The Researches Of Two-phase Test Suite Reduction Based On Requirement Anlysis

With the rapid development of software technology and widespread popularityof software application, users proposes deeper requirements for software reliabilityand safety.Software testing is a necessary process and procedures to ensure thesecurity and the correctness of software. But as the scale and the complicationincreasing more and more, the structure complexity of the software of unceasingenhancement, the consumption of software testing grows inevitably. So, research ontesting process optimization is an important way to reduce workload and the cost intesting process.The number of test cases influts testing workload directly. On the one side,execution time increases as the growth of the test case set. On the other side, thenumber of little cannot the test completeness cann’t be guaranteed without enoughcases. The object of test cases set reduction is to find a subset of the original cases set,which has the least number of cases, also the same coverage and ability to find errors.This paper presents a two-phase approach based on requirement alysis, which containsdemand slicing and cases set optimizing. As fowllows:First phase, presents a slicing algorithm to structured program, which accordingto formalized description of demand. The mehod dosen’t have to establish proceduredependence graph, but base dependence table (BDT) and function dependence table(FDT). This program begins at the the final executive statements, then, move backuntil the initialization position of requirement, at last, judge the interrelationsbetween requirement slices. Experimental results show that our algorithm is lesscomplex, compared with similar synchronous dynamic slicing algorithm, but hassimilar accuracy.Second phase, public path reduction algorithm is (PPR) is shown at first, whichis is based on DDGRAPH, and used for dealing with the result of the last phase. Oncethe the minimalist path been got, it can be criteria to deside wheather a case cover thewhole path. By this way, the complexity is reduced. At last, requirement-caseadjusting algorithm is present, by which the final result of cases set is got. Experimentshows that effect is better than G and GRE, based on greedy algorithm.