1989
DOI: 10.2307/2008351
|View full text |Cite
|
Sign up to set email alerts
|

Improved Condition Number for Spectral Methods

Abstract: Abstract.For the known spectral methods (Galerkin, Tau, Collocation) the condition number behaves like 0(N4) (N: maximal degree of polynomials).We introduce a spectral method with an 0(N2) condition number. The advantages with respect to propagation of rounding errors and preconditioning are demonstrated.A direct solver for constant coefficient problems is given. Extensions to variable coefficient problems and first-order problems are discussed. Numerical results are presented, showing the effectiveness of our… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
30
0

Year Published

1997
1997
2014
2014

Publication Types

Select...
6

Relationship

0
6

Authors

Journals

citations
Cited by 18 publications
(30 citation statements)
references
References 9 publications
0
30
0
Order By: Relevance
“…The iteration (19) can be combined with multigrid, in which case it serves the role of a smoother, or accelerated by Krylov-based projection methods such as conjugate gradients (if M is SPD) or GMRES [36]. One also can use multigrid as the preconditioner for conjugate gradients (CG) or GMRES.…”
Section: Iterative Solution Methodsmentioning
confidence: 99%
See 1 more Smart Citation
“…The iteration (19) can be combined with multigrid, in which case it serves the role of a smoother, or accelerated by Krylov-based projection methods such as conjugate gradients (if M is SPD) or GMRES [36]. One also can use multigrid as the preconditioner for conjugate gradients (CG) or GMRES.…”
Section: Iterative Solution Methodsmentioning
confidence: 99%
“…Figure 3 illustrates the error behavior for three variants of the hybrid Schwarz scheme for a 2 × 2 array of elements with N = 16. Figure 3(a) shows the error after a single application of the smoothing step (19) with the (additive) Schwarz preconditioner (28) with minimal overlap as depicted in Fig. 2(b).…”
Section: Richardson Lcs Gmres Lcs/gmresmentioning
confidence: 99%
“…This leads in general to full matrices, which is numerically inefficient. The modification of the standard Chebyshev-Galerkin method is basically a combination of ideas posed by Heinrichs, [11,12], Shen, [29], and also Pop, [23]. Key feature in Galerkin methods is the use of shape and test functions that satisfy a priori the boundary conditions.…”
Section: A Appendix : a Modified Chebyshev-galerkin Methodsmentioning
confidence: 99%
“…The result is projected on a finite dimensional space in order to get a finite algebraic system. Let T k (ζ ) denote the Chebyshev polynomial of degree k. We define the shape functions ( [11,12])…”
Section: A Appendix : a Modified Chebyshev-galerkin Methodsmentioning
confidence: 99%
See 1 more Smart Citation