Case-Based Generation of Design Concepts

Abstract

Abstract Knowledge accumulated from past design experience would help the design engineer in one of two ways: relating the functional requirements of the artifact being designed to parts, assemblies, and systems, or assembling independent parts, assemblies, or systems to achieve the desired functional requirements. In the natural thinking process of most of the engineering designs, those parts, assemblies, systems are associated with some typical design cases. In the perspective of domain knowledge transformation (DKT), the process of case-based reasoning (CBR) is a knowledge state transformation from old design cases to a new design solution. This paper attempts to highlight CBR in generating design concepts (solutions) by knowledge-based design system. Starting with a brief introduction of CBR, the discussions focus on the major challenges of deploying the CBR paradigm in the conceptual design of mechanical systems, viz. how to describe cases, represent and retrieve past cases to be used as candidates of a solution to a new design problem, and so forth. These issues are discussed particularly in the light of the design of positioning systems. A case study is presented to illustrate CBR in generating design solutions.