Input-Output Based Relations T-Way Test Suite Generation Strategy Based on Ant Colony Optimization Algorithm (iTTSGA)
DOI:
https://doi.org/10.37934/araset.63.1.116Keywords:
T-way testing, input-output based relations, metaheuristic optimization algorithm, ant colony algorithmAbstract
Input-output based relations (IOR) testing is one of the important support interactions in t-way testing (t is referring to the interaction strength). For the past 20 years, a number of IOR test suite generation strategies have been put forth in the literature (e.g., Density, ParaOrder, ReqOrder, UNION, Greedy, ITTDG, and AURA). Although these strategies can produce a sizable IOR test suite, very little of them use the metaheuristic search technique. According to the literature, compared to other search techniques, most metaheuristic-based t-way strategies can produce smaller test suite sizes for uniform and variable strength interactions. Because the T-way test suite generation problem is NP-hard, no strategy can guarantee that it can produce the ideal test suite size for every possible system configuration. Motivated by these challenges, this paper presents an IOR test suite generation strategy based on a metaheuristic algorithm, Ant Colony Optimization (ACO) called iTTSGA. The results of two benchmark experiments for IOR uniform and non-uniform configurations were compared with those of other IOR strategies that have been published. The performance of iTTSGA has been statistically analyzed using the Friedman test and Wilcoxon Sum test. Results show that, with the exception of |R| = 10 and the top ranking in the Friedman Test, iTTSGA outperforms other strategies in the majority of uniform configurations. In addition, the iTTSGA performs exceptionally well in non-uniform configuration for |R| = 30 and 60 and ranks second in the Friedman test. It demonstrates that iTTSGA can produce test suites with a smaller file size for IOR t-way testing.