A first comparison of abstract argumentation systems: A computational perspective

In this paper we introduce an initial comparison among three different implementations of Abstract Argumentation Systems: ASPARTIX, ConArg, and Dung-O-Matic. These tools are tested over four different kinds of interaction graphs, corresponding to Erdo?s-?enyi networks, scale-free Barabasi networks, Watts-Strogatz, and Kleinberg small-world networks. Our final goal is to thoroughly evaluate their performance (in this work we test complete and stable semantics only), and to find the most efficient one, but also, more in general, to better study this kind of problems from the computational point of view.