Ideal Agents Sharing (some!) Knowledge.
Alessio R. Lomuscio, Mark Dermot Ryan · 1998
. A well-known framework by Fagin, Halpern, Moses and Vardi models knowledge-based agents as "Interpreted Systems". In this paper we analyse a particular class of interpreted systems, which we call hypercube systems, that share information among themselves. Hypercube systems arise by taking the full Cartesian product of the state space of interpreted systems. We analyse hypercube systems by taking their semantically equivalent Kripke frames and we present a sound and complete axiomatisation for them. The logic thus obtained, which we study in some detail, is a proper extension of the system S5n , commonly studied for modelling knowledge for a community of ideal agents. 1 INTRODUCTION The need for specifications of complex systems in AI, as in mainstream computer science, has brought forward the use of logic as formal tool for reasoning and proving properties about systems. In this respect, Multi-Agent Systems (MAS) constitute no exception and in the last thirty years many logics for m...