Applying analogical problem solving to mechanical design

Abstract

Design automation holds great benefits for mechanical-product development. In addition to saving engineers from having to carry out redundant tasks, mechanical design automation can also provide embodiment of knowledge, reduced dowstream manufacturing costs, reduced manual errors and more reliable designs. Most of the approaches to mechanical design automation, thus far, have required a large amount of domain-specific knowledge (e.g. expert systems), and/or have had to presume a particular style of design problem solving (e.g. top-down decomposition, bottom-up constructive). The paper proposes analogical problem solving as an approach for alleviating some of these inherit problems in mechanical design automation. Analogical problem solving is based on the fundamental principle that problem solving can be assisted by the review of solutions to past problems that have been attempted. The technique and issues related to the application of analogical problem solving to mechanical design are presented.