AE Climb: Manual Design to Automated Design to Generative Design
Design automation is the ability to script computational processes that reflect the creative and analytical methods of human driven design using specific rules and instructions. That’s a mouthful that boils down to taking a manual process and letting a computer help you do it faster. Time is spent up front making the script which will save time in the long run by automating repetitive tasks. Autodesk Dynamo works great for design automation in architecture and engineering. Using Dynamo, designers can create visual and graphical logic to explore conceptual Revit designs that update parametrically. Projects can be visualized using precision surface and solid geometry, and iterations can be explored and analyzed faster.
Generative design uses the rules laid out with design automation in order to create many versions and variations of possible designs. Instead of one outcome being evaluated at a time, the process can be completed at a much faster rate with computer assistance. These include options the user may not have even thought of due to the limited ability to create numerous iterations and compare them. Generative design helps lead to design optimization by comparing different outcomes of parameter variables to the given goals. As Applied Software Senior Specialist Christopher Riddell describes it, “Often you will hear generative design referred to as co-design. The computer and the human collaborate to produce an outcome. The human sets the rules and constraints and the computer uses its processing power to iterate through and optimize based on the given constraints.”
Applied Software Developer Carlo Licea explains, “Generative design also reduces cost, not only through optimization of performance and material use, but also by reducing the time dedicated to finding optimal solutions. Dynamo plus Refinery is a common technology combination for achieving generative design. Refinery uses multiple runs of Dynamo scripts to graph values of parameters and show different outcomes.”
Autodesk used the Dynamo+Refinery technology a few years ago to generatively design the layout of two of the three office floors in its new Toronto office. Referred to as Project Discover, the design approached multiple goals – daylight, space adjacencies, square footages, and form constraints. A 2017 article in the Journal of the American Institute of Architects describes the process as software using algorithms primed with occupant preferences to determine ideal office arrangements.
Design automation and generative design are both user-defined processes where the user is setting the rules and flow of the design parameters. Licea says, “Numerous times I’ve encountered people who conceptually thought of generative design as a means of letting a computer come up with a design for them, but it’s actually the user(s) telling the computer how to design and letting the computer create the possible outcomes within given parameters.” So it’s not “artificial intelligence” that spontaneously comes up with a design. The capabilities are limited to the rules and data supplied by the user.
The next step? Machine learning is the next logical stage beyond generative design. With it, the code uses the results of generative design to “predict” optimal outcomes from the design patterns encountered. No industry is exempt from machine learning and development of disruptive technologies; see the Applied Software blog for “How Manufacturing is Evolving.”
The experts at Applied Software stay up-to-date on the latest technology that affects AE, construction and manufacturing design processes. To learn more about Dynamo or the other software tools that can speed up and improve your design workflow, contact Applied today for a quick discovery call. Applied Software will be your handrail on the AE climb.
