The documentation shows different problems implemented in OpenMDAO but they aren't clear what MDO architecture they are using. I would like to essentially recreate the 2013 paper comparing architectures to learn more about MDO in general. That mentions an architecture class that only appears to be in older versions.
Is there a way to easily implement different architectures or easily available examples for doing so?
OpenMDAO doesn't directly support distributed MDO architectures any more. You can easily implement any of the monolithic architectures (IDF, MDF, and SAND) in OpenMDAO without any special effort.
For the distributed architectures, you can still be implement them using sub-problems. You can check out the nested optimization examples from the 2020 reverse hackathon for some guidance.
Note: For a clear description of distributed vs monolithic architectures see the 2013 Martins and Lamb paper
