A Case Study of Adaptive Combinatorial Testing

The ability of Combinatorial Testing (CT) to detect and locate the interaction triggered failure has been well studied. But CT still suffers from many challenges, such as modeling for CT, sampling mechanisms for test generation, applicability and effectiveness. To overcome these issues of CT, adaptive combinatorial testing (ACT) is proposed in this paper, which improves the traditional CT with a well established adaptive testing method as the counter part of adaptive control and aims to make CT more flexible and practical. ACT can significantly enhance testing quality, software reliability and support testing strategy adjustment dynamically. To support further investigation, a preliminary form of concrete strategy for ACT is given as a heuristic guideline, and a case study is presented to illustrate its operations.