An Application Research For The Automatic Test Data Generation Technology Based On Genetic Algorithm

With the expansion of software size and software complexity improved, software testing technology continues to development, test data generation technology is widely introduced. Automatic test data generation is an effective way to reduce the test cost and to improve testing efficiency. Automatic test data generation for software unit testing is a critical part of automation test, so it is an important guarantee means to ensure the quality of the software products. Some methods are already applied to the practice of automatically generated test data, and also continued to explore, so their study has certain theoretical and practical value.This paper is focus on the methods how to use the genetic algorithm to generate test data automatically. There are several points in this paper:First, the paper introduce the software testing technology and research,the status of related basic theory and technology of software testing, as well as path-based testing techniques and the basic method of the test case generation.Then, it introduces the basic concepts and characteristics of genetic algorithm and analyzed detailed general process of genetic algorithm and an important factor affecting the genetic algorithm. In this based on the analysis of test data using genetic algorithm to generate the feasibility of the genetic algorithm to solve the test data automatically generation.Based on the research and analysis of genetic algorithm application in the test data generated, Analysis of genetic algorithm in automatic test data generation and application of the general process steps, process instrumentation for testing, fitness function design, the design of genetic operators on issues such as research and analysis, create a test tool model so that it automatically according to actual needs generate test data.Through the process instrumentation combined with the genetic algorithm, it is implemented on the path to cover the automatic generation of test data. Finally, design created by implementing a genetic algorithm-based automatic test data generation tool model to verify the practicability of tool model. In the actual system platform targeted for the experiment, analysis the factors on the genetic algorithm of the test data generation, the generation efficiency of genetic algorithm. By drawing a large number of experimental data, it can have the analysis and comparison of experimental data relevant conclusions drawn. Hoping that the conclusions has some practical value and help for further study in later using this genetic algorithm tool produced automatic generation of test data.