On the implementation of search based approach to mutation testing

Abstract

As a method to measure a test case, mutation testing has become a subject of research, mostly about optimizing it using a search-based approach. At the time, there was no research about the potential obstacle of implementing a search-based approach in mutation testing. In this paper, using our previous work, we explored some alternative method to handle mutant redundancy. We use cache to improve our mutation testing tool performance, but surprisingly the cache can have other uses, to measure a search-based approach. We proposed the use of two alternative indicators to measure a search based mutation process: unique undetected mutant in population, and undetected mutant when the cache is full.