Fuzzy adaptive teaching learning-based optimization strategy for pairwise testing

Abstract

Pairwise strategies have tested effectively a range of software and hardware systems. These testing strategies offer solutions that can substitute exhaustive testing. In simple terms, a pairwise testing strategy significantly minimizes large input parameter values (or configuration options) of a system into a smaller set based on pairwise interaction (or combination). Fuzzy Adaptive Teaching Learning-based Optimization (ATLBO) algorithm is an improved form of Teaching Learning-based Optimization (TLBO) algorithm. ATLBO employs Mamdani fuzzy inference system to select adaptively either teacher phase or learner phase based on performance instead of blind sequential application as in original TLBO. In this paper, two pairwise testing strategies based on ATLBO and TLBO are proposed. Experimental results suggest that the proposed strategies are capable to be part of testers' toolkit as they outperformed competing meta-heuristic based pairwise testing strategies and tools on many pairwise benchmarks. Moreover, ATLBO based strategy generated optimal pairwise test suites than the one based on TLBO.