Elegant users forum

Hi,

I am trying to optimize a 3 chicane linac system, but I am getting unexpected results from ELEGANT.

While keeping the bunch length and beam energy constant at the exit of the 3rd chicane, I would like ELEGANT to optimize the emittance (enx). I am wondering if I have the correct format in my optimization_setup script which I have attached.

Also, I will be running this in parallel using Pelegant, does that change any of the optimization_setup format?

Thank you in advance!

One problem is that the optimization term for the emittance will return very small values, which will be dwarfed by the values from the other terms. Since the terms are all summed to create a single penalty function, the emittance is not likely to be improved. Try replacing the term "enx" with "enx 1e8 *".

--Michael

In answer to your second question, you shouldn't need any changes when using Pelegant. It will track in parallel.

--Michael

Thanks for the speedy help Michael (as always!). Quick question:

I am trying to learn a bit more about this optimization routine (I am using simplex) that ELEGANT is implementing in order to get a better understanding (and confidence) of the solutions it spits out. The manual is a bit sparse on details, do you recommend any sources for learning more, or know where the routine is located in the code that I can dig through?

Thanks again

The simplex optimization routine is in one of the libraries. You can find the source code in epics/extensions/src/SDDS/mdbmth/simplex.c

It's a pretty standard implementation of the simplex algorithm. Like any simplex algorithm, it is not guaranteed to find the optimum solution, but only a local minimum.

--Michael