The evolutionary algorithm such as genetic algorithm (GA) is highly efficient for various optimization problems. One such application of GA is in automatic test case generation for a subject under test. The primary requirement of GA to work depends on creation of fitness function and the adjustment of various optimization parameters such as - crossover probability, mutation probability and initial population size. The auto adjustment of optimization parameters is a subject of research as it is done manually. Careful auto adjustment of parameters results in more optimized outputs in less time.

This paper discusses the effect of various optimization parameters need to be adjusted while generating test cases for software under test using genetic algorithm. Branch based distance approach is used to form the fitness function and auto adjustment in GA parameters is achieved by observing the behavior of outputs, this results in highly efficient and accurate test data in less number of function evaluations as compared to manual approach.