EVOLUTIONARY DESIGN SYNTHESIS COMPARISON: GROWTH & DEVELOPMENT VS. FIXED-MESH CELLS

DS 68-10: Proceedings of the 18th International Conference on Engineering Design (ICED 11), Impacting Society through Engineering Design, Vol. 10: Design Methods and Tools pt. 2, Lyngby/Copenhagen, Denmark, 15.-19.08.2011

Year: 2011
Editor: Culley, S.J.; Hicks, B.J.; McAloone, T.C.; Howard, T.J. & Dong, A.
Author: Yogev, Or; Shapiro, Andrew A.; Antonsson, Erik K.
Series: ICED
Section: Design Methods and Tools Part 2
Page(s): 402-412

Abstract

Two design synthesis techniques, both utilizing evolutionary computation, are compared. One employs growth and development of a model composed of finite element cells; the other employs a pre-defined fixed mesh of finite element cells. Both of these processes use a similar set of rules. In the fixed-mesh case, some of the rules present in the growth development case have been disabled to preserve the fixed mesh. The development process starts with a single cell while the fixed-mesh method starts with a predefined volume of material comprising 640 cells. The fitness of individuals in both approaches is determined by evaluating a combination of four different performance objectives: a desired resonant frequency; an ability to support a stochastically varying external load; a desired vertical height; and a minimum weight. A solution with both types of models is found in a relatively small number of generations. However, the growth development approach generated design configurations with valuable engineering characteristics, including symmetry, smoothness, and regularity, while the fixed-mesh approach generated non-smooth configurations with porosity and inhomogeneity.

Keywords: INDIRECT ENCODING; GENETIC ALGORITHM; EVOLUTION; DEVELOPMENT; DESIGN SYNTHESIS

Download

Please sign in to your account

This site uses cookies and other tracking technologies to assist with navigation and your ability to provide feedback, analyse your use of our products and services, assist with our promotional and marketing efforts, and provide content from third parties. Privacy Policy.