Indirect Encoding of the Genes of a Closed Curve for interactively create innovative car Silhouettes

Bernard Yannou (1) , Markus Dihlmann (2) , François Cluzel (1)

1 LGI - Laboratoire Génie Industriel - EA 2606
2 US - University of Stuttgart

An Interactive Genetic Algorithm system is proposed for designing a car silhouette while involving the style designer in the evaluation process of a population of individuals. This IGA is based on the principle of an indirect encoding of a closed curve genome using a primary Fourier decomposition. A crossing over operator is proposed for mixing the parents' genes by a random weighted average into a new child's genome. A perceived similarity index between two genomes is built. It allows to check that our IGA is able to converge toward a targeted individual which was not present within the initial population and, consequently, that any silhouette in a style designer's mind could be revealed by our system.

Mots clés: Interactive Genetic Algorithms; shape design; car profile; Fourier decomposition; evolutionary design

Domaines: Mécanique [physics.med-ph]; Intelligence artificielle [cs.AI]

Bernard Yannou, Markus Dihlmann, François Cluzel. Indirect Encoding of the Genes of a Closed Curve for interactively create innovative car Silhouettes. 10th International Design Conference - DESIGN 2008, May 2008, Dubrovnik, Croatia. pp.1243-1254. ⟨hal-00796974⟩