Generative design is a cutting-edge software that allows engineers and manufacturers to determine the “best possible” designs. Buildings and objects created by generative design algorithms bear a striking similarity to organic shapes like bones and branches, suggesting that those shapes are some of the strongest and most efficient possible.
But engineers working with generative design algorithms aren’t the first to realize this. One architect, working at the beginning of the 20th century, arrived at generative design’s conclusions through painstaking work with physical models.
Antoni Gaudi was a Spanish architect remembered for his fantastical, one-of-a-kind buildings. His unpredictable curves seem like an extravagant luxury, but there’s a purpose to them: Gaudi designed his buildings based on organic forms because of their strength and economy of material.
“Nothing is invented, for it's written in nature first,” Gaudi is famous for saying, and he also famously used living things (like plants and animals) as references for his designs. Their influence is obvious: his Casa Batllo has columns shaped like bones, and the railings of his Casa Vicens are shaped like palm leaves.
But Gaudi didn’t draw on nature just because he thought it was beautiful; to him, form needed to serve function. To create the strongest and most efficient designs, the methodical Gaudi spent hours in his workshop creating and testing 3D models of his designs. These models, made of clay, stone and rope, let him test and scrap designs. Through this testing process, Gaudi found that organically-inspired designs were the strongest and most efficient.
Perhaps one of the best examples of nature influencing design was his unfinished masterpiece, Barcelona’s Sagrada Familia cathedral. Gaudi wasn’t keen on the large buttresses of gothic architecture, and wanted to save material while creating a stronger design. In his final design, the columns looked like trees with branches at the top, with the “branches” inclined at the perfect angle to support the roof without buttresses. Designing this way allowed Gaudi to replace the cornerstones with skylights to let in the sun, and also ensured that the building wouldn’t collapse if a single column fell. Using meticulous testing and drawing on ideas from nature, Gaudi developed an entirely new way of constructing columns.
After Gaudi’s death, his architecture was seen as too complex and ornamented to be truly practical. But his organic shapes are getting a boost from an unexpected place: the world of generative design.
In generative design, the human designer first decides on the constraints and objectives of their final design: for example, if you’re designing a chair, you might decide that it needs four legs and it needs to be able to support an adult human, and then also decide you want to minimize the amount of material used. The designer then puts those parameters into a generative design algorithm, which can test thousands of possible designs in seconds and decide on the one that fits the parameters the best. In our chair example, the algorithm would run all the possible chair models with four legs and the capacity to support an adult and decide on the one that used the least material.
Many people expected generatively-designed items to look minimalist and clean, with straight lines and flat planes. Instead, the designs look asymmetrical and curvy, filled with holes and meshes. They look, in other words, like the bones and branches that support natural structures.
But why do these designs, produced by an artificial algorithm, look so natural? Like a generative design algorithm, natural selection has been able to “experiment” with multiple forms for things like bones and wings and has “settled on” the forms most adapted to the creature’s environment.
“Just like nature has optimized for weight and improved stiffness, so too have these algorithms,” says Jordan Brandt, Technology Futurist at Autodesk. “We’re essentially running accelerated artificial evolution," says Brandt.
For now, generative design has been popular in manufacturing, but hasn’t quite made the jump to architecture. But industry experts believe that programs like Autodesk’s DreamCatcher might help design the buildings of tomorrow. “[Generative design] is your future,” Steve Jones, director of Dodge Data and Analytics, told engineering.com. “After 13 years of study in this field, this is the best advice I can give you right now.”
Gaudi, working with his physical models, probably couldn’t have imagined the digital future of architecture today. But his ideas have been vindicated by time and technology, and his influence lingers on.