Optimal Transfer Operators in Algebraic Two-Level Methods for Nonsymmetric and Indefinite Problems
SIAM Journal on Matrix Analysis and Applications · 30.06.2026 · 47 (2) · 824-851
DOI: 10.1137/25m179436x ↗Kurz gefasst
Fachartikel in der Zeitschrift SIAM Journal on Matrix Analysis and Applications, online erschienen am 19.06.2026. Verfasst von Oliver A. Krzysik und 5 weiteren Autor:innen. Die Arbeit nennt 47 Literaturangaben. Gefördert von: Los Alamos National Laboratory.
Abstract (Originalsprache)
Consider an algebraic two-level method applied to the [Formula: see text]-dimensional linear system [Formula: see text] using fine-space preconditioner (i.e., “relaxation” or “smoother”) [Formula: see text], with [Formula: see text], restriction and interpolation [Formula: see text] and [Formula: see text], and algebraic coarse-space operator [Formula: see text]. Then, what are the best possible transfer operators [Formula: see text] and [Formula: see text] of a given dimension [Formula: see text]? Brannick et al. [ 12 ] showed that when [Formula: see text] and [Formula: see text] are Hermitian positive definite (HPD), the optimal interpolation is such that its range contains the [Formula: see text] smallest generalized eigenvectors of the matrix pencil [Formula: see text]. Recently, in Ali et al. [ 5 ] we generalized this framework to the non-HPD setting, by considering both right (interpolation) and left (restriction) generalized eigenvectors of [Formula: see text] and defining corresponding nonsymmetric transfer operators [Formula: see text]. Tight convergence bounds for [Formula: see text] are derived in spectral radius, as well as a proof of pseudo-optimality. Note, [Formula: see text] are typically complex valued, which is not practical for real-valued problems. Here we build on [ 5 ], first characterizing all inner products in which the coarse-space correction defined by [Formula: see text] is orthogonal. We then develop tight two-level convergence bounds in these norms, and prove that the underlying transfer operators [Formula: see text] are genuinely optimal. As a special case, our theory both recovers and extends the HPD results from [ 12 ]. Finally, we show how to construct optimal, real-valued transfer operators in the case of that [Formula: see text] and [Formula: see text] are real valued, but are not HPD. Numerical examples arising from a discretized advection-reaction equation, wave-equation, and Stokes equations are used to verify and illustrate the theory.
Förderung
- Los Alamos National Laboratory
Zitieren
APA
Krzysik, O. A., Southworth, B. S., Wimmer, G. A., Ali, A., Brannick, J., & Kahl, K. (2026). Optimal Transfer Operators in Algebraic Two-Level Methods for Nonsymmetric and Indefinite Problems. SIAM Journal on Matrix Analysis and Applications, 47(2), 824-851. https://doi.org/10.1137/25m179436x
Vancouver
Krzysik OA, Southworth BS, Wimmer GA, Ali A, Brannick J, Kahl K. Optimal Transfer Operators in Algebraic Two-Level Methods for Nonsymmetric and Indefinite Problems. SIAM Journal on Matrix Analysis and Applications. 2026;47(2):824-851. doi:10.1137/25m179436x
BibTeX
@article{Krzysik2026,
author = {Krzysik, Oliver A. and Southworth, Ben S. and Wimmer, Golo A. and Ali, Ahsan and Brannick, James and Kahl, Karsten},
title = {Optimal Transfer Operators in Algebraic Two-Level Methods for Nonsymmetric and Indefinite Problems},
journal = {SIAM Journal on Matrix Analysis and Applications},
year = {2026},
volume = {47},
number = {2},
pages = {824--851},
doi = {10.1137/25m179436x}
}
EUVIMED