Refinement of Anguin's Machine Learning Algorithm to learning models from TTCN-3 Test Cases

Christian Otto


When developing software, software tests and structured proceeding is essentially needed. Often, because of financial restrictions, the budget to plan, analyze, design and test the software will be reduced. A consequence of this restrictions will be an incomplete specification of the software. Incomplete specifications can also occure if software updates were not or not completely documented. In most cases it is not easy to reconstruct these software specifications. In summer term 2008, S. Polonski wrote a masters thesis called “Learning of protocol-based automata“ (5). This thesis worked on the using the learning algorithm of Angluin to learn potocol-based automata that can be used to restore software specifications. This bachelors thesis will discuss enhancements of the programm of S. Polonski. We will discuss the use of the data format of the TTCN-3 test suite, the handling of TTCN-3 verdicts, ”any“ statements and loops.
Machine learning, TTCN-3
Document Type: 
Bachelor's Theses
Göttingen, Germany
Institute of Computer Science, Georg-August-Universität Göttingen
2020 © Software Engineering For Distributed Systems Group

Main menu 2