@article{MAKHILLAJIT201615236543,
    title = {Test Case Generation for Pairwise+Testing},
    journal = {Asian Journal of Information Technology},
    volume = {15},
    number = {23},
    pages = {4800-4805},
    year = {2016},
    issn = {1682-3915},
    doi = {ajit.2016.4800.4805},
    url = {https://makhillpublications.co/view-article.php?issn=1682-3915&doi=ajit.2016.4800.4805},
    author = {V. and},
    keywords = {Combinatorial testing,pairwise testing,pairwise+ testing,test case generation,3-way testing},
    abstract = {Software system faults can be completely detected only through exhaustive testing. But, it cannot
be performed on many of the real-life systems because it may be too expensive as it may consume an enormous
amount of time. The t-way combinatorial testing enables testing to be performed at low cost and less time. The
value of t starts from 2 and takes 3, 4, 5 and so on. As the value of t increases, the size of the test suite increases
and thus it takes more time for testing the system. In pairwise testing (t = 2), every pair of parameter values in
the input domain is covered by at least one test case. It is highly effective in detecting up to 70% of faults
triggered by a system. In 3-way testing, every triple of parameter values in input domain is covered by at least
one test case. It can detect up to 90% of faults triggered by a system but the size of test suite is relatively larger.
Therefore, pairwise testing is widely used in industry. When a test suite is generated for pairwise testing, it may
contain a lot of gaps. The gaps have to be filled with some arbitrary values in order to proceed for testing. In
this paper, we present an approach called pairwise+ testing in which the gaps in the test suite are filled in such
a way that many of the triples that are useful for a 3-way testing are accommodated in the test suite. Thus the
test suite generated for pairwise testing is enriched to cover a part of 3-way testing. Depending on the amount
of coverage, the percentage of failures triggered by pairwise+ testing can be estimated to be anywhere within
the range of 70 and 90%.}
    }