Bio-inspired Knowledge Representation Framework for Decision Making in Product Design

Abstract

A lot of information, mostly disorganized, is available to the designer during early stages of design in the form of raw data. To extract useful information from raw data, its storage and analysis is important, for which knowledge representation plays prominent role. While new product variants are generated, a few previous designs are generally accessible to the designers to take inspiration. If efficient knowledge representation of relationships among previous designs exists, it would help the designers in new product development. In this work, analogy is sought from biological phenomenon in nature, i.e., phylogenetic, to store, depict, and retrieve knowledge from previous design concepts in the form of phylogenetic tree. The tree is developed using two Phenetic approaches, i.e., unweighted pair group method with arithmetic mean (UPGMA) and neighbor joining algorithm. The tree of UPGMA represents similarity of product features for different design concepts, and the tree of neighbor joining depicts number of modifications a designer performs to develop a new variant from the previous variants. An example of power drill is taken to illustrate the application of both the algorithms in product design.