On the impact of configuration on abstract argumentation automated reasoning

Abstract

In this paper we consider the impact of configuration of abstract argumentation reasoners both when using a single solver and choosing combinations of framework representation-solver options; and also when composing portfolios of algorithms. To exemplify the impact of the framework-solver configuration we consider one of the most configurable solvers, namely ArgSemSAT-runner-up of the last competition on computational models of argumentation (ICCMA-15)-for enumerating preferred extensions. We discuss how to configure the representation of the argumentation framework in the input file and show how this coupled framework-solver configuration can have a remarkable impact on performance. As to the impact of configuring differently structured portfolios of abstract argumentation solvers, we consider the solvers submitted to ICCMA-15, which provided the community with a heterogeneous panorama of approaches for handling abstract argumentation frameworks. A superficial reading of the results of ICCMA-15 is that reduction-based systems (either SAT-based or ASP-based) are always the most efficient. Our investigation, concerning the enumeration of stable and preferred extensions, shows that this is not true in full generality and suggests the areas where the relatively underdeveloped non reduction-based systems should focus more to improve their performance. Moreover, it also highlights that the state-of-the-art solvers are very complementary and can be successfully combined in portfolios.

How we designed winning algorithms for abstract argumentation and which insight we attained

“…10 http://argumentationcompetition.org/2017/results.html (on 6th February 2019). 11 We refer any reader interested in knowing the effect of different SAT solvers configuration parameters when addressing argumentation problems to [13]. (Figure 8), while still being the one with highest IPC score ( Figure 5).…”

“…This is hardly a surprise since those benchmarks were specifically designed to have a large number of preferred extensions, but only one semi-stable extension. 13 Therefore for these benchmarks providing additional constraints on the undecided arguments proves to be substantially advantageous.…”

“…We refrain from further comparison with state of the art approaches as this has been the subject of ICCMA 2015 [27,53]; a re-run of the competition to further analyse the results [13]; and a survey of various approaches [54], which in turn has been superseded by a chapter in the Handbook of Formal Argumentation [55], also published as a journal paper in [56]. We refer an interested reader to those references for further analysis on other techniques for solving problems in abstract argumentation.…”

“…While different approaches turn out to be complementary and can be successfully combined in portfolios [13], reduction-based approaches, most of them based on SAT, had the best performance during ICCMA 2017, 2 the last competition. It is worth noticing that this is not a universal truth as we already discussed at length in [13].…”

“…http://argumentationcompetition.org/2017/SemBuster.pdf (on 08 February 2019)13 Semi-stable semantics attempts to minimise the number of undecided arguments.…”