Automated Concept Generation Based on Function-Form Synthesis

Abstract

This work hypothesizes that enhancing next generation products’ distinctiveness through function-form synthesis results in feasible design concepts for designers. A data mining driven methodology that searches for novel function and form candidates suitable to include in next generation product design is introduced in this work. The methodology employs a topic modeling algorithm to search for functional relationships between the current product design and designs from related/unrelated domains. Combining the current product design and candidate products’ form and function, which is acquired from related/unrelated domains, generates next generation design concepts. These resulting design concepts are not only distinct from their parent designs but are also likely to be implemented in the real world by containing novel functions and form features. A hybrid marine model, which is differentiated from both the current design and candidate products in related/unrelated domains, is introduced in the case study in order to demonstrate the proposed methodology’s potential to develop concepts for novel product domains. By comparing the form and function similarity values between generated design concepts, an existing hybrid marine model (Wing In Ground effect ship: WIG), and source products, this research verifies the feasibility of these design concepts.