Study Of Automatically Generate Software Structural Test Data Based On Evolutionary Algorithm

The goal of software testing is to find potential errors in the program, the key factor ishow to quickly generate test data automatically, the traditional manual testing cost a lotof manpower and material resources, so the automation testing technology was studiedby more and more people.Test data automatic generation technology can automatically generate test dataaccording to the requirement of the application under test, and improve the testefficiency and reduce the cost of software development. In recent years, the research ofautomatic test data generation technique more and more widely, obtained certainresearch results in many areas, and had achieved good application in function testing,structure testing, performance testing and object-oriented testing and other fields.Evolutionary testing technology using evolutionary algorithms to generate test data,it has many advantages compared with the traditional testing technique. The study ofautomatically generate test data based on evolutionary algorithm method focuses on thepath coverage direction. The key problem is how to construct a suitable and has a goodorientation of the fitness function to evaluate the quality of a test data. This paperintroduces several evolution test methods, shows its advantages and problems it caneffectively solve, and proposes an improved evolutionary test data generation method,which can generate better test data.This paper mainly studied in the following aspects:1) Studied structural test data generation for path coverage method home andabroad, studied a variety of structure method automatic generation of test data,summarized the advantages and disadvantages of each method;2) Studied and analyzed the evolutionary algorithm, such as parameters,complexity and convergence, and put forward a type of fitness function constructorbased on the current study;3) An evolutionary testing environment was built, which can automatically obtaintest data. The test environment can promote the research of evolutionary testing.