Study On Test Cases Generation Based On Combination Of Genetic Algorithm And Ant Algorithm

With the computer handling targets are more and more complexity,the corresponding function of the required software should be much stronger,the quality and reliability of software is particularly important,and software testing is its vital guarantee.Software testing is a critical step of software development process.The selection and automatic generation for test cases is one of the difficulties of software testing.Meanwhile,during the automatic generation of test cases, each test requirement generally produces the corresponding test suite that perhaps have many redundant test cases. So it is necessary to simplify the test suite.And Intelligent Algorithms will be able to optimize a huge number of generated test cases,display theirs unique advantages and high efficiency.This paper launched a study on these.Firstly, this paper adopted the "structure tree" to express the software structure, based on the path coverage criteria of the structure tree, made a research on the automatic generation of test cases using the combination of Genetic Algorithm and Ant Algorithm(CG3A) from the parameter coding strategy,coverage criteria table,the construction of fitness function and the key operator setting.Through the simulation experiments with an example of the triangle classification,the test suite under different parameters value were analyzed.Secondly,used on the automatic generation of test cases with the Simple Genetic Algorithm(SGA),the Improved Immune Genetic Algorithm(IIGA) and the Simulated Annealing Genetic Algorithm(SAG-A),compared three algorithms with the CG3A through the simulation experiments in the aspects of test coverage,convergence and the numbers of test cases,the results showed that the CG3A was better than others.Finally,this paper analyzed the existing test suite reduction methods.Based on the Greedy algorithm(G) and H algorithm,the author introduced the contribution degree of test cases,explored a test suite minimization algorithm which can simplify the generated test suite by the triangle classification based on the CG3A,and got the minimal test suite that met all the test requirements.The validity of the algorithm was proved by example analysis.