Re: Free Nonlinear Optimization Package?
- From: andy2O@xxxxxxxxxxx
- Date: 3 May 2005 05:45:06 -0700
shahehe@xxxxxxxxx wrote:
> I have an optimization problem where the objective function is
> nonlinear and there is no functional constraint. The feasible region
is
> a box, i.e., the optimization variables satisfy the constraints
>
> xi_min <= x_i <= xi_max
>
> where xi_min and xi_max are the bounds for variable xi.
>
> Currently, I am using a general optimization package CFSQP to solve
the
> problem. I would like to know if there is a more robust and more
> efficient free optimization package implemented in C or C++ that can
be
> used to solve my problem, i.e., "nonlinear objective function, no
> functional constraint, and box shaped feasible region".
>
> Thanks.
Have a look at the bound-constrained optimisation section of the NEOS
Optimisation software guide:
http://www-fp.mcs.anl.gov/otc/Guide/SoftwareGuide/index.html
If you can link your code to Fortran, or just use Fortran - it's simple
enough, then I highly recommend the code TNBC linked at the above URL,
or directly at
http://www-fp.mcs.anl.gov/otc/Guide/SoftwareGuide/Blurbs/tn_tnbc.html
It proved really very robust for my work - more so than many other
library codes available. (You'll probably get a much wider choice of
good codes if you are able to use/link to Fortran - there's a lot of
well tested robust old code out there in FORTRAN 77 which will do your
job. You could look at running it through the f2c program to turn it
into C perhaps - but look for advice from others on this first...)
andy
.
- References:
- Free Nonlinear Optimization Package?
- From: shahehe
- Free Nonlinear Optimization Package?
- Prev by Date: Solving a pde in spherical coordinates
- Next by Date: Re: nnz in convariance of a sparse matrix
- Previous by thread: Re: Free Nonlinear Optimization Package?
- Next by thread: New mathematics/physical sciences positions at http://jobs.phds.org, May 02, 2005
- Index(es):
Relevant Pages
|
