MPRGP for Bound Constraints

Zdeněk Dostál ()
Additional contact information
Zdeněk Dostál: VŠB - Technical University Ostrava

Chapter 10 in Optimal Quadratic Programming and QCQP Algorithms with Applications, 2025, pp 215-247 from Springer

Abstract: Abstract The core of this chapter is the description and analysis of the MPRGP (modified proportioning with reduced gradient projection) algorithm for solving bound-constrained quadratic programming problems. This active-set-based algorithm expands the active set by the free gradient projection and carries out minimization in the face by the conjugate gradient method. The algorithm balances the norms of the free and chopped gradients to ensure that the reduced free gradient is always sufficiently large in the CG iterations. The modified algorithm enjoys the R-linear rate of convergence of both the cost function and norm of projected gradient. Moreover, it has the finite termination property even for dual degenerate QP problems. We show that conjugate projector preconditioning (deflation) can improve the convergence rate. We also show how to adapt MPRGP to solve problems with SPS Hessian. The performance of the algorithms, including scalability, is illustrated by solving academic benchmarks and particle simulation.

Date: 2025
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:spochp:978-3-031-95167-1_10

Ordering information: This item can be ordered from
http://www.springer.com/9783031951671

DOI: 10.1007/978-3-031-95167-1_10

Access Statistics for this chapter

More chapters in Springer Optimization and Its Applications from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().