# Automatically generated file --- DO NOT EDIT
2007-05-24 23:38  goffioul

	* doc/Makefile: fix dvipdf calls

2007-04-19 10:57  mcreel

	* src/samin.cc: further fixes to help string and comments

2007-04-19 10:27  mcreel

	* inst/samin_example.m: very minor improvements

2007-04-19 10:25  mcreel

	* src/samin.cc: minor improvements to help string

2007-03-29 13:26  mcreel

	* inst/bs_gradient.m: this file has been just a backward
	  compatibility a placeholder for a long time

2007-03-23 17:14  adb014

	* inst/battery.m, inst/bfgsmin.m, inst/bfgsmin_example.m,
	  inst/dfdp.m, inst/fmin.m, inst/fmins.m, inst/fminsearch.m,
	  inst/fzero.m, inst/leasqr.m, inst/leasqrdemo.m,
	  inst/rosenbrock.m, inst/samin_example.m, src/__bfgsmin.cc,
	  src/numgradient.cc, src/numhessian.cc, src/samin.cc: Update the
	  FSF address

2007-03-04 21:46  adb014

	* Makefile: Add explicit taret for sub-directories of packages

2007-03-02 22:00  adb014

	* Makefile: double-colon in pattern match rule is a termination, so
	  remove pattern match from pre-pkg and post-pkg rules

2007-02-10 15:29  mcreel

	* Makefile: revert ver 1.18 to 1.17, as requested

2007-02-10 10:59  gnumuthu

	* Makefile: NIST based physical constants functions. The main file
	  not included, as I dont know terms of distribution from NIST. Python
	  generator is included, but not a serious candidate for real-time
	  generation

2007-01-30 23:23  adb014

	* DESCRIPTION: Add Autoload field to DESCRIPTION of main/ packages

2007-01-27 17:32  adb014

	* Makefile: Remove pkg.mk and incorporate in Makeconf.base. Update
	  Makefiles accordingly. Use cvs2cl to create ChangeLog files for all
	  of the packages if a ChangeLog file doesn't exist. Package build
	  machine must therefore have access to the CVS

2007-01-25 22:09  goffioul

	* src/Makefile: Add "clean" rules to makefiles and discard exit
	  status

2006-12-28 07:15  pkienzle

	* inst/wpolyfit.m: Corrections to comments regarding F-test for
	  polynomial degree

2006-12-03 22:54  adb014

	* INDEX: Update INDEX files for new functions

2006-11-05 11:31  sis-sou

	* inst/fminsearch.m: Initial commint into CVS.

2006-10-22 16:54  rlaboiss

	* inst/leasqr.m: Improved documentation.  Thanks to Francesco
	  Potorti for the patch.

2006-10-20 19:55  adb014

	* doc/Makefile: COpy latex2html nav buttons

2006-10-19 19:42  adb014

	* inst/fzero.m: function_handle and not function handle

2006-10-16 21:11  adb014

	* INDEX: Update category of INDEX for consistency

2006-10-14 19:39  hauberg

	* COPYING: Removed unvalid characters from all GPL licenses in the
	  packages

2006-10-10 09:56  mcreel

	* inst/bfgsmin.m: correct erroneous information in help string

2006-10-10 09:56  mcreel

	* inst/bfgsmin_example.m: nicer format for output

2006-10-09 16:26  mcreel

	* src/samin.cc: fix bug with scalar parameters

2006-10-07 21:02  adb014

	* doc/Makefile: Add -verbosity=0 to latex2html commands so that the
	  url link to a non existent file is not added

2006-10-07 00:57  adb014

	* INDEX: more indexing fixes

2006-10-06 22:49  adb014

	* INDEX: more INDEX fixes

2006-10-05 14:34  adb014

	* doc/.cvsignore: ignore certain tex temporary files

2006-10-05 14:32  adb014

	* doc/Makefile: Documentation seems to auto build correctly

2006-10-05 12:44  adb014

	* Makefile, doc/Makefile, doc/optim-mini-howto-2.lyx,
	  doc/optim-mini-howto-2.tex, doc/optim-mini-howto.2.html,
	  doc/optim-mini-howto.2.lyx, doc/optim-mini-howto.2.tex: First
	  attempt at automatic doc build

2006-10-05 12:24  adb014

	* doc/figures/: 2D_slice-3.eps, 2D_slice-3.eps2: Ok my fault.. fix
	  it

2006-10-05 12:23  adb014

	* doc/figures/: 2D_slice-3.eps, 2D_slice-3.eps2,
	  optim_tutorial_slice.eps, optim_tutorial_slice.eps2: Not all
	  latex versions look for *.eps2 files

2006-10-03 14:56  hauberg

	* INDEX: Fixed various INDEX files

2006-10-03 11:02  mcreel

	* src/: Makefile, __bfgsmin.cc, numgradient.cc, numhessian.cc,
	  samin.cc: remove use of celleval, use a more efficient direct
	  call to feval

2006-10-03 11:01  mcreel

	* src/celleval.cc: remove the celleval hack, unnecessary duplicate
	  call to feval

2006-10-02 21:47  adb014

	* DESCRIPTION: Latest mega package manager update

2006-09-14 10:08  mcreel

	* src/samin.cc: simplified temperature search, direct use of feval
	  rather than though celleval

2006-09-10 23:18  adb014

	* Makefile: Simplify package Makefiles, package all default package
	  files, and remove the redundant Makefiles

2006-09-07 15:37  mcreel

	* src/: __bfgsmin.cc, samin.cc: minor improvements

2006-08-28 22:06  adb014

	* .cvsignore: More package manager stuff

2006-08-25 08:52  adb014

	* Makefile: Further package manager updates. All but miscellaneous
	  and audio should build in main/. Some broken symlinks in image and
	  octcdf still

2006-08-21 20:39  hauberg

	* inst/.cvsignore, src/.cvsignore: Added a lot of .cvsignore's in
	  main

2006-08-20 15:37  hauberg

	* COPYING, DESCRIPTION, Makefile, __bfgsmin.cc, adsmax.m,
	  battery.m, bfgsmin.m, bfgsmin_example.m, bs_gradient.m, cdiff.m,
	  celleval.cc, d2_min.m, deriv.m, dfdp.m, expdemo.m, expfit.m,
	  fmin.m, fminbnd.m, fmins.m, fminunc.m, fzero.m, leasqr.m,
	  leasqrdemo.m, leval.cc, line_min.m, mdsmax.m, minimize.m,
	  nelder_mead_min.m, nmsmax.m, nrm.m, numgradient.cc,
	  numhessian.cc, optimset.m, poly_2_ex.m, polyconf.m, rosenbrock.m,
	  samin.cc, samin_example.m, test_d2_min_1.m, test_d2_min_2.m,
	  test_d2_min_3.m, test_fminunc_1.m, test_min_1.m, test_min_2.m,
	  test_min_3.m, test_min_4.m, test_minimize_1.m,
	  test_nelder_mead_min_1.m, test_nelder_mead_min_2.m,
	  test_wpolyfit.m, wpolyfit.m, wpolyfitdemo.m, wsolve.m,
	  inst/adsmax.m, inst/battery.m, inst/bfgsmin.m,
	  inst/bfgsmin_example.m, inst/bs_gradient.m, inst/cdiff.m,
	  inst/d2_min.m, inst/deriv.m, inst/dfdp.m, inst/expdemo.m,
	  inst/expfit.m, inst/fmin.m, inst/fminbnd.m, inst/fmins.m,
	  inst/fminunc.m, inst/fzero.m, inst/leasqr.m, inst/leasqrdemo.m,
	  inst/line_min.m, inst/mdsmax.m, inst/minimize.m,
	  inst/nelder_mead_min.m, inst/nmsmax.m, inst/nrm.m,
	  inst/optimset.m, inst/poly_2_ex.m, inst/polyconf.m,
	  inst/rosenbrock.m, inst/samin_example.m, inst/test_d2_min_1.m,
	  inst/test_d2_min_2.m, inst/test_d2_min_3.m,
	  inst/test_fminunc_1.m, inst/test_min_1.m, inst/test_min_2.m,
	  inst/test_min_3.m, inst/test_min_4.m, inst/test_minimize_1.m,
	  inst/test_nelder_mead_min_1.m, inst/test_nelder_mead_min_2.m,
	  inst/test_wpolyfit.m, inst/wpolyfit.m, inst/wpolyfitdemo.m,
	  inst/wsolve.m, src/Makefile, src/__bfgsmin.cc, src/celleval.cc,
	  src/leval.cc, src/numgradient.cc, src/numhessian.cc,
	  src/samin.cc: Changed the directory structure to match the
	  package system

2006-08-14 01:17  mcreel

	* bfgs.m: remove bfgs.m, bfgsmin.m does this job

2006-08-14 01:14  mcreel

	* numgradient.cc, numhessian.cc: direct incorporation of the work
	  previously done by finitedifference.cc

2006-08-14 01:13  mcreel

	* Makefile: addition of  __bfgsmin.cc, removal of bfgsmin.cc and
	  finitedifference.cc

2006-08-14 01:11  mcreel

	* __bfgsmin.cc: backend for bfgs and lbfgs minimization

2006-08-14 01:10  mcreel

	* bfgsmin.cc: remove bfgsmin.cc, replace with __bfgsmin.cc and
	  bfgsmin.m frontend

2006-08-14 01:08  mcreel

	* bfgsmin.m: .m file frontend for __bfgsmin.cc

2006-08-14 01:06  mcreel

	* bfgsmin_example.m: minor improvements

2006-08-14 01:05  mcreel

	* finitedifference.cc: remove finitedifference - functionality is
	  incorporated directly in needed places

2006-07-28 00:26  jwe

	* test_wpolyfit.m: transpose result from polyfit

2006-05-19 08:58  jwe

	* leval.cc: [no log message]

2006-05-12 15:22  adb014

	* leasqr.m: Initialization error in leasqr (For Thomas Weber /
	  Debian)

2006-04-10 10:58  adb014

	* Makefile, lp_test.m: other cleanups for removal of lp.cc

2006-04-08 16:25  adb014

	* lp.cc: Remove lp.cc to close long standing bug report as it is
	  superseded by glpk.cc in 2.9.x

2006-04-07 14:15  mcreel

	* Makefile: forgot to update with last bfgsmin change, sorry

2006-04-07 14:14  mcreel

	* bfgsmin.cc: bug in function name

2006-03-23 16:04  mcreel

	* bisectionstep.cc: remove bisectionstep.cc, it's now inside
	  bfgsmin.cc

2006-03-23 16:01  mcreel

	* newtonstep.cc: Remove newtonstep.cc, it is now incorporated in
	  bfgsmin.cc

2006-03-23 15:40  mcreel

	* rosenbrock.m: Multidimensional Rosenbrock function, used as
	  example in bfgsmin_example.m

2006-03-23 15:30  mcreel

	* bfgsmin_example.m: Different (better?) example, also more
	  informative usage message.

2006-03-23 15:29  mcreel

	* bfgsmin.cc: Incorporate stepsize, numeric gradient and finite
	  difference functions. This makes them all use common argument
	  checking, and simplifies code. Gives roughly 15% performance
	  improvement.

2006-03-20 10:01  pkienzle

	* leasqr.m: [for Olaf Till] better handling of step limits.

2006-01-07 06:11  adb014

	* .cvsignore: patch to allow 2.9.x autoload functionality to
	  replace symbolic links

2006-01-06 03:58  pkienzle

	* test_wpolyfit.m: Accept either row or column as returned
	  coefficients vector.

2005-12-31 11:07  pkienzle

	* wsolve.m: Allow for multiple y values when calculating degrees of
	  freedom.

2005-11-28 10:33  sjvdw

	* fmins.m: [with Mark van Rossum] Don't ignore specified options.

2005-11-01 16:39  pkienzle

	* polyconf.m: Clarify docs for control parameter 1-alpha

2005-10-30 13:20  aadler

	* fzero.m: no warning when user explicitly says options = []

2005-10-12 17:13  pkienzle

	* bfgsmin.cc: [for Orion Poplawski] Support 64-bit indexing.

2005-10-10 06:21  pkienzle

	* wpolyfit.m: Return the polynomial as a row vector.

2005-10-07 15:07  pkienzle

	* wpolyfit.m: If a third output argument is requested, center and
	  scale the data first.

2005-10-07 03:13  pkienzle

	* test_wpolyfit.m: Make it easier to swap between wpolyfit and
	  polyfit in the tests.

2005-09-28 09:12  mcreel

	* bfgsmin_example.m, samin_example.m: minor improvements

2005-09-08 05:07  pkienzle

	* d2_min.m, minimize.m, optimset.m: [for Bill Denney] is_struct ->
	  isstruct

2005-09-08 04:00  pkienzle

	* d2_min.m, deriv.m, fminunc.m, fzero.m, polyconf.m,
	  test_fminunc_1.m, test_minimize_1.m, wpolyfit.m: [for Bill
	  Denney] isstr -> ischar

2005-05-30 17:18  mcreel

	* bfgsmin.cc: cleanup tabbing

2005-05-29 18:00  pkienzle

	* fzero.m: Document extended arguments

2005-05-29 17:41  pkienzle

	* fzero.m: Remove spurious abs() from asserts

2005-05-29 17:40  pkienzle

	* fzero.m: [for Ben Barrowes] accept extra arguments for objective
	  function

2005-05-25 05:43  pkienzle

	* adsmax.m, cdiff.m, d2_min.m, expdemo.m, fzero.m, lp_test.m,
	  mdsmax.m, nelder_mead_min.m, nmsmax.m, test_d2_min_1.m,
	  test_d2_min_2.m, test_d2_min_3.m, test_nelder_mead_min_1.m,
	  test_nelder_mead_min_2.m, test_wpolyfit.m, wpolyfitdemo.m:
	  Author/Copyright consistency

2005-05-17 15:00  mcreel

	* bfgsmin.cc, bisectionstep.cc, celleval.cc, finitedifference.cc,
	  newtonstep.cc, numgradient.cc, numhessian.cc, samin.cc: fix tabs

2005-04-24 18:33  pkienzle

	* wsolve.m: Fix copyright notices.

2005-04-05 17:04  pkienzle

	* leasqrdemo.m: Restore matlab compatibility

2005-04-05 17:02  pkienzle

	* leasqr.m: [with Timothe Lecomte] use sparse matrices to save
	  storage.

2005-03-23 05:51  pkienzle

	* minimize.m: Don't use getfields.

2005-03-01 13:45  pkienzle

	* optimset.m: [for Heikki Junes] remove obsolete reference to
	  set_field

2005-03-01 04:27  pkienzle

	* fminunc.m, optimset.m: Use struct() rather than setfields() for
	  an empty structure

2005-01-14 10:18  mcreel

	* bfgsmin_example.m: improve explanatory comments

2005-01-14 10:17  mcreel

	* bfgsmin.cc: add check for obj fn value, fix lbfgs memory bug

2004-10-26 14:24  mcreel

	* bfgsmin.cc: add iters as output

2004-09-30 21:17  pkienzle

	* fminunc.m, minimize.m, optimset.m: Change to use
	  getfields/setfields

2004-09-28 12:23  mcreel

	* bfgsmin.cc, numgradient.cc, numhessian.cc: more arg checking

2004-09-17 11:55  mcreel

	* bfgsmin.cc, samin.cc: fix a recently-introduced bug (severely
	  C-challenged)

2004-09-14 14:16  mcreel

	* bfgsmin.cc, bisectionstep.cc, newtonstep.cc: final round arg
	  checking for these functions (famous last words!)

2004-09-11 16:40  przykry2004

	* fzero.m: Minor changes; octave-2.1.58's fsolve makes the 8th demo
	  break down...

2004-09-09 14:17  mcreel

	* samin-example.m: [no log message]

2004-09-09 14:16  mcreel

	* bfgsmin.cc, finitedifference.cc, samin.cc: more complere arg
	  checking

2004-09-09 02:26  adb014

	* Makefile: Remove deleted target

2004-09-08 18:19  pkienzle

	* fzero.m: Clean up argument checking and tests so they work in
	  2.1.58

2004-09-08 16:12  mcreel

	* lbfgsmin_example.m: [no log message]

2004-09-08 16:10  mcreel

	* bfgsmin.cc, bfgsmin_example.m: Integrate bfgsmin and lbfgsmin
	  into same file so convergence, etc. is same. Complete argument
	  checking (Go ahead - hit me!). Beautify output.

2004-09-08 16:07  mcreel

	* bfgsmin-example.m: [no log message]

2004-09-08 15:58  mcreel

	* lbfgsmin-example.m, lbfgsmin.cc: [no log message]

2004-09-07 12:31  mcreel

	* bfgsmin_example.m, lbfgsmin_example.m, samin_example.m: renamed
	  example files

2004-09-07 12:30  mcreel

	* bfgsmin.cc, lbfgsmin.cc, newtonstep.cc, samin.cc: Fix newtonstep
	  bug introduced in July, and add pointers to examples to algorithms

2004-09-06 15:50  mcreel

	* bisectionstep.cc, finitedifference.cc, newtonstep.cc: full
	  argument checking (number and types) added

2004-09-06 14:48  mcreel

	* bfgsmin-example.m, bfgsmin.cc, bisectionstep.cc,
	  finitedifference.cc, lbfgsmin.cc, newtonstep.cc: Improvement of
	  help strings and argument checking for internal use functions. Still
	  need to check types.

2004-09-03 16:25  przykry2004

	* fzero.m: Complete replacement for the old wrapper function. Now
	  fzero actually implements the Brent's root-finding method or, if
	  necessary, falls back to fsolve(). This function has been written by
	  a student of mine, Lukasz Bodzon.

2004-07-27 21:46  rlaboiss

	* fmins.m: Put the description of the options vector inside a
	  @verbatim environment.  I did not check whether this would look okay
	  when processed by TeX but, at least, the output of "help fmins" is
	  not gibberish anymore.

2004-07-12 16:56  pkienzle

	* polyconf.m: Added copyright notices

2004-07-09 11:03  mcreel

	* samin.cc: prevent temperature from becoming infinite in first
	  stage

2004-07-07 17:25  pkienzle

	* polyconf.m: Correct tests; hide unimplemented features and known
	  bugs in sparse

2004-07-07 11:38  adb014

	* Makefile: Allow dist, distclean and clean targets to run without
	  Makeconf. Replace include with sinclude in Makefiles. Build fixed
	  and comm docs as part of dist target

2004-06-30 11:35  mcreel

	* INDEX: more accurate sorting

2004-06-29 11:45  mcreel

	* lbfgsmin.cc: more realistic default tolerance for gradient - was
	  too picky

2004-06-29 11:44  mcreel

	* bfgsmin.cc: more realistic default tolerance for gradient, was
	  too picky

2004-06-28 14:11  mcreel

	* samin-example.m: applied insecticide, function minimum really is
	  zero now

2004-06-28 13:57  mcreel

	* bfgsmin.cc, bisectionstep.cc, celleval.cc, finitedifference.cc,
	  lbfgsmin.cc, newtonstep.cc, numgradient.cc, numhessian.cc,
	  samin.cc: re-indented using emacs

2004-06-21 18:29  mcreel

	* doc/: mintoolkit.lyx, mintoolkit.tex, readme.txt: commit
	  mintoolkit documentation

2004-06-19 22:18  pkienzle

	* INDEX: categorize new functions

2004-06-16 16:51  etienne

	* bfgsmin.cc, minimize.m, test_min_4.m, test_minimize_1.m: cruft
	  for bfgs -> bfgsmin transition

2004-06-14 20:07  pkienzle

	* wpolyfit.m: Explain what 'good fit' means.  Describe F-test for
	  determining polynomial order.

2004-06-04 02:27  pkienzle

	* minimize.m: use struct rather than setfield

2004-06-01 14:32  pkienzle

	* wpolyfit.m: Report probability rather than raw chi^2 numbers.

2004-06-01 02:02  pkienzle

	* wpolyfit.m, wsolve.m: details: sumsq( (y-f(x))/sigma ) is chisq,
	  chisq/df is normalized chisq

2004-05-31 20:12  pkienzle

	* wpolyfit.m: Update documentation; provide access to the internal
	  plot function

2004-05-31 19:31  pkienzle

	* polyconf.m: Add test against Hocking data

2004-05-31 04:48  pkienzle

	* polyconf.m, test_wpolyfit.m, wpolyfit.m, wsolve.m: wpolyfit now
	  returns a struction compatible with polyfit.

2004-05-26 15:29  mcreel

	* bfgsmin.cc, lbfgsmin.cc: improvements in determining convergence
	  when stepsize goes to zero

2004-05-26 15:28  mcreel

	* bfgsmin-example.m, lbfgsmin-example.m, samin-example.m: add
	  mintoolkit example files

2004-05-25 21:39  pkienzle

	* wpolyfit.m: Support matrix inputs: wpolyfit([x,y,dy],n).  Correct
	  comments.  Use 'uncertainty' instead of 'error'.  Remove GPL license
	  (trivial amounts of John's code or comments left).

2004-05-25 21:35  pkienzle

	* wsolve.m: Avoid assumptions about inv(Q) during derivation; check
	  for underdetermined systems.

2004-05-24 12:57  mcreel

	* Makefile: new .cc files

2004-05-24 12:56  mcreel

	* celleval.cc: oops, forgot to add it

2004-05-22 02:01  pkienzle

	* wpolyfit.m, wsolve.m: Separate system solver from observation
	  matrix construction; the solver is useful by itself, for example in
	  solving weighted systems.

2004-05-21 22:57  pkienzle

	* test_wpolyfit.m, wpolyfit.m: Correct the fit-through-origin test;
	  add monte carlo test; warn in documentation that monte carlo results
	  do not match fit results.

2004-05-21 21:57  pkienzle

	* wpolyfit.m: More accurate and faster computation of uncertainty
	  in parameters.

	  For weighted fits, use weights to compute uncertainty in parameters.
	  Now computing the standard deviation on the parameters from a monte
	  carlo simulation of an unweighted fit to data selected from a
	  gaussian distribution according to weights matches the uncertainty
	  computed from the weighted fit.

	  For unweighted data, continue to estimate uncertainty from chisq.

2004-05-20 14:09  mcreel

	* battery.m, bfgsmin.cc, bisectionstep.cc, finitedifference.cc,
	  lbfgsmin.cc, newtonstep.cc, numgradient.cc, numhessian.cc,
	  samin.cc: add the .cc files from mintoolkit

2004-05-20 14:05  mcreel

	* Makefile, bfgs.m, bs_gradient.m, test_min_1.m, test_min_2.m,
	  test_min_3.m: move ./optim mintoolkit to ./optim, replace bfgs.m
	  and bs_gradient.m

2004-05-20 02:35  pkienzle

	* Makefile: Build and install mintoolkit subdirectory (build works,
	  install untested)

2004-05-20 02:29  pkienzle

	* Makefile: Use lo_ieee_isnan rather than isnan.

2004-05-18 13:55  mcreel

	* mdsmax.m: fix crash when stopit is not provided (failure to use
	  default)

2004-05-18 13:54  mcreel

	* adsmax.m: fix crash when either stopit or P are not provided
	  (failure to use defaults)

2003-12-12 23:05  pkienzle

	* leval.cc: Distribute tests hardcoded in batch_test.m to their
	  respective function definition files

2003-12-08 09:21  pkienzle

	* lp.cc: Use lo_ieee_* functions for finite, isinf, isnan,
	  inf_value, nan_value

2003-11-11 22:40  pkienzle

	* INDEX: Transform "use <code>fn</code>" to "use <f>fn</f>", and
	  make it link to fn.  Fix links to octave-forge CVS Add links to
	  octave CVS We now walk the octave source tree instead of using
	  DOCSTRINGS, so some function descriptions are broken.  Perhaps we
	  could separate DOCSTRINGS from file locations somehow so everything
	  works nicely together.

2003-10-20 19:31  pkienzle

	* lp.cc, lp_test.m: Better argument checking (allow [] and scalar
	  as lower/upper bound).  A bit nicer code to transform bounds to
	  0-based bounds.  Fix array out of bounds reference (use 0-origin
	  array indices).  Warn that solution does not always work (like the
	  new test 6 for example).

2003-09-12 16:22  adb014

	* lp.cc, test_fminunc_1.m: Changes to allow use with latest CVS of
	  octave (do_fortran_indexing, etc)

2003-08-28 23:17  pkienzle

	* INDEX: Add some missing functions

2003-06-18 17:30  aadler

	* fmins.m, nmsmax.m: change default parameters

2003-06-06 20:30  pkienzle

	* wpolyfit.m: (Rolf Fabian) Handle complex-valued x,y.

2003-06-01 21:36  pkienzle

	* .cvsignore: Ignore build targets when doing cvs -q update -dP ---
	  makes it easier to notice that files have been added.

2003-05-15 21:56  etienne

	* test_min_2.m, test_min_3.m, test_min_4.m: Make tests reproducible

2003-05-15 21:55  etienne

	* line_min.m: Tweaking

2003-05-15 21:55  etienne

	* bfgs.m: Tweaking and Michael Creel's bugfix

2003-05-05 10:00  etienne

	* d2_min.m, minimize.m, nelder_mead_min.m: Remove 'interface
	  subject to change' message

2003-05-05 01:58  aadler

	* adsmax.m, mdsmax.m: maximize and minimize

2003-05-04 18:48  aadler

	* adsmax.m, fmins.m, mdsmax.m, nmsmax.m: functions to support fmins
	  (from www.maths.man.ac.uk/~higham)

2003-05-04 18:45  aadler

	* fminbnd.m: remove space at top - interferes with help

2003-03-09 03:21  pkienzle

	* lp.cc: [Stephane Kian (?)] Fix range error --- someone else will
	  have to confirm that the fix is reasonable.

2003-03-05 16:31  pkienzle

	* lp.cc: Backport to octave-2.1.36

2003-01-21 18:11  etienne

	* bfgs.m, d2_min.m: Soothe operator precedence warnings

2003-01-21 16:20  aadler

	* cdiff.m, fminunc.m, minimize.m, nelder_mead_min.m,
	  test_d2_min_3.m, test_fminunc_1.m, test_min_4.m,
	  test_minimize_1.m: minor modifications to work with --traditional

2002-12-24 07:33  pkienzle

	* lp.cc: No more implicit conversion from value to
	  octave_value_list.

2002-12-10 10:17  etienne

	* d2_min.m: Misc small changes

2002-12-08 19:12  etienne

	* bfgs.m, d2_min.m, nelder_mead_min.m: Typo in option name in help
	  text

2002-12-04 20:16  etienne

	* cdiff.m: ... -> varargin

2002-12-04 16:21  pkienzle

	* fmin.m, deriv.m, fminunc.m, minimize.m, nrm.m, optimset.m: use
	  varargin rather than ... in the function declarations

2002-12-03 23:11  pkienzle

	* fminbnd.m: ... -> varargin.  Also, x=1:3; x(2) = 4; no longer
	  works.

2002-11-27 09:40  pkienzle

	* expfit.m: author/license updates

2002-11-27 09:39  pkienzle

	* INDEX: Make sure all new functions are indexed

2002-11-27 06:26  pkienzle

	* lp_test.m: Use assert rather than error messages for tests

2002-11-16 22:12  pkienzle

	* lp.cc: Use c++ replacements for standard C headers.

2002-11-16 21:38  pkienzle

	* lp.cc: Windows dll and gcc 3.2 problems

2002-11-04 21:09  etienne

	* minimize.m, nelder_mead_min.m: varargout changes

2002-11-04 10:22  etienne

	* bfgs.m, bs_gradient.m, bs_gradient2.m, cdiff.m, deriv.m, fmin.m,
	  fminbnd.m, fminunc.m, line_min.m, minimize.m, nelder_mead_min.m,
	  nrm.m, optimset.m, test_fminunc_1.m, test_min_1.m,
	  test_minimize_1.m, doc/optim-mini-howto.2.html: varargin changes

2002-11-03 18:18  pkienzle

	* expfit.m: (for Rolf Fabian) correct handling of complex
	  exponentials

2002-10-04 16:12  etienne

	* doc/: optim-mini-howto.2.lyx, optim-mini-howto.2.tex,
	  figures/2D_slice-3.eps2, figures/optim_tutorial_slice.eps2:
	  Updates to tutorial

2002-09-20 18:09  etienne

	* INDEX: Remove removed files, add added files.

2002-07-31 16:48  pkienzle

	* wpolyfit.m: return values as row vectors for compatibility, and
	  cosmetic changes to graph.

2002-06-23 11:40  etienne

	* bfgs.m: Accept gradient of any shape

2002-06-02 17:38  pkienzle

	* test_wpolyfit.m, wpolyfit.m: Oops --- removed the need for
	  flipud(p) but didn't remove flipud(p) This didn't completely fix the
	  problem with my dataset, so I added the dataset to test_wpolyfit.

2002-05-30 19:27  pkienzle

	* wpolyfit.m, wpolyfitdemo.m: Option to force polynomial through
	  the origin.  Only return p,dp (yf is easy enough to calculate).
	  Tidy output when no polynomial is returned.  Parameterize demo to
	  take polynomial coefficients.

2002-05-17 21:00  etienne

	* test_min_1.m, test_cg_min_1.m, test_cg_min_2.m, test_cg_min_3.m,
	  test_cg_min_4.m, test_min_2.m, test_min_3.m, test_min_4.m: give
	  more generic name to test_cg_min_N.m

2002-05-17 20:32  etienne

	* __quasi_func__.m, bracket_min.m, brent_line_min.m, bs_gradient.m,
	  cg_min.m, dfp.m, semi_bracket.m: remove obsolete files and
	  replaced bs_gradient.m

2002-05-17 20:27  etienne

	* bfgs.m, d2_min.m, nelder_mead_min.m: misc changes

2002-05-07 18:14  etienne

	* minimize.m, test_minimize_1.m: add 'ndiff' option

2002-05-01 20:19  etienne

	* fminunc.m, optimset.m, test_fminunc_1.m: Add m*tlab compatibility
	  functions

2002-05-01 20:00  etienne

	* minimize.m, test_minimize_1.m: Accept struct or list of options

2002-04-28 21:57  etienne

	* minimize.m, test_minimize_1.m: new jac, hess and ihess options

2002-04-27 14:02  etienne

	* bfgs.m, bs_gradient2.m, test_cg_min_1.m, test_cg_min_2.m,
	  test_cg_min_3.m, test_cg_min_4.m: Change API: default=num. deriv,
	  accept other ways of differentiating

2002-04-23 23:18  pkienzle

	* test_wpolyfit.m: new file

2002-04-23 00:00  pkienzle

	* lp.cc: use more conventional octave error handling

2002-04-21 15:35  etienne

	* bfgs.m, line_min.m: small fixes in doc

2002-04-21 10:43  etienne

	* minimize.m, test_cg_min_1.m, test_cg_min_2.m, test_cg_min_3.m,
	  test_cg_min_4.m, test_minimize_1.m: use bfgs rather than cg_min

2002-04-21 10:41  etienne

	* bfgs.m: new interface

2002-04-11 06:28  pkienzle

	* INDEX: More categorization --- mostly complete now

2002-04-10 07:47  pkienzle

	* fminbnd.m: Tune the first sentence heuristic

2002-04-09 22:49  pkienzle

	* INDEX: Split main.idx into individual subdirectories to make it
	  more manageable

2002-04-07 18:34  etienne

	* cdiff.m: remove spurious sayif() calls

2002-04-07 18:03  etienne

	* cdiff.m: quicker code (smaller and less variables, less brackets)

2002-03-30 22:20  etienne

	* minimize.m: use new read_options()

2002-03-30 22:16  etienne

	* test_minimize_1.m: Fix typo

2002-03-29 17:42  etienne

	* test_minimize_1.m: test function for minimize

2002-03-29 17:41  etienne

	* minimize.m: change treatment of options, of backends, check
	  coherence of args

2002-03-29 13:11  pkienzle

	* leasqr.m: Update comment string

2002-03-28 22:16  etienne

	* test_cg_min_1.m, test_cg_min_2.m, test_cg_min_3.m: Change
	  messages

2002-03-28 21:09  etienne

	* d2_min.m: accept struct 'ctl' argument. Do multiple termination
	  criteria.

2002-03-28 21:08  etienne

	* test_d2_min_1.m, test_d2_min_2.m, test_d2_min_3.m: Change
	  messages

2002-03-28 20:08  etienne

	* nelder_mead_min.m: Fix doc. Twindle termination criterion
	  selection.

2002-03-28 16:52  etienne

	* test_nelder_mead_min_2.m: Test whether termination criteria work
	  correctly

2002-03-28 16:51  etienne

	* nelder_mead_min.m: Accept verbose option and multiple termination
	  criteria; Fix simplex volume criterion

2002-03-26 06:06  pkienzle

	* wpolyfit.m, wpolyfitdemo.m: Weighted least squares polynomial
	  fitting with error estimates on the parameters

2002-03-23 21:23  etienne

	* nelder_mead_min.m, doc/optim-mini-howto.2.html,
	  doc/optim-mini-howto.2.tex: add mini-howto on unconstrained
	  optimizations

2002-03-23 21:15  etienne

	* minimize.m: optimization front-end and dependencies

2002-03-23 21:13  etienne

	* cg_min.m: fix low-gradient termination criterion

2002-03-23 21:11  etienne

	* bracket_min.m, brent_line_min.m, poly_2_ex.m, semi_bracket.m:
	  dependencies for cg_min

2002-03-23 13:14  etienne

	* cg_min.m, test_cg_min_1.m, test_cg_min_2.m, test_cg_min_3.m: add
	  conj. grad. backend

2002-03-18 08:43  pkienzle

	* Makefile: Makefiles more consistent; accept optional "install"
	  target in tertiary makes

2002-03-16 18:15  etienne

	* d2_min.m: Accept 1st differential that is not row vector

2002-03-13 22:50  etienne

	* Makefile, cdiff.m, d2_min.m, leval.cc, nelder_mead_min.m,
	  test_d2_min_1.m, test_d2_min_2.m, test_d2_min_3.m,
	  test_nelder_mead_min_1.m: Add functions nelder_mead_min, d2_min
	  and cdiff + dependencies

2002-03-13 19:51  pkienzle

	* nrm.m: optimize update expression

2002-03-13 19:23  pkienzle

	* nrm.m: save two function evaluations by inlining the derivatives
	  pass all_va_args through to the function

2002-01-31 07:06  alnd

	* leasqr.m, leasqrdemo.m, dfdp.m:

	  Cosmetic changes in the comment section for 'Copyright' and/or
	  'Author' lines to enable parsing by the admin/get_authors script and
	  automatic generation of the AUTHORS file.  No changes to executable
	  code!

2002-01-29 17:53  pkienzle

	* expdemo.m, expfit.m: fit y = \sum{ a_i exp(c_i x) }

2002-01-04 16:53  pkienzle

	* lp.cc: Changes required to compile for gcc-3.0 in debian
	  hppa/unstable

2001-10-30 20:36  pkienzle

	* fminbnd.m: modified for use with functions which take more than
	  one parameter

2001-10-10 21:54  pkienzle

	* Makefile, __quasi_func__.m, bfgs.m, bs_gradient.m, deriv.m,
	  dfdp.m, dfp.m, fmin.m, fminbnd.m, fzero.m, leasqr.m,
	  leasqrdemo.m, lp.cc, lp_test.m, nrm.m: Initial revision

2001-10-10 21:54  pkienzle

	* Makefile, __quasi_func__.m, bfgs.m, bs_gradient.m, deriv.m,
	  dfdp.m, dfp.m, fmin.m, fminbnd.m, fzero.m, leasqr.m,
	  leasqrdemo.m, lp.cc, lp_test.m, nrm.m: revised heirarchy

