Abstract
Regression testing is a useful technique for software testing. There are three mostly used applications: regression test selection, test case minimization, and test case prioritization. This thesis focused on test case prioritization problem. Traditionally, there are several techniques for test case prioritization; two of the most used techniques are Greedy and Additional Greedy Algorithm (GA and AGA). However, there exists a drawback on forward two techniques that it does not consider severity while prioritizing test cases. Thus, we provided a technique: Enhanced Additional Greedy Algorithm (EAGA) that was modified from AGA for test case prioritization. We design an experiment with eight subject programs to investigate the effects of different techniques under different criteria and fault severity. The results of this thesis were: EAGA outperformed the other techniques in terms of “units-of-fault-severity-detected-per-unit-test-cost” when severity was taken into account; EAGA performed equally well as AGA in terms of “fault-detection-of-test suite” and “decision-detection-of-test suite” in large size program when severity was not taken into consideration, although AGA excelled in small size programs. In summary, while severity was taken into account, EAGA outperformed AGA and GA in test case prioritization problem.