Multi-heuristic based algorithm for test case prioritization

Abstract

Regression testing is the process of retesting the software after it has been modified and ensuring that there is no new errors have been introduced in the software due to these modifications. As the size of the software projects increases, the regression testing became a very costly process, so the need of detecting the faults in the software project as fast as possible became more and more important. Test case prioritization arranges test cases for execution to increase the probability of early fault detection during the regression testing. In this paper, three simple test case prioritization heuristics are presented, where every heuristic calculates the average number faults found per each test case. The three heuristics are combined together to develop a multi-heuristic based algorithm that arrange test cases based on their priorities using the scores obtained from the three heuristics. The effectiveness of the three heuristics and the multi-heuristic based algorithm are illustrated with the help of APFD (Average Percentage Faults Detected) metric. The main aim of this paper is to show how using simple heuristics for test cases prioritization would help in error early detection during regression testing, and to show how the proposed multi-heuristic based algorithm has significant increase in terms of APFD even if the algorithm is using simple heuristics. © 2014 Springer International Publishing.