GENE_ARCH: An evolution-based generative design system for sustainable architecture

Abstract

GENE_ARCH is an evolution-based Generative Design System that uses adaptation to shape energy-efficient and sustainable architectural solutions. The system applies goal-oriented design, combining a Genetic Algorithm (GA) as the search engine, with DOE2. IE building simulation software as the evaluation module. The GA can work either as a standard GA or as a Pareto GA, for multicriteria optimization. In order to provide a full view of the capacities of the software, different applications are discussed: 1) Standard GA: testing of the software; 2) Standard GA: incorporation of architecture design intentions, using a building by architect Alvaro Siza; 3) Pareto GA: choice of construction materials, considering cost, building energy use, and embodied energy; 4) Pareto GA: application to Siza's building; 5) Standard GA: Shape generation with single objective function; 6) Pareto GA: shape generation with multicriteria; 7) Pareto GA: application to an urban and housing context. Overall conclusions from the different applications are discussed. © Springer-Verlag Berlin Heidelberg 2006.