| description abstract | Product family design (PFD) is a widely adopted strategy for product realization, especially when design requirements are diversified and multifaceted. Due to everchanging customer needs and the increasingly complex and integrated product design structure, PFD and its optimization have been concerned more about a rapid and contextual product analysis and variant derivation based on a multiobjective optimization scheme subject to design concerns, which are often cross disciplinary, such as product service, carbon footprint, user experience, esthetics, etc. Existing PFD modeling approaches, which are primarily structured using component attributes and assembly relationships, possess notable limitations in representing complex component and design relationships. Hence, it has restricted comprehensive PFD analysis in an agile and contextual manner. Previously, we have studied and demonstrated the feasibility of using ontology for product family modeling and have suggested a framework of faceted information search and retrieval for product family design. In this paper, several new perspectives towards PFD based on ontology modeling are presented. Firstly, new metrics of ontologybased commonality that better reveal conceptual similarity under various design perspectives are formed. Secondly, faceted concept ranking is proposed as a new ranking approach for ontologybased component search under complex and heterogeneous design requirements. Thirdly, using these ranked results, a platform selection approach that considers a maximum aggregated ranking with a minimal platform modification among various platform choices is researched. From the selected platform and the newly proposed metrics, a modified multiobjective evolutionary algorithm with an embedded feature of configuration incompatibility check is studied and deployed for the optimal selection of components. A case study of PFD using four laptop computer families is reported as our first attempt to showcase how faceted component analysis, selection, and optimization can be accomplished based on the proposed family ontology. | |
