Software compatibility test system and method based on AI

Abstract

The invention discloses an AI-based software compatibility test system and method, and belongs to the field of software testing, and the method specifically comprises the steps: collecting historical test data, market demands and product update information, predicting compatibility test demands by using a machine learning algorithm, automatically distributing and scheduling test tasks, monitoring a test environment in real time, and automatically configuring the test environment; different use scenes are simulated, when an automatic test script is executed, the progress is monitored in real time, abnormity is processed, and the system generates a detailed test report by learning a test rule, positioning a compatibility problem, adjusting a test strategy and optimizing product configuration, so that the test efficiency and accuracy are improved, and intelligent management of the test process is realized.

Description

Technical Field

[0001] The present invention belongs to the field of software testing, and particularly relates to an AI-based software compatibility testing system and method. Background Art

[0002] With the diversification and complexity of software applications, software compatibility testing has become increasingly important. Traditional compatibility testing methods usually rely on manual testing or automated testing based on fixed rules, which are inefficient and difficult to handle complex testing scenarios. In the prior art, there is a lack of a compatibility testing method that can intelligently analyze software behavior, automatically generate test cases, and dynamically adjust test strategies.

[0003] For example, Chinese Patent Application with Publication No. CN116257441A discloses a configuration software compatibility testing method and system, including: S100: establishing a configuration software compatibility testing design model; S200: starting from two dimensions of compatibility result impact, test cycle, and resource requirements, establishing a two-dimensional distribution map of test content that affects compatibility testing quality and efficiency; obtaining key test content with large compatibility result impact, long test cycle, and high resource requirements in the two-dimensional distribution map; S300: performing gray-box automated testing based on the underlying design logic for the compatibility testing of new and old versions of configuration software and the compatibility testing of configuration software and supporting industrial application software, which can comprehensively, systematically, accurately, and efficiently carry out the compatibility testing of industrial configuration software, improve the quality of compatibility testing of software products, and avoid unpredictable serious consequences caused by software upgrades on-site.

[0004] For example, Chinese Patent with Authorization Announcement No. CN111459729B discloses a server compatibility testing and analysis method and system, which includes: obtaining basic data required for server compatibility analysis; performing compatibility analysis on the obtained basic data according to a pre-configured dependency and mutual exclusion relationship table between product configuration and backplane and a dependency and mutual exclusion relationship table between product configuration and components to generate compatibility test configuration data; sending the generated compatibility test configuration data to the tester's terminal, so as to greatly ensure the accuracy of test configuration and improve test efficiency through data collation and system analysis, and the test results of each configuration are clearly visible and highly traceable.

[0005] The above prior arts all have the following problems: 1) Although the two-dimensional distribution map considers two dimensions of compatibility result impact, test cycle, and resource requirements, it ignores the difficulty of technical implementation and the stability factor of the test environment; 2) It cannot fully cover all test scenarios and abnormal situations, and has limitations. Summary of the Invention

[0006] In view of the deficiencies of the prior art, the present invention proposes an AI-based software compatibility testing system and method. The testing management module predicts compatibility testing requirements and automatically assigns tasks. The task scheduling unit formulates an execution plan in combination with resource availability. The adaptive testing environment module monitors the environmental status in real time and automatically configures and simulates different scenarios. The test execution module executes automated scripts, providing real-time feedback on progress and exceptions. The intelligent learning unit extracts compatibility features and testing rules through machine learning. The data analysis and problem identification unit identifies potential problems. The strategy adjustment unit optimizes the testing strategy and product configuration. The performance optimization unit provides performance suggestions. The resource management module monitors the device status, automatically schedules devices, and backs up data. Finally, the report generation unit generates a detailed test report, improving the testing efficiency and accuracy.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] An AI-based software compatibility testing system, comprising: a testing management module, an adaptive testing environment module, a test execution module, an intelligent analysis and optimization module, a resource management module, and a user interaction module;

[0009] The testing management module is responsible for the planning, organization, and monitoring of the entire testing process;

[0010] The adaptive testing environment module is used to automatically configure and simulate a testing environment according to testing requirements;

[0011] The test execution module is used to automatically or manually execute compatibility testing tasks;

[0012] The intelligent analysis and optimization module is used to analyze the test results and propose optimization suggestions;

[0013] The resource management module is used to manage various resources involved in the testing process;

[0014] The user interaction module is used to provide a user interface and interaction functions.

[0015] Specifically, the testing management module includes: a requirement prediction and analysis unit, a task assignment unit, a task scheduling unit, and a progress monitoring unit. The adaptive testing environment module includes: an environment perception unit, an automatic configuration unit, and a scenario simulation unit. The test execution module includes: an automated testing unit and an exception detection and handling unit;

[0016] The demand forecasting and analysis unit is used to forecast and analyze test requirements; the task assignment unit is used to reasonably assign test tasks to testers or test teams according to the test requirements; the task scheduling unit is used to schedule and manage the already assigned tasks; the progress monitoring unit is used to monitor the execution status of test tasks in real time; the environment perception unit is used to perceive and collect relevant information about the test environment; the automatic configuration unit is used to automatically configure the test environment according to the information provided by the environment perception unit; the scenario simulation unit is used to simulate various actual usage scenarios to test the compatibility and stability of the product in different scenarios; the automated testing unit is used to be responsible for automatically executing test cases; the anomaly detection and handling unit monitors the test execution process in real time to detect and handle anomalies that occur during the test.

[0017] Specifically, the intelligent analysis and optimization module includes: an intelligent learning unit, a data analysis unit, a problem identification unit, a strategy adjustment unit, and a performance optimization unit, and the resource management module includes: a data storage unit, a device status monitoring unit, a device scheduling unit, a data backup unit, and a report generation unit;

[0018] The intelligent learning unit is used to learn and analyze the test historical data using machine learning algorithms to extract feature information and patterns; the data analysis unit is used to comprehensively and deeply analyze the data generated during the test process to extract key metrics and trends; the problem identification unit is used to perform pattern matching and anomaly detection on the test data using preset rules and algorithms according to the results of the data analysis to automatically identify compatibility problems and performance bottleneck problems that occur during the test; the strategy adjustment unit is used to automatically adjust the test strategy and optimize the test resource allocation according to the results of the problem identification; the performance optimization unit is used to provide optimization suggestions and solutions for the performance problems found during the test; the data storage unit is used to store all the data generated during the test process; the device status monitoring unit is used to monitor the status of test devices in real time; the device scheduling unit is used to automatically schedule and manage test devices according to the test requirements; the data backup unit is used to regularly back up the test data; the report generation unit is used to generate a detailed test report according to the test results and data.

[0019] Specifically, the task scheduling unit adopts the critical path strategy, and the specific steps include:

[0020] A1: Develop a project plan based on the test tasks assigned by the task assignment unit and split the project to obtain manageable sub-projects , where i represents the index of the split sub-project;

[0021] A2: Draw a project network diagram, where each node represents a sub-project, the arrows between nodes represent the dependencies between sub-projects, and use the three-point estimation method to estimate each sub-project to determine the time required for completion;

[0022] A3: Starting from the start node of the project, calculate the earliest start time of each sub-project in turn according to the time required for completion and the dependencies of the sub-projects and the earliest finish time , and starting from the end node of the project, calculate the latest start time of each sub-project by reverse deduction and the latest finish time ;

[0023] A4: According to , , , , calculate the float time of each sub-project , if , then reorganize the sub-project corresponding to to obtain the critical path.

[0024] Specifically, the automated test unit executes the automated test script in a concurrent execution and data-driven execution manner.

[0025] Specifically, the anomaly detection and handling unit adopts a statistics-based anomaly detection strategy, and the specific steps include:

[0026] B1: Perform missing value processing and preliminary outlier screening on the data during the test execution process to obtain a new data set , and extract features from the new data set , where j represents the number of data in the new data set, represents the number of extracted features, represents the jth new test execution data, represents the th extracted feature;

[0027] B2: Establish a statistical model based on the extracted features and perform parameter estimation;

[0028] B3: Set a threshold H to detect outliers in the statistical model. If , then this data is an outlier. If , then this data is a normal value;

[0029] B4: Evaluate the detected outliers to determine whether they are true outliers. If they are true outliers, analyze the causes of the outliers through data validation, and perform data repair or replacement and model adjustment. If they are not true outliers, perform algorithm optimization, reselect or improve features to obtain a new dataset with all normal values. , where represents the k-th normal data, and k represents the number of normal data.

[0030] Specifically, the features extracted in B1 include: graph features, composite features, statistical features, model-based features, spatial features, and text features.

[0031] Specifically, the intelligent learning unit adopts a data exploration and pattern discovery strategy, and the specific steps include:

[0032] C1: Receive the new dataset with all normal values obtained through the anomaly detection and processing unit , and combine with machine learning algorithms to automatically identify and select box plots and scatter plot matrices to display the distribution and relationship of the data;

[0033] C2: Use unsupervised learning algorithms to automatically determine the optimal number of clusters or reduce the dimension, perform cluster analysis and dimensionality reduction analysis, introduce a deep learning model, combine non-linear mapping and feature transformation to capture complex structures and hidden patterns in the data, and automatically adjust the number of clusters or reduce the dimension according to the characteristics of the data;

[0034] C3: Establish a pattern interpretation model based on deep learning to automatically extract and interpret key patterns and features in the data.

[0035] AI-based software compatibility testing method, including:

[0036] S1: Input historical test data and perform preprocessing; the historical test data includes compatibility test results, error logs, and performance data of the product in different environments and configurations;

[0037] S2: Use machine learning algorithms to perform data analysis on the preprocessed historical test data to predict compatibility test requirements;

[0038] S3: According to the predicted compatibility test requirements, automatically allocate test tasks, and formulate a project plan and a task scheduling plan; the project plan includes the splitting of test tasks, determination of dependencies between sub-projects, and time estimation;

[0039] S4: According to the project plan, adopt the critical path strategy to formulate a task scheduling plan, determine the start and end times of each sub-project, and automatically configure the test environment according to the test requirements;

[0040] S5: In the test environment, automatically execute test cases, record test results, and monitor and handle abnormal situations in real time during the test process;

[0041] S6: Analyze the test results, put forward optimization suggestions, adjust the test strategy and product configuration according to the optimization suggestions, and generate and output a test report.

[0042] Compared with the prior art, the beneficial effects of the present invention are:

[0043] 1. The present invention proposes a software compatibility testing system based on AI, and has optimized improvements in the architecture, operation steps and processes. The system has the advantages of simple process, low investment and operation costs, and low production work costs.

[0044] 2. The present invention proposes a software compatibility testing system based on AI. The intelligent learning unit uses machine learning technology to extract compatibility features and testing rules, making the testing process more intelligent. The adaptive test environment module can monitor the environmental status in real time and automatically configure, simulate different scenarios for testing, enabling the system to perform tests in multiple environments, improving the comprehensiveness and accuracy of the tests. The test execution module can provide real-time feedback on progress and anomalies, facilitating testers to timely understand the test situation and adjust the test strategy.

[0045] 3. The present invention proposes a software compatibility testing method based on AI. Through automated and intelligent task allocation and testing processes, it reduces manual operations and interventions, improves the automation level and testing efficiency of the testing process. Through accurate problem identification and solution, it improves the quality of the tests, ensures the compatibility and stability of the product. At the same time, by optimizing resource management and improving testing efficiency, it reduces the human, material and time costs required for testing. By solving compatibility problems, it improves the running stability and performance of the product on different platforms, devices and environments, thereby enhancing the user experience and market competitiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is the architecture diagram of the software compatibility testing system based on AI of the present invention;

[0047] Figure 2 It is the working flow diagram of the test management module of the software compatibility testing system based on AI of the present invention;

[0048] Figure 3 It is the working flow diagram of the intelligent analysis and optimization module of the software compatibility testing system based on AI of the present invention;

[0049] Figure 4 It is the algorithm flow diagram of the abnormal detection and handling unit of the software compatibility testing system based on AI of the present invention. Detailed implementation manners

[0050] Embodiment 1

[0051] Please refer to Figure 1 - Figure 2 , an embodiment provided by the present invention: an AI-based software compatibility testing system, including:

[0052] A test management module, an adaptive test environment module, a test execution module, an intelligent analysis and optimization module, a resource management module, and a user interaction module;

[0053] The test management module is responsible for planning, organizing, and monitoring the entire test process;

[0054] The adaptive test environment module is used to automatically configure and simulate a test environment according to test requirements;

[0055] The test execution module is used to automatically or manually execute compatibility test tasks;

[0056] The intelligent analysis and optimization module is used to analyze test results and propose optimization suggestions;

[0057] The resource management module is used to manage various resources involved in the test process;

[0058] The user interaction module is used to provide a user interface and interaction functions.

[0059] The test management module includes: a requirement prediction and analysis unit, a task assignment unit, a task scheduling unit, and a progress monitoring unit. The adaptive test environment module includes: an environment perception unit, an automatic configuration unit, and a scenario simulation unit. The test execution module includes: an automated test unit and an exception detection and handling unit;

[0060] The requirement prediction and analysis unit is used to predict and analyze test requirements to ensure the accuracy of test objectives and scopes. The task assignment unit is used to reasonably assign test tasks to testers or test teams according to test requirements, considering the skills, experience of testers, as well as the complexity and urgency of tasks, to ensure the fairness and efficiency of task assignment. The task scheduling unit is used to schedule and manage the already assigned tasks, dynamically adjust the task execution order according to factors such as test progress and resource status, handle the dependencies between tasks, avoid task conflicts and delays, ensure the continuity and stability of the test work, and improve the overall efficiency of the test. The progress monitoring unit is used to monitor the execution status of test tasks in real time, and provide progress reports and warning information by collecting and analyzing data in the test process, including task completion status and resource usage.

[0061] The environmental perception unit is used to perceive and collect relevant information of the test environment, monitor the hardware resources, operating system version, and network status information of the test environment, and update the environment status in real time to ensure the accuracy and consistency of the test environment; The automatic configuration unit is used to automatically configure the test environment according to the information provided by the environmental perception unit. It mainly adjusts the parameters and settings of the test environment automatically according to the test plan and test objectives, including installing necessary software, configuring network parameters, and setting environment variables, reducing the workload of manual configuration and improving test efficiency; The scenario simulation unit is used to simulate various actual usage scenarios and test the compatibility and stability of the product under different scenarios. Among them, scenario simulation includes: simulating different network environments, hardware platforms, and operating system versions, simulating user operations and load change factors. Through scenario simulation, the system can comprehensively test the performance of the product under different conditions and discover potential problems and faults;

[0062] The automated testing unit is responsible for automatically executing test cases, automatically loading and executing predefined test cases, covering multiple operating systems, browsers, and devices to ensure the compatibility and stability of the software under different platforms and environments. Through automated testing, the normal operation of various functions of the product can be quickly verified, improving the accuracy and reliability of testing; The anomaly detection and handling unit monitors the test execution process in real time, detects and handles anomalies that occur during testing. This unit can monitor various indicators during the test process in real time, such as response time and error rate. Once anomalies are found, such as crashes and error messages, it can immediately record, analyze, and handle them. At the same time, it can also provide anomaly warning and fault location functions, quickly locate the cause of the problem, and take measures to repair it.

[0063] The intelligent analysis and optimization module includes: an intelligent learning unit, a data analysis unit, a problem identification unit, a strategy adjustment unit, and a performance optimization unit. The resource management module includes: a data storage unit, a device status monitoring unit, a device scheduling unit, a data backup unit, and a report generation unit;

[0064] The intelligent learning unit is used to learn and analyze the historical test data by using machine learning algorithms, and extract feature information and patterns. The extracted feature information and patterns mainly refer to those features that can reveal the internal laws, correlations, and potential problems of the data. Among them, the feature information includes: 1) performance patterns, such as response time and resource utilization under different loads; 2) error patterns, such as error types, triggering conditions, and error paths; 3) input-output associations; 4) system behavior characteristics, such as system state transitions and event sequences; 5) anomaly detection patterns; and 6) user behavior patterns, such as user usage habits and preferences. This unit can automatically analyze historical test cases, error logs, and performance data, learn the behavior patterns of the product and potential compatibility problems, and gradually improve the understanding of the test scenario through learning, thereby enhancing the pertinence and efficiency of testing. The data analysis unit is used to comprehensively and deeply analyze the data generated during the testing process, and extract key indicators and trends. This unit can perform statistics, comparison, and visual display on the test data to help users quickly understand the distribution of test results, abnormal points, and performance bottlenecks. The problem identification unit is used to perform pattern matching and anomaly detection on the test data according to the results of data analysis by using preset rules and algorithms, and automatically identify compatibility problems and performance bottleneck problems that occur during the testing. It can also combine historical data and expert knowledge to improve the accuracy and reliability of problem identification. The strategy adjustment unit is used to automatically adjust the test strategy and optimize the test resource allocation according to the results of problem identification. At the same time, it can dynamically adjust the execution order, priority, and resource allocation of test cases according to the severity and occurrence frequency of the problems, ensuring that key problems and high-risk points are given priority. It can also continuously optimize the test strategy based on historical data and test results to improve the accuracy and coverage of testing. The performance optimization unit is used to analyze the reasons for performance bottlenecks for the performance problems found during the testing, and provide optimization suggestions and solutions, including adjusting parameter configurations and optimizing algorithm implementations.

[0065] The data storage unit is used to store all data generated during the testing process, including test cases, test results, and performance data. It mainly adopts efficient storage technologies to ensure data integrity and accessibility. The device status monitoring unit is used to monitor the status of the test device in real time, including the running status and performance parameters. Once device anomalies or performance degradation are detected, an alarm is immediately issued to avoid interruptions and failures during the testing process. The device scheduling unit is used to automatically schedule and manage test devices according to test requirements. When a device is idle, it can be automatically assigned to other pending test tasks, thereby improving device utilization. The data backup unit is used to regularly back up test data. According to the predetermined backup strategy, important test data is backed up to secure storage media such as hard disks and cloud storage, and data recovery functions are supported. Once data is lost or damaged, the data can be quickly restored to its previous state. The report generation unit is used to generate detailed test reports based on test results and data.

[0066] Please refer to Figure 3 , the specific working processes of the corresponding units in the test management module, adaptive test environment module, test execution module, intelligent analysis and optimization module, and resource management module include:

[0067] The requirement prediction and analysis unit in the test management module collects the historical test data, market demand, and product update information of the product, performs data analysis using machine learning algorithms, predicts the compatibility test requirements of the product, and transmits the prediction results and analysis reports to the task assignment unit. The task assignment unit receives the prediction results and analysis reports provided by the requirement prediction and analysis unit, automatically assigns test tasks according to the skills and experience of the testers and the complexity and priority of the tasks, and notifies the task scheduling unit and the testers of the assignment results. The task scheduling unit schedules and sorts the tasks according to the task assignment results and the availability of test resources, determines the execution order and time arrangement of the tasks, generates a task scheduling plan, and sends the scheduling plan to the test execution module and the progress monitoring unit. The progress monitoring unit receives the test progress data sent by the test execution module in real time, monitors the progress and completion status of the tasks. If abnormal progress or delays are found, it promptly notifies the task scheduling unit for adjustment;

[0068] The adaptive test environment module monitors the parameters and status of the test environment in real time through the environment perception unit and sends the perceived environmental data to the automatic configuration unit and the scenario simulation unit. The automatic configuration unit automatically adjusts the configuration of the test environment according to the data provided by the environment perception unit and the test requirements, and notifies the configuration results to the scenario simulation unit and the test execution module. The scenario simulation unit simulates different usage scenarios and environmental conditions according to the test requirements and the configuration results of the automatic configuration unit, and sends the simulated scenario environment information to the test execution module;

[0069] The automated test unit of the test execution module executes the automated test script according to the scheduling plan provided by the task scheduling unit, and sends the data and log information during the test execution process to the progress monitoring unit and the anomaly detection and handling unit in real time. The anomaly detection and handling unit detects and handles anomalies during the test execution process in real time, and sends the anomaly information and handling results to the intelligent analysis and optimization module;

[0070] The intelligent learning unit in the intelligent analysis and optimization module collects the data provided by the test execution module, learns using machine learning algorithms, extracts the compatibility features and test rules of the product, and sends the learning results to the data analysis unit and the problem identification unit. The data analysis unit deeply analyzes the collected test data, identifies potential compatibility problems, and generates a data analysis report. The problem identification unit combines the learning results of the intelligent learning unit and the analysis report of the data analysis unit, accurately locates the compatibility problems, and sends the problem information to the policy adjustment unit and the performance optimization unit. The policy adjustment unit automatically adjusts the test policy and optimizes the test resource configuration according to the problem information provided by the problem identification unit, and sends the adjusted policy and optimization suggestions to the test management module and the test execution module. The performance optimization unit provides product performance optimization suggestions by considering multiple performance indicators based on the results of intelligent learning and data analysis, and sends the optimization suggestions to the development team;

[0071] The data storage unit in the resource management module stores all the data generated during the test process, monitors the running status and performance parameters of the test equipment in real time through the equipment status monitoring unit, and sends the equipment status information to the equipment scheduling unit. The equipment scheduling unit automatically schedules the test equipment according to the test requirements and the information provided by the equipment status monitoring unit. At the same time, the data backup unit regularly backs up the stored data, and the report generation unit generates a detailed test report by collecting the data and results provided by the test execution module and the intelligent analysis and optimization module.

[0072] Embodiment 2

[0073] Please refer to Figure 4 , another embodiment provided by the present invention: an AI-based software compatibility testing system, including:

[0074] The task scheduling unit adopts the critical path strategy, and the specific steps include:

[0075] A1: Formulate a project plan based on the test tasks assigned by the task assignment unit and split the project to obtain manageable sub-projects , where i represents the index of the split sub-project;

[0076] A2: Draw a project network diagram, where each node represents a sub-project, the arrows between nodes represent the dependencies between sub-projects, and use the three-point estimation method to estimate each sub-project to determine the time required for completion;

[0077] The three-point estimation method takes into account the following three time estimates:

[0078] (1) Optimistic time : The shortest time in which the task might be completed if everything goes well, usually based on best conditions, unhindered processes, and maximum work efficiency;

[0079] (2) Most likely time : The time in which the task is most likely to be completed under normal circumstances, usually based on normal work efficiency, regular work processes, and some minor obstacles that might be encountered;

[0080] (3) Pessimistic time : The longest time in which the task might be completed if everything goes against it, usually based on worst-case scenarios such as low work efficiency, process interruptions, and resource shortages.

[0081] Based on these three time estimates, calculate the expected completion time and standard deviation :

[0082]

[0083] where the expected completion time is a weighted average estimate based on the optimistic, most likely, and pessimistic times, and the standard deviation represents the uncertainty or risk of the task completion time.

[0084] A3: Starting from the project start node, calculate the earliest start time and earliest finish time of each sub-project in sequence according to the time required for completion and dependencies of the sub-projects, and starting from the project end node, calculate the latest start time and latest finish time of each sub-project by backward calculation;

[0085] A4: According to , , , , calculate the float time of each sub-project. If , then reorganize the sub-project corresponding to to obtain the critical path. The float time calculation formula is:

[0086]

[0087] The automated test unit executes the automated test script in a concurrent execution and data-driven execution manner.

[0088] The anomaly detection and handling unit adopts a statistics-based anomaly detection strategy, and the specific steps include:

[0089] B1: Process the missing values and preliminarily screen the outlier values in the data during the test execution process to obtain a new data set , and extract features from the new data set , where j represents the number of data in the new data set, represents the number of extracted features, represents the j-th new test execution data, represents the -th extracted feature;

[0090] B2: Establish a statistical model based on the extracted features and perform parameter estimation;

[0091] B3: Set a threshold H to detect outlier values for the statistical model. If , then this data is an outlier value. If , then this data is a normal value;

[0092] B4: Evaluate the detected outlier values to determine whether they are real outlier values. If they are real outlier values, analyze the reasons for the outlier values through data verification, and perform data repair or replacement and model adjustment. If they are not real outlier values, perform algorithm optimization, re-select or improve features to obtain a new data set with all normal values , where, represents the k-th normal data, and k represents the number of normal data.

[0093] The present invention adopts a statistics-based anomaly detection strategy to detect anomalies in the test execution process in real time. Once an anomaly is detected, the anomaly detection and handling unit will immediately trigger a processing mechanism, such as recording anomaly information, attempting to resume test execution, and notifying relevant personnel, having the effects of accuracy and real-time performance.

[0094] The features extracted in B1 include: graph features, composite features, statistical features, model-based features, spatial features, and text features.

[0095] Among them, 1) graph features include: node degree, the number of connections of nodes in the graph; clustering coefficient, the connection density between the neighbors of a node; path length, the shortest path between nodes; 2) composite features include: feature combination, the product or ratio of two or more simple features; principal component analysis score, the main components obtained after dimensionality reduction by PCA; 3) statistical features include: mean value; standard deviation; skewness, the degree of skewness of data distribution; kurtosis, the sharpness of data distribution; quartiles, describing different parts of data distribution; 4) model-based features include: residuals, the differences between actual observed values and model predicted values; model coefficients; 5) spatial features include: location information, such as longitude and latitude, grid position; density or distance, the distance or density from neighboring points; clustering coefficient, describing the degree of aggregation between spatial points; 6) text features include: word frequency, the frequency of occurrence of specific words in the text; TF-IDF, a statistical method for evaluating the importance of a word in a document; n-gram, a sequence of n consecutive words.

[0096] Specifically, the intelligent learning unit adopts a data exploration and pattern discovery strategy, and the specific steps include:

[0097] C1: Receive a new data set all of whose values are normal obtained through the anomaly detection and processing unit, and in combination with machine learning algorithms, automatically identify and select box plots and scatter plot matrices to display the distribution and relationship of the data;

[0098] C2: Use unsupervised learning algorithms to automatically determine the optimal number of clusters or reduce the dimension, perform clustering analysis and dimensionality reduction analysis, introduce a deep learning model, combine non-linear mapping and feature transformation, capture the complex structure and hidden patterns in the data, and automatically adjust the number of clusters or reduce the dimension according to the characteristics of the data;

[0099] C3: Establish a pattern interpretation model based on deep learning to automatically extract and interpret the key patterns and features in the data.

[0100] The present invention adopts a data exploration and pattern discovery strategy, which is beneficial to discovering the internal structure and grouping in the data, revealing the hidden patterns and relationships in the data, and at the same time, avoiding the curse of dimensionality.

[0101] Embodiment 3

[0102] Another embodiment provided by the present invention: an AI-based software compatibility testing method, including:

[0103] S1: Input historical test data and perform preprocessing; the historical test data includes compatibility test results, error logs, and performance data of the product under different environments and configurations;

[0104] S2: Use machine learning algorithms to perform data analysis on the preprocessed historical test data to predict compatibility test requirements;

[0105] S3: According to the predicted compatibility test requirements, automatically allocate test tasks, and formulate a project plan and a task scheduling plan; the project plan includes the splitting of test tasks, the determination of dependencies between sub-projects, and time estimation;

[0106] S4: According to the project plan, adopt the critical path strategy to formulate a task scheduling plan, determine the start and end times of each sub-project, and automatically configure the test environment according to the test requirements;

[0107] S5: In the test environment, automatically execute test cases, record test results, and monitor and handle abnormal situations in the test process in real time;

[0108] S6: Analyze the test results, put forward optimization suggestions, and adjust the test strategy and product configuration according to the optimization suggestions, and generate and output a test report.

[0109] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments without departing from the spirit of the present invention and the scope protected by the claims. These all fall within the protection scope of the present invention.

Claims

1. AI-based software compatibility testing system, characterized by: include: Test management module, adaptive test environment module, test execution module, intelligent analysis and optimization module, resource management module, user interaction module; The test management module is responsible for planning, organizing and monitoring the entire test process; The adaptive test environment module is used to automatically configure and simulate the test environment according to test requirements; The test execution module is used to automatically or manually execute compatibility test tasks; The intelligent analysis and optimization module is used to analyze the test results and make optimization suggestions; The resource management module is used to manage various resources involved in the testing process; The user interaction module is used to provide a user interface and interactive functions.

2. The AI-based software compatibility testing system according to claim 1, characterized in that: The test management module includes: a demand prediction and analysis unit, a task allocation unit, a task scheduling unit, and a progress monitoring unit; the adaptive test environment module includes: an environment perception unit, an automatic configuration unit, and a scenario simulation unit; the test execution module includes: an automated test unit and an anomaly detection and processing unit; The demand prediction and analysis unit is used to predict and analyze test requirements; the task allocation unit is used to reasonably allocate test tasks to testers or test teams according to test requirements; the task scheduling unit is used to schedule and manage the assigned tasks; the progress monitoring unit is used to monitor the execution of test tasks in real time; the environment perception unit is used to perceive and collect relevant information of the test environment; the automatic configuration unit is used to automatically configure the test environment according to the information provided by the environment perception unit; the scenario simulation unit is used to simulate various actual usage scenarios and test the compatibility and stability of the product in different scenarios; the automated testing unit is responsible for the automated execution of test cases; the anomaly detection and processing unit monitors the test execution process in real time, detects and handles anomalies that occur during the test.

3. The AI-based software compatibility testing system according to claim 2, characterized in that: The intelligent analysis and optimization module includes: an intelligent learning unit, a data analysis unit, a problem identification unit, a strategy adjustment unit, and a performance optimization unit; the resource management module includes: a data storage unit, a device status monitoring unit, a device scheduling unit, a data backup unit, and a report generation unit; The intelligent learning unit is used to use machine learning algorithms to learn and analyze test history data and extract feature information and patterns; the data analysis unit is used to conduct a comprehensive and in-depth analysis of the data generated during the test process and extract key indicators and trends; the problem identification unit is used to perform pattern matching and anomaly detection on the test data based on the results of data analysis and using preset rules and algorithms, and automatically identify compatibility problems and performance bottleneck problems that occur during the test; the strategy adjustment unit is used to automatically adjust the test strategy and optimize the test resource configuration based on the results of problem identification; the performance optimization unit is used to provide optimization suggestions and solutions for performance problems found in the test; the data storage unit is used to store all data generated during the test; the equipment status monitoring unit is used to monitor the status of the test equipment in real time; the equipment scheduling unit is used to automatically schedule and manage the test equipment according to the test requirements; the data backup unit is used to regularly back up the test data; the report generation unit is used to generate a detailed test report based on the test results and data.

4. The AI-based software compatibility testing system according to claim 3, characterized in that: The task scheduling unit adopts a critical path strategy, and the specific steps include: A1: Develop a project plan based on the test tasks assigned by the task allocation unit and split the project into manageable sub-projects , where i represents the index of the split sub-item; A2: Draw a project network diagram, where each node represents a sub-project, and the arrows between nodes represent the dependencies between sub-projects. Use the three-point estimation method to estimate each sub-project and determine the time required to complete it. A3: Starting from the project start node, calculate the earliest start time of each sub-project in turn based on the time required to complete the sub-project and its dependencies. and earliest end time , and start from the project end node and work backwards to calculate the latest start time of each sub-project and the latest end time ; A4: According to , , , , calculate the float time for each sub-project ,like , then The corresponding sub-projects are reorganized to obtain the critical path.

5. The AI-based software compatibility testing system according to claim 4, characterized in that: The anomaly detection and processing unit adopts a statistical-based anomaly detection strategy, and the specific steps include: B1: Process the missing values ​​and preliminarily screen the outliers during the test execution to obtain a new data set , and extract features from the new dataset , where j represents the number of data in the new data set, represents the number of extracted features, represents the jth new test execution data, Indicates The extracted features; B2: Establish a statistical model based on the extracted features and perform parameter estimation; B3: Set the threshold H to detect outliers in the statistical model. , then the data is an abnormal value, if , then the data is the normal value; B4: Evaluate the detected outliers to determine whether they are true outliers. If they are true outliers, analyze the causes of the outliers through data verification, and perform data repair or replacement and model adjustment. If they are not true outliers, optimize the algorithm, reselect or improve the features to obtain a new data set with all normal values. ,in, represents the kth normal data, and k represents the number of normal data.

6. The AI-based software compatibility testing system according to claim 5, characterized in that: The intelligent learning unit adopts a data exploration and pattern discovery strategy, and the specific steps include: C1: Receive a new data set containing all normal values ​​obtained by the anomaly detection and processing unit , combined with machine learning algorithms, automatically identifies and selects box plots and scatter plot matrices to display the distribution and relationship of data; C2: Use unsupervised learning algorithms to automatically determine the optimal number of clusters or reduce the dimension. Conduct cluster analysis and dimensionality reduction analysis, introduce deep learning models, combine nonlinear mapping and feature transformation, capture complex structures and hidden patterns in the data, and automatically adjust the number of clusters or reduce the dimension according to the characteristics of the data; C3: Build a deep learning-based pattern interpretation model to automatically extract and interpret key patterns and features in the data.

7. The AI-based software compatibility testing system according to claim 6, characterized in that: The automated testing unit executes automated testing scripts in a concurrent and data-driven manner.

8. The AI-based software compatibility testing system according to claim 7, characterized in that: The features extracted in B1 include: graph features, composite features, statistical features, model-based features, spatial features, and text features.

9. An AI-based software compatibility testing method, which is implemented based on the AI-based software compatibility testing system according to any one of claims 1 to 8, characterized in that: include: S1: Input historical test data and perform preprocessing; the historical test data includes compatibility test results, error logs, and performance data of the product under different environments and configurations; S2: Use machine learning algorithms to analyze pre-processed historical test data and predict compatibility test requirements; S3: Automatically assign test tasks based on predicted compatibility test requirements, and formulate project plans and task scheduling plans; the project plans include splitting of test tasks, determination of sub-project dependencies, and time estimation; S4: According to the project plan, the critical path strategy is used to formulate a task scheduling plan, determine the start and end time of each sub-project, and automatically configure the test environment according to the test requirements; S5: In the test environment, automatically execute test cases, record test results, and monitor and handle abnormal situations during the test in real time; S6: Analyze the test results, make optimization suggestions, adjust the test strategy and product configuration according to the optimization suggestions, and generate and output the test report.

Images