Research On The System Model Of Automatic Test Case Generation Based On Genetic Algorithm

An important means to ensure software quality is to systematically test software and topertinently test in each stage of the software development process. So testing plays animportant role in this process. The key of software testing is to find representative test cases,which can ensure that the defects can be found from software more efficiently. However, withthe growing size of software, conventional testing techniques are unable to meet the need,which takes a huge space for development of automatical software testing technology. Inorder to reduce cost of software testing and improve testing efficiency and automation, thispaper mainly studies the automated generation of test cases and provides systematic solution.This paper mainly studies the genetic algorithm which is applied in automaticallygenerating the test case. On the basis of existing software test technology and test casegeneration methods, an improved genetic algorithm is proposed in this paper, which is appliedto the automatic generation of test case, and improve the inefficient defects of existing geneticalgorithm to generate test cases. Based on the domestic and foreign research, firstly, this paperanalyzes the theory of software testing and methods of automatical test case generation andfocuses on studying the lack of test case generation method of path software, including theexisting defects in auto-generated technical of test cases based on genetic algorithms. Then,the paper researches the basic principle and the realization technology of the geneticalgorithm and analyzes the key technologies solved in the application of genetic algorithm toautomatically generate the test case, which includes: the parameter coding, fitness functiondesign, genetic operator design etc, and gives specific implementation method for the issuesrelated in the experiments. And then, for the adaptive adjustment of relevant parameters in thegenetic algorithm operating, an improved adaptive genetic algorithm is proposed based on thetwo existing technology of adaptive genetic algorithm. It can improve the probability settingof crossover and mutation operator, optimize selection strategy. What’s more, the geneticalgorithm can fully play the global search ability after improving, for the purpose to avoid local convergence, premature and such phenomena as far as possible.Finally the paper studies the system of genetic algorithm for test case generation and thefunction of each module, realizes the core module of the system based on JAVA In the Eclipseplatform, namely the genetic algorithm package. Make a contrast test respectively on theimproved algorithm and standard genetic algorithm by using the triangle classificationprocedure as an under test software.Comparing the experimental results, we can obtain thatthe quality and efficiency ofthe improved genetic algorithm to generate test casesautomatically are better than the standard genetic algorithm for the tested program, whichdemonstrates the feasibility of the improved genetic algorithm.