Automated and Scalable Mutation Testing

Abstract

Symbolic execution and mutation testing have been demonstrated to be effective in revealing program defects. Despite their potential, their application cost, scalability and robust operation are obstacles to their practical use. Symbolic execution has the problem of path explosion, which results from the vast number of program paths, while mutation testing involves a large number of program variants to be analysed and executed. Both these problems result in scalability issues when applied in real world programs. These problems are to by addressed by the present research. The key point to be addressed is the design of few but effective mutations. These will be selected based on symbolic execution. The present research will 1) evaluate the level of con?dence or the guarantees that can be established by mutation testing, 2) design a technique to effectively detect useful mutants, 3) automate the whole mutation testing process.