A multi-dimensional model enabling autonomic reasoning for context-aware pervasive applications

A fundamental requirement for autonomic computing is to be able to automatically infer how human users react in similar contextual conditions. This paper examines the problem of autonomic reasoning for adapting context-aware applications in mobile and pervasive computing environments. In this type of systems, both the context and the adaptation possibilities must be modeled appropriately to enable the adaptation reasoning engine to infer decisions on which adaptations to perform. It is assumed that multiple cross-cutting concerns affect such decisions, and thus we introduce a multi-dimensional, utility-based model which attempts to simulate the user's reasoning mechanisms. The proposed model is applied to component-based mobile and pervasive applications, and is being evaluated through a detailed scenario. It is argued that the proposed model provides a novel and promising approach for designing context-aware, self-adaptive systems, in particular with respect to mapping the adaptive behavior to the system.