Researches On The Optimizing Of Regression Test Case Prioritization Based On Additional Strategy

Software testing is an effective measure for ensuring software quality. In recent years, with the development of software technology, software size is growing up more and more huge, and the number of test cases are increasing sharply. Since time, human being and funds cost are limited, the test cases always cannot be executed wholly with limited resources. We have to select some of the most worthy of test cases to execute from a huge set of test cases. Researchers have put forward many optimization methods about selecting test cases for the regression testing, which can be divided into three aspects:test cases selection, test cases reduction and test cases prioritization. These optimization methods have achieved certain results in practical applications. This paper focuses on two typical strategies of the test cases prioritization. The first one is Additional strategy which is used in the technology of the test case prioritization based on code. Firstly, the weights of test cases are set according to the history coverage information of them. Secondly, these test cases are sorted based on Additional strategy. In each time, it tends to choose the test case which has not yet been covered in order to achieve high coverage as soon as possible. During the sorting process, the coverage information of the remaining test cases is adjusted dynamically. The other is PORT₆ strategy which is used in the technology of the test case prioritization based on requirements. Firstly, some factors related to software requirements are collected. Secondly, the prioritization of requirements is calculated. Finally, the test cases are sorted. Based on in-depth study of these two methods, this paper accomplishes the following two tasks:1) We find that Additional strategy randomly selects a test case in each time when the test cases have the same highest coverage rate. And these test cases have the same coverage may find none or some bugs. In order to change its random selectivity and give priority to test cases which can find bugs, this paper proposes AdditionalUL strategy. When two or more test cases have the same coverage rate, the proposed strategy firstly analyzes test case from four aspects:whether the test case can find bugs and the severity of the bugs, whether the code changes, whether the code belongs to the core code, whether the code has potential errors. Then, each test case is quantified according to these analyses. Finally, the test case with the highest priority is chose. Experimental results show that the proposed strategy can improve the effectiveness of the sorted results of Additional strategy and has a higher error finding rate.2) This paper studies PORT₆ strategy which quantifies and sorts the test cases from six aspects of requirements. By analyzing its features and deficiencies, we propose PORT₄ strategy based on Additional. Firstly, the requirements are measured based on four aspects, i.e., core level of requirements, fault impact of requirements, requirement changes and requirement fallibilities. Secondly, the test cases are sorted based on Additional strategy, which incorporate the logical relationship among requirements. Finally, the priorities of test cases are adjusted dynamically in the testing process. The basic idea is if one test case finds bugs, the priorities of the test cases which can cover the same requirement with the test case are reduced. Through experimental verification, it shows that this method can improve defection and inspection bug rate.