Skip to Main Content
This paper presents a software approach to augmenting existing real-time systems with self-adaptation capabilities. In this approach, based on the control loop paradigm commonly used in industrial control, self-adaptation is decomposed into observing system events, inferring necessary changes based on a system's functional model, and activating appropriate adaptation procedures. The solution adopts an architectural decomposition that emphasizes independence and separation of concerns. It encapsulates observation, modeling and adaptation into separate modules to allow for easier customization of the adaptive behavior and flexibility in selecting implementation technologies.