Extended Defeasible Reasoning for Common Goals in n-Person Argumentation Games

Duy Hoang Pham (Posts and Telecommunications Institute of Technology, Vietnam)

Guido Governatori (NICTA, Queensland Research Laboratory, Australia)

Subhasis Thakur (Griffith University, Australia)

Abstract: Argumentation games have been proved to be a robust and flexible tool to resolveconflicts among agents. An agent can propose its explanation and its goal known as a claim, which can be refuted by other agents. The situation is more complicated when there are morethan two agents playing the game.

We propose a weighting mechanism for competing premises to tackle with conflicts from multipleagents in an n-person game. An agent can defend its proposal by giving a counter-argument to change the "opinion" of the majority of opposing agents. Furthermore, using the extendeddefeasible reasoning an agent can exploit the knowledge that other agents expose in order to promote and defend its main claim.

Keywords: argumentation systems, artificial intelligence, defeasible reasoning

Categories: I.2.4