Research And Realization Of Intelligent System For Automated Testing Of Android Mobile Phone Performance

It has been sought after by many mobile phone manufacturers and smart hardware companies since the emergence of Android system because of its openness and stability. And many terminal equipments based on Android system continue to emerge. Due to the complexity of the application to be tested and the limited resources of mobile devices, meanwhile the major android equipment manufacturers have depth customization on Android,gradually increase the difficulty of testing. Android software testing often contains a lot of repetitive work, the traditional manual testing requires a large number of testers to meet the test requirements. To solve these problems, Google make Uiautomator framework known to public, in order to solve the problem of the Android software repeatability test by means of automated procedures, thereby saving testing manpower. Automated test framework improves the efficiency of the test in a certain extent. But it does not solve the test case selection, the potential failure of intelligent scheduling and use of large-scale mobile phone testing and other issues. Based on the above pain points, this paper presents an intelligent test system for Android mobile phone performance.The contributions of this paper include: (1) This paper presents a new intelligent test system of Android mobile performance through analyzing the shortcomings of traditional testing technology based on the introduction of Android automatic test technology. (2) Based on the research and implementation of web crawler-based test case extraction technology: get the software information in the vertical application market through web crawler. Sort the popular application according to the software score,download number and the views. Then write test cases based on popular applications. (3) Introduce the collaborative filtering recommendation algorithm to rapidly mining the potential failure test cases. And improve the efficiency of vulnerabilities mining through the scheduling module to test the intelligent scheduling of cases. (4) Optimize hadoop for small file storage and job scheduling to improve the efficiency of test log analysis. (5)System validation and testing.In order to evaluate the feasibility of the system, this paper builds a test platform to verify the mainstream mobile phone function. The experimental results show that the intelligent system of Android mobile performance test can effectively exploit the potential vulnerabilities under the Android system,and avoid the repetitive manual labor effectively. Ensure the test case of scientific and quality based on the network crawler extract test cases. The proposed method can improve the efficiency of mining failure test cases by introducing collaborative filtering algorithm. In addition, a log storage and parsing engine for testing scenarios is proposed, which greatly improves the efficiency of test log analysis. The system effectively improves the accuracy and efficiency of vulnerabilities mining, and single-person multi-machine large-scale mobile phone testing has become possible. At the same time, it provides an important reference value for the follow-up mobile phone test research.