The Research And Application Of Combinatorial Test Generation

At present society, computer technologies have been widely used in all industries. As the soul of computer, software affects various industries directly. Ensuring the quality of software and preventing the accident triggered by software faults are a very important work in the process of software development. Software testing is a main way to assure the quality of software. Software testing is a process for discovering the faults in the software. Designing and generating test cases is an important aspect of software testing. As the restrictions of time and cost, the key problem in software testing is how to generate a small test case set that can find most faults.Combinatorial testing is a common and relatively effective functional testing method. By studying the knowledge of combinatorial testing and doing the simulation experiments, deep understanding about combinatorial testing is generated. The main work of this paper is studying the generation methods of coverage of the paired combinations. By analyzing the current research, this paper set a new model of the generation of coverage of paired combination and put forward some improvements on the generation methods.The main contributions of this paper are listed as follows:(1) Using set theory, a mathematical model of the generation of coverage of the paired combination is build. The generation of coverage of the paired combination has close contact with set theory.(2) Aiming at generating minimum test cases, Heuristic Genetic Algorithm (HGA) based on the dynamic solution space is presented in this paper. According the changes of the solution space, heuristic operator is added to Genetic Algorithm, which makes the generation of the test case with the local optimal coverage in current environment more efficient.Different metamorphosis probability is applied in the Heuristic Genetic Algorithm, which can keep the diversity of the colony and can introduce many new gene combinations at same time.(3) By constructing the searching model of combinatorial space, a new test suite minimization method is presented which based on ant colony algorithm merging differential evolution with dynamic heuristic information (DEACA). The new method overcomes the premature convergence effectively.Through the experiment, it is verified that HAD and DEACA produce more optimal test suite than the original methods and has some merits compared with other methods.