optimization method

[weix...@umn.edu]

Hi all

For the kind of equation in the attached fie, which is the optimization method in the parameter estimation is the best one to use in Copasi? Thanks! 

Best

Ning 

[Pedro Mendes]

What are you trying to maximize or minimize in that function?

In general there is no "best" algorithm, though for some problems there
may be (however it has never been studied systematically).

In your case I think you could use Levenberg-Marquardt, Hooke-Jeeves or
Particle Swarm.

If these algorithms don't provide a satisfactory result, try changing
their tuning parameters. For example I would set many more iterations
for Levenberg-Marquardt than the default (say 20000) and fix the
tolerance to 1e-12.

Other algorithms could also work, so it is a matter of trying them out
(which is why we provide an easy way to switch them)

Pedro

> --
> You received this message because you are subscribed to the Google
> Groups "COPASI User Forum" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to copasi-user-fo...@googlegroups.com
> <mailto:copasi-user-fo...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/copasi-user-forum/f9fc8622-9302-4c6e-9ed1-c7cd232027e5n%40googlegroups.com
> <https://groups.google.com/d/msgid/copasi-user-forum/f9fc8622-9302-4c6e-9ed1-c7cd232027e5n%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Ning Wei]

Dear Pedro

Got it. Thanks! We want to minimize the cost function between experimental data and numerical prediction for A for example. The ODE for A is the attached file. 

Best

Ning 

To unsubscribe from this group and stop receiving emails from it, send an email to copasi-user-fo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/copasi-user-forum/01533336-72aa-06f5-39b8-8296398dfd4d%40copasi.org.