Using defeasible argumentation in progression and regression planning

Abstract

The aim of this work is to study an argumentation-based formalism that an agent could use for constructing plans. Elsewhere we have introduced a formalism for agents to represent knowledge about their environment in Defeasible Logic Programming, and a set of actions that they are capable of executing in order to change the environment where they are performing their tasks. We have also shown that action selection is more involved than expected when combined with a defeasible argumentation formalism. In this paper we will develop a novel way of using argumentation in the definition of actions and combining those actions to form plans. Since our interest here lies in exploring the important issues that need to be addressed, the main contribution will be to show meaningful examples where those issues are exhibited and not in improving current planning implementations. Therefore, we will use simple planning algorithms in an effort to reduce the complexity of the examples. Nevertheless, as the different ways of constructing plans introduce interesting details we will be considering progression and regression planning.