Actions and arguments: preliminaries and examples

Abstract

The goal of this work involves the developing of an argumentation-based formalism that an agent could use for constructing plans. In this formalism, agents will have certain knowledge about the world, and a set of actions that they will be capable of execution. The agent’s knowledge will be represented by a knowledge base, using Defeasible Logic Programming. Actions will provide agents the ability of change their world, adding or removing facts from their knowledge base. We will develop here a formalism for representing actions in a defeasible argumentation environment. We will show how an agent may select a sequence of actions in order to satisfy some goal, and this action selection is more involved than expected.