; Based on the network performance indicators, the base station layout optimization plan and edge computing node configuration strategy are dynamically adjusted to obtain a 5G communication network deployment plan.

2. The communication network deployment method according to claim 1, characterized in that: Before the pre-processed multi-dimensional traffic demand dataset is fed into the pre-trained multi-class network demand prediction model, the following steps are also included: Based on the geographical environment factors of different network coverage areas, the density of user distribution, and the network service demand characteristics of each area, communication sensors and communication network protocols are configured in each network coverage area to build a multi-source information collection network; Utilizing the multi-source information collection network to collect user location, traffic demand data, network load, environmental data, and device type data, to obtain a historical multi-dimensional traffic demand dataset; The missing values in the historical multi-dimensional traffic demand dataset are processed by adopting a group mean filling strategy, and the processed historical multi-dimensional traffic demand dataset is normalized by a minimum and maximum scaling strategy to obtain the pre-processed historical multi-dimensional traffic demand dataset; The pre-processed historical multi-dimensional traffic demand dataset is organized into a time series data format, and the ARIMA model and the LSTM model are trained; The trained ARIMA model and LSTM model are used as the network demand prediction model.

3. The communication network deployment method according to claim 2, characterized in that: The steps of arranging the preprocessed historical multi-dimensional traffic demand dataset into a time series data format and training the ARIMA model and the LSTM model include: Arranging the pre-processed historical multi-dimensional traffic demand data set according to predetermined time intervals, establishing an association relationship between each historical traffic demand data point and a corresponding timestamp, and obtaining traffic demand time series data; Dividing the traffic demand time series data into a training set and a validation set according to a predetermined ratio, and using the training set to train an initial ARIMA model of selected model parameters, and estimating parameter values of the ARIMA model by a maximum likelihood estimation strategy; The trained ARIMA model is evaluated using the validation set, and the parameter values of the ARIMA model are adjusted according to the evaluation results to obtain the trained ARIMA model; The preprocessed historical multi-dimensional traffic demand dataset is converted into a three-dimensional tensor form and the LSTM model is trained.

4. The communication network deployment method according to claim 1, characterized in that: The steps of fusing the network demand forecast results output by the various types of network demand forecast models to generate a network status report include: Dynamically adjust the corresponding model weight ratio according to the prediction error of each type of network demand prediction model in different time periods; Based on the weight ratio of the model, a weighted average strategy is adopted to fuse the network demand forecast results output by each type of network demand forecast model to obtain the future network demand; The future network demand is compared with historical network demand data to generate the network status report.

5. The communication network deployment method according to claim 1, characterized in that: The steps of formulating a base station layout based on the network status report and various factors affecting network traffic to obtain a base station layout optimization solution include: The network status report is analyzed in different dimensions to obtain high network demand areas, base station location areas to be optimized, and key business focus areas, wherein the high network demand area refers to an area corresponding to a preset traffic demand threshold value in the network status report where the traffic demand assessment value is greater than the traffic demand threshold value, and the traffic demand assessment value is determined based on the user-side traffic demand parameter, population density, and the current regional signal strength; the base station location area to be optimized refers to an area corresponding to a preset base station coverage threshold value in the network status report where the base station coverage assessment value is lower than the preset base station coverage threshold value, and the base station coverage assessment value is determined based on predetermined survey factors and network planning objectives, and the predetermined survey factors include the current geographical location of the base station, surrounding environmental factors, and equipment operating status; the key business focus area refers to an area corresponding to a specific business assessment value in the network status report that is higher than the preset business threshold value, and the specific business assessment value is determined based on the traffic proportion of the specific business, business type distribution, and business growth trend parameters; Determine new base station plans based on areas with high network demand, develop base station migration plans based on areas where base station locations are to be optimized, and develop base station parameter adjustment plans based on the demand characteristics of key business focus areas; The base station layout optimization plan is formulated by integrating the new base station plan, the base station migration plan and the base station parameter adjustment plan.

6. The communication network deployment method according to claim 1, characterized in that: The step of formulating an edge computing node configuration strategy based on the network status report and the base station layout optimization plan includes: Analyze the network status report and the base station layout optimization plan to obtain the network demand characteristics of different regions, network change trends, and the impact of base station layout on the configuration of edge computing nodes in each region, and obtain the first type of node influencing factors; Obtain the network demand forecast results, the current user terminal distribution status, and the impact of each service characteristic on the edge computing node configuration in each area to obtain the second type of node influencing factors; Based on the first type of node influencing factors and the second type of node influencing factors, the overall area is divided into different types of sub-areas, and corresponding node deployment strategies are formulated according to the regional characteristics of the different types of sub-areas; Computing resources are allocated according to regional business needs and the number of users to obtain a node resource allocation plan for each region. The node deployment strategies of different regions and the node resource allocation plans for each region are combined to obtain the edge computing node configuration strategy.

7. The communication network deployment method according to claim 1, characterized in that: The linear regression model is constructed in the following way: Extracting historical user feedback data and historical network performance data, and dividing the historical user feedback data and historical network performance data into a model training set and a model test set according to a chronological order or a random sampling principle; Initializing the parameters of a linear regression model based on predefined independent variables and dependent variables, wherein the independent variables include: network delay, throughput, and packet loss rate, and the dependent variable includes a target network performance indicator; The initialized linear regression model is trained using the model training set, and the parameters of the linear regression model are iteratively adjusted by a gradient descent method until the loss function converges or a preset number of iterations is reached, thereby completing the training of the linear regression model and obtaining a trained linear regression model, wherein, in each iteration, the loss function value under the current parameters is calculated, and the parameters of the linear regression model are updated according to the loss function value; The trained linear regression model is evaluated using the model test set, the evaluation index value is calculated, and the hyperparameters of the linear regression model are adjusted to obtain the verified linear regression model.

8. The communication network deployment method according to claim 1, characterized in that: The steps of dynamically adjusting the base station layout optimization plan and the edge computing node configuration strategy based on the network performance indicators to obtain a 5G communication network deployment plan include: Divide the 5G network coverage area into different performance areas according to the network performance indicators, analyze the network performance indicators of different performance areas, and obtain performance characteristics, user needs, and user distribution changes of different performance areas; Dynamically adjusting the base station layout optimization solution based on the performance characteristics of different performance areas and the user needs; Dynamically adjust the edge computing node configuration scheme based on changes in user distribution in different performance areas; The dynamically adjusted base station layout optimization plan and the edge computing node configuration plan are integrated to obtain the 5G communication network deployment plan.

9. A 5G-based communication network deployment device, characterized in that: include: A multi-class network demand prediction unit is used to input the pre-processed multi-dimensional traffic demand data set into a pre-trained multi-class network demand prediction model, and use the network demand prediction model to evaluate the network demand within a specified time period in the future; a demand fusion unit, configured to fuse the network demand forecast results output by each type of network demand forecast model, generate a network status report, and formulate a base station layout based on the network status report and various network traffic influencing factors to obtain a base station layout optimization plan; a node configuration unit, configured to formulate an edge computing node configuration strategy based on the network status report and the base station layout optimization plan, wherein the edge computing node configuration strategy is used to configure network node resources in each area; A network operation monitoring unit is used to monitor the network operation status after configuring regional network nodes, collect user feedback data, and analyze the network operation status and the user feedback data using a linear regression model to obtain network performance indicators; A configuration adjustment unit is used to dynamically adjust the base station layout optimization plan and edge computing node configuration strategy based on the network performance indicators to obtain a 5G communication network deployment plan.

10. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the 5G-based communication network deployment method described in any one of claims 1 to 8.

Images