On the Computation of Naive-based Argumentation Semantics

Master's thesis by Javanshir Alammadli

Supervisor Sebastian Rudolph, Sarah Alice Gaggl
Computational Logic
8 September 2015 – 14 März 2016

The increasing interest in Argumentation over the last two decades led to several proposals for formalisms. They differ in many aspects, therefore, implementations are needed for a better analysis, evaluation and comparison.

These implementation methods can be divided into two categories, the reduction and direct approach. The idea of reduction approach is to translate the problem into another language, for which sophisticated systems already exist. In direct approach the methods and algorithms are implemented from scratch. For instance, implementation of algorithms based on argument labelling functions. While there exist several implementations for the semantics introduced by Dung, there are not so many implementations when it comes to naive-based semantics. Our contribution is to develop a program using serial and concurrent computation techniques in the programming language Java for computing all naive, cf2, stage and stage2 labellings for a given AF and compare its performance with ASPARTIX, which is the only system computing all naive-based extensions to date.