Knowledge-based assessment of behavior in dynamic environments

In this paper we propose a knowledge-based approach for thesupport of adaptive and context-aware behavior in autonomic systems. In order to providemeaningful and adequate behavior in the presence of dynamic (or adverse) environmentssystems must possess the ability to assess their current state on the basis of knowledge about themselves andtheir environments. We propose to use ontologies for therepresentation of knowledge and its integration into a general systemsarchitecture supporting autonomic behavior. Especially we rely on fuzzy description logics in order to support robust automatic reasoning w.r.t. environmental conditions.We distinguish several aspects of knowledge and use multiset term rewritingto describe mechanisms of coordination between them.We discuss the advantages of our approach and give examples forparadigmatic applications.