Social Adaptation at Runtime

One of the main goals of software adaptation is that users get their dynamic requirements met efficiently and correctly. Adaptation is traditionally driven by changes in the system internally and its operational environment. An adaptive system has to monitor and analyse such changes and, if needed, switch to the right behaviour to meet its requirements. In this paper, we advocate another essential driver for adaptation which is the collective judgement of users on the different behaviours of a system. This judgement is based on the feedback iteratively collected from users at run-time. Users feedback should be related to their main interest which is the ability and quality of the system in reaching their requirements. We propose a novel approach to requirements-driven adaptation that gives the collective judgement of users, inferred from their individual feedback, a primary role in planning and guiding adaptation. We apply our approach on a case study and report on the results.