nlpSolve
Solves general linear and nonlinear
programming problems.
Main features:

nlpSolve is a Matlab implementation of the
Filter SQP by Roger Fletcher and Sven Leyffer,
Nonlinear Programming without a penalty function,
Dundee NA/171, 1997.

It is an advantage if the user can supply a routine
to compute the second derivatives of the constraints.

nlpSolve is using second order information, if available
and is using a general problem formulation with lower/upper bound format,
It is an advantage if the user can supply a routine
to compute the second derivatives of the constraints.

The user can select to run either the standard
Filter SQP
or a quasiNewton BFGS Filter SQP,
estimating the Hessian matrix using only the gradient information.
If the gradient is estimated numerically, the user must select to
use the BFGS, otherwise numerical problems are likely.

In Tomlab /MINOS nlpSolve is using QPOPT (or MINOS) as QP solver
giving speed and robustness.

It has
explicit treatment of simple bounds, linear and nonlinear constraints.

The Filter SQP algorithm
is also available in the
TOMLAB /MINLP toolbox
as filterSQP,
a Fortran version written by Fletcher and Leyffer.
