Re: [Gretl-devel] limited memory bfgs
Am 03.04.2016 um 12:29 schrieb oleg_komashko(a)ukr.net:
Dear all,
The User's Guide, 31.1, says
Limited-memory variant
See Byrd et al. (1995) (. . . FIXME: expand a little here . . . )
Is it not implemented yet, or just not described?
If the latter is true, how to use it?
I don't know any details, but see the 'set' command with these options:
lbfgs: on or off (the default). Use the limited-memory version of BFGS
(L-BFGS-B) instead of the ordinary algorithm. This may be advantageous
when the function to be maximized is not globally concave.
• lbfgs_mem: an integer value in the range 3 to 20 (with a default value
of 8). This determines the number of corrections used in the limited
memory matrix when L-BFGS-B is employed.
-- I guess there could be a cross-link from the bfgs() function doc to
these 'set' options.
hth,
sven