Towards Formalizations in Case-Based Reasoning for Synthesis

Similarity and Adaptation

This paper presents the formalization of a novel approach to structural similarity assessment and adaptation in case-based reasoning (CBR) for synthesis. The approach has been informally presented, exemplified, and implemented for the domain of industrial building design (B6rner 1993). By relating the approach to existing theories we provide the foundation of its systematic evaluation and appropriate usage. Cases, the primary repository of knowledge, are represented structurally using an algebraic approach. Similarity relations provide structure preserving case modifications modulo the underlying algebra and an equational theory over the algebra (so available). This representation of a modeled universe of discourse enables theory-based inference of adapted solutions. The approach enables us to {ncorporate formally generalization, abstraction, geometrical transformation, and their combinations into CBR.