| Nowadays, software testing has become vital part of software development which could not be substituted. Software testing has been proved the best way for guaranteeing the software product quality. In real software project, in order to see the test result quickly, engineers usually carries on the black box test first, namely function test. In the function test process, sometimes the input parameters are in inter-combination situation. Based on the experience and the observation, the researcher discovered that, the parameter pair combination test is the most effective test method, since many software mistakes are caused by the mutually affected parameter pairs. Thus, focus on how to generate and choose the test cases, the researchers proposed a series of algorithms and the research results, including the pairwise testing and its related strategies. Pairwise testing can efficiently find bugs which caused by pairs of parameter values, it has been proved one of the most effective way of testing. Besides, the test cases generated by pairwise testing are comparatively tiny compare with the full combination of all parameter values, and therefore is less expensive.Under this premise, this paper proved the pairwise testing is essentially the pair- cover issue. And based on this theory, this paper introduced and analyzed several pairwise algorithms, such as the AETG, IPO strategy, the orthogonal Latin square and so on, which are the most popular ones at present, and also proposed some algorithms such as input parameter constraint, equivalent class, etc. for their improvement in the decreasing of both quantity of generated test cases and generation time cost.Finally, we designed and implemented an automation test tool which meets the requirements of our team, and could produce pairwise test cases automatically, proving our method performs better. |
