Reducing Mutation Testing Endeavor and Deporting the Equivalent Mutants

Abstract

Software testing is an important technique for the assurance of software quality. Mutation testing is the white-box fault-based testing technique for unit testing. It is a software testing technique that ameliorates the quality and reliability of the critical software. Mutation testing deals with mutating parts of the program intentionally and then detecting them. The purpose is not to find the faults but to generate an effective test suite, which can detect all the faults in the program. Mutation testing suffers from the problem of high computational cost. The computational cost is high due to many mutants are generated for even small programs. We have experimented with different programs of variable sizes, and were able to come up with the solution of the problem. Mutation Testing is one of the most dexterous testing techniques in retracing the faults. In order to evaluate the exact mutation score in mutation testing, the vital question is whether a mutant is equivalent to its program. Unfortunately, the answer to this question is not always possible. In this thesis, we are introducing mutation operators and conditions that deport the mutants to behave as equivalently. Based on the specific criteria, detection of equivalent mutants of the program becomes ingenious. They are utilitarian in calculating the mutation score of the program accurately. This thesis presents a mutation testing technique based on the concept of the same mutation operator under similar conditions occur at different locations in the program. In the proposed technique, we assemble the tantamount behavior mutants under a group and a single mutant is contemplated from the group for performing mutation testing. The benefits of the proposed approach are reduction in time, effort and cost.