Design Concept Generation: A Hierarchical Coevolutionary Approach

Abstract

As design problems become more complex and design lead time more pressing, designers need effective support tools to expand their design space exploration. In this paper, a hierarchical coevolutionary approach is proposed to support designers by automatically generating design concepts based on the designers’ inputs. The approach adopts a zigzag design process in which function structures and their corresponding solution principles coevolve in parallel across different levels of an abstraction hierarchy. A grammar-based approach is applied to decompose higher-level functions and generate an initial population of function structures. From this initial population, a coevolutionary algorithm is devised to coevolve more function structures and their corresponding solution principles. A case study of designing a mechanical personal transporter is presented to demonstrate the effectiveness and features of the proposed approach.