Loading…
Accurate error estimation in CG
In practical computations, the (preconditioned) conjugate gradient (P)CG method is the iterative method of choice for solving systems of linear algebraic equations A x = b with a real symmetric positive definite matrix A . During the iterations, it is important to monitor the quality of the approxim...
Saved in:
Published in: | Numerical algorithms 2021-11, Vol.88 (3), p.1337-1359 |
---|---|
Main Authors: | , , |
Format: | Article |
Language: | English |
Subjects: | |
Citations: | Items that this one cites Items that cite this one |
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
cited_by | cdi_FETCH-LOGICAL-c319t-8dddeefa9b45263e928b1c8424abf6b702fc83e7a5b7b9c5325447134eb86b1f3 |
---|---|
cites | cdi_FETCH-LOGICAL-c319t-8dddeefa9b45263e928b1c8424abf6b702fc83e7a5b7b9c5325447134eb86b1f3 |
container_end_page | 1359 |
container_issue | 3 |
container_start_page | 1337 |
container_title | Numerical algorithms |
container_volume | 88 |
creator | Meurant, Gérard Papež, Jan Tichý, Petr |
description | In practical computations, the (preconditioned) conjugate gradient (P)CG method is the iterative method of choice for solving systems of linear algebraic equations
A
x
=
b
with a real symmetric positive definite matrix
A
. During the iterations, it is important to monitor the quality of the approximate solution
x
k
so that the process could be stopped whenever
x
k
is accurate enough. One of the most relevant quantities for monitoring the quality of
x
k
is the squared
A
-norm of the error vector
x
−
x
k
. This quantity cannot be easily evaluated; however, it can be estimated. Many of the existing estimation techniques are inspired by the view of CG as a procedure for approximating a certain Riemann–Stieltjes integral. The most natural technique is based on the Gauss quadrature approximation and provides a lower bound on the quantity of interest. The bound can be cheaply evaluated using terms that have to be computed anyway in the forthcoming CG iterations. If the squared
A
-norm of the error vector decreases rapidly, then the lower bound represents a tight estimate. In this paper, we suggest a heuristic strategy aiming to answer the question of how many forthcoming CG iterations are needed to get an estimate with the prescribed accuracy. Numerical experiments demonstrate that the suggested strategy is efficient and robust. |
doi_str_mv | 10.1007/s11075-021-01078-w |
format | article |
fullrecord | <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_journals_2918610119</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2918610119</sourcerecordid><originalsourceid>FETCH-LOGICAL-c319t-8dddeefa9b45263e928b1c8424abf6b702fc83e7a5b7b9c5325447134eb86b1f3</originalsourceid><addsrcrecordid>eNp9kE9LxDAUxIMouK5-AS8WPEfz8qdJjkvRVVjYy3oOSZpIF23XpGXx2xut4M3Tm8PMvOGH0DWQOyBE3mcAIgUmFDApSuHjCVqAkBRrWovToglIDEyrc3SR854UF6FygW5W3k_JjqEKKQ2pCnns3u3YDX3V9VWzvkRn0b7lcPV7l-jl8WHXPOHNdv3crDbYM9AjVm3bhhCtdlzQmgVNlQOvOOXWxdpJQqNXLEgrnHTaC0YF5xIYD07VDiJbotu595CGj6msMPthSn15aagGVZf9oIuLzi6fhpxTiOaQytz0aYCYbxBmBmEKCPMDwhxLiM2hXMz9a0h_1f-kvgAjtl9A</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2918610119</pqid></control><display><type>article</type><title>Accurate error estimation in CG</title><source>Springer Nature</source><creator>Meurant, Gérard ; Papež, Jan ; Tichý, Petr</creator><creatorcontrib>Meurant, Gérard ; Papež, Jan ; Tichý, Petr</creatorcontrib><description>In practical computations, the (preconditioned) conjugate gradient (P)CG method is the iterative method of choice for solving systems of linear algebraic equations
A
x
=
b
with a real symmetric positive definite matrix
A
. During the iterations, it is important to monitor the quality of the approximate solution
x
k
so that the process could be stopped whenever
x
k
is accurate enough. One of the most relevant quantities for monitoring the quality of
x
k
is the squared
A
-norm of the error vector
x
−
x
k
. This quantity cannot be easily evaluated; however, it can be estimated. Many of the existing estimation techniques are inspired by the view of CG as a procedure for approximating a certain Riemann–Stieltjes integral. The most natural technique is based on the Gauss quadrature approximation and provides a lower bound on the quantity of interest. The bound can be cheaply evaluated using terms that have to be computed anyway in the forthcoming CG iterations. If the squared
A
-norm of the error vector decreases rapidly, then the lower bound represents a tight estimate. In this paper, we suggest a heuristic strategy aiming to answer the question of how many forthcoming CG iterations are needed to get an estimate with the prescribed accuracy. Numerical experiments demonstrate that the suggested strategy is efficient and robust.</description><identifier>ISSN: 1017-1398</identifier><identifier>EISSN: 1572-9265</identifier><identifier>DOI: 10.1007/s11075-021-01078-w</identifier><language>eng</language><publisher>New York: Springer US</publisher><subject>Accuracy ; Algebra ; Algorithms ; Approximation ; Computer Science ; Eigenvalues ; Errors ; Iterative methods ; Linear algebra ; Lower bounds ; Mathematical analysis ; Matrices (mathematics) ; Matrix algebra ; Numeric Computing ; Numerical Analysis ; Original Paper ; Quadratures ; Robustness (mathematics) ; Stieltjes integral ; Theory of Computation</subject><ispartof>Numerical algorithms, 2021-11, Vol.88 (3), p.1337-1359</ispartof><rights>The Author(s), under exclusive licence to Springer Science+Business Media, LLC part of Springer Nature 2021</rights><rights>The Author(s), under exclusive licence to Springer Science+Business Media, LLC part of Springer Nature 2021.</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c319t-8dddeefa9b45263e928b1c8424abf6b702fc83e7a5b7b9c5325447134eb86b1f3</citedby><cites>FETCH-LOGICAL-c319t-8dddeefa9b45263e928b1c8424abf6b702fc83e7a5b7b9c5325447134eb86b1f3</cites><orcidid>0000-0001-6008-4056 ; 0000-0002-6036-3482</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,780,784,27924,27925</link.rule.ids></links><search><creatorcontrib>Meurant, Gérard</creatorcontrib><creatorcontrib>Papež, Jan</creatorcontrib><creatorcontrib>Tichý, Petr</creatorcontrib><title>Accurate error estimation in CG</title><title>Numerical algorithms</title><addtitle>Numer Algor</addtitle><description>In practical computations, the (preconditioned) conjugate gradient (P)CG method is the iterative method of choice for solving systems of linear algebraic equations
A
x
=
b
with a real symmetric positive definite matrix
A
. During the iterations, it is important to monitor the quality of the approximate solution
x
k
so that the process could be stopped whenever
x
k
is accurate enough. One of the most relevant quantities for monitoring the quality of
x
k
is the squared
A
-norm of the error vector
x
−
x
k
. This quantity cannot be easily evaluated; however, it can be estimated. Many of the existing estimation techniques are inspired by the view of CG as a procedure for approximating a certain Riemann–Stieltjes integral. The most natural technique is based on the Gauss quadrature approximation and provides a lower bound on the quantity of interest. The bound can be cheaply evaluated using terms that have to be computed anyway in the forthcoming CG iterations. If the squared
A
-norm of the error vector decreases rapidly, then the lower bound represents a tight estimate. In this paper, we suggest a heuristic strategy aiming to answer the question of how many forthcoming CG iterations are needed to get an estimate with the prescribed accuracy. Numerical experiments demonstrate that the suggested strategy is efficient and robust.</description><subject>Accuracy</subject><subject>Algebra</subject><subject>Algorithms</subject><subject>Approximation</subject><subject>Computer Science</subject><subject>Eigenvalues</subject><subject>Errors</subject><subject>Iterative methods</subject><subject>Linear algebra</subject><subject>Lower bounds</subject><subject>Mathematical analysis</subject><subject>Matrices (mathematics)</subject><subject>Matrix algebra</subject><subject>Numeric Computing</subject><subject>Numerical Analysis</subject><subject>Original Paper</subject><subject>Quadratures</subject><subject>Robustness (mathematics)</subject><subject>Stieltjes integral</subject><subject>Theory of Computation</subject><issn>1017-1398</issn><issn>1572-9265</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2021</creationdate><recordtype>article</recordtype><recordid>eNp9kE9LxDAUxIMouK5-AS8WPEfz8qdJjkvRVVjYy3oOSZpIF23XpGXx2xut4M3Tm8PMvOGH0DWQOyBE3mcAIgUmFDApSuHjCVqAkBRrWovToglIDEyrc3SR854UF6FygW5W3k_JjqEKKQ2pCnns3u3YDX3V9VWzvkRn0b7lcPV7l-jl8WHXPOHNdv3crDbYM9AjVm3bhhCtdlzQmgVNlQOvOOXWxdpJQqNXLEgrnHTaC0YF5xIYD07VDiJbotu595CGj6msMPthSn15aagGVZf9oIuLzi6fhpxTiOaQytz0aYCYbxBmBmEKCPMDwhxLiM2hXMz9a0h_1f-kvgAjtl9A</recordid><startdate>20211101</startdate><enddate>20211101</enddate><creator>Meurant, Gérard</creator><creator>Papež, Jan</creator><creator>Tichý, Petr</creator><general>Springer US</general><general>Springer Nature B.V</general><scope>AAYXX</scope><scope>CITATION</scope><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>AFKRA</scope><scope>ARAPS</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>GNUQQ</scope><scope>HCIFZ</scope><scope>JQ2</scope><scope>K7-</scope><scope>L6V</scope><scope>M7S</scope><scope>P62</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PTHSS</scope><orcidid>https://orcid.org/0000-0001-6008-4056</orcidid><orcidid>https://orcid.org/0000-0002-6036-3482</orcidid></search><sort><creationdate>20211101</creationdate><title>Accurate error estimation in CG</title><author>Meurant, Gérard ; Papež, Jan ; Tichý, Petr</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c319t-8dddeefa9b45263e928b1c8424abf6b702fc83e7a5b7b9c5325447134eb86b1f3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Accuracy</topic><topic>Algebra</topic><topic>Algorithms</topic><topic>Approximation</topic><topic>Computer Science</topic><topic>Eigenvalues</topic><topic>Errors</topic><topic>Iterative methods</topic><topic>Linear algebra</topic><topic>Lower bounds</topic><topic>Mathematical analysis</topic><topic>Matrices (mathematics)</topic><topic>Matrix algebra</topic><topic>Numeric Computing</topic><topic>Numerical Analysis</topic><topic>Original Paper</topic><topic>Quadratures</topic><topic>Robustness (mathematics)</topic><topic>Stieltjes integral</topic><topic>Theory of Computation</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Meurant, Gérard</creatorcontrib><creatorcontrib>Papež, Jan</creatorcontrib><creatorcontrib>Tichý, Petr</creatorcontrib><collection>CrossRef</collection><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science & Engineering Collection</collection><collection>ProQuest Central</collection><collection>Advanced Technologies & Aerospace Collection</collection><collection>ProQuest Central Essentials</collection><collection>AUTh Library subscriptions: ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central</collection><collection>ProQuest Central Student</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Computer Science Collection</collection><collection>Computer Science Database</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>ProQuest Advanced Technologies & Aerospace Collection</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>Engineering Collection</collection><jtitle>Numerical algorithms</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Meurant, Gérard</au><au>Papež, Jan</au><au>Tichý, Petr</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Accurate error estimation in CG</atitle><jtitle>Numerical algorithms</jtitle><stitle>Numer Algor</stitle><date>2021-11-01</date><risdate>2021</risdate><volume>88</volume><issue>3</issue><spage>1337</spage><epage>1359</epage><pages>1337-1359</pages><issn>1017-1398</issn><eissn>1572-9265</eissn><abstract>In practical computations, the (preconditioned) conjugate gradient (P)CG method is the iterative method of choice for solving systems of linear algebraic equations
A
x
=
b
with a real symmetric positive definite matrix
A
. During the iterations, it is important to monitor the quality of the approximate solution
x
k
so that the process could be stopped whenever
x
k
is accurate enough. One of the most relevant quantities for monitoring the quality of
x
k
is the squared
A
-norm of the error vector
x
−
x
k
. This quantity cannot be easily evaluated; however, it can be estimated. Many of the existing estimation techniques are inspired by the view of CG as a procedure for approximating a certain Riemann–Stieltjes integral. The most natural technique is based on the Gauss quadrature approximation and provides a lower bound on the quantity of interest. The bound can be cheaply evaluated using terms that have to be computed anyway in the forthcoming CG iterations. If the squared
A
-norm of the error vector decreases rapidly, then the lower bound represents a tight estimate. In this paper, we suggest a heuristic strategy aiming to answer the question of how many forthcoming CG iterations are needed to get an estimate with the prescribed accuracy. Numerical experiments demonstrate that the suggested strategy is efficient and robust.</abstract><cop>New York</cop><pub>Springer US</pub><doi>10.1007/s11075-021-01078-w</doi><tpages>23</tpages><orcidid>https://orcid.org/0000-0001-6008-4056</orcidid><orcidid>https://orcid.org/0000-0002-6036-3482</orcidid></addata></record> |
fulltext | fulltext |
identifier | ISSN: 1017-1398 |
ispartof | Numerical algorithms, 2021-11, Vol.88 (3), p.1337-1359 |
issn | 1017-1398 1572-9265 |
language | eng |
recordid | cdi_proquest_journals_2918610119 |
source | Springer Nature |
subjects | Accuracy Algebra Algorithms Approximation Computer Science Eigenvalues Errors Iterative methods Linear algebra Lower bounds Mathematical analysis Matrices (mathematics) Matrix algebra Numeric Computing Numerical Analysis Original Paper Quadratures Robustness (mathematics) Stieltjes integral Theory of Computation |
title | Accurate error estimation in CG |
url | http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-06T23%3A05%3A01IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Accurate%20error%20estimation%20in%20CG&rft.jtitle=Numerical%20algorithms&rft.au=Meurant,%20G%C3%A9rard&rft.date=2021-11-01&rft.volume=88&rft.issue=3&rft.spage=1337&rft.epage=1359&rft.pages=1337-1359&rft.issn=1017-1398&rft.eissn=1572-9265&rft_id=info:doi/10.1007/s11075-021-01078-w&rft_dat=%3Cproquest_cross%3E2918610119%3C/proquest_cross%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c319t-8dddeefa9b45263e928b1c8424abf6b702fc83e7a5b7b9c5325447134eb86b1f3%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=2918610119&rft_id=info:pmid/&rfr_iscdi=true |