Dynamic and flexible software systems have become the backbone of our society, running everything from manufacturing to media. A major challenge lies, however, in ensuring that software remains of high quality by adapting to user requirements or changes in the environment, which can be addressed through self-adaptation. Although much progress has been achieved in this area, it is important to fill important technical and scientific gaps that will facilitate these adaptions through a single centralised entity. The EU-funded FEDERATES (A Foundation for engineering decentralized self-adaptive software systems) project successfully elaborated a new approach for developing decentralised self-adaptive systems. To achieve this, the project team studied available literature regarding current claims on self-adaptation, latest self-adaptive systems, use of the variability principle for self-adaptation and handling uncertainty in such systems. It then conducted tests to prove that external feedback loops improve the design of self-adaptive systems, developing a model that combines adaptation and evolution to ensure sustainability. Looking subsequently at different decentralised control patterns and self-adaptation principles for building multi-agent and service-based systems, FEDERATES designed collective intelligent systems that place humans in the loop. This was followed by a new research roadmap on middleware support for engineering decentralised adaptive systems. The third part of the project involved developing a unifying reference model for self-adaptive systems involving applicable research in intelligent traffic monitoring and mobile learning. This led to new design templates for self-adaptive systems as well as to ActivFORMS, a concrete instantiation of perpetual assurances. Lastly, the team validated its research through various applications related to mobile technology, mobile learning, robotics, and remote e-health for the elderly. These results have been disseminated via tutorials, scientific events, webinars, publications and workshops. Several partner companies have embraced the results to develop new technologies, including better applications in fields such as remote elderly care and rehabilitation. The project’s outcomes have no doubt contributed to refining decentralised self-adaptive software systems across the board.
Self-adaptive software, elderly care, FEDERATES, collective intelligent systems, robotics