Re: Question about fmins!

[Michael Creel]

I ran across some Python bindings to the Fortran lbfgsb code by Nocedal and 
co-authors (http://www.ece.northwestern.edu/~nocedal/lbfgsb.html), at 
http://setiathome.ssl.berkeley.edu/~aparsons/python/aipy.lbfgsb.html

As far as I know, lbfgsb is about the state-of-the-art for minimization of 
differentiable functions of high-dimensional parameters, subject to bounds. 
Maybe someone who has experience with Octave/Fortran would like to try to make 
some bindings?

Otherwise, samin in octave-forge does bounds, without using derivatives. It's a 
stochastic method. In spite of that, it is pretty quick for many problems. 
Etienne's suggestion is probably the first thing to try, though.
M.

address@hidden wrote:
>   Hello,
>
> On Mon, June 11, 2007 01:33, rambovn wrote:
> #
> # Hi,
> # I am newbie in octave so I have many questions :)
> # Now I am want to find the minimum of a function of several variables. # I am 
> trying with fmins but the results are not like I expected :,( # I
> want to limit the result for example: My function have x , y as variables, # 
> and I want that x value is in the range from 0 to 1, and y is positive.
> How # can I do that with fmins? Or is there any other function in octave that 
> can # do it for me?
>
> I don't know that Octave has a constrained multi-variable function (does anyone 
> know otherwise?), but you can transform your function, say 'f',
> defined on [0,1] x [0,+infty] so it's domain is [-infty,+infty] x 
> [-infty,+infty].
>
> function z = g (xy)
>   z = f (1/2 + tanh (x(1))/2, exp (x(2));
> end
>
> That's a well-known trick, see an optimization textbook. Then, with a little 
> luck (your mileage may vary, as with all local minimization methods)
>
>   xy_0 = fmins (g, xy_init)
>
> should give you x_min = 1/2 + tanh (xy_0(1))/2 and y_min = exp (xy_0(2)).
>
>   Hth,
>
>   Etienne
>
> # Thank you very much for any help.
> # Regards
> # Anh
> # --
> # View this message in context: 
> http://www.nabble.com/Question-about-fmins%21-tf3900212.html#a11056777 # Sent 
> from the Octave - General mailing list
> archive at Nabble.com. #
> # _______________________________________________
> # Help-octave mailing list
> # address@hidden
> # https://www.cae.wisc.edu/mailman/listinfo/help-octave
> #