Leveraging State-Based User Preferences in Context-Aware Reconfigurations for Self-Adaptive Systems

Applications in ubiquitous environments need to adapt to a range of fluid factors, like user preferences, context, and various system configurations. In this paper, we address the problem of system adaptation in order to continuously achieve high user benefit while keeping reconfiguration costs low. To this end, the presented approach leverages not only the immediate context but also future transitions. In contrast to existing approaches that either maximize benefit or minimize reconfiguration costs, our proposed decision support mechanism achieves a trade-off between those factors. Considering user preferences, deployment constraints, and probabilistic context state transitions, we propose a multi-objective utility function to determine the best reconfiguration choices. Experimental results show that the proposed approach achieves high user benefit while keeping reconfigurations costs low.