Researches On The Test Case Prioritization Techniques In Regression Testing

The information society construction of our country greatly promoted the development of software technology. Along with the expansion of the scale of software system, the increase of the complexity of structure, software quality assurance has become an important problem in the software industry, so the software testing, which can guarantee the software quality, has been highly developed. As an important software testing method, regression testing can enhance the confidence of the software quality when the existing software changes. In the whole process of software testing, regression testing is a large proportion, so it is very valuable to improve the efficiency of regression testing in order to ensure the relative quality of software system under limited resources and reduce the cost of regression testing.Now, the improvement method of regression testing efficiency is mainly to optimize the test case set, and the common means of test cases set optimization mainly in the test case reduction, test case selection and test case prioritization. Through the n-depth analysis of the three means and the consideration of technology and operation, this paper mainly concerned the test case prioritization technique based on source code, and proposed two new test case prioritization techniques in the process, which combining with the actual test operation.One of the new test case prioritization technique is test case prioritization technique based on Additional-DS. Through the study and analysis of Additional strategy,one of the test case prioritization technique based on code, we found that when Additional strategy facing multiple test cases have the same coverage of code, it will randomly select a test case to give the priority in sorting so that it makes it possible that the test case, which could detected defect, could be put in the back of sorting, thus reducing the detection rate of software defects. In order to improve the shortcoming of Additional strategy, the Additional-DS strategy is proposed by this paper, the strategy abandon the previous random selection in the face of multiple test cases with the same coverage, and it uses defect severity as a criterion to guide the prioritization of test cases. The experimental results of this paper shows that Additional-DS strategy enables superior test cases to be performed prior and ensures the stability of the test case prioritization results in a certain degree.The other new test case prioritization technique is test case prioritization technique based DC-Additional. Through the study and analysis of the defect correlation, we found the existence of positively defect correlation and defect with large defect influence degree will have a certain negative impact on the software testing. Based on this, DC-Additional strategy is proposed in this paper for eliminating the influence of defect correlation. The strategy will collect historical information in the process of testing, then it will sort the test cases set corresponding defect by the length of correlated defects set from large to small, if multiple correlated defects have the same length of correlated defects set, then use the influence degree of correlated defects from large to small to further set the priority of the test case sets. To those test cases in each test cases set, using Additional strategy as a guide to finish the test cases prioritization. Finally, the validity and practicability of DC-Additional have been tested and verified by related experiment.