Research On Genetic Algorithm-Based Generating Test Case For Component-Based Software

With the development of software scale, high-efficient and high-quality software development becomes the research hotspot of modern software engineering. Component-Based Software Engineering, in short CBSE is the technology to ensure high-efficient and high-quality software development. But for the Component-Based Software, how to ensure the reliability of software becomes the difficult problem for software engineer, software testing is a necessary means of ensuring software reliability. So the testing technology of Component-Based Software will attract the testing engineer's attention.In Component-Based Software testing technologies, efficient test case generation is the necessary means of simplification of software testing, improving testing efficiency and lower testing cost. Because the test case generated as a result of the initial volume, testing inefficient, so it's requiring the use of a powerful optimization algorithm to optimize test case initially generated. Genetic Algorithm as a strong search method, showing a unique advantage and high efficiency in resolving the problems of high complexity such as: large space, multi-hillside, nonlinear and optimization of the overall. But for the local search space problem Genetic Algorithm is not very effective, reducing the diversity of the individual quickly. To overcome these shortcomings, Chun, based on somatic cell theory and the theory of immune network made an Immune Genetic Algorithm [1]. In this algorithm, target function as antigen and antibody as answer, the affinity between antigen and antibody as a joint intensity of the answer, showing a unique advantage and high efficiency. Through the improvements of the algorithm it can achieve the requirements of optimizing test case.This paper first introduces the testing technology and the automatic test case generation technologies of Component-Based Software. Then introduces the basic idea of Immune Genetic Algorithm, combined with concurrent idea and self improve idea suggests an Improved Iimmune Genetic Algorithm (IIGA),Combined with component's contract checking idea, and suggests a Model of Test Case Generation of Component-Based Software (MTCGCBS) based on the IIGA, introduces the idea of IIGA in detail and realizes it. Through the comparison with Traditional Genetic Algorithm and Random Algorithm of the quantity and the generation efficiency of test case which achieves 100% coverage of the component interface methods for experimental Component-Based Software. IIGA is further proved to be superior in generating and optimizing test case for Component-Based Software, and at the same time, the correctness and feasibility of the MTCGCBS is validated.