Nonlinear Programming Packages

QPOPT Authors Philip Gill
Walter Murray
Michael Saunders
Language Fortran 77
Algorithm  
Input Format May be called from a driver program, typically in Fortran, C or MATLAB
Modeling Languages link Matlab interface included
Commercial Status Licensing Contact: Hans Wiesendanger, Licensing Associate, Stanford University Office of Technology Licensing E-Mail: hans@otlmail.stanford.edu, Phone: (650) 723-0692
Platform Any machine with a reasonable amount of memory and a Fortran compiler
Remarks QPOPT is a software package for solving dense linear and quadratic programs. If the quadratic objective function is convex (definite or semidefinite), the solution obtained is a global optimum. For non-convex problems, the solution may be a local optimum or a dead-point (or unbounded).
The quadratic form x'Qx is defined by a user routine that computes Qx for a given vector x. (Hence some advantage arises if Q is sparse.) Linear constraints and bounds on the variables are treated separately by an active-set method. If the problem has no feasible solution, QPOPT minimizes the sum of the constraint and bound violations.
Numerically stable algorithms. Implicit definition of the quadratic objective via Qx products. Warm start capability. Elastic bounds on variables and constraints (for infeasible problems).
General-purpose dense linear programming. Convex or non-convex quadratic programming
References P. E. Gill, W. Murray and M. A. Saunders, User's guide for QPOPT 1.0: A Fortran package for quadratic programming, Report SOL 95-4, Systems Optimization Laboratory, Stanford University (1995). (Same as Report NA 95-1, Dept of Mathematics, University of California, San Diego, 1995).