Dealing with Incompatibilities among Goals

An intelligent agent may in general pursue multiple goals at the same time, which leads to arise some conflict among them. In this paper, we focus on these conflicts or incompatibilities among goals. Our approach is based on the model of Castelfranchi and Paglieri, in which, three forms of incompatibility and the criteria for selection of goals are defined. We characterise computationally the different forms of incompatibility (terminal, instrumental and superfluity) and propose a way, based on abstract argumentation theory, for selecting those goals that will continue to be pursued.