Self-stabilizing Distributed Algorithms for Defeat Status Computation in Argumentation

Argumentation is receiving an increasing attention as a technique for practical and uncertain reasoning underlying the realization of intelligent autonomous agents. In line with the evolution of agent architectures towards distribution, we propose a distributed approach to argumentation, where independent processes construct arguments and revise their defeat status by exploiting local information only. We present and compare two distributed self-stabilizing algorithms for defeat status computation, one specifically tailored to rebutting defeat, and the other able to handle any form of defeat. The property of self-stabilization is enforced to ensure a globally correct behavior of the system.