Automatic Test Data Generation for Java Card Applications.
The main objective of testing software is to have the highest likelihood of finding the most faults with a minimum amount of time and effort. Genetic algorithms have been successfully used by researchers in software testing to automatically generate test data. Genetic algorithms have never been used in testing Java Smart Card (JSC) applications. In this thesis, a genetic algorithm is applied using branch coverage criterion to generate the least possible set of test data to test JSC applications. Results show that applying genetic algorithm achieves better performance in terms of average number of test data generations, execution time, and percentage of branch coverage.