MTRON : A Matlab interface to the TRON package of C.-J. Lin and J. Moreé.
Chih-Jen Lin and Jorge Moreé, Newton's method for large
bound-constrained optimization problems, SIAM Journal on Optimization,
Volume 9, Number 4, pp. 1100-1127, 1999
(as of 28 November 2006)
MTRON was written by Christoph Ortner
TRON is a trust region method for the solution of large scale
bound-constraint optimization problems, i.e. problems of the form
min f(x) subject to a_j \leq x_j \leq b_j
where f : R^n --> R^n should be twice differentiable.
TRON works by reverse communication, i.e., the user must write the
optimization loop and calls dtron.f at each iteration. mtron.c
simply translates dtron.f into Matlab. Its calling convention
is described in mtron.m. test_mtron.m is an example for using
MTRON in this way.
Alternatively, the interface routine itron.m may be used which is
modeled after the optimization toolbox routine fmincon (though
itron only supports bound-constraints)
readme.txt : this file
mtron.c : gateway routine
mtron.m : helptext for using mtron.c
itron.m : driver routine, which encapsulates the reverse communication
principle of TRON and mtron.
make_mtron.m : script for compiling mtron.c
test_mtron.m : implemenation of a simple test problem (uses only mtron.c
and not the driver routine itron)
test_itron.m : test problem for itron
plap.m : definition of the test problem used in test_mtron and test_itron