Optimal status enforcement in abstract argumentation
Andreas Niskanen, Johannes Peter Wallner, Matti Järvisalo · International Joint Conference on Artificial Intelligence · 2016
We present complexity results and algorithms for optimal status enforcement in abstract argumentation. Status enforcement is the task of adjusting a given argumentation framework (AF) to support given positive and negative argument statuses, i.e., to accept and reject specific arguments. We study optimal status enforcement as the problem of finding a structurally closest AF supporting given argument statuses. We establish complexity results for optimal status enforcement under several central AF semantics, develop constraint-based algorithms for NP and second-level complete variants of the problem, and empirically evaluate the procedures.