Skip to Main content Skip to Navigation
Reports

QPAL -- A solver of convex quadratic optimization problems, using an augmented Lagrangian approach

Abstract : QPAL is a piece of software that aims at solving a convex quadratic optimization problem with linear equality and inequality constraints. The implemented algorithm uses an augmented Lagrangian approach, which relaxes the equality constraints and deals explicitely with the bound constraints on the original and slack variables. The generated quadratic functions are minimized on the activated faces by a truncated conjugate gradient algorithm, interspersed with gradient projection steps. When the optimal value is finite, convergence occurs at a linear speed that can be prescribed by the user. Matrices can be strored in dense or sparse structures; in addition, the Hessian of the quadratic objective function may have a direct of inverse $\ell$-BFGS form. QPAL is written in Fortran-2003.
Complete list of metadata

Cited literature [6 references]  Display  Hide  Download

https://hal.inria.fr/inria-00442295
Contributor : Jean Charles Gilbert <>
Submitted on : Saturday, December 19, 2009 - 3:08:48 PM
Last modification on : Friday, May 25, 2018 - 12:02:03 PM
Long-term archiving on: : Thursday, June 17, 2010 - 11:57:07 PM

File

RT-0377.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : inria-00442295, version 1

Collections

Citation

Jean Charles Gilbert. QPAL -- A solver of convex quadratic optimization problems, using an augmented Lagrangian approach. [Technical Report] RT-0377, INRIA. 2009, pp.22. ⟨inria-00442295⟩

Share

Metrics

Record views

298

Files downloads

246