2008
DOI: 10.1007/978-3-540-75199-1_34
|View full text |Cite
|
Sign up to set email alerts
|

The All-floating BETI Method: Numerical Results

Abstract: Summary. The all-floating BETI method considers all subdomains as floating subdomains and improves the overall asymptotic complexity of the BETI method. This effect and the scalability of the method are shown in numerical examples.

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1

Citation Types

0
4
0

Year Published

2008
2008
2013
2013

Publication Types

Select...
4
3

Relationship

0
7

Authors

Journals

citations
Cited by 11 publications
(4 citation statements)
references
References 13 publications
0
4
0
Order By: Relevance
“…However, the flux density |B| may vary strongly along subdomain boundaries and large values of |B| appear mostly at singularities of the potential u, e. g., near material corners. Contrary to the usual suggestion to choose subdomain partitions that resolve material interfaces in order to obtain robustness (for numerical examples see [4,5]), our new bounds (7), (9) suggest that it might be more advantageous to put each peak of |B| and thus each material corner into the center of a subdomain. Fig.…”
Section: Application To Magnetostatic Problemsmentioning
confidence: 83%
“…However, the flux density |B| may vary strongly along subdomain boundaries and large values of |B| appear mostly at singularities of the potential u, e. g., near material corners. Contrary to the usual suggestion to choose subdomain partitions that resolve material interfaces in order to obtain robustness (for numerical examples see [4,5]), our new bounds (7), (9) suggest that it might be more advantageous to put each peak of |B| and thus each material corner into the center of a subdomain. Fig.…”
Section: Application To Magnetostatic Problemsmentioning
confidence: 83%
“…Following [10] (see also [36]) and taking the all-floating idea into account [26,27] (see also [7] for a closely related approach called total FETI), we tear the global skeleton potential vector u C on all subdomain boundaries Γ i including the Dirichlet parts by introducing the individual local unknowns u C,i = R i u C . The global continuity of the potentials on Γ i \ Γ and the Dirichlet boundary conditions on Γ D = Γ are now enforced by the constraints…”
Section: Coupled Beti -Feti Systemsmentioning
confidence: 99%
“…, p, where the coefficient a(•) is slightly varying, we use the upper bounds a i for the entries in Q. The matrix L 0 is defined by the representation of λ in the form λ = L 0 λ 0 + λ e (27) with known λ e = QG(G QG) −1 e fulfilling the constraints G λ e = e, and unknown L 0 λ 0 ∈ ker G , i. e., G L 0 λ 0 = 0. The right-hand side of BETI-FETI-2 system (24) is defined by the relations d C,i = b C,i − B C,i λ e for i = 1, .…”
Section: Coupled Beti -Feti Systemsmentioning
confidence: 99%
“…In [14,20], we suggested the "all-floating" BETI method. Related numerical results were presented in [21]. In the meantime, the all-floating formulation was also applied for the coupling of FETI and BETI methods [15,25].…”
Section: Introductionmentioning
confidence: 99%