pith. sign in

arxiv: 2606.31662 · v1 · pith:E2O2CF4Bnew · submitted 2026-06-30 · 🧮 math.NA · cs.NA

Positivity-preserving dynamical low-rank methods for the Vlasov equation

Katharina Kormann , Murtazo Nazarov , Junjie Wen This is my paper

Pith reviewed 2026-07-01 04:03 UTC · model grok-4.3

classification 🧮 math.NA cs.NA
keywords positivity preservationlow-rank approximationVlasov equationquadratic programmingVlasov-Poissonspectral discretizationmass conservationRunge-Kutta
0
0 comments X

The pith

Low-rank Vlasov approximations regain positivity and other physical properties through a minimal correction found by quadratic programming.

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

The paper develops correction procedures that restore positivity to dynamical low-rank solutions of the Vlasov equation. Structural requirements are cast as constraints, and a quadratic program locates the smallest additive term that satisfies them. The resulting solution stays close to the original low-rank form yet obeys the enforced properties. One of the two schemes also conserves total mass and momentum. The methods are tested on Vlasov-Poisson and Vlasov-Poisson-BGK problems discretized spectrally in space and advanced with explicit Runge-Kutta time stepping.

Core claim

We introduce positivity-preserving correction methods for low-rank approximations of the Vlasov equation. The key idea is to formulate structural properties, including positivity-preservation, as constraints and to seek a minimal correction term that is added to the low-rank solution, by solving a quadratic programming problem. As a result, the corrected solution satisfies the constraints and preserves these properties, while remaining close to the original low-rank solution. Two positivity-preserving schemes are proposed in this work, and one of them also preserves the total mass and momentum of the system.

What carries the argument

The quadratic programming problem that computes the smallest correction term enforcing positivity and additional structural constraints on the low-rank solution.

If this is right

  • The corrected low-rank solution satisfies the positivity constraint at every point.
  • One of the two schemes additionally preserves total mass and momentum.
  • The corrected solution remains close to the uncorrected low-rank approximation.
  • The procedures apply directly to Vlasov-Poisson and Vlasov-Poisson-BGK systems under spectral spatial discretization and explicit Runge-Kutta time integration.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same correction framework could be adapted to enforce other invariants such as energy or entropy in kinetic models.
  • The approach may extend to other hyperbolic or kinetic equations where low-rank methods produce unphysical negative values.
  • Efficient quadratic-program solvers tailored to the low-rank tensor format would be required before the method scales to higher-dimensional problems.

Load-bearing premise

The quadratic programming problem for the correction can be solved efficiently at each time step without destroying the low-rank structure or the overall accuracy of the approximation.

What would settle it

A computation in which the time required to solve the quadratic program at each step exceeds the computational savings of the low-rank representation, or in which the size of the correction term becomes comparable to the low-rank solution itself.

Figures

Figures reproduced from arXiv: 2606.31662 by Junjie Wen, Katharina Kormann, Murtazo Nazarov.

Figure 1
Figure 1. Figure 1: Landau damping: evolution of the electric energy (a), rank (b), the relative [PITH_FULL_IMAGE:figures/full_fig_p014_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Landau damping with BGK operator: evolution of the electric energy (a), [PITH_FULL_IMAGE:figures/full_fig_p015_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Two-stream instability: evolution of the electric energy (a), rank (b), the [PITH_FULL_IMAGE:figures/full_fig_p016_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Two-stream instability: evolution of the lower bound of the distribution func [PITH_FULL_IMAGE:figures/full_fig_p017_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Two-stream instability: distribution functions at [PITH_FULL_IMAGE:figures/full_fig_p018_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Two-stream instability: the ∥ · ∥F of the correction term δF for PP-KL and MMC-PP-S methods, respectively. The number of degrees of freedom is Nx = Nv = 128 [PITH_FULL_IMAGE:figures/full_fig_p018_6.png] view at source ↗
read the original abstract

In this manuscript, we introduce positivity-preserving correction methods for low-rank approximations of the Vlasov equation. The key idea is to formulate structural properties, including positivity-preservation, as constraints and to seek a minimal correction term that is added to the low-rank solution, by solving a quadratic programming problem. As a result, the corrected solution satisfies the constraints and preserve these properties, while remaining close to the original low-rank solution. Two positivity-preserving schemes are proposed in this work, and one of them also preserves the total mass and momentum of the system. We apply the proposed methods to a Vlasov--Poisson and Vlasov--Poisson-BGK employing a spectral discretization in space and an explicit Runge--Kutta scheme in time. Numerical experiments demonstrate the effectiveness of the proposed methods.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The manuscript introduces two positivity-preserving correction schemes for dynamical low-rank approximations of the Vlasov equation. Structural properties (positivity, and optionally mass and momentum) are encoded as constraints in a quadratic program whose solution yields a minimal correction added to the low-rank solution; the corrected field is asserted to satisfy the constraints while remaining close to the original low-rank approximant. The schemes are combined with a spectral spatial discretization and explicit Runge-Kutta time stepping and are tested on Vlasov-Poisson and Vlasov-Poisson-BGK problems; numerical experiments are stated to demonstrate effectiveness.

Significance. A method that reliably enforces positivity (and conservation) inside a dynamical low-rank framework without sacrificing the O(r²N) complexity would be a useful addition to structure-preserving integrators for kinetic equations. The quadratic-programming correction idea is conceptually clean and could extend to other low-rank settings if the efficiency question is resolved.

major comments (2)
  1. [Method description (abstract and §3)] The quadratic-programming correction is formulated without any indication that the optimization is performed on the tangent space of the low-rank manifold or that the resulting correction is immediately re-compressed to rank r. For a spectral discretization the decision variables of the QP live in a space whose dimension is exponential in the number of velocity dimensions, while the low-rank factors are only size N×r; a dense correction immediately produces an O(N^d) update whose storage and subsequent dynamical-low-rank step destroy the complexity advantage that motivates the entire approach. This point is load-bearing for the claim that the corrected solution “remains close to the original low-rank solution” while preserving the low-rank structure.
  2. [Numerical experiments section] No quantitative error tables, convergence rates, or timing breakdowns are supplied that would show the cost of the QP solve relative to the low-rank update or that the positivity guarantee survives any re-compression step. Without such data the assertion that the schemes are effective cannot be assessed against the central efficiency requirement of dynamical low-rank methods.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the thorough review and valuable feedback on our manuscript. We address the major comments point by point below. We will revise the manuscript to incorporate clarifications and additional data as outlined in our responses.

read point-by-point responses

  1. Referee: [Method description (abstract and §3)] The quadratic-programming correction is formulated without any indication that the optimization is performed on the tangent space of the low-rank manifold or that the resulting correction is immediately re-compressed to rank r. For a spectral discretization the decision variables of the QP live in a space whose dimension is exponential in the number of velocity dimensions, while the low-rank factors are only size N×r; a dense correction immediately produces an O(N^d) update whose storage and subsequent dynamical-low-rank step destroy the complexity advantage that motivates the entire approach. This point is load-bearing for the claim that the corrected solution “remains close to the original low-rank solution” while preserving the low-rank structure.

    Authors: The referee correctly identifies that the manuscript does not explicitly describe the optimization being restricted to the tangent space or the immediate re-compression of the correction to rank r. In our approach, the QP is solved over the full discretized phase space to find the minimal correction that enforces positivity (and optionally conservation). The corrected solution is then used as input to the subsequent dynamical low-rank step, which inherently involves projection onto the low-rank manifold. However, to ensure the complexity advantage is maintained, we agree that explicit re-compression after correction is important and will be detailed in the revised §3. We will also clarify that the method does not constrain the correction to the tangent space but relies on the minimal nature of the correction to keep the solution close to the low-rank approximant. Regarding the exponential dimension, for the tested problems with moderate velocity dimensions, the QP is solved using efficient solvers, but we acknowledge this may limit scalability and will discuss this in the revision. revision: yes

  2. Referee: [Numerical experiments section] No quantitative error tables, convergence rates, or timing breakdowns are supplied that would show the cost of the QP solve relative to the low-rank update or that the positivity guarantee survives any re-compression step. Without such data the assertion that the schemes are effective cannot be assessed against the central efficiency requirement of dynamical low-rank methods.

    Authors: We agree with the referee that the numerical section lacks sufficient quantitative information. In the revised manuscript, we will add tables with error norms, observed convergence rates, and CPU timing comparisons between the low-rank update and the QP correction step. Additionally, we will include verification that the positivity (and conservation) properties hold after the full time step including any re-compression. revision: yes

Circularity Check

0 steps flagged

No circularity: method is a direct constructive correction via QP

full rationale

The paper introduces a new algorithmic construction: formulate positivity (and optionally mass/momentum) as constraints on the low-rank solution, then obtain a minimal correction by solving a quadratic program. No derivation step reduces by construction to its own inputs, no fitted parameter is relabeled as a prediction, and no load-bearing claim rests on a self-citation chain. The approach is presented as an independent numerical technique whose correctness is checked by experiments on Vlasov-Poisson and Vlasov-Poisson-BGK problems; the derivation chain remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No free parameters, axioms, or invented entities are identifiable from the abstract; the method builds on standard low-rank techniques and quadratic programming without introducing new postulated entities.

pith-pipeline@v0.9.1-grok · 5669 in / 1208 out tokens · 61575 ms · 2026-07-01T04:03:56.822206+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

27 extracted references · 22 canonical work pages

  1. [1]

    Kormann, A semi-Lagrangian Vlasov solver in tensor train format, SIAM J

    K. Kormann, A semi-Lagrangian Vlasov solver in tensor train format, SIAM J. Sci. Comput. 37 (4) (2015) B613–B632.doi:10.1137/140971270

    work page doi:10.1137/140971270 2015

  2. [2]

    Ehrlacher, D

    V. Ehrlacher, D. Lombardi, A dynamical adaptive tensor method for the Vlasov-Poisson system, J. Comput. Phys. 339 (2017) 285–306.doi:10.1016/j.jcp.2017.03.015

    work page doi:10.1016/j.jcp.2017.03.015 2017

  3. [3]

    Guo, J.-M

    W. Guo, J.-M. Qiu, A low rank tensor representation of linear transport and nonlinear Vlasov solutions and their associated flow maps, J. Comput. Phys. 458 (2022) Paper No. 111089, 24.doi:10.1016/j.jcp.2022.111089

    work page doi:10.1016/j.jcp.2022.111089 2022

  4. [4]

    Lubich, I

    C. Lubich, I. V. Oseledets, A projector-splitting integrator for dynamical low-rank approxima- tion, BIT 54 (1) (2014) 171–188.doi:10.1007/s10543-013-0454-0

    work page doi:10.1007/s10543-013-0454-0 2014

  5. [5]

    Einkemmer, C

    L. Einkemmer, C. Lubich, A low-rank projector-splitting integrator for the Vlasov-Poisson equation, SIAM J. Sci. Comput. 40 (5) (2018) B1330–B1360.doi:10.1137/18M116383X

    work page doi:10.1137/18m116383x 2018

  6. [6]

    Ceruti, J

    G. Ceruti, J. Kusch, C. Lubich, A rank-adaptive robust integrator for dynamical low-rank approximation, BIT 62 (4) (2022) 1149–1174.doi:10.1007/s10543-021-00907-7

    work page doi:10.1007/s10543-021-00907-7 2022

  7. [7]

    J. A. Rossmanith, D. C. Seal, A positivity-preserving high-order semi-Lagrangian discontinuous Galerkin scheme for the Vlasov-Poisson equations, J. Comput. Phys. 230 (16) (2011) 6203– 6232.doi:10.1016/j.jcp.2011.04.018

    work page doi:10.1016/j.jcp.2011.04.018 2011

  8. [8]

    Qiu, C.-W

    J.-M. Qiu, C.-W. Shu, Positivity preserving semi-Lagrangian discontinuous Galerkin formula- tion: theoretical analysis and application to the Vlasov-Poisson system, J. Comput. Phys. 230 (23) (2011) 8386–8409.doi:10.1016/j.jcp.2011.07.018

    work page doi:10.1016/j.jcp.2011.07.018 2011

  9. [9]

    Zheng, X

    N. Zheng, X. Cai, J.-M. Qiu, J. Qiu, A conservative semi-Lagrangian hybrid Hermite WENO scheme for linear transport equations and the nonlinear Vlasov-Poisson system, SIAM J. Sci. Comput. 43 (5) (2021) A3580–A3606.doi:10.1137/20M1363273

    work page doi:10.1137/20m1363273 2021

  10. [10]

    Xiong, J.-M

    T. Xiong, J.-M. Qiu, Z. Xu, A. Christlieb, High order maximum principle preserving semi- Lagrangian finite difference WENO schemes for the Vlasov equation, J. Comput. Phys. 273 (2014) 618–639.doi:10.1016/j.jcp.2014.05.033

    work page doi:10.1016/j.jcp.2014.05.033 2014

  11. [11]

    Kiechle, E

    Y. Kiechle, E. Chudzik, C. Helzel, A positivity-preserving active flux method for the Vlasov- Poisson system, J. Comput. Phys. 524 (2025) Paper No. 113693, 23.doi:10.1016/j.jcp. 2024.113693

    work page doi:10.1016/j.jcp 2025

  12. [12]

    Zhang, X

    X. Zhang, X. Cai, W. Cao, The positivity-preserving high-order semi-lagrangian spectral vol- ume method for vlasov-poisson equations (2025).arXiv:2504.04501

    arXiv 2025

  13. [13]

    Zhang, C.-W

    X. Zhang, C.-W. Shu, On maximum-principle-satisfying high order schemes for scalar conserva- tion laws, J. Comput. Phys. 229 (9) (2010) 3091–3120.doi:10.1016/j.jcp.2009.12.030

    work page doi:10.1016/j.jcp.2009.12.030 2010

  14. [14]

    Kormann, M

    K. Kormann, M. Nazarov, J. Wen, A second-order structure- and positivity-preserving convex limiting method for the vlasov equations, submitted to SISC (2026)

    2026

  15. [15]

    Einkemmer, K

    L. Einkemmer, K. Kormann, J. Kusch, R. G. McClarren, J.-M. Qiu, A review of low-rank methods for time-dependent kinetic simulations, J. Comput. Phys. 538 (2025) Paper No. 114191, 27.doi:10.1016/j.jcp.2025.114191

    work page doi:10.1016/j.jcp.2025.114191 2025

  16. [16]

    E. Ye, N. F. Loureiro, Quantized tensor networks for solving the vlasov–maxwell equations, J. of Plasma Phys. 90 (3) (2024) 805900301.doi:10.1017/S0022377824000503

    work page doi:10.1017/s0022377824000503 2024

  17. [17]

    Uschmajew, A

    A. Uschmajew, A. Zeiser, Dynamical low-rank approximation of the Vlasov-Poisson equation with piecewise linear spatial boundary, BIT 64 (2) (2024) Paper No. 19, 26.doi:10.1007/ s10543-024-01019-8

    2024

  18. [18]

    J. Wen, M. Nazarov, An anisotropic nonlinear stabilization for finite element approximation of Vlasov-Poisson equations, J. Comput. Phys. 536 (2025) Paper No. 114079, 21.doi: 10.1016/j.jcp.2025.114079

    work page doi:10.1016/j.jcp.2025.114079 2025

  19. [19]

    Kormann, M

    K. Kormann, M. Nazarov, J. Wen, A structure-preserving finite element framework for the Vlasov–Maxwell system, Comput. Methods Appl. Mech. Engrg. 446 (2025) Paper No. 118290.doi:10.1016/j.cma.2025.118290

    work page doi:10.1016/j.cma.2025.118290 2025

  20. [20]

    Einkemmer, I

    L. Einkemmer, I. Joseph, A mass, momentum, and energy conservative dynamical low-rank scheme for the Vlasov equation, J. Comput. Phys. 443 (2021) Paper No. 110495, 16.doi: 10.1016/j.jcp.2021.110495

    work page doi:10.1016/j.jcp.2021.110495 2021

  21. [21]

    Einkemmer, A

    L. Einkemmer, A. Ostermann, C. Scalone, A robust and conservative dynamical low-rank algorithm, J. Comput. Phys. 484 (2023) Paper No. 112060, 20.doi:10.1016/j.jcp.2023. 112060. URL https://doi.org/10.1016/j.jcp.2023.112060

    work page doi:10.1016/j.jcp.2023 2023

  22. [22]

    C. Liu, D. Milesis, C.-W. Shu, X. Zhang, Efficient optimization-based invariant-domain- preserving limiters in solving gas dynamics equations, J. Comput. Phys. 558 (2026) Paper No. 114839, 26.doi:10.1016/j.jcp.2026.114839. URL https://doi.org/10.1016/j.jcp.2026.114839

    work page doi:10.1016/j.jcp.2026.114839 2026

  23. [23]

    Liu, C.-W

    C. Liu, C.-W. Shu, X. Zhang, Efficient admissible set projection in optimization-based invariant- POSITIVITY-PRESERVING LOW-RANK METHODS21 domain-preserving limiters for ideal mhd (2026).arXiv:2605.10929. URL https://arxiv.org/abs/2605.10929

    Pith/arXiv arXiv 2026

  24. [24]

    O. Koch, C. Lubich, Dynamical low-rank approximation, SIAM Journal on Matrix Analysis and Applications 29 (2) (2007) 434–454

    2007

  25. [25]

    Ceruti, C

    G. Ceruti, C. Lubich, An unconventional robust integrator for dynamical low-rank approxima- tion, BIT 62 (1) (2022) 23–44.doi:10.1007/s10543-021-00873-0

    work page doi:10.1007/s10543-021-00873-0 2022

  26. [26]

    Crouseilles, F

    N. Crouseilles, F. Filbet, Numerical approximation of collisional plasmas by high order methods, J. Comput. Phys. 201 (2) (2004) 546–572.doi:10.1016/j.jcp.2004.06.007

    work page doi:10.1016/j.jcp.2004.06.007 2004

  27. [27]

    Campos Pinto, K

    M. Campos Pinto, K. Kormann, E. Sonnendr¨ ucker, Variational framework for structure- preserving electromagnetic particle-in-cell methods, J. Sci. Comput. 91 (2) (2022) Paper No. 46, 39.doi:10.1007/s10915-022-01781-3

    work page doi:10.1007/s10915-022-01781-3 2022