More objective functions in topology optimization module

[ysyysy115]

Hi,

Is it possible to have an example of modifying the objective function in the top_opt module, i.e. eigenfrquency, p-norm von-Mises stress and etc..
Thank you!

Best,
Jason

[tianjuxue]

We have the two (outdated) examples from our old repo for eigenfrquency and p-norm von-Mises stress.

[ysyysy115]

Hi Xue, Thank you for the information. I am also curious if any alternative optimizers could be used, like Adam in Optax. Since JAX has a huge advantage in AD. Best, Jason

…

On Sun, Nov 3, 2024 at 4:54 AM Tianju Xue ***@***.***> wrote: We have the two (outdated) examples from our old repo for eigenfrquency <https://github.com/tianjuxue/jax-am/blob/main/applications/fem/top_opt/eigen.py> and p-norm von-Mises stress <https://github.com/tianjuxue/jax-am/blob/main/applications/fem/top_opt/L_shape.py> . — Reply to this email directly, view it on GitHub <#40 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ASYH2OQKPPGUUM3VHM6PZ5DZ6YMIJAVCNFSM6AAAAABQ5RMUWSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINJTGQYTQMBTHA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[tianjuxue]

It's fine to use Adam, but that does not work as good as MMA (usually) for TO problems.