Android软件可靠性测试用例自动生成的设计研究

ON DESIGNING AUTOMATIC GENERATION OF ANDROID SOFTWARE RELIABILITY TEST CASES

Android是近几年新兴的智能手机操作系统,同时也出现了各种类型的Android应用软件,竞争激烈。对其进行可靠性测试是保证Android软件质量和可靠性的重要途径,同时也可帮助开发方在满足用户期望和尽快上市发布产品两者之间找到合理的平衡。而测试用例的生成是软件可靠性测试的重要环节。针对Android软件可靠性测试用例的自动生成问题,首先提出收集用户对Android软件使用的习惯数据,以便于构建软件的马尔科夫链模型的方法,并基于马尔科夫链模型,提出和实现了设置状态间转移概率算法和测试用例生成算法,最终根据以上的研究内容设计了带有自动记录用户操作习惯模块的Android文件管理器和Android软件可靠性测试用例自动生成工具,给出了Android软件可靠性测试用例设计的总体思路。实验结果表明,该方法大大降低了Android软件测试用例生成的难度,同时节约了测试用例生成的时间和成本,提高了可靠性测试的效率。

Intelligent mobile phone operating system, Android, is emerging in recent years, and various types of Android software have become a flood with fierce competition. Reliability testing is the important way to ensure the quality and reliability of Android software, which can also help the developers to catch a balance between the expectations of users and releasing the software as earlier as possible. And test case generation is an important part of software reliability testing. The main contents Of this paper are as follows ： First, we proposed a method to collect usage data of users on Android software so as to build software Markov chain model. Secondly, based on Markov chain model, we proposed and implemented the algorithms of setting transition probabilities between the states and the test case generation. Finally, according to the above two proposals, we designed the Android file manager software which contains a module automatically recording users＇ operation habits and an automatic generation tool for Android software reliability test cases, and gave the general, idea of designing Android software reliability test cases. Experimental result showed that the method greatly reduced the difficulty of Android software reliability test cases and saved the time and cost of test cases generation as well, and the efficiency of the reliability test was increased as well.