Nonmonotonic Reasoning about Action and Change

John Gooday and Antony Galton

In Hans Jürgen Ohlbach (editor), Temporal Logic: Proceedings of the ICTL Workshop, Max-Planck-Institut für Informatik, technical report MPI-I-94-230, June 1994, pages 77-84.

Abstract

Reasoning about action and change is of fundamental importance to many areas of AI, such as planning, diagnosis, and qualitative modelling. In this paper we present a new formalism for representing and solving action and change problems in a straightforward manner. In particular, we introduce a preference strategy that assigns penalty values to models containing information that cannot be explained by observations, inertia, motivated actions or ramifications. We demonstrate the simplicity and effectiveness of the system by applying it to a number of representative problems, and compare our approach with other formalisms.