An Incremental Algorithm for Computing the Grounded Extension of Dynamic Abstract Argumentation Frameworks

Several formalisms have been introduced to model disputes between agents. Abstract argumentation is a simple, yet powerful formalism for modeling disputes by abstracting from the internal structure of arguments. Much work has been done to characterize fast algorithms for ‘static’ argumentation frameworks which are assumed to be fixed, in the sense that they do not change during a dispute. However, argumentation frameworks are highly dynamic in practice. For instance, applications of argumentation for negotiation and persuasion are usually based on protocols where agents state their arguments and attacks one after the other in a dynamic process during which the outcome of the debate evolves. We focus on one of the most popular argumentation semantics, namely the grounded semantics, and deal with the problem of recalculating the extensions of argumentation frameworks after adding or deleting attacks or arguments. In particular, we propose an incremental algorithm for the efficient computation of the grounded semantics, useful in dynamic contexts where argumentation frameworks are continuously updated to consider new information. We report on experiments showing that our incremental algorithm is on average faster than CoQuiAAS, the solver that won the last edition of the international competition on computational models of argumentation for the task of computing the grounded extension of an argumentation framework.