Mina Nazari - Faculty of Computer Engineering, Razi University Kermanshah, Iran
Amirhossein Damia - Faculty of Computer Engineering, K. N. Toosi University of Technology Tehran, Iran

Software testing is a very expensive and time consuming process. Test methods which generate test data based on the program’s internal structure are intensively used. Structural testing is used for automatic test data generation, in which the problem becomes a search problem. The use of metaheuristic global search techniques for software test data generation has been the focus of researchers in recent years. Genetic algorithm (GA) is one class of search algorithms that has been used widely in this field. This paper proposes a comparison between GA and random search for automatic test data generation, a technique that forces the execution of a desired path of the program called target path. Results show that our approach to software test data generation is more efficient compared with the random search method.

Software testing, structural testing, test data generation, global search, genetic algorithm