About Generative Design

Generative systems offer a methodology and philosophy that view the world in terms of dynamic processes and their outcomes. In the terminology of Thomas Kuhn (1996), it offers a paradigm shift for the process of design and the expression of that process. For designers, it involves a reconsideration of the static artifact and the actions that manipulate it. Conceptualization shifts from the primacy of objects to envisaging interacting components, systems and processes, which in turn generate new artifacts, with special properties (outlined below). The generative methodology offers an unconventional way of conceptualizing and working in design. Research in generative systems is closely tied to the general concept of synthesis, most viscerally apparent in nature and natural systems. Nature has devised a specific mechanism for generalized synthesis, using the physical apparatus of DNA, protein synthesis and biochemistry. The diversity and adaptability of life on Earth demonstrates the potential of these mechanisms to overcome problems in design and to generate novelty and diversity from relatively simple units. The key properties of generative systems can be summarized as:
• The ability to generate complexity, many orders of magnitude greater than their specification. This is commonly referred to as database amplification, whereby interacting components of a given complexity generate aggregates of far greater behavioral and/or structural complexity. Such aggregates may in turn generate their own interactions forming new aggregates of even higher sophistication and complexity. This is referred to as a dynamic hierarchy. A poignant example being complex multi-cellular organisms, whose hierarchy can be summarized: atom; molecule; organelle; cell; organ; organism; ecosystem.
• The complex and interconnected relationship between organism and environment. Organisms not only evolve and adapt to their environment, their presence and number may effect and change the environment itself. Inter- and intra-species dependencies form a complex web of relations (an ecosystem), within which there are often many feedback loops. These systems are typically homeostatic. That is, they actively maintain their state in order to offset environmental changes. 
• The ability to self-maintain and self-repair. Human-designed structures are typically brittle either in a physical or functional sense. As stated above, generative systems may adapt themselves to maintain stable configurations within a changing environment. Swarm systems for example can overcome significant disruption and individual loss, reforming and adapting their behavioral function to survive. They exhibit fault-tolerance and have a high degree of internal redundancy, giving them the ability to overcome changes that would limit a more fragile design.
• The ability to generate novel structures, behaviors, outcomes or relationships. Novelty used in this sense means the quality of being new, original and different from anything else before it. There are of course, different degrees of novelty. RNA and DNA were novel in that they introduced a completely new mechanism for replication and encoding of protein synthesis. Artists and designers are always seeking novelty (the opposite of which is mimicry or copying, something depreciated in the art and design world). Artistic novelty may not have such a significant impact as, for example, DNA, but the key concept is that of the new — generative systems have the potential to give rise to genuinely new properties. This is why they are often referred to as emergent systems. These new properties typically fall outside the designer’s expectations or conceptualizations for the design, resulting in functionality or outcomes that were not anticipated. This of course raises the issue of control, a problematic issue for generative design, particularly if the designer is accustomed to organizing outcomes in a predictable way.

Jon McCormack, Alan Dorin and Troy Innocent, 2004

References

Thomas Kuhn, The Structure of Scientific Revolutions (Third Edition), University of Chicago Press, Chicago, Ill., 1996.

Jon McCormack, Alan Dorin and Troy Innocent, Generative Design: a paradigm for design research, in: J. Redmond, et. al. (eds), Proceedings of Futureground, Design Research Society, Melbourne, 2004.