Pith. sign in

REVIEW 3 major objections 7 minor 83 references

Spacetime Wavelet Method for Linear Boundary-Value Problems in Sylvester Matrix Equation Form

T0 review · 3 major / 7 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read A spacetime wavelet discretization turns linear initial-boundary value problems into Sylvester matrix equations solved by Global GMRES with recursive initial guesses, achieving predicted high-order convergence at lower memory cost than…

desk verdict The accuracy results are solid and the recursive initialization is a real small contribution, but the paper's own scaling fits undercut the headline performance claim, so the comparison needs rework before acceptance. read the letter →

arxiv 2509.02720 v1 pith:CA2VDF3C submitted 2025-09-02 math.NA cs.NAphysics.comp-ph

classification math.NAcs.NAphysics.comp-ph MSC 65T6065F1065M7065M12
keywords spacetimewaveletmethodSylvestermatrixequationGlobalGMRESlinearinitial-boundaryvalueproblemsDeslauriers-Dubucwaveletsderivativeconnectioncoefficientsrecursiveinitial-guessalgorithmhigh-orderconvergence
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes solving linear initial-boundary value problems by discretizing both space and time simultaneously with Deslauriers-Dubuc wavelets, so the PDE becomes a Sylvester matrix equation AX + XB = C rather than a single large Kronecker-product system. To solve it, the authors use Global GMRES and feed it recursively generated wavelet-synthesized initial guesses. They report high-order convergence rates, matching wavelet theory, for both solutions and derivatives on diffusion and convection-diffusion test problems, and report that the Sylvester/Global GMRES route is consistently faster than restarted GMRES on the Kronecker formulation. A sympathetic reader would care because spacetime methods avoid timestep stability limits and offer a priori error control, while the Sylvester form avoids the memory blow-up of vectorized systems.

What carries the argument

The central objects are the Deslauriers-Dubuc wavelet basis and its derivative connection coefficients $\Gamma$, which discretize derivatives directly in space and time. Applying these operators to a linear PDE puts the discrete system in Sylvester form $AX+XB=C$; semi-orthogonal permutation matrices $P_x$, $P_t$ enforce initial and boundary conditions while keeping the reduced system in the same Sylvester structure. Global GMRES with a Modified Global Arnoldi process, which generates orthogonal matrices rather than vectors, solves the equation, and the paper's recursive algorithm synthesizes coarser-grid solutions into informed initial guesses for finer grids.

What would settle it

Run the method on an advection-dominated linear problem, such as convection-diffusion with $c \gg \nu$, or on a hyperbolic or oscillatory linear PDE, at successive resolution levels while tracking Global GMRES iterations and cumulative wall time against the zero-initial-guess baseline. If iteration counts grow superlinearly with level $j$, or the recursive solver becomes slower than solving directly at $j_{\max}$, the efficiency claim is refuted.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that a fully wavelet-based spacetime discretization of linear PDEs produces Sylvester equations that can be solved accurately and efficiently: numerical solutions and derivative approximations converge at the rates predicted by wavelet theory, with user-prescribed error estimates, and the Sylvester form can be solved with Global GMRES rather than the Kronecker-product vectorized system. The recursive wavelet algorithm synthesizes the solution at one resolution level into the initial guess for the next level, reducing the number of Global GMRES iterations and cumulative time to solution compared with starting from zero at the target level. The paper further claims that this Sylvester route outperforms the standard Kronecker-product formulation, which requires storing orders of magnitude more nonzero entries.

Load-bearing premise

The load-bearing premise is that Global GMRES converges quickly enough on the Sylvester systems produced by the spacetime wavelet discretization, without preconditioning and with the fixed restart rule $m=30(j+1)$, across the class of linear PDEs the method claims to target.

Editorial extensions

If this is right

  • The discretization provides a built-in a priori error estimate: truncation error is $O(\Delta q^{p-\alpha})$, dictated by the weakest derivative order in the PDE, so users can predict convergence order before solving.
  • The full spacetime formulation avoids timestep stability restrictions, sidestepping the explicit/implicit timestep trade-off of conventional semi-discretizations.
  • For the diffusion test at level $j=5$, the Sylvester matrices together store roughly 4,449 nonzeros while the equivalent Kronecker product matrix stores 1,267,589 nonzeros, giving a concrete memory advantage.
  • The recursive wavelet-synthesized initial-guess procedure reduces Global GMRES iteration counts and cumulative wall time relative to a zero initial guess at every tested resolution level.
  • Boundary and initial condition enforcement via permutation matrices preserves well-posedness, with uniqueness guaranteed when the spectra of $A$ and $-B$ are disjoint.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the convergence behavior is generic, the method could be applied to stiff regimes such as small-diffusion, strong-advection linear problems where explicit timestepping fails, since the spacetime formulation has no visible CFL restriction.
  • A natural next step, absent from the paper, would be a convergence theory for Global GMRES on this class of Sylvester systems linking the spectra of $A$ and $B$ to iteration bounds; the paper only provides empirical spectra.
  • The recursive initial-guess idea should transfer to nonlinear problems, where prior-level solutions could initialize Newton-like iterations for both the state and the Jacobian.
  • The K-matrix spectra for the convection-diffusion case show a vertical band absent in the diffusion case, suggesting that advection-dominated limits may hit a performance cliff requiring different restart values or preconditioning; this is a testable boundary of the efficiency claim.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 7 minor

Summary. The paper proposes a spacetime wavelet discretization for linear initial-boundary value problems, based on Deslauriers-Dubuc wavelets and dense wavelet derivative operators. The discretization produces a Sylvester matrix equation, solved here with global GMRES (Gl-GMRES) using a modified global Arnoldi process; boundary and initial conditions are enforced with permutation matrices, and a recursive wavelet-synthesis algorithm supplies initial guesses for successive resolution levels. Verification studies on one-dimensional linear diffusion and convection-diffusion equations with manufactured solutions report high-order convergence rates for both solution and derivative approximations, consistent with the a priori estimates cited from wavelet theory. The paper also claims that the Gl-GMRES Sylvester-form solver outperforms restarted GMRES applied to the Kronecker-product form, and that the recursive initial-guess technique reduces iteration counts and wall time.

Significance. If fully supported, the paper would be a useful practical contribution: it demonstrates an alternative to Kronecker-form spacetime discretizations, where the Sylvester form substantially reduces the number of stored nonzeros, and it provides a simple recursive initialization that accelerates Gl-GMRES. The convergence portion is credible and well executed: manufactured solutions are used, the observed rates match the predicted p-alpha orders, derivative superconvergence is documented, and the algorithms are given in sufficient detail to be reproduced. The main weakness is that the headline performance claim rests on a narrow, confounded comparison in Section 4.3; the authors' own complexity fits suggest a possible crossover at larger problem sizes, so the claim that the Sylvester formulation is 'consistently faster' is not yet established. The convergence results are sound, but the efficiency claim needs additional work before the paper can be accepted as a demonstration of the Sylvester form's superiority.

major comments (3)
  1. [§4.3, Fig. 8] The claim that Gl-GMRES on the Sylvester form is 'consistently faster' is not supported beyond the tested range. The text reports fitted scalings of approximately O(N^2.5) for Gl-GMRES and slightly below O(N^2) for restarted GMRES on the Kronecker system. If these fits describe the asymptotic behavior, the advantage shown in Fig. 8a must eventually reverse at sufficiently large N. Please provide results at larger N, or present the fits with confidence intervals and an explicit crossover estimate; otherwise the claim should be restricted to the tested problem sizes.
  2. [§4.3] The performance comparison changes two variables simultaneously: the algebraic formulation (Sylvester versus Kronecker) and the iterative solver (Gl-GMRES versus restarted GMRES). This does not isolate the benefit of the Sylvester form. To substantiate the abstract and conclusion claims that the Sylvester formulation outperforms the Kronecker-product approach, the comparison should either hold the solver fixed across formulations, include a solver appropriate to the vectorized Kronecker system (for example a preconditioned GMRES variant), or be explicitly framed as a comparison of the authors' combined solver choice, not of the algebraic form alone.
  3. [§3.3 and §4.3] The efficiency and recursive-initialization results depend on the heuristically chosen restart parameter m = 30(j+1) and on the tolerances tol_H = 10^{-8} and residual tolerance 10^{-8}. The paper states that similar results are obtained for other values of m, but gives no sensitivity data and no convergence analysis for the restarted Gl-GMRES iteration. Since the central performance claim relies on this restart rule, please include a sensitivity study (e.g., iteration counts and times for several m values at each level) or a theoretical justification for the rule; otherwise the generality of the conclusions is unclear.
minor comments (7)
  1. [§4.4, Fig. 9 and Fig. 10] The text states that 'Fig. 9 shows that the number of iterations is drastically attenuated,' but Fig. 9 shows relative time; the iteration counts appear in Fig. 10. Please correct the citation.
  2. [Fig. 7 caption] The caption for Fig. 7 says the convection-diffusion problem is at j = 2 with px = 6, pt = 4, while the surrounding text states that px = pt = 8 is used for this example. The caption and text should be made consistent.
  3. [Algorithms 1 and 2] If the modified global Arnoldi loop breaks early because H(z+1,z) < tol_H, the least-squares problem in Algorithm 1 is written for y in R^m and H of size (m+1) x m, while the actual H has smaller dimensions. Please clarify how the reduced H and V are handled during a restart cycle.
  4. [Tables 1 and 2 and Figs. 3 and 6] The level ranges in the captions are inconsistent: Table 1 and Fig. 3 use j = 3, 4, 5, while Table 2 and Fig. 6 use different ranges (j = 2, 3, 4 in the text). Please harmonize the level ranges in the captions and bodies.
  5. [Abstract and §2.3] The abstract promises 'user-prescribed error estimates,' but the paper presents a priori asymptotic rates of the form O(delta q^{p-alpha}), not user-specified tolerance-controlled error estimates. Please rephrase to avoid overstating what is demonstrated.
  6. [Fig. 8b] The vertical axis of Fig. 8b is labeled 'number of operations,' but the text does not define how this quantity is computed. Please specify whether it counts matrix-vector products, flops, or some other measure, and how the fits were obtained.
  7. [Title, page 1] The title contains a spacing artifact, 'Proble ms'; this should be corrected in the final version.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the accuracy and convergence claims are verified against manufactured solutions and external wavelet theory, and the efficiency claims are empirical comparisons rather than construction-equivalent predictions.

full rationale

The paper's derivation chain is self-contained. The spacetime wavelet discretization (Eqs. 3-8) is built from standard Deslauriers-Dubuc multiresolution analysis and connection coefficients from external wavelet literature; the convergence estimate in Eq. (9) is cited to [28,45,56] and is not fitted to the numerical results in Section 4. Accuracy is checked against manufactured analytical solutions (Eqs. 18 and 20), so the reported high-order rates are independent of the method's own assumptions. Boundary-condition enforcement via permutation matrices follows [44], and the Sylvester solution theory (unique solution when spectra of A and -B are disjoint) is cited to [23]. The Gl-GMRES solver and Modified Global Arnoldi algorithm are presented as Algorithms 1-2 from the established literature [53]. Self-citations to [20,45,46,47] identify the MRWT software and the prior nonlinear spacetime discretization on which this linear formulation builds, but the linear reduction to AX+XB=C, the recursive-initial-guess algorithm, and the comparisons in Section 4.3 rest on the paper's own constructed systems and measurements, not on those citations. The restart parameter m=30(j+1) and absence of preconditioning are heuristic choices affecting performance generalization, and the §4.3 complexity fits (Gl-GMRES near O(N^2.5), restarted GMRES below O(N^2)) may indicate a crossover at larger N; these are robustness/correctness caveats, not circular reductions. No equation is defined in terms of a result it is used to prove, and no fitted parameter is renamed as a prediction.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

No new physical entities are introduced. The paper relies on standard wavelet approximation theory and published iterative solvers; its free parameters are algorithmic choices (tolerances, restart size, basis orders) rather than fitted physical constants.

free parameters (5)
  • restart parameter m = 30(j+1)
    Hand-chosen in Section 4 to balance memory and convergence; not derived from theory.
  • Gl-GMRES residual tolerance = 1e-8
    Stopping criterion set in Section 4.
  • Arnoldi break tolerance tol_H = 1e-8
    Set in Section 4.
  • basis orders (px, pt) = (6,4) diffusion; (8,8) convection-diffusion
    User-selected in Sections 4.1 and 4.2 to control accuracy and cost; not fitted to observed convergence.
  • recursive starting level j = 0
    Start of recursion in Algorithm 3; determines which coarse solves contribute to the cumulative time in Fig. 9.
assumptions (5)
  • standard math Dense wavelet derivative approximation error obeys the bound in Eq. (9), with order p_i - alpha.
    Cited from [28,45,56]; used to predict convergence orders in Sections 2.3 and 4.
  • standard math Deslauriers-Dubuc wavelets have sufficient differentiability for first and second derivatives.
    Invoked when constructing connection coefficients in Section 2.3; based on [25,65].
  • standard math The Sylvester equation AX+XB=C has a unique solution when the spectra of A and -B are disjoint.
    Stated in Section 1 with citation [23]; guarantees well-posedness after BC enforcement.
  • domain assumption Permutation matrix enforcement of Dirichlet conditions yields a well-posed reduced system.
    Adopted from [44] in Section 3.2; no proof or conditioning analysis is given for this discretization.
  • domain assumption Global GMRES converges for the nonsymmetric Sylvester systems arising here.
    The method is assumed to converge within the restart budget; demonstrated only empirically in Section 4.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Spacetime Wavelet Method for Linear Boundary-Value Problems in Sylvester Matrix Equation Form." pith.science (2026). https://pith.science/paper/CA2VDF3C

@misc{pith2026250902720,
  author       = {Pith},
  title        = {Pith review of: Spacetime Wavelet Method for Linear Boundary-Value Problems in Sylvester Matrix Equation Form},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CA2VDF3C}},
  note         = {Machine review of arXiv:2509.02720}
}
read the original abstract

We present a high-order spacetime numerical method for discretizing and solving linear initial-boundary value problems using wavelet-based techniques with user-prescribed error estimates. The spacetime wavelet discretization yields a system of algebraic equations resulting in a Sylvester matrix equation. We solve this system with a Global Generalized Minimal Residual (GMRES) method in conjunction with a wavelet-based recursive algorithm to improve convergence. We perform rigorous verification studies using linear partial differential equations (PDEs) with both convective and diffusive terms. The results of these simulations show the high-order convergence rates for the solution and derivative approximations predicted by wavelet theory. We demonstrate the utility of solving the Sylvester equation through comparisons to the commonly-used Kronecker product formulation. We show that our recursive wavelet-based algorithm that generates initial guesses for the iterative Global GMRES method improves the performance of the solver.

Figures

Figures reproduced from arXiv: 2509.02720 by the authors.

Figure 1
Figure 1. Sparsity patterns of the A, B, (Eq. (12)) and K (Eq. (13)) matrices obtained by discretizing the linear diffusion equation, Eq. 11. We see that all matrices are square and have a sparse banded structure. However the K matrix shown in Fig. 1c is significantly larger than the A and B matrices (Figs. 1a, 1b) and contains more nonzero entries. The A and B matrices at j = 2 contain 403 and 126 nonzeros, respectively, whi… view at source ↗
Figure 2
Figure 2. Spacetime solution for the linear diffusion proble [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Solution and derivative convergence for the linea [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: shows the eigenvalue spectra of the A, B, and K matrices. (a) Eigenvalue spectrum of the A matrix, Eq. (12). (b) Eigenvalue spectrum of the B matrix, Eq. (12). (c) Eigenvalue spectrum of the K matrix, Eq. (13) [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Spacetime solution for the convection-diffusion p [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Solution and derivative convergence for the conve [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: Eigenvalue spectra of the A, B, and K matrices obtained by discretizing the convection-diffusion equation (Eq. 22) at j = 2 with px = 6, pt = 4. We see that the addition of the convection term results in complex eigenvalues in the A matrix (Fig. 7a) and the appearance …
Figure 8
Figure 8. Figure 8: Performance comparisons of the Gl-GMRES and stand [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Relative time comparison of the Gl-GMRES solver wi [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]
Figure 10
Figure 10. Figure 10: Number of Gl-GMRES iterations required to reach i [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

83 extracted references · 61 canonical work pages

  1. [1]

    Hawker, Robert B

    Reza Abedi, Morgan A. Hawker, Robert B. Haber, and Karel M atouˇ s. An adaptive spacetime dis- continuous Galerkin method for cohesive models of elastody namic fracture. International journal for numerical methods in engineering , 81(10):1207–1241, 2010

  2. [2]

    Alam, Nicholas K.-R

    Jahrul M. Alam, Nicholas K.-R. Kevlahan, and Oleg V. Vasi lyev. Simultaneous space–time adaptive wavelet solution of nonlinear parabolic differential equat ions. Journal of Computational Physics , 214 (2):829–857, 2006

  3. [3]

    Implicit high-order time marching schemes for the linearized Euler equations

    George Arabatzis, Panagiotis Vavilis, Ioannis Toulopo ulos, and John A Ekaterinaris. Implicit high-order time marching schemes for the linearized Euler equations. AIAA journal , 45(8):1819–1826, 2007

  4. [4]

    A wavelet based space-time adaptive numerical method for partial differential equations

    Emmanual Bacry, Stephane Mallat, and George Papanicola ou. A wavelet based space-time adaptive numerical method for partial differential equations. ESAIM: Mathematical Modelling and Numerical Analysis, 26(7):793–834, 1992

  5. [5]

    A technique for accelerating the con- vergence of restarted GMRES

    Allison H Baker, Elizabeth R Jessup, and Thomas Manteuffe l. A technique for accelerating the con- vergence of restarted GMRES. SIAM Journal on Matrix Analysis and Applications , 26(4):962–984, 2005

  6. [6]

    Trends in large space structure control theo ry: fondest hopes, wildest dreams

    Mark Balas. Trends in large space structure control theo ry: fondest hopes, wildest dreams. IEEE Transactions on automatic control , 27(3):522–535, 2003

  7. [7]

    Bartels and George W

    Richard H. Bartels and George W. Stewart. Solution of the matrix equation AX + XB = C [F4]. Communications of the ACM , 15(9):820–826, 1972

  8. [8]

    Beattie, Mark Embree, and Danny C

    Christopher A. Beattie, Mark Embree, and Danny C. Sorens en. Convergence of polynomial restart Krylov methods for eigenvalue computations. SIAM review , 47(3):492–515, 2005

Show all 83 references
  1. [9]

    Asynchronous multi-do main variational integrators for nonlinear hyperelastic solids

    Michal Beneˇ s and Karel Matouˇ s. Asynchronous multi-do main variational integrators for nonlinear hyperelastic solids. Computer Methods in Applied Mechanics and Engineering , 199(29-32):1992–2013, 2010

  2. [10]

    S Bertoluzza and G. Naldi. A wavelet collocation method for the numerical solution of partial differential equations. Applied and Computational Harmonic Analysis , 3(1):1–9, 1996

  3. [11]

    On the adaptive num erical solution of nonlinear partial differ- ential equations in wavelet bases

    Gregory Beylkin and James M/ Keiser. On the adaptive num erical solution of nonlinear partial differ- ential equations in wavelet bases. Journal of computational physics , 132(2):233–259, 1997

  4. [12]

    Convex constrained optimization for large-scale generalized sylvester equations

    A Bouhamidi, K Jbilou, and M Raydan. Convex constrained optimization for large-scale generalized sylvester equations. Computational Optimization and Applications , 48(2):233–253, 2011

  5. [13]

    A note on the numerical approximate solutions for generalized Sylvester matrix equations with applications

    Abderrahman Bouhamidi and Khalide Jbilou. A note on the numerical approximate solutions for generalized Sylvester matrix equations with applications . Applied Mathematics and Computation , 206 (2):687–694, 2008

  6. [14]

    A precon- ditioned block Arnoldi method for large Sylvester matrix eq uations

    Abderrahman Bouhamidi, Mustapha Hached, Mohammed Hey ouni, and Khalide Jbilou. A precon- ditioned block Arnoldi method for large Sylvester matrix eq uations. Numerical Linear Algebra with Applications, 20(2):208–219, 2013

  7. [15]

    Semi-supervised multi-label learning by solving a Sylvester equation

    Gang Chen, Yangqiu Song, Fei Wang, and Changshui Zhang. Semi-supervised multi-label learning by solving a Sylvester equation. In Proceedings of the 2008 SIAM International Conference on Da ta Mining, pages 410–419. SIAM, 2008

  8. [16]

    Matrix preconditioning techniques and applications

    Ke Chen. Matrix preconditioning techniques and applications . Number 19. Cambridge University Press, 2005

  9. [17]

    Meshless methods in solid mechanics , volume 9

    Youping Chen, James D Lee, and Azim Eskandarian. Meshless methods in solid mechanics , volume 9. Springer, 2006. 15

  10. [18]

    A projection method and Krone cker product preconditioner for solving Sylvester tensor equations

    Zhen Chen and LinZhang Lu. A projection method and Krone cker product preconditioner for solving Sylvester tensor equations. Science China Mathematics , 55:1281–1292, 2012

  11. [19]

    O n the ⋆-Sylvester equationAX±X ⋆B⋆ = C

    Chun-Yueh Chiang, Eric King-Wah Chu, and Wen-Wei Lin. O n the ⋆-Sylvester equationAX±X ⋆B⋆ = C. Applied Mathematics and Computation , 218(17):8393–8407, 2012

  12. [20]

    Cochran and Karel Matouˇ s

    Cody D. Cochran and Karel Matouˇ s. Spacetime wavelet me thod for the solution of nonlinear partial differential equations. International Journal for Numerical Methods in Engineerin g, 126(13):e70076, 2025

  13. [21]

    ¨Uber die partiellen differenzengleichungen der mathematischen physik

    Richard Courant, Kurt Friedrichs, and Hans Lewy. ¨Uber die partiellen differenzengleichungen der mathematischen physik. Mathematische annalen , 100(1):32–74, 1928

  14. [22]

    A practical method for numerical evaluation of solutions of par- tial differential equations of the heat-conduction type

    John Crank and Phyllis Nicolson. A practical method for numerical evaluation of solutions of par- tial differential equations of the heat-conduction type. In Mathematical proceedings of the Cambridge philosophical society, volume 43, pages 50–67. Cambridge University Press, 1947

  15. [23]

    Arnoldi methods for la rge Sylvester-like observer matrix equations, and an associated algorithm for partial spectrum assignmen t

    Biswa Nath Datta and Youcef Saad. Arnoldi methods for la rge Sylvester-like observer matrix equations, and an associated algorithm for partial spectrum assignmen t. Linear Algebra and its Applications , 154: 225–244, 1991

  16. [24]

    Ten lectures on wavelets

    Ingrid Daubechies. Ten lectures on wavelets . SIAM, 1992

  17. [25]

    De Villiers, Karin M

    Johan M. De Villiers, Karin M. Goosen, and Ben M. Herbst. Dubuc–Deslauriers subdivision for finite sequences and interpolation wavelets on an interval. SIAM Journal on Mathematical analysis , 35(2): 423–452, 2003

  18. [26]

    David L. Donoho. Interpolating wavelet transforms. Preprint, Department of Statistics, Stanford Uni- versity, 2(3):1–54, 1992

  19. [27]

    F ASTEN: Fast Sylvester equation solver for graph mining

    Boxin Du and Hanghang Tong. F ASTEN: Fast Sylvester equation solver for graph mining. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledg e Discovery & Data Mining , pages 1339–1347, 2018

  20. [28]

    Dubos and N

    T. Dubos and N. K.-R. Kevlahan. A conservative adaptive wavelet method for the shallow-water equations on staggered grids. Quarterly Journal of the Royal Meteorological Society , 139(677):1997– 2020, 2013

  21. [29]

    El Guennouni, Khalide Jbilou, and A.J

    A. El Guennouni, Khalide Jbilou, and A.J. Riquet. Block Krylov subspace methods for solving large Sylvester equations. Numerical Algorithms, 29:75–96, 2002

  22. [30]

    Elleithy, Husain J

    Wael M. Elleithy, Husain J. Al-Gahtani, and M. El-Gebei ly. Iterative coupling of BE and FE methods in electrostatics. Engineering Analysis with Boundary Elements , 25:685–695, 2001

  23. [31]

    The tortoise and the hare restart GMRES

    Mark Embree. The tortoise and the hare restart GMRES. SIAM review, 45(2):259–266, 2003

  24. [32]

    Martin J. Gander. 50 years of time parallel time integra tion. In Multiple shooting and time domain decomposition methods, pages 69–113. Springer, 2015

  25. [33]

    Gander and Stefan Vandewalle

    Martin J. Gander and Stefan Vandewalle. Analysis of the parareal time-parallel time-integration method. SIAM Journal on Scientific Computing , 29(2):556–578, 2007

  26. [34]

    Van Veen, and Susan C

    Fuqiang Gao, Barry D. Van Veen, and Susan C. Hagness. Sen sitivity of the distorted born iterative method to the initial guess in microwave breast imaging. IEEE Transactions on Antennas and Propa- gation, 63(8):3540–3547, 2015

  27. [35]

    Gardiner, Alan J

    Judith D. Gardiner, Alan J. Laub, James J. Amato, and Cle ve B. Moler. Solution of the Sylvester matrix equation AXB T +CXD T =E. ACM Transactions on Mathematical Software (TOMS) , 18(2): 223–231, 1992. 16

  28. [36]

    A new class of high ly accurate solvers for ordinary differential equations

    Andreas Glaser and Vladimir Rokhlin. A new class of high ly accurate solvers for ordinary differential equations. Journal of Scientific Computing , 38(3):368–399, 2009

  29. [37]

    Finite differenc e method for numerical computation of dis- continuous solutions of the equations of fluid dynamics

    Sergei K Godunov and Ihor Bohachevsky. Finite differenc e method for numerical computation of dis- continuous solutions of the equations of fluid dynamics. Matematiˇ ceskij sbornik, 47(3):271–306, 1959

  30. [38]

    Wavelets and their application for t he solution of partial differential equations in physics

    Stefan Goedecker. Wavelets and their application for t he solution of partial differential equations in physics. Max-Planck Institute for Solid State Research, Stuttgart, Germany, 2009

  31. [39]

    Sˆ onia M. Gomes. Convergence estimates for the wavelet -Galerkin method: superconvergence at the node points. Advances in Computational Mathematics , 4:261–282, 1995

  32. [40]

    An essay on the application of mathematical analysis to the t heories of electricity and magnetism, volume 3

    George Green. An essay on the application of mathematical analysis to the t heories of electricity and magnetism, volume 3. Mayer & M¨ uller, 1889

  33. [41]

    Eigen v3

    Ga¨ el Guennebaud, Beno ˆ ıt Jacob, et al. Eigen v3. http: //eigen.tuxfamily.org, 2010

  34. [42]

    Gunzburger and Angela Kunoth

    Max D. Gunzburger and Angela Kunoth. Space-time adapti ve wavelet methods for optimal control problems constrained by parabolic evolution equations. SIAM journal on control and optimization , 49 (3):1150–1170, 2011

  35. [43]

    Th´ eorie des ´ equations aux d´ eriv´ees partielles lin´ eaires hyperboliques et du probl` eme de Cauchy

    Jacques Hadamard. Th´ eorie des ´ equations aux d´ eriv´ees partielles lin´ eaires hyperboliques et du probl` eme de Cauchy. Acta mathematica, 31(1):333–380, 1908

  36. [44]

    Sylvester equations a nd the numerical solution of partial fractional differential equations

    Matthew Harker and Paul O’Leary. Sylvester equations a nd the numerical solution of partial fractional differential equations. Journal of Computational Physics , 293:370–384, 2015

  37. [45]

    Adapti ve wavelet algorithm for solving nonlinear initial–boundary value problems with error control

    Cale Harnish, Karel Matous, and Daniel Livescu. Adapti ve wavelet algorithm for solving nonlinear initial–boundary value problems with error control. International Journal for Multiscale Computational Engineering, 16(1), 2018

  38. [46]

    A multiresolution adaptive wavelet method for nonlinear partial differential equations

    Cale Harnish, Luke Dalessandro, Karel Matous, and Dani el Livescu. A multiresolution adaptive wavelet method for nonlinear partial differential equations. International Journal for Multiscale Computational Engineering, 19(2), 2021

  39. [47]

    An adaptive wavelet method for nonlinear partial differential equations with applicat ions to dynamic damage modeling

    Cale Harnish, Luke Dalessandro, Karel Matouˇ s, and Dan iel Livescu. An adaptive wavelet method for nonlinear partial differential equations with applicat ions to dynamic damage modeling. Journal of Computational Physics , 479:112002, 2023

  40. [48]

    Scientific computing: an introductory survey, revised seco nd edition

    Michael T Heath. Scientific computing: an introductory survey, revised seco nd edition. SIAM, 2018

  41. [49]

    Hughes and James R

    Thomas J.R. Hughes and James R. Stewart. A space-time fo rmulation for multiscale phenomena. Journal of Computational and Applied Mathematics , 74(1-2):217–229, 1996

  42. [50]

    Hulbert and Thomas J.R

    Gregory M. Hulbert and Thomas J.R. Hughes. Space-time fi nite element methods for second-order hyperbolic equations. Computer methods in applied mechanics and engineering , 84(3):327–348, 1990

  43. [51]

    Implicitly re started Krylov subspace methods for stable partial realizations

    Imad M Jaimoukha and Ebrahim M Kasenally. Implicitly re started Krylov subspace methods for stable partial realizations. SIAM Journal on Matrix Analysis and Applications , 18(3):633–652, 1997

  44. [52]

    On the wavelet based differentiation ma trix

    Leland Jameson. On the wavelet based differentiation ma trix. Journal of Scientific Computing , 8(3): 267–305, 1993

  45. [53]

    Global FOM and GMRES algorithms for matrix equations

    Khalide Jbilou, Abderrahim Messaoudi, and Hassane Sad ok. Global FOM and GMRES algorithms for matrix equations. Applied Numerical Mathematics , 31(1):49–63, 1999

  46. [54]

    Implicitly restarted global FOM and GMRES f or nonsymmetric matrix equations and Sylvester equations

    Yi-Qin Lin. Implicitly restarted global FOM and GMRES f or nonsymmetric matrix equations and Sylvester equations. Applied mathematics and computation , 167(2):1004–1025, 2005. 17

  47. [55]

    Comparative performance of exponential, implicit, and explicit integrators for stiff systems of ODEs

    John Loffeld and Mayya Tokman. Comparative performance of exponential, implicit, and explicit integrators for stiff systems of ODEs. Journal of Computational and Applied Mathematics , 241:45–67, 2013

  48. [56]

    Kent McCormick and Raymond O. Wells. Wavelet calculus a nd finite difference operators. Mathematics of Computation , 63(207):155–173, 1994

  49. [57]

    Merkle and Yun-Ho Choi

    Charles L. Merkle and Yun-Ho Choi. Computation of low-s peed compressible flows with time-marching procedures. International Journal for Numerical Methods in Engineerin g, 25(2):293–311, 1988

  50. [58]

    Fundamentals of engineering numerical analysis

    Parviz Moin. Fundamentals of engineering numerical analysis . Cambridge University Press, 2010

  51. [59]

    Naudet and Matthew J

    Charles J. Naudet and Matthew J. Zahr. A space-time high -order implicit shock tracking method for shock-dominated unsteady flows. Journal of Computational Physics , 501:112792, 2024

  52. [60]

    Nazareth

    John L. Nazareth. Conjugate gradient method. Wiley Interdisciplinary Reviews: Computational Statis- tics, 1(3):348–353, 2009

  53. [61]

    Zikoski, and Temistocle Gr enga

    Samuel Paolucci, Zachary J. Zikoski, and Temistocle Gr enga. W AMR: An adaptive wavelet method for the simulation of compressible reacting flow. Part II. The pa rallel algorithm. Journal of Computational Physics, 272:842–864, 2014

  54. [62]

    Zikoski, and Damrongsak Wi rasaet

    Samuel Paolucci, Zachary J. Zikoski, and Damrongsak Wi rasaet. W AMR: an adaptive wavelet method for the simulation of compressible reacting flow. Part I. Acc uracy and efficiency of algorithm. Journal of Computational Physics , 272:814–841, 2014

  55. [63]

    A sp ace–time discontinuous Galerkin method for the solution of the wave equation in the time domain

    Steffen Petersen, Charbel Farhat, and Radek Tezaur. A sp ace–time discontinuous Galerkin method for the solution of the wave equation in the time domain. International Journal for Numerical Methods in Engineering, 78(3):275–295, 2009

  56. [64]

    Computationally efficient multi-time-step method for partitioned time integration of highly nonlinea r structural dynamics

    Arun Prakash, Ertugrul Taciroglu, and Keith D Hjelmsta d. Computationally efficient multi-time-step method for partitioned time integration of highly nonlinea r structural dynamics. Computers & Struc- tures, 133:51–63, 2014

  57. [65]

    Simple regularity criteria for subdivi sion schemes

    Olivier Rioul. Simple regularity criteria for subdivi sion schemes. SIAM Journal on Mathematical Analysis, 23(6):1544–1576, 1992

  58. [66]

    A convergence analysis of GMRES and FOM methods for Sylvester equations

    Micka¨ el Robb´ e and Miloud Sadkane. A convergence analysis of GMRES and FOM methods for Sylvester equations. Numerical Algorithms, 30:71–89, 2002

  59. [67]

    Youcef Saad and Martin H. Schultz. GMRES: A generalized minimal residual algorithm for solving nonsymmetric linear systems. SIAM Journal on scientific and statistical computing , 7(3):856–869, 1986

  60. [68]

    Iterative methods for sparse linear systems

    Yousef Saad. Iterative methods for sparse linear systems . SIAM, 2003

  61. [69]

    Iterative methods for linear systems of eq uations: A brief historical journey

    Yousef Saad. Iterative methods for linear systems of eq uations: A brief historical journey. arXiv preprint arXiv:1908.01083, 2019

  62. [70]

    Block-Arnoldi and Davidson methods fo r unsymmetric large eigenvalue problems

    Miloud Sadkane. Block-Arnoldi and Davidson methods fo r unsymmetric large eigenvalue problems. Numerische Mathematik , 64:195–211, 1993

  63. [71]

    Code verification by th e method of manufactured solutions

    Kambiz Salari and Patrick Knupp. Code verification by th e method of manufactured solutions. Technical report, Sandia National Lab.(SNL-NM), Albuquerque, NM (Un ited States); Sandia . . . , 2000

  64. [72]

    Space-time adapti ve wavelet methods for parabolic evolution problems

    Christoph Schwab and Rob Stevenson. Space-time adapti ve wavelet methods for parabolic evolution problems. Mathematics of Computation , 78(267):1293–1318, 2009

  65. [73]

    On the convergence of restarted Kry lov subspace methods

    Valeria Simoncini. On the convergence of restarted Kry lov subspace methods. SIAM Journal on Matrix Analysis and Applications , 22(2):430–452, 2000. 18

  66. [74]

    Asynchronous space–tim e algorithm based on a domain decomposition method for structural dynamics problems on non-matching me shes

    Waad Subber and Karel Matouˇ s. Asynchronous space–tim e algorithm based on a domain decomposition method for structural dynamics problems on non-matching me shes. Computational Mechanics , 57(2): 211–235, 2016

  67. [75]

    Asynchronous space–tim e domain decomposition method with local- ized uncertainty quantification

    Waad Subber and Karel Matouˇ s. Asynchronous space–tim e domain decomposition method with local- ized uncertainty quantification. Computer Methods in Applied Mechanics and Engineering , 325:369–394, 2017

  68. [76]

    Tezduyar

    Kenji Takizawa and Tayfun E. Tezduyar. Space-time fluid -structure interaction methods. Mathematical Models and Methods in Applied Sciences , 22, 2012

  69. [77]

    Choice o f initial guess in iterative solution of series of systems arising in fluid flow simulations

    Damien Tromeur-Dervout and Yuri Vassilevski. Choice o f initial guess in iterative solution of series of systems arising in fluid flow simulations. Journal of Computational Physics , 219(1):210–227, 2006

  70. [78]

    Tsai, Robert P.K

    Angela Y.J. Tsai, Robert P.K. Chan, and Shixiao Wang. Tw o-derivative Runge–Kutta methods for PDEs using a novel discretization approach. Numerical Algorithms, 65:687–703, 2014

  71. [79]

    Van Loan

    Charles F. Van Loan. The ubiquitous Kronecker product. Journal of computational and applied math- ematics, 123(1-2):85–100, 2000

  72. [80]

    Vasilyev and Christopher Bowman

    Oleg V. Vasilyev and Christopher Bowman. Second-gener ation wavelet collocation method for the solution of partial differential equations. Journal of Computational Physics , 165(2):660–693, 2000

  73. [81]

    Jan G. Verwer. Explicit Runge-Kutta methods for parabo lic partial differential equations. Applied Numerical Mathematics, 22(1-3):359–379, 1996

  74. [82]

    Improvin g initial guess for the iterative solution of linear equation systems in incompressible flow

    Shuai Ye, Yufei Lin, Liyang Xu, and Jiaming Wu. Improvin g initial guess for the iterative solution of linear equation systems in incompressible flow. Mathematics, 8(1):119, 2020

  75. [83]

    Partial differential equations of applied mathematics

    Erich Zauderer. Partial differential equations of applied mathematics . John Wiley & Sons, 2011. 19 Appendices A. Matrices for Boundary/Initial Condition Enforcement Px =   ← ··· ⃗0 ··· → In−2 ← ··· ⃗0 ··· →  ∈ Rn×(n−2) Pt =   ← ··· ⃗0 ··· → Is−1  ∈ Rs×(s−1) XD =    ...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.