Reduzindo o Custo do Teste de Mutação com base em Informações de Análise Estática

Abstract

To guarantee the quality of the software, static and dynamic analysis techniques can be used. Both have advantages and disadvantages and should be used together to improve the quality of the results obtained. In this work, we present a strategy for the application of a set of mutation operators in the software test, which represents a dynamic technique, based on the difficulty that automated static analyzers have in the detection of model defects. In other words, we investigate which defect categories, represented by mutation operators; an automated static analyzer can recognize and thus prioritize the mutation test considering only the set of mutation operators that model defects that are difficult to identify by the static analyzer used. We compare our subset with other subsets of already proposed mutation operators and. With the data collected, the statistical analysis showed that there are statistical differences between the mutation score and the cost reduction of the proposed strategy and the strategies present in the literature. The results obtained are promising since the subset of mutation operators identified by the proposed approach presents similar cost and mutation score to the subsets of mutation operators compared but still aggregate information not previously captured by these subsets, i.e., the intersection with defects that can be identified employing static analysis at a lower cost.