SEANET-based nuclear emergency communication rapid recovery method

Abstract

The invention discloses a SEANET-based core emergency communication rapid recovery method, and relates to the field of core emergency communication, and the method comprises the steps: obtaining network equipment, links, bandwidth configuration information and log historical data of core communication based on a SEANET network; setting a resource priority calling level according to priority use conditions of network equipment, links and bandwidths; on the basis of a machine learning algorithm model, whether the network devices, links and bandwidth configurations are abnormal or not is judged; according to a judgment result of the machine learning algorithm model, screening out a suitable kernel communication rapid recovery method based on a deep reinforcement learning algorithm model; and based on an ICMP protocol, daily nuclear emergency communication and communication quality after rapid recovery are detected through a machine learning algorithm model. The method has the advantages that a suitable nuclear communication quick recovery method can be quickly screened out, and the nuclear emergency communication quick recovery efficiency and the emergency response capability are improved.

Description

Technical Field

[0001] The invention relates to the field of nuclear emergency communications, and in particular to a nuclear emergency communications rapid recovery method based on SEANET. Background Art

[0002] During a sudden nuclear accident, network facilities may be damaged, resulting in untimely communication, delayed fault information investigation and processing, and the degree of nuclear pollution damage will be deepened. Therefore, it is necessary to attach great importance to nuclear safety, strengthen the supervision and maintenance of nuclear facilities, and ensure that various network devices can communicate in time after a sudden nuclear accident. With the development of SEANET network technology, the monitoring and data collection capabilities of network devices using SEANET network technology can effectively achieve rapid communication recovery and command and dispatch in nuclear emergency situations. By combining SEANET network technology with intelligent algorithms, it is of great significance to ensure the rapid recovery of nuclear emergency communications.

[0003] Existing technologies rarely involve the rapid recovery of nuclear emergency communications, and mainly focus on monitoring nuclear radiation concentrations through monitoring equipment and devices for collecting nuclear radiation concentrations. There is a lack of risk prevention for the damage of communication equipment caused by sudden nuclear accidents, which leads to untimely communication between various network devices, resulting in delayed nuclear fault information investigation and emergency response actions, making it difficult to quickly contain nuclear pollution and causing greater damage and losses. Summary of the invention

[0004] In order to solve the above technical problems, a method for rapid recovery of nuclear emergency communications based on SEANET is provided. The technical solution solves the problem proposed in the above background technology that there is a lack of risk prevention against the damage of communication equipment caused by sudden nuclear accidents, resulting in untimely communication between various network devices, thereby causing delays in nuclear fault information investigation and emergency response behavior, making it difficult to quickly contain nuclear pollution and causing greater damage and losses.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is: A nuclear emergency communication rapid recovery method based on SEANET, comprising: Based on the SEANET network, obtain the network equipment, links, bandwidth configuration information and log history data of the core communication; Set resource priority levels based on the priority usage of network devices, links, and bandwidth; Based on the machine learning algorithm model, determine whether there are abnormalities in each network device, link, and bandwidth configuration; According to the judgment results of the machine learning algorithm model, a suitable nuclear communication rapid recovery method is screened out based on the deep reinforcement learning algorithm model; Based on the ICMP protocol, the quality of daily nuclear emergency communications and communications after rapid recovery is detected through a machine learning algorithm model.

[0006] Preferably, the acquisition of network equipment, links, bandwidth configuration information and log history data of core communication based on the SEANET network specifically includes: Based on the SEANET network, obtain the network equipment, link, and bandwidth configuration distribution information in daily nuclear communication; Based on the SEANET network, the log data of each network device is collected and analyzed in real time to record the operating status, configuration changes, and error information of the device; According to the network equipment, link, bandwidth configuration information of the core communication and the historical data of the network equipment log, a core communication network equipment operation configuration relationship table is established.

[0007] Preferably, the setting of resource priority call levels according to the priority usage of network devices, links, and bandwidth specifically includes: According to the core communication network equipment operation configuration relationship table, network equipment, links, and bandwidth are layered, and network equipment, links, and bandwidths with the same function or containing the same function are divided into the same layer. Among them, network equipment, links, and bandwidths with multiple functions can be cross-linked in the layering, and can exist in multiple layers; Set the resource priority call level based on the priority usage of each layer of network equipment, links, and bandwidth after stratification.

[0008] Preferably, judging whether there is an abnormality in each network device, link, or bandwidth configuration based on a machine learning algorithm model specifically includes: According to the historical data of the nuclear communication network equipment operation configuration relationship table and ICMP protocol detection log, a nuclear communication training sample set and a test set are established. Among them, ICMP protocol detection includes: connectivity detection, delay evaluation, packet loss rate detection and jitter analysis; According to the connectivity detection, delay assessment, packet loss rate detection and jitter analysis detection results of the core communication training sample set, the results are quantified, digitized and the data is standardized; Based on the processed data, a linear regression equation for the operating status of network equipment, links, and bandwidth is established; According to the gradient descent algorithm, the update equations of network equipment, link and bandwidth operation status parameters are established; Based on the calculation results of the machine learning algorithm model, determine whether there are any abnormalities in the operating status of each network device, link, and bandwidth; The linear regression equation for the network equipment, link, and bandwidth operation status is: In the formula, The operating status of each network device, link, and bandwidth. is the error constant term, is the data after connectivity detection processing, To delay the evaluation of processed data, This is the data after packet loss rate detection processing. Detect processed data for jitter analysis, are the parameters of the data after connectivity detection processing, To delay the evaluation of parameters on processed data, is the parameter of the data after packet loss rate detection processing, Detect parameters of processed data for jitter analysis; The network equipment, link, and bandwidth operating state parameter update equations are: In the formula, For the The updated parameters under each test result are: For the The parameters under each test result are: is the number of sample groups in the core communication training sample set, is the learning rate, For the The test results are as follows The operating status results of network devices, links, and bandwidth.

[0009] Preferably, the method of selecting a suitable nuclear communication rapid recovery method based on a deep reinforcement learning algorithm model according to the judgment result of the machine learning algorithm model specifically includes: Determine the state transition probability matrix based on the resource priority call level of each layer and the judgment result of the machine learning algorithm model; Based on the simulation experiment of nuclear communication history data and channel link for rapid nuclear communication recovery, the reward value of channel link for rapid nuclear communication recovery is set based on big data analysis; According to the state transition probability matrix and the channel link reward value of fast recovery nuclear communication, a deep reinforcement learning model for fast recovery nuclear communication is established based on the Bellman equation; Based on the training and learning results of the deep reinforcement learning model for rapid recovery of nuclear communications, channel links are autonomously established for sudden nuclear emergency communications to rapidly restore nuclear communications; In the autonomous selection process of the deep reinforcement learning model for fast recovery of nuclear communication, it is determined whether all state transition probability values ​​in the same layer are 0. If so, it means that it is difficult for the model to autonomously recover nuclear communication and a warning is issued. If not, it means that the model can achieve autonomous recovery of nuclear communication. The state transition probability is calculated as follows: In the formula, Select the state transition probability value of a certain network device, link, and bandwidth for core communication, The operating status values ​​of each network device, link, and bandwidth calculated by the machine learning algorithm model. The call probability value is determined based on big data according to the priority call level of each layer of resources; The Bellman equation is expressed as: In the formula, To select The state value function value under the detection result state, For the Select the first The probability of a level of resources, For the Test result status and select The probability of feedback reward value under level resources, For the Test result status and select The feedback reward value under the resource level, is the reward decay factor, For the Test result status and select The next state value function value under level resources, To select The next state value function value under the detection result state.

[0010] Preferably, the detection of the quality of daily nuclear emergency communication and communication after rapid recovery by a machine learning algorithm model based on the ICMP protocol specifically includes: Based on the ICMP protocol, the daily detection data of the ICMP protocol is obtained through a machine learning algorithm model to determine whether there are abnormalities in the daily operation of each network device, link, and bandwidth configuration; After an emergency occurs, the ICMP protocol is used to actively send Ping commands to obtain the operating information of each network device, link, and bandwidth. The machine learning algorithm model is used to determine whether there are any abnormalities in the operation of each network device, link, and bandwidth, and the deep reinforcement learning model is used to restore nuclear emergency communications. After the nuclear emergency communication is restored, each network device continues to actively send Ping commands, and uses the machine learning algorithm model to determine whether each network device, link, and bandwidth are operating normally; if so, it means that the nuclear emergency communication has been restored to normal; if not, it means that the nuclear emergency communication has been restored abnormally, and a warning is issued and abnormal information is recorded, and the deep reinforcement learning model is refreshed and repaired again.

[0011] Compared with the prior art, the present invention has the following beneficial effects: According to the monitoring and data collection capabilities of network equipment of SEANET network technology, historical data of nuclear communications are obtained. By processing the obtained data, the machine learning algorithm model of linear regression is used to model and train the previous data. When the detection results of the ICMP protocol are received, it is judged whether there are abnormalities in the operation of each network device, link, and bandwidth. Secondly, according to the judgment results of the operation of each network device, link, and bandwidth, the probability values ​​of the deep reinforcement learning model taking different actions under different states are determined. By training the reward values ​​after completing a set of state selections, suitable nuclear communication rapid recovery methods are effectively screened out, which effectively improves the risk resistance of nuclear communications and improves the rapid recovery efficiency and emergency response capabilities of nuclear emergency communications. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A flow chart of a method for rapid recovery of nuclear emergency communications based on SEANET according to the present invention; Figure 2 Based on the machine learning algorithm model of the present invention, it is a flowchart for judging whether there are abnormalities in each network device, link, and bandwidth configuration; Figure 3 A flowchart of a method for rapidly recovering nuclear communications that is suitable for screening out the results of a machine learning algorithm model based on a deep reinforcement learning algorithm model according to the present invention; Figure 4 This is a flow chart of the present invention, which detects the quality of daily nuclear emergency communication and communication after rapid recovery through a machine learning algorithm model based on the ICMP protocol; Figure 5 It is a schematic diagram of the structure of the electronic device of the present invention; Figure 6 It is a schematic diagram of the computer-readable storage medium structure of the present invention. DETAILED DESCRIPTION

[0013] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0014] Reference Figure 1 As shown, a nuclear emergency communication rapid recovery method based on SEANET includes: Based on the SEANET network, obtain the network equipment, links, bandwidth configuration information and log history data of the core communication; Set resource priority levels based on the priority usage of network devices, links, and bandwidth; Based on the machine learning algorithm model, determine whether there are abnormalities in each network device, link, and bandwidth configuration; According to the judgment results of the machine learning algorithm model, a suitable nuclear communication rapid recovery method is screened out based on the deep reinforcement learning algorithm model; Based on the ICMP protocol, the quality of daily nuclear emergency communications and communications after rapid recovery is detected through a machine learning algorithm model.

[0015] It can be explained that this scheme is based on the monitoring capability and data collection capability of the network equipment of the SEANET network technology to obtain the historical data of nuclear communication, and by processing the obtained data, using the machine learning algorithm model of linear regression, and modeling and training the previous data, when receiving the detection results of the ICMP protocol, it is possible to judge whether there are abnormalities in the operation of each network device, link, and bandwidth. Secondly, according to the judgment results of the operation of each network device, link, and bandwidth, the probability values ​​of the deep reinforcement learning model taking different actions under different states are determined, and by training the reward value after completing a set of state selections, a suitable nuclear communication rapid recovery method is effectively screened out.

[0016] Reference Figure 2 As shown, the method of judging whether there are abnormalities in each network device, link, and bandwidth configuration based on the machine learning algorithm model specifically includes: According to the historical data of the nuclear communication network equipment operation configuration relationship table and ICMP protocol detection log, a nuclear communication training sample set and a test set are established. Among them, ICMP protocol detection includes: connectivity detection, delay evaluation, packet loss rate detection and jitter analysis; According to the connectivity detection, delay assessment, packet loss rate detection and jitter analysis detection results of the core communication training sample set, the results are quantified, digitized and the data is standardized; Based on the processed data, a linear regression equation for the operating status of network equipment, links, and bandwidth is established; According to the gradient descent algorithm, the update equations of network equipment, link and bandwidth operation status parameters are established; Based on the calculation results of the machine learning algorithm model, determine whether there are any abnormalities in the operating status of each network device, link, and bandwidth; The linear regression equation for the network equipment, link, and bandwidth operation status is: In the formula, The operating status of each network device, link, and bandwidth. is the error constant term, is the data after connectivity detection processing, To delay the evaluation of processed data, This is the data after packet loss rate detection processing. Detect processed data for jitter analysis, are the parameters of the data after connectivity detection processing, To delay the evaluation of parameters on processed data, is the parameter of the data after packet loss rate detection processing, Detect parameters of processed data for jitter analysis; The network equipment, link, and bandwidth operating state parameter update equations are: In the formula, For the The updated parameters under each test result are: For the The parameters under each test result are: is the number of sample groups in the core communication training sample set, is the learning rate, For the The test results are as follows The operating status results of network devices, links, and bandwidth.

[0017] It can be explained that based on SEANET network technology, ICMP protocol detection can accurately obtain the detection results of various factors for judging the operating status of network devices, links, and bandwidth. By analyzing the weights of these factors affecting the operating status of network devices, links, and bandwidth, that is, parameter values, a linear relationship between the influencing factors and the operating status of network devices, links, and bandwidth is constructed, and the gap between the calculated value and the actual value is reduced by gradient descent. By receiving various detection results detected by the ICMP protocol, it is possible to judge whether the operating status of network devices, links, and bandwidth at each layer is abnormal. Among them, there are only two situations for connectivity detection results, either connected or not connected, which can be represented by 0 or 1 in the quantification and digitization process. Therefore, when establishing a linear regression equation, it is necessary to discuss the impact of other detection results in the connected state.

[0018] Reference Figure 3 As shown, according to the judgment result of the machine learning algorithm model, the method for selecting a suitable nuclear communication rapid recovery method based on the deep reinforcement learning algorithm model specifically includes: Determine the state transition probability matrix based on the resource priority call level of each layer and the judgment result of the machine learning algorithm model; Based on the simulation experiment of nuclear communication history data and channel link for rapid nuclear communication recovery, the reward value of channel link for rapid nuclear communication recovery is set based on big data analysis; According to the state transition probability matrix and the channel link reward value of fast recovery nuclear communication, a deep reinforcement learning model for fast recovery nuclear communication is established based on the Bellman equation; Based on the training and learning results of the deep reinforcement learning model for rapid recovery of nuclear communications, channel links are autonomously established for sudden nuclear emergency communications to rapidly restore nuclear communications; In the autonomous selection process of the deep reinforcement learning model for fast recovery of nuclear communication, it is determined whether all state transition probability values ​​in the same layer are 0. If so, it means that it is difficult for the model to autonomously recover nuclear communication and a warning is issued. If not, it means that the model can achieve autonomous recovery of nuclear communication. The state transition probability is calculated as follows: In the formula, Select the state transition probability value of a certain network device, link, and bandwidth for core communication, The operating status values ​​of each network device, link, and bandwidth calculated by the machine learning algorithm model. The call probability value is determined based on big data according to the priority call level of each layer of resources; The Bellman equation is expressed as: In the formula, To select The state value function value under the detection result state, For the Select the first The probability of a level of resources, For the Test result status and select The probability of feedback reward value under level resources, For the Test result status and select The feedback reward value under the resource level, is the reward decay factor, For the Test result status and select The next state value function value under level resources, To select The next state value function value under the detection result state.

[0019] It can be explained that in the deep reinforcement learning algorithm model, it is important to determine the state transition probability and reward value. This scheme uses the machine learning algorithm model to judge the operating status of network devices, links, and bandwidths at different levels of each layer. If it is judged that the operating status of a certain network device, link, or bandwidth is abnormal, it can be said that the probability of the resource being selected in the state transition process is 0. Otherwise, its probability can be determined according to the priority level of the resource. Therefore, the state transition probability matrix is ​​determined by the priority call level of each layer of resources and the judgment result of the machine learning algorithm model. Secondly, for the reward value of the deep reinforcement learning algorithm model, according to the expectation of the final result, based on the simulation experiment of the nuclear communication historical data and the channel link of the rapid recovery of nuclear communication, the reward value of the channel link of the rapid recovery of nuclear communication is set. In particular, the deep reinforcement learning algorithm model is a continuous learning process. By setting different reward values ​​in the state transition process of different environments, the optimal strategy selection is obtained. With the updating and learning of the model, the model can accept more strategies and cover a wider range, which is suitable for the complex and changeable nuclear emergency communication environment.

[0020] Reference Figure 4 As shown, the detection of the quality of daily nuclear emergency communication and communication after rapid recovery based on the ICMP protocol through a machine learning algorithm model specifically includes: Based on the ICMP protocol, the daily detection data of the ICMP protocol is obtained through a machine learning algorithm model to determine whether there are abnormalities in the daily operation of each network device, link, and bandwidth configuration; After an emergency occurs, the ICMP protocol is used to actively send Ping commands to obtain the operating information of each network device, link, and bandwidth. The machine learning algorithm model is used to determine whether there are any abnormalities in the operation of each network device, link, and bandwidth, and the deep reinforcement learning model is used to restore nuclear emergency communications. After the nuclear emergency communication is restored, each network device continues to actively send Ping commands, and uses the machine learning algorithm model to determine whether each network device, link, and bandwidth are operating normally; if so, it means that the nuclear emergency communication has been restored to normal; if not, it means that the nuclear emergency communication has been restored abnormally, and a warning is issued and abnormal information is recorded, and the deep reinforcement learning model is refreshed and repaired again.

[0021] It can be explained that this scheme uses the ICMP protocol to actively send Ping commands to each network device to obtain the detection results of various influencing factors, and uses the machine learning algorithm model to judge the operating status of each network device, link, and bandwidth. According to the judgment results, the nuclear emergency communication is restored based on the deep reinforcement learning model. Therefore, this scheme effectively improves the risk resistance of nuclear communication and improves the rapid recovery efficiency and emergency response capability of nuclear emergency communication through the establishment of machine learning algorithm models and deep reinforcement learning models. Among them, the detection of daily nuclear emergency communication and communication quality after rapid recovery is divided into three steps, specifically: S1. Send Ping commands regularly to determine the operating status of each network device, link, and bandwidth based on the machine learning algorithm model. S2. After a sudden nuclear accident, send a Ping command to each network device to detect the operating status of each network device, link, and bandwidth after the nuclear accident, and restore nuclear emergency communications through a deep reinforcement learning model; S3. After the nuclear emergency communication is quickly restored, the Ping command is sent to each network device again, and the machine learning algorithm model is used to determine whether the nuclear emergency communication has returned to normal and remains stable.

[0022] Furthermore, the method according to the embodiment of the present application can also be performed by Figure 5 The electronic device architecture shown in FIG. Figure 5 As shown, the electronic device 500 may include a bus 501, one or more CPUs 502, a read-only memory (ROM) 503, a random access memory (RAM) 504, a communication port 505 connected to a network, an input / output component 506, a hard disk 507, etc. The storage device in the electronic device 500, such as the ROM 503 or the hard disk 507, may store a nuclear emergency communication rapid recovery method based on SEANET provided in the present application. The electronic device 500 may also include a user interface 508. Of course, Figure 5 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 5 One or more components of an electronic device are shown.

[0023] Figure 6 Schematic diagram of a computer-readable storage medium structure provided by an embodiment of the present application. Figure 6As shown, a computer-readable storage medium 600 according to an embodiment of the present application is shown. Computer-readable instructions are stored on the computer-readable storage medium 600. When the computer-readable instructions are executed by the processor, a method for rapid recovery of nuclear emergency communications based on SEANET according to an embodiment of the present application described with reference to the above figures can be executed. The storage medium 600 includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0024] In summary, the advantages of the present invention are: based on the monitoring capability and data collection capability of the network equipment of the SEANET network technology, the historical data of nuclear communication is obtained, and the obtained data is processed and modeled and trained with the machine learning algorithm model of linear regression, so that when the detection result of the ICMP protocol is received, it is judged whether there is an abnormality in the operation of each network device, link, and bandwidth. Secondly, according to the judgment results of the operation of each network device, link, and bandwidth, the probability values ​​of the deep reinforcement learning model taking different actions under different states are determined, and by training the reward value after completing a set of state selections, a suitable nuclear communication rapid recovery method is effectively screened out, which effectively improves the risk resistance of nuclear communication and improves the rapid recovery efficiency and emergency response capability of nuclear emergency communication.

[0025] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A nuclear emergency communication rapid recovery method based on SEANET, characterized in that: include: Based on the SEANET network, obtain the network equipment, links, bandwidth configuration information and log history data of the core communication; Set resource priority levels based on the priority usage of network devices, links, and bandwidth; Based on the machine learning algorithm model, determine whether there are abnormalities in each network device, link, and bandwidth configuration; According to the judgment results of the machine learning algorithm model, a suitable nuclear communication rapid recovery method is screened out based on the deep reinforcement learning algorithm model; Based on the ICMP protocol, the quality of daily nuclear emergency communications and communications after rapid recovery is detected through a machine learning algorithm model.

2. A method for rapid recovery of nuclear emergency communications based on SEANET according to claim 1, characterized in that: The acquisition of network equipment, links, bandwidth configuration information and log history data of core communication based on the SEANET network specifically includes: Based on the SEANET network, obtain the network equipment, link, and bandwidth configuration distribution information in daily nuclear communication; Based on the SEANET network, the log data of each network device is collected and analyzed in real time to record the operating status, configuration changes, and error information of the device; According to the network equipment, link, bandwidth configuration information of the core communication and the historical data of the network equipment log, a core communication network equipment operation configuration relationship table is established.

3. A method for rapid recovery of nuclear emergency communications based on SEANET according to claim 2, characterized in that: The setting of resource priority call levels according to the priority usage of network devices, links, and bandwidth specifically includes: According to the core communication network equipment operation configuration relationship table, network equipment, links, and bandwidth are layered, and network equipment, links, and bandwidths with the same function or containing the same function are divided into the same layer. Among them, network equipment, links, and bandwidths with multiple functions can be cross-linked in the layering, and can exist in multiple layers; Set the resource priority call level based on the priority usage of each layer of network equipment, links, and bandwidth after stratification.

4. A method for rapid recovery of nuclear emergency communications based on SEANET according to claim 3, characterized in that: The method of judging whether there are abnormalities in the configuration of each network device, link, or bandwidth based on the machine learning algorithm model specifically includes: According to the historical data of the nuclear communication network equipment operation configuration relationship table and ICMP protocol detection log, a nuclear communication training sample set and a test set are established. Among them, ICMP protocol detection includes: connectivity detection, delay evaluation, packet loss rate detection and jitter analysis; According to the connectivity detection, delay assessment, packet loss rate detection and jitter analysis detection results of the core communication training sample set, the results are quantified, digitized and the data is standardized; Based on the processed data, a linear regression equation for the operating status of network equipment, links, and bandwidth is established; According to the gradient descent algorithm, the update equations of network equipment, link and bandwidth operation status parameters are established; Based on the calculation results of the machine learning algorithm model, determine whether there are any abnormalities in the operating status of each network device, link, and bandwidth; The linear regression equation for the network equipment, link, and bandwidth operation status is: In the formula, The operating status of each network device, link, and bandwidth. is the error constant term, is the data after connectivity detection processing, To delay the evaluation of processed data, This is the data after packet loss rate detection processing. Detect processed data for jitter analysis, are the parameters of the data after connectivity detection processing, To delay the evaluation of parameters on processed data, is the parameter of the data after packet loss rate detection processing, Detect parameters of processed data for jitter analysis; The network equipment, link, and bandwidth operating status parameter update equations are: In the formula, For the The updated parameters under each test result are: For the The parameters under each test result are: is the number of sample groups in the core communication training sample set, is the learning rate, For the The test results are as follows The operating status results of network devices, links, and bandwidth.

5. A method for rapid recovery of nuclear emergency communications based on SEANET according to claim 4, characterized in that: The method of selecting a suitable nuclear communication rapid recovery method based on the deep reinforcement learning algorithm model according to the judgment result of the machine learning algorithm model specifically includes: Determine the state transition probability matrix based on the resource priority call level of each layer and the judgment result of the machine learning algorithm model; Based on the simulation experiment of nuclear communication history data and channel link for rapid nuclear communication recovery, the reward value of channel link for rapid nuclear communication recovery is set based on big data analysis; According to the state transition probability matrix and the channel link reward value of fast recovery nuclear communication, a deep reinforcement learning model for fast recovery nuclear communication is established based on the Bellman equation; Based on the training and learning results of the deep reinforcement learning model for rapid recovery of nuclear communications, channel links are autonomously established for sudden nuclear emergency communications to rapidly restore nuclear communications; In the autonomous selection process of the deep reinforcement learning model for fast recovery of nuclear communication, it is determined whether all state transition probability values ​​in the same layer are 0. If so, it means that it is difficult for the model to autonomously recover nuclear communication and a warning is issued. If not, it means that the model can achieve autonomous recovery of nuclear communication. The state transition probability is calculated as follows: In the formula, Select the state transition probability value of a certain network device, link, and bandwidth for core communication, The operating status values ​​of each network device, link, and bandwidth calculated by the machine learning algorithm model. The call probability value is determined based on big data according to the priority call level of each layer of resources; The Bellman equation is expressed as: In the formula, To select The state value function value under the detection result state, For the Select the first The probability of a level of resources, For the Test result status and select The probability of feedback reward value under level resources, For the Test result status and select The feedback reward value under the resource level, is the reward decay factor, For the Test result status and select The next state value function value under level resources, To select The next state value function value under the detection result state.

6. A method for rapid recovery of nuclear emergency communications based on SEANET according to claim 5, characterized in that: The detection of the quality of daily nuclear emergency communication and communication after rapid recovery by machine learning algorithm model based on ICMP protocol specifically includes: Based on the ICMP protocol, the daily detection data of the ICMP protocol is obtained through a machine learning algorithm model to determine whether there are abnormalities in the daily operation of each network device, link, and bandwidth configuration; After an emergency occurs, the ICMP protocol is used to actively send Ping commands to obtain the operating information of each network device, link, and bandwidth. The machine learning algorithm model is used to determine whether there are any abnormalities in the operation of each network device, link, and bandwidth, and the deep reinforcement learning model is used to restore nuclear emergency communications. After the nuclear emergency communication is restored, each network device continues to actively send Ping commands, and uses the machine learning algorithm model to determine whether each network device, link, and bandwidth are operating normally; if so, it means that the nuclear emergency communication has been restored to normal; if not, it means that the nuclear emergency communication has been restored abnormally, and a warning is issued and abnormal information is recorded, and the deep reinforcement learning model is refreshed and repaired again.

7. An electronic device, characterized in that: include: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the SEANET-based nuclear emergency communication rapid recovery method as described in any one of claims 1-6.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the SEANET-based nuclear emergency communication rapid recovery method described in any one of claims 1 to 6 is implemented.

Images