Daniel Richards and Martyn Amos
School of Computing, Mathematics and Digital Technology, Manchester Metropolitan University, Manchester, United Kingdom
In this chapter, we consider the application of evolutionary algorithms and regulatory networks to problems in architectural design. The “cybernetic theory of architecture” dates back to 1969, when Pask  predicted that “Various computer-assisted (or even computer-directed) design procedures will be developed into useful instruments” for the design of buildings and cities. Pask’s ideas about control and communication were developed by architects such as John Frazer, who were interested in how concepts of adaptation might be applied to the design, construction and evolution of building form and performance. In An Evolutionary Architecture, Frazer sets out how natural processes might be harnessed as “the generating force for architectural form” . Importantly, he seeks to go beyond the rigid and specific “blueprint” metaphor, and to develop a “genetic language of architecture”, in which form-generating processes give rise to structure and behavior. Since then, evolutionary design in architecture has become well-established [6, 7, 16]. However, many such syntheses have still been modelled on what Oxman  calls the cyclical “generate and test” paradigm, in which conceptual design generation is followed by performance evaluation. Oxman argues ...