1 NLopt 1.1 (28 October 2009)
3 * Added --without-octave and --without-matlab configure options to
4 disable Octave and Matlab plugins, respectively.
6 * Modified COBYLA algorithm to have better support for bound
9 * Experimental support for nonlinear equality constraints via
10 augmented-Lagrangian method.
12 * Support for compiling under Windows (ideally with MinGW) as a
13 DLL, although you have to manually add #define NLOPT_DLL
14 to nlopt.h *after* installing (after compiling NLopt).
16 * Workaround for incompatible qsort_r functions in BSD and GNU libc
17 by always using my own version; thanks to Wendy Vandoolaeghe
18 and Philippe Preux for the bug report and explanation.
20 * Workaround for gcc 3.4.x conflict with HUGE_VAL definition in Solaris
23 * Better identification of Matlab-plugin installation directory.
25 * Fixed identification of Octave-plugin installation directory for
26 recent Octave versions.
28 NLopt 1.0.1 (13 Nov. 2008)
30 * Allow user to override Matlab-plugin installation directory with
33 * Bug fix in my DIRECT code that prevented convergence (DIRECT-L unaffected).
35 * MLSL needs a nonzero default ftol_rel and/or xtol_rel to ensure that
36 its local searches terminate; use roughly machine precision as defaults.
38 NLopt 1.0 (11 Nov. 2008)
40 * Initial public release.