Generative design is an intriguing new creative paradigm emerging in the field of design which mirrors similar developments in art and music. The modernists of the early 20th century emphasized the importance of the processes used to create a painting, a poem, or a composition. Generative design pushes this procedural method to the conclusion that could only have been imagined by many of the modernists. In generative design, designers set up a process — they write the rules for a system — but the end result is produced by the process itself. Like a chain of
Like a chain of dominoes where each domino determines the position of the next, the designer has only indirect control over the end result of the creative process. One might even go as far as saying that, in generative design, the creative work is split between the designer and tools they are working with. In most cases, those tools are a combination of hardware and software.
What Is Generative Design?
Let us start with the basics, something most of us already have a bit of understanding: a computer. Computers are, among other things, machines with the ability to precisely calculate complex mathematical problems within split seconds. Computers are actually a lot better at performing these kinds of equations that people are. Which is why, even with the technology available in 1997, IBM could design a computer that could defeat chess grandmaster Garry Kasparov.
For the computer to do that, a programmer (or a large team of programmers) write the right program to tell the computer exactly what it needs to do in order to accomplish the end goal. Let us put it this way: in chess, the end goal to checkmate the opponent, but there are millions or more possibilities with each consecutive move. The program tells the computer the most effective way to accomplish the goal relative to every move that the opponent makes during a game; this is where computer algorithm comes in. In other words, the computer can only accomplish these incredible feats when it’s been given a set of rules to follow by a designer.
An algorithm allows a programmer to tell a computer not only what to do, but also how to do it. When this is applied to the world of engineering and design, what you get is generative design. When algorithms can be applied in manufacturing industries, infrastructure, and basically everything that requires impeccable combination between form and function; we call it generative design. The following diagram helps to explain the principle:
Based on the diagram, we can easily tell that idea is the only part of the process. From this idea, a set of rules are created, and it is these rules that the computer uses to generate a solution to the design problem being posed. The important and pivotal aspect of this process is that, once let loose, the computer can and probably will come up with solutions that are entirely unanticipated by the designers. And this is what makes it a generative process.
When it comes to design, an idea is always a solution to some kind of problem. When someone has an idea, it does not necessarily mean that the person has found a way to solve the problem at hand. Sometimes, an idea can simply be an insight into the nature of the problem itself — the solution often remains opaque or prohibitively complex. But some things which are incredibly complex to us are simply cheesecake to a computer.
The designer starts with an objective (or a set of objectives) and runs them through an algorithm. While they are the architects of the system, it is the system which produces the final result.
For example, a shoe designer wants to create a running shoe capable of providing comfortable cushioning and maximum weight support of 300 pounds; or an architect who wants to build a structure with massive solar energy conversion without sacrificing aesthetics. In conventional design, the designer or architect would sketch their ideas, create blueprints, make a 3D model or a prototype, and work out how the final product should look and be made. Changes would be made by iterating these steps, until a design is arrived at which the designer believes satisfies all of their project goals.
All these processes are called “explicit design”. If they put their ideas or problems into computer algorithms and let the computer create design iterations for them, what they do is called generative design. A computer can perform infinitely many iterations much, much faster than any human being could.
A good example of how the technology works was set by the works of Lightning Motorcycle below.
|
1. Swing Arm Prototype – Source |
2. Developed Prototype – Source |
Autodesk, with its Dreamcatcher software, made the prototype of a swing arm for Lightning, a motorcycle company based in San Carlos, California. The objective was to create the lightest swing-arm possible for a motorcycle without sacrificing structural integrity. The swing-arm must support the weight of the engine and rider. Of course, with any vehicle design, reducing weight leads to higher efficiency.
Thanks to generative design, engineers were able to set starting parameters such as strength and weight, and allowed algorithms to potential solutions through mathematical analysis. Dreamcatcher worked on the given parameters and came up with numerous potential design solutions, which were then evaluated by the design team so that the optimal solution could be identified.
Notice how the swing arm in that first design is… not what you’d expect? That’s because the algorithms were provided with weight and strength requirements as their primary constraints — the generative system was not concerned with how the thing should look, and so it produced an unexpected result. In fact, these kinds of generative algorithms often produce these kinds of organic-looking results.
Image number 2 shows a refined version of the prototype. Engineers probably tinkered with the algorithm during the process to make the end product more realistic for the envisioned application. While the end result is a lot more conventional than the algorithms earlier iteration, it still demonstrates a kind of mathematical structure that is a hallmark of generative design processes.
Generative Design does not refer to one designing platform only, such as Dreamcatcher. According to Autodesk, the four most common methods are:
- Form Synthesis: the method employed by Dreamcatcher. It allows designers to input certain parameters of an object such as expected goals and limitations. Leveraging recent developments in AI technology, complex algorithms generate countless design solutions, select those which seem optimal, and present those results to the designer.
- Lattice and Surface Optimization: this method uses an already existing structure. Instead of making new objects, generative design software is run to optimize the surface structure of the object so it becomes stronger yet lighter. Generative design algorithms are great at coming up with complex asymmetrical grids and meshes for optimum strength with minum material use, for example.
- Topology Optimization: similar to the previous method, topology optimization aims to create the most material-efficient design by removing unnecessary parts without degrading performance.
- Trabecular Structures: used mainly in medical engineering, the method scales and distributes microscopic pores through solid materials to mimic bones. The design helps patient heal faster after an implant procedure.
Regardless of the methods, the procedures are more or less similar to each other. In basic terms, it goes as follows:
- It all starts with human input: the first part of the method requires a designer to set the raw ideas or problems to solve. In manufacturing industries, the parameters revolve around typical factors such as weight, strength, type, and cost.
- Artificial Intelligence takes over: the algorithm runs its own ‘reasoning’ and collects data from multiple sources to generate hundreds (if not thousands) of potential design solutions. Thanks to the Internet, particularly the rapid development of cloud-computing, a connected computer can now collect and process an incredible amount of data, a key aspect of the emergence of generative design technology.
- Humans play a role again: a powerful computer with the right software can come up with relevant solutions, but it still requires human intervention for the design to be financially and technically possible with the existing manufacturing facility. Based on the multiple choices of designs, human modifies the parameters to narrow down the options and then refine the available ones.
- While this step isn’t necessary, strictly speaking, at this point the designers will usually produce a 3D printed prototype for real-world evaluation. The combination of these two futuristic technologies — generative design and 3D printing — is a reminder of the truly remarkable technological era that we live in.
How Will Generative Design Change the Industry?
The idea of having computers become partners in the design process might strike some as being a bit frightening. Maybe a little too Terminator-y. But people felt the same way about factory automation and trains. The first thing to remember is that humans are still involved in this process, and it is the collaboration between human intellect and computational power that gives generative design its edge. We’re a long way off from developing computers with the intuitive power of your average person.
It’s also important to remember the ways in which generative design, while producing better and more creative solutions, can also free up the human designer for less mundane tasks. A computer can go through hundreds of iterations of a design in the time that it would take a designer to do maybe one. But there’s a lot more to designing than iterating concepts to find the best mechanical configuration. And generative design frees up innovators to pursue those other, less banal activities. This includes determining what the design objectives should be – something which an algorithm simply is not capable of doing. Yet.
It also means that many things will become more affordable due to the reduced cost of research and development. Creative design ideas will thrive and inventions should appear to be easy things any designer can make. All in all, the technology offers a wide range of benefits including but not limited to:
- Time-efficient designing process: a collaboration of creative designers and cloud-based computing greatly reduces the time needed to create multiple designs.
- A boost in creative design ideas: engineers and designers can really think outside of the box and see what kinds of solutions are generated by the AI.
- Cost-saving manufacturing process: because almost all test procedures are all covered in the initial designing process, there is no need for major overhauls in later stages of productions. With all the right parameters implemented, the only thing left for testing is the real-life performance of the design/object.
As with all new pieces of technology, generative design will become more accessible over time. When even small-scale businesses can make use of the technology, there may be an explosion of creative, unforeseen innovations in product design and development.
Are you looking for a great designer to help you design your product? Or maybe you need to find someone who can introduce you to the right algorithm. Well, Cad Crowd has got you covered! We have thousands of expert, cutting-edge design professionals ready to help you with your project. Looking for a human-algorithm partnership? Well, get in touch for a free quote. We’ll connect you with a leading CAD design professional at the competitive prices you need to realize your goals.
