Abstract

Expert systems (ES) usually generate extensive inference-trees before showing to users a definitive result related to a complex dynamic system (DS) behavior. These inference-trees are not included in the results but it could provide additional information to understand the overall performance of a DS. They contain a set of statements that describe the knowledge about the truths of the DS plus a set of constraints that can give statements that must be true in the DS behavior. This document describes a method to generate explanations based on the conclusions reached by an ES respect to the DS behavior, using a specific ontology and discourse patterns. The input of the method is an intermediate- state tree (the inferencetree) and a specific knowledge-domain represented by the ontology. The document describes the software architecture to generate the explanations and the testing cases designed to validate the results in a complex real domain, such as the copper bioleaching domain.