An effective similarity determination model for case-based reasoning in support of low-carbon product design

Abstract

In the initial stage of low-carbon product design, design information is always uncertain and incomplete, as well as the coupling properties between design attributes, thus it requires retrospective coordination for design conflicts resulting from the inclusion of low-carbon requirements. Reusing the prior design knowledge can promote design efficiency, however, the acquisition of similar cases knowledge not only needs to consider the similarity of design problems, but also the adaptability of candidate cases. This study presents an effective similarity determination model to support low-carbon product design, and targets of the proposed model are (1) to reasonably determine design ranges of attribute values for product cases retrieval by representing the uncertain design attributes with fuzzy set theory; (2) to construct an efficient indexing structure to generate the index set of similar cases based on the improved discretized highest similarity method by proposing two effective strategies; and, (3) to establish similarity estimation models for different types of attributes, and it calculates the information content of each attribute to evaluate the adaptability of cases based on the Information Axiom. The applicability of the proposed model is demonstrated through a case study of similar cases retrieval for the vacuum pump low-carbon design.