Handling uncertainty in self-adaptive systems in presence of multiple concerns: an architectural approach

In her thesis, Sara Mahdavi Hezavehi sheds light on the notion of uncertainty in self-adaptive systems through investigation of state-of-the-art and community’s perspective on the topic, and she identifies the most important concerns in the domain and devises methods to equip the system with the capability of balancing these concerns while handling uncertainty at runtime.

Mahdavi Hezavehi first conducted an extensive research of the existing architecture-based self-adaptive approaches dealing with uncertainty in the presence of multiple concerns. Then she shifted her focus on investigating and grasping the concept of uncertainty itself. The next step was to conduct an extensive study to look into the perception of the community on the concept of uncertainty, approaches to handle uncertainty, and open challenges in this area. 

Mahdavi Hezavehi concluded the research by proposing an architectural view- point for benefit-cost-risk (i.e., multiple concerns) aware decision-making in self- adaptive systems. The aim of this viewpoint is to assist architects to better address the trade-off analysis both at design time and runtime. The viewpoint allows strategic identification of benefit, cost, and risk factors of the system, which for instance can be translated into domain specific quality properties, and implementation of solutions which eventually will be applied at runtime.

Dissertation: https://hdl.handle.net/11370/31666484-5742-4888-8ecc-7318790dd55b