Merging Aesthetics with Functionality: An Interactive Genetic Algorithm Based on the Principle of Weighted Mutation
This paper proposes an algorithm through which the development of computationally generated forms can be externally directed towards both functional objectives and intuitive design targets. More precisely, it presents a special version of Interactive Genetic Algorithm, which introduces Weighted Mutation as a method to support the long life of genes corresponding to favored phenotypic characteristics. At the same time, optimization processes towards operational goals are also enabled. A set of experiments is conducted on the case study of a building façade, which appears to provide a suitable means for the investigation of functional, as well as aesthetic issues. The results are positively assessed, since they prove that this new methodology broadens the capacities of standard Interactive Genetic Algorithms, shedding light on how a constructive human-machine relationship can benefit the design process.
KeywordsInteractive Genetic Algorithm Weighted Mutation Human Evaluation Algorithmic Design Control
Unable to display preview. Download preview PDF.
- 1.Baker, E., Seltzer, M.: Evolving Line Drawings. In: Proceedings of the Fifth International Conference on Genetic Algorithms, pp. 91–100. Morgan Kaufmann Publishers, San Francisco (1994)Google Scholar
- 2.Cho, S.-B.: Towards Creative Evolutionary Systems with Interactive Genetic Algorithm. In: Applied Intelligence, vol. 16, pp. 129–138. Kluwer Academic Publishers, Dordrecht (2002)Google Scholar
- 3.Dawkins, R.: The Blind Watchmaker. Longman Scientific and Technical, Harlow (1986)Google Scholar
- 7.Sasaki, M.: Morphogenesis of Flux Structure. From the realization of Free 3D Forms by Architects to Evolving the Most Logical Structure by Computer Calculation. In: Sakamoto, T., Ferré, A.T. (eds.) From Control to Design. Parametric/Algorithmic Architecture, pp. 68–69. Actar, Barcelona (2007)Google Scholar
- 9.Torres, S.L., Sakamoto, Y.: Façade Design Optimization for Daylight with a Simple Genetic Algorithm. In: International Building Performance Simulation Association, pp. 1162–1167 (2007), http://www.ibpsa.org/proceedings/BS2007/p117_final.pdf