Intelligent Techniques for Configuration Knowledge Evolution

Abstract

Automated testing and debugging of knowledge bases (such as configuration knowledge bases and feature models) is an important contribution to manage knowledge evolution efficiently. However, existing approaches rely on the assumption of consistent test suites which are always kept up-to-date within the scope of different knowledge base maintenance cycles. In this paper we introduce diagnosis techniques that actively guide stakeholders (knowledge engineers and domain experts) in the process of testing and debugging knowledge bases. These techniques take into account faulty test cases and constraints and recommend diagnoses which are the source of a given inconsistency.