Modular Surface Evolution

Biraj Ruvala
Emergent Technology and Design
Architectural Association

Abstract

This research began with the use of GENR8, a software package that combines growth languages with evolutionary search methods and that uses grammatical evolution to specify the growth and evolution of surfaces. The software plugs into Wavefront's Maya, and combines 3D map L-systems that are extended to an abstract physical environment with evolutionary computation. The aim of this experiment was to set up the parameters that are used to evolve a family of surfaces, to understand their morphology and to relate this to the concepts of emergence and evolution. The parameters used in the experimental set up are determined through the growth model. Using this, an algorithm is developed to generate an entire array of surfaces. These surfaces are selected based on the fitness criteria options available within GENR8. The surfaces, being the genomes, are introduced into environments. The environment is a boundary box of a fixed size and the boundary conditions of each are recorded. Using this information, a global surface is set up using the phenotype surface as a tiling component. These global surfaces then are reused as genomes and are tested based on a secondary fitness criteria.

The parameters for the evolution of the surfaces through a MEL script

Analysing the Edge Conditions