(Dis)Belief Change Based on Messages Processing

Abstract

This paper focuses on the features of belief change when agents have to consider information received from other agents. We focus on belief change operators when agents have to process messages about a static world. We propose to consider agents’ belief state as a set of pairs 〈belief, origin of the belief〉 combined with a preference relation over the agents embedded in the multi-agent system. The belief revision procedure for handling received messages is a safe base revision procedure where messages are considered in their syntactic form. According to the reliability of the sources of the conflicting belief, agents remove the less reliable belief in order to handle the received message. Notice that the less reliable source can be the sender of the message itself. In order not to loose precious information conflicting belief are not removed but considered as potential belief. As the agent changes its belief, potential belief is reconsidered and may be reinstated as current belief. In a similar way, messages can concern statements that should not be believed, called disbelief. As belief, disbelief can become potential. These different kinds of belief enables us to propose a new semantics for a modal based language for describing (dis)belief. Agents may handle sequences of messages since the proposed belief change operators handle iterated belief change.