Mobile terminal network optimization method based on multi-login-point expected real-time virtual networking algorithm in communication network

Abstract

The invention discloses a mobile terminal network optimization method based on a multi-login-point expected real-time virtual networking algorithm in a communication network, and the method comprises the steps: S1, dividing interactive network participants into the same virtual network, constructing a user network node graph, and analyzing the connection weight and relevance between nodes; s2, predicting the future value of each node, dynamically judging the activeness, importance and association degree of the node, and determining whether to replace the existing node; s3, dynamically adjusting the network structure, and optimizing the connection relationship between the nodes; s4, predicting a vehicle and a moving path of the user, and dynamically adjusting node connection; and S5, judging the importance of the node in the network, and intelligently selecting to join or replace the existing node. According to the method, the basic data of the user is extracted, the data of the network nodes to be connected by the user is obtained in advance, the virtual network resources are adjusted in advance, the delay and interruption phenomena during network switching are remarkably reduced, and the method is particularly suitable for network interaction scenes with high real-time performance requirements.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless communication algorithms, and in particular to a mobile terminal network optimization method based on a multi-login point expected real-time virtual networking algorithm in a communication network. Background Art

[0002] In current multi-person interactive network applications, network architecture optimization technologies based on mobile communications, including virtual networks and network slicing, are often used to improve the network stability of real-time interactions. However, the network connection of mobile terminals faces frequent network switching problems in high-speed mobile scenarios (such as high-speed rail, buses, etc.), especially the rapid switching between base stations, which makes it difficult for existing network optimization technologies to ensure stable network connections and low-latency interactive experiences. In addition, the current network switching mechanism usually relies on manual intervention or passive waiting by users, lacks automatic detection of user mobility status and advance networking, and cannot achieve smooth network switching. To this end, it is necessary to study and design a dynamic network connection networking algorithm suitable for this scenario to improve and optimize the network connection of mobile terminals in multi-person interactive network applications. Summary of the invention

[0003] In view of the above problems, the present invention provides a mobile terminal network optimization method based on a multi-login point expected real-time virtual networking algorithm in a communication network, comprising the following steps:

[0004] S1, in interactive network interaction, participants are divided into the same virtual network and a user network node graph is constructed, each user's access mode, mobile path and other data are mapped to the user network node graph, the connection weight and correlation between nodes are analyzed, and a prediction basis is provided for node joining and replacement;

[0005] S2, predicts the future value of each node through the node selection algorithm, so as to select the best replacement node when adding a new node, and dynamically judges the activity, importance and relevance of the node in combination with the replacement algorithm to decide whether to replace the existing node;

[0006] S3, centrally manage user nodes in the new communication network, dynamically adjust the network structure, and optimize the connection relationship between nodes;

[0007] S4, predicts the user’s transportation and movement path, and dynamically adjusts node connections;

[0008] S5, when a new node joins, it determines the importance of the node in the network and intelligently chooses to join or replace the existing node.

[0009] As a further illustration of the present invention, in the user network node optimization process, the algorithm performance is optimized in real time, and the parameters are continuously adjusted through big data analysis and Bayesian optimization.

[0010] Furthermore, in step S1, through historical data analysis, the historical behavior of the user node is used as the underlying base point data to optimize the subsequent node selection strategy in the communication network.

[0011] Furthermore, in step S1, the connection weights and associations between nodes are analyzed by a graph neural network (GNN).

[0012] Furthermore, in step S2, the future value of each node is predicted by DQN, and the best replacement node is selected when a new node is added. In combination with the policy gradient algorithm, the activity, importance, and relevance of the node are dynamically judged to decide whether to replace the existing node.

[0013] Furthermore, in step S3, the SDN controller is used to centrally manage user nodes, placing all users under the management of the SDN controller.

[0014] Furthermore, in step S3, the SDN controller dynamically adjusts the network structure based on real-time data traffic, geographic location information, etc., and optimizes the connection relationship between nodes through automated routing.

[0015] Furthermore, in step S4, the node connections are dynamically adjusted in combination with the LSTM model.

[0016] Furthermore, in step S4, DQN and graph neural network are used to determine the importance of nodes in the network and intelligently select to join or replace existing nodes. The algorithm predicts the user's future connection needs and traffic patterns and selects the node combination that best suits the current network structure.

[0017] Furthermore, in step S5, by predicting the user's future connection needs and traffic patterns, the node combination that best suits the current network structure is selected; when predicting future network traffic needs and node changes, the rail transit prediction model is used to optimize base stations and nodes.

[0018] Beneficial effects of the present invention:

[0019] The method of the present invention is applied in large-scale interactive network interaction. Participants are divided into the same virtual network through SDN and other technical means, the login status of base stations or wifi is dynamically monitored, the mobile status classification is fitted, and the network node data to which the user is about to connect is obtained in advance in combination with basic data. When the mobile terminal frequently switches networks, the virtual network resources are adjusted in advance, and the established SDN set is adjusted, which significantly reduces the delay and interruption phenomenon during network switching and realizes seamless switching, so as to ensure the network connection stability and interaction fluency of users in high-speed movement in the new communication network, especially in network interaction scene applications with high real-time requirements such as online games and video conferencing. It has significant advantages. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 The present invention is a flowchart of a mobile terminal network optimization method based on an expected real-time virtual networking algorithm of multiple login points in a communication network. DETAILED DESCRIPTION

[0021] The following is a detailed description of the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments.

[0022] In the description of the present invention, it should be understood that the terms "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside", "outside", "first", "second", etc., indicating directions or positions or sequential relationships, are based on the directions or positions or sequential relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention.

[0023] A mobile terminal network optimization method based on a multi-login point expected real-time virtual networking algorithm in a communication network is invented, comprising the following steps:

[0024] S1, in interactive network interactions, participants are divided into the same virtual network and a user network node graph is constructed. Each user's access mode, mobile path and other data are mapped to the user network node graph, and the connection weights and correlations between nodes are analyzed to provide a prediction basis for node joining and replacement. Specifically, through historical data analysis, the connection weights and correlations between nodes are further analyzed through the graph neural network (GNN) to provide a prediction basis for node joining and replacement. The historical behavior of user nodes is used as the underlying base point data to optimize subsequent node selection strategies in the communication network. Graph neural network is a deep learning model suitable for graph structured data. By learning the connection relationship and weights between nodes, GNN can be used to analyze the correlation between user nodes and provide data support for node selection and replacement.

[0025] S2, design a node selection algorithm and a replacement algorithm. The node selection algorithm is used to predict the future value of each node so that the best replacement node can be selected when a new node is added. The replacement algorithm is used to dynamically judge the activity, importance and relevance of the node to decide whether to replace the existing node. A node selection and replacement algorithm based on DQN and policy gradient can be designed. The future value of each node is predicted by DQN so that the best replacement node can be selected when a new node is added. Combined with the policy gradient algorithm, the activity, importance and relevance of the node are dynamically judged to decide whether to replace the existing node, ensuring the optimization of network performance after the new node is added. DQN is a deep reinforcement learning algorithm based on Q learning. It approximates the Q value function through a neural network and is used to find the optimal action in a complex state space. In the present invention, DQN is used to predict the future value of the node and help select the node to be replaced or retained. Policy gradient is a deep reinforcement learning method that directly optimizes the strategy and outputs the optimal behavior by learning a policy function. The policy gradient algorithm is applicable to the continuous action space, which helps the present invention select the optimal strategy during node replacement and resource allocation to ensure the adaptability and stability of the network.

[0026] Specifically, a graph set is constructed to represent the relationship between nodes as a directed graph, including a node set and an edge set. The node represents the out-degree when it is a starting point, and represents the in-degree when it is an arrival point. The weight of each edge is set according to the flow between the corresponding nodes, and the network flow parameters of the network graph are constructed based on this.

[0027] Node selection algorithm: According to the maximum flow algorithm, the future value of each node is predicted. In the constructed network graph, the maximum flow path is calculated, and the nodes connected on the path are identified as the best nodes. At the same time, the weight of the node will be dynamically adjusted according to the changes in network status and flow, ensuring that the selected node always has a high utility.

[0028] Replacement algorithm: The replacement algorithm is a dynamic scoring and evaluation mechanism used to determine whether the current node needs to be replaced and select the optimal replacement node to optimize the network structure and overall performance. The replacement algorithm of this embodiment first establishes a scoring framework based on the MRU mode, which is subdivided into three core indicators: node activity, survivability and optimal weight. Activity is used to measure the recent use frequency of the node and reflect its current use value; survivability predicts its future continued effectiveness by analyzing the state change trend of the node in the network; the optimal weight comprehensively evaluates the performance of the node in the network flow parameters and its contribution to the overall function of the network. The replacement algorithm calculates the three scoring indicators of all nodes in real time and obtains the comprehensive score of the node according to the set scoring weight. When the comprehensive score of the node is lower than the preset threshold, it indicates that the use value of the node is no longer sufficient to meet the network needs, and the replacement algorithm triggers the replacement operation. In addition, when a new node joins the network, if the comprehensive score of the new node is significantly higher than that of the existing node, the replacement algorithm automatically determines it as a potential priority replacement object. During the replacement process, the replacement algorithm gives priority to the node with the highest score and the greatest contribution to the overall network as the replacement object based on the comprehensive score of the node. After the replacement is completed, the replacement algorithm dynamically updates the scoring indicators of each node according to the latest state of the network, thereby ensuring that the scoring framework is always consistent with the actual situation of the network. Through this dynamic evaluation and replacement mechanism, the replacement algorithm can continuously optimize the network structure, so that it can always maintain an efficient and stable operating state in a constantly changing environment.

[0029] S3, centrally manage user nodes in the new communication network, dynamically adjust the network structure, and optimize the connection relationship between nodes. Specifically, the SDN controller can be used in the new communication network to centrally manage user nodes and put all users under the management of the SDN controller. The SDN controller dynamically adjusts the network structure based on real-time data traffic, geographic location information, etc., optimizes the connection relationship between nodes through automated routing, and ensures the user's connection stability and data transmission efficiency in the new communication network.

[0030] S4, predict the user's means of transportation and movement path, and dynamically adjust the node connection. Specifically, the node connection can be dynamically adjusted in combination with the LSTM model. LSTM is a time series model that is suitable for the prediction of long-term dependent data. It remembers the information of long time series through its unique structure and is used to predict the changes in the user's movement path. The recurrent neural network RNN ​​is a neural network model for processing sequence data, which processes time series data through the cyclic structure between nodes. In this embodiment, the user's movement path is predicted by the recurrent neural network RNN.

[0031] S5, when a new node is added, the importance of the node in the network is judged, and the existing node is intelligently selected to join or replace. In this embodiment, DQN and graph neural network are used to judge the importance of the node in the network, and the existing node is intelligently selected to join or replace. The algorithm predicts the user's future connection needs and traffic patterns, selects the node combination that best suits the current network structure, and ensures efficient and stable resource allocation of the new communication network. In this embodiment, the state space model is used to predict future node positions and traffic demands. The state space model is used to describe the dynamic changes in the system state, and the future behavior of the system can be predicted through state variables and observation variables. When predicting future network traffic demands and node changes, the rail transit prediction model is used to optimize base stations and nodes. The state space model is used to analyze the future three-stage traffic status in advance, and a prediction model is constructed to optimize the node connection order to ensure a smooth transition of connections when the network structure is adjusted.

[0032] In this embodiment, in the user network node optimization process, the algorithm performance is optimized in real time, and the parameters are continuously adjusted through big data analysis and Bayesian optimization. Bayesian optimization is an optimization algorithm based on the Bayesian probability model, which is suitable for tuning complex system parameters. The present invention adjusts parameters such as node replacement and bandwidth allocation through Bayesian optimization to improve network performance. The response speed of the evaluation algorithm, the accuracy of node replacement and the stability of the network are evaluated, and the strategies such as bandwidth, node importance weight and path allocation are optimized in combination with the AI ​​model to ensure that the performance in the new communication network remains robust under different environments.

[0033] The method of the present invention is applied in large-scale interactive network interaction. Participants are divided into the same virtual network through SDN and other technical means, the login status of base stations or wifi is dynamically monitored, the mobile status classification is fitted, and the network node data to which the user is about to connect is obtained in advance in combination with basic data. When the mobile terminal frequently switches networks, the virtual network resources are adjusted in advance, and the established SDN set is adjusted, which significantly reduces the delay and interruption phenomenon during network switching and realizes seamless switching, so as to ensure the network connection stability and interaction fluency of users in high-speed movement in the new communication network, especially in network interaction scene applications with high real-time requirements such as online games and video conferencing. It has significant advantages.

[0034] The above only describes the preferred embodiments of the present invention, but it cannot be understood as limiting the claims. The present invention is not limited to the above embodiments, and its specific structure is allowed to be changed. In short, all changes made within the protection scope of the independent claims of the present invention are within the protection scope of the present invention.

Claims

1. A mobile terminal network optimization method based on a multi-login point expected real-time virtual networking algorithm in a communication network, characterized in that: The following steps are involved: S1, in interactive network interaction, participants are divided into the same virtual network and a user network node graph is constructed, each user's access mode, mobile path and other data are mapped to the user network node graph, the connection weight and correlation between nodes are analyzed, and a prediction basis is provided for node joining and replacement; S2, predicts the future value of each node through the node selection algorithm, and dynamically determines the activity, importance and relevance of the node in combination with the replacement algorithm to decide whether to replace the existing node; S3, centrally manage user nodes in the new communication network, dynamically adjust the network structure, and optimize the connection relationship between nodes; S4, predicts the user’s transportation and movement path, and dynamically adjusts node connections; S5, when a new node joins, it determines the importance of the node in the network and intelligently chooses to join or replace the existing node.

2. According to claim 1, the mobile terminal network optimization method based on the multi-login point expected real-time virtual networking algorithm in the communication network is characterized by: In the user network node optimization process, the algorithm performance is optimized in real time, and the parameters are continuously adjusted through big data analysis and Bayesian optimization.

3. According to claim 1, the mobile terminal network optimization method based on the multi-login point expected real-time virtual networking algorithm in the communication network is characterized by: In step S1, through historical data analysis, the historical behavior of the user node is used as the underlying base point data to optimize the subsequent node selection strategy in the communication network.

4. The mobile terminal network optimization method based on the multi-login point expected real-time virtual networking algorithm in the communication network according to claim 1 is characterized by: In step S1, the connection weights and correlations between nodes are analyzed through the graph neural network GNN.

5. According to claim 1, the mobile terminal network optimization method based on the multi-login point expected real-time virtual networking algorithm in the communication network is characterized by: In step S2, the future value of each node is predicted by DQN, and the best replacement node is selected when a new node is added. The activity, importance, and relevance of the node are dynamically judged in combination with the policy gradient algorithm to decide whether to replace the existing node.

6. The mobile terminal network optimization method based on the multi-login point expected real-time virtual networking algorithm in the communication network according to claim 1 is characterized by: In step S3, the SDN controller is used to centrally manage user nodes, and all users are placed under the management of the SDN controller.

7. The mobile terminal network optimization method based on the multi-login point expected real-time virtual networking algorithm in the communication network according to claim 7 is characterized by: In step S3, the SDN controller dynamically adjusts the network structure based on real-time data traffic, geographic location information, etc., and optimizes the connection relationship between nodes through automated routing.

8. The mobile terminal network optimization method based on the multi-login point expected real-time virtual networking algorithm in the communication network according to claim 1 is characterized by: In step S4, the node connections are dynamically adjusted in combination with the LSTM model.

9. The mobile terminal network optimization method based on the multi-login point expected real-time virtual networking algorithm in the communication network according to claim 1 is characterized by: In step S4, DQN and graph neural network are used to determine the importance of nodes in the network and intelligently select to join or replace existing nodes. The algorithm predicts the user's future connection needs and traffic patterns and selects the node combination that best suits the current network structure.

10. The mobile terminal network optimization method based on the multi-login point expected real-time virtual networking algorithm in the communication network according to claim 1, characterized in that: In step S5, by predicting the user's future connection needs and traffic patterns, the node combination that best suits the current network structure is selected; when predicting future network traffic needs and node changes, the rail transit prediction model is used to optimize base stations and nodes.

Images