The Research Of State-based Class Testing Technology

With the rapid development of information technology, applications of the computer software penetrate into every aspect of people's life; people's understanding of software quality has been increased to a new level. Software testing gets widespread attention as it is an important means to assurance software quality. Object-oriented technology provides better software development methods, but also the characteristics of object-oriented software have brought new challenges for the software testing.This article takes the class of object-oriented software as the research subject, for the state-based testing method, studing the test sequence and test data generation method based on consistency EFSM mode and selecting the UIO sequences as state identification sequences.Firstly, for the problem that the UIO tree generated by traditional UIO sequence generation algorithm is too big, based on the study of the existing algorithms, a new UIO sequence generation algorithm is proposed using the invertible sequence and the translation table. The new UIO sequence generation algorithm uses a variety of pruning strategies to control the UIO tree's size and the chain-node search technique to improve the efficiency of UIO sequences output.Secondly, in order to reduce the length of test sequence, a better test sequence generation algorithm is developed by expanding existing methods, based on the study of the two main methods for reducing the length of the test sequence. The new algorithm use invertible sequences as connection sequences, and find a shortest traversal path through the rural postman problem algorithm, so as to constitute a shorter test sequence. The test sequence is characterized by that the UIO sequence for each state is included in the UIO sequence of the other states, so that the overlap between test segments is used effectively.Finally, in order to improve the efficiency of test case generation, genetic algorithm is selected to construct the required test data for performing the test sequence. The application of genetic algorithm in the test method is studied, and a certain improvement measures are adopted to speed up the convergence rate of genetic algorithm.