Defeasible Argumentation-based Epistemic Planning with Preferences

Abstract

Many real-world applications of intelligent systems involve solving planning problems of different nature, oftentimes in dynamic environments and having to deal with potentially contradictory information, leading to what is commonly known as epistemic planning. In this context, defeasible argumentation is a powerful tool that has been developed for over three decades as a practical mechanism that allows for flexible handling of preferences and explainable reasoning. In this article, we first motivate the need to develop argumentationbased epistemic planning frameworks that can be leveraged in real-world applications, describe the related literature, and then provide an overview of a recently-proposed approach to incorporate defeasible argumentation and preferences into automated planning processes. In particular, the framework incorporates conditional expressions to select and change priorities regarding information upon which plans are constructed. We describe its main properties, analyze its strengths and limitations using an illustrative use case, and discuss several future research directions that can be taken to further develop it.