2019
DOI: 10.1155/2019/7527590
|View full text |Cite
|
Sign up to set email alerts
|

Nonrecursive Equivalent of the Conjugate Gradient Method without the Need to Restart

Abstract: A simple alternative to the conjugate gradient (CG) method is presented; this method is developed as a special case of the more general iterated Ritz method (IRM) for solving a system of linear equations. This novel algorithm is not based on conjugacy, i.e. it is not necessary to maintain overall orthogonalities between various vectors from distant steps. This method is more stable than CG, and restarting techniques are not required. As in CG, only one matrix-vector multiplication is required per step with app… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
5
0
4

Year Published

2019
2019
2024
2024

Publication Types

Select...
5
2

Relationship

2
5

Authors

Journals

citations
Cited by 7 publications
(9 citation statements)
references
References 13 publications
0
5
0
4
Order By: Relevance
“…For such a small example, an exact check of the algorithm stability is possible. Because the number of operations is small and the rounding error is negligible, both methods are intentionally perturbed by δ = 1 [4], added (for example) to the seventh component of p (1) (at the end of the first step): p 7,(1) ← p 7,(1) + δ…”
Section: Check Of the Algorithm Stabilitymentioning
confidence: 99%
See 1 more Smart Citation
“…For such a small example, an exact check of the algorithm stability is possible. Because the number of operations is small and the rounding error is negligible, both methods are intentionally perturbed by δ = 1 [4], added (for example) to the seventh component of p (1) (at the end of the first step): p 7,(1) ← p 7,(1) + δ…”
Section: Check Of the Algorithm Stabilitymentioning
confidence: 99%
“…As a consequence, restarting of IRM-CG is not needed. Further, preconditioning-like techniques [11] can be adopted easily [4]. One more advantage of this formulation is natural adoption of the relaxation factor.…”
mentioning
confidence: 99%
“…Bez obzira na k(A), tijek točnoga proračuna podudara se za oba postupka i završava nakon 188 koraka. Riječ je o broju nepoznanica umanjenom za četiri, koliko je "neaktivnih" l j , onih kojima pripada v j • b jednak (u egzaktnoj aritmetici cjelobrojnoj) nuli [25,26]. Ili drugačije, potreban broj koraka jednak je broju "aktivnih" vlastitih vrijednosti, za koje je v j • b ≠ 0.…”
Section: Rezultati Proračuna Praktičnih Modelaunclassified
“…Primjena biblioteke Eigen omogućava gotovo doslovan prijepis pseudokôda (iz, primjerice, [4]) u programski kôd:…”
Section: Realizacija U Programskom Jeziku C++unclassified