Evolution of supershapes for the generation of three-dimensional designs

Preen, Richard J.; Bull, Larry

Abstract

This paper explores the evolution of three-dimensional objects with a simple generative encoding, known as the Superformula. Evolving three-dimensional objects has long been of interest in a wide array of disciplines, from engineering (e.g., robotics) to biology (e.g., studying morphological evolution). While many representations have been presented, ranging from direct encodings to complex graphs and grammars, the vast majority have possessed complex underlying encodings, which were necessary to produce varied morphologies. Here we explore the target-based evolution of Superformula, which is simply encoded as a vector of reals and show that is possible to generate very closely matching designs of a number of complex three-dimensional objects.

• Journal Article Type: Article
• Publication Date: Apr 18, 2012
• Journal: arXiv
• Peer Reviewed: Not Peer Reviewed
• Keywords: evolution, supershapes, neural and evolutionary computing, computational geometry, three dimensional designs, encoding, superformula
• Public URL: https://uwe-repository.worktribe.com/output/948135
• Publisher URL: http://arxiv.org