Combining variability, RCA and feature model for context-awareness
Anne Marie Amja, Abdel Obaid, Hafedh Mili · 2016
As the notion of context-awareness evolves in different paradigms, the development of context-aware systems involves several processes. These processes include, inter alia, context modeling and reasoning as well as adaptation. In [1], we presented an approach that consists of context modeling and reasoning hand in hand based on relational concept analysis and descriptive logic, respectively. An essential aspect that is often neglected in context modeling and also triggers adaptation is the context variability. In this paper, we propose an approach that lies to answer this matter based on software product line. Our approach creates a semantic link between a context RCA-based model and a feature model, and uses the MAPE-K adaptation loop to determine the appropriate SPL configurations to deploy with regards to context changes. Thus, we used ontology to represent a combined context and feature model. Thereafter, the reasoning is done via descriptive logic. We also defined context rules based on SWRL and applied them to valid SPL configurations. Furthermore, we implemented the MAPE-K adaptation loop with Prolog.