Limit number of optimization iterations for running GitHub Action tests

[markfuge]

Currently the tests can time out if we test all optimize functions for many steps. We decided that this is not necessary for the purposes of the github action tests. We can reduce the number of iterations used by each optimize step.

• Check that our problem optimize functions can each change the number of steps
• Change the github action to only run it with this non-default value of number of steps

[g-braeunlich]

Atm I am not sure how to set max iterations for all problems:

• airfoil: only see adjointMaxIter in templates/airfoil_opt.py is this the correct parameter @cashend ?
• beams2d: This should be Config.max_iter
• headconduction2d / 3d: ? @MiladHB
• photonics2d: This should be Conditions.num_optimization_steps
• power_electronics: No optimize
• thermoelastic2d / 3d: ? @gapaza

[gapaza]

For thermoelastic2d / 3d, the number of optimization iterations is determined by both the update threshold (how large the last gradient update was) and a hard upper bound (I think around 120 iterations for each). If the update threshold is not met before the hard upper bound, the optimization stops. I could modify the code such that, when optimize is called, an extra parameter can be defined in the config dictionary that sets the number of optimization iterations. Would this be an acceptable patch?

[g-braeunlich]

@gapaza Yes, that would be great

[g-braeunlich]

Or I can do that myself. Found it in model/fea_model.py

[g-braeunlich]

@MiladHB , @gapaza : I have implemented the changes in #219 -> you dont need to work on this anymore.