Greedy Interaction Elements Coverage Analysis for AI-Based T-Way Strategies

Abstract

Recently, many researchers have started to adopt Artificial Intelligence AI-based strategies for t-way testing. Here, each interaction is covered at most once whenever possible. In many AI-based strategies, sampling for the most optimal test cases is given utmost priority, but measuring of the interaction coverage metric per test case is often neglected. In the situation where not all test cases can be executed due to constraints on project deadline, the availability of interaction coverage metric per test case can be a useful indicator on how greedy each AI-based strategy of interests is. In this manner, test engineers can make informed decision on the selection of suitable strategies for use. In this paper, this study presents a systematic analysis of existing AIbased strategies including that of Hill Climbing HC, Simulated Annealing SA, Tabu Search TS, Great Flood GF, Particle Swarm Optimization PSTG and Harmonic Search Strategy HSS as far as its rate of coverage per test case is concerned. In doing so, this paper demonstrates that HSS, in most cases, gives competitive interaction coverage rate as compared to competing AI-based strategies but with less number of iterations.