A social reasoning mechanism based on dependence networks

Abstract

This paper describes the fundamental concepts of a social reasoning mechanism, designed to be part of an agent's internal model, in a multi-agent systems (MAS) context. It enables an agent to reason about the others using information about their goals, actions, resources and plans. Every agent stores this information in a data structure called external description. We have formally defined and implemented the concepts of external description, dependence relation, and dependence network. One of the main contributions of this work is that an agent can infer his dependence on others using either his own plans or those of the others. As a result, we have defined a preliminary taxonomy of dependence situations regarding the goal being analysed (unilateral, mutual or reciprocal) and the sets of plans used in this reasoning mechanism (mutually or locally believed). We have used this model to build a dependence network simulator, called DEPNET, which is also briefly described in this paper...