An evolutionary multi population approach for test data generation

In this paper we propose an approach for test data generation using genetic algorithm. Our objective is to design a multi-population genetic algorithm using uniform crossover. In this paper we analyze the performance of proposed uniform crossover multi population genetic algorithm method with different combinations of factors that influence the test data generation strategy. For implementing multi-population genetic algorithm, random migration is used and individuals are added to the existing subpopulation. Here we have also compared the single population approach and multi-population approach to determine which of these are effective towards generation of test data. By combining the individuals in the subpopulation using uniform cross over the test data generated will have better chance of existence.