Preferential Multi-Agent Nonmonotonic Logics: Preliminary Report

This paper presents a way of constructing a multi-agent nonmonotonic logic from any propositional preference logic (the base logic). The multi-agent logic will "correctly" extend the base logic for each agent, that is, each agent will reason in the base logic, and the reasoning of each agent is independent of each other. If the base logic is propositional circumscription then the extended logic will also correctly deal with default formulas that mix the knowledge of more than one agent. Some of the properties of the extended logics and an example of nonmonotonic multi-agent reasoning are presented. 1 Introduction The term agent is frequently used in the area of knowledge representation and reasoning. Several different systems have been proposed to model an agent reasoning about its environment, an environment that in many applications includes other agents. Thus, some logics have been proposed to deal with the knowledge of more than one agent [HM92, FHV91], but these log..