Search-based mutation testing to improve performance tests

Performance bugs are common and can cause a significant deterio ration in the behaviour of a program, leading to costly issues. To detect them and reduce their impact, performance tests are typi cally applied. However, there is a lack of mechanisms to evaluate the quality of performance tests, causing many of these bugs re main unrevealed. Mutation testing, a fault-based technique to assess and improve test suites, has been successfully studied with func tional tests. In this paper, we propose the use of mutation testing together with a search-based strategy (evolutionary algorithm) to find mutants that simulate performance issues. This novel approach contributes to enhance the confidence on performance tests while reducing the cost of mutation testing.