A Numerical Optimization Search Strategy for Exploring Morphological Charts

Abstract

We present a numerical optimization search strategy for exploring design concepts from morphological charts. An initial pool of creatively produced design concepts are decomposed in a general manner, based on function, form, or aesthetics. Feature options that solve a common function are grouped into rows of the morphological chart. Traditionally, designers manually select promising combinations of features to form alternative concepts. The numerical search operations presented here are dened to automatically generate new concepts by recombining interchangeable features and descriptive traits. Genetic algorithms are used to computationally combine features from the morphological chart and nd the best performing concepts according to calculated objective values. The level of resolution in the results is determined by the level to which the concepts are decomposed by the designer and performance is described by the objective’s equations. The strategy couples the speed of computational searches with creativity and insight of human designers. It is demonstrated with a simple product example to generate new, well suited concepts that designers may not have otherwise discovered. A nal set of non-dominated designs is presented to the designer with the quantitative objective values and simplied visual representations of the concepts to aid in the qualitative evaluation of the designs. The proposed framework brings the ability to quickly generate numerous, optimally performing solutions across a wide solution space, in an ecient and human-centered way, and does so in the early stages of design.