History-Based Dynamic Test Case Prioritization for Requirement Properties in Regression Testing

Abstract

Regression testing is an important but extremely costly and time-consuming process. Because of limited resources in practice, test case prioritization focuses on the improvement of testing efficiency. However, traditional test case prioritization techniques emphasize only one-time testing without considering huge historical data generated in regression testing. This paper proposes an approach to prioritizing test cases based on historical data. Requirements are a significant factor in the testing process, the priorities of test cases are initialized based on requirement priorities in our history-based approach, and then are calculated dynamically according to historical data in regression testing. To evaluate our approach, an empirical study on an industrial system is conducted. Experimental results show an improved performance for our proposed method using measurements of Average Percentage of Faults Detected and Fault Detection Rate.