Two Roles of Preferences in Argumentation Frameworks

摘要

In this paper, we show that preferences intervene twice in argumentation frameworks: i) to compute standard solutions (i.e. extensions), and ii) to refine those solutions (i.e. to return only the preferred extensions). The two roles are independent and obey to distinct postulates. After introducing and studying the postulates, we provide an example of a formal framework which models the two roles and verifies all the proposed postulates.