Raciocínio sobre crenças utilizando a lógica de confiança em tópicos

Abstract

Trust is considered one of the vital concepts of multi-agent systems, in which autonomous agents have their own interests, which may differ from the interests of a group which they belong. In order to make a decision, the parties involved must be able to exchange information, so that the group’s point of view is in accordance with the information collected by all participants. Eventually, agents can tell untruths, either through ignorance or in order to get an advantage on the group to which they belong. Based on this situation, agents must decide whether information generated by another agent is reliable or not, so that the group’s point of view supports a decision that reflects its needs. In this work, we proposed an extension of a logic that allows to formalize and reason about beliefs that involve the concept of trust between agents with respect to information. In our extension, we introduced two new features in this logic: (i) the trust between agents on topics and (ii) the distrust between agents on formulas and topics.