Survey on Test Case Prioritization Techniques for Regression Testing

Objective: The main intent of this research is to provide prioritization of test cases in regression testing for various applications. For that, several test case prioritization techniques which are classified based on various parameters are investigated. Methods/Statistical Analysis: In this manuscript, a survey has been made on various test case prioritization techniques for regression testing. Several test case prioritization techniques are presented for regression testing of various applications. One of the suggested technique is Reinforcement Learning (RL) based Hidden Markov Model (HMM) method for prioritizing test cases during regression testing of Graphical User Interface (GUI) applications. Results: This survey comprehensively studies the issues in test case prioritization techniques for regression testing. The performance of different methods is compared with various parameters such as Average Percentage Faults Detected (APFD), effect size, statistical testing. The mean values of APFD for RL-Based HMM model method is 0.68, for accumulated Q-value method is 0.62 and for statement coverage method is 0.61. Findings: The major findings in this survey are that test case prioritization is still at its primitive stages and more research is required to make it applicable in today’s world. It can be applied to all streams of computer engineering if made practically feasible. Currently test case prioritization is only applied where there is no consideration of cost like safety critical software. Many existing methods are surveyed in this work and the findings suggest that model based test case prioritization is best in all aspects. Application/Improvement: Test case prioritization can be applied to all kinds of software once it is cost effective and practically feasible but currently its application is limited to some software components only. Conclusion: This survey investigates several test case prioritization techniques and provides the idea for efficient methods for future work.