Ontology-based product knowledge framework and its application to case-based reasoning

Abstract

Participants in a product lifecycle want to share comprehensive product knowledge without ambiguity or heterogeneity. In addition, they want to infer new knowledge using reasoning services. Although numerous product data managing tools and expert systems have recently been introduced to manufacturing companies, knowledge is not well organized among these systems. Comprehensiveness, heterogeneity, and indefiniteness of product knowledge complicate organization of the product knowledge. These problems restricts expert systems’ inference ability. To make matters worse, some expert systems use different reasoning services from others without considering characteristics of product knowledge. We adopt ontology-based knowledge modeling and reasoning paradigm for product knowledge modeling and reasoning. The ontology approach enables us to clarify the structure of product knowledge as well as hybrid knowledge reasoning such as axiom-based reasoning, rule-based reasoning and case-based reasoning. In this thesis, we propose a systematic lattice structure of product knowledge; an ontology-based multi-layered knowledge framework (OBMKF) which has three product knowledge types and four layers. The three knowledge types are axioms, knowledge maps, and specialized knowledge for a domain. The four layers consist of a product context model, a product specific model, a product planning model, and a product manufacturing model. These layers represent different abstraction levels of the product knowledge conceptualization. In addition, we suggest a product ontology and product ontology language to reduce user’s effort for his ontology construction. Our product ontology is constructed based on a top-level ontology so that its axioms include fundamental properties of concepts. Our product ontology language has product knowledge primitives defined in the product ontology, and it can be transformed to OWL. So users can define the specific product knowledge based on our langua...