Hierarchical Diagnosis of Large Configurator Knowledge Bases

Debugging, validation, and maintenance of configurator knowledge bases are important tasks for the successful deployment of product configuration systems. Consistency-based diagnosis has shown to be a promising approach for detecting faulty parts in the knowledge bases and explaining unexpected behavior of the configurator, whereby (partial) configurations are used as test cases. In this paper we show how hierarchical diagnosis can be employed to cope with the complexity of debugging large configurator knowledge bases. A framework for hierarchical diagnosis on different levels of abstraction is presented as well as an algorithm for the calculation of diagnoses on those levels. The presented approach aims at the reuse of existing special purpose configuration systems. We show that the exploitation of hierarchies in such problem domains leads to a significant efficiency enhancement thus broadening the applicability of consistency-based diagnosis.