Recently, I've been looking recently into what uses AI - specifically evolutionary algorithms - may have in automating engineering design. For a long time, there have been algorithms that solve constraint satisfaction problems, and, to me, it makes sense to consider engineering problems as a superset of constraint satisfaction problems. In spite of this, I haven't been able to find any cases of engineering design being automated other than a couple of cases of genetic algorithms being used to optimise structural members.

So my question is, why can't I find any examples?

The first thing that springs to mind is that I just haven't been looking hard enough - if this is the case, could anyone point me in the right direction?

The other obvious answer is that it isn't a widely researched area - if so, why not? Is it just due to lack of interest or are there technical hurdles (abstraction, complex logic & reasoning, etc.) that make this a much more difficult problem than computer vision, games, and so on?


Examples of GA implementations:

  1. NASA’s spacecraft evolved antenna
  2. Solution for TSP with a big number of connections. Here is an example with Python and here with C#.
  3. From my experience, GA can be even used for a training of small artificial neural networks, but with less efficiency than with traditional deep learning approaches

Also, I heard about successful projects with using GA for scheduling, packing and logistics

Therefore, everything, that can be represented in a form of digital chromosomes (can be serialized to and deserialized back from a sequence of bits) and we can define a function to evaluate its efficiency – fitness or reward function can be optimized by using of GA


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.