Research Of Test Suite Optimization Method On Regressing Test Model

Software testing is a very important part of software development process. The huger ofsoftware scale, the more costs of time and manpower and resources would be spent. The testcost and efficient is decided by test suites quantity. It is a key issue of software testing of howto get the best testing effects with least test cases, when software test quantity is assured andkey operates is tested fully. In this paper it is discussed how to generate test cases and reducethe test suite scale so that the test cost and time can be reduced while efficiency can beimproved on regression model in some degree.In this paper, we first introduce the background and significance of this project, and thenmake a profound study about test suite reduction. After that, we introduce the basic theoreticalknowledge about software test, regression test and the generation and optimization of testsuite. Based on the analysis of existing domestic and international test suite reductionalgorithms, an improved optimization method based on regression test model is proposed.In the paper a new optimization method which contraposes test suite is put forward onthe basis of analysis and summarization of original test suite optimization. First of all,questions can not be revealed by running original test cases, which aims at functions ofmodified or new added in software. It is necessary that new test cases be superadded to testnew functions or characters. There are many test cases generation methods, where pair-wisecombination test case generation method is adopted in this paper. Improved AETG methodand ant colony arithmetic are adopted to generate new test cases. Secondly, it is studied howto reduce the test suite scale on regression model. Redundant and disabled test cases areincluded in original test case library on regression test model, so it would be deducted. Theremust be test cases that satisfy test requirements in newly generated test suite NT, which alsoexists in the original one, so that reduction is needed. Aiming at regression test model, when test suite is optimized, OT and NT should be reduced at first separately. Greedy arithmetic ismainly adopted to depress the complexity between test cases as well as between test case andtest requirements. And then, OT and NT should be united. The whole test suite will beoptimized mainly in improved greedy arithmetic method. As a result, repeated and redundanttest cases are cut down and the test cost is depressed.The combination functions are considered fully in the improved AETG combination.Schema matching is adopted in the selecting value of factor process, which reduced selectingtime of values and numbers of redundant test cases. Ant colony arithmetic provides a newmethod to solve the problem also, by which less test cases can generated while morecombinations can be covered. On the basis of original and new test suite, improved greedyarithmetic can cut down the scale of test suite and the test cost .So the methods proposed inthe paper can generate less test cases and the scale can be reduced efficiently, which aims atthe goal of optimization.