A preliminary study of computational complexity in non-monotonic reasoning

Abstract

In this work we analyze existing complexity results in the area of non-monotonic reasoning in general and argumentation in particular. Even though the area of argumentation is based on solid theoretical foundations, its main problems rely on the computational complexity of the system that have so far been developed. In order to use argumentation in real time scenarios we must find an implementation with a reasonable response time. Complexity analysis of argument systems is an indispensable tool for addressing this taks. We expect that the development of this research line will result in a general analysis of the issues in complexity of argument systems, leading to an efficient implementation of a particular formalism, observation-based defeasible logic programming, that could be integrated in an intelligent agent architecture.