Pith. sign in

REVIEW 3 major objections 4 minor 53 references

Multiscale approximation and two-grid preconditioner for extremely anisotropic heat flow

T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that local spectral eigenfunctions aligned with magnetic field lines give an accurate coarse-grid approximation and an anisotropy-independent two-grid preconditioner for extremely anisotropic heat flow, with iteration…

desk verdict Strong numerical case for spectral multiscale coarse spaces on extreme anisotropic heat flow, but the two-grid proof does not deliver the claimed O(1) anisotropy independence. read the letter →

arxiv 2412.08355 v1 pith:44FERQKC submitted 2024-12-11 math.NA cs.NA

Maria Vasilyeva , Golo A. Wimmer , Ben S. Southworth This is my paper · ORCID
classification math.NAcs.NA MSC 65N5565N3065F10
keywords anisotropicdiffusionmagneticconfinementfusiongeneralizedmultiscalefiniteelementsspectralbasistwo-gridpreconditionergapheattransport
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 sets out to show that a coarse space built from local spectral eigenproblems of the anisotropic diffusion operator can do double duty: it approximates extremely anisotropic heat flow accurately on a coarse grid, and it makes a two-grid preconditioner converge in $O(1)$ iterations, independent of the ratio $k_\parallel/k_\perp$ up to $10^{12}$. The key difficulty this addresses is that temperature fields follow magnetic field lines that do not align with the mesh, so ordinary coarse spaces and algebraic multigrid fail or need expensive line relaxation. The paper claims that the lowest eigenvectors of a normalized anisotropic Laplacian naturally align with the field, so a handful of them per coarse node captures the parallel heat flow. If the claim holds, the same offline-computable basis serves as both a reduced-order surrogate model and a robust preconditioner for implicit fusion plasma simulations.

What carries the argument

The central object is the local generalized eigenvalue problem (9) in each coarse-grid subdomain $\omega_i$, with $A^{\omega_i}_h$ the local anisotropic stiffness matrix and $D^{\omega_i}_h$ the diagonal of that matrix. The eigenvectors belonging to the $J$ smallest eigenvalues become the multiscale basis functions after multiplication by a linear partition of unity $\chi_i$; they encode transport along the magnetic field within the coarse basis. These basis functions define the interpolation operator $P$ and restriction $R = P^T$, from which the coarse system matrices are obtained by Galerkin projection. The two-grid convergence argument then runs through the approximation properties (16), whose constants involve the local spectral gap $\lambda_{J+1}$, leading to the condition-number bound $K_{TG} = (1 + C/\tau) H^2/\Lambda_*$, where $\Lambda_* = H^2 \lambda_{J+1}$.

What would settle it

Evaluate the local eigenvalues of (9) on the closed-field-line configuration of Test 1 at fixed $J$ for anisotropy ratios $k_\parallel/k_\perp = 10^3, 10^6, 10^9, 10^{12}$ and track $\Lambda_* = H^2\lambda_{J+1}$. If $\Lambda_*$ decreases significantly with anisotropy, then $K_{TG} = (1+C/\tau)H^2/\Lambda_*$ grows and the two-grid iteration count would be expected to rise, contradicting the claimed anisotropy-independent $O(1)$ convergence.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that the local generalized eigenvalue problem $A^{\omega_i}_h \varphi = \lambda D^{\omega_i}_h \varphi$, where $A^{\omega_i}_h$ is the anisotropic stiffness matrix on a subdomain and $D^{\omega_i}_h$ its diagonal, produces eigenvectors whose low modes are aligned with the magnetic field. These eigenvectors, multiplied by a linear partition of unity, form a multiscale space that gives a highly accurate coarse-grid approximation of the evolving temperature field, with errors that decrease as the number $J$ of basis functions per coarse node grows. When this same coarse space is embedded as the coarse-grid correction in a two-level preconditioner with pointwise Jacobi or Gauss-Seidel smoothing, the paper reports preconditioned conjugate gradient convergence in $O(1)$ iterations for anisotropy ratios up to $10^{12}$, across three magnetic field configurations including closed and open field lines.

Load-bearing premise

The paper's convergence claim rests on the premise that the scaled spectral gap $\Lambda_* = H^2\lambda_{J+1}$ stays sufficiently large, or grows, as the anisotropy ratio $k_\parallel/k_\perp$ increases while the number $J$ of local basis functions is held fixed; the two-grid bound $K_{TG} = (1+C/\tau)H^2/\Lambda_*$ only delivers $O(1)$ iterations if that gap does not shrink with anisotropy, and the paper demonstrates this condition numerically rather than proving it.

Editorial extensions

If this is right

  • Since the basis depends only on the magnetic field and the mesh, it can be computed once offline and reused across many implicit time steps, amortizing the construction cost.
  • The same coarse space serves simultaneously as a reduced-order surrogate model, shrinking degrees of freedom from roughly 200,000 to a few thousand, and as a preconditioner for the full fine-scale solve.
  • The solver needs only standard pointwise Jacobi or Gauss-Seidel smoothing; no line relaxation, semi-coarsening, or field-aligned meshing is required.
  • On the 20x20 coarse grid with enough basis functions, iteration counts fall to 6-10 and stay there as the anisotropy ratio grows, so adding more local basis functions monotonically improves convergence rather than stalling.
  • The method handles configurations with closed magnetic field lines, the hard case for the paper's earlier block-preconditioning approach, as well as open field lines.

Reading between the lines

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

  • A rigorous proof that the scaled spectral gap $\Lambda_* = H^2\lambda_{J+1}$ stays bounded below (or grows) with $k_\parallel/k_\perp$ would turn the numerically observed $O(1)$ convergence into a theorem; the topological difference between closed and open field lines is a plausible controlling factor for that gap.
  • The same normalized-Laplacian construction could transfer to other transport problems with a distinguished direction, such as advection-dominated flow, drift-reduced plasma models, or anisotropic permeability in porous media, where the magnetic field is replaced by the relevant flow direction.
  • Because the paper reports only online solve times, the total time-to-solution including offline basis construction, which the authors note is parallelizable across subdomains, remains unquantified; a parallel-overhead study would show whether the two-grid gain survives when basis construction is included.
  • A testable prediction is that configurations with a higher fraction of closed field lines require more basis functions $J$ to reach a given accuracy; the monotone improvement seen in the iteration tables suggests the local spectral gap is the quantity to monitor.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper proposes a GMsFEM-type spectral multiscale coarse space for strongly anisotropic heat flow. Local generalized eigenproblems (9) with D=diag(A) are solved on coarse-node neighborhoods; eigenvectors with the smallest eigenvalues are multiplied by partition-of-unity functions to form a multiscale space. This space is used both as a reduced-order model and as the coarse-grid correction in a two-grid preconditioner with Jacobi or Gauss-Seidel smoothing. The authors derive local and global approximation estimates (14)-(16), an error estimate for the multiscale solution (17), and a two-grid condition-number bound K_TG=(1+C/τ)H²/Λ* in Appendix B, asserting anisotropy-independent convergence. Numerical experiments for three magnetic-field configurations and contrast ratios up to 10^12 show errors decreasing with the number J of basis functions and iteration counts that stabilize, in contrast to standard AMG solvers.

Significance. If the O(1) iteration claim were rigorously established, this would be a practically important contribution for magnetic-fusion simulations: a field-aligned spectral coarse space serving both as a reduced model and as a preconditioner, with robustness over twelve orders of anisotropy. The numerical evidence is substantial and systematic, covering three fields, four contrast levels, two coarse grids, two smoothers, and comparisons with three AMG variants; the trends in Tables 1-6 are internally consistent. However, the proof of the headline claim is incomplete. It rests on the unquantified behavior of Λ*=H²λ_{J+1}, and for the actual normalization D=diag(A) the scaling analysis indicates that the displayed bounds are mesh-dependent. The contribution is therefore promising, but it needs either a corrected proof with a genuine lower bound on the spectral gap or a weakened statement of the theoretical claims.

major comments (3)
  1. [§3.3, Eq. (17), Appendix B] The proof's central quantity Λ*=H²λ_{J+1} is never analyzed. For eigenproblem (9) with D=diag(A), on a patch of diameter H the first nonconstant eigenvalues of D^{-1}A scale like (h/H)²: already in 1D with P1 elements, λ_2=1-cos(π h/H)≈π² h²/(2H²), and the 2D case behaves the same way. Consequently Λ* is O(h²), and H²/Λ* is O((H/h)²). The bounds (17) and K_TG=(1+C/τ)H²/Λ* therefore contain a mesh-dependent factor and do not establish O(1) iteration counts. The sentence in §4 that 'given a sufficient number of multiscale basis functions, we can obtain anisotropy-independent convergence' is not a consequence of the displayed theorem. The authors should either use a normalization for which λ_{J+1} is O(1) on a unit domain (for example a mass-type weighted inner product as in standard GMsFEM) or supply a quantitative lower bound on λ_{J+1} in terms of H, h, and the problem data.
  2. [§3.3, §4, Appendix B] No estimate is given for how λ_{J+1} (or Λ*) depends on the anisotropy ratio. For a magnetic field aligned with the mesh, there are nonconstant finite-element functions v with b·∇v=0 exactly; for these v, vᵀA_aniso v=0 while vᵀD_aniso v>0, so the Rayleigh quotient in (9) tends to 0 as k∥/k⊥→∞. Thus λ_{J+1} and Λ* collapse, and the proven bound diverges in a regime within the method's stated scope. The numerical O(1) iteration counts for non-aligned fields are credible, but they are not backed by the proof; a lower bound on Λ* in terms of H, h, and the angle between b and the mesh edges is needed before the claimed theorem can support the anisotropy-independence statement.
  3. [Appendix B, K_TG derivation] The step ||v-Πv||²_DQ ≤ (1+C/τ)||v-Πv||²_Dh is not justified by standard finite-element scaling. Since DQ=diag(Qh)=(1/τ)diag(Mh)+diag(Ah) and Dh=diag(Ah), while diag(Mh)~h² and diag(Ah)~O(1) for P1/P2 elements, the factor should be of the form 1+C h²/τ and not 1+C/τ. This does not affect the numerical results, but it is another place where the displayed proof claims a stronger bound than the algebra supports.
minor comments (4)
  1. [§3.3] The bound |∇χ_i|≤1/H² should read |∇χ_i|≤C/H (or |∇χ_i|²≤C/H²); as written the scaling is inconsistent with the factor 1/H² appearing in (16).
  2. [§5.1, text after Figure 4] The sentence 'Generally, the error increases as J and coarse grid resolution increases' states the opposite of what Tables 1-3 show; the error decreases as J increases.
  3. [Table 6] In the 20×20 coarse-grid block, J=32, Gauss-Seidel smoother, k∥/k⊥=10³, the reported time tms=318.7 is an order of magnitude larger than the neighboring entries and appears to be a typo or an anomalous run; please verify.
  4. [§3.1, Eq. (9)] The phrase 'SSPD' is unusual; presumably 'symmetric positive definite' is meant. Also, D=diag(A) is symmetric positive definite only if all diagonal entries are positive, which holds here, but the terminology should be clarified.

Circularity Check

0 steps flagged · score 1.0 of 10

No circular derivation found: the coarse space and two-grid convergence estimates are expressed through the spectral gap of a local eigenproblem, not through fitted outputs or self-referential definitions.

full rationale

The paper's central claims—local spectral eigenvectors align with the magnetic field and give an accurate coarse approximation, and the two-grid preconditioner converges in O(1) iterations independent of anisotropy—are validated against an independent fine-grid reference solution and against external AMG solvers (mg1–mg3), rather than being derived from the target outputs. The convergence estimates in (16), (17), and Appendix B are all stated in terms of the spectral gap λ_{J+1}, or Λ* = H²λ_{J+1}, which is a quantity defined by the local generalized eigenproblem (9) and is not fitted to the numerical results. The method is adapted from the authors' earlier graph GMsFEM work [41,46], and the normalized-Laplacian construction with D = diag(A) is attributed to that line of work; however, those citations supply a construction template rather than the paper's conclusions. The decisive approximation inequalities are cited to the external reference [1], and the two-grid framework to [47,8,20,52]. No equation is defined in terms of the quantity it is supposed to predict, and no fitted parameter is renamed as a prediction. The main weakness is a rigor gap rather than circularity: Appendix B proves K_TG = (1 + C/τ)H²/Λ* but does not analyze how Λ* depends on k∥/k⊥, so the formal proof of anisotropy-independence is incomplete even though the numerical experiments support that conclusion. That concern belongs to correctness risk, not to circularity analysis.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The method introduces no fitted constants: accuracy depends on the user-chosen truncation rank J and coarse mesh size H, which are discretization choices rather than fitted parameters. The constant C in the two-grid bound is a generic equivalence constant, not fit to data. The proof relies on cited local projection inequalities, standard two-grid theory, and an unstated regularity condition.

assumptions (6)
  • standard math Local projection inequalities (14) for generalized eigenproblems with D=diag(A), taken from Abreu-Diaz-Galvis [1].
    Invoked in Section 3.3 to derive the global weak, strong, and fractional approximation properties in (16); the inequalities are not proved in the paper.
  • standard math Two-grid convergence framework and spectral equivalence of smoothers to DQ=diag(Qh), from [47,8,20,52].
    Appendix B reduces the two-grid condition number to sup_v ||v-Pi v||^2_DQ / ||v||^2_Qh and then applies (16); this standard machinery is cited, not derived.
  • domain assumption A_h and D_h are SSPD, with the first local eigenvalue zero and constant eigenvector.
    Section 3.1 states this 'by nature of the underlying problem and discretization'; it ensures real nonnegative eigenvalues and an orthogonal eigenbasis.
  • domain assumption The fine grid Th sufficiently resolves the magnetic field and temperature so that a P2 CG solution is an accurate reference.
    Section 2 assumes the fine grid has sufficient resolution of b and T; all error and iteration numbers in Section 5 are measured against this reference.
  • ad hoc to paper Additional regularity and appropriate initial conditions needed for the final error estimate (17).
    Appendix A closes with 'under some additional regularity and appropriate initial conditions' without specifying them; the final O(tau H^2/Lambda*) estimate depends on this unstated condition.
  • domain assumption The magnetic field is fixed in time during offline basis construction.
    Section 3 builds the spectral multiscale basis for a fixed field; all three test cases use time-independent field distributions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multiscale approximation and two-grid preconditioner for extremely anisotropic heat flow." pith.science (2026). https://pith.science/paper/44FERQKC

@misc{pith2026241208355,
  author       = {Pith},
  title        = {Pith review of: Multiscale approximation and two-grid preconditioner for extremely anisotropic heat flow},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/44FERQKC}},
  note         = {Machine review of arXiv:2412.08355}
}
abstract

We consider anisotropic heat flow with extreme anisotropy, as arises in magnetized plasmas for fusion applications. Such problems pose significant challenges in both obtaining an accurate approximation as well in the construction of an efficient solver. In both cases, the underlying difficulty is in forming an accurate approximation of temperature fields that follow the direction of complex, non-grid-aligned magnetic fields. In this work, we construct a highly accurate coarse grid approximation using spectral multiscale basis functions based on local anisotropic normalized Laplacians. We show that the local generalized spectral problems yield local modes that align with magnetic fields, and provide an excellent coarse-grid approximation of the problem. We then utilize this spectral coarse space as an approximation in itself, and as the coarse-grid in a two-level spectral preconditioner. Numerical results are presented for several magnetic field distributions and anisotropy ratios up to $10^{12}$, showing highly accurate results with a large system size reduction, and two-grid preconditioning that converges in $O(1)$ iterations, independent of anisotropy.

Figures

Figures reproduced from arXiv: 2412.08355 by the authors.

Figure 1
Figure 1. Illustration of the 10 × 10 coarse grid TH with local domain ωi and linear partition of unity function χi For the construction of a basis that provides an accurate approximation on the coarse grid TH, we will construct spectral multiscale basis functions motivated by the Generalized Multiscale Finite Element Method (GMsFEM) [19, 17, 15]. In the presented approach, we design an eigenvalue problem based on the normali… view at source ↗
Figure 2
Figure 2. Local domain with magnetic field and corresponding eigenvectors for three different anisotropy ratios. [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Reference solutions with magnetic field lines at final time for Test 1, 2 and 3, [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Error vs DOFH for Tests 1, 2 and 3 (from left to right). posed preconditioner with several classic multilevel AMG solvers from the PyAMG library. We consider three AMG preconditioners with pre- and post-smoothing to match the spectral two-level method, and default para…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

53 extracted references · 49 canonical work pages

  1. [1]

    A convergence analysis of generalized multiscale finite element methods

    Eduardo Abreu, Ciro D ´ ıaz, and Juan Galvis. A convergence analysis of generalized multiscale finite element methods. Journal of Computational Physics , 396:303–324, 2019

    work page 2019

  2. [2]

    Multiscale model reduction for shale gas transport in poroelastic fractured media

    I Yucel Akkutlu, Yalchin Efendiev, Maria Vasilyeva, and Yuhe Wang. Multiscale model reduction for shale gas transport in poroelastic fractured media. Journal of Computational Physics , 353:356–376, 2018

    work page 2018

  3. [3]

    Multiscale model reduction for shale gas trans- port in fractured media

    IY Akkutlu, Yalchin Efendiev, and Maria Vasilyeva. Multiscale model reduction for shale gas trans- port in fractured media. Computational Geosciences, pages 1–21, 2015

    work page 2015

  4. [4]

    Pyamg: Algebraic multigrid solvers in python

    Nathan Bell, Luke N Olson, and Jacob Schroder. Pyamg: Algebraic multigrid solvers in python. Journal of Open Source Software , 7(72):4142, 2022

    work page 2022

  5. [5]

    Pyamg: algebraic multigrid solvers in python

    Nathan Bell, Luke N Olson, Jacob Schroder, and Ben Southworth. Pyamg: algebraic multigrid solvers in python. Journal of Open Source Software , 8(87):5495, 2023

    work page 2023

  6. [6]

    The Mixed Virtual Element Discretization for highly-anisotropic problems: the role of the boundary degrees of freedom

    Stefano Berrone, Stefano Scial` o, and Gioana Teora. The mixed virtual element discretization for highly-anisotropic problems: the role of the boundary degrees of freedom. arXiv preprint arXiv:2307.16474, 2023

    work page Pith review arXiv 2023

  7. [7]

    Brandt, J

    A. Brandt, J. Brannick, K. Kahl, and I. Livshits. Algebraic distance for anisotropic diffusion prob- lems: multilevel results, 2014. arXiv:1409.4702

    arXiv 2014

  8. [8]

    Smoothed aggregation spectral element agglomeration amg: Sa- ρ amge

    Marian Brezina and Panayot S Vassilevski. Smoothed aggregation spectral element agglomeration amg: Sa- ρ amge. In International Conference on Large-Scale Scientific Computing , pages 3–15. Springer, 2011

    work page 2011

Show all 53 references
  1. [9]

    Briggs, Van Emden Henson, and Steve F

    William L. Briggs, Van Emden Henson, and Steve F. McCormick. A Multigrid Tutorial, Second Edition. Society for Industrial and Applied Mathematics, second edition, 2000

    2000

  2. [10]

    First order hyper- bolic approach for anisotropic diffusion equation

    Amareshwara Sainadh Chamarthi, Hiroaki Nishikawa, and Kimiya Komurasaki. First order hyper- bolic approach for anisotropic diffusion equation. Journal of Computational Physics , 396:243–263, 2019

    2019

  3. [11]

    Spectral amge (ρ amge)

    Tim Chartier, Robert D Falgout, VE Henson, J Jones, T Manteuffel, S McCormick, J Ruge, and Panayot S Vassilevski. Spectral amge (ρ amge). SIAM Journal on Scientific Computing , 25(1):1–26, 2003

    2003

  4. [12]

    Spectral element agglomerate amge

    Timothy Chartier, Robert Falgout, Van Emden Henson, Jim E Jones, Tom A Manteuffel, John W Ruge, Steve F McCormick, and Panayot S Vassilevski. Spectral element agglomerate amge. In Domain decomposition methods in science and engineering XVI , pages 513–521. Springer, 2007

    2007

  5. [13]

    Constraint energy minimizing general- ized multiscale finite element method

    Eric T Chung, Yalchin Efendiev, and Wing Tat Leung. Constraint energy minimizing general- ized multiscale finite element method. Computer Methods in Applied Mechanics and Engineering , 339:298–319, 2018

    2018

  6. [14]

    Non-local multi-continua upscaling for flows in heterogeneous fractured media

    Eric T Chung, Yalchin Efendiev, Wing Tat Leung, Maria Vasilyeva, and Yating Wang. Non-local multi-continua upscaling for flows in heterogeneous fractured media. Journal of Computational Physics, 372:22–34, 2018. 22

    2018

  7. [15]

    Generalized multiscale finite el- ement methods for problems in perforated heterogeneous domains

    Eric T Chung, Yalchin Efendiev, Guanglian Li, and Maria Vasilyeva. Generalized multiscale finite el- ement methods for problems in perforated heterogeneous domains. Applicable Analysis, 95(10):2254– 2279, 2016

    2016

  8. [16]

    Multiscale model reduction for transport and flow problems in perforated domains

    Eric T Chung, Wing Tat Leung, Maria Vasilyeva, and Yating Wang. Multiscale model reduction for transport and flow problems in perforated domains. Journal of Computational and Applied Mathematics, 330:519–535, 2018

    2018

  9. [17]

    Generalized multiscale finite element methods (gmsfem)

    Yalchin Efendiev, Juan Galvis, and Thomas Y Hou. Generalized multiscale finite element methods (gmsfem). Journal of computational physics , 251:116–135, 2013

    2013

  10. [18]

    Multiscale spectral amge solvers for high-contrast flow problems

    Yalchin Efendiev, Juan Galvis, and Panayot S Vassilevski. Multiscale spectral amge solvers for high-contrast flow problems. ISC-Preprint, Texas A&M University , 2012

    2012

  11. [19]

    Multiscale finite element methods for high-contrast problems using local spectral basis functions

    Yalchin Efendiev, Juan Galvis, and Xiao-Hui Wu. Multiscale finite element methods for high-contrast problems using local spectral basis functions. Journal of Computational Physics , 230(4):937–955, 2011

    2011

  12. [20]

    On two-grid convergence esti- mates

    Robert D Falgout, Panayot S Vassilevski, and Ludmil T Zikatanov. On two-grid convergence esti- mates. Numerical linear algebra with applications , 12(5-6):471–494, 2005

    2005

  13. [21]

    Domain decomposition preconditioners for multiscale flows in high-contrast media

    Juan Galvis and Yalchin Efendiev. Domain decomposition preconditioners for multiscale flows in high-contrast media. Multiscale Modeling & Simulation , 8(4):1461–1483, 2010

    2010

  14. [22]

    A new smoothed aggregation multigrid method for anisotropic problems

    Michael W Gee, Jonathan J Hu, and Raymond S Tuminaro. A new smoothed aggregation multigrid method for anisotropic problems. Numerical Linear Algebra with Applications , 16(1):19–37, 2009

    2009

  15. [23]

    Gmsh: A 3-d finite element mesh generator with built-in pre-and post-processing facilities

    Christophe Geuzaine and Jean-Fran¸ cois Remacle. Gmsh: A 3-d finite element mesh generator with built-in pre-and post-processing facilities. International journal for numerical methods in engineer- ing, 79(11):1309–1331, 2009

    2009

  16. [24]

    A high-order non field-aligned approach for the discretization of strongly anisotropic diffusion operators in magnetic fusion

    Giorgio Giorgiani, Hugo Bufferand, Fr´ ed´ eric Schwander, Eric Serre, and P Tamain. A high-order non field-aligned approach for the discretization of strongly anisotropic diffusion operators in magnetic fusion. Computer Physics Communications , 254:107375, 2020

    2020

  17. [25]

    An efficient high-order numerical solver for diffusion equations with strong anisotropy

    David Green, Xiaozhe Hu, Jeremy Lore, Lin Mu, and Mark L Stowell. An efficient high-order numerical solver for diffusion equations with strong anisotropy. Computer Physics Communications, 276:108333, 2022

    2022

  18. [26]

    Finite element and higher order difference for- mulations for modelling heat transport in magnetised plasmas

    Sibylle G¨ unter, Karl Lackner, and C Tichmann. Finite element and higher order difference for- mulations for modelling heat transport in magnetised plasmas. Journal of Computational Physics , 226(2):2306–2316, 2007

    2007

  19. [27]

    A constrained finite element method sat- isfying the discrete maximum principle for anisotropic diffusion problems

    Dmitri Kuzmin, Mikhail J Shashkov, and Daniil Svyatskiy. A constrained finite element method sat- isfying the discrete maximum principle for anisotropic diffusion problems. Journal of Computational Physics, 228(9):3448–3463, 2009

    2009

  20. [28]

    Automated solution of differential equations by the finite element method: The FEniCS book , volume 84

    Anders Logg, Kent-Andre Mardal, and Garth Wells. Automated solution of differential equations by the finite element method: The FEniCS book , volume 84. Springer Science & Business Media, 2012. 23

    2012

  21. [29]

    Nonsymmetric reduction-based algebraic multigrid

    Thomas A Manteuffel, Steffen M¨ unzenmaier, John Ruge, and Ben Southworth. Nonsymmetric reduction-based algebraic multigrid. SIAM Journal on Scientific Computing , 41(5):S242–S268, 2019

    2019

  22. [30]

    A root-node–based algebraic multigrid method

    Thomas A Manteuffel, Luke N Olson, Jacob B Schroder, and Ben S Southworth. A root-node–based algebraic multigrid method. SIAM Journal on Scientific Computing , 39(5):S723–S756, 2017

    2017

  23. [31]

    Nonsymmetric algebraic multigrid based on local approximate ideal restriction (lAIR)

    Thomas A Manteuffel, John Ruge, and Ben S Southworth. Nonsymmetric algebraic multigrid based on local approximate ideal restriction (lAIR). SIAM Journal on Scientific Computing , 40(6):A4105– A4130, 2018

    2018

  24. [32]

    Convergence in norm of nonsymmetric algebraic multigrid

    Tom Manteuffel and Ben S Southworth. Convergence in norm of nonsymmetric algebraic multigrid. SIAM Journal on Scientific Computing , 41(5):S269–S296, 2019

    2019

  25. [33]

    The partition of unity finite element method: basic theory and applications

    Jens M Melenk and Ivo Babuˇ ska. The partition of unity finite element method: basic theory and applications. Computer methods in applied mechanics and engineering , 139(1-4):289–314, 1996

    1996

  26. [34]

    Convergence analysis of perturbed two-grid and multigrid methods

    Yvan Notay. Convergence analysis of perturbed two-grid and multigrid methods. SIAM journal on numerical analysis, 45(3):1035–1044, 2007

    2007

  27. [35]

    Algebraic multigrid

    John W Ruge and Klaus St¨ uben. Algebraic multigrid. In Multigrid methods, pages 73–130. SIAM, 1987

    1987

  28. [36]

    Smoothed aggregation solvers for anisotropic diffusion

    Jacob B Schroder. Smoothed aggregation solvers for anisotropic diffusion. Numerical Linear Algebra with Applications, 19(2):296–312, 2012

    2012

  29. [37]

    Preserving monotonicity in anisotropic diffusion

    Prateek Sharma and Gregory W Hammett. Preserving monotonicity in anisotropic diffusion. Journal of Computational Physics , 227(1):123–142, 2007

    2007

  30. [38]

    Air algebraic multigrid for a space- time hybridizable discontinuous galerkin discretization of advection (-diffusion)

    Abdullah Ali Sivas, Ben S Southworth, and Sander Rhebergen. Air algebraic multigrid for a space- time hybridizable discontinuous galerkin discretization of advection (-diffusion). SIAM Journal on Scientific Computing , 43(5):A3393–A3416, 2021

    2021

  31. [39]

    Nonlinear magnetohydrodynamics simulation using high- order finite elements

    Carl R Sovinec, AH Glasser, TA Gianakon, DC Barnes, RA Nebel, SE Kruger, DD Schnack, SJ Plimpton, A Tarditi, MS Chu, et al. Nonlinear magnetohydrodynamics simulation using high- order finite elements. Journal of Computational Physics , 195(1):355–386, 2004

    2004

  32. [40]

    Multigrid

    Ulrich Trottenberg, Cornelius W Oosterlee, and Anton Schuller. Multigrid. Academic press, 2001

    2001

  33. [41]

    Generalized multiscale finite element method for discrete network (graph) models

    Maria Vasilyeva. Generalized multiscale finite element method for discrete network (graph) models. Journal of Computational and Applied Mathematics , 457:116275, 2025

    2025

  34. [42]

    Multiscale dimension reduction for flow and transport problems in thin domain with reactive boundaries

    Maria Vasilyeva, Valentin Alekseev, Eric T Chung, and Yalchin Efendiev. Multiscale dimension reduction for flow and transport problems in thin domain with reactive boundaries. Journal of Computational Physics , 442:110512, 2021

    2021

  35. [43]

    Upscaling of the single-phase flow and heat transport in fractured geothermal reservoirs using nonlocal multicontinuum method

    Maria Vasilyeva, Masoud Babaei, Eric T Chung, and Valentin Alekseev. Upscaling of the single-phase flow and heat transport in fractured geothermal reservoirs using nonlocal multicontinuum method. Computational Geosciences, 23:745–759, 2019

    2019

  36. [44]

    Decoupled multiscale numerical approach for reactive transport in marine sediment column

    Maria Vasilyeva, Richard B Coffin, and Ingo Pecher. Decoupled multiscale numerical approach for reactive transport in marine sediment column. Computer Methods in Applied Mechanics and Engineering, 428:117087, 2024. 24

    2024

  37. [45]

    Multiscale model reduction for pore- scale simulation of li-ion batteries using gmsfem.Journal of Computational and Applied Mathematics, 344:73–88, 2018

    Maria Vasilyeva, Aashutosh Mistry, and Partha P Mukherjee. Multiscale model reduction for pore- scale simulation of li-ion batteries using gmsfem.Journal of Computational and Applied Mathematics, 344:73–88, 2018

    2018

  38. [46]

    An adaptive two-grid preconditioner for flow in fractured porous media

    Maria Vasilyeva, Ben S Southworth, and Shubin Fu. An adaptive two-grid preconditioner for flow in fractured porous media. arXiv preprint arXiv:2411.17903 , 2024

    2024 arXiv

  39. [47]

    Multilevel block factorization preconditioners: Matrix-based analysis and algorithms for solving finite element equations

    Panayot S Vassilevski. Multilevel block factorization preconditioners: Matrix-based analysis and algorithms for solving finite element equations . Springer Science & Business Media, 2008

    2008

  40. [48]

    Coarse spaces by algebraic multigrid: Multigrid convergence and upscaling error estimates

    Panayot S Vassilevski. Coarse spaces by algebraic multigrid: Multigrid convergence and upscaling error estimates. Advances in Adaptive Data Analysis , 3(01n02):229–249, 2011

    2011

  41. [49]

    Scipy 1.0: fundamental algorithms for scientific computing in python

    Pauli Virtanen, Ralf Gommers, Travis E Oliphant, Matt Haberland, Tyler Reddy, David Cour- napeau, Evgeni Burovski, Pearu Peterson, Warren Weckesser, Jonathan Bright, et al. Scipy 1.0: fundamental algorithms for scientific computing in python. Nature methods, 17(3):261–272, 2020

    2020

  42. [50]

    A fast algebraic multigrid solver and accurate discretization for highly anisotropic heat flux i: open field lines

    Golo A Wimmer, Ben S Southworth, Thomas J Gregory, and Xian-Zhu Tang. A fast algebraic multigrid solver and accurate discretization for highly anisotropic heat flux i: open field lines. SIAM Journal on Scientific Computing , 46(3):A1821–A1849, 2024

    2024

  43. [51]

    An accurate SUPG- stabilized continuous Galerkin discretization for anisotropic heat flux in magnetic confinement fusion

    Golo A Wimmer, Ben S Southworth, Koki Sagiyama, and Xian-Zhu Tang. An accurate SUPG- stabilized continuous Galerkin discretization for anisotropic heat flux in magnetic confinement fusion. (in review), 2024

    2024

  44. [52]

    The method of alternating projections and the method of sub- space corrections in hilbert space

    Jinchao Xu and Ludmil Zikatanov. The method of alternating projections and the method of sub- space corrections in hilbert space. Journal of the American Mathematical Society , 15(3):573–597, 2002

    2002

  45. [53]

    Algebraic multigrid methods

    Jinchao Xu and Ludmil Zikatanov. Algebraic multigrid methods. Acta Numerica, 26:591–721, 2017. 25

    2017

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.