AMPL: A Modeling Language for Mathematical Programming

L-BFGS-B Directives for AMPL

To set these directives, assign a string specifying their values to the AMPL option lbfgsb_options. For example:
     ampl: option lbfgsb_options 'pgtol=1e-06 maxit=50';

Keyword Value
factrfactr * macheps = tolerance for relative change in function values; default 100.
iprintprint level:
-1 ==> no printing (default)
0 ==> 1 line at last iteration
0 < iprint < 99 ==> f and |proj g| every iprint iterations
100 ==> also changes of active set and final x
> 100 ==> details every iteration.
m number of corrections, 3 <= m <= 20 recommended; default = 10.
maxfun maximum function and gradient evaluations (default 2000).
maxit maximum iterations (default 1000).
nprob objective choice: 1 (default) = 1st.
outlev synonym for iprint.
pgtol tolerance for max(|proj g_i|), default 1e-12.
wantsol solution report without -AMPL: sum of
1 ==> write .sol file
2 ==> print variable values
8 ==> do not print solution message

Comments or questions?
Write to or use our comment form.

Return to the AMPL remote access page.

Return to the AMPL home page.