optimization contain misalign error

[Hwang, Ji-Gwang]

Hello. Dear All.
I have some questions. So, I leave this message.
If someone know the this, Please give me some comment.

Now, I want to make .ele file for optimization the lattice of contain the misalign error
I first add the error of lattice and I using "&save_lattice = %.new" commend at end of error.
But In this case, Not saved the error effect in new lattice file.
So, I just put the error commend pervious to optimization command.
But I this case the result is not so good.
Which method is good for save the error effect?

I have one more question. Runingtime of attached file is very long. How to decrease the runing time?

Sincerely yours. Have a nice day~!

[michael_borland]

Ji-Gwang,

You can save error values using the error_log parameter of error_control. You can later reload them using load_parameters. Errors are not saved to the lattice file because they are considered perturbations, not part of the lattice definition.

What you are attempting to do is basically lattice correction using the optimizer. I'm not surprised that this takes a very long time to run. One issue is that the tolerances you are trying to achieve are very small (0.01%). Since you have more sources of error than quadrupoles to perform correction, it may simply not be possible to achieve this level of correction. I would increase this to 1% and see if that helps.

--Michael

[Hwang, Ji-Gwang]

Thank you for your reply.

error log and load_parameter command is good for me.

But, When I contain the error, optimization result is don't good.
I've achieve to 1% result compare with bare lattice in case of not contain error effect.
I have change the target to near 1%. But Still, It is don't get to good result. Result of this file has error near 10%.
Is this need more fitpoint for optimization? or change the some command in optimization?

From Hwang, Ji-Gwang

[michael_borland]

Ji-Gwang,

Can you post (or email me) the lattice file?

--Michael

[Hwang, Ji-Gwang]

Dear Michael

Thank you for your reply. I attached the lattice file.

From Hwang, Ji-Gwang

[michael_borland]

Ji-Gwang,

Attached you'll find my revision of your files. The convergence is much faster and I get RMS beats below 1%.

There's a Notebook file that explains things.

--Michael

[Hwang, Ji-Gwang]

Dear Michael

Thank you for your concerned. This method is very helpful for me.

From Hwang, Ji-Gwang