基于行为树模型的Android应用GUI自动随机测试方法  

AUTOMATED RANDOM GUI TESTING OF ANDROID APPLICATION BASED ON BEHAVIORAL TREE MODEL

作　　者：佘学文 陈向宇 刘东启 She Xuewen;Chen Xiangyu;Liu Dongqi(Lingnan Institute of Technology,Guangzhou 510663,Guangdong,China;School of Computer Science and Engineering,South China University of Technology,Guangzhou 510006,Guangdong,China)

机构地区：[1]广东岭南职业技术学院,广东广州510663 [2]华南理工大学计算机科学与工程学院,广东广州510006

出　　处：《计算机应用与软件》2021年第5期21-25,48,共6页Computer Applications and Software

基　　金：国家自然科学基金项目(61771007);广东省科技创新战略专项项目(2016A030305001)。

摘　　要：基于马尔可夫链等统计模型进行GUI测试是目前最有效的GUI自动化测试方法之一,但目前的方法在开始测试前需要很多前置条件,且要花费大量时间做静态分析等准备工作,从而导致测试效率和自动化水平不高。提出一种基于行为树的GUI自动化随机测试方法,该方法不需要进行静态分析,只需通过统计模型从用户使用日志中创建行为树模型,且可以动态地对行为树模型进行更新,以增加用户很少或从未使用过的事件的概率。对三种开源Android应用程序的评估,以及和其他GUI自动测试方法的比较结果显示:提出的Android应用GUI自动化随机测试方法具有很好的效果。Automated GUI testing based on behavior tree model such as Markov chain is one of the most effective testing methods.However,there are many prerequisites before testing,and requires a lot of time doing static analysis,which leads to low test efficiency and low automation level.An automated random GUI testing method based on behavior tree is proposed.This method does not need static analysis,but creates behavior tree model from usage log by statistical model.The behavior tree model can be dynamically updated to increase the probability of events that are rarely or never used.Through the evaluation of three open source Android applications and the comparison with other automatic test methods of GUI,the results show that the proposed method of automated random GUI testing of Android application has good effect.

关 键 词：Android 软件测试 GUI测试 自动化测试 行为树模型 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]