Merging Argumentation Systems

摘要

The aim of this paper is to define an arbitration merging operator for argumentation frameworks. As it is known, an argumentation framework is a collection of defeasible proofs, called arguments, and a relation attack between these arguments. Many of such arguments can be put forward by different agents and represent their different points of view. The problem is how to merge these frameworks to obtain a unique framework reflecting the arguments of the group. We overcome it by resorting to a semantic approach that selects those arguments and their attacks of an agent that vary the least from the arguments and attacks of the other agents. Then, we proved our proposal satisfy some reasonable postulates and show a procedure to build an argumentation framework resulting from this arbitration merging operator.