A Systematic Literature Review on Test Case Prioritization and Regression Test Selection

Regression testing is a crucial component of software testing and a crucial tool for ensuring the quality of software. An appropriate optimization method is essential for maximizing productivity and reducing expenses in regression testing. Test case prioritization (TCP) and regression test selection (RTS) are two popular methods in regression testing. This paper provides a qualitative analysis of 18 TCP and 17 RTS publications from the last five years. This paper presents four main issues. The first covers the most popular TCP techniques, the second covers the most popular RTS methods, the third covers the most popular metrics for measuring TCP and RTS, and the fourth covers data sources. Based on this study, we draw the following conclusions: (1) Defect prediction and machine learning-based TCP methods, machine learning, multi-objective, and model-based RTS methods will receive additional attention in future. (2) Defects4J is the most commonly used data set in TCP in the past five years. SIR and GitHub are the most commonly used datasets in RTS. (3) The most widely used measurement methods in TCP and RTS are APFD and cost, respectively. In future, researchers will use these two indicators to conduct a more comprehensive evaluation together with cost, fault detection capability, and test coverage.