An adaptive choice of primal constraints for BDDC domain decomposition algorithms

Juan G. Calvo, Olof B. Widlund

Research output: Contribution to journalArticle

Abstract

An adaptive choice for primal spaces based on parallel sums is developed for BDDC deluxe methods and elliptic problems in three dimensions. The primal space, which forms the global, coarse part of the domain decomposition algorithm and which is always required for any competitive algorithm, is defined in terms of generalized eigenvalue problems related to subdomain edges and faces; selected eigenvectors associated to the smallest eigenvalues are used to enhance the primal spaces. This selection can be made automatic by using tolerance parameters specified for the subdomain faces and edges. Numerical results verify the results and provide a comparison with primal spaces commonly used. They include results for cubic subdomains as well as subdomains obtained by a mesh partitioner. Different distributions for the coefficients are also considered with constant coefficients, highly random values, and channel distributions.

Original languageEnglish (US)
Pages (from-to)524-544
Number of pages21
JournalElectronic Transactions on Numerical Analysis
Volume45
StatePublished - 2016

Fingerprint

Decomposition Algorithm
Domain Decomposition
Face
Generalized Eigenvalue Problem
Smallest Eigenvalue
Space Form
Coefficient
Elliptic Problems
Eigenvector
Tolerance
Three-dimension
Mesh
Verify
Numerical Results

Keywords

  • Adaptive primal constraints
  • BDDC deluxe preconditioners
  • Domain decomposition
  • Elliptic problems

ASJC Scopus subject areas

  • Analysis

Cite this

An adaptive choice of primal constraints for BDDC domain decomposition algorithms. / Calvo, Juan G.; Widlund, Olof B.

In: Electronic Transactions on Numerical Analysis, Vol. 45, 2016, p. 524-544.

Research output: Contribution to journalArticle

@article{436dbae00670479d82ae1918168842ff,
title = "An adaptive choice of primal constraints for BDDC domain decomposition algorithms",
abstract = "An adaptive choice for primal spaces based on parallel sums is developed for BDDC deluxe methods and elliptic problems in three dimensions. The primal space, which forms the global, coarse part of the domain decomposition algorithm and which is always required for any competitive algorithm, is defined in terms of generalized eigenvalue problems related to subdomain edges and faces; selected eigenvectors associated to the smallest eigenvalues are used to enhance the primal spaces. This selection can be made automatic by using tolerance parameters specified for the subdomain faces and edges. Numerical results verify the results and provide a comparison with primal spaces commonly used. They include results for cubic subdomains as well as subdomains obtained by a mesh partitioner. Different distributions for the coefficients are also considered with constant coefficients, highly random values, and channel distributions.",
keywords = "Adaptive primal constraints, BDDC deluxe preconditioners, Domain decomposition, Elliptic problems",
author = "Calvo, {Juan G.} and Widlund, {Olof B.}",
year = "2016",
language = "English (US)",
volume = "45",
pages = "524--544",
journal = "Electronic Transactions on Numerical Analysis",
issn = "1068-9613",
publisher = "Kent State University",

}

TY - JOUR

T1 - An adaptive choice of primal constraints for BDDC domain decomposition algorithms

AU - Calvo, Juan G.

AU - Widlund, Olof B.

PY - 2016

Y1 - 2016

N2 - An adaptive choice for primal spaces based on parallel sums is developed for BDDC deluxe methods and elliptic problems in three dimensions. The primal space, which forms the global, coarse part of the domain decomposition algorithm and which is always required for any competitive algorithm, is defined in terms of generalized eigenvalue problems related to subdomain edges and faces; selected eigenvectors associated to the smallest eigenvalues are used to enhance the primal spaces. This selection can be made automatic by using tolerance parameters specified for the subdomain faces and edges. Numerical results verify the results and provide a comparison with primal spaces commonly used. They include results for cubic subdomains as well as subdomains obtained by a mesh partitioner. Different distributions for the coefficients are also considered with constant coefficients, highly random values, and channel distributions.

AB - An adaptive choice for primal spaces based on parallel sums is developed for BDDC deluxe methods and elliptic problems in three dimensions. The primal space, which forms the global, coarse part of the domain decomposition algorithm and which is always required for any competitive algorithm, is defined in terms of generalized eigenvalue problems related to subdomain edges and faces; selected eigenvectors associated to the smallest eigenvalues are used to enhance the primal spaces. This selection can be made automatic by using tolerance parameters specified for the subdomain faces and edges. Numerical results verify the results and provide a comparison with primal spaces commonly used. They include results for cubic subdomains as well as subdomains obtained by a mesh partitioner. Different distributions for the coefficients are also considered with constant coefficients, highly random values, and channel distributions.

KW - Adaptive primal constraints

KW - BDDC deluxe preconditioners

KW - Domain decomposition

KW - Elliptic problems

UR - http://www.scopus.com/inward/record.url?scp=85010677175&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85010677175&partnerID=8YFLogxK

M3 - Article

VL - 45

SP - 524

EP - 544

JO - Electronic Transactions on Numerical Analysis

JF - Electronic Transactions on Numerical Analysis

SN - 1068-9613

ER -