Representing and communicating context in multiagent systems

Abstract: Context-aware agents operating in a cooperative multi agent system (MAS) can benefit from establishing a shared view of their context, since this increases coherence and consistency in the system’s behavior. To this end, agents must share contextual knowledge with each other. In our prior work on context-mediated behavior, agents used frame-based contextual schemas (c-schemas) to explicitly represent and reason about context.  While an expressively rich approach, the lack of formal structure poses problems for mutual understanding of c-schemas among agents in a MAS. As we are interested MASs with heterogeneous agents, not only will agents represent c-schemas in idiosyncratic ways, but the set of c-schemas known by each agent will differ. In this paper we propose anew, related representation of contextual knowledge using description logic and a shared ontology, and we present a technique for communicating contextual knowledge while respecting bandwidth limitations.

Citation: Sonia Rode and Roy M. Turner.Proceedings of the Ninth International and Interdisciplinary Conference on Modeling and Using Context (CONTEXT’15), published as a volume in Springer’s Lecture Notes in Artificial Intelligence, Larnaca, Cyprus, November 2015.