Based On The Test Cases Of The Gray Correlation Topsis Evaluation Model

As a key step in software quality assurance, software testing is an important part in the software life cycle. According to Boehm's statistics, the cost of testing is about 30% to 50% of the total cost of software development. Therefore, it is practical to enhance the efficiency of software testing and reduce the test cost. In recent years, with the development of reuse and component technology, granularity of reuse have enhanced greatly. The concept of component has broken through the limitations of program code, and extended the entire life cycle of software development. The test plan and test case in software testing have become an important part of software reuse. Each test case indicates a specific test scheme, which guides the implementation of the testing process, the design of test input, the evaluation of measurement baseline of test results and etc. Reuse of test cases can greatly improve the efficiency and quality of testing, while reducing the cost of testing.As the improvement of testing, the scale of test cases is more and more huge. And the difficulty of test case reuse is also increasing. Then test case prioritization techniques have emerged as an assistant technology of test case reuse. This technique is mainly to set priority of test cases based on a certain criteria, and reuse test cases in the library according to priority. Then the speed of error detection will increase. The key problem of this technique is how to set the priority criteria of test case.This dissertation aimes at solving the test case prioritization based on the perspective of Multi-attribute Decision. There are two important problems in the Multi-attribute Decision. One is how to set indicators and the other is how to give weight of indicator. Through analyzing the factors impact the success of test case reuse, this dissertation finds thirteen indicators for test case, including Experience,Type,Terse,Coverage,DiscNewError, ErrorRank,ErrorCount,Efficiency,Improvement,DocIntegrity,ReuseDegree,Independence and Credit. We give the combinational weight of indicators using the Triangular Fuzzy Number AHP and Entropy Weighting method.The weights not only stand for the experts' experience, but also show the intrinsic rules of indicators'data. At last, this dissertation evaluates the test cases using a grey correlation and similarity to ideal solution evaluation method. A test case evaluation tool was designed and implemented. Finally, large numbers of experiments are executed to validate the effectiveness of this evaluation model.