An Internet operation and maintenance method optimized based on artificial intelligence algorithms

Abstract

The present invention discloses an Internet operation and maintenance method optimized based on artificial intelligence algorithms, which relates to the field of Internet technologies and includes: mining the implicit connection between Internet operation failures and performance index parameters; performing correlation analysis on two operation failures to generate at least one regular combination of failures; forming a judgment criterion for a suspected regular combination of failures; obtaining at least one suspected regular combination of failures that meets the judgment criterion of the suspected regular combination of failures for the performance index parameters; obtaining a target regular combination of failures; and processing the operation failures in the target regular combination of failures. By mining the implicit connection between Internet operation failures and performance index parameters, obtaining the target regular combination of failures, and processing the operation failures in the target regular combination of failures, it is possible to set maintenance parameters in a targeted manner. At the same time, according to their severity, the maintenance order is set, so as to better meet the operation and maintenance requirements.

Description

Technical Field

[0001] The present invention relates to the field of Internet technologies, and more particularly to an Internet operation and maintenance method optimized based on artificial intelligence algorithms. Background Art

[0002] At present, the development of the Internet presents the characteristics of high integration, intelligence, and ecologicalization, and is widely used in multiple fields such as manufacturing, energy, transportation, and healthcare. Internet operation and maintenance personnel are responsible for ensuring the efficient and secure operation of Internet systems by performing tasks such as network connectivity, data collection and processing, identification and resolution application, platform application optimization, and system security maintenance. Among them, it is particularly important to identify and handle various operating faults in Internet operation and maintenance.

[0003] However, due to the large number of Internet fault categories, the unknown situation of mixed operating faults, and the different degrees of each fault, it poses great challenges for identification and repair. Moreover, during maintenance, it is difficult to sort the maintenance according to the severity level, and it cannot well meet the requirements of operation and maintenance. Summary of the Invention

[0004] To solve the above technical problems, an Internet operation and maintenance method optimized based on artificial intelligence algorithms is provided, and this technical solution solves the problems raised in the above background art.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] An Internet operation and maintenance method optimized based on artificial intelligence algorithms, comprising:

[0007] Using a smart terminal to collect real-time data of at least one performance index parameter of the Internet in real time;

[0008] Determining the data transmission method, configuring the data transmission protocol, establishing a data transmission channel, and performing automated data collection and transmission;

[0009] Obtaining at least one operating fault in the Internet and mining the implicit relationship between the Internet operating fault and the performance index parameter;

[0010] Based on the implicit relationship, performing a correlation analysis on two operating faults, and generating at least one fault regular combination based on the correlation between the two operating faults;

[0011] Forming a judgment criterion for a suspected fault regular combination;

[0012] Obtaining at least one suspected fault regular combination that meets the judgment criterion of the suspected fault regular combination for the performance index parameter;

[0013] Screen at least one suspected fault regular combination to obtain a target fault regular combination;

[0014] Based on big data, process the operating faults in the target fault regular combination.

[0015] Preferably, the determining the data transmission method, configuring the data transmission protocol, establishing the data transmission channel, and performing data automatic acquisition and transmission specifically include:

[0016] Adopt the wireless ZigBee technology for transmission between the Internet and the intelligent terminal;

[0017] The data acquisition software reads various data on the Internet in real time and converts them into a format for transmission;

[0018] Configure the data trigger mechanism to be timed trigger. When the trigger condition is met, the data acquisition software automatically acquires data and sends it to the intelligent terminal.

[0019] Preferably, the mining of the implicit relationship between the Internet operating faults and the performance index parameters specifically includes:

[0020] When an operating fault occurs, the performance index parameters that change by more than the preset amplitude are used as the target performance index parameters, and the performance index parameters that change by no more than the preset amplitude are used as the non-target performance index parameters. The preset amplitude is the normal fluctuation amplitude of the performance index parameters;

[0021] There is an implicit relationship between the target performance index parameters and the operating faults, and there is no implicit relationship between the non-target performance index parameters and the operating faults;

[0022] Equally divide the time range when the operating fault occurs to obtain at least one time point. Pair and fit the time point with the value of the non-target performance index parameter at the time point to obtain a preliminary function, and take the average of all preliminary functions to obtain a non-target function;

[0023] Pair and fit the time point with the value of the target performance index parameter at the time point to obtain a target function;

[0024] Integrate the ratio of the target function to the non-target function within the time range when the operating fault occurs to obtain the implicit coefficient of the target performance index parameter and the operating fault;

[0025] Set the implicit coefficient of the non-target performance index parameter and the operating fault to 0;

[0026] Combine the situations of the non-target performance index parameters and the target performance index parameters to obtain the implicit coefficient of the performance index parameters and the operating faults.

[0027] Preferably, the performing the correlation analysis on two operating faults based on the implicit relationship specifically includes:

[0028] Record the two operating failures as the first operating failure and the second operating failure respectively;

[0029] Take the difference between the implicit coefficient of the first operating failure and the performance index parameter and the implicit coefficient of the second operating failure and the performance index parameter, and take the absolute value to obtain the actual difference;

[0030] Accumulate the actual differences of all performance index parameters to obtain the overall difference;

[0031] Accumulate the implicit coefficients of the first operating failure and all performance index parameters and the implicit coefficients of the second operating failure and all performance index parameters to obtain the implicit sum;

[0032] Subtract 1 from the quotient of the overall difference divided by the implicit sum and take the absolute value to obtain the correlation coefficient.

[0033] Preferably, generating at least one fault regular combination based on the correlation between the two operating failures specifically includes:

[0034] Take all possible combinations of all operating failures as the to-be-verified fault combinations respectively;

[0035] Select any one operating failure from the to-be-verified fault combinations as the characteristic operating failure;

[0036] Based on big data, obtain the occurrence probability of the characteristic operating failure;

[0037] Pair the characteristic operating failure with the remaining operating failures in the to-be-verified fault combination to obtain at least one characteristic operating failure group;

[0038] Divide the correlation between the two operating failures in the characteristic operating failure group by the occurrence probability of the characteristic operating failure to obtain the independent probability;

[0039] Multiply the occurrence probability by the independent probabilities of all characteristic operating failure groups to obtain the to-be-identified probability;

[0040] Take the maximum value of the occurrence probabilities of the operating failures with the occurrence frequency less than 1 in a year as the recognition threshold;

[0041] Take the to-be-verified fault combination with the to-be-identified probability exceeding the recognition threshold as the fault regular combination.

[0042] Preferably, forming a judgment criterion for the suspected fault regular combination includes the following steps:

[0043] Based on big data, obtain the maximum value of the values of the performance index parameters when there is an operating failure as the maximum characteristic value, and obtain the minimum value of the values of the performance index parameters when there is an operating failure as the minimum characteristic value;

[0044] Take the maximum value of the maximum eigenvalues of the performance index parameters of all operating faults in the conventional fault combination as the upper threshold of the performance index parameters;

[0045] Take the minimum value of the minimum eigenvalues of the performance index parameters of all operating faults in the conventional fault combination as the lower threshold of the performance index parameters;

[0046] The lower threshold and the upper threshold of the performance index parameters form the evaluation range of the performance index parameters;

[0047] The judgment criterion for the suspected fault conventional combination is: when the real-time data of the performance index parameters respectively belong to the evaluation range of the performance index parameters in the conventional fault combination, then the conventional fault combination is taken as the suspected fault conventional combination.

[0048] Preferably, the screening of at least one suspected fault conventional combination to obtain the target fault conventional combination includes the following steps:

[0049] Obtain the fault degree range of the operating fault, equally spaced divide the fault degree range to obtain at least one identification point;

[0050] Randomly match at least one identification point to the operating faults in the suspected fault conventional combination, and each matching method forms a suspected allocation scheme;

[0051] Under the condition of the suspected allocation scheme, calculate the conditional value of the performance index parameter, and the conditional value of the performance index parameter is equal to the result of multiplying the performance index parameter by the implicit coefficient of the operating fault and the identification point of the operating fault in the suspected fault conventional combination and then accumulating;

[0052] Arrange the performance index parameters in ascending order according to the conditional value of the performance index parameter to obtain a conditional sequence;

[0053] Arrange the performance index parameters in ascending order according to the real-time data of the performance index parameter to obtain a real-time sequence;

[0054] If there is a condition sequence that is the same as the real-time sequence, then take the suspected fault conventional combination corresponding to the condition sequence as the target fault conventional combination;

[0055] Take the fault degree of the operating fault in the suspected allocation scheme for generating the target fault conventional combination as the real-time approximate fault degree of the operating fault.

[0056] Preferably, the processing of the operating faults in the target fault conventional combination based on big data includes the following steps:

[0057] Identify the urgency of the operating fault, multiply the urgency of the operating fault by the real-time approximate fault degree of the operating fault to obtain the severity of the operating fault;

[0058] According to the severity of the operating faults, the operating faults in the conventional combination of target faults are sorted from large to small to obtain an operating fault sequence;

[0059] Based on big data, obtain the parameters of the benchmark processing scheme when the severity of the operation fault is a preset value, and the preset value is any value within the value range of the severity of the operation fault;

[0060] The severity of the operating fault is divided by the preset value to obtain the target multiple;

[0061] The processing parameters are set according to the target multiples of the benchmark processing scheme parameters for the operating faults in the target fault conventional combination.

[0062] Preferably, the step of identifying the urgency of the operating failure comprises the following steps:

[0063] When the fault degree of the operation fault is 0, obtaining first sample data of the performance indicator parameter;

[0064] When the fault degree of the operation fault is a preset value, obtaining second sample data of the performance indicator parameter;

[0065] Based on the analytic hierarchy process, the weights of performance indicator parameters for Internet operation are obtained;

[0066] Accumulate the first sample data and the second sample data of all performance indicator parameters to obtain a comprehensive evaluation value;

[0067] According to the weight of the performance indicator parameter, the absolute value of the difference between the first sample data and the second sample data of the performance indicator parameter is accumulated to obtain a key coefficient;

[0068] The critical coefficient is divided by the comprehensive evaluation value to obtain the urgency of the operating failure.

[0069] Compared with the prior art, the present invention has the following beneficial effects:

[0070] By mining the implicit connection between Internet operation failures and performance indicator parameters, performing correlation analysis on two operation failures, generating at least one general combination of failures, obtaining a target general combination of failures and processing the operation failures in the target general combination of failures, the general combination of failures can be generated and then preliminarily screened in combination with pre-set standards. At the same time, through further screening, the target general combination of failures can be obtained, and the degree of the operation failures in the target general combination of failures can be determined, so that targeted maintenance parameters can be set. At the same time, the maintenance sequence can be set according to its severity, so that the operation and maintenance needs can be better met. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 Flow schematic diagram of the Internet operation and maintenance method optimized based on artificial intelligence algorithm of the present invention;

[0072] Figure 2 Flow schematic diagram of determining the data transmission mode, configuring the data transmission protocol, establishing the data transmission channel, and performing automatic data collection and transmission of the present invention;

[0073] Figure 3 Flow schematic diagram of mining the implicit connection between Internet operation faults and performance index parameters of the present invention;

[0074] Figure 4 Flow schematic diagram of performing correlation analysis on two operation faults based on the implicit connection of the present invention;

[0075] Figure 5 Flow schematic diagram of generating at least one fault regular combination based on the correlation between two operation faults of the present invention;

[0076] Figure 6 Flow schematic diagram of forming a judgment criterion for a suspected fault regular combination of the present invention;

[0077] Figure 7 Flow schematic diagram of screening at least one suspected fault regular combination to obtain a target fault regular combination of the present invention;

[0078] Figure 8 Flow schematic diagram of processing operation faults in the target fault regular combination based on big data of the present invention;

[0079] Figure 9 Flow schematic diagram of identifying the urgency of operation faults of the present invention. Detailed implementation manners

[0080] 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 in the following description are only examples, and those skilled in the art can think of other obvious variations.

[0081] Referring to Figure 1 as shown, an Internet operation and maintenance method optimized based on artificial intelligence algorithm includes:

[0082] Using an intelligent terminal to collect real-time data of at least one performance index parameter of the Internet in real time;

[0083] Determining the data transmission mode, configuring the data transmission protocol, establishing the data transmission channel, and performing automatic data collection and transmission;

[0084] Obtain at least one operation fault in the Internet, and mine the implicit connection between the Internet operation fault and the performance indicator parameter;

[0085] Based on the implicit connection, a correlation analysis is performed on the two operating faults, and based on the correlation of the two operating faults, at least one fault regular combination is generated;

[0086] Form judgment criteria for conventional combinations of suspected faults;

[0087] Obtaining at least one suspected fault conventional combination, which satisfies the judgment criteria that the performance indicator parameters meet the suspected fault conventional combination;

[0088] Screening at least one suspected fault regular combination to obtain a target fault regular combination;

[0089] Based on big data, operational faults in the regular combination of target faults are processed.

[0090] In this scheme, the main focus is on identifying situations where multiple faults occur in parallel. The main basis is that different operating faults have different effects on performance index parameters. Therefore, based on this, the existing operating faults can be determined by comparing with the real-time data of the performance index parameters. Therefore, it is necessary to analyze the relationship between the performance index parameters and the operating faults, and use the relationship obtained from the analysis to set up the corresponding algorithm to obtain the existing operating faults and their degree.

[0091] Reference Figure 2 As shown, determining the data transmission mode, configuring the data transmission protocol, establishing the data transmission channel, and performing automatic data collection and transmission specifically include:

[0092] Wireless ZigBee technology is used for transmission between the Internet and smart terminals;

[0093] The data collection software reads various data from the Internet in real time and converts them into a format for transmission;

[0094] Configure the data trigger mechanism as a timed trigger. When the trigger conditions are met, the data acquisition software automatically collects data and sends it to the smart terminal.

[0095] Reference Figure 3 As shown in Figure 1, mining the implicit relationship between Internet operation failures and performance indicator parameters specifically includes:

[0096] When an operation fault occurs, the performance indicator parameter that changes by more than a preset range will be used as the target performance indicator parameter, and the performance indicator parameter that changes by no more than a preset range will be used as the non-target performance indicator parameter. The preset range is the normal fluctuation range of the performance indicator parameter;

[0097] There is a hidden connection between the target performance index parameters and the operating faults, while there is no hidden connection between the non-target performance index parameters and the operating faults;

[0098] Divide the time range when the operating faults occur at equal intervals to obtain at least one time point. Pair the time points with the values of the non-target performance index parameters at the time points and fit them to obtain a preliminary function. Take the mean of all the preliminary functions to obtain the non-target function;

[0099] Pair the time points with the values of the target performance index parameters at the time points and fit them to obtain the target function;

[0100] Integrate the ratio of the target function to the non-target function within the time range when the operating faults occur to obtain the hidden coefficient between the target performance index parameters and the operating faults;

[0101] Set the hidden coefficient between the non-target performance index parameters and the operating faults to 0;

[0102] Combine the situations of the non-target performance index parameters and the target performance index parameters to obtain the hidden coefficient between the performance index parameters and the operating faults.

[0103] The hidden connection is mainly characterized by the hidden coefficient, and the hidden coefficient is obtained by comparing the target function and the non-target function. It mainly uses the non-target performance index parameters as the object of comparison, so as to characterize the relationship between different target performance index parameters. Here, since the non-target performance index parameters are all approximate and are hardly affected by the operating faults, their data are basically the same. Therefore, choosing any non-target performance index parameter for reference, the results are basically the same.

[0104] Reference Figure 4 As shown in

[0105] Record the two operating faults as the first operating fault and the second operating fault respectively;

[0106] Take the absolute value of the difference between the hidden coefficient of the first operating fault and the performance index parameters and the hidden coefficient of the second operating fault and the performance index parameters to obtain the actual difference;

[0107] Accumulate the actual differences of all the performance index parameters to obtain the overall difference;

[0108] Accumulate the hidden coefficients of the first operating fault and all the performance index parameters and the hidden coefficients of the second operating fault and all the performance index parameters to obtain the hidden total;

[0109] Take the absolute value of the difference between the result of dividing the overall difference by the hidden total and 1 to obtain the correlation coefficient.

[0110] The analysis of the correlation analysis is used to obtain the correlation coefficient, which is mainly obtained through the summary analysis of the gap between the implicit coefficients of the first operating fault and the second operating fault and each performance index parameter. The principle is that when the correlation between the first operating fault and the second operating fault is higher, the gap between their implicit coefficients with respect to the same performance index parameter is smaller, and thus the correlation is higher. Therefore, according to this principle, the correlation coefficient is obtained.

[0111] Refer to Figure 5 As shown, based on the correlation between two operating faults, generating at least one fault regular combination specifically includes:

[0112] Take all possible combinations of all operating faults as the fault combinations to be verified respectively;

[0113] Select any one operating fault in the fault combination to be verified as the characteristic operating fault;

[0114] Based on big data, obtain the occurrence probability of the characteristic operating fault;

[0115] Pair the characteristic operating fault with the remaining operating faults in the fault combination to be verified to obtain at least one characteristic operating fault group;

[0116] Divide the correlation between the two operating faults in the characteristic operating fault group by the occurrence probability of the characteristic operating fault to obtain the independent probability;

[0117] Multiply the occurrence probability by the independent probabilities of all characteristic operating fault groups to obtain the probability to be identified;

[0118] Take the maximum value of the occurrence probabilities of the operating faults with an occurrence frequency less than 1 in a year as the identification threshold;

[0119] Take the fault combination to be verified with the probability to be identified exceeding the identification threshold as the fault regular combination.

[0120] Since the combinations of faults are infinite, but in actual situations, only a finite number of fault combinations are worthy of consideration. Because, many fault combinations have too low occurrence probabilities, such as occurring once a year, so there is no need to consider them as regular Internet operation and maintenance at all. Therefore, set corresponding classification criteria to determine the fault regular combination.

[0121] Refer to Figure 6 As shown, forming the judgment criteria for the suspected fault regular combination includes the following steps:

[0122] Based on big data, obtain the maximum value of the value of the performance index parameter when there is an operating fault as the maximum eigenvalue, and obtain the minimum value of the value of the performance index parameter when there is an operating fault as the minimum eigenvalue;

[0123] Take the maximum value of the maximum eigenvalues of the performance index parameters of all operating faults in the conventional fault combination as the upper threshold of the performance index parameters;

[0124] Take the minimum value of the minimum eigenvalues of the performance index parameters of all operating faults in the conventional fault combination as the lower threshold of the performance index parameters;

[0125] The lower threshold and the upper threshold of the performance index parameters form the evaluation range of the performance index parameters;

[0126] The judgment criterion for the suspected fault conventional combination is: when the real-time data of the performance index parameters respectively belong to the evaluation range of the performance index parameters in the conventional fault combination, then take the conventional fault combination as the suspected fault conventional combination.

[0127] When the real-time data of the performance index parameters are determined, accordingly, a preliminary screening of the conventional fault combination can be carried out first, so as to reduce the amount of data analysis for subsequent detailed analysis. Because the operating faults in the conventional fault combination will have the maximum and minimum values of influence on the performance index parameters, therefore, summarize their overall situation to form the evaluation range of the performance index parameters corresponding to the conventional fault combination. Here, the maximum values of the influence of each operating fault on the performance index parameters are not superimposed. Because the acquisition environment is when there are operating faults, this is the actual environment, which must include the situation where various other operating faults occur in parallel. Therefore, there is no need to superimpose, just take the maximum value. For the minimum value situation, just take the minimum value accordingly. Then, a preliminary screening of the suspected fault conventional combination can be carried out according to the formed standard.

[0128] Refer to Figure 7 As shown, the steps for screening at least one suspected fault conventional combination to obtain the target fault conventional combination include:

[0129] Obtain the fault degree range of the operating fault, equally spaced divide the fault degree range to obtain at least one identification point;

[0130] Randomly match at least one identification point to the operating faults in the suspected fault conventional combination, and each matching method forms a suspected allocation plan;

[0131] Under the condition of the suspected allocation plan, calculate the conditional value of the performance index parameter. The conditional value of the performance index parameter is equal to the result of multiplying the performance index parameter by the implicit coefficient of the operating fault and the identification point of the operating fault in the suspected fault conventional combination and then accumulating;

[0132] Arrange the performance index parameters in ascending order according to the conditional value of the performance index parameter to obtain a conditional sequence;

[0133] Arrange the performance index parameters in ascending order according to their real-time data to obtain a real-time sequence;

[0134] If there is a condition sequence that is the same as the real-time sequence, then use the suspected fault conventional combination corresponding to the condition sequence as the target fault conventional combination;

[0135] Use the fault degree of the operating fault in the suspected allocation plan for generating the target fault conventional combination as the real-time approximate fault degree of the operating fault.

[0136] Here, since the fault degree of the operating fault in the Internet is unknown, for the suspected fault conventional combination to form a suspected allocation plan, multiple suspected allocation plans form all possible combinations of the fault degree of the operating fault in the suspected fault conventional combination. Because when the distance of the recognition point segmentation is small enough, any value of the fault degree of the operating fault in the suspected fault conventional combination can be approximated by a suspected allocation plan;

[0137] After that, it is necessary to screen the suspected allocation plan and the suspected fault conventional combination. Since the implicit relationship between the operating fault and the performance index parameters was determined before, therefore, according to the fault degree of the operating fault and the implicit relationship, calculate the numerical value of each performance index parameter and accumulate them to obtain a condition sequence. Then, the sorting relationship of the performance index parameters calculated from the suspected allocation plan and the suspected fault conventional combination corresponding to the actual situation at this time must be consistent with the sorting relationship of at least one performance index parameter sorted by real-time data. Therefore, the required target fault conventional combination can be selected.

[0138] Refer to Figure 8 As shown, based on big data, the processing of the operating fault in the target fault conventional combination includes the following steps:

[0139] Identify the urgency of the operating fault, and multiply the urgency of the operating fault by the real-time approximate fault degree of the operating fault to obtain the severity of the operating fault;

[0140] Arrange the operating faults in the target fault conventional combination in descending order according to the severity of the operating fault to obtain an operating fault sequence;

[0141] Based on big data, obtain the benchmark processing plan parameters when the severity of the operating fault is a preset value, and the preset value is any value within the value range of the severity of the operating fault;

[0142] Divide the severity of the operating fault by the preset value to obtain a target multiple;

[0143] Set the processing parameters according to the target multiple of the benchmark processing plan parameters of the operating fault in the target fault conventional combination.

[0144] When processing, it is necessary to allocate the processing order according to the severity of the operation failure. However, the severity of the operation failure is not only related to the urgency of the operation failure, but also related to the degree of the operation failure. The urgency of the operation failure is to evaluate the attributes of the operation failure, not to evaluate the severity of the failure. Taking two situations as examples, one is a very urgent failure, but the degree of its failure is very low, and the other is a non-urgent failure, but the degree of its failure is very serious. Therefore, corresponding algorithm settings are required to sort them. Otherwise, it is very difficult to make an accurate judgment.

[0145] Refer to Figure 9 As shown, identifying the urgency of the operation failure includes the following steps:

[0146] When the degree of the operation failure is 0, obtain the first sample data of the performance index parameters;

[0147] When the degree of the operation failure is a preset value, obtain the second sample data of the performance index parameters;

[0148] Based on the analytic hierarchy process, obtain the weight of the performance index parameters for Internet operation;

[0149] Accumulate the first sample data and the second sample data of all performance index parameters to obtain a comprehensive evaluation value;

[0150] According to the weight of the performance index parameters, accumulate the absolute value of the difference between the first sample data and the second sample data of the performance index parameters to obtain a key coefficient;

[0151] Divide the key coefficient by the comprehensive evaluation value to obtain the urgency of the operation failure.

[0152] Furthermore, this solution also proposes a storage medium, on which a computer-readable program is stored. When the computer-readable program is called, it executes the above-mentioned Internet operation and maintenance method optimized based on artificial intelligence algorithms.

[0153] It can be understood that the storage medium can be a magnetic medium, such as a floppy disk, a hard disk, a magnetic tape; an optical medium such as a DVD; or a semiconductor medium such as a solid state disk (SSD), etc.

[0154] In summary, the advantages of the present invention are as follows: by exploring the implicit connection between Internet operation failures and performance index parameters, analyzing the correlation between two operation failures, generating at least one regular combination of failures, obtaining the target regular combination of failures, and processing the operation failures in the target regular combination of failures, the generation of the regular combination of failures can be used to preliminarily screen the combination of failures in combination with the preset standards. At the same time, through further screening, the target regular combination of failures is obtained, and the degree of the operation failures in the target regular combination of failures is determined. Furthermore, targeted maintenance parameter settings can be made, and at the same time, according to its severity, the maintenance order is set, so as to better meet the operation and maintenance requirements.

[0155] The foregoing has shown and described 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 by the above embodiments. What is described in the above embodiments and the specification is only the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection required by the present invention is defined by the appended claims and their equivalents.

Claims

1. An Internet operation and maintenance method optimized based on artificial intelligence algorithms, characterized in that, Including: Using an intelligent terminal to collect real-time data of at least one performance metric parameter of the Internet in real time; Determining the data transmission method, configuring the data transmission protocol, establishing a data transmission channel, and performing automated data collection and transmission; Obtaining at least one operating fault in the Internet, and mining the implicit relationship between the Internet operating fault and the performance metric parameter; Based on the implicit relationship, performing a correlation analysis on two operating faults, and generating at least one regular fault combination based on the correlation of the two operating faults; Forming a judgment criterion for a suspected regular fault combination; Obtaining at least one suspected regular fault combination that satisfies the judgment criterion of the suspected regular fault combination for the performance metric parameter; Screening at least one suspected regular fault combination to obtain a target regular fault combination; Processing the operating faults in the target regular fault combination based on big data; The specific process of mining the implicit relationship between the Internet operating fault and the performance metric parameter includes: When an operating fault occurs, the performance metric parameter whose change exceeds the preset amplitude is taken as the target performance metric parameter, and the performance metric parameter whose change does not exceed the preset amplitude is taken as the non-target performance metric parameter. The preset amplitude is the normal fluctuation amplitude of the performance metric parameter; There is an implicit relationship between the target performance metric parameter and the operating fault, and there is no implicit relationship between the non-target performance metric parameter and the operating fault; Equally spaced dividing the time range when the operating fault occurs to obtain at least one time point, pairing and fitting the time point with the value of the non-target performance metric parameter at the time point to obtain a preliminary function, and taking the average of all preliminary functions to obtain a non-target function; Pairing and fitting the time point with the value of the target performance metric parameter at the time point to obtain a target function; Integrating the ratio of the target function to the non-target function within the time range when the operating fault occurs to obtain the implicit coefficient between the target performance metric parameter and the operating fault; The implicit coefficient between the non-target performance metric parameter and the operating fault is set to 0; Combining the situations of the non-target performance metric parameter and the target performance metric parameter to obtain the implicit coefficient between the performance metric parameter and the operating fault.

2. The Internet operation and maintenance method optimized based on artificial intelligence algorithm according to claim 1, characterized in that, The specific process of determining the data transmission method, configuring the data transmission protocol, establishing a data transmission channel, and performing automated data collection and transmission includes: Adopting the wireless ZigBee technology for transmission between the Internet and the intelligent terminal; The data acquisition software reads various data of the Internet in real time and converts them into a format for transmission; Configuring the data trigger mechanism as a timed trigger. When the trigger condition is met, the data acquisition software automatically collects data and sends it to the intelligent terminal.

3. The Internet operation and maintenance method optimized based on artificial intelligence algorithm according to claim 2, characterized in that, The specific process of performing a correlation analysis on two operating faults based on the implicit relationship includes: Denoting the two operating faults as the first operating fault and the second operating fault respectively; Taking the absolute value of the difference between the implicit coefficient of the first operating fault and the performance metric parameter and the implicit coefficient of the second operating fault and the performance metric parameter to obtain the actual difference; Accumulating the actual differences of all performance metric parameters to obtain the overall difference; Accumulating the implicit coefficients of the first operating fault and all performance metric parameters and the implicit coefficients of the second operating fault and all performance metric parameters to obtain the implicit sum; The absolute value of the difference between 1 and the overall gap divided by the implicit sum is taken to obtain the correlation coefficient.

4. The Internet operation and maintenance method optimized based on the artificial intelligence algorithm according to claim 3, wherein, Generating at least one fault regular combination based on the correlation between two operating faults specifically includes: Regarding all possible combinations of all operating faults as candidate fault combinations to be verified respectively; Selecting any one operating fault from the candidate fault combinations to be verified as the characteristic operating fault; Based on big data, obtaining the occurrence probability of the characteristic operating fault; Pairing the characteristic operating fault with the remaining operating faults in the candidate fault combination to be verified to obtain at least one group of characteristic operating faults; Dividing the correlation between the two operating faults in the group of characteristic operating faults by the occurrence probability of the characteristic operating fault to obtain the independent probability; Multiplying the occurrence probability by the independent probabilities of all groups of characteristic operating faults to obtain the probability to be identified; Taking the maximum value of the occurrence probabilities of the operating faults with occurrence frequencies less than 1 in a year as the identification threshold; Regarding the candidate fault combination to be verified with the probability to be identified exceeding the identification threshold as the fault regular combination.

5. A method for optimizing Internet operation and maintenance based on artificial intelligence algorithm according to claim 4, characterized in that, Forming a judgment criterion for the suspected fault regular combination includes the following steps: Based on big data, obtaining the maximum value of the values of the performance index parameters when there are operating faults as the maximum characteristic value, and obtaining the minimum value of the values of the performance index parameters when there are operating faults as the minimum characteristic value; Taking the maximum value of the maximum characteristic values of the performance index parameters of all operating faults in the fault regular combination as the upper threshold of the performance index parameters; Taking the minimum value of the minimum characteristic values of the performance index parameters of all operating faults in the fault regular combination as the lower threshold of the performance index parameters; The lower threshold and the upper threshold of the performance index parameters form the evaluation range of the performance index parameters; The judgment criterion for the suspected fault regular combination is: when the real-time data of the performance index parameters respectively belong to the evaluation range of the performance index parameters in the fault regular combination, then regarding the fault regular combination as the suspected fault regular combination.

6. The Internet operation and maintenance method optimized based on the artificial intelligence algorithm according to claim 5, wherein, Screening at least one suspected fault regular combination to obtain the target fault regular combination includes the following steps: Obtaining the fault degree range of the operating faults, equally spacing and dividing the fault degree range to obtain at least one identification point; Randomly matching at least one identification point to the operating faults in the suspected fault regular combination, and each matching method forms a suspected allocation plan; Under the condition of the suspected allocation plan, calculating the conditional value of the performance index parameter, and the conditional value of the performance index parameter is equal to the result of multiplying and accumulating the performance index parameter by the implicit coefficient of the operating fault and the identification point of the operating fault in the suspected fault regular combination; Arranging the performance index parameters in ascending order according to the conditional values of the performance index parameters to obtain a conditional sequence; Arranging the performance index parameters in ascending order according to the real-time data of the performance index parameters to obtain a real-time sequence; If there is a case where the conditional sequence is the same as the real-time sequence, then regarding the suspected fault regular combination corresponding to the conditional sequence as the target fault regular combination; Taking the fault degree of the operating fault in the suspected allocation plan for generating the target fault regular combination as the real-time approximate fault degree of the operating fault.

7. An Internet operation and maintenance method optimized based on an artificial intelligence algorithm according to claim 6, characterized in that Processing the operating faults in the target fault regular combination based on big data includes the following steps: Identify the urgency level of the operating fault, multiply the urgency level of the operating fault by the real-time approximate fault level of the operating fault to obtain the severity level of the operating fault; Sort the operating faults in the target fault conventional combination from large to small according to the severity level of the operating fault to obtain an operating fault sequence; Based on big data, obtain the benchmark processing scheme parameters when the severity level of the operating fault is a preset value, and the preset value is any value within the value range of the severity level of the operating fault; Divide the severity level of the operating fault by the preset value to obtain the target multiple; Set the processing parameters according to the target multiple of the benchmark processing scheme parameters of the operating fault in the target fault conventional combination.

8. An Internet operation and maintenance method optimized based on an artificial intelligence algorithm according to claim 7, characterized in that The identification of the urgency level of the operating fault includes the following steps: When the fault level of the operating fault is 0, obtain the first sample data of the performance index parameters; When the fault level of the operating fault is a preset value, obtain the second sample data of the performance index parameters; Based on the analytic hierarchy process, obtain the weight of the performance index parameters for Internet operation; Accumulate the first sample data and the second sample data of all performance index parameters to obtain a comprehensive evaluation value; According to the weight of the performance index parameters, accumulate the absolute value of the difference between the first sample data and the second sample data of the performance index parameters to obtain a key coefficient; Divide the key coefficient by the comprehensive evaluation value to obtain the urgency level of the operating fault.

Images