Knowledge-Based Self-Adaptation

Self-adaptation is the ability of a system to adapt its behavior and/or computational structures to changes in the execution environment. The paradigm requires that the system engages in various interactions where important structural and dynamic aspects of the environment are perceived. In this paper, we present an approach to implementing selfadaptation capabilities with KnowLang, a special framework for knowledge representation and reasoning. KnowLang provides for a special knowledge context and a special reasoner operating in that context. The reasoner communicates with the host system via special ASK and TELL operators allowing for knowledge queries and updates. Whereas TELL Operators feed the knowledge context with important information driven by errors, executed actions, new sensory data, etc., ASK Operators provide the system with awareness-based conclusions about the current state of the system or the environment and ideally with behavior models for self-adaptation.