Operationalization of Interpretation Models for Model-Based Diagnosis

Interpretation models are generic models of problem solving, describing abstractly data types and inference steps of the problem solving process that process these data. They are a frequently used instrument to specify knowledge level models. This report surveys interpretation models for model-based (consistency-based) diagnostic systems that are based on the General Diagnostic Engine (GDE) paradigm. Starting with an interpretation model for standard GDE, a model of a GDE with fault models (GDE+FM) and one of a GDE with fault models and hierarchical refinement (GDE+FM+H) are presented. We will also discuss possible views (links of interpretation models) to existing implementations within the GenDE framework, which we call operationalization. The groundwork for this approach (theory and implementations published in literature) is described. Finally, problems of this type of operationalization are discussed and conclusions for the contribution of these interpretation models to repositories in order to support their reuse and for the further development of the conceptual modeling language MoMo are derived.