Margaret H. Wright
||NPSOL uses an SQP algorithm in which the subproblems have the same linearized constraints as in MINOS, but the objective is a quadratic approximation to the Lagrangian. (Hence, no function or gradient values are needed during the solution of each QP.) A merit function promotes convergence from arbitrary starting points.
||May be called from a driver program, typically in Fortran, C or MATLAB
||Modeling Languages link
||Matlab interface included. CUTE interface available
||Licensing Contact: Hans Wiesendanger, Licensing Associate, Stanford University Office of Technology Licensing E-Mail: email@example.com, Phone: (650) 723-0692
||Any machine with a reasonable amount of memory and a Fortran compiler
||NPSOL is a software package for solving constrained optimization problems (nonlinear programs). It employs a dense SQP algorithm and is especially effective for nonlinear problems whose functions and gradients are expensive to evaluate. The functions should be smooth but need not be convex. An augmented Lagrangian merit function ensures convergence from an arbitrary point.
Numerically stable algorithms. Global convergence. Needs only first derivatives. Can estimate them by differences. Automatic computation of finite difference intervals if necessary. Warm start capability.
||P. E. Gill, W. Murray, M. A. Saunders and M. H. Wright, User's Guide for NPSOL (Version 4.0): a Fortran package for nonlinear programming, Report SOL 86-2, Systems Optimization Laboratory, Stanford University (1986).
P. E. Gill, W. Murray, M. A. Saunders and M. H. Wright, Some theoretical properties of an augmented Lagrangian merit function, in P. M. Pardalos (ed.), Advances in Optimization and Parallel Computing, North-Holland, Amsterdam, 101--128 (1992).