A NFR-Based Framework for User-Centered Adaptation

Pervasive environments support users' daily routines in an invisible and unobtrusive way. To do so, they include a technical pervasive infrastructure, which is aware of and adaptive to both the operational context and the users at hand. Non-Functional Requirements (NFRs) have been effectively used to inform decision-making in software engineering: functional alternatives are compared in terms of their contribution to NFRs satisfaction. In this work, we consider user preferences over NFRs as a key driver for the adaptation of a pervasive infrastructure. We devise a model-driven framework for building pervasive systems that maximize fitness with the context and the user. Our contributions are: (i) adaptive task models, a conceptual model to describe user routines that accounts for user preferences over NFRs; and (ii) an adaptation framework, which uses our models at runtime to guide a pervasive infrastructure in adapting its behaviour to user preferences and context.