<< sifipopt CUTEr sifqld >>

CUTEr >> CUTEr > sifoptim

sifoptim

apply optim function to a SIF problem

Calling Sequence

[fopt,xopt,gopt]=sifoptim(siffile)
[fopt,xopt,gopt]=sifoptim(siffile [,algo] [,Stop] [imp=iflag] )
[fopt,xopt,gopt]=sifoptim(siffile [,algo=value] [,Stop=value] [imp=iflag] )

Parameters

siffile

path of the SIF problem file

algo

string "qn" or "gc" optim method selection.

Stop

a vector with at most 4 elements: [nap iter epsg epsf]

or a list with at most 5 entries: list(nap,iter,epsg,epsf,epsx)

napgives the maximum number of function calls, niter the maximum number of iteration to do, epsg the threshold on gradient norm, epsf the threshold controlling decreasing of the criterion and epsx the threshold array controlling decreasing of the variables (see optim)

"imp=iflag"

named argument used to set the trace mode. The possible values for iflag are 0,1,2 and >2. Use this option with caution : most of these reports are written on the Scilab standard output.

  • iflag=0: nothing (except errors) is reported (this is the default),

  • iflag=1: initial and final reports,

  • iflag=2: adds a report per iteration,

  • iflag>2: add reports on linear search.

  • iflag<0: calls the cost function with ind=1 every -imp iterations.

fopt

scalar, criterion at the found optimum

gopt

vector, Gradient at the found optimum

xopt

vector, optimum location

Description

The sifoptim function decodes, compile, link the given problem and then call the Scilab function optim in order to solve the given problem.

Examples

p=get_sif_path()
[fopt,xopt,gopt]=sifoptim(p+'/sif/BROWNDEN.SIF')

See Also

Authors


Report an issue
<< sifipopt CUTEr sifqld >>