Abstract
This paper presents an automatic test-data generation technique that uses a genetic algorithm (GA) to generate test data that satisfy data-flow coverage criteria. The technique applies the concepts of dominance relations between nodes to define a new multi-objective fitness function to evaluate the generated test data. The paper also presents the results of a set of empirical studies conducted on a set of programs that evaluate the effectiveness of our technique compared to the random-testing technique. The studies show the effective of our technique in achieving coverage of the test requirements, and in reducing the size of test suites, the search time, and the number of iterations required to satisfy the data-flow criteria.