摘要： 针对自适应随机测试(ART)存在的高维和距离度量问题，提出一种改进的软件自适应随机测试策略。分析ART的局限性，扩展ART的前提假设，基于测试用例特征相似性度量选择最佳的测试用例。实验结果表明，与传统随机测试策略相比，改进策略能更多地触发软件失效，测试效果更优。

关键词: 软件测试, 自适应随机测试, 测试用例, 测试序列, 相似性

Abstract: Adaptive Random Testing(ART) is an enhanced version of Random Testing(RT). There are two factors that restrict the performance of ART, high-dimension data and distance metric. In this paper, an advanced strategy of ART based on test case similarity is presented by analyzing the limitation of ART and enlarging the hypothesis. Experimental result shows that the advanced ART, to some extent, can compensate for the limitations of existing ART and performs better than random test.

Key words: software testing, Adaptive Random Testing(ART), testing case, testing sequences, similarity

中图分类号: 

• TP311