Abstract
We propose a knowledge management method using indexing based on physical quantity dimension space. All physical quantities appearing in the past, present and future design problems can be represented by seven-dimensional vectors based on seven fundamental units of SI. Our key idea is to use this seven-dimensional vector (quantity dimension vector) for indexing knowledge from the viewpoint of physical phenomena. We mathematically define similarity between two physical quantities based on quantity dimension vector. Then we classify design knowledge into three patterns and define index representing those pattern similarity and physical quantity similarity. We prepare some sample knowledge and use our method to estimate similarity between knowledge and to retrieve knowledge relevant to current design problem. As a result, we confirmed that our approach is feasible for design knowledge management.
