ABSTRACT
This chapter discusses the first class of granular models. It introduces a concept of granular models and shows that information granularity is an inherent conceptual and algorithmic component of such models. The chapter discusses various scenarios of design of granular models, especially distinguishing between passive and active modes of interaction. It presents one of the interesting scenarios for capturing sources of knowledge in the case when the individual sources of knowledge-fuzzy models are treated as a collection of fuzzy rule-based models. In the active approach the original sources of knowledge involved become affected and could be modified based on the feedback mechanism already formed at the higher level of the overall hierarchy. Various architectural versions of granular modeling are covered as well. The chapter distinguishes two general approaches in the hierarchical reconciliation of knowledge, which depend on the way in which the knowledge is being utilized. These are passive approach and active approach.
