The Research And Application Of Test-suite Reduction Methods

Today, all trades are using different functions of software, the type, size and complexity of software are constantly increasing to a certain degree, will directly cause the increment of error probability, software error could cause extraordinary losses, it would sometimes even pose a major threat to human security, so, people have aroused highly attention about software testing, which can guarantee the quality and reliability of software. Testing play important roles in software development, how to effectively short the period of testing and lower the developing cost with high quality of the software is a hot field in test industry.With the development of computer science and technology, leading to software testing meanwhile, from original examine the source program to establish test groups, from manual tests to semi-automation, nowadays, software testing has developed a series of specific process, the development of the industry is standard unification. How to select test cases from original test suite and keep the cost to a minimum, that is the main objective of testing.If you want to select the minimum cost test cases with 100% coverage from original test suite, that must keep the cost of repeat test to a minimum, this obviously a NP problem, these questions is difficult to answer with traditional algorithm, so, this paper proposes to overcome those combinational optimization problems by the intelligent algorithm, this algorithm improved the ant colony algorithm, and introduced the monolithic transfer mechanism of path, and combined suitable parameters to guarantee dynamics and randomicity of the search, use this new algorithm to realize the test-suite reduction, that can get the global optimal solution and avoid the state of premature and stagnation. This algorithm can get the approximate optimal solution even the optimal solution, that have an important value in the theory research and application.Finally, this paper use amount of experiments and design the value of parameters in adaptive ant colony algorithm, and applied in test-suite reduction, then designed the automatic test case generating system, it can select test case subset from two table(the relational table of test-suite and requirements, costs' table)and count the cost of subset, compared all subsets'costs with this paper's method and other classical algorithms, the results show this article proposes the algorithm is an effective and advisable method, is superior than other algorithms, it can assure the test coverage and reduced human resources efficiently too, inherently reduced the total cost of software development.