Generating Minimal Pair-wise Covering Test Suites

Software is ubiquitous and needs to be reliable. Properly testing a software system informs about its quality and reliability so as to prevent unexpected behavior during system execution. Therefore, software testing plays an important role in software development. One of the methods to prevent failures consists in testing a system under all possible input values, but when the number of parameters increases, an exponential number of test cases might result. In software testing, pair-wise testing is a combinatorial technique that focuses on making combinations of pairs of input values to generate test cases. Pair-wise covering test suites are sets of test cases that contain all such pairs of possible input values. By considering pairs of input values, instead of all possible combinations, the size of the covering test suite is significantly reduced. We propose and present an algorithm to generate pair-wise test suites based on iterations of clever rotational patterns, which result in a logarithmic growth of the covering test suite. As a result, we prove that all pairs of inputs are covered by this logarithmic growth approach. We also report the sizes of the generated test suites, and the number of pairs covered in the covering test suite.