OpenSpace - AI Techniques for Testing Highly-Variant Software

The increasing degree of parameterizability and resulting degree of software variability leads to new challenges regarding software development- and test processes. In order to be able to deal with an increasing software variability, different modeling approaches such as feature models are provided. For providing a personalized software variant to the customer, the resulting models are taken into account in build processes or during runtime. This increasing degree of software mass customization makes, for example, the search for relevant test cases more complex. A major goal of OpenSpace is the development of Machine Learning approaches that efficiently support different tasks related to software testing & debugging in order to be able to deal with the increasing complexity of software. In this context, OpenSpace will develop variability-aware test methods which play a major role in the context of software product lines. From the viewpoint of research, OpenSpace will develop new approaches 1) to the automated analysis of variability models and corresponding test case generation and 2) to the machine learning based identification of faulty software components and faulty/suboptimal parametrizations which could lead to inefficiencies or erroneous behavior. The share of testing efforts in software projects is around 25-40%. Improvements in related quality assurance processes are of extreme importance especially in the context of software product line development. Due to the increasing complexity and variability of the provided software components, the OpenSpace application partner wants to make test processes variability-aware in order to achieve a sustainable improvement in software development practices.