Variations over Test Suite Reduction

Abstract

Abstract. This paper deals with the problem of effective test suite reduction. In its original form this problem is equivalent to the set covering problem, which has already been extensively studied and many strategies such as greedy or branch and bound for computation of an approximative optimal solution to this NP-complete problem are known. All of these algorithms only focus on one objective which is the minimization of the number of action calls within the test suite reduction. However, practical experience shows that balancing out the distribution of action calls is another objective which should be considered when choosing an efficient test suite. We will therefore introduce and evaluate different extensions of the standard techniques which incorporate action call distribution. We will see that these adjusted strategies can compute a reduced test suite with a smoother distribution over function calls within an acceptable amount of additional time in comparison to the classic algorithms.