Combinatorial testing, random testing, and adaptive random testing for detecting interaction triggered failures

Changhai Nie,Huayao Wu,Xintao Niu,Fei-Ching Kuo,Hareton Leung,Charles J. Colbourn
DOI: https://doi.org/10.1016/j.infsof.2015.02.008
IF: 3.9
2015-01-01
Information and Software Technology
Abstract:Although ART can greatly enhance RT and t-way CT is better than ART and RT, normally in hitting t′-MFS:(1)There are only from 10% to 30% difference when t=t′, see Fig. 6, Fig. 16, Fig. 26.(2)The difference become quite smaller when t′<t, see Fig. 15, Fig. 24, Fig. 25.(3)When t′>t, it may not be better, see Fig. 7, Fig. 8, Fig. 17 .So we can understand why some researchers claim that there is no significant difference between RT and CT. (See the results from the above figures, where Sn(t)(t′-cov), t, t′=2, 3, or 4 represents the t′-way combination coverage of the same size test suite of the t-way covering array generated by different methods, more details can be found from paper.)
What problem does this paper attempt to address?