Stepan Shevtsov started his PhD in 2014 in the AdaptWise research group. One key challenge in the field of self-adaptive systems is to provide evidence to assure the required quality properties during system operation. To tackle this challenge, Stepan studies an approach that combines formal modeling and verification with online learning. He takes an architecture-centric perspective on self-adaptation, which provides the right level of abstraction to manage complexity and the required generality of solutions. To design feedback loops, he employs principles from control theory, which provide a mathematical basis for analyzing key properties of self-adaptation, incl. stability and transient behavior. Online learning provides the means to enable the self-adaptive system to acquire knowledge about design time uncertainties during system operation. E.g., the system learns over time the reliability of a set of services, which enables better service selection. The research results will be validated in one of the domains of that are studied in the ongoing research projects of the research group.


Article in journal (Refereed)

Conference paper (Refereed)

Article, review/survey (Refereed)

Doctoral thesis, comprehensive summary (Other academic)

Licentiate thesis, comprehensive summary (Other academic)