Randomization for FractionalFactorials

[jduerholt]

This PR adds an option for randomizing the order of candidates proposed by the FractionaFactorialStrategy.

[jduerholt]

@dlinzner-bcs: also the https://github.com/experimental-design/bofire/blob/main/tutorials/doe/basic_examples.ipynb is failing. Any idea, why?

[Osburg]

Hey Johannes :)

To me it looks like it is running without an error, am I wrong?

Cheers,
Aaron

[jduerholt]

@Osburg: you are right. The error stems from another one, and I solved it in a seperate one. If you are already looking at this PR, can you also just review ;)

[jduerholt]

@Osburg: here it is failing again; https://github.com/experimental-design/bofire/actions/runs/13155774927/job/36712478341?pr=511

[Osburg]

think it looks good :)

[Osburg]

@Osburg: here it is failing again; https://github.com/experimental-design/bofire/actions/runs/13155774927/job/36712478341?pr=511

Will have a look at it

[Osburg]

The notebook fails since ipopt sometimes has problems to fulfill the more complicated nonlinear constraints (like the cones) to the (fixed) tolerance level in the validation of candidates. I would suggest to set "raise_validation_error=False" for the cone examples (since the constraints are only mildly violated, i.e. above the threshold but still on the cone surface when visually inspecting the design). An alternative would be to increase the number of max iterations, but then the notebook will run longer, which is not worth it i think, the jobs are running for a very long time now already... What do you think?