Multi-objective direct optimization of DA and MA

[ranamonikaa]

Hi,
Hi, I am running the MOGA example for optimizing DA and MA. The program generates "optim1" files including optim1-000000, optim1-000011, optim1-000012, optim1-000015, optim1-000024, optim1-000025, optim1-000029, optim1-000033, and optim1-000035, however, it stops after completing 9 jobs, even though it continues to occupy nodes. What could be causing this issue, and what kind of results should I expect?

Thanks.

[michael_borland]

It is hard to say without more details, but a possible cause is that the required files are not created by the jobs. Do you see the final processed output files, which should have extension .proc? Any errors in the individual log files?

--Michael

[ranamonikaa]

yes there are .proc files and I see no errors in the individual log files. Can you please tell me how many jobs should the program run or what output should I expect for this particular example.
Thanks.

[michael_borland]

You should see several additional files created by the geneticOptimizer script with names matching optim1.*. In particular, look for optim1.sort, which will have summary data for all runs that has been sorted. There will be a column Rank that gives the ranks from the non-dominated sort. The geneticOptimizer script will use the solutions with Rank=1 to generate new candidates, which it should submit for evaluation.

--Michael

[ranamonikaa]

Actually at the starting point in the output file, it is
Fri Apr 25 11:32:38 IST 2025: Jobs: current=40, done=0, proc'd=0, toProc=0, running=40.
what I understand is, 40 jobs were lunched. where does this number come from?
and after these 40 jobs are completed, it is
Fri Apr 25 21:38:16 IST 2025: Jobs: current=40, done=40, proc'd=40, toProc=0, running=0.
then the program hangs for forever and shows no progress. probably the program is waiting for another step but some file that MOGA expects is missing. where can I download all the moga files from?

Thanks.

-Monika

[michael_borland]

Monika,

I don't understand why it isn't submitting more jobs. Can you verify that you have a set of files with extension .proc? Also, can you upload the input file you use with geneticOptimizer?

I'm looking for the MOGA example files. They used to be on our web page but I don't see them now.

--Michael

[ranamonikaa]

Yes, I do have a set of .proc files in place. I am uploading one of the .proc files with the input file I'm using with geneticOptimizer so you can take a look.
Thanks.

-Monika

[michael_borland]

Monika,

I don't see anything obviously wrong with the processed file. Can you upload the input file you are using with geneticOptimizer?

--Michael

[ranamonikaa]

Michael,

I have sent all the files I am using. I am only running the aps files for now that I have attached in the previous reply.
Thanks.

-Monika

[ranamonikaa]

I have attached the optim1.sdds file. Sorry it was not in the previous attachment. I think this sets the total jobs to 40. while running this optimization for a different lattice I am getting this error in the main.log file:
** Starting 30-line aperture search
Error:
Unable to update page--new number of rows less than previous number (SDDS_UpdateBinaryPage)
Problem setting SDDS parameter values (do_aperture_search)

It doesn't do dynamic aperture calculations. what could be the issue?

Thanks.

-Monika