Metaheuristic Based Approach to Regression Testing

Abstract — Regression testing is a very expensive and time consuming process as there may be insufficient resources to re-execute all the test cases in resource and time constrained environment. It acquires a lot of human effort, if done manually. Lot of techniques have been reported on how to select regression tests so that the number of test cases do not boast up too high as the software evolves. The techniques include regression test selection, test minimization and test prioritization. In this paper, various metaheuristic approaches have been studied to examine their potential benefits to regression testing. The paper also addresses the problem of choice of fitness metric and determination of the most suitable search technique to apply. Genetic algorithm performs well, although Greedy approaches are surprisingly effective, given the multimodal nature of the landscape. It may be accomplished that Cuscuta ordering which is inspired by intelligent behaviour of plants gives same results as given by the optimal and ACO ordering but better than unordered, random and reverse order. The study also reveals that ABC outperforms the other approaches i.e. GA, ACO, BCO and PSO in test suite optimization process as parallel behaviour of the bees is used to reach the solution generation faster