Comparative analysis of regression test case prioritization techniques

Abstract

Software testing is an activity performed for evaluating the system by various testing techniques. Regression test case selection selects those test cases which focuses on the modified part of the software. It generally filters test cases instead of removing them. Test case minimization eliminates certain test cases thus, reducing the number of test cases and lays stress on using a subset of the test suite from an economy point of view. This sometimes reduces the rate of fault detection. Elimination of test cases is not performed in test case prioritization, rather it arranges them according to priority and they are thus executed with higher priority cases first followed by the lower priority ones. Test case prioritization can be search-based or fault-based or coverage based. They may also be risk-based, fault-based, history-based or requirement based. This paper has presented a comparative analysis of various TCP approaches according to the prioritization objectives framed for this study. The aim is not to show that a particular approach has some limitations as the performance of an approach varies on multiple factors such as testing scenarios, testing needs, size of the program upon which they were applied and the testing environment. The factors governing the need for the development of a certain approach are fast execution of test cases or early fault detection or making testing cost-effective, although researchers try to achieve most of them if not all. The approaches developed by numerous researchers are comparatively studied and results are presented systematically.