Generation of test data based on genetic algorithms and program dependence analysis

A novel approach to generating test data for branch testing is presented. First, we propose an approach to decrease the number of branches by branch selection based on CDG (Control-Dependence Graph) and priority assignment. Then, we present path selection algorithm to obtain the path constraints, which are relatively easier for test data generator based on GA (Genetic Algorithm). Finally, the technique transforms the path constraints into the appropriate fitness function, and then GA is used to generate multiple test cases. Comparing with existing approaches based on GA, the proposed approach can effectively improve the efficiency of program coverage and test data generation.