An evolutionary testing approach for test data generation from EFSM model with string data input

Abstract: The thesis has aimed to test data generation from EFSM model with string data input. In testing area a very few work is done to generate test data with string data input. So this topic is interesting to the testing arena. To reach the goal a genetic algorithm (GA) tool is developed. A study was carried out to choose the best fitness function for string data input; resulting modified edit distance algorithm was used as a fitness function. Firstly, string and alphanumeric values with different lengths were passed through the GA tool and evaluated the result. Then three EFSM models were designed and deployed to the GA tool where most of cases the whole path is passed. This work was limited to string equality and there is a scope to work with string ordering in future.