Communication optimization method, system and device of power distribution network terminal

Abstract

The invention relates to the technical field of network communication. The invention discloses a communication optimization method, system and device for a power distribution network terminal. The method comprises the following steps: obtaining to-be-sent data and first verification information of the to-be-sent data; predicting predicted network state data corresponding to the power distribution network terminal through the target real-time network state prediction model; generating a data transmission target strategy corresponding to the power distribution network terminal according to the predicted network state data; sending the to-be-sent data and the first verification information to the power distribution network terminal based on the data transmission target strategy, so that the power distribution network terminal processes the to-be-sent data according to the first verification information; feedback information sent by the power distribution network terminal is received; and adjusting the data transmission target strategy according to the feedback information. The data transmission efficiency can be optimized, the packet loss rate can be reduced, and the resource consumption can be reduced.

Description

Technical Field

[0001] The present application relates to the field of network communication technologies, and in particular, to a communication optimization method, system and device for a distribution network terminal. Background Art

[0002] In modern communication systems, network fluctuations are a common phenomenon. Such fluctuations can cause delays, out-of-order or loss of data packets during transmission, thereby having an adverse impact on the reliability and consistency of communication. In related technologies, these problems are usually alleviated by retransmission mechanisms, data verification or redundant designs. However, directly using the retransmission mechanism for data retransmission will increase the communication delay and cause a decrease in transmission efficiency due to additional occupation of bandwidth resources; redundant designs will significantly increase unnecessary stored information, and the increase in the overhead of computing resources will also lead to large resource consumption; simply using the data verification method cannot maintain data consistency in multiple application scenarios.

[0003] In summary, the technical problems existing in related technologies need to be improved. Summary of the Invention

[0004] The main purpose of the embodiments of the present application is to propose a communication optimization method, system and device for a distribution network terminal, which can optimize data transmission efficiency, reduce the packet loss rate and reduce resource consumption.

[0005] To achieve the above object, on the one hand, an embodiment of the present application proposes a communication optimization method for a distribution network terminal, and the method includes the following steps:

[0006] Obtain the data to be sent and the first verification information of the data to be sent;

[0007] Predict the predicted network state data corresponding to the distribution network terminal through a target real-time network state prediction model;

[0008] Generate a data transmission target strategy corresponding to the distribution network terminal according to the predicted network state data;

[0009] Send the data to be sent and the first verification information to the distribution network terminal based on the data transmission target strategy, so that the distribution network terminal processes the data to be sent according to the first verification information;

[0010] Receive the feedback information sent by the distribution network terminal;

[0011] Adjust the data transmission target strategy according to the feedback information.

[0012] In some embodiments, the obtaining the data to be sent and the first verification information of the data to be sent includes:

[0013] The original data is segmented to obtain segmented data;

[0014] Generate first check information for the segmented data.

[0015] In some embodiments, predicting the predicted network state data corresponding to the distribution network terminal by the target real-time network state prediction model includes:

[0016] Obtain the real-time network state data corresponding to the distribution network terminal at the current moment;

[0017] Update the first historical network state data corresponding to the distribution network terminal according to the real-time network state data to obtain second historical network state data;

[0018] Train the real-time network state prediction model at the current moment through the second historical network state data to obtain a target real-time network state prediction model;

[0019] Predict the predicted network state data corresponding to the distribution network terminal through the target real-time network state prediction model.

[0020] In some embodiments, the target real-time network state prediction model includes a time series analysis model and a neural network model, and the data output of the time series analysis model is the data input of the neural network model.

[0021] In some embodiments, generating the data transmission target strategy corresponding to the distribution network terminal according to the predicted network state data includes:

[0022] Determine a target transmission path and a target time window according to the predicted network state data;

[0023] Calculate the estimated packet loss rate of the target transmission path within the target time window;

[0024] Determine the working state of the dynamic retransmission strategy according to the estimated packet loss rate;

[0025] Wherein, the data transmission target strategy includes the target transmission path, the target time window, and the working state of the dynamic retransmission strategy.

[0026] In some embodiments, determining the working state of the dynamic retransmission strategy according to the estimated packet loss rate includes:

[0027] Determine the dynamic retransmission frequency of the dynamic retransmission strategy according to the estimated packet loss rate;

[0028] Control the working state of the dynamic retransmission strategy based on the dynamic retransmission frequency.

[0029] In some embodiments, the calculation formula for the dynamic retransmission frequency is as follows:

[0030]

[0031] Wherein, R t represents the dynamic retransmission frequency, P loss represents the estimated packet loss rate, R max represents the maximum frequency of retransmission, α represents the weight of the fixed transmission cost, β represents the sensitivity of the packet loss rate to the retransmission frequency, and P th represents the packet loss rate threshold.

[0032] In some embodiments, the calculation formula for the estimated packet loss rate is as follows:

[0033]

[0034] Wherein, N represents the total number of network state samples, F(x i ) represents the transmission result function of each data packet, and x i represents the network state sample.

[0035] On the other hand, an embodiment of the present invention provides a communication optimization system for a distribution network terminal, including:

[0036] A data processing module, configured to obtain the data to be sent and the first check information of the data to be sent;

[0037] A network state prediction module, configured to predict the predicted network state data corresponding to the distribution network terminal through a target real-time network state prediction model;

[0038] A policy generation module, configured to generate a data transmission target policy corresponding to the distribution network terminal according to the predicted network state data;

[0039] A policy sending module, configured to send the data to be sent and the first check information to the distribution network terminal based on the data transmission target policy, so that the distribution network terminal processes the data to be sent according to the first check information;

[0040] A data receiving module, configured to receive the feedback information sent by the distribution network terminal;

[0041] A policy optimization module, configured to adjust the data transmission target policy according to the feedback information.

[0042] On the other hand, an embodiment of the present invention provides a computer device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it is used to implement the communication optimization method of the distribution network terminal.

[0043] The embodiments of the present application at least include the following beneficial effects:

[0044] In this embodiment, the target real-time network state prediction model is used to predict the predicted network state data corresponding to the distribution network terminal, and the data transmission target policy corresponding to the distribution network terminal is generated according to the predicted network state data. At the same time, the data transmission target policy can be adjusted according to the feedback information sent by the distribution network terminal, so as to adapt to the data transmission process in different network states, improve the data transmission efficiency, and reduce the resource consumption occupied by redundant design. Then, based on the data transmission target policy, the data to be sent and the first verification information are sent to the distribution network terminal, and the distribution network terminal processes the data to be sent according to the first verification information, thereby improving the stability and reliability of data transmission. Description of the Drawings

[0045] Figure 1 is a flowchart of a communication optimization method for a distribution network terminal provided by an embodiment of the present application;

[0046] Figure 2 is a schematic diagram of a communication optimization system for a distribution network terminal provided by an embodiment of the present application;

[0047] Figure 3 is a schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present application. Detailed Embodiments

[0048] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application. They are only examples of devices and methods that are consistent with some aspects of the embodiments of the present application.

[0049] The terms "at least one", "multiple", "each", "any one", etc. used in the present application, at least one includes one, two or more, multiple includes two or more, each refers to each of the corresponding multiple, and any one refers to any one of the multiple.

[0050] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0051] Before elaborating on the embodiments of the present application in detail, some nouns and terms involved in the embodiments of the present application are first explained. The nouns and terms involved in the embodiments of the present application are applicable to the following explanations:

[0052] Retransmission mechanism: In the process of data communication, when data packets are lost or in error due to various reasons (such as network congestion, signal interference, physical link failure, etc.), the strategy and process by which the sender retransmits this data. It is an important means to ensure the reliability of data transmission.

[0053] ARIMA (Autoregressive Integrated Moving Average Model): A commonly used time series prediction model for analyzing and predicting data sequences with a time sequence. It combines three components: autoregression (AR), differencing (I), and moving average (MA), and can effectively handle non-stationary time series data.

[0054] MLP (Multi-Layer Perceptron): An artificial neural network structure composed of an input layer, one or more hidden layers, and an output layer, used for performing complex non-linear mappings on input data to achieve tasks such as classification and regression.

[0055] LSTM (Long Short-Term Memory): A special type of Recurrent Neural Network (RNN). Recurrent Neural Networks are mainly used to process sequential data, and their characteristic is that the output of the network depends not only on the current input but also on the previous input states, enabling better capture of long-distance dependency information in sequential data.

[0056] ADF (Augmented Dickey-Fuller Test): A statistical test method used to test whether time series data is stationary. The statistical characteristics (such as mean, variance, etc.) of stationary time series data do not change over time, while non-stationary time series may have trends, seasonality, etc. that cause changes in their statistical characteristics.

[0057] ReLU (Rectified Linear Unit): An activation function widely used in artificial neural networks (especially deep neural networks). The main role of the activation function is to introduce non-linearity into the neural network. Because without an activation function, the output of each layer of a multi-layer neural network is only a linear combination of the input of the previous layer, which can be simplified to a linear model and cannot effectively handle complex non-linear problems.

[0058] A communication optimization method for a distribution network terminal provided by an embodiment of the present application relates to the field of network communication technologies. The communication optimization method for a distribution network terminal provided by an embodiment of the present application can be applied to a terminal, a server, or software running on a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, a vehicle-mounted terminal, etc., but is not limited thereto; the server side can be configured as an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or can also be configured as a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network; the software can be an application for implementing the communication optimization method of the distribution network terminal, etc., but is not limited to the above forms.

[0059] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0060] Figure 1 is an optional flowchart of a communication optimization method for a distribution network terminal provided by an embodiment of the present application, Figure 1 The method in may include but is not limited to steps S100 to S600:

[0061] Step S100, obtain the data to be sent and the first verification information of the data to be sent;

[0062] Step S200, predict the predicted network state data corresponding to the distribution network terminal through a target real-time network state prediction model;

[0063] Step S300, generate a data transmission target strategy corresponding to the distribution network terminal according to the predicted network state data;

[0064] Step S400: Send the data to be sent and the first verification information to the distribution network terminal based on the data transmission target policy, so that the distribution network terminal processes the data to be sent according to the first verification information;

[0065] Step S500: Receive the feedback information sent by the distribution network terminal;

[0066] Step S600: Adjust the data transmission target policy according to the feedback information.

[0067] In some embodiments, in step S100, the data to be sent may include, but is not limited to, file data, video stream data, or audio stream data. Since the sizes of these data to be sent are different and the transmission capacity of the target network is limited, therefore, in this embodiment, the data to be sent can be segmented before data transmission. For example, when the data to be sent is large model file data, the appropriate segment size can be determined according to the maximum transmission unit (MTU) of the network, and then the large model file data can be segmented and then transmitted, so as to avoid ineffective transmission due to too large data segments. Another example is in an Ethernet environment. Since the MTU is usually about 1500 bytes, in order to avoid fragmentation, after setting the segment size to be less than 1500 bytes, the data to be sent can be segmented and then transmitted. It can be understood that when segmenting the data to be sent, it can start from the beginning of the data and divide the data in turn according to the determined segment size. For example, for a data to be sent with a length of 5000 bytes, if the segment size is 1000 bytes, it can be divided into 5 data segments. Specifically, during the division process, it is necessary to ensure that the order information of the data can be recorded so that the distribution network terminal can correctly reconstruct the content of the data to be sent according to the received data.

[0068] In some embodiments, this embodiment can also select an appropriate verification algorithm based on factors such as the importance of the data to be sent, the sensitivity to errors, and computing resources to generate the first verification information. Exemplarily, if extremely high requirements are placed on data accuracy and computing resources permit, the cyclic redundancy check (CRC) algorithm can be selected to generate the first verification information for the data to be sent. More specifically, for example, the CRC-32 algorithm in the cyclic redundancy check algorithm can be selected to generate the first verification information for the data to be sent. If more savings are required for resources, the checksum algorithm can be selected to generate the first verification information for the data to be sent. It can be understood that, taking the generation of the first verification information for the data to be sent by CRC-32 as an example, this embodiment can generate a 32-bit verification code as the first verification information by performing complex polynomial operations on the data segments corresponding to the data to be sent. Then, after sending each data segment and the first verification information corresponding to the data segment to the distribution network terminal, the distribution network terminal can verify the data based on the first verification information. Exemplarily, taking the generation of the first verification information for the data to be sent by the checksum algorithm as an example, this embodiment can obtain a fixed-length checksum as the first verification information by performing simple summation operations or carry processing on all bytes of all data segments in the data to be sent. After associating and recording the generated first verification information with the corresponding data segments, this association information can be stored by creating a table or adding fields in the data packet header, so that the distribution network terminal can find the corresponding first verification information according to the association information for verification after receiving the data segments, thereby improving the reliability of data transmission.

[0069] In some embodiments, in step S200, during the data transmission process, the network status is monitored in real time, and at the same time, the real-time network status data corresponding to the distribution network terminal at the current moment is obtained. Specifically, the real-time network status data can include the packet loss rate of data transmission in the network. The packet loss rate can be collected from routers, switches, server network cards, and network monitoring tools in the network. Among them, in order to improve the fluctuation details during the data transmission process in a short period of time, this embodiment can set the recording frequency during the collection process to once every few seconds. It can be understood that when collecting the real-time network status data, this embodiment can also synchronously collect auxiliary data including but not limited to the real-time bandwidth of the network interface, the current number of connections, the load conditions of each link, and the traffic type proportion in different time periods, so as to analyze the reasons for packet loss through these auxiliary data, so that the target real-time network status prediction model can predict more accurately in the later training. Among them, the traffic type proportion includes information such as the proportion of video streams, file downloads, and instant messaging traffic. It can be understood that the acquisition period of the real-time network status data should be long enough, and the acquisition period should cover the peak and trough periods during the network usage process, such as continuously collecting data for a week or even a month, so as to accumulate enough samples to reflect the network fluctuations of various networks.

[0070] In some embodiments, the process of predicting the predicted network state data corresponding to the distribution network terminal through the target real-time network state prediction model may be as follows: after obtaining the real-time network state data corresponding to the distribution network terminal at the current moment, update the first historical network state data corresponding to the distribution network terminal according to the real-time network state data to obtain the second historical network state data; then, after training the real-time network state prediction model at the current moment through the second historical network state data to obtain the target real-time network state prediction model, predict the predicted network state data corresponding to the distribution network terminal through the target real-time network state prediction model.

[0071] In some embodiments, the target real-time network state prediction model may include, but is not limited to, a time series analysis model and a neural network model. The data output of the time series analysis model is the data input of the neural network model. It can be understood that during the training process of the real-time network state prediction model, the second historical network state data is input into the time series analysis model to obtain the first model data, and then the first model data and the preprocessed auxiliary data are input into the neural network model to obtain the second model data, and the transmission path and time window are determined according to the second model data.

[0072] Exemplarily, taking the ARIMA model as the time series analysis model and the MLP model as the neural network model as an example, the process of training the real-time network state prediction model through the second historical network state data includes, but is not limited to, the following steps:

[0073] Perform data cleaning and data standardization on the second historical network state data. Specifically, during the data cleaning process, data with obvious errors or outliers are removed by setting a reasonable threshold. Such abnormal data may be caused by reasons such as temporary device failures and instant network attacks. Exemplarily, when the reasonable threshold is set to 90%, if the packet loss rate of the second historical network state data is higher than 90% and the duration is extremely short and is abnormal, then the second historical network state data is removed; if the packet loss rate data of a certain period in the second historical network state data is missing, it is filled by the linear trend or mean of the data in the previous and subsequent periods to maintain the continuity of the second historical network state data. For data unified standardization, all the second historical network state data is converted into a standard form with a mean of 0 and a standard deviation of 1 through the formula where: x is the second historical network state data, μ is the mean, and σ is the standard deviation. Through data unified standardization, it is ensured that the second historical network state data of different magnitudes has reasonable weights in model training, and the model learning process is prevented from being misled due to data scale differences.

[0074] Input the processed second historical network status data into the ARIMA model to obtain the original packet loss rate time series graph through the ARIMA model, so as to observe whether there are obvious trends or seasonal fluctuations through the time series graph, and then adjust the model parameters of ARIMA.

[0075] Specifically, in the process of adjusting the model parameters of ARIMA, if there is a linear growth or decline trend in the original packet loss rate time series graph, perform a first-order difference operation, that is, y′ t = y t - y t-1 to generate a new sequence y′ t ; then re-detect the stationarity through the unit root test. Among them, the unit root test can specifically be the ADF test algorithm; if it is still not stationary, increase the difference order according to the amplitude of the linear growth or decline trend until the sequence is stationary.

[0076] The first model data obtained by processing the packet loss rate in the future for a period of time by the adjusted ARIMA model and the preprocessed auxiliary data together constitute the input vector of the MLP to adjust the model parameters of the MLP through the input vector. It can be understood that for the network structure of the MLP model, the number of input layer nodes matches the dimension of the input vector, the hidden layer is 2 - 3 layers, and the ReLU activation function is used to enhance the non-linear fitting ability of the MLP model. The output layer is set to a single node. The target value is the comprehensive score of different target transmission paths as the second model data. Calculate the comprehensive score of each network interface according to past experience and experimental setting rules. The higher the comprehensive score value, the more suitable the transmission path and time window corresponding to the network interface.

[0077] After completing the training of the ARIMA model and the MLP model, this embodiment can also comprehensively train multiple ARIMA-MLP combined models with different parameter settings by changing the ARIMA order or the number of hidden layers and nodes of the MLP. In the comprehensive training process, calculate the mean value or weighted mean value of the outputs of multiple combined models with different parameter settings, so as to enhance the robustness of the prediction and reduce the influence of the error of a single model. In addition, this embodiment can also finely adjust the hyperparameters through cross-validation to determine the optimal model configuration, and at the same time increase or decrease the features input to the MLP model according to the actual network operation and maintenance feedback according to the requirements, so as to achieve the effect of continuously optimizing the model performance.

[0078] In the process of training the data in the second historical network state data by the above real-time network state prediction model, the ARIMA model has a more intuitive explanation for the linear time series characteristics, making it easier to understand the basic trend of the change in the state of the second historical network state data; the MLP model supplements the explanation of the non-linear part on the basis of the ARIMA model; and the interpretability and prediction ability of the target real-time network state prediction model are balanced through the combination of the two models. Then, the parameters of the ARIMA model and the network structure of the MLP model are adjusted according to the characteristics of the second historical network state data, which can adapt to different network environments and the prediction tasks of the target transmission path and the target time window.

[0079] In some embodiments, the target real-time network state prediction model can also be constructed only by using a time series analysis model. Specifically, a time series analysis model including but not limited to the ARIMA or LSTM algorithm model can be used to effectively predict the network change trend in the future time period. According to the prediction result, the target transmission path and the target time window with the optimal network conditions are dynamically selected, so as to optimize the data transmission efficiency and reduce the packet loss rate.

[0080] Specifically, the packet loss rate in the network state data is monitored in real time through a network detection tool, the monitoring function of the network device itself, or a proxy-based monitoring system, and the packet loss rate threshold is set according to past experience and experiments. When it is detected that the packet loss rate L in the real-time network state data exceeds the preset packet loss rate threshold T, the retraining of the target real-time network state prediction model is triggered to improve the accuracy of the model's prediction. After updating the prediction model, the sender is notified to use the new model for subsequent predictions and apply the new model to the network state prediction of the next target transmission path and target time window; at the same time, the sender records the time and specific parameters of the model update to ensure the traceability of subsequent verification and debugging. In addition, the results of each prediction and the actual network state data will be stored for long-term monitoring and optimization of the prediction accuracy of the model.

[0081] In some embodiments, by recording the update time of the target real-time network state prediction model, the parameters of the updated model, and the prediction results, reference data is provided for evaluating the effectiveness of the target real-time network state prediction model. In addition, in this embodiment, the prediction error E of the model can be obtained by comparing the actual network state with the prediction results of the target real-time network state prediction model; the prediction accuracy can be calculated based on the prediction error; then, the error threshold F can be set according to the prediction accuracy and industry experience, and it can be determined whether it is necessary to re-evaluate the effectiveness of the prediction model based on the magnitude relationship between the prediction error E and the error threshold F. Specifically, when the specific condition E > F is met, the sender automatically triggers the re-evaluation process of the model, analyzes the effectiveness of the current model, and determines whether the model can be optimized or replaced to further improve the prediction accuracy and reliability. It can be understood that during the evaluation process, the sender will collect and organize recent prediction error data to analyze the performance of the model. Among them, the performance of the model includes the prediction accuracy, stability, and adaptability to the latest network state of the model. If the evaluation result shows that the current model cannot meet the actual requirements, then the option can be to continue to optimize the current model or develop a new model.

[0082] In some embodiments, in step S300, after determining the target transmission path and the target time window according to the predicted network state data, the estimated packet loss rate can be obtained by obtaining the packet loss situation in the target transmission path and the target time window predicted by the target real-time network state prediction model, and then the working state of the dynamic retransmission policy can be determined according to the estimated packet loss rate. It can be understood that the core goal of the dynamic retransmission policy is to improve the reliability and efficiency of data transmission in a complex network environment. The dynamic retransmission policy dynamically adjusts the retransmission frequency and method by monitoring the network state in real time, according to the estimated packet loss probability and transmission delay information, avoiding the bandwidth waste caused by frequent retransmissions, and enhancing the data recovery ability under high packet loss rates. Among them, the data transmission target policy includes the target transmission path, the target time window, and the working state of the dynamic retransmission policy.

[0083] In some embodiments, in determining the working state of the dynamic retransmission policy according to the estimated packet loss rate, the dynamic retransmission frequency of the dynamic retransmission policy can be determined according to the estimated packet loss rate, and the working state of the dynamic retransmission policy can be controlled based on the dynamic retransmission frequency. Specifically, when the target transmission path and the target time window are determined, the dynamic retransmission frequency is determined by the dynamic retransmission policy, and then the data to be sent and the first check information are sent to the distribution network terminal according to the dynamic retransmission frequency in combination with the target transmission path and the target time window. It can be understood that the dynamic retransmission frequency can dynamically adapt to network changes according to the estimated packet loss rate and the situation of the data to be sent, avoiding excessive retransmissions or insufficient retransmissions, and improving the transmission efficiency.

[0084] In some embodiments, the calculation formula for the dynamic retransmission frequency is:

[0085]

[0086] Among them, R t represents the dynamic retransmission frequency, P loss represents the estimated packet loss rate, R max represents the maximum frequency of retransmission, α represents the weight of the fixed transmission cost, β represents the sensitivity of the packet loss rate to the retransmission frequency, P th represents the packet loss rate threshold.

[0087] In some embodiments, the calculation formula for the estimated packet loss rate is:

[0088]

[0089] Among them, N represents the total number of network status samples, F(x i ) represents the transmission result function of each data packet, x i represents the network status sample.

[0090] In some embodiments, during the calculation of the dynamic retransmission frequency, the network delay situation can be adaptively adjusted when adjusting the retransmission frequency. Specifically, when the network delay is large, overly frequent retransmissions may cause network congestion and further deteriorate the network condition. To avoid the deteriorated network condition, the maximum frequency of retransmission R max can be set according to the network delay. Similarly, during the calculation of the dynamic retransmission frequency, the network bandwidth limit can also be adaptively adjusted. Specifically, if the bandwidth is already close to saturation, frequent retransmissions will exhaust the bandwidth and affect other normal data transmissions. To avoid affecting normal transmissions, the retransmission strategy is adjusted by monitoring the bandwidth utilization rate. When the bandwidth utilization rate exceeds a certain threshold, the sensitivity β of the packet loss rate to the retransmission frequency can be reduced, and the impact on the bandwidth can be reduced when calculating the dynamic retransmission frequency.

[0091] In some embodiments, in step S400, based on the data transmission target strategy, the data to be sent and the first check information are sent to the distribution network terminal, so that the distribution network terminal processes the data to be sent according to the first check information.

[0092] Specifically, when the data transmission target policy has been determined, the distribution network terminal continuously listens to the network through the network interface to receive the data packets transmitted in the network, avoiding the situation of missing data reception. Among them, the data packet includes the data to be sent and the first check information. When the data to be sent and the first check information are monitored and received by the distribution network terminal, the distribution network terminal extracts the key information from the header of each received data packet. The key information includes: the data segment sequence number, the data packet length, and the first check information. These key information can improve the accuracy and efficiency of the subsequent checksum and recombination process. The received data packets are temporarily stored in the receive buffer, waiting for further processing.

[0093] After the distribution network terminal receives the data packet, it checks the data packet according to the check algorithm used by the sending end. Exemplarily, taking the checksum algorithm as an example to generate the first check information for the data to be sent, the distribution network terminal also uses the same checksum algorithm. For each received data to be sent and the first check information, the check value is recalculated using the extracted original data. In the specific check process, all the bytes in the data segment are added together to obtain a locally calculated checksum; the locally calculated check value is compared with the first check information extracted from the data packet header; if the two are equal, it indicates that the data segment is likely not to have errors during transmission; if they are not equal, it indicates that the data may have been damaged, and corresponding measures can be taken. Among them, the measures taken include but are not limited to requesting the sending end to resend the data segment.

[0094] In some embodiments, after the distribution network terminal receives the data packet, according to the sequence number information extracted from the data packet header, it sorts the data segments received and saved by the distribution network terminal in the buffer to ensure that the data segments are arranged in the order sent by the sending end, improving the accuracy of restoring the original data. Specifically, by judging through the total number information of the data segments or according to the range of the data segment sequence numbers, if it is found that there are missing data segments, the distribution network terminal can send a request to the sending end, asking the sending end to resend the missing data segments; if it is confirmed that the data segments are complete and in the correct order, starting from the first data segment, the data contents of the subsequent data segments are connected in sequence to form the complete data.

[0095] In some embodiments, in step S500, after the distribution network terminal completes the checksum and recombination, it will send feedback information to the sending end. Specifically, the sending end can receive the feedback message from the distribution network terminal according to a dedicated feedback reception mechanism. Among them, these feedback messages may include the status of receiving data, the network status of the distribution network terminal, and the evaluation information of the data transmission quality. It can be understood that the status of receiving data can include whether the reception is successful or whether there are missing or damaged data segments; the network status of the distribution network terminal can include the current bandwidth utilization rate, latency, and packet loss rate; the evaluation information of the data transmission quality can include clarity or integrity.

[0096] In some embodiments, in step S600, the sending end parses and extracts the received feedback message, and stores the feedback information after the parsing and extraction operations in a suitable data structure as a basis for adjusting the transmission strategy. Then, based on the feedback information after the parsing and extraction operations, the sending end checks whether there are any lost or damaged data segments. If the distribution network terminal requests the retransmission of certain data segments, the positions and importance of these data segments in the original data can be determined. Exemplarily, for file transfer, lost data segments may affect the integrity of the file. For video stream transmission, damaged data segments may cause mosaics or lags in the video. The sending end analyzes the network status parameters including bandwidth, latency, and packet loss rate fed back by the distribution network terminal to understand the current network performance, determine whether the network is congested or whether there are areas with high latency or high packet loss, and also consider the evaluation of the transmission quality by the distribution network terminal, especially for applications with high quality requirements. Exemplarily, when the distribution network terminal feeds back poor quality, such as blurred video or noisy audio in video stream transmission, it can be determined that there are factors affecting the quality. Among them, the factors affecting the quality include insufficient bandwidth or network fluctuations, etc.

[0097] In some embodiments, regarding the sending rate of data at the sending end, when the feedback information indicates network congestion or limited processing capacity of the distribution network terminal, the sending end can adjust the reception and processing of the distribution network terminal by reducing the sending rate of data. When the feedback information indicates good network status, sufficient reception and processing capacity, and bandwidth of the distribution network terminal, the sending rate can be appropriately increased to accelerate data transmission.

[0098] Please refer to Figure 2 As shown in the figure, an embodiment of the present application provides a communication optimization system for a distribution network terminal, including:

[0099] A data processing module for obtaining the data to be sent and the first verification information of the data to be sent;

[0100] A network status prediction module for predicting the predicted network status data corresponding to the distribution network terminal through a target real-time network status prediction model;

[0101] A policy generation module for generating a data transmission target policy corresponding to the distribution network terminal according to the predicted network status data;

[0102] A policy sending module for sending the data to be sent and the first verification information to the distribution network terminal based on the data transmission target policy, so that the distribution network terminal processes the data to be sent according to the first verification information;

[0103] A data receiving module for receiving feedback information sent by a distribution network terminal;

[0104] A policy optimization module for adjusting the data transmission target policy according to the feedback information.

[0105] It can be understood that the content in the above method embodiments is applicable to the system embodiments of the present application. The functions specifically implemented in the system embodiments of the present application are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those in the above method embodiments.

[0106] An embodiment of the present application also provides a computer device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the above-mentioned large model-assisted learning processing method is implemented. The computer device can be any intelligent terminal including a tablet computer, a vehicle-mounted computer, etc.

[0107] It can be understood that the content in the above method embodiments is applicable to the device embodiments of the present application. The functions specifically implemented in the device embodiments of the present application are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those in the above method embodiments.

[0108] Please refer to Figure 3 as shown in Figure 3 which shows the hardware structure of a computer device according to another embodiment. The computer device includes:

[0109] A processor 410, which can be implemented in a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0110] A memory 420, which can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 420 can store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of the present specification through software or firmware, the relevant program codes are stored in the memory 420 and are called by the processor 410 to execute the large model-assisted learning processing method of the embodiments of the present application;

[0111] An input / output interface 430 for implementing information input and output;

[0112] A communication interface 440, which is used to implement the communication interaction between this device and other devices. It can achieve communication through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.);

[0113] A bus 450, which transmits information between various components of the device (such as a processor 410, a memory 420, an input / output interface 430, and a communication interface 440);

[0114] Among them, the processor 410, the memory 420, the input / output interface 430, and the communication interface 440 achieve communication connections with each other inside the device through the bus 450.

[0115] The embodiments described in the embodiments of this application are for more clearly explaining the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. Those skilled in the art know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are equally applicable to similar technical problems.

[0116] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than those shown in the figures, or combine some steps, or different steps.

[0117] The device embodiments described above are only illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0118] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules in the systems and devices, can be implemented as software, firmware, hardware, and their appropriate combinations.

[0119] It should be understood that in this application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that there can be three relationships. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist at the same time. Here, A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one (individual) of the following" or its similar expressions refer to any combination of these items, including any combination of single item (individual) or plural items (individuals). For example, at least one (individual) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0120] In several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the above division of units is only a logical function division. In actual implementation, there can be other division methods. For example, 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 displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces. The indirect coupling or communication connection of devices or units can be in electrical, mechanical or other forms.

[0121] In addition, each functional unit in various embodiments of this application can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0122] 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 this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in various embodiments of this application. And the aforementioned storage medium includes: USB flash drive, mobile hard disk, read-only memory (ROM for short), random access memory (RAM for short), magnetic disk or optical disc and other various media that can store programs.

[0123] The preferred embodiments of the embodiments of the present application have been described above with reference to the accompanying drawings. This does not limit the scope of the rights of the embodiments of the present application. Any modifications, equivalent replacements, and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall fall within the scope of the rights of the embodiments of the present application.

Claims

1. A communication optimization method for a distribution network terminal, characterized in that: The method comprises the following steps: Acquire data to be sent and first verification information of the data to be sent; Predicting predicted network status data corresponding to the distribution network terminal by using a target real-time network status prediction model; Generating a data transmission target strategy corresponding to the distribution network terminal according to the predicted network status data; Sending the data to be sent and the first verification information to the distribution network terminal based on the data transmission target strategy, so that the distribution network terminal processes the data to be sent according to the first verification information; Receiving feedback information sent by the distribution network terminal; The data transmission target strategy is adjusted according to the feedback information.

2. The method according to claim 1, characterized in that The obtaining of the data to be sent and the first verification information of the data to be sent includes: Segmenting the original data to obtain segmented data; Generate first verification information of the segmented data.

3. The method according to claim 1, characterized in that The predicting of the predicted network status data corresponding to the distribution network terminal by using the target real-time network status prediction model includes: Obtaining real-time network status data corresponding to the distribution network terminal at the current moment; Update the first historical network status data corresponding to the distribution network terminal according to the real-time network status data to obtain the second historical network status data; The current real-time network status prediction model is trained by using the second historical network status data to obtain a target real-time network status prediction model; The target real-time network status prediction model is used to predict the predicted network status data corresponding to the distribution network terminal.

4. The method according to claim 3, characterized in that The target real-time network status prediction model includes a time series analysis model and a neural network model, and the data output of the time series analysis model is the data input of the neural network model.

5. The method according to claim 1, characterized in that The generating of the data transmission target strategy corresponding to the distribution network terminal according to the predicted network status data comprises: Determine a target transmission path and a target time window according to the predicted network status data; Calculating an estimated packet loss rate of the target transmission path within the target time window; Determining a working state of a dynamic retransmission strategy according to the estimated packet loss rate; The data transmission target strategy includes the target transmission path, the target time window and the working status of the dynamic retransmission strategy.

6. The method according to claim 5, characterized in that The determining the working state of the dynamic retransmission strategy according to the estimated packet loss rate includes: Determining a dynamic retransmission frequency of a dynamic retransmission strategy according to the estimated packet loss rate; The working state of the dynamic retransmission strategy is controlled based on the dynamic retransmission frequency.

7. The method according to claim 6, characterized in that The calculation formula of the dynamic retransmission frequency is: Among them, R t Indicates the dynamic retransmission frequency, P loss represents the estimated packet loss rate, R max represents the maximum frequency of retransmission, α represents the weight of fixed transmission cost, β represents the sensitivity of packet loss rate to retransmission frequency, P th Indicates the packet loss rate threshold.

8. The method according to claim 7, characterized in that The calculation formula for estimating the packet loss rate is: Where N represents the total number of network status samples, F(x i ) represents the transmission result function of each data packet, x i Represents a network status sample.

9. A communication optimization system for distribution network terminals, characterized in that: include: A data processing module, used for acquiring data to be sent and first verification information of the data to be sent; A network status prediction module, used to predict the predicted network status data corresponding to the distribution network terminal through a target real-time network status prediction model; A strategy generation module, used to generate a data transmission target strategy corresponding to the distribution network terminal according to the predicted network status data; a strategy sending module, configured to send the data to be sent and the first verification information to the distribution network terminal based on the data transmission target strategy, so that the distribution network terminal processes the data to be sent according to the first verification information; A data receiving module, used for receiving feedback information sent by the distribution network terminal; A strategy optimization module is used to adjust the data transmission target strategy according to the feedback information.

10. A computer device, characterized in that: The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 8 when executing the computer program.

Images