WO2022095381A1

WO2022095381A1 - Performance monitoring and evaluation method and apparatus, computer device, and readable storage medium

Info

Publication number: WO2022095381A1
Application number: PCT/CN2021/091338
Authority: WO
Inventors: 李桄宇
Original assignee: 平安科技（深圳）有限公司
Priority date: 2020-11-03
Filing date: 2021-04-30
Publication date: 2022-05-12
Also published as: CN112306808A; CN112306808B

Abstract

A performance monitoring and evaluation method and apparatus, a computer device, and a readable storage medium, relating to the technical field of artificial intelligence. The method comprises: obtaining performance data of a mobile terminal according to anomaly information sent from the mobile terminal (S102); training an initial deep learning network by means of the performance data and benchmark data to obtain a mature deep learning network (S106); generating a test result by calculating adjustable data by means of the mature deep learning network, calculating the difference between the test result and an equilibrium result to obtain a test loss value, adjusting the adjustable data according to the test loss value until the test result generated by the mature deep learning network is consistent with the equilibrium result, and setting the adjustable data as a dynamic indicator (S107); and monitoring a performance indicator of the mobile terminal and using same as a real-time indicator, and evaluating the real-time indicator according to the dynamic indicator to obtain an evaluation result (S109). The present method can accurately reflect the performance indicators of the mobile terminal in the normal state and at the critical point of the abnormal state, thereby ensuring the reliability of mobile terminal monitoring and evaluation.

Description

Performance monitoring and evaluation method, device, computer equipment and readable storage medium

This application claims the priority of the Chinese patent application with the application number CN 202011208472.8 and the title of "Performance Monitoring and Evaluation Method, Apparatus, Computer Equipment and Readable Storage Medium" filed on November 3, 2020, the entire contents of which are incorporated by reference in this application.

technical field

The present application relates to the technical field of intelligent decision-making of artificial intelligence, and in particular, to a performance monitoring and evaluation method, apparatus, computer equipment and readable storage medium, which use the deep learning technology of artificial intelligence.

Background technique

With the popularity of smart mobile terminals promoting the rapid development of mobile applications, human activities can basically be completed through mobile applications, and the performance of the applications is worthy of attention.

The mobile terminal runs a large number of apps all the time. Therefore, it is necessary to monitor the performance indicators of the mobile terminal through performance monitoring tools to prevent the mobile terminal from entering an abnormal state due to running a large number of apps (such as: freeze, freeze, system crash, etc.). ); current performance monitoring tools usually use preset performance thresholds to monitor various performance indicators of the mobile terminal and issue an alarm.

However, the inventor realized that once the mobile terminal is used for a long time, the performance of the mobile terminal will be degraded (for example, the hardware is aging, the hard disk stores too much data, a large number of APPs are downloaded, etc.), or when the user replaces the mobile terminal, Then the preset performance threshold will no longer match the performance index when the mobile terminal enters an abnormal state, so that it cannot accurately issue an alarm to the mobile terminal, resulting in low reliability of the performance alarm.

SUMMARY OF THE INVENTION

The purpose of this application is to provide a performance monitoring and evaluation method, device, computer equipment and readable storage medium, which are used to solve the problem that after long-term use of the mobile terminal in the prior art, the preset performance threshold will no longer be compatible with the mobile terminal. When the terminal enters an abnormal state, the performance indicators match, so that it cannot accurately send an alarm to the mobile terminal, resulting in the problem of low reliability of the performance alarm.

In order to achieve the above purpose, the present application provides a performance monitoring and evaluation method for monitoring and evaluating the performance indicators of the mobile terminal, including:

Obtain performance data of the mobile terminal according to the abnormal information sent by the mobile terminal, wherein the performance data reflects the performance index of the mobile terminal in an abnormal state;

A mature deep learning network is obtained by training the preset initial deep learning network through the performance data and the preset benchmark data, wherein the benchmark data reflects the performance index of the mobile terminal in a normal state;

Calculate the preset adjustable data through the mature deep learning network to generate the test result, and calculate the difference between the test result and the preset equalization result to obtain the test loss value, and adjust the adjustable data according to the test loss value. Data until the test result generated by the mature deep learning network is consistent with the equilibrium result, and the adjustable data is set as a dynamic index;

The performance index of the mobile terminal is monitored and taken as a real-time index, an evaluation result is obtained by evaluating the real-time index according to the dynamic index, and alarm information is sent to the mobile terminal according to the evaluation result.

In order to achieve the above purpose, the present application also provides a performance monitoring and evaluation device, comprising:

an abnormality acquisition module, configured to acquire performance data of the mobile terminal according to the abnormality information sent by the mobile terminal, wherein the performance data reflects the performance index of the mobile terminal in an abnormal state;

a network training module for training a preset initial deep learning network to obtain a mature deep learning network through the performance data and preset benchmark data, wherein the benchmark data reflects the performance index of the mobile terminal in a normal state;

The dynamic adjustment module is used to calculate the preset adjustable data through the mature deep learning network to generate the test result, and calculate the difference between the test result and the preset equalization result to obtain the test loss value, according to the test loss Adjust the adjustable data until the test result generated by the mature deep learning network is consistent with the equilibrium result, and set the adjustable data as a dynamic index;

The monitoring and alarming module is used to monitor the performance index of the mobile terminal and use it as a real-time index, evaluate the real-time index according to the dynamic index to obtain an evaluation result, and send alarm information to the mobile terminal according to the evaluation result.

In order to achieve the above object, the present application also provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and running on the processor, which is implemented when the processor of the computer device executes the computer program. The above performance monitoring and evaluation methods;

The performance monitoring and evaluation method includes:

In order to achieve the above purpose, the present application also provides a computer-readable storage medium, where a computer program is stored on the readable storage medium, and the above-mentioned performance monitoring is implemented when the computer program stored in the readable storage medium is executed by a processor. Evaluation method, the performance monitoring evaluation method includes:

The performance index of the mobile terminal is monitored and taken as a real-time index, an evaluation result is obtained by evaluating the real-time index according to the dynamic index, and alarm information is sent to the mobile terminal according to the evaluation result. The performance monitoring and evaluation method, device, computer equipment and readable storage medium provided by this application train the initial neural network through the performance data of the mobile terminal in normal and abnormal states, and obtain the ability to accurately judge the mobile terminal according to the performance index of the mobile terminal. Whether the terminal is in a normal state or an abnormal state of a mature deep learning network, the mature deep learning network generates dynamic indicators by adjusting the adjustable data.

Therefore, the adjustable data corresponding to the critical point between the normal state and the abnormal state of the mobile terminal is dynamically identified based on the mature deep learning network, and the adjustable data is set as a dynamic index, so that the mobile terminal is Whether the performance of the mobile terminal is degraded due to long-term use, or the mobile terminal is replaced, the dynamic index can correctly reflect the performance index of the mobile terminal at the critical point of normal state and abnormal state, so as to achieve Accurately send alarm information to the mobile terminal to ensure the reliability of mobile terminal monitoring and evaluation.

Description of drawings

Fig. 1 is the flow chart of the first embodiment of the performance monitoring and evaluation method of the present application;

Fig. 2 is the environmental application schematic diagram of the performance monitoring and evaluation method in the second embodiment of the performance monitoring and evaluation method of the application;

Fig. 3 is the specific method flow chart of the performance monitoring and evaluation method in the second embodiment of the performance monitoring and evaluation method of the present application;

4 is a schematic diagram of a program module of Embodiment 3 of the performance monitoring and evaluation device of the present application;

FIG. 5 is a schematic diagram of a hardware structure of a computer device in Embodiment 4 of the computer device of the present application.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present application more clearly understood, the present application will be described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, but not to limit the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

The performance monitoring and evaluation method, device, computer equipment and readable storage medium provided by this application are suitable for the technical field of intelligent decision-making of artificial intelligence, and provide a method based on abnormal acquisition module, network training module, dynamic adjustment module and monitoring and alarm module. Performance monitoring and evaluation methods. This application obtains the performance data of the mobile terminal according to the abnormal information sent by the mobile terminal; trains the initial deep learning network to obtain a mature deep learning network through the performance data and benchmark data; The difference between the result and the equilibrium result is the test loss value, and the adjustable data is adjusted according to the test loss value until the test result generated by the mature deep learning network is consistent with the equilibrium result, and the adjustable data is set as a dynamic index; monitor the performance index of the mobile terminal, And evaluate the performance index according to the dynamic index to obtain the evaluation result, and send the alarm information to the mobile terminal according to the evaluation result.

Example 1:

Referring to FIG. 1, a performance monitoring and evaluation method of the present embodiment is used to monitor and evaluate the performance indicators of the mobile terminal, including:

S102: Acquire performance data of the mobile terminal according to the abnormality information sent by the mobile terminal, wherein the performance data reflects the performance index of the mobile terminal in an abnormal state.

S106: Train a preset initial deep learning network to obtain a mature deep learning network by using the performance data and preset benchmark data, wherein the benchmark data reflects the performance index of the mobile terminal in a normal state.

S107: Generate a test result by calculating preset adjustable data through the mature deep learning network, and calculate the difference between the test result and a preset equalization result to obtain a test loss value, and adjust the test loss value according to the test loss value Adjust the data until the test result generated by the mature deep learning network is consistent with the equilibrium result, and set the adjustable data as a dynamic index.

S109: Monitor the performance index of the mobile terminal and use it as a real-time index, evaluate the real-time index according to the dynamic index to obtain an evaluation result, and send alarm information to the mobile terminal according to the evaluation result.

In this embodiment, the performance data of the mobile terminal is obtained through a performance monitoring tool, and the user information of the mobile terminal is obtained. The performance monitoring tool is a computer program for monitoring the performance of the mobile terminal, so as to obtain the abnormal state of the mobile terminal. performance data below.

Based on the performance data and the benchmark data used to reflect the performance index of the mobile terminal in a normal state, the initial neural network is trained to obtain a mobile terminal that can accurately judge whether the mobile terminal is in a normal state or an abnormal state according to the performance index of the mobile terminal. Mature deep learning network to realize the purpose of judging the normal state and abnormal state of the mobile terminal based on artificial intelligence technology. By entering the adjustable data into the input layer of the mature deep learning network, calling the hidden layer of the mature deep learning network to calculate the adjustable data of the input layer to obtain the test results, and outputting the test results to the mature deep learning network The output layer of the deep learning network; the test loss value is obtained by the preset test loss function and the difference between the test result and the preset equalization result, and the gradient descent method is used to adjust the test loss value according to the test loss value. Adjust the data until the mature deep learning network generates the test result and the equilibrium result through the adjustable data, and set the adjustable data as a dynamic index. Therefore, once the performance index of the mobile terminal exceeds the dynamic index, it can be regarded as abnormal, which greatly improves the judgment efficiency of the abnormal state of the mobile terminal, and further greatly reduces the monitoring of the performance index of the mobile terminal, and the judgment of the performance index in the said performance index. It can realize the technical effect of monitoring and judging the performance indicators of a large number of mobile terminals.

At the same time, due to the long-term use of the mobile terminal, the performance of the mobile terminal will be degraded (such as hardware aging, too much data stored on the hard disk, downloaded a large number of apps, etc.), or when the user replaces the mobile terminal, the performance index of the mobile terminal when it enters an abnormal state will also change accordingly. Therefore, the adjustable data corresponding to the critical point between the normal state and the abnormal state of the mobile terminal is dynamically identified based on the mature deep learning network, and the adjustable data is set as a dynamic index, so that the mobile terminal is Whether the performance of the mobile terminal is degraded due to long-term use, or the mobile terminal is replaced, the dynamic index can correctly reflect the performance index of the mobile terminal at the critical point of normal state and abnormal state, so as to achieve Accurately send alarm information to the mobile terminal to ensure the reliability of mobile terminal monitoring and evaluation.

The performance indicators of the mobile terminal are monitored by a performance monitoring tool, and dynamic indicators associated with the user information of the mobile terminal are obtained, the performance indicators are evaluated according to the dynamic indicators to obtain an evaluation result, and the mobile terminal is obtained according to the evaluation results. Whether there may be an abnormal state, and send alarm information to the mobile terminal that may be abnormal, so as to realize the real-time monitoring and evaluation of the mobile terminal, and the technical effect of real-time alarming the mobile terminal that may be in an abnormal state.

Embodiment 2:

This embodiment is a specific application scenario of the above-mentioned Embodiment 1. Through this embodiment, the method provided in this application can be described more clearly and specifically.

Hereinafter, the method provided by this embodiment will be specifically described by taking as an example that a mature deep learning network is used to adjust the adjustable data to generate a dynamic index in a server running the performance monitoring and evaluation method. It should be noted that this embodiment is only exemplary, and does not limit the protection scope of the embodiment of this application.

FIG. 2 schematically shows a schematic diagram of an environmental application of the performance monitoring and evaluation method according to the second embodiment of the present application.

In the exemplary embodiment, the server 2 where the performance monitoring and evaluation method is located is respectively connected to the mobile terminal 4 through the network 3; the server 2 may provide services through one or more networks 3, and the network 3 may include various network devices, such as Routers, switches, multiplexers, hubs, modems, bridges, repeaters, firewalls, proxy devices and/or etc. The network 3 may include physical links such as coaxial cable links, twisted pair cable links, fiber optic links, combinations thereof, and/or the like. The network 3 may include wireless links, such as cellular links, satellite links, Wi-Fi links and/or the like; the mobile terminal 4 may be computer equipment such as smart phones, tablet computers, notebook computers, desktop computers, etc.

FIG. 3 is a flowchart of a specific method of a performance monitoring and evaluation method provided by an embodiment of the present application, and the method specifically includes steps S200 to S209.

S200: Create an upload interface for monitoring abnormal information of the mobile terminal, and execute step S202;

In order to know the abnormal state of the mobile terminal during operation in time, so as to obtain the performance data of the mobile terminal in the abnormal state in time, this step creates an upload interface to monitor the abnormal information uploaded by the mobile terminal in real time. When abnormal, upload abnormal information to the upload interface, so as to realize the rapid discovery of the abnormal state of the mobile terminal, so as to obtain the performance data under the abnormal state in time; wherein, the abnormal information is the mobile terminal judging its own running state is in an abnormal state, and the sent message is used to reflect that the mobile terminal is currently in an abnormal state.

In this embodiment, by setting a button on the mobile terminal and associating it with the upload interface, the proxy event corresponding to the button is used to send abnormal information to the upload interface, so that the user only needs to click the button. The sending of abnormal information realizes the effect of one-click abnormal feedback.

S201: Send a status request to a mobile terminal, and receive status information sent by the mobile terminal according to the status request; determine whether the status information has an abnormal tag; if so, determine that the status information is abnormal information, and execute S202; if not, end.

In order to be able to find out more situations where the mobile terminal has an abnormal state, this step sends a status request to the mobile terminal, and receives the status information sent by the mobile terminal according to the status request; The pop-up box is used as a status request, and the mobile terminal generates status information with an abnormal label by clicking the abnormal button; if the current state of the mobile terminal is normal, the mobile terminal will click the normal button in the pop-up box to send a message with normal status. The status information of the tag, or the status information with the cancellation tag is sent by clicking the close button in the pop-up box. The abnormal information is a message sent by the user to determine that the mobile terminal is currently in an abnormal state, and is used to reflect that the mobile terminal is currently in an abnormal state.

S202: Acquire performance data of the mobile terminal according to the abnormality information sent by the mobile terminal, wherein the performance data reflects the performance index of the mobile terminal in an abnormal state.

In order to obtain the performance data of the mobile terminal in an abnormal state, and to make the mobile terminal and the performance data correspond to each other, in this step, the performance data of the mobile terminal is obtained through a performance monitoring tool, and the user information of the mobile terminal is obtained. Monitoring tools are computer programs used to monitor mobile performance, such as PerfDog, MobilePerformance Soloπ, Testin, etc. The performance data refers to performance indicators that reflect the abnormal state of the mobile terminal, such as CPU usage, memory usage, power consumption, and the like.

S203: Acquire user information of the mobile terminal, associate and save the performance data with the user information.

In this step, the user information is used as data information for marking the user identity using the mobile terminal, for example: the user's login account on the mobile terminal, the user's mobile phone number, and the like. Based on the key-value key-value pair technology, the user information is used as a primary key and the performance data is used as a key value to construct a key-value pair, so as to realize the association between the performance data and the user information, and combine the performance data with the user information. User information is sent to the preset exception database for preservation.

S204: Call the exception database to aggregate performance data belonging to the same user information to form a performance set, and delete performance data that exceeds a preset time limit in the performance set.

Since the performance of the mobile terminal will be degraded with the long-term use of the mobile terminal, or the user changes the mobile terminal, the performance data corresponding to the user information will no longer match the replaced mobile terminal. After a performance set is formed from the performance data obtained from the set, the performance data that exceeds the preset time limit in the performance set is deleted, so as to avoid the impact of the performance data before a long period on the dynamic indicators currently judging the mobile terminal.

In this embodiment, the step of deleting performance data that exceeds a preset time limit in the performance set includes:

S41: Extract the current time, where the current time is data information reflecting the current date.

S42: Subtract the current time and the time limit to obtain a deadline.

In this step, the time limit can be counted as day, month and year. Exemplarily, if the current time is 2020/3/1 and the time limit is 2 months, then the deadline is 2020/1/1.

S43: Delete performance data whose date is earlier than the deadline in the performance collection.

In this step, the performance data in the performance set has a time stamp, and the time stamp reflects the time when the performance data is acquired, the time stamp of each performance data in the performance set is identified, and the time stamp with a date earlier than the deadline is deleted. corresponding performance data.

S205: Calculate the number of performance data in the performance set in the abnormal database, and determine whether the number of performance data reaches a preset training threshold; if so, extract the performance data in the performance set, and execute S206; if not, end .

In order to avoid the situation that a small amount of performance data cannot make the deep learning network stably and accurately determine the dynamic indicators, this step calculates the number of performance data in the performance set in the abnormal database, and determines whether the number of performance data reaches the training threshold. ; If the quantity of performance data reaches the training threshold, the deep learning network can be trained; if the quantity of performance data does not reach the training threshold, end until the quantity of the performance data reaches the training threshold.

In FIG. 3, the S205 is shown with the following labels:

S51: Calculate the quantity of performance data in the performance set in the abnormal database, and determine whether the quantity of the performance data reaches a preset training threshold;

S52: If yes, extract the performance data in the performance set, and execute S206;

S53: If no, end.

S206: Train a preset initial deep learning network to obtain a mature deep learning network by using the performance data and preset benchmark data, wherein the benchmark data reflects the performance index of the mobile terminal in a normal state.

In order to realize the purpose of judging the normal state and abnormal state of the mobile terminal based on artificial intelligence technology, in this step, the initial neural network is trained by the performance data and the benchmark data used to reflect the performance index of the mobile terminal in the normal state, In order to obtain a mature deep learning network that can accurately judge whether the mobile terminal is in a normal state or an abnormal state according to the performance indicators of the mobile terminal.

In a preferred embodiment, the step of training a preset initial deep learning network to obtain a mature deep learning network by using the performance data and preset benchmark data includes:

S61: obtain benchmark data;

S62: According to the sub-data items in the benchmark data and the performance data, construct the neurons of the input layer in the initial deep learning network; enter the data corresponding to each sub-data item in the benchmark data into the corresponding input layer In the neuron, the data corresponding to each sub-data item in the performance data is entered into the corresponding neuron in the input layer;

In this step, the sub-data item is the metadata of the sub-data in the benchmark data and the performance data, and the sub-data is the smallest data unit that constitutes the benchmark data and the performance data; for example, the performance data is CPU usage: 35%, Use memory: 550M, power consumption 0.26mAh/s; sub-data items of this performance data include: CPU usage: 35%, used memory: 550M, power consumption 0.26mAh/s; sub-data items of this performance data include: CPU usage, memory usage, power consumption. The benchmark data is: CPU usage: 1%, memory usage: 10M, power consumption 0.01mAh/s; sub-data of this benchmark data include: CPU usage rate: 1%, memory usage: 10M, power consumption 0.01mAh/s s; the sub-data items of the benchmark data include: CPU usage, memory usage, and power consumption.

S63: Invoke the hidden layer of the initial deep learning network, obtain the input vector of the input layer and perform operations on it to obtain an output vector, and output the output vector to the output layer of the initial deep learning network.

In this step, the output layer includes normal neurons and abnormal neurons

Among them, the normal neuron is used to express the probability that the mobile terminal is in a normal state, and the abnormal neuron is used to express the probability that the mobile terminal is in an abnormal state.

Further, the initial deep learning network is trained by the benchmark data and performance data at the same time, the purpose is to avoid training the initial deep learning network only by the benchmark data or performance data, resulting in the obtained mature deep learning network can only judge, Whether the mobile terminal is in a normal state or whether it is in an abnormal state occurs, and it is impossible to determine the normal state and the abnormal state at the same time. Therefore, it is helpful to obtain the critical point when the mobile terminal is in a normal state and an abnormal state.

S64: Calculate the difference between the output vector of the output layer and the abnormal information corresponding to the performance data by using a preset training loss function, and obtain a training loss value.

S65: Adjust the weight and paranoia value of the hidden layer through the back-propagation algorithm and according to the training loss value until the training loss value is less than a preset training loss threshold to obtain a mature deep learning network.

It should be noted that using BP (Back Propagation) neural network as the initial deep learning network, BP network (Back Propagation), is a multi-layer feedforward network trained according to the error back propagation algorithm, and is one of the most widely used neural network models at present. BP network can learn and store a large number of input-output pattern mapping relationships without revealing the mathematical equations describing this mapping relationship in advance.

The loss function is a way to measure the predicted and actual values of an artificial neural network (ANN). It is used for training the deep learning network, so that the predicted value of the deep learning network is close to or even in line with the actual value.

The back-propagation method refers to a supervised learning algorithm, which is often used to train a multilayer perceptron and a forward neural network. The back-propagation algorithm (BP algorithm) mainly consists of two links (excitation propagation, weight update) iterating repeatedly until the network's response to the input reaches a predetermined target range.

S207: Generate a test result by calculating preset adjustable data through the mature deep learning network, and calculate the difference between the test result and the preset equalization result to obtain a test loss value, and adjust the test loss value according to the test loss value Adjust the data until the test result generated by the mature deep learning network is consistent with the equilibrium result, and set the adjustable data as a dynamic index.

If a mature deep learning network is used to monitor the performance indicators of the mobile terminal in real time, a large part of the computing power of the server will be consumed, and if a large number of performance indicators of the mobile terminal are monitored, it is even more impossible to achieve; in this step, the preset Adjustable data is entered into the input layer of the mature deep learning network, and the hidden layer of the mature deep learning network is called to calculate the adjustable data of the input layer to obtain test results, and the test results are output to the mature deep learning network The output layer of the network; the test loss value is obtained by the preset test loss function and the difference between the test result and the preset equalization result, and the adjustable data is adjusted according to the test loss value by using the gradient descent method , until the mature deep learning network generates the test result and the equilibrium result through the adjustable data, and sets the adjustable data as a dynamic index. The equilibrium result reflects a critical result between a normal state and an abnormal state, for example, the normal probability value of the output layer is 0.5, and the abnormal probability value is 0.5. In this embodiment, the adjustable data and the metadata of the performance data are in one-to-one correspondence, and the data set in which the value corresponding to each metadata in the adjustable data can be adjusted includes CPU usage. Rate, memory usage, power consumption, for example: adjustable data include: CPU usage: X%, memory usage: YM, power consumption ZAh/s; where X, Y, Z can be any real number. Further, the test result is a data set generated by the mature deep learning network calculating the adjustable data, and the test result is in one-to-one correspondence with the metadata of the test result, which includes normal probability values and abnormal values. The probability value, for example, the adjustable data is CPU usage: 35%, memory usage: 550M, power consumption 0.26mAh/s; the obtained test results are the normal probability value of 0.2 and the abnormal probability value of 0.8.

Therefore, by obtaining a dynamic index that can be determined by a mature deep learning network as a critical point between a normal state and an abnormal state (ie, the equilibrium result), therefore, once the performance index of the mobile terminal exceeds the dynamic index, it can be identified as abnormal, The efficiency of judging the abnormal state of the mobile terminal is greatly improved, which in turn greatly reduces the computing power consumed by monitoring the performance indicators of the mobile terminal and judging whether the mobile terminal is abnormal under the performance indicators. Performance indicators to monitor and judge the technical effect.

It should be noted that the gradient descent method (Gradient Descent) is an optimizer commonly used in machine learning, which is used to predict the experimental results of mature deep learning networks according to adjustable data, and optimally adjust the adjustable as an input layer according to the experimental results. Adjust the data, and finally match the test results with the equilibrium results. Based on the above example, if the adjustable data is CPU usage: 35%, memory used: 550M, and power consumption 0.26mAh/s; the test results obtained are the normal probability value of 0.2 and the abnormal probability value of 0.8; then the normal probability value is 0.8. The value of 0.2 is 0.3 lower than the normal probability of the balanced result at 0.5, and the abnormal probability value of 0.8 is 0.3 higher than the abnormal probability value of 0.5 of the balanced result, so it is necessary to reduce the CPU usage in the tunable data and/or The values of memory and/or power consumption are used to increase the normal probability value of the test result and decrease the abnormal probability value of the test result until the value of the test result agrees with the value of the equilibrium result.

S208: Associate and save the dynamic indicator with the user information of the mobile terminal.

In order to ensure that the dynamic index can be repeatedly called in a long time, and to ensure that the dynamic index can accurately correspond to the corresponding mobile terminal; this step ensures that the user with the user information can accurately call by associating the dynamic index with the user information. Corresponding dynamic indicators are used to evaluate the performance indicators of the mobile terminal used, which ensures the directivity and accuracy of dynamic indicators evaluation, and sends the associated dynamic indicators and user information to the preset indicator server, so that the dynamic indicators can be It is reused for a long time to ensure the reliability of the performance index monitoring of the mobile terminal.

In this embodiment, the key-value key-value pair technology is used to associate the dynamic indicator with the user information, wherein the user information is used as the primary key and the dynamic indicator is used as the key value to form a mutual association. key-value pair.

Optionally, after associating and saving the dynamic indicator with the user information of the mobile terminal, the method further includes:

Upload the dynamic indicator and the user information to the blockchain.

It should be noted that the corresponding summary information is obtained based on the dynamic index and the user information. Specifically, the summary information is obtained by hashing the dynamic index and the user information, for example, by using the sha256s algorithm. Uploading summary information to the blockchain ensures its security and fairness and transparency to users. User equipment can download the summary information from the blockchain to verify whether dynamic indicators and user information have been tampered with. The blockchain referred to in this example is a new application mode of computer technology such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm. Blockchain, essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information to verify its Validity of information (anti-counterfeiting) and generation of the next block. The blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.

S209: Monitor the performance index of the mobile terminal and use it as a real-time index, evaluate the real-time index according to the dynamic index to obtain an evaluation result, and send alarm information to the mobile terminal according to the evaluation result.

In order to monitor and evaluate the performance indicators of the mobile terminal in real time, and send alarm information to the mobile terminal that may be in an abnormal state in real time, in this step, the performance indicators of the mobile terminal are monitored by a performance monitoring tool, and user information related to the mobile terminal is obtained. The associated dynamic index, the performance index is evaluated according to the dynamic index to obtain an evaluation result, and according to the evaluation result, it is known whether the mobile terminal may be in an abnormal state, and an alarm message may be sent to the mobile terminal that may be abnormal, In order to realize the technical effect of real-time monitoring and evaluation of the mobile terminal, and real-time alarm to the mobile terminal that may appear abnormal state.

In a preferred embodiment, the monitoring of the performance index of the mobile terminal, and evaluating the performance index according to the dynamic index to obtain an evaluation result, and the step of sending alarm information to the mobile terminal according to the evaluation result includes the following steps: :

S91: Monitor the performance indicators of the mobile terminal and use them as real-time indicators, and acquire user information of the mobile terminal and extract dynamic indicators associated with the user information.

In this step, the performance monitoring tool is used to monitor the performance indicators of the mobile terminal, and user information of the mobile terminal is acquired, and an indicator server is invoked to acquire dynamic indicators associated with the user information.

S92: Set the real-time index whose value exceeds the dynamic index as an abnormal index, and set the number of the abnormal index as an evaluation result. In this step, the performance index is compared with the dynamic index, the performance index whose value exceeds the dynamic index is identified, and the performance index is set as the abnormal index; the number of the abnormal index is calculated, and the number is set as the evaluation result.

Exemplarily, it is assumed that the dynamic indicators associated with the user information of the mobile terminal A include: CPU: 30%, power consumption: 0.238mAh/s, memory: 500M;

The performance indicators obtained by monitoring mobile terminal A include: CPU: 35%, power consumption: 0.26mAh/s, memory: 550M; then, since the device CPU: 35% > CPU threshold 30%; the device power consumption 0.26mAh/s > The power consumption threshold is 0.238mAh/s; the device memory 550M > the memory threshold 500M. Therefore, an evaluation result of 3 was obtained.

S93: Determine whether the evaluation result exceeds a preset abnormal threshold;

S94: if yes, send alarm information to the mobile terminal;

S95: If no, end.

Exemplarily, assuming that the abnormality threshold is 2, and based on the above example, the obtained evaluation result is 3, it is determined that the mobile terminal may appear or has experienced an abnormal state, and therefore, alarm information is sent to the mobile terminal.

Embodiment three:

Please refer to FIG. 4 , a performance monitoring and evaluation device 1 of this embodiment includes:

An abnormality acquisition module 12, configured to acquire performance data of the mobile terminal according to the abnormality information sent by the mobile terminal, wherein the performance data reflects the performance index of the mobile terminal in an abnormal state;

The network training module 16 is used for training the preset initial deep learning network to obtain a mature deep learning network through the performance data and the preset benchmark data, wherein the benchmark data reflects the performance index of the mobile terminal in a normal state ;

The dynamic adjustment module 17 is configured to calculate the preset adjustable data through the mature deep learning network to generate the test result, and calculate the difference between the test result and the preset equalization result to obtain the test loss value, according to the test The loss value is adjusted to the adjustable data until the test result generated by the mature deep learning network is consistent with the equilibrium result, and the adjustable data is set as a dynamic index;

The monitoring and alarming module 19 is used to monitor the performance index of the mobile terminal and use it as a real-time index, evaluate the real-time index according to the dynamic index to obtain an evaluation result, and send alarm information to the mobile terminal according to the evaluation result.

Optionally, the performance monitoring and evaluation device 1 further includes:

The creation module 10 is used to create an upload interface for monitoring the abnormal information of the mobile terminal, and call the abnormal acquisition module 12;

The information request module 11 is used for sending a status request to the mobile terminal, and receiving the status information sent by the mobile terminal according to the status request; judging whether the status information has an abnormal label; if yes, then judging that the status information is exception information, and call the exception acquisition module 12; if not, end.

The information association module 13 is configured to acquire user information of the mobile terminal, and associate and save the performance data with the user information.

The information deletion module 14 is configured to call the exception database to aggregate performance data belonging to the same user information to form a performance set, and delete performance data exceeding a preset time limit in the performance set.

The training trigger module 15 is used to calculate the quantity of performance data in the performance set in the abnormal database, and judge whether the quantity of the performance data reaches a preset training threshold; if so, extract the performance data in the performance set, and invoke network training Module 16,; if not, end.

The association saving module 18 is configured to associate and save the dynamic indicator with the user information of the mobile terminal.

This technical solution is applied to the field of intelligent decision-making of artificial intelligence, and the performance data of the mobile terminal is obtained according to the abnormal information sent by the mobile terminal; through the performance data and the benchmark data, the neural network is used as the initial deep learning network for training, and a mature deep learning network is obtained and its As a classification model; the test result is generated by calculating the adjustable data through the mature deep learning network, and the test loss value is obtained by calculating the gap between the test result and the equilibrium result, and the adjustable data is adjusted according to the test loss value until the test result generated by the mature deep learning network. Consistent with the equilibrium result, the adjustable data is set as the dynamic index; the performance index of the mobile terminal is monitored, and the performance index is evaluated according to the dynamic index to obtain the evaluation result, and alarm information is sent to the mobile terminal according to the evaluation result.

Embodiment 4:

In order to achieve the above purpose, the present application also provides a computer device 5, the components of the performance monitoring and evaluation device 1 of the third embodiment can be dispersed in different computer devices, and the computer device 5 can be a smart phone, tablet computer, Notebook computers, desktop computers, rack servers, blade servers, tower servers or rack servers (including independent servers, or server clusters composed of multiple application servers), etc. The computer device in this embodiment at least includes but is not limited to: a memory 51 and a processor 52 that can be communicatively connected to each other through a system bus, as shown in FIG. 5 . It should be pointed out that FIG. 5 only shows a computer device having a component -, but it should be understood that it is not required to implement all the shown components, and more or less components may be implemented instead.

In this embodiment, the memory 51 (ie, a readable storage medium) includes flash memory, hard disk, multimedia card, card-type memory (eg, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), Read-Only Memory (ROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Programmable Read-Only Memory (PROM), Magnetic Memory, Magnetic Disk, Optical Disk, etc. In some embodiments, the memory 51 may be an internal storage unit of a computer device, such as a hard disk or a memory of the computer device. In other embodiments, the memory 51 may also be an external storage device of a computer device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), Secure Digital (SD) card, Flash Card (Flash Card), etc. Of course, the memory 51 may also include both the internal storage unit of the computer device and its external storage device. In this embodiment, the memory 51 is generally used to store the operating system and various application software installed on the computer equipment, such as the program code of the performance monitoring and evaluation apparatus of the third embodiment. In addition, the memory 51 can also be used to temporarily store various types of data that have been output or will be output.

The processor 52 may be a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 52 is typically used to control the overall operation of the computer device. In this embodiment, the processor 52 is used for running the program code or processing data stored in the memory 51, for example, running a performance monitoring and evaluation apparatus, so as to implement the performance monitoring and evaluation methods of the first and second embodiments.

Embodiment 5:

In order to achieve the above purpose, the present application also provides a computer-readable storage medium. In this embodiment, the computer-readable storage medium may be volatile or non-volatile, such as flash memory, hard disk, multimedia card, Card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), programmable Read only memory (PROM), magnetic memory, magnetic disk, optical disk, server, App application mall, etc., on which computer programs are stored, and when the programs are executed by the processor 52, corresponding functions are realized. The computer-readable storage medium of this embodiment is used to store the performance monitoring and evaluation apparatus, and when executed by the processor 52, implements the performance monitoring and evaluation methods of the first embodiment and the second embodiment.

The above-mentioned serial numbers of the embodiments of the present application are only for description, and do not represent the advantages or disadvantages of the embodiments.

From the description of the above embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus a necessary general hardware platform, and of course hardware can also be used, but in many cases the former is better implementation.

The above are only the preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the contents of the description and drawings of the present application, or directly or indirectly applied in other related technical fields , are similarly included within the scope of patent protection of this application.

Claims

A performance monitoring and evaluation method is used to monitor and evaluate the performance indicators of the mobile terminal, including:

Obtain performance data of the mobile terminal according to the abnormal information sent by the mobile terminal, wherein the performance data reflects the performance index of the mobile terminal in an abnormal state;

A mature deep learning network is obtained by training the preset initial deep learning network through the performance data and the preset benchmark data, wherein the benchmark data reflects the performance index of the mobile terminal in a normal state;

Calculate the preset adjustable data through the mature deep learning network to generate the test result, and calculate the difference between the test result and the preset equalization result to obtain the test loss value, and adjust the adjustable data according to the test loss value. Data until the test result generated by the mature deep learning network is consistent with the equilibrium result, and the adjustable data is set as a dynamic index;

The performance index of the mobile terminal is monitored and taken as a real-time index, an evaluation result is obtained by evaluating the real-time index according to the dynamic index, and alarm information is sent to the mobile terminal according to the evaluation result.
The performance monitoring and evaluation method according to claim 1, wherein before acquiring the performance data of the mobile terminal according to the abnormal information sent by the mobile terminal, the method further comprises:

Create an upload interface for monitoring exception information on the mobile terminal.
The performance monitoring and evaluation method according to claim 1, wherein before acquiring the performance data of the mobile terminal according to the abnormal information sent by the mobile terminal, the method further comprises:

Send a status request to the mobile terminal, and receive the status information sent by the mobile terminal according to the status request; determine whether the status information has an abnormal label; if so, determine that the status information is abnormal information; if not, then Finish.
The performance monitoring and evaluation method according to claim 1, wherein after acquiring the performance data of the mobile terminal according to the abnormal information sent by the mobile terminal, the method further comprises:

acquiring user information of the mobile terminal, associating and saving the performance data with the user information;

Call the exception database to aggregate the performance data belonging to the same user information to form a performance set, and delete the performance data in the performance set that exceeds the preset time limit;

Calculate the number of performance data in the performance set in the abnormal database, and determine whether the number of performance data reaches a preset training threshold; if so, extract the performance data in the performance set; if not, end.
The performance monitoring and evaluation method according to claim 1, wherein the step of training a preset initial deep learning network to obtain a mature deep learning network by using the performance data and the preset benchmark data comprises:

Get benchmark data;

According to the sub-data items in the benchmark data and performance data, construct the neurons of the input layer in the initial deep learning network; enter the data corresponding to each sub-data item in the benchmark data into the corresponding neurons in the input layer In the unit, the data corresponding to each sub-data item in the performance data is entered into the corresponding neuron in the input layer;

Calling the hidden layer of the initial deep learning network, obtaining the input vector of the input layer and operating it to obtain an output vector, and outputting the output vector to the output layer of the initial deep learning network;

Calculate the difference between the output vector of the output layer and the abnormal information corresponding to the performance data through the preset training loss function, and obtain the training loss value;

A mature deep learning network is obtained by adjusting the weight and paranoia value of the hidden layer through the back-propagation algorithm and according to the training loss value until the training loss value is less than a preset training loss threshold.
The performance monitoring and evaluation method according to claim 1, wherein after the adjustable data is set as a dynamic index, the method further comprises:

associating and saving the dynamic indicator with the user information of the mobile terminal;

After associating and saving the dynamic indicator with the user information of the mobile terminal, the method further includes:

Upload the dynamic indicator and the user information to the blockchain.
The performance monitoring and evaluation method according to claim 1, wherein the performance index of the monitoring mobile terminal is used as a real-time index, an evaluation result is obtained by evaluating the real-time index according to the dynamic index, and an evaluation result is sent to the mobile terminal according to the evaluation result. The step of sending alarm information by the mobile terminal includes:

Monitor the performance index of the mobile terminal and use it as a real-time index, and obtain the user information of the mobile terminal and extract the dynamic index associated with the user information;

Set the real-time index whose value exceeds the dynamic index as an abnormal index, and set the number of the abnormal index as an evaluation result; determine whether the evaluation result exceeds a preset abnormal threshold;

If so, send alarm information to the mobile terminal;

If not, end.
A performance monitoring and evaluation device, comprising:

an abnormality acquisition module, configured to acquire performance data of the mobile terminal according to the abnormality information sent by the mobile terminal, wherein the performance data reflects the performance index of the mobile terminal in an abnormal state;

a network training module for training a preset initial deep learning network to obtain a mature deep learning network through the performance data and preset benchmark data, wherein the benchmark data reflects the performance index of the mobile terminal in a normal state;

The dynamic adjustment module is used to calculate the preset adjustable data through the mature deep learning network to generate the test result, and calculate the difference between the test result and the preset equalization result to obtain the test loss value, according to the test loss Adjust the adjustable data until the test result generated by the mature deep learning network is consistent with the equilibrium result, and set the adjustable data as a dynamic index;

The monitoring and alarming module is used to monitor the performance index of the mobile terminal and use it as a real-time index, evaluate the real-time index according to the dynamic index to obtain an evaluation result, and send alarm information to the mobile terminal according to the evaluation result.
A computer device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor of the computer device implements a performance monitoring and evaluation method when executing the computer program, and the Performance monitoring and evaluation methods, including:

Obtain performance data of the mobile terminal according to the abnormal information sent by the mobile terminal, wherein the performance data reflects the performance index of the mobile terminal in an abnormal state;

A mature deep learning network is obtained by training the preset initial deep learning network through the performance data and the preset benchmark data, wherein the benchmark data reflects the performance index of the mobile terminal in a normal state;

Calculate the preset adjustable data through the mature deep learning network to generate the test result, and calculate the difference between the test result and the preset equalization result to obtain the test loss value, and adjust the adjustable data according to the test loss value. Data until the test result generated by the mature deep learning network is consistent with the equilibrium result, and the adjustable data is set as a dynamic index;

The performance index of the mobile terminal is monitored and taken as a real-time index, an evaluation result is obtained by evaluating the real-time index according to the dynamic index, and alarm information is sent to the mobile terminal according to the evaluation result.
The computer device according to claim 9, wherein, before acquiring the performance data of the mobile terminal according to the abnormal information sent by the mobile terminal, the method further comprises:

Create an upload interface for monitoring the abnormal information of the mobile terminal;

Before acquiring the performance data of the mobile terminal according to the abnormal information sent by the mobile terminal, the method further includes:

Send a status request to the mobile terminal, and receive the status information sent by the mobile terminal according to the status request; determine whether the status information has an abnormal label; if so, determine that the status information is abnormal information; if not, then Finish.
The computer device according to claim 9, wherein, after obtaining the performance data of the mobile terminal according to the abnormal information sent by the mobile terminal, the method further comprises:

acquiring user information of the mobile terminal, associating and saving the performance data with the user information;

Call the exception database to aggregate the performance data belonging to the same user information to form a performance set, and delete the performance data in the performance set that exceeds the preset time limit;

Calculate the number of performance data in the performance set in the abnormal database, and determine whether the number of performance data reaches a preset training threshold; if so, extract the performance data in the performance set; if not, end.
The computer device according to claim 9, wherein the step of training a preset initial deep learning network to obtain a mature deep learning network by using the performance data and the preset benchmark data comprises:

Get benchmark data;

According to the sub-data items in the benchmark data and performance data, construct the neurons of the input layer in the initial deep learning network; enter the data corresponding to each sub-data item in the benchmark data into the corresponding neurons in the input layer In the unit, the data corresponding to each sub-data item in the performance data is entered into the corresponding neuron in the input layer;

Calling the hidden layer of the initial deep learning network, obtaining the input vector of the input layer and operating it to obtain an output vector, and outputting the output vector to the output layer of the initial deep learning network;

Calculate the difference between the output vector of the output layer and the abnormal information corresponding to the performance data through the preset training loss function, and obtain the training loss value;

A mature deep learning network is obtained by adjusting the weight and paranoia value of the hidden layer through the back-propagation algorithm and according to the training loss value until the training loss value is less than a preset training loss threshold.
The computer device according to claim 9, wherein after setting the adjustable data as a dynamic index, the method further comprises:

associating and saving the dynamic indicator with the user information of the mobile terminal;

After associating and saving the dynamic indicator with the user information of the mobile terminal, the method further includes:

Upload the dynamic indicator and the user information to the blockchain.
The computer device according to claim 9, wherein the performance index of the monitoring mobile terminal is used as a real-time index, an evaluation result is obtained by evaluating the real-time index according to the dynamic index, and an evaluation result is sent to the The steps of sending the alarm information by the mobile terminal include:

Monitor the performance index of the mobile terminal and use it as a real-time index, and obtain the user information of the mobile terminal and extract the dynamic index associated with the user information;

Set the real-time index whose value exceeds the dynamic index as an abnormal index, and set the number of the abnormal index as an evaluation result; determine whether the evaluation result exceeds a preset abnormal threshold;

If so, send alarm information to the mobile terminal;

If not, end.
A computer-readable storage medium on which a computer program is stored, wherein the computer program stored in the readable storage medium is executed by a processor to implement a performance monitoring and evaluation method

The performance monitoring and evaluation method includes:

Obtain performance data of the mobile terminal according to the abnormal information sent by the mobile terminal, wherein the performance data reflects the performance index of the mobile terminal in an abnormal state;

A mature deep learning network is obtained by training the preset initial deep learning network through the performance data and the preset benchmark data, wherein the benchmark data reflects the performance index of the mobile terminal in a normal state;

Calculate the preset adjustable data through the mature deep learning network to generate the test result, and calculate the difference between the test result and the preset equalization result to obtain the test loss value, and adjust the adjustable data according to the test loss value. Data until the test result generated by the mature deep learning network is consistent with the equilibrium result, and the adjustable data is set as a dynamic index;

The performance index of the mobile terminal is monitored and taken as a real-time index, an evaluation result is obtained by evaluating the real-time index according to the dynamic index, and alarm information is sent to the mobile terminal according to the evaluation result.
The computer-readable storage medium according to claim 15, wherein, before acquiring the performance data of the mobile terminal according to the abnormality information sent by the mobile terminal, the method further comprises:

Create an upload interface for monitoring the abnormal information of the mobile terminal;

Before acquiring the performance data of the mobile terminal according to the abnormal information sent by the mobile terminal, the method further includes:

Send a status request to the mobile terminal, and receive the status information sent by the mobile terminal according to the status request; determine whether the status information has an abnormal label; if so, determine that the status information is abnormal information; if not, then Finish.
The computer-readable storage medium according to claim 15, wherein after acquiring the performance data of the mobile terminal according to the abnormal information sent by the mobile terminal, the method further comprises:

acquiring user information of the mobile terminal, associating and saving the performance data with the user information;

Call the exception database to aggregate the performance data belonging to the same user information to form a performance set, and delete the performance data in the performance set that exceeds the preset time limit;

Calculate the number of performance data in the performance set in the abnormal database, and determine whether the number of performance data reaches a preset training threshold; if so, extract the performance data in the performance set; if not, end.
The computer-readable storage medium according to claim 15, wherein the step of training a preset initial deep learning network to obtain a mature deep learning network by using the performance data and the preset benchmark data comprises:

Get benchmark data;

According to the sub-data items in the benchmark data and performance data, construct the neurons of the input layer in the initial deep learning network; enter the data corresponding to each sub-data item in the benchmark data into the corresponding neurons in the input layer In the unit, the data corresponding to each sub-data item in the performance data is entered into the corresponding neuron in the input layer;

Calling the hidden layer of the initial deep learning network, obtaining the input vector of the input layer and operating it to obtain an output vector, and outputting the output vector to the output layer of the initial deep learning network;

Calculate the difference between the output vector of the output layer and the abnormal information corresponding to the performance data through the preset training loss function, and obtain the training loss value;

A mature deep learning network is obtained by adjusting the weight and paranoia value of the hidden layer through the back-propagation algorithm and according to the training loss value until the training loss value is less than a preset training loss threshold.
The computer-readable storage medium of claim 15, wherein after setting the adjustable data as a dynamic indicator, the method further comprises:

associating and saving the dynamic indicator with the user information of the mobile terminal;

After associating and saving the dynamic indicator with the user information of the mobile terminal, the method further includes:

Upload the dynamic indicator and the user information to the blockchain.
The computer device according to claim 15, wherein the performance index of the monitoring mobile terminal is used as a real-time index, an evaluation result is obtained by evaluating the real-time index according to the dynamic index, and an evaluation result is sent to the The steps of sending the alarm information by the mobile terminal include:

Monitor the performance index of the mobile terminal and use it as a real-time index, and obtain the user information of the mobile terminal and extract the dynamic index associated with the user information;

Set the real-time index whose value exceeds the dynamic index as an abnormal index, and set the number of the abnormal index as an evaluation result; determine whether the evaluation result exceeds a preset abnormal threshold;

If so, send alarm information to the mobile terminal;

If not, end.