Research Of Software Component Testing Method Based On Genetic Algorithm

Software testing is an significant measure to ensure the quality and vitality of software development. With the expansion of the construction scale of modern softwa re, and the extensive using of the ideology of the component-based software development, the efficiency of the development of software has greatly improved. However, it is inconformity to the speed of the component-based development, due to some characteristics of the software component, the user of the components is unable to know the interior construction, the traditional software testing cannot be applied to the component testing.This paper aimed at the problem that the difficult of the software compone nts testing, analyzed the factors that affect the implementation of component testing, introduced the thought of component built in testing, improved the testability of the components by established observation points in the components. And designed a component testing model based on improved genetic algorithm, which can overwriting detect the logical path of components, achieved to generate the cases of the testing automatically, and applied the method in actual development process.The contents of this paper include:According to the difficulties of the component testing and the factors that affect component testing, proposed the thought that introduce the built-in-test to ensure the testability of the components, which can leave testing points in the inte rior during the component developing and describe the critical information of the observation points, so that to support the testing ability of the system.According to the difficulties of detection aimed at the complex logical construction in the interior of the components, overwriting detected the logical path of components by using the strong nonlinear searching ability of genetic algorithms. In order to solve the problem that the traditional genetic algorithms using the fixed genetic operators are prone to premature convergence or to be in a torpid state, improved the genetic algorithm in its adaptive aspects, so that the genetic algorithm can effectively avoid falling into local optimum and to maximize the retention of the efficient search ability.To combine the development method of component testing and the generation technique of the genetic algorithm path-coverage testing data, established a generation model of the path coverage testing case. confirmed the practicality and efficiency of the componen t testing development and genetic algorithm of the path-coverage testing by using the model to develop a testing system to test the software component in the oilfield.