Report a bug: UQLink cannot support Soft Constraints

Dear UQLab,

Hope you find this question informative;
I want to link OpenSees with RBDO module using UQLink.
Without defining soft constraints, my RBDO problem works well. However, after defining a simple soft constraint using either string or mfile I face the follwong error:

Too many output arguments.
Error in uq_Kriging_initialize (line 108)

It seems uq_SoftConstraints cannot be utilized when we want to use third-party software via UQLink.
It’s worth mentioning that hard cosntraint works properly.
Last but not least, I am most grateful for your kind suggestions/recommendations. As always, I will be honored if I can benefit from Dr. @moustapha’s fruitful suggestions. I also can send the code.

Kind regards,

Hi @ali ,

Thanks for your message. Can you please provide us with more details, e.g., which RBDO method are you using? at which line of the main code does it fail? The error seems to occur at initialization, while creating the Kriging model. I don’t think it is directly linked to uq_evalSoftConstraints.m.

If you are using the OpenSees example from UQLab, it would be helpful if you could share the main file you are using to set up your analysis.


1 Like

Hi Dr. @moustapha

First and foremost, I am most grateful for your sustainable support and I am deeply sorry for the late response.
I have already sent you the code via email. In summary, from the RBDO module, I applied QMC with Kriging as the metamodel. Moreover, UQLink is utilized to link UQLab and OpenSees. As you mentioned, it seems the RBDO could not initiate with having “Soft Constraints”. As outlined in the previous post, without having soft constraints, the code can generate both initial (via MCS) and added (active learning phase) samples. I’d hazard a guess, maybe something is wrong with connecting soft constraints and OpenSees.
If I can be of assistance, please do not hesitate to contact me. I will be honored if I can benefit from your suggestions, as always. Last but not least, I really appreciate your priceless efforts and other RSUQ members in advance.

Kind regards,