An Efficient Regression Test Suite Optimization Approach Using Hybrid Spider Monkey Optimization Algorithm

Abstract

Regression testing validates the modified software and safeguards against the introduction of new errors during modification. A number of test suite optimization techniques relying on meta-heuristic techniques have been proposed to find the minimal set of test cases to execute for regression purposes. This paper proposes a hybrid spider monkey optimization based regression test suite optimization approach and empirically compares its performance with three other approaches based on bat search, ant colony, and cuckoo search. The authors conducted an empirical study with various subjects retrieved from software artifact infrastructure repository. Fault coverage and execution time of algorithm are used as fitness measures to meet the optimization criteria. Extensive experiments are conducted to evaluate the performance of the proposed approach with other search-based approaches under study using various statistical tests like m-way ANOVA and post hoc tests including odds ratio. Results indicate the superiority of the proposed approach in most of the cases and comparable in others.