An Empirical Examination of Abstract Test Case Prioritization Techniques

Abstract test case prioritization (ATCP) aims at ordering abstract test case in order to increase the speed at which faults are detected, potentially increasing the fault detection rate. This paper empirically examines possible ATCP techniques, according to the following four categories: non-information-guided prioritization (NIGP), interaction coverage based prioritization (ICBP), input-model mutation based prioritization (IMBP), and similarity based prioritization (SBP). We found that the ICBP category has better testing effectiveness than others, according to fault detection rates. Surprisingly, we found that NIGP can achieve similar performance to IMBP, and that SBP can sometimes achieve even better rates of fault detection than some ICBP techniques.