Artificial intelligence-based mobile phone application performance automatic test system and method

Abstract

The application discloses a mobile phone application performance automatic test system and method based on artificial intelligence, and relates to the technical field of electric digital data processing.The system comprises a sampling condition monitoring module, an automatic sampling adjustment module and a test feedback module.The application performs automatic test on the mobile phone application performance based on an artificial intelligence algorithm in the current test period, monitors the sampling condition of test data in real time, then determines whether to perform automatic sampling adjustment according to the result of sampling frequency adaptation determination, if yes, monitors the data noise condition of test data after automatic sampling adjustment to perform data noise degree determination, and then performs test feedback, and simultaneously determines whether to perform secondary sampling adjustment, otherwise, continues to monitor the data sampling condition, thereby improving the adaptation degree of the test condition of test data and the sampling frequency, and solving the problem of low adaptation degree of the sampling frequency and the corresponding test condition in the process of testing the mobile phone application performance in the prior art.

Description

Technical Field

[0001] This invention relates to the field of electronic digital data processing technology, and in particular to an automated testing system and method for the performance of mobile applications based on artificial intelligence. Background Technology

[0002] Existing AI-driven mobile phone performance testing systems generally consist of several parts: a data acquisition layer, a scenario-driven layer, an AI (Artificial Intelligence) analysis and modeling layer, and an optimization and feedback layer. The data acquisition layer uses system interfaces, hardware monitoring APIs, and performance monitoring tools provided by Android / iOS (iPhone Operating System) to collect metrics such as CPU (Central Processing Unit), memory, GPU (Graphics Processing Unit), power consumption, FPS (Frames Per Second), and network latency. The scenario-driven layer executes test cases, using automated scripts or simulated user behavior, combined with AI models (reinforcement learning, behavior prediction models) to intelligently generate operation steps, covering more user paths. Common frameworks include Appium, UIAutomator (a user interface automation testing framework provided by Android), and XCUITest (Xcode UI Test, a user interface testing framework provided by Apple). The AI ​​analysis and modeling layer uses AI algorithms to analyze the collected data, including anomaly detection (LSTM-Long Short-Term...). The system includes a Long Short-Term Memory (LSTM) network, an AutoEncoder to capture performance fluctuations, convergence analysis (training the model to evaluate whether performance tuning has stabilized), and root cause localization (using graph models or feature importance analysis to identify performance bottlenecks) to predict application performance under different hardware environments and loads. The tuning and feedback layer automatically generates performance optimization suggestions based on AI analysis results, or directly adjusts sampling frequency, task scheduling strategies, and power control parameters, supporting closed-loop optimization from testing to analysis, then to tuning, and then back to testing.

[0003] For example, Chinese invention patent with publication number CN115145797A discloses a method, apparatus, device, and storage medium for application performance testing, including: determining a preset number of target processes running in the device based on the operating status monitoring information of the device where the application to be tested is located; determining the running occupancy rate of each target process in the device; and obtaining performance test data of the application to be tested.

[0004] For example, Chinese invention patent CN109710521B discloses a multimedia application performance testing method, apparatus, computer device, and storage medium, which includes: obtaining a multimedia application test request, wherein the test request includes an identifier of a target multimedia application; launching the target multimedia application corresponding to the identifier of the target multimedia application on the current device; during the operation of the target multimedia application, obtaining the performance parameters of the current device and the first operation information executed by the target multimedia application; and determining the performance of the target multimedia application on the current device based on the performance parameters of the current device and the first operation information executed by the target multimedia application.

[0005] The above-mentioned technology has at least the following technical problems:

[0006] Due to the complexity of the mobile phone environment, such as system processes, network fluctuations, and background tasks, performance indicators (CPU, power consumption, etc.) fluctuate significantly, causing artificial intelligence to mistakenly perceive abnormal data noise. At the same time, the mobile operating system is highly dynamic, its background processes are uncontrollable, and it also has resource reclamation and scheduling mechanisms, which limits data collection. If the sampling frequency is too high, short-term fluctuations may occur, misleading artificial intelligence to amplify the corresponding anomalies. Conversely, it may miss key anomalies. Therefore, how to match the sampling frequency in the process of testing mobile application performance with the corresponding test conditions is one of the urgent problems to be solved. Summary of the Invention

[0007] To address the technical problem of low adaptability between sampling frequency and corresponding test conditions in existing technologies for testing mobile application performance, this invention provides an automated mobile application performance testing system and method based on artificial intelligence. The technical solution is as follows:

[0008] On one hand, an AI-based automated testing system for mobile application performance is provided. This system includes a sampling monitoring module, an automated sampling adjustment module, and a test feedback module. The sampling monitoring module performs automated testing of mobile application performance based on AI algorithms during the current testing cycle and monitors the sampling status of the corresponding test data in real time to determine the sampling frequency adaptation and assess the degree of fit between the sampling frequency and the current testing conditions. The automated sampling adjustment module determines whether to implement automated sampling adjustments to improve sampling adaptation based on the sampling frequency adaptation results. The test feedback module monitors the data noise of the test data after automated sampling adjustment if such adjustments are implemented, and provides corresponding test feedback after determining the level of data noise. It also decides whether to perform secondary sampling adjustments; otherwise, it continues to monitor the data sampling status of the automated mobile application performance test.

[0009] On the other hand, an automated testing method for mobile application performance based on artificial intelligence is provided. The specific steps are as follows: In the current testing cycle, the performance of the mobile application is automatically tested based on artificial intelligence algorithms, and the sampling status of the corresponding test data is monitored in real time to determine the sampling frequency adaptation and the degree of consistency between the sampling frequency and the current testing status; the result of the sampling frequency adaptation determination determines whether to perform automated sampling adjustment to improve the sampling adaptation; if automated sampling adjustment is performed, the data noise of the test data is monitored after the automated sampling adjustment, and the corresponding test feedback is given after the data noise level is determined, and it is decided whether to perform secondary sampling adjustment; otherwise, the data sampling status of the automated test of mobile application performance continues to be monitored.

[0010] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0011] 1. During the current testing cycle, the system performs automated performance testing of mobile applications based on artificial intelligence algorithms, and monitors the sampling of corresponding test data in real time. This allows for sampling frequency adaptation judgment, determining the degree of consistency between the sampling frequency and the current testing conditions. This helps to more accurately quantify the sampling precision of the test data. The result of the sampling frequency adaptation judgment then determines whether to execute automated sampling adjustment, further improving the fit between the sampling frequency and the testing conditions. If automated sampling adjustment is executed, the system monitors the data noise after adjustment to understand the optimization effect on data noise and continue to improve the optimization. After determining the level of data noise, corresponding test feedback is provided to reflect the sampling status of the test data in a timely manner. Simultaneously, it determines whether to perform secondary sampling adjustment, achieving closed-loop control of the sampling frequency and improving the fit between the test data and the sampling frequency. Otherwise, the system continues to monitor the data sampling status of the automated mobile application performance test, ensuring the continuity of test data sampling monitoring.

[0012] 2. First, test data is collected based on the initial test sampling frequency, and the applicable analysis parameters for the test sampling frequency of the current test cycle are obtained. This helps to understand the current test situation more accurately. Then, the sampling frequency analysis factor is obtained based on the applicable analysis parameters for the test sampling frequency, thereby more accurately quantifying the sampling frequency requirements of the test data in the current test cycle. Then, the corresponding optimized sampling frequency range is obtained based on the sampling frequency analysis factor. This helps to automatically determine whether the sampling frequency is in line with the current test situation. If the initial test sampling frequency belongs to the corresponding optimized sampling frequency range, it means that the current sampling frequency matches the test situation. Then, the sampling accuracy is optimized to further improve the sampling accuracy of the test data. Otherwise, the adaptive automatic sampling adjustment is executed to ensure that the sampling frequency matches the sampling requirements of the test data.

[0013] 3. After automated sampling and adjustment, monitoring the data noise of the test data and obtaining the corresponding data noise evaluation parameters helps to understand the current noise situation more intuitively. Then, the data noise evaluation parameters are compared with the extracted data noise judgment parameters, and the corresponding sampling frequency secondary adjustment score is recorded. This not only realizes the automated judgment of the pass rate of data noise, but also more accurately quantifies the impact of data noise. Finally, the cumulative value of the sampling frequency secondary adjustment score is calculated to determine the degree of data noise, so as to realize the automated decision on whether to perform secondary adjustment. This not only improves the efficiency of system decision-making, but also further improves the sampling accuracy of test data.

[0014] 4. If the initial test sampling frequency is lower than the minimum value of the optimized sampling frequency range, the difference between the initial test sampling frequency and the minimum value of the optimized sampling frequency range is obtained. This quantifies the degree to which the sampling frequency needs to be increased and automatically adjusts the sampling frequency to limit the increase. This not only limits the increase in sampling frequency but also helps ensure the stability of the sampling frequency increase process. If the initial test sampling frequency is higher than the maximum value of the optimized sampling frequency range, the difference between the initial test sampling frequency and the maximum value of the optimized sampling frequency range is obtained. This quantifies the degree to which the sampling frequency needs to be reduced and automatically adjusts the sampling frequency to limit the reduction. This limits the reduction in sampling frequency and thus reduces the impact of sampling frequency adjustment on system stability. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the structure of the AI-based automated testing system for mobile application performance provided in an embodiment of the present invention;

[0017] Figure 2 This is a schematic diagram of the sampling frequency adaptation determination provided in an embodiment of the present invention;

[0018] Figure 3 This is a flowchart illustrating the data noise situation of the monitoring test data provided in an embodiment of the present invention;

[0019] Figure 4 This is a flowchart illustrating the automated performance testing method for mobile applications based on artificial intelligence provided in this embodiment of the invention. Detailed Implementation

[0020] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0021] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0022] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0023] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0024] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0025] This invention provides an automated testing system and method for mobile application performance based on artificial intelligence. For example... Figure 1 The diagram shows the structure of an AI-based automated testing system for mobile application performance. The system includes: a sampling monitoring module, an automated sampling adjustment module, and a test feedback module.

[0026] The sampling monitoring module is used to perform automated testing of mobile application performance based on artificial intelligence algorithms in the current testing cycle, and to monitor the sampling status of the corresponding test data in real time to determine the sampling frequency adaptation and the degree of consistency between the sampling frequency and the current testing status.

[0027] The automated sampling adjustment module is used to determine whether to perform automated sampling adjustment to improve sampling fit based on the specific results of the sampling frequency adaptation judgment.

[0028] The test feedback module is used to monitor the data noise of the test data after automatic sampling adjustment if automatic sampling adjustment is performed, and to provide corresponding test feedback after judging the level of data noise. It also decides whether to perform secondary sampling adjustment. Otherwise, it continues to monitor the data sampling of the automated test of mobile application performance.

[0029] In this embodiment, artificial intelligence algorithms are used to automate the performance testing of mobile applications and monitor the sampling of test data in real time. This allows for dynamic perception of the matching degree between the sampling frequency and the current test scenario, avoiding oversampling (wasting resources) or undersampling (missing performance fluctuations) caused by a fixed sampling frequency. This ensures that the sampling frequency closely matches the actual test requirements. Based on the sampling frequency adaptation judgment, the system determines whether automated sampling adjustment is needed, enabling adaptive adjustment of the sampling frequency. This allows the system to automatically optimize the sampling strategy based on the volatility and stability of performance data, thereby improving sampling accuracy and efficiency and reducing manual intervention costs. After automated sampling adjustment, the noise level of the sampled data continues to be monitored and data noise is judged. Test feedback is then provided, and a decision is made on whether to perform secondary sampling adjustment. Through feedback and further adjustment of data noise, the stability and reliability of the sampled data are ensured, improving the accuracy and reproducibility of performance test results. Simultaneously, excessive redundant sampling is avoided, enhancing the intelligence and robustness of the testing system.

[0030] like Figure 2 The diagram shown is a flowchart illustrating the sampling frequency adaptation determination process provided in this embodiment of the invention. The specific logic is as follows: Test data is collected based on each initial test sampling frequency, and applicable analysis parameters for the current test cycle's sampling frequency are obtained. A sampling frequency analysis factor is obtained based on the applicable analysis parameters, and the corresponding optimized sampling frequency range is output based on the sampling frequency analysis factor. The initial test sampling frequency is matched with the corresponding optimized sampling frequency range: if the initial test sampling frequency belongs to the corresponding optimized sampling frequency range, sampling accuracy optimization is performed to improve the sampling accuracy of the test data; otherwise, adaptive automatic sampling adjustment is executed. The specific content of the sampling accuracy optimization is as follows:

[0031] The median of the optimized sampling frequency range is denoted as the optimized sampling median. The difference between the optimized sampling median and the initial test sampling frequency is quantified to obtain the optimized-initial sampling difference value. The optimized-initial sampling difference value is compared with the set sampling median error range. If the optimized-initial sampling difference value is within the sampling median error range, no sampling accuracy optimization is performed; otherwise, the corresponding optimal sampling frequency adjustment range is obtained by quantizing the distance between the optimized-initial sampling difference value and the optimal sampling frequency. If the optimized-initial sampling difference value is greater than the optimization direction threshold, the adjustment is increased based on the optimal sampling frequency adjustment range; if the optimized-initial sampling difference value is less than the optimization direction threshold, the adjustment is decreased based on the optimal sampling frequency adjustment range. Through the above process, not only is the fit between the test conditions and the data sampling frequency more accurately quantified, but it also helps to adjust for mismatches in a timely manner, thereby improving sampling accuracy.

[0032] Furthermore, the specific process for determining the sampling frequency adaptation is as follows:

[0033] R1 collects test data based on each initial test sampling frequency and obtains the applicable analysis parameters for the test sampling frequency of the current test cycle. The applicable analysis parameters for the test sampling frequency include CPU utilization, memory usage, power consumption, and FPS.

[0034] It should be added that if the phone's operating system is Android, CPU usage can be obtained using the `top / dumpsys cpuinfo` command and `adb shell dumpsys meminfo`.<package_name> Metrics such as PSS, RSS, and Heap are obtained to represent memory usage. Power consumption is obtained via `adb shell dumpsys batterystats`, and FPS is obtained via `adb shell dumpsys gfxinfo`.<package_name> If the phone system is iOS, CPU usage is obtained using the Xcode Instruments command-line tool, memory usage is obtained through performance diagnostics in Xcode Organizer, power consumption is obtained through the power consumption report provided by Xcode, and FPS is obtained through Instruments (Core Animation tool).

[0035] R2, based on the applicable analysis parameters of the test sampling frequency, obtains the sampling frequency analysis factor used to quantify the sampling frequency requirements of the test data in the current test cycle. Based on the sampling frequency analysis factor in a pre-constructed sampling frequency projection sequence used to fit the mapping relationship between the sampling frequency analysis factor and the optimized sampling frequency interval, the corresponding optimized sampling frequency interval is output.

[0036] Specifically, the sampling frequency projection sequence is pre-built in a preset database. The sampling frequency analysis factors are input into the trained sampling frequency projection sequence to output the corresponding optimized sampling frequency range. The training data used in this sampling frequency projection sequence comes from sampling frequency analysis factors acquired in historical time periods, as well as optimized sampling frequency ranges set by professional technicians based on empirical rules. This data is used to fit the mapping relationship between the sampling frequency analysis factors and the optimized sampling frequency ranges, thereby enabling a more accurate analysis of the sampling frequency fit.

[0037] R3 matches the initial test sampling frequency with the corresponding optimized sampling frequency range: if the initial test sampling frequency belongs to the corresponding optimized sampling frequency range, it means that the current sampling frequency matches the test situation, and the sampling accuracy is optimized to improve the sampling accuracy of the test data; otherwise, the adaptive automatic sampling adjustment is executed.

[0038] In this embodiment, by collecting and acquiring analysis parameters (CPU, memory, power consumption, FPS) based on the initial sampling frequency, the rationality of the sampling frequency can be comprehensively evaluated from multiple performance indicators, ensuring that the sampling frequency determination is comprehensive and representative, and avoiding bias caused by a single indicator. At the same time, by analyzing sampling frequency factors and mapping them to the optimized sampling frequency range, and using a pre-constructed sampling frequency projection sequence, a precise mapping between the sampling frequency requirement and the optimized sampling range is achieved. This can adapt to the test requirements under different scenarios, improve the scientificity and intelligence of the determination, match and determine the initial sampling frequency with the optimized range, and quickly determine whether the current sampling frequency is reasonable through a matching mechanism. If they match, only the sampling accuracy is optimized; if they do not match, automatic sampling adjustment is triggered, thereby achieving dynamic matching between the sampling frequency and the test scenario. When matching, the sampling accuracy is further improved; when they do not match, automatic adjustment is performed, ensuring that the sampling can meet the needs of capturing performance fluctuations without wasting resources due to oversampling, thus achieving a balance between test efficiency and sampling accuracy.

[0039] Furthermore, the specific method for obtaining the sampling frequency analysis factor is as follows:

[0040] R21 performs data normalization on the test sampling frequency applicable analysis parameters and extracts the pre-stored test sampling frequency applicable analysis weights, which include CPU utilization weight, memory utilization weight, power consumption weight, and FPS weight.

[0041] It should be explained that the test sampling frequency applicable analysis weight is extracted from the test sampling frequency mapping matching table in the preset database. The test sampling frequency mapping matching table is a data table that reflects the mapping relationship between the test sampling frequency applicable analysis parameters and the corresponding test sampling frequency applicable analysis weight. By inputting the real-time test sampling frequency applicable analysis parameters into the test sampling frequency mapping matching table, the corresponding CPU utilization ratio, memory utilization ratio, power consumption ratio, and FPS ratio can be output. These represent the degree of influence of CPU utilization, memory consumption, power consumption, and FPS on the sampling frequency analysis factor, thereby obtaining a more accurate sampling frequency analysis factor.

[0042] R22 is a sampling frequency analysis factor obtained by weighting the applicable analysis parameters of the test sampling frequency based on the weight of the applicable analysis of the test sampling frequency.

[0043] The specific constraint expression for the sampling frequency analysis factor is as follows:

[0044] ;

[0045] In the formula, x a Indicates CPU utilization, x b Indicates memory usage, x c Indicates power consumption, x d Indicates FPS, i a Indicates the CPU utilization rate, i b Indicates the proportion of memory usage, i c Indicates the power consumption ratio, i d y represents the FPS weight, and y represents the sampling frequency analysis factor.

[0046] In this embodiment, the algorithm combines the applicable analysis parameters of the test sampling frequency with the corresponding weight of the applicable analysis of the test sampling frequency to obtain the sampling frequency analysis factor. In the formula, as the CPU utilization, memory usage, power consumption and FPS increase, it indicates that the test situation in the current test cycle may change rapidly, and the demand for the sampling frequency of the test data may be higher, and the corresponding sampling frequency analysis factor may increase accordingly. At the same time, the applicable analysis parameters of the test sampling frequency are also interrelated. For example, when the CPU frequency increases and the number of cores increases, power consumption will increase because the current and voltage load is greater. When memory is tight, the CPU needs to spend more time on garbage collection, memory allocation or page swapping, resulting in increased CPU utilization. Excessive power consumption will cause the device to heat up, and the temperature control mechanism will trigger thermal throttling, resulting in a decrease in FPS. Furthermore, when memory is insufficient or the CPU is overloaded, the application will stutter, resulting in a decrease in FPS. Through the above analysis, it is helpful to more accurately understand the degree of fit between the test situation and the sampling frequency, so as to take corresponding optimization adjustments based on the analysis results, thereby improving the fit between the sampling requirements of the test data and the actual sampling frequency.

[0047] Furthermore, the specific details of the sampling accuracy optimization are as follows:

[0048] R31, obtain the median of the optimized sampling frequency range and denot it as the optimized sampling median. Quantize the difference between the optimized sampling median and the initial test sampling frequency to obtain the optimized-initial sampling difference value, which represents the subtraction operation between the optimized sampling median and the initial test sampling frequency.

[0049] R32 compares the optimized-initial sampling difference value with the pre-set sampling median error range. If the optimized-initial sampling difference value falls within the sampling median error range, no sampling accuracy optimization is performed. Otherwise, the distance quantization value of the optimized-initial sampling difference value is input into the pre-fitted sampling frequency fitting function to output the corresponding optimal sampling frequency adjustment range. Here, the distance quantization value represents the absolute value of the optimized-initial sampling difference value.

[0050] It needs to be explained that the sampling frequency fitting function is obtained by linearly fitting a linear function based on the sampling frequency fitting training data. The sampling frequency fitting training data includes the distance quantization value of the optimization-initial sampling gap and the optimal sampling frequency adjustment range set by the preset staff. First, a linear function (such as a linear function y=ax+b) is assumed based on the sampling frequency fitting training data. Then, the parameters (i.e., parameters a and b in the aforementioned linear function) are calculated using the least squares method. Finally, the corresponding sampling frequency fitting function is obtained by fitting the distance quantization value of the optimization-initial sampling gap and the optimal sampling frequency adjustment range using numpy.polyfit() and / or scipy.optimize.curve_fit() in Python. This function is used to fit the mapping relationship between the distance quantization value of the optimization-initial sampling gap and the optimal sampling frequency adjustment range.

[0051] R33, if the optimization-initial sampling difference value is greater than the optimization direction threshold value, the initial test sampling frequency will be increased based on the optimal sampling frequency adjustment range; where the optimization direction threshold value is preset by professional test technicians and stored in a preset database, with 0 as the optimization direction threshold value.

[0052] R34, if the optimization-initial sampling difference value is less than the optimization direction critical value, then the initial test sampling frequency will be reduced based on the optimal sampling frequency adjustment range.

[0053] In this embodiment, by calculating the difference between the optimized sampling median and the initial sampling frequency, the deviation of the current sampling frequency from the optimal sampling interval can be accurately quantified, providing a scientific basis for subsequent adjustments. Introducing the sampling median error range for judgment and setting a corresponding error tolerance range can avoid unnecessary frequent adjustments when the difference is small, thereby maintaining system stability and reducing the overhead caused by sampling frequency jitter. Through a pre-fitted sampling frequency function, the difference value is transformed into the optimal adjustment range, achieving precise adjustment of the sampling frequency and ensuring that the adjustment range is neither insufficient nor excessive, thereby improving the accuracy and reliability of sampling optimization. Simultaneously, based on the comparison results between the difference value and the critical value, the sampling frequency is increased or decreased respectively, which can flexibly adapt to different performance testing needs, ensuring that the sampling frequency adjustment is directional and targeted, and improving the effectiveness of the test results.

[0054] Furthermore, the specific process of performing adaptive automated sampling adjustment is as follows:

[0055] On the one hand, if the initial test sampling frequency is lower than the minimum value of the optimized sampling frequency range, the amount of difference to be increased between the initial test sampling frequency and the minimum value of the optimized sampling frequency range, as well as the corresponding increase limit used to limit the increase of the sampling frequency, are obtained for automatic sampling adjustment; where the amount of difference to be increased is the difference between the initial test sampling frequency and the minimum value of the optimized sampling frequency range.

[0056] It should be added that the increase limit is pre-set by professional testing technicians based on the specific requirements of mobile application performance testing and stored in a preset database.

[0057] On the other hand, if the initial test sampling frequency is higher than the maximum value of the optimized sampling frequency range, the amount of the difference to be reduced between the initial test sampling frequency and the maximum value of the optimized sampling frequency range, as well as the corresponding reduction limit used to limit the reduction of the sampling frequency, are obtained for automatic reduction of sampling adjustment; wherein, the amount of the difference to be reduced is the difference between the initial test sampling frequency and the maximum value of the optimized sampling frequency range.

[0058] It should be added that the reduction limit is preset by professional testing technicians based on the specific requirements of mobile application performance testing and stored in a preset database.

[0059] In this embodiment, the increase adjustment when the sampling frequency is lower than the minimum value of the optimization range can automatically calculate the gap and make reasonable increases when the sampling frequency is insufficient to reflect performance fluctuations, avoiding the omission of key performance data due to insufficient sampling, thereby ensuring the integrity and accuracy of performance testing. Conversely, the decrease adjustment when the sampling frequency is higher than the maximum value of the optimization range can automatically identify and reduce the sampling frequency in a controlled manner when excessive sampling frequency causes waste of system resources, avoiding the additional overhead caused by oversampling, thereby improving testing efficiency and system energy efficiency. Through amplitude limit constraints, the sampling frequency is not adjusted too much at once, which would cause instability in the testing process, thus achieving gradual adjustment and improving the safety and robustness of sampling adjustment.

[0060] Furthermore, the specific steps for automating sampling adjustment are as follows:

[0061] If the gap to be improved is not greater than the improvement limit, then in the next test cycle, the initial test sampling frequency will be increased according to the gap to be improved.

[0062] Specifically, the gap to be improved is input into a preset improvement mapping set to output the corresponding improvement magnitude. The improvement magnitude is multiplied by the initial test sampling frequency to obtain the corresponding optimized improvement sampling frequency. Sampling is performed based on the optimized improvement sampling frequency. The improvement mapping set is a dataset that reflects the mapping relationship between the gap to be improved and the improvement magnitude. It is trained using improvement magnitude training data from historical data. The improvement magnitude training data includes the gap to be improved within the historical time period and the improvement magnitude set by professional test technicians based on experience rules, thereby outputting a more accurate improvement magnitude and further adjusting the initial test sampling frequency more precisely.

[0063] T2. If the gap to be improved is greater than the improvement limit, the initial test sampling frequency will be increased according to the improvement limit in the next test cycle, and the difference between the gap to be improved and the improvement limit will be obtained to get the improvement difference value.

[0064] It should be explained that the increase limit is input into a pre-defined increase limit projection sequence to obtain the corresponding increase limit magnitude. The increase limit magnitude is then multiplied by the initial test sampling frequency to obtain the corresponding optimized increase limit sampling frequency. Sampling is performed based on the optimized increase limit sampling frequency. The increase limit projection sequence is pre-built in a preset database. The increase limit is input into the trained increase limit projection sequence to output the corresponding increase limit magnitude. The training data used in this increase limit projection sequence comes from increase limits acquired in historical periods and increase limit magnitudes set by professionals based on empirical rules. This data is used to fit the mapping relationship between the increase limit and the increase limit magnitude, thereby allowing for more precise adjustment of the initial test sampling frequency.

[0065] T3 inputs the increase difference into the test cycle mapping table in the database, which is a mapping table between the increase difference and the test cycle reduction ratio, and outputs the corresponding test cycle reduction ratio.

[0066] It should be noted that the test cycle mapping table is pre-stored in a preset database to reflect the mapping relationship between the improvement difference and the test cycle reduction ratio. It is trained based on test cycle reduction training data, which includes the improvement difference within a preset time period in historical data, as well as the test cycle reduction ratio set by professional test technicians based on experience rules. In use, the real-time improvement difference is input into the test cycle mapping table, and the corresponding test cycle reduction ratio can be output, thus realizing the reduction of the test cycle.

[0067] T4 reduces the test cycle for automated testing of mobile application performance based on artificial intelligence by reducing the test cycle ratio; the reduction process means multiplying the test cycle reduction ratio by the test cycle.

[0068] In this embodiment, when the difference to be improved does not exceed the amplitude limit, an equal increase can be made directly in the next test cycle, ensuring the rapid responsiveness and real-time performance of the sampling frequency adjustment and improving the sensitivity of the test. Furthermore, when the difference to be improved exceeds the amplitude limit, a limited increase is adopted and the remaining difference is recorded to avoid system instability caused by excessively large one-time adjustments to the sampling frequency. Simultaneously, the continuity and controllability of the adjustment are maintained through the accumulation of the difference. At the same time, a preset difference-test cycle reduction ratio mapping table is used to convert the remaining improvement difference into a shortening of the test cycle, thereby accelerating test convergence and ensuring that even when the adjustment is insufficient, more refined data can still be obtained through a higher sampling frequency. Finally, by reducing the test cycle, the AI-based automated performance test can run at a higher frequency, improving sampling accuracy and the ability to capture performance fluctuations, thus improving the accuracy and robustness of the test results.

[0069] Furthermore, the specific steps for automating sampling adjustment are as follows:

[0070] D1. If the amount of difference to be reduced is not greater than the reduction limit, the initial test sampling frequency will be reduced according to the amount of difference to be reduced in the next test cycle.

[0071] Specifically, the amount of difference to be reduced is input into a preset reduction mapping set to output the corresponding reduction magnitude. The reduction magnitude is then multiplied by the initial test sampling frequency to obtain the corresponding optimized reduction sampling frequency. Sampling is performed based on the optimized reduction sampling frequency. The reduction mapping set is a dataset that reflects the mapping relationship between the amount of difference to be reduced and the reduction magnitude. The reduction magnitude training data is used in historical data. The reduction magnitude training data includes the amount of difference to be reduced within the historical time period and the reduction magnitude set by professional test technicians based on experience rules, thereby outputting a more accurate reduction magnitude and further adjusting the initial test sampling frequency more precisely.

[0072] D2. If the amount of difference to be reduced is greater than the reduction limit, the initial test sampling frequency will be reduced according to the reduction limit in the next test cycle, and the difference between the amount of difference to be reduced and the reduction limit will be obtained as the reduction difference.

[0073] It should be explained that the reduction limit is input into a pre-defined reduction limit projection sequence to obtain the corresponding reduction limit magnitude. The reduction limit magnitude is then multiplied by the initial test sampling frequency to obtain the corresponding optimized reduction limit sampling frequency. Sampling is performed based on this optimized reduction limit sampling frequency. The reduction limit projection sequence is pre-built in a preset database. The reduction limit is input into the trained reduction limit projection sequence to output the corresponding reduction limit magnitude. The training data used in this reduction limit projection sequence comes from reduction limits acquired in historical periods and reduction limit magnitudes set by professionals based on empirical rules. This data is used to fit the mapping relationship between the reduction limit and the reduction limit magnitude, thereby allowing for more precise adjustment of the initial test sampling frequency.

[0074] D3 inputs the reduction difference into the test cycle mapping table of the reduction difference and the test cycle amplification ratio in the database, and outputs the corresponding test cycle amplification ratio.

[0075] It should be noted that the test cycle mapping table is pre-stored in a preset database to reflect the mapping relationship between the reduction difference and the test cycle reduction ratio. It is trained based on test cycle reduction training data, which includes the reduction difference within a preset time period in historical data, as well as the test cycle reduction ratio set by professional test technicians based on experience rules. In use, the real-time reduction difference is input into the test cycle mapping table, and the corresponding test cycle reduction ratio can be output, thus realizing the reduction of the test cycle.

[0076] D4 amplifies the test cycle for automated testing of mobile application performance based on artificial intelligence by increasing the test cycle amplification ratio; the amplification process means multiplying the test cycle amplification ratio by the test cycle.

[0077] In this embodiment, when the amount of difference to be reduced does not exceed the reduction limit, the same amount can be reduced directly in the next test cycle, ensuring rapid adaptability of the sampling frequency and avoiding resource waste caused by oversampling. Furthermore, when the amount of difference to be reduced exceeds the reduction limit, a limited reduction method is adopted, retaining the remaining difference to avoid incomplete performance fluctuation capture caused by a one-time excessive reduction in the sampling frequency, while ensuring the gradualness and stability of the sampling adjustment. Then, through a preset difference-test cycle amplification ratio mapping table, the undigested reduction difference is transformed into an extension of the test cycle, achieving dual dynamic optimization of the test cycle and sampling frequency. This reduces sampling overhead while maintaining the integrity of data acquisition. By extending the test cycle, AI-based automated performance testing can still maintain a grasp of performance trends even at lower sampling frequencies, thereby improving resource utilization in the testing process and ultimately increasing the sampling accuracy of the test data.

[0078] like Figure 3 The diagram illustrates a flowchart of the process for monitoring data noise in test data according to an embodiment of the present invention. The specific logic is as follows: The data noise of the test data is monitored, and corresponding data noise evaluation parameters are obtained. These parameters are compared with extracted data noise judgment parameters. If the variance is greater than the maximum variance limit, the secondary adjustment score of the sampling frequency is recorded as an effective adjustment value; otherwise, it is recorded as an invalid adjustment value. If the coefficient of variation is greater than the maximum coefficient of variation limit, the secondary adjustment score of the sampling frequency is recorded as an effective adjustment value; otherwise, it is recorded as an invalid adjustment value. If the signal-to-noise ratio is less than the maximum signal-to-noise ratio limit, the secondary adjustment score of the sampling frequency is recorded as an effective adjustment value; otherwise, it is recorded as an invalid adjustment value. The cumulative value of the secondary adjustment score of the sampling frequency is statistically analyzed to determine the degree of data noise. Through this process, not only is the optimized data noise more accurately quantified, but it also helps the system to make more efficient automated decisions for secondary adjustment.

[0079] Furthermore, the specific process for monitoring the data noise of the test data is as follows:

[0080] The first step is to monitor the data noise of the test data and obtain the corresponding data noise evaluation parameters, including variance, coefficient of variation, and signal-to-noise ratio.

[0081] Specifically, the mean of a sampled data set (e.g., 60 FPS samples within 1 second) is calculated, and then the sum of squared deviations is calculated to obtain the corresponding variance. In AI systems, a sliding window variance is generally used. Based on the variance calculation, the standard deviation is divided by the mean to obtain the corresponding coefficient of variation. Finally, the mean of the collected indicators is used as the signal strength, and the standard deviation is used as the noise strength. The signal strength and noise strength are then compared to obtain the corresponding signal-to-noise ratio.

[0082] The second step is to compare the data noise assessment parameters with the extracted data noise judgment parameters, which include the maximum limit of variance, the maximum limit of coefficient of variation, and the minimum limit of signal-to-noise ratio.

[0083] It should be noted that the data noise judgment parameters are extracted from a preset database and are usually preset by professional test technicians based on experience rules.

[0084] The third step is to record the second adjustment fraction of the sampling frequency as the effective adjustment value if the variance is greater than the maximum limit value of variance; otherwise, record the second adjustment fraction of the sampling frequency as the invalid adjustment value. The effective adjustment value is usually 1, and the invalid adjustment value is usually 0.

[0085] It should be noted that a larger variance indicates that the data noise fluctuation may be greater. When the variance is greater than the variance limit, it means that the sampling frequency may need to be adjusted. The corresponding second adjustment fraction of the sampling frequency is the effective adjustment value. On the other hand, if the variance is not greater than the maximum variance limit, it means that no adjustment may be needed. The corresponding second adjustment fraction of the sampling frequency is the ineffective adjustment value.

[0086] Fourth step: If the coefficient of variation is greater than the maximum limit of the coefficient of variation, then record the second adjustment fraction of the sampling frequency as the effective adjustment value; otherwise, record the second adjustment fraction of the sampling frequency as the invalid adjustment value.

[0087] Fifth step: If the signal-to-noise ratio is less than the maximum limit of the signal-to-noise ratio, then record the second adjustment fraction of the sampling frequency as the effective adjustment value; otherwise, record the second adjustment fraction of the sampling frequency as the invalid adjustment value.

[0088] The sixth step is to determine the level of data noise by calculating the cumulative value of the second adjustment fraction of the sampling frequency.

[0089] In this embodiment, by simultaneously introducing three types of parameters—variance, coefficient of variation, and signal-to-noise ratio—a comprehensive assessment of the noise level of the sampled data can be conducted from different perspectives, avoiding bias caused by a single indicator and thus improving the comprehensiveness and reliability of noise assessment. By comparing the assessment parameters with the maximum / minimum limit values, quantitative noise assessment can be achieved, enhancing the scientific rigor and objectivity of the assessment process. Furthermore, the use of a secondary adjustment score valid / invalid labeling method can clearly quantify the impact of different noise assessment results on sampling frequency adjustment, ensuring the operability and traceability of the adjustment assessment. By accumulating the results of multiple indicators, a multi-factor comprehensive assessment can be achieved, helping to reduce the risk of misjudgment caused by anomalies in a single indicator, thereby improving the stability and accuracy of data noise level assessment.

[0090] Furthermore, the specific process for determining the level of data noise is as follows:

[0091] If the cumulative value of the secondary adjustment score of the sampling frequency is greater than the preset adjustment score threshold, the initial test sampling frequency is adjusted a second time. The secondary adjustment ratio is obtained by mapping the current cumulative value of the secondary adjustment score of the sampling frequency. The initial test sampling frequency is adjusted again by the secondary adjustment ratio, and the data noise of the current test cycle is fed back.

[0092] The adjustment score threshold is preset by professional testing technicians, who typically set it based on empirical rules and standard requirements for sampling frequency. The secondary adjustment ratio is obtained by matching from a pre-set secondary adjustment mapping table. By inputting the cumulative value of the secondary adjustment score at the sampling frequency into the mapping table, the corresponding secondary adjustment ratio is output. Sampling is then performed based on the product of the secondary adjustment ratio and the initial test sampling frequency. The secondary adjustment mapping table is a data table used to fit the mapping relationship between the cumulative value of the secondary adjustment score at the sampling frequency and the secondary adjustment ratio. This mapping table is obtained after training with secondary adjustment training data, which includes the cumulative value of the secondary adjustment score at the sampling frequency acquired within a historical time period, and the secondary adjustment ratio set by professional testing technicians based on empirical rules.

[0093] If the cumulative value of the secondary adjustment score of the sampling frequency is not greater than the preset adjustment score threshold, no secondary adjustment will be performed, and the automated testing process of the mobile application performance will continue to be monitored in the next test cycle, and the data noise of the current test cycle will be reported as qualified.

[0094] In this embodiment, by comparing the cumulative value of the secondary adjustment fraction of the sampling frequency with a preset threshold, quantitative threshold control of data noise can be achieved, avoiding unnecessary frequent adjustments and ensuring the stability of system operation. Furthermore, when the noise level exceeds the threshold, the sampling frequency is adjusted by the secondary adjustment ratio obtained by mapping the cumulative fraction value, ensuring that the adjustment amplitude is proportional to the severity of the noise, achieving targeted and adaptive optimization. At the same time, feedback on the severity of noise after secondary adjustment, or feedback on noise compliance when the threshold is not exceeded, can provide a real-time monitoring and feedback closed loop for subsequent automated testing, improving the controllability and traceability of the testing process. Even if secondary adjustment is not triggered, subsequent testing cycles continue to be monitored, achieving dynamic and continuous monitoring of the testing process and ensuring that the sampling frequency can maintain a long-term match with the testing scenario.

[0095] like Figure 4 As shown, Figure 4 This is a flowchart illustrating the automated performance testing method for mobile applications based on artificial intelligence provided in an embodiment of the present invention. The specific steps are as follows:

[0096] In the current testing cycle, the performance of mobile applications is automatically tested based on artificial intelligence algorithms, and the sampling of corresponding test data is monitored in real time to determine the sampling frequency adaptation and the degree of consistency between the sampling frequency and the current testing conditions.

[0097] The specific results of the sampling frequency adaptation determination determine whether to implement automated sampling adjustment to improve sampling adaptation.

[0098] If automated sampling adjustment is performed, the data noise of the test data will be monitored after the automated sampling adjustment, and corresponding test feedback will be given after the data noise level is determined. At the same time, it will be decided whether to perform secondary sampling adjustment. Otherwise, the data sampling of the automated test of mobile application performance will continue to be monitored.

[0099] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the flow or function according to the embodiments of the present invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. A computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0100] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0101] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0102] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0103] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0104] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0105] In the embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0106] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0107] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0108] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0109] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An automated performance testing system for mobile applications based on artificial intelligence, characterized in that, include: Sampling monitoring module, automated sampling adjustment module, and test feedback module; The sampling monitoring module is used to perform automated testing of mobile application performance based on artificial intelligence algorithms in the current testing cycle, and to monitor the sampling status of the corresponding test data in real time to determine the sampling frequency adaptation and the degree of consistency between the sampling frequency and the current testing status. The specific process for determining the sampling frequency adaptation is as follows: Test data is collected based on each initial test sampling frequency, and applicable analysis parameters for the current test cycle's sampling frequency are obtained. These parameters include CPU utilization, memory usage, power consumption, and FPS. A sampling frequency analysis factor is obtained based on the applicable analysis parameters to quantify the sampling frequency requirements of the test data in the current test cycle. Based on the sampling frequency analysis factor, a pre-constructed sampling frequency projection sequence is used to fit the mapping relationship between the sampling frequency analysis factor and the optimized sampling frequency interval, and the corresponding optimized sampling frequency interval is output. The initial test sampling frequency is matched with the corresponding optimized sampling frequency interval: if the initial test sampling frequency belongs to the corresponding optimized sampling frequency interval, sampling accuracy optimization is performed to improve the sampling accuracy of the test data; otherwise, adaptive automatic sampling adjustment is executed. The specific content of the sampling accuracy optimization is as follows: The median of the optimized sampling frequency range is obtained and recorded as the optimized sampling median. The difference between the optimized sampling median and the initial test sampling frequency is quantized to obtain the optimized-initial sampling gap value. The optimized-initial sampling gap value is compared with the set sampling median error range. If the optimized-initial sampling gap value falls within the sampling median error range, no sampling accuracy optimization is performed; otherwise, the distance quantization value of the optimized-initial sampling gap value is input to a pre-fitted sampling frequency fitting function to output the corresponding optimal sampling frequency adjustment range. If the optimized-initial sampling gap value is greater than the optimization direction threshold, the initial test sampling frequency is increased based on the optimal sampling frequency adjustment range. If the optimized-initial sampling gap value is less than the optimization direction threshold, the initial test sampling frequency is decreased based on the optimal sampling frequency adjustment range. The automated sampling adjustment module is used to determine whether to perform automated sampling adjustment to improve sampling adaptability based on the result of the sampling frequency adaptation judgment. The test feedback module is used to monitor the data noise of the test data after automatic sampling adjustment if automatic sampling adjustment is performed, and to provide corresponding test feedback after judging the degree of data noise. At the same time, it decides whether to perform secondary sampling adjustment. Otherwise, it continues to monitor the data sampling of the automatic test of mobile application performance.

2. The automated performance testing system for mobile applications based on artificial intelligence according to claim 1, characterized in that, The specific method for obtaining the sampling frequency analysis factor is as follows: The test sampling frequency applicable analysis parameters are normalized, and the pre-stored test sampling frequency applicable analysis weights are extracted. The test sampling frequency applicable analysis weights include CPU utilization weight, memory utilization weight, power consumption weight, and FPS weight. The sampling frequency analysis factor is obtained by weighting the applicable analysis parameters of the test sampling frequency based on the weight of the applicable analysis proportion of the test sampling frequency.

3. The automated performance testing system for mobile applications based on artificial intelligence according to claim 1, characterized in that, The specific process of performing the adaptive automated sampling adjustment is as follows: If the initial test sampling frequency is lower than the minimum value of the optimized sampling frequency range, the difference between the initial test sampling frequency and the minimum value of the optimized sampling frequency range to be improved, as well as the corresponding improvement limit used to limit the improvement of the sampling frequency, are obtained for automatic improvement of sampling adjustment. If the initial test sampling frequency is higher than the maximum value of the optimized sampling frequency range, the amount of the difference between the initial test sampling frequency and the maximum value of the optimized sampling frequency range to be reduced, as well as the corresponding reduction limit used to limit the reduction of the sampling frequency, are obtained for automatic reduction of sampling adjustment.

4. The automated performance testing system for mobile applications based on artificial intelligence according to claim 3, characterized in that, The specific steps for automating and improving sampling adjustment are as follows: If the gap to be improved is not greater than the improvement limit, the initial test sampling frequency will be increased according to the gap to be improved in the next test cycle. If the gap to be improved is greater than the improvement limit, the initial test sampling frequency will be increased according to the improvement limit in the next test cycle, and the difference between the gap to be improved and the improvement limit will be obtained to obtain the improvement difference. The increase difference is input into a test cycle mapping table that establishes the increase difference and the test cycle reduction ratio, so as to output the corresponding test cycle reduction ratio; The test cycle is reduced by shortening the test cycle ratio.

5. The automated performance testing system for mobile applications based on artificial intelligence according to claim 3, characterized in that, The specific steps for the automated reduction sampling adjustment are as follows: If the amount of difference to be reduced is not greater than the reduction limit, the initial test sampling frequency will be reduced according to the amount of difference to be reduced in the next test cycle. If the amount of difference to be reduced is greater than the reduction limit, the initial test sampling frequency will be reduced according to the reduction limit in the next test cycle, and the difference between the amount of difference to be reduced and the reduction limit will be obtained as the reduction difference. Input the reduction difference into the test cycle mapping table that establishes the reduction difference and the test cycle amplification ratio, and output the corresponding test cycle amplification ratio. The test cycle is amplified by increasing the test cycle ratio.

6. The automated performance testing system for mobile applications based on artificial intelligence according to claim 1, characterized in that, The specific process for monitoring the data noise of the test data is as follows: Monitor the data noise of the test data and obtain the corresponding data noise evaluation parameters, including variance, coefficient of variation and signal-to-noise ratio; The data noise assessment parameters are compared with the extracted data noise judgment parameters, which include the maximum limit of variance, the maximum limit of coefficient of variation, and the minimum limit of signal-to-noise ratio. If the variance is greater than the maximum variance limit, the second adjustment fraction of the sampling frequency is recorded as an effective adjustment value; otherwise, the second adjustment fraction of the sampling frequency is recorded as an invalid adjustment value. If the coefficient of variation is greater than the maximum limit of the coefficient of variation, then the second adjustment fraction of the sampling frequency is recorded as an effective adjustment value; otherwise, the second adjustment fraction of the sampling frequency is recorded as an invalid adjustment value. If the signal-to-noise ratio is less than the maximum limit of the signal-to-noise ratio, then the second adjustment fraction of the sampling frequency is recorded as an effective adjustment value; otherwise, the second adjustment fraction of the sampling frequency is recorded as an invalid adjustment value. The degree of data noise is determined by the cumulative value of the second adjustment of the statistical sampling frequency.

7. The automated performance testing system for mobile applications based on artificial intelligence according to claim 6, characterized in that, The specific process for determining the level of data noise is as follows: If the cumulative value of the secondary adjustment score of the sampling frequency is greater than the preset adjustment score threshold, the initial test sampling frequency is adjusted a second time. The secondary adjustment ratio is obtained by mapping the current cumulative value of the secondary adjustment score of the sampling frequency. The initial test sampling frequency is adjusted again by the secondary adjustment ratio, and the data noise of the current test cycle is fed back. If the cumulative value of the secondary adjustment score of the sampling frequency is not greater than the preset adjustment score threshold, no secondary adjustment will be performed, and the automated testing process of the mobile application performance will continue to be monitored in the next test cycle, and the data noise of the current test cycle will be reported as qualified.

8. An automated testing method for mobile application performance based on artificial intelligence, applied to the automated testing system for mobile application performance based on artificial intelligence as described in any one of claims 1-7, characterized in that, The specific steps are as follows: In the current testing cycle, the performance of mobile applications is automatically tested based on artificial intelligence algorithms, and the sampling of corresponding test data is monitored in real time to determine the sampling frequency adaptation and the degree of consistency between the sampling frequency and the current testing conditions. The result of the sampling frequency adaptation determination determines whether to implement automated sampling adjustment to improve sampling adaptation. If automated sampling adjustment is performed, the data noise of the test data will be monitored after the automated sampling adjustment, and corresponding test feedback will be given after the data noise level is determined. At the same time, it will be decided whether to perform secondary sampling adjustment. Otherwise, the data sampling of the automated test of mobile application performance will continue to be monitored.

Images