‘Ambha’: A Metaheuristic Framework for Test Case Suite Optimization

Abstract

Background:To deal with resource and time constraints in retesting of modified software, it is really vital as well as a significant issue to look for regression testing strategies that keeps up with performance as well as quality of modified software. To meet this objective, regression test case suites should be optimized to check functionalities of modified software.

Objectives: In regression testing scenario, this study aims to identify an optimized test case suite in order to maximize its fault revealing potential as well as reduction in its execution time.

Methods: Authors developed a metaheuristic framework that consists of three phases, i.e., (i) test case suite reduction using ant system-based classification rule discovery (ii) test case suite selection using Whales as search agents (iii) test case suite prioritization using grey wolf optimization algorithm. The ‘Ambha’ framework is implemented on 12 test case suites taken from GitHub and TravisCI code repositories.

Results:The results reveal that ‘Ambha’ framework has significantly improve five performance metrices. Specifically, Average Percentage of Fault Detection metric is improved up to 56.4% and execution time is reduced by 19.7%.

Conclusions: The proposed metaheuristic framework ‘Ambha’ has shown promising results for regression testing problems as compared to framework (ACO+WOA+GWO).