Mercurial > octave-nkf

comparison scripts/optimization/fminunc.m @ 9627:5bcfa0b346e8

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
fix extra outputs in fminunc
author Jaroslav Hajek <highegg@gmail.com>
date Tue, 08 Sep 2009 13:39:39 +0200
parents bc0739d02724
children 73e6ad869f08
comparison
equal deleted inserted replaced
9626:bccba774af8b 9627:5bcfa0b346e8
18 ## 18 ##
19 ## Author: Jaroslav Hajek <highegg@gmail.com> 19 ## Author: Jaroslav Hajek <highegg@gmail.com>
20 20
21 ## -*- texinfo -*- 21 ## -*- texinfo -*-
22 ## @deftypefn{Function File} {} fminunc (@var{fcn}, @var{x0}, @var{options}) 22 ## @deftypefn{Function File} {} fminunc (@var{fcn}, @var{x0}, @var{options})
23 ## @deftypefnx{Function File} {[@var{x}, @var{fvec}, @var{info}, @var{output}, @var{fjac}]} = fminunc (@var{fcn}, @dots{}) 23 ## @deftypefnx{Function File} {[@var{x}, @var{fvec}, @var{info}, @var{output}, @var{grad}, @var{hess}]} = fminunc (@var{fcn}, @dots{})
24 ## Solve a unconstrained optimization problem defined by the function @var{fcn}. 24 ## Solve a unconstrained optimization problem defined by the function @var{fcn}.
25 ## @var{fcn} should accepts a vector (array) defining the unknown variables, 25 ## @var{fcn} should accepts a vector (array) defining the unknown variables,
26 ## and return the objective function value, optionally with gradient. 26 ## and return the objective function value, optionally with gradient.
27 ## In other words, this function attempts to determine a vector @var{x} such 27 ## In other words, this function attempts to determine a vector @var{x} such
28 ## that @code{@var{fcn} (@var{x})} is a local minimum. 28 ## that @code{@var{fcn} (@var{x})} is a local minimum.
57 ## @item 0 57 ## @item 0
58 ## Iteration limit exceeded. 58 ## Iteration limit exceeded.
59 ## @item -3 59 ## @item -3
60 ## The trust region radius became excessively small. 60 ## The trust region radius became excessively small.
61 ## @end table 61 ## @end table
62 ##
63 ## Optionally, fminunc can also yield a structure with convergence statistics
64 ## (@var{output}), the output gradient (@var{grad}) and approximate hessian
65 ## (@var{hess}).
62 ## 66 ##
63 ## Note: If you only have a single nonlinear equation of one variable, using 67 ## Note: If you only have a single nonlinear equation of one variable, using
64 ## @code{fminbnd} is usually a much better idea. 68 ## @code{fminbnd} is usually a much better idea.
65 ## @seealso{fminbnd, optimset} 69 ## @seealso{fminbnd, optimset}
66 ## @end deftypefn 70 ## @end deftypefn
306 310
307 output.iterations = niter; 311 output.iterations = niter;
308 output.successful = nsuciter; 312 output.successful = nsuciter;
309 output.funcCount = nfev; 313 output.funcCount = nfev;
310 314
315 if (nargout > 5)
316 hess = hesr'*hesr;
317 endif
318
311 endfunction 319 endfunction
312 320
313 ## An assistant function that evaluates a function handle and checks for 321 ## An assistant function that evaluates a function handle and checks for
314 ## bad results. 322 ## bad results.
315 function [fx, gx] = guarded_eval (fun, x) 323 function [fx, gx] = guarded_eval (fun, x)