Any one into Generative Design ?

Without igniting a huge flame war, this reminds me of (most) designers’ impressions of rapid prototyping. Both generative design and rapid protyping technology seem to have a “well, it lessens my importance as a professional, so it must not positive” reaction.

That said, a couple thoughts on generative design. I don’t know much about it, but I can certainly imagine that it would be useful in design (as opposed to art). Assuming that I understand it correctly, it seems likely that various inputs (length, height, width, specific features, etc.) could be put together with standard template-like functions to create a product (most likely in less time/with less effort than a designer could give.) Of course, that wouldn’t work for extremely specific things.

Art, on the other hand… maybe not. I appreciate art because it took skill to make it, not necessarily because it “looks cool” or follows artistic algorithms. That video about generative art didn’t do much for me. I’d appreciate it more if a person actually made it. But maybe that’s just me, and I’m not really one to think that mass-produced objects are as “artistic” as individual pieces, so that probably distances me from some people.

I found this example on vimeo of a chair whose seat angle and height can be changed. Although I’m not necessarily of fan of the resulting aesthetic, I think it makes for a good case study.

I remember a thread on this forum a few years ago when a Pro - E module (Gene-ometry? (now I am reading your blog and realize it was your company Genometri)) was making the rounds. Unfortunately most of the examples were pretty bad, so it was a hard sell, but at that time I remember making the case (on a core77 discussion) that in the hands of a good designer, it would be a very powerful tool for exploring variations on a theme (at the very least) once a design direction was underway. I don’t see it as a threat to designers at all, just another tool. This is essentially what Josh Davis and others have been doing with graphics for awhile now.

here’s an interesting link for a spirograph type thing that I think is pertinent
http://www.subblue.com/projects/guilloche

Can you provide some case studies where this was used and led to a positive outcome?
How are architects using it?

These guys are using it for jewelry. Really beautiful.

xylem_front.jpg
I’ve also seen a few chairs and tables recently, but don’t have them bookmarked. I’m not as convinced of its usefulness in architecture. I don’t see blobby buildings ever being anything more than a curiosity.

According to the book Space Craft: Developments in Architectural Computing (RIBA Publishing, 2008), > "Irish physicists Denis Weare and Robert Phelan were able to calculate that the most efficient way to divide a space into cells of equal volume while minimizing the surface area between them > was to use a stacked arrangement composed of 75% 14-sided shapes and 25% 12-sided shapes." So far so good, but since the resulting structure would have 22,000 steel members connected at 12,000 nodes, generating an actual model based on the idea exceeded the reach of conventional design.

Instead, according to the book, to manipulate this complex geometric system dynamically, Arup wrote parametric software that automated the drawing and analysis process. Based on specified design constraints and less than 190 loading scenarios, the algorithm iteratively checked the distribution of forces through the entire structure based on specific member sizes, allowing the team to test different design configurations and receive feedback within 25 minutes. > The result was a spectacular building with a sophisticated structure that is optimized in terms of material weight-to-strength ratio, > and it was achieved with relative ease. In addition to the structural advantages, Arup estimated that > it saved $10 million on design costs alone > compared with traditional design methods.

I certainly seem more interest 5 years latter. I guess, man have heard about it now. But the ship has left the port for IDers. Generative Design is now fully embraced by the cutting edge architectural practices to make both totally meaningless blobby shapes but also to optimize performances to a very hi degree. In Architecture, generative design is going main stream

Need to accept that doing blobby stuff ain’t easy if you have to crank out working drawings and resolve connection details. There is certainly some clever thinking and new capabilities required. Such work is not out sourced as architects unlike IDers have created new skills and new value - through the adaptation of advanced design technologies which exploit the creative capacity of computers. In contrast, we are witnessing the demise of of shape and color play and with it the demise of sectors of profession focused purely on it. I have blogged about it.

In some ways the idea of using the computer to determine a best-fit solution given constraints is hardly anything new - it’s the main strength of using the computer for any particular process!

exactly

exactly.

The understanding of design as a management of constrains is not new. The understanding that nature’s designs are based on generative designs are not new. The understanding that computers can devise strategies that beat the best human players is not new. The use of computers as intelligent design tools is not new.

But for most IDe’rs the concept that the computer is more than a 3d drawing tool seems to be very new and hard to digest.

So to answer the question, what is generative design ? its about using the computers for design exploration.

I know I’ve had happy accidents in modeling programs before. Does that count? =)

Just imagine if you can choose from 1000’s of accidents. That is how nature designs.

That’s not really a fair comparison, because nature designs by incorporating subtle mutations and relying on natural-selection over many generations to pass on the strongest attributes.

I don’t understand why you’d want a computer to generate meaningless design for a designer to choose from. As a designer, I want more meaning to go into form-giving, not less.

Again, can you provide a case study?

SK, this is brilliant! Thanks for sharing. I have seen loads of these architecture concepts around and I use it for recent project inspiration.

I’ m jumpin’ on.

Grasshopper, right?

Why not Chris, this is easily done using computers. They have been doing this over many decades now, using what is called Evolutionary Algorithms. There is an excellent book by Peter J Bently “Evolutionary Design by Computers” it has many examples of applications in engineering. Most engineering applications use what is called fitness functions to select the strongest attribute. This is problematic when the numerical evaluation of the attribute becomes uncomputable - as in design. So this process has to be modified for seeking the designers judgment and it has been.

…for simple reasons. It can do a far better job, faster and cheaper. It can extend human imagination. This is what architects are using generative design for.

Wonder if you will find a cheap turbocharded CAD monkey, that can work by itself - of some assistance to you ?
You don’t have to pay this monkey, but need to guide it (but too much).

As I have said - the ship has left the harbor quite some time ago. You will find about 300,000 references in google now, and about 100,000 images if you search for “generative design”

Yes, almost all architecture students all over the world are using generative design . You don’t need to be a programmer to be able to do generative design. Rhino has introduced grasshopper that makes it possible for people with no programming experience to rig out generative schemes and it has been wildly successful.

Brook, there are many views and many approaches for generative design and many definitions too. For a long time engineers have been using what is called parametric design - to be able to model “properly” in a way that variations can be created using design tables. They use this approach a lot at the optimization stage of the design. You can invert this same method and use if for generating designs. But most people who currently do generative design have not got to this stage yet. Architects have embraced parametric design with “proper” modeling techniques. This allows them to explore 100’s of design variations. They often call this generative design. I don’t. But from here true generative design is only a simple step a way. Once you have a properly structured CAD model, you can use random numbers within limits to create design variations. That’s all there is to generative design.

But then, this changes the game - entirely, opens up so many other issues, that you will discover for yourself.

You may look up http://www.opengenerativedesign.com an open design initiative I have launched with some Rhino and SolidWorks examples, which may be of interest to you. Hope you will enjoy playing with those models and hope it will open up a new world of possibilities.

Maybe the problem with ID users not picking up on this en masse is that as-is, it doesn’t contribute much to the design process. A “Randomize Parameters” tool is a rather different beast than “Fill in this space with stuff, make it strong, and use as little material as possible.”

Architects seem to be using the latter approach in your given Beijing Olympics examples. While ID and Architecture share many aspects (Both are designing things ultimately for the benefit, comfort, and safety of people, taking materials and space/shape in consideration) there’s a big difference in scale. Simply making random designs is nice, but what if you could teach the computer styling and form? The architects weren’t just getting random width and height walls and windows on a concrete rectangle, and until the tools for ID can move past that stage I see this being rather hit or miss for benefits.

Perhaps if you could teach the computer about form and style, give it a parts library it can work from or build off of.

In generative design, the designer is very much is control of style. Generated designs in most cases apear to belong to a design family, very often with strong stylistic commonalty.

The big difference between those who use generative design and those who remain skeptical is that the first sector focuses on what it can do and the second in what it cannot - for now.

I am disappointed by the responses you are getting from this topic SK. If so called designers cannot think of a use for this tool, then what is the state of our design thinking? Man ppl, you need to up the ante big time in your creative levels and progressive thinking.

Now the cat is out of the bag might as well throw out a few uses for this tool: footwear design, speakers and grills, micro textures, textures in general, button layouts, grills in general, wheels, tread patterns, accessories, bags & cases, clothing and the list is infinite to the power of dope. Only limit is that which is lacking in imagination.

Not to mention furniture and the most obvious, graphics. Oh I forgot, thats not in the id realm.