ABSTRACT 

To early discover faults in source code, test case ordering has to be properly chosen. To this aim test prioritization techniques can be used. 

Several of these techniques leave out the execution cost of test cases and exploit a single objective function (e.g., code or requirements coverage). In this paper, we present a multi-objective test prioritization technique that determines sequences of test cases that maximize the number of discovered faults that are both technical and business critical. The technique uses the information related to the code and requirements coverage, as well as the execution cost of each test case. The approach also uses recovered traceability links among source code and system requirements via the Latent Semantic Indexing technique. We evaluated our proposal against both a random prioritization technique and two single prioritization techniques on two software applications implemented in Java. The results indicate that our proposal outperforms the baseline techniques and that additional improvements are still possible.

BIO

Mahfuzul Islam is MSc student in Computer Science at University of Trento, Italy. He received his bachelor degree in Computer Science & Engineering at American International University Bangladesh. After his bachelor Mahfuzul worked for 2 years as a software engineer at Code71 Bangladesh Ltd. Currently, he is visiting Lero to finalize his master thesis, which is titled "A Multi-Objective Technique to Prioritize Test Cases". In particular, Mahfuzul will improve the work of his thesis with a notion of context variability of software requirements. His research interests are on software testing, maintenance and requirements engineering.

If you would like to attend please RSVP to  patsy.finn@lero.ie