Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

restart multi obj #133

Open
wgyang opened this issue Feb 12, 2023 · 3 comments
Open

restart multi obj #133

wgyang opened this issue Feb 12, 2023 · 3 comments

Comments

@wgyang
Copy link

wgyang commented Feb 12, 2023

I tried to make a bi-obj optimization start from the previous state when the process was interrupted instead of starting from the beginning. Following the tutorial, I added the following to the parameter input file:

CACHE_FILE cache.txt
HOT_RESTART_ON_USER_INTERRUPT true
HOT_RESTART_READ_FILES true
HOT_RESTART_WRITE_FILES true
HOT_RESTART_FILE hotrestart.txt

But DMULTIMADS couldn't restart from the previous iteration when I killed the optimization. Similarly, when I ran bi-obj again without changing my param.txt, it made another MAX_BB_EVAL black box runs instead of being stopped at the beginning due to reaching max bb runs. For a single obj, the parameters above can make NOMAD read the hot restart file first. The example problems I tested were taken from nomad/examples/basic/batch/multi_obj/param.txt and /basic/batch/examples1/param.txt.

Thank you for your help!

@ctribes
Copy link
Contributor

ctribes commented Feb 13, 2023

Hot restart is not yet implemented for DMultiMads.
I have added a warning message (instead of an exception) whenever user wants to stop before restart.

Thanks for reporting this

@wgyang
Copy link
Author

wgyang commented Feb 16, 2023

Thanks for your response. I am wondering if there is any other bi-obj algorithm in NOMAD that allows me to restart an interrupted process because my function evaluations are expensive and my optimization job that couples NOMAD to a CFD solver is run on a cluster with a wall time limit. Thanks!

@ctribes
Copy link
Contributor

ctribes commented Feb 16, 2023

There is no other bi-objective algorithm in Nomad 4.

I suggest to save all evaluations in a cache file and/or history file (see user guide for details).

When saved in a cache files, evaluation points will be considered when re-launching Nomad on the same problem.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants