Search-Based Higher Order Mutation Testing: A Mapping Study

Abstract

Higher Order Mutation Testing (HOMT) uses mutants derived by applying mutation operators more than once in the program under test. This kind of test increases efficacy, but can be more expensive and lead to some challenges such as the large space of mutants. The challenges can be managed in the Search-Based Software Engineering field to seek only interesting mutants that satisfy some desired conditions regarding, for example, type of revealed faults or reduction of equivalent mutants. For this reason, we observe a growing interest in search-based HOMT and an increasing number of related studies. To better characterize such studies, this paper presents results of a mapping on search-based HOMT. We found 25 primary studies on this subject and identified the most common search-based algorithms, preferred fitness functions, addressed programming languages, and evaluation aspects. In addition to this, some trends and research opportunities are identified that allow researchers to direct future investigations.