How to set a weighted multi-response objective in topology optimization (Compliance & Frequency)

Naoki

Hi everyone,

I understand that in OptiStruct you can only define one objective per analysis. However, I’ve heard that by combining multiple responses with appropriate weighting, it’s possible to simulate a multi-objective optimization.

For example, suppose I want to minimize compliance and maximize frequency in the same run. I was thinking of defining an equation like:

f = 0.5 * (Compliance / Comp0) + 0.5 * (1 / Frequency) / (1 / Freq0)

(where Comp0 and Freq are the initial compliance and frequency values used for the sake of normalization)

My questions are:

1. Is there a a way in the GUI to set an objective like this, where "f" is a weighted combination of multiple responses? (I am using the 2025 version hypermesh)
2. How can I obtain the initial compliance and frequency values( Comp0, Freq0) for use in this equation?

Any guidance on how to implement this in Hypermesh/OptiStruct would be greatly appreciated.

Thanks in advance!

loistf

Hi! All this that you mention should be possible. Bqt have you checked out "compliance index" in the responses section in the doc? That responses is, if i understand correctly, rather similar to what you are trying to implement in a custom way.

Having said that, for custom responses, you need dresp2 (response of type function in HM) referring to other responses you have defined previously.

and to get the values, i recommend you define a dummy optimization problem (doptprm, desmax, 0)

i hope it helps