A Tree Based Strategy for Test Data Generation and Cost Calculation for Uniform and Non-Uniform Parametric Values

Software testing is a very important phase of software development to ensure that the developed system is reliable. Due to huge number of possible combinations involved in testing and the limitation in the time and resources, it is usually too expensive and sometimes impossible to test systems exhaustively. To reduce the number of test cases to an acceptable level, combinatorial software interaction testing has been suggested and used by many researchers in the software testing field. It is also reported in literature that pairwise (2-way) combinatorial interaction testing can detect most of the software faults. In this paper we propose a new strategy for test data generation, a Tree Based Test Case Generation and Cost Calculation strategy (TBGCC) that supports uniform and non-uniform values, for input parameters (i.e. parameters with same and different number of values). Our strategy is distinct from others work since we include only the test cases which covers the maximum number of pairs in the covering array at every iteration. Additionally, the whole set of test cases will be checked as one block at every iteration only until the covering array is covered. Other strategies check each test case (N-1) times, where N is the maximum number of the input parameters. A detail description of the tree generation strategy, the iterative cost calculation strategy and efficient empirical results are presented.