which optimization method should I use?

I tried to look around the help sections but got a little lost

I have a multi variable (~8-10 variables) function that i want to minimise.All the variables are bound between approximately 1-2 orders of magnitudes. (for example, one variable is bound between 1-10, another between 1e-4 and 1e-2, etc). I need to find the minimum of thee function within these bounds.

function evaluation takes a long time, about 1 minute on average. I wouldnt mind having matlab run for a few days, but not more than that.

I will be happy to share any additional details on the function. the function itself i cannot share as it is part of a huge code.

should I use the global optimization toolbox or the optimization toolbox? I saw that surrogateopt is good for time consuming functions. should I use this?

Nathan

Answer 1

[Stephan]

Following your statements surrogateopt appears to be the correct choice.

[Walter Roberson]

Especially as surrogateopt specifically talks about how it creates checkpoint files so that it can resume in case of system problem.

[Nathan Bblanc]

my thanks to both of you