A New Method Of Constrained Combination Of Test Case Generation

Combinatorial testing is a scientific and effective method of software test. On the premise of software quality assurance, it can detect the variables and interactions among variables in system under test with as little as possible test cases. It has been shown to have high error detection capability in practice.Currently, the combination of test case generation problem is one of hot issues in the study of combinatorial testing. The traditional combination of test case generation method is based on covering array. However, the number of test cases generated is still too much by this method. So, researchers have presented some methods, which add some constraints in combinations of variables values. The type of constraints include: (1) Seed test, some combinations of variables values must be tested. (2) Variable/Mixed strength combinatorial testing,which allow the covering intensity among the different variables is different. (3) Specific test scenarios, some of the tested procedures are not only multiple inputs, there may be multiple outputs, then the test set need to cover any combinations of input variables,which impact any one of output values. (4) Constraints between variables, some combinations of variables values is not allowed or invalid.The article refers to the constraints between variables. We apply it to the combination of test case generation process reasonably when such constraints in the larger test suite. On the premise of software quality assurance, this approach can eliminate the invalid combinations of variables values, reduce test suite size, improve the test efficiency, reduce the cost of test.The main content of the paper is how to deal with the constraints between the variables.The processes are: First, we convert constraints into Boolean satisfiability problem. We first convert the constraints into the form of Boolean expression, and then convert Boolean expression into CNF form. Then, we using the zChaff which one of SAT solvers to find out the results, which are satisfity the constraints. Finally, we designed a new algorithm—AETG-SAT algorithm, which can obtain the better combinatorial test suite. The experimental results show that the superiority of the algorithm.