A New Method of Reducing Pair-wise Combinatorial Test Suite

Abstract

The biggest problem for combinatorial test is a numerous number of combinations of input parameters by combinatorial explosion. Pair-wise combinatorial coverage testing is an effective method which can reduce the test cases in a suite and is able to detect about 70% program errors. But, under many circumstances, the parameters in programs under test (PUTs) have relations with each other. So there are some ineffective test cases in pair-wise combinatorial test suites. In this paper, we propose a method of reducing ineffective combinatorial test cases from pair-wise test suite. The main ideas of the method is that we firstly analyzes the dependent relationships among input parameters, then use the relationships to reduce ineffective pair-wise combinations of input parameters, and lastly generate the pair-wise combinatorial coverage test suite. The experiments show that the method is feasible and effective, and considerably reduce the number of pair-wise combinatorial test cases for some programs under test.