Communication network optimization control system and method based on multi-dimensional data driving

Abstract

The invention discloses a communication network optimization control system and method based on multi-dimensional data driving. The method comprises the steps that a communication collection module collects multiple pieces of network flow data, user behavior data and exchange mode data at all communication nodes in a communication network in real time; the first analysis unit inputs the network traffic data into a pre-selected trained traffic prediction model, and predicts to obtain network predicted traffic after a preset time period; the second analysis unit analyzes and processes the behavior data of the users to obtain network communication habits of the users; the third analysis unit classifies the exchange mode data according to a preset flow classification algorithm to obtain multiple pieces of classification data; the comprehensive processing unit obtains a comprehensive analysis result according to the network prediction flow, the network communication habits and the classified data; and the communication adjusting module dynamically adjusts bandwidth resources, routing selection and flow distribution at each communication node according to the comprehensive analysis result. According to the invention, optimal control of network communication resources is realized.

Description

Technical Field

[0001] The present invention relates to the field of network communication technology, and in particular to a communication network optimization control system and method based on multi-dimensional data drive. Background Art

[0002] With the rapid development of the Internet and communication technologies, global network traffic continues to grow, and the complexity of network communications continues to increase. Especially driven by emerging technologies such as 5G and the Internet of Things (IoT), the number of users, devices, data traffic, and communication modes in the network have exploded. Traditional network management methods and resource allocation strategies can no longer meet the needs of modern communication networks for efficiency, intelligence, and dynamic adjustment.

[0003] In traditional network management, traffic monitoring and control often rely on simple traffic thresholds and static policies. Although these methods guarantee the basic stability of the network to a certain extent, they cannot cope with the ever-changing network demands and complex communication patterns. For example, when the network experiences a traffic peak, traditional methods often cannot predict traffic changes in advance, and thus cannot respond in time, resulting in increased network latency, bandwidth waste, or local network congestion. When a large number of user behaviors change in the network (such as sudden high bandwidth demands or different user usage habits), traditional methods are even less able to make flexible adjustments.

[0004] In addition, the switching patterns in modern communication networks are becoming more and more complex. The congestion levels of different network paths, the switching frequencies, and the diversity of path selection make network congestion more difficult to predict. As the network topology changes, static routing and bandwidth allocation methods cannot ensure the best network performance in all cases.

[0005] In order to solve these problems, network management systems need more intelligent and dynamic technologies to process and analyze data from different network nodes in real time. These data include not only traditional network traffic data, but also user behavior data and exchange pattern data. By comprehensively analyzing these multi-dimensional data, network management systems can more accurately predict network traffic, identify user needs, discover potential network bottlenecks, and make dynamic adjustments based on this, thereby effectively improving the utilization efficiency of network resources and enhancing user experience.

[0006] At present, although there are some network optimization technologies based on traffic prediction and user behavior analysis, these technologies still face several key problems:

[0007] Existing network control technologies are usually based on simple analysis of historical data, lacking real-time and intelligence, and unable to cope with the instantaneous changing network environment;

[0008] In complex networks, how to combine different types of data for comprehensive analysis and optimize network resource allocation accordingly remains a technical challenge.

[0009] Therefore, how to flexibly respond to changes in user behavior and optimize network resource allocation through comprehensive analysis of multi-dimensional data has become a key technical issue that needs to be urgently solved in the field of network communication management. Summary of the invention

[0010] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a communication network optimization control system and method driven by multi-dimensional data, which is used to realize comprehensive analysis of multi-dimensional data and achieve optimal control of network communication resources.

[0011] To achieve the above object, the present invention provides the following technical solution: a communication network optimization control system driven by multi-dimensional data, comprising:

[0012] A communication collection module is used to collect multiple network traffic data, user behavior data and exchange mode data at each communication node in the communication network in real time;

[0013] The data analysis module is connected to the communication acquisition module and includes:

[0014] A first analysis unit, used for inputting each of the network traffic data into a pre-selected and trained traffic prediction model to predict the network traffic after a preset time period;

[0015] A second analysis unit is used to analyze the user behavior data to obtain the network communication habits of each user;

[0016] A third analysis unit, configured to classify each of the exchange mode data according to a preset traffic classification algorithm to obtain a plurality of classification data;

[0017] A comprehensive processing unit, connected to the first analysis unit, the second analysis unit and the third analysis unit respectively, for obtaining a comprehensive analysis result according to the network predicted traffic, the network communication habits and the classification data processing;

[0018] The communication adjustment module is connected to the data analysis module and is used to dynamically adjust the bandwidth resources, route selection and traffic diversion at each of the communication nodes according to the comprehensive analysis results.

[0019] Furthermore, the network traffic data includes bandwidth utilization, packet loss rate, delay, data packet size and traffic distribution, the user behavior data includes access frequency, access type, access time, device type and geographic location, and the exchange mode data includes data packet exchange frequency, exchange path and path congestion level.

[0020] Furthermore, it also includes:

[0021] A detection module, connected to the data analysis module, for detecting the device load at each of the communication nodes in real time;

[0022] A calculation module, connected to the communication acquisition module and the data analysis module, for inputting the bandwidth utilization, the packet loss rate, the delay and the data packet size into a preset comprehensive congestion calculation formula to calculate a communication congestion index;

[0023] A storage module, connected to the data analysis module, for storing a plurality of historical training data of each of the communication nodes, wherein the historical training data includes historical load, historical congestion index and historical traffic data;

[0024] The first analysis unit comprises:

[0025] A training subunit is used to introduce an initial traffic model, and take each of the historical loads, each of the historical congestion indexes and each of the historical traffic data as input, and take each of the historical traffic data after the next time period as output, retrain the initial traffic model, and obtain the traffic prediction model;

[0026] The prediction subunit is connected to the training subunit and is used to input the device load, the network traffic data and the communication congestion index at the current moment into the traffic prediction model to predict the network predicted traffic after a preset time period.

[0027] Furthermore, the comprehensive congestion calculation formula is configured as:

[0028]

[0029] Among them, C i Used to represent the communication congestion index, U b (t) is used to represent the bandwidth utilization at time t, L p (t) is used to represent the packet loss rate at time t, L t (t) is used to represent the delay at time t, S p (i) is used to represent the size of the i-th data packet, T is used to represent the time window of calculation, α, β, γ, δ are respectively used to represent constants that control the influence of various parameters on the communication congestion index, and N is used to represent the total number of data packets.

[0030] Furthermore, the comprehensive processing unit includes:

[0031] A fusion subunit is used to fuse the network predicted traffic, the network communication habits and the classified data into a preset data fusion model to obtain multi-source fusion data;

[0032] The analysis subunit is connected to the fusion subunit and is used to analyze the multi-source fusion data to obtain comprehensive analysis data.

[0033] Furthermore, the formula configuration of the data fusion model is:

[0034]

[0035] Among them, M f It is used to represent the multi-source fusion data, N f (t) is used to represent the predicted network traffic at time t, C f (t) is used to represent the traffic classification data at time t, R i It is used to represent the data source signal strength at the communication node, α2, β2, γ2, δ2, η2, κ2 are used to represent the flow prediction influence constant, communication habit influence constant, flow classification influence constant, time attenuation influence constant and signal strength influence constant, respectively, μ(H c (i)) is used to indicate the volatility of the network communication habits.

[0036] Furthermore, the detection module is also used to detect multiple external environment data and multiple network security event data at each of the communication nodes in real time;

[0037] The comprehensive processing unit also includes an optimization subunit, connected to the fusion subunit, for inputting each of the external environment data and the communication habit influence constant into a preset first constant optimization formula to obtain a communication habit optimization influence constant, and inputting each of the network security event data and the traffic prediction influence constant into a preset second constant optimization formula to obtain a traffic prediction optimization influence constant;

[0038] The data fusion model is updated based on the communication habit optimization influence constant and the traffic prediction optimization influence constant.

[0039] Furthermore, the external environment data includes temperature variation, terrain interference, environmental magnetic field, and environmental noise;

[0040] The network security event data includes password cracking data and virus infection data.

[0041] Furthermore, the first constant optimization formula is configured as:

[0042]

[0043] Among them, β2′ is used to represent the communication habit optimization influence constant, ΔT is used to represent the temperature change, φ is used to represent the influence weight of the temperature change, G d is used to represent the terrain interference amount, χ is used to represent the influence weight of the terrain interference amount, and B e is used to represent the environmental magnetic field, ψ is used to represent the influence weight of the environmental magnetic field, N e is used to represent the environmental noise, and ω is used to represent the influence weight of the environmental noise;

[0044] The second constant optimization formula is configured as:

[0045]

[0046] Among them, α2′ is used to represent the flow prediction optimization influence constant, P k Used to represent the password cracking data, T k Used to represent the time delay associated with password cracking, a k 、b k They are used to represent the first cracking related constant and the second cracking related constant related to password cracking, V j Used to represent the virus infection data, E j Used to represent environmental factors data related to viral infection, c j d j They are respectively used to represent a first infection-related constant and a second infection-related constant related to virus infection, n is used to represent the total amount of the password cracking data, and m is used to represent the total amount of the virus infection data.

[0047] A communication network optimization control method based on multi-dimensional data drive, applied to the above-mentioned communication network optimization control system based on multi-dimensional data drive, comprising:

[0048] Step S1, the communication collection module collects multiple network traffic data, user behavior data and exchange mode data at each communication node in the communication network in real time;

[0049] Step S2, the first analysis unit inputs each of the network traffic data into the pre-selected and trained traffic prediction model to predict the network traffic after a preset time period, the second analysis unit analyzes each of the user behavior data to obtain the network communication habits of each user, and the third analysis unit classifies each of the exchange mode data according to a preset traffic classification algorithm to obtain multiple classification data;

[0050] Step S3, the comprehensive processing unit obtains a comprehensive analysis result according to the network predicted traffic, the network communication habits and the classification data;

[0051] Step S4: the communication regulation module dynamically adjusts the bandwidth resources, routing selection and traffic diversion at each of the communication nodes according to the comprehensive analysis results.

[0052] Beneficial effects of the present invention:

[0053] The present invention realizes the optimization control of network communication resources by comprehensively analyzing multi-dimensional data (including network traffic, user behavior and exchange mode data). It can predict future network traffic based on real-time data and dynamically adjust bandwidth, routing and traffic distribution to avoid network congestion and overload, and ensure the stability and efficiency of network operation.

[0054] In addition, the present invention can intelligently analyze user behavior and accurately identify the communication habits of different users, so that network resources can be allocated in a personalized manner according to user needs. This not only improves user experience, but also improves resource utilization efficiency and ensures high-quality network services.

[0055] Finally, the present invention refines traffic classification and dynamic adjustment mechanisms so that different types of network traffic can be given reasonable priority and bandwidth allocation, avoiding resource waste and reducing network operating costs, thereby improving the overall network service quality and operating efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a structural schematic diagram of a communication network optimization control system driven by multi-dimensional data in the present invention;

[0057] Figure 2 is a schematic structural diagram of the first analysis unit in the present invention;

[0058] Figure 3 It is a structural schematic diagram of the comprehensive processing unit in the present invention;

[0059] Figure 4 It is a schematic diagram of the steps of the communication network optimization control method driven by multi-dimensional data in the present invention.

[0060] Figure numerals: 1. Communication acquisition module; 2. Data analysis module; 21. First analysis unit; 211. Training subunit; 212. Prediction subunit; 22. Second analysis unit; 23. Third analysis unit; 24. Comprehensive processing unit; 241. Fusion subunit; 242. Analysis subunit; 243. Optimization subunit; 3. Communication adjustment module; 4. Detection module; 5. Calculation module; 6. Storage module. DETAILED DESCRIPTION

[0061] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. The same parts are represented by the same reference numerals. It should be noted that the words "front", "rear", "left", "right", "upper" and "lower" used in the following description refer to directions in the accompanying drawings, and the words "bottom surface" and "top surface", "inner" and "outer" refer to directions toward or away from the geometric center of a specific component, respectively.

[0062] Example 1, reference Figure 1 , which is the first embodiment of the present invention, provides a communication network optimization control system driven by multi-dimensional data, which can realize comprehensive analysis of multi-dimensional data and realize optimization control of network communication resources, including:

[0063] The communication collection module 1 is used to collect multiple network traffic data, user behavior data and exchange mode data at each communication node in the communication network in real time;

[0064] The data analysis module 2 is connected to the communication acquisition module 1 and includes:

[0065] The first analysis unit 21 is used to input each network traffic data into a pre-selected and trained traffic prediction model to predict the network traffic after a preset time period;

[0066] The second analysis unit 22 is used to analyze the behavior data of each user and obtain the network communication habits of each user;

[0067] The third analysis unit 23 is used to classify each exchange mode data according to a preset traffic classification algorithm to obtain a plurality of classification data;

[0068] The comprehensive processing unit 24 is connected to the first analysis unit 21, the second analysis unit 22 and the third analysis unit 23 respectively, and is used to obtain a comprehensive analysis result according to the network prediction flow, network communication habits and various classification data processing;

[0069] The communication adjustment module 3 is connected to the data analysis module 2 and is used to dynamically adjust the bandwidth resources, routing selection and traffic diversion at each communication node according to the comprehensive analysis results.

[0070] Preferably, network traffic data includes bandwidth utilization, packet loss rate, latency, data packet size and traffic distribution, user behavior data includes access frequency, access type, access time, device type and geographic location, and exchange mode data includes data packet exchange frequency, exchange path and path congestion level.

[0071] Working principle of embodiment 1:

[0072] The communication acquisition module 1 is used to collect various data of multiple nodes in the communication network in real time. Specifically, the communication acquisition module 1 collects network flow data, user behavior data and exchange mode data at each communication node in real time through network sensors or network devices (such as switches and routers).

[0073] The data analysis module 2 is composed of three analysis units, namely the first analysis unit 21, the second analysis unit 22 and the third analysis unit 23. The first analysis unit 21 processes the collected network traffic data using the trained traffic prediction model to predict the network traffic in the future. The prediction results can be used to dynamically adjust the network bandwidth and resource allocation to avoid network congestion.

[0074] The second analysis unit 22 collects and analyzes the communication behaviors of users to extract the communication habits of users. For example, some users have higher bandwidth requirements in a specific time period, while other users have lower bandwidth requirements. Based on these analysis results, appropriate network resources can be allocated to different user groups to provide personalized network services.

[0075] The third analysis unit 23 classifies the exchange mode data according to a preset traffic classification algorithm, distinguishing different types of network traffic such as real-time traffic, batch traffic, video traffic, etc. Based on these classifications, the system can prioritize high-priority traffic, such as real-time video or voice communication, to ensure its quality.

[0076] The comprehensive processing unit 24 receives the output data from the three analysis units, including network traffic prediction results, user behavior analysis results and traffic classification data. It combines these data, conducts multi-dimensional comprehensive analysis, and generates a comprehensive comprehensive analysis result. The comprehensive analysis result provides decision support for the allocation of network resources to ensure that the system can achieve optimal resource allocation.

[0077] According to the comprehensive analysis results output by the comprehensive processing unit 24, the communication adjustment module 3 dynamically adjusts each communication node in the communication network. Specifically, the communication adjustment module 3 can adjust bandwidth allocation, routing selection and traffic diversion strategies. For example, when it is predicted that the traffic of a node will exceed its bandwidth limit, the system will automatically adjust the routing path of the node, or give priority to important traffic according to the traffic classification results to avoid network performance degradation.

[0078] Preferably, it also includes:

[0079] The detection module 4 is connected to the data analysis module 2 and is used to detect the equipment load at each communication node in real time;

[0080] The calculation module 5 is connected to the communication acquisition module 1 and the data analysis module 2, and is used to input the bandwidth utilization, packet loss rate, delay and data packet size into a preset comprehensive congestion calculation formula to calculate the communication congestion index;

[0081] The storage module 6 is connected to the data analysis module 2 and is used to store a plurality of historical training data of each communication node, wherein the historical training data includes historical load, historical congestion index and historical traffic data;

[0082] like Figure 2 As shown, the first analysis unit 21 includes:

[0083] The training subunit 211 is used to introduce the initial traffic model, and take each historical load, each historical congestion index and each historical traffic data as input, take each historical traffic data after the next time period as output, retrain the initial traffic model, and obtain a traffic prediction model;

[0084] The prediction subunit 212 is connected to the training subunit 211 and is used to input the current device load, network traffic data and communication congestion index into the traffic prediction model to predict the network predicted traffic after a preset time period.

[0085] Specifically, in this embodiment, the detection module 4 is connected to the data analysis module 2 for real-time monitoring of the equipment load of each communication node in the network. The module detects the equipment load of each node in real time by acquiring the equipment operation status of each node (such as CPU utilization, memory occupancy and other indicators) in real time, ensuring that the data analysis module 2 can accurately grasp the equipment status of each node, and providing an important basis for subsequent traffic prediction and resource adjustment. The calculation module 5 is connected to the communication acquisition module 1 and the data analysis module 2, and its main function is to calculate the congestion index of the communication network. The calculation module 5 takes the bandwidth utilization, packet loss rate, delay and data packet size as inputs, and substitutes them into the comprehensive congestion calculation formula to obtain the communication congestion index of each node. The congestion index is an important indicator reflecting the load and communication quality of network nodes, and provides decision support for traffic prediction and dynamic resource allocation. The storage module 6 is connected to the data analysis module 2 for storing and managing the historical training data of each communication node. The historical training data includes historical load, historical congestion index and historical traffic data, which are used to train the traffic prediction model and help analyze the long-term operation trend of the node. Through the storage module 6, the system can continuously optimize the flow prediction based on historical data and improve the accuracy of the prediction.

[0086] Preferably, the comprehensive congestion calculation formula is configured as:

[0087]

[0088] Among them, C iUsed to represent the communication congestion index, U b (t) is used to represent the bandwidth utilization at time t, L p (t) is used to represent the packet loss rate at time t, L t (t) is used to represent the delay at time t, S p (i) is used to represent the size of the i-th data packet, T is used to represent the time window of calculation, α, β, γ, δ are used to represent the constants that control the influence of each parameter on the communication congestion index, and N is used to represent the total number of data packets.

[0089] Specifically, in this embodiment, the bandwidth utilization is obtained through SNMP (Simple Network Management Protocol) or a traffic monitoring tool (such as NetFlow, sFlow, etc.). The packet loss rate is obtained through an ICMP echo request (ping) or a traffic counting function of a network device. The delay is obtained by measuring the round-trip time from the source to the destination of the network using a ping command or a delay measurement tool based on a protocol (such as ICMP, TCP). The packet size can be obtained through a protocol analysis tool (such as Wireshark) or through the traffic analysis function of the device. In the data preprocessing stage, the collected raw data will be converted into a form suitable for formula calculation: the data of bandwidth utilization, packet loss rate, delay and packet size are normalized or standardized so as to have a consistent scale in the calculation. For indicators with time variations such as packet loss rate and delay, a sliding window (for example, every 5 seconds) is used to calculate and store the network status at the current moment. Then enter the comprehensive congestion calculation formula for calculation. Bandwidth utilization is usually in the range of [0, 1], packet loss rate is usually in the range of [0, 1], and delay, in milliseconds, is usually a positive number. The size of each data packet is in bytes. The constants selected in this embodiment are: α = 1.2: bandwidth utilization has a greater impact on congestion, β = 2.5: packet loss rate has a stronger impact on the congestion index, γ = 0.05: delay has a slight negative impact on the congestion index, δ = 0.8: packet size has a smaller impact on congestion, but still has a certain impact. Through simulation or actual network environment testing, the network congestion index can be calculated and output in real time. The results show that when the bandwidth utilization exceeds 90%, the packet loss rate exceeds 5%, and the delay exceeds 100ms, the communication congestion index rises sharply, prompting the network administrator to intervene.

[0090] In a specific embodiment, in a network environment with high bandwidth utilization and high packet loss rate, the comprehensive congestion index may reach 0.85, while in a normal network environment, the congestion index is below 0.3.

[0091] Preferably, Figure 3 As shown, the comprehensive processing unit 24 includes:

[0092] The fusion subunit 241 is used to fuse the network predicted traffic, network communication habits and the data fusion model preset according to each classification data to obtain multi-source fusion data;

[0093] The analysis subunit 242 is connected to the fusion subunit 241 and is used to analyze the multi-source fusion data to obtain comprehensive analysis data.

[0094] Specifically, in this embodiment, the fusion subunit 241 receives multiple input data sources from the data analysis module 2, including network predicted traffic, network communication habits, and classified data obtained based on traffic classification analysis. It uses a preset data fusion model to perform multi-source data fusion on these data, and uniformly processes heterogeneous data from different sources to obtain a comprehensive multi-source fusion data. The process of multi-source data fusion enables the system to comprehensively consider the relevance and complementarity of different data sources, thereby improving the accuracy and comprehensiveness of the analysis results. The analysis subunit 242 is connected to the fusion subunit 241 and is responsible for analyzing the multi-source fusion data after fusion processing. Through multi-dimensional analysis, the analysis subunit 242 extracts comprehensive analysis data, which provides a basis for subsequent network resource scheduling and dynamic adjustment. The comprehensive analysis data may include network load prediction, user demand analysis, network bottleneck identification, etc., to provide strong support for the decision-making of the communication regulation module 3. In a specific application scenario: it is applied to a high-traffic online video platform, which needs to process the viewing data of millions of users in real time and dynamically adjust according to different user needs and network conditions.

[0095] Multi-source data fusion: The fusion subunit 241 receives data from network traffic forecast, user behavior analysis, and traffic classification. For example, network traffic forecast indicates that the platform will experience peak traffic during certain periods, user behavior analysis reveals the viewing preferences of different users during peak hours, and traffic classification results distinguish different types of traffic such as video traffic and real-time interactive traffic. The fusion subunit 241 fuses this information to generate multi-source fusion data, helping the system to fully understand the current network status and user needs.

[0096] Comprehensive analysis: The analysis subunit 242 performs in-depth analysis on the multi-source fusion data and extracts key indicators from it. For example, the analysis subunit 242 may find that the number of viewers of certain video content surges in a specific time period, or that certain user groups prefer to play in high definition at a specific time. Based on these comprehensive analysis data, the system can predict future network loads and take measures to optimize resources.

[0097] Dynamic adjustment: The communication adjustment module 3 dynamically adjusts network resources based on the comprehensive analysis data provided by the analysis subunit 242. For example, when the system foresees that network congestion will occur during a certain period of time, the communication adjustment module 3 will allocate more bandwidth to high-priority traffic (such as high-definition video streaming) in advance, and direct part of the traffic to nodes with lower network loads through load balancing strategies to avoid network bottlenecks.

[0098] Optimize user experience: Through the fusion analysis of multi-source data, the system of the present invention can respond to network changes and user needs in real time, ensuring that users can still get a smooth viewing experience during the peak period of the platform. The system can perform intelligent optimization based on user viewing habits, improve video playback quality, reduce delays and freezes, and ensure efficient use of platform resources.

[0099] Preferably, the formula configuration of the data fusion model is:

[0100]

[0101] Among them, M f Used to represent multi-source fusion data, N f (t) is used to represent the predicted network traffic at time t, C f (t) is used to represent the traffic classification data at time t, R i It is used to represent the signal strength of the data source at the communication node. α2, β2, γ2, δ2, η2, and κ2 are used to represent the flow prediction influence constant, the communication habit influence constant, the flow classification influence constant, the time attenuation influence constant, and the signal strength influence constant, respectively. c (i)) is used to indicate the volatility of online communication habits.

[0102] Specifically, in this embodiment, the parameters are selected as follows: α2=2, β2=0.5, γ2=1.2, δ2=0.1, η2=1.5, κ2=0.8;

[0103] When N f (t) = 100 Mbps, H c (t) = 0.8, C f When (t) = 0.4, then M f (t)=100 2 ·(1+e -0 . 5·0.8 )·e -1.2·0.4 ·e -0.1·|t-12| , when time t = 12, assuming M = 3, R1 = 0.9, R2 = 1.1, R3 = 1.2, the final multi-source fusion data is the sum of the integral part and the signal source weighted part:

[0104]

[0105] Finally, the fused network traffic prediction data M is calculated. f =320.75Mbps.

[0106] Embodiment 2 is the second embodiment of the present invention. Different from the previous embodiment, this embodiment provides an optimization subunit 243, which can optimize and correct the communication habit influence constant and the traffic prediction influence constant according to the external environment data and the network security event data, respectively, so as to improve the calculation accuracy of the multi-source fusion data, wherein the detection module 4 is also used to detect multiple external environment data and multiple network security event data at each communication node in real time;

[0107] The comprehensive processing unit 24 also includes an optimization subunit 243, connected to the fusion subunit 241, for inputting each external environment data and the communication habit influence constant into a preset first constant optimization formula to obtain the communication habit optimization influence constant, and inputting each network security event data and the flow prediction influence constant into a preset second constant optimization formula to obtain the flow prediction optimization influence constant;

[0108] The data fusion model is updated based on the communication habit optimization influence constant and the traffic prediction optimization influence constant;

[0109] External environmental data include temperature change, terrain interference, environmental magnetic field, and environmental noise;

[0110] Cybersecurity incident data includes password cracking data and virus infection data.

[0111] Working principle of embodiment 2:

[0112] Changes in the external environment may cause performance degradation of communication equipment and unstable signal propagation, which in turn affects the user's communication habits. By introducing the external environment data into the first constant optimization formula, the constant affecting communication habits is optimized to obtain the communication habit optimization influence constant, which filters out the influence of the external environment on the communication habit influence constant and improves the calculation accuracy of multi-source fusion data. Network security incidents will have a significant impact on the use of traffic in the network communication process. For example, when a network communication node is attacked by a relatively serious network security incident, the traffic used by users may drop sharply in the short term, which will seriously interfere with the prediction of traffic data. By introducing the network security event input into the second constant optimization formula, the constant affecting traffic prediction is corrected to obtain the traffic prediction optimization influence constant, which filters out the influence of network security time on the traffic prediction constant, and thus improves the calculation accuracy of multi-source fusion data.

[0113] Preferably, the first constant optimization formula is configured as:

[0114]

[0115] Among them, β2′ is used to represent the influence constant of communication habit optimization, ΔT is used to represent the temperature change, φ is used to represent the influence weight of the temperature change, G d is used to represent the amount of terrain disturbance, χ is used to represent the influence weight of the amount of terrain disturbance, and B e It is used to represent the environmental magnetic field, ψ is used to represent the influence weight of the environmental magnetic field, N e It is used to represent environmental noise, and ∈ is used to represent the impact weight of environmental noise;

[0116] Specifically, in this embodiment, if the communication habit optimization influence constant is close to 0, it means that the influence of environmental factors is very strong, and the communication habit influence constant has been greatly optimized; if the communication habit optimization influence constant is close to the communication habit influence constant, it means that the influence of environmental factors is small, and the optimized constant is basically the same as the initial value. The above-mentioned first constant optimization formula uses an exponential decay function to describe the influence of environmental factors. The weights of various environmental factors can be flexibly adjusted according to actual conditions, so that the formula can adapt to different network environments and communication requirements, and the optimization effect can be fine-tuned for specific situations. The formula can take into account the impact of various factors such as temperature changes, terrain interference, environmental magnetic fields, noise, etc. on communication habits, ensuring that the communication system can operate stably in different environments, especially in environments with complex interference, and has strong environmental adaptability.

[0117] The second constant optimization formula is configured as:

[0118]

[0119] Among them, α2′ is used to represent the flow prediction optimization influence constant, P k Used to represent password cracking data, T k Used to represent the time delay associated with password cracking, a k 、b k They are used to represent the first cracking related constant and the second cracking related constant related to password cracking, V j Used to represent virus infection data, E j Used to represent environmental factors data related to viral infection, c j d j They are respectively used to represent a first infection-related constant and a second infection-related constant related to virus infection, n is used to represent the total amount of password cracking data, and m is used to represent the total amount of virus-infected data.

[0120] Specifically, in this embodiment, the first infection-related constant and the second infection-related constant respectively represent the intensity of the impact of virus infection data on traffic prediction and the regulating factor of the impact of environmental factors on traffic prediction. If the traffic prediction optimization impact constant is close to 0, it indicates that the impact of security events (such as password cracking, network intrusion, etc.) on traffic prediction is very large, and the system's traffic prediction ability is greatly suppressed; if the traffic prediction optimization impact constant is close to the traffic prediction impact constant, it indicates that the impact of these security events is small, the traffic prediction constant is close to the initial value, and the system can operate normally. The above formula comprehensively considers the impact of security events on traffic prediction by combining multiple data sources such as password cracking, network intrusion, leakage data, and virus infection data. It is not limited to the consideration of a single factor, so that the above formula can be widely used in traffic prediction in complex network environments. The constants in the above formula can be optimized and adjusted according to the actual network environment, historical data or experimental results, so as to accurately adapt to different communication scenarios.

[0121] A communication network optimization control method based on multi-dimensional data driving is applied to the above-mentioned communication network optimization control system based on multi-dimensional data driving, such as Figure 2 As shown, including:

[0122] Step S1, the communication collection module 1 collects multiple network traffic data, user behavior data and exchange mode data at each communication node in the communication network in real time;

[0123] Step S2, the first analysis unit 21 inputs each network traffic data into the pre-selected and trained traffic prediction model to predict the network traffic after a preset time period, the second analysis unit 22 analyzes each user behavior data to obtain the network communication habits of each user, and the third analysis unit 23 classifies each exchange mode data according to a preset traffic classification algorithm to obtain multiple classification data;

[0124] Step S3, the comprehensive processing unit 24 obtains a comprehensive analysis result according to the network predicted traffic, network communication habits and various classified data processing;

[0125] Step S4, the communication adjustment module 3 dynamically adjusts the bandwidth resources, routing selection and traffic diversion at each communication node according to the comprehensive analysis results.

[0126] The above are only preferred embodiments of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A communication network optimization control system based on multi-dimensional data drive, characterized in that: include: A communication collection module (1) is used to collect multiple network flow data, user behavior data and exchange mode data at each communication node in the communication network in real time; The data analysis module (2) is connected to the communication acquisition module (1) and comprises: A first analysis unit (21) is used to input each of the network traffic data into a pre-selected and trained traffic prediction model to predict the network traffic after a preset time period; A second analysis unit (22) is used to analyze the user behavior data and obtain the network communication habits of each user; A third analysis unit (23) is used to classify each of the exchange mode data according to a preset traffic classification algorithm to obtain a plurality of classification data; A comprehensive processing unit (24), connected to the first analysis unit (21), the second analysis unit (22) and the third analysis unit (23), respectively, for obtaining a comprehensive analysis result according to the network predicted traffic, the network communication habits and the classification data processing; The communication adjustment module (3) is connected to the data analysis module (2) and is used to dynamically adjust the bandwidth resources, route selection and traffic diversion at each communication node according to the comprehensive analysis results.

2. The communication network optimization control system based on multi-dimensional data drive according to claim 1 is characterized in that: The network traffic data includes bandwidth utilization, packet loss rate, latency, packet size and traffic distribution.

3. The communication network optimization control system based on multi-dimensional data drive according to claim 2 is characterized in that: Also includes: A detection module (4), connected to the data analysis module (2), for detecting the device load at each of the communication nodes in real time; A calculation module (5), connected to the communication acquisition module (1) and the data analysis module (2), for inputting the bandwidth utilization, the packet loss rate, the delay and the data packet size into a preset comprehensive congestion calculation formula to calculate a communication congestion index; A storage module (6), connected to the data analysis module (2), and used to store a plurality of historical training data of each of the communication nodes, wherein the historical training data includes historical load, historical congestion index and historical traffic data; The first analysis unit (21) comprises: A training subunit (211) is used to introduce an initial traffic model, and use each of the historical loads, each of the historical congestion indexes and each of the historical traffic data as input, and use each of the historical traffic data after the next time period as output, to retrain the initial traffic model and obtain the traffic prediction model; The prediction subunit (212) is connected to the training subunit (211) and is used to input the device load, the network traffic data and the communication congestion index at the current moment into the traffic prediction model to predict the network predicted traffic after a preset time period.

4. The communication network optimization control system based on multi-dimensional data drive according to claim 2 is characterized in that: The comprehensive congestion calculation formula is configured as: Among them, C i Used to represent the communication congestion index, U b (t) is used to represent the bandwidth utilization at time t, L p (t) is used to represent the packet loss rate at time t, L t (t) is used to represent the delay at time t, S p (i) is used to represent the size of the i-th data packet, T is used to represent the time window of calculation, α, β, γ, δ are respectively used to represent constants that control the influence of various parameters on the communication congestion index, and N is used to represent the total number of data packets.

5. The communication network optimization control system based on multi-dimensional data drive according to claim 3 is characterized in that: The comprehensive processing unit (24) comprises: A fusion subunit (241) is used to fuse the network predicted traffic, the network communication habits and the classified data into a preset data fusion model to obtain multi-source fusion data; The analysis subunit (242) is connected to the fusion subunit (241) and is used to analyze the multi-source fusion data to obtain comprehensive analysis data.

6. The communication network optimization control system based on multi-dimensional data drive according to claim 5 is characterized in that: The formula configuration of the data fusion model is: Among them, M f It is used to represent the multi-source fusion data, N f (t) is used to represent the predicted network traffic at time t, C f (t) is used to represent the traffic classification data at time t, R i It is used to represent the signal strength of the data source at the communication node. α2, β2, γ2, δ2, η2, and k2 are used to represent the flow prediction influence constant, the communication habit influence constant, the flow classification influence constant, the time attenuation influence constant, and the signal strength influence constant, respectively. c (i)) is used to indicate the volatility of the network communication habits.

7. The communication network optimization control system based on multi-dimensional data drive according to claim 6 is characterized in that: The detection module (4) is also used to detect multiple external environment data and multiple network security event data at each of the communication nodes in real time; The integrated processing unit (24) further comprises an optimization subunit (243), connected to the fusion subunit (241), for inputting each of the external environment data and the communication habit influence constant into a preset first constant optimization formula to obtain a communication habit optimization influence constant, and for inputting each of the network security event data and the traffic prediction influence constant into a preset second constant optimization formula to obtain a traffic prediction optimization influence constant; The data fusion model is updated based on the communication habit optimization influence constant and the traffic prediction optimization influence constant.

8. The communication network optimization control system based on multi-dimensional data drive according to claim 7 is characterized in that: The external environment data includes temperature change, terrain interference, environmental magnetic field, and environmental noise; The network security event data includes password cracking data and virus infection data.

9. The communication network optimization control system based on multi-dimensional data drive according to claim 8, characterized in that: The first constant optimization formula is configured as: Among them, β2′ is used to represent the communication habit optimization influence constant, ΔT is used to represent the temperature change, φ is used to represent the influence weight of the temperature change, G d is used to represent the terrain interference amount, χ is used to represent the influence weight of the terrain interference amount, B e is used to represent the environmental magnetic field, ψ is used to represent the influence weight of the environmental magnetic field, N e is used to represent the environmental noise, and ω is used to represent the influence weight of the environmental noise; The second constant optimization formula is configured as: Among them, α2′ is used to represent the flow prediction optimization influence constant, P k Used to represent the password cracking data, T k Used to represent the time delay associated with password cracking, a k , D k They are used to represent the first cracking related constant and the second cracking related constant related to password cracking, V j Used to represent the virus infection data, E j Used to represent environmental factors data related to viral infection, c j ,d j They are respectively used to represent a first infection-related constant and a second infection-related constant related to virus infection, n is used to represent the total amount of the password cracking data, and m is used to represent the total amount of the virus infection data.

10. A communication network optimization control method based on multidimensional data drive, applied to the communication network optimization control system based on multidimensional data drive according to any one of claims 1 to 9, characterized in that: include: Step S1, the communication collection module (1) collects multiple network flow data, user behavior data and exchange mode data at each communication node in the communication network in real time; Step S2, the first analysis unit (21) inputs each of the network traffic data into a pre-selected and trained traffic prediction model to predict the network traffic after a preset time period, the second analysis unit (22) analyzes each of the user behavior data to obtain the network communication habits of each user, and the third analysis unit (23) classifies each of the exchange mode data according to a preset traffic classification algorithm to obtain a plurality of classification data; Step S3, the comprehensive processing unit (24) obtains a comprehensive analysis result according to the network predicted traffic, the network communication habits and the classification data; Step S4, the communication adjustment module (3) dynamically adjusts the bandwidth resources, route selection and traffic diversion at each of the communication nodes according to the comprehensive analysis results.

Images