Genetic Algorithm-Based Approach for Adequate Test Data Generation

Abstract

Software testing is an important phase in software development. It involves two activities, test data generation and test execution. Test data generation is a NP-complete problem as we have to find a lot of test data to validate our system. Also those test data should be adequate in nature. In this paper, we present a method to generate test data automatically from initial test data and then testing these test data against the software under test (SUT) for adequacy criteria. First, we generate a test data set randomly. Then, we apply genetic algorithm to find a better test data set iteratively. We stop at the position where our test data set satisfies the stopping condition or it completed maximum iterations. We test the generated test data against the software to check its adequacy. The test data generated by our approach are more capable of finding the synchronization and loop faults. A case study is given to illustrate our approach.