Study And Application Of Unit Testing Based On Symbolic Execution And Genetic Algorithm

ABSTRACT:With the rapid development of software technology, the quality of software has been the subject of the software industry. Software testing which is an important means of guaranteeing software quality plays a vital role in the process of software development. Software testing occupies more than 50 percent of the cost in software projects. In recent years, testing has gradually become the hotspot in the software field because of its complexity and importance. Improving the efficiency and performance of software testing is of great significance for the software quality and the process of software development.This paper introduces the basic methods of software testing and the importance of automated unit testing. Also this paper describes the process of parameterized unit testing which uses the technology of symbolic execution. Symbolic execution is an importantly formal and automated analysis technique. It plays an import role in automated testing. During the development of symbolic execution, Dynamic symbolic execution is arisen because of the drawbacks of symbolic execution. While testing the program, Dynamic symbolic execution makes the actual execution and symbolic execution simultaneously. To generate test cases, we should combine the symbolic execution with constraint solver. In this paper, Pex is selected to do experiment. Pex is a parameterized unit testing software which is implemented by the technology of DSE. From the experiment, we can validate the advantages of the automated unit testing.In theory, symbolic execution can achieve absolutely code coverage. However, because of the complexity of the program, it is difficult to achieve complete coverage of the code. To address this problem, this paper describes the genetic algorithm in the field of software testing, also it describes the genetic algorithm which is combined with the simulated annealing algorithm. Finally, this paper gives an experiment of a typical example to validate the performance of the improved genetic algorithm.