Pith. sign in

REVIEW 4 major objections 7 minor 75 references

Extending the Numerical Flow Iteration to the multi-species Vlasov-Maxwell system through Hamiltonian Splitting

T0 review · 4 major / 7 minor · reviewed 2026-08-03 · 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 the Numerical Flow Iteration—storing field histories and tracing characteristics backward to the initial time—extends from the electrostatic Vlasov-Poisson system to the multi-species electromagnetic Vlasov-Maxwell sys

desk verdict Plausible and useful extension of NuFI to Vlasov–Maxwell, but the printed update equations have sign/typo inconsistencies that must be fixed before the method is reproducible from the paper. read the letter →

arxiv 2511.11322 v3 pith:JQFND5EQ submitted 2025-11-14 physics.plasm-ph cs.NAmath.NA

Rostislav-Paul Wilhelm , Fabio Bacchini , Sebastian Schöps , Manuel Torrilhon , Melina Merkel , Matthias Kirchhart This is my paper · ORCID
classification physics.plasm-phcs.NAmath.NA MSC 35Q8365P1065M70 PACS 52.65.Ff52.35.Qz
keywords NumericalFlowIterationVlasov-MaxwellequationsHamiltoniansplittingstructure-preservingintegratorsemi-LagrangianmethodskineticplasmasimulationWeibelinstabilityfilamentation
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's aim is to show that the Numerical Flow Iteration (NuFI)—a phase-space solver that stores only the time-evolution of the electromagnetic fields and reconstructs the distribution function by tracing characteristics backward to the initial time—can be lifted from the electrostatic Vlasov-Poisson system to the full multi-species Vlasov-Maxwell system. The lifting works by splitting the Vlasov-Maxwell Hamiltonian into electric-, magnetic-, and kinetic-energy parts, each of which yields a linear transport equation with an explicit analytic solution; composing those flows produces two schemes, NuFI-Ham and NuFI-PC. The authors claim the structure-preserving features carry over: L^p norms and entropy are preserved (entropy with errors below 1e-5 and total energy below 1e-3 in the reported Weibel conservation test), so the method avoids the artificial dissipation of grid-based solvers. The numerical benchmarks—weak Landau damping, streaming Weibel instability, and filamentation instability—reproduce the expected damping/growth rates and display the NuFI signature ability to resolve velocity-space structure below the grid spacing. A sympathetic reader would care because, if correct, this provides a low-memory, conservative route to electromagnetic kinetic plasma simulations.

What carries the argument

The key object is the Hamiltonian splitting of the Vlasov-Maxwell Hamiltonian into three linear sub-Hamiltonians H_E, H_B, and H_f, each with closed-form characteristic flow: acceleration by E, rotation by the matrix exponential of the cross-product matrix J_B, and free streaming with current feedback. NuFI uses these flows in two ways: to advance fields on a coarse spatial grid, and to evaluate f at arbitrary phase-space points by composing the inverse flows backward from t_n to t=0 and applying them to the initial distribution. This composition is the mechanism that avoids storing any phase-space grid and, because it is symplectic, transfers the exact entropy/L^p preservation and drift-fre

What would settle it

Run the no-restart streaming Weibel simulation at 16^3 with Δt=1/200 exactly as Algorithm 1 is printed, and monitor total energy, entropy, and ∇·E-ρ/∇·B every step to t=100. If the printed equations are those actually implemented, total energy error should stay below 1e-3 and entropy error below 1e-5; if instead the code follows eq. (30) rather than eqs. (33)/(35), the current feedback sign flips and the growth rate will change. Comparing a single field-update step against the analytic sub-flows on a smooth initial current is the quickest check.

Watch

Extended reading notes

Core claim

The paper's central claim is that the Numerical Flow Iteration idea—reconstruct f(t,x,v) by composing backward characteristic maps all the way to t=0 and storing only the histories of E and B instead of the six-dimensional per-species distribution—remains valid and structure-preserving when the magnetic field and multiple species are added. The construction is the Hamiltonian splitting H_VM = H_E + H_B + H_f: the electric-field part translates velocities by E while B changes by -t∇×E; the magnetic-field part rotates velocities by the matrix exponential exp(-J_B t) while E changes by t∇×B; the kinetic part advects positions by -tv and updates E by the cumulative current. Composing these flows

Load-bearing premise

The whole construction rests on the printed analytic sub-Hamiltonian flows being correct; as written, eq. (30) gives ∂_t E=-j while eqs. (33)/(35) give ∂_t E=+j, eq. (52) has an extra Δt q/m factor, eq. (53) has sign errors, and the claimed Gauss-law preservation in Remark 2.1 is not numerically checked—if those formulas are wrong, the reported conservation and growth-rate results are not reproduced by the stated algorithm.

Editorial extensions

If this is right

  • If the extension is correct, electromagnetic kinetic runs can be made with dramatically less memory, because only field histories (and in NuFI-Ham a shifted current density) are stored, not 6D distribution functions per species.
  • The zoom property—recovering noise-free, sub-grid velocity-space detail after the simulation—extends to electromagnetic instabilities, as shown for the Weibel case.
  • NuFI-PC can be used as a subcycling/restart engine inside existing semi-Lagrangian codes; it relaxes the time-step restriction to Δt ≲ Δx but sacrifices exact conservation.
  • For high-fidelity conservation, NuFI-Ham is the better choice: entropy stays at quadrature-error level and total energy error remains below 1e-3 over the reported time span.
  • Restarting every n_r time steps reduces the computational cost from quadratic to linear in the number of steps, making long simulations feasible; the remaining dissipation comes from the interpolation used in the restart snapshot.

Reading between the lines

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

  • The paper's conservation tests are empirical, not proof for all times: entropy is exactly preserved only until about t≈45 in the 16^3 test and then stays under 1e-5; a long-time 6D run or a run with multiple species would be needed to see whether these small-error plateaus persist.
  • The printed formulas contain an inconsistency—eq. (30) defines ∂_t E=-j while eqs. (33)/(35) give ∂_t E=+j, eq. (52) carries an extra Δt q/m factor, and eq. (53) has sign issues on the second-order terms. If these are not typographical, the algorithms as written do not describe the Vlasov-Maxwell system; this is a flag for readers, not part of the paper's stated claims.
  • Remark 2.1 says the splitting preserves Gauss' law by construction, but the manuscript presents no numerical check of ∇·E-ρ or ∇·B for the electromagnetic benchmarks; a single run monitoring these residuals would directly test that assertion.
  • Since the restart routine uses linear interpolation, which the paper itself calls dissipative, a natural next step for long-time structure preservation is to store a compressed flow map rather than a distribution snapshot—a direction the paper points to in Remark 2.4. Such a variant would be expected to get closer to the exact-conservation behavior of the no-restart scheme.
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

4 major / 7 minor

Summary. The manuscript extends the Numerical Flow Iteration (NuFI) semi-Lagrangian scheme from the electrostatic Vlasov–Poisson system to the multi-species Vlasov–Maxwell system. Two variants are presented: NuFI-Ham, based on a first-order Lie splitting of the Hamiltonian sub-flows H_E, H_B, and H_f, and NuFI-PC, in which the same splitting is used for the Vlasov advection while Maxwell's equations are advanced by a predictor-corrector scheme. Numerical benchmarks are reported for weak Landau damping, the streaming Weibel instability, the filamentation instability, and for conservation of energy and entropy. The central claim is that the structure-preserving, memory-slim properties of electro-static NuFI carry over to the electromagnetic case. The numerical results are plausible, but the printed derivation contains sign and factor errors in the core update formulas, so the algorithm as written is not reproducible from the text alone.

Significance. The underlying idea is genuinely interesting: NuFI stores a low-dimensional field history instead of the full phase-space distribution, and its exact characteristic tracing gives sub-grid resolution and good conservation properties. A Hamiltonian-splitting extension to Vlasov–Maxwell is a natural and useful step, and the paper provides a first proof-of-concept with comparisons to analytic growth rates and to reference simulations by Cheng et al. and Kormann et al. No parameters are fitted to the target results, which is a strength. However, the contribution is not self-contained as submitted: the one-step map in Section 2.3 is not uniquely determined because of internal inconsistencies, and the implementation is cited only via an unversioned GitHub branch. The paper can be made acceptable after the derivation is corrected and the numerical experiments are repeated with the corrected formulas.

major comments (4)
  1. [§2.2.3, Eq. (30) vs. (33)/(35)] The exact H_f flow is inconsistent with its own governing equation. Eq. (30) states ∂_t E = -j, so the exact solution must be E(t,x) = E_0(x) - ∫_0^t j(s,x) ds. Equations (33) and (35) instead give E(t,x) = E_0(x) + ∫_0^t j(s,x) ds. The later discrete formulas (39)–(40) use the correct minus sign. Since (33)/(35) define the analytic sub-flow used in the splitting, this sign error must be corrected; as printed the derivation is internally contradictory.
  2. [§2.3, Eq. (52)] The spatial argument of the composed flow contains an extra factor Δt q/m. Composing (37), (43), and (49) gives f_3(x,v) = f_0( x - Δt exp(-q/m Δt J_{B0})(v - Δt q/m (E_0 - ĵ + Δt ∇×B_0)), exp(-q/m Δt J_{B0})(v - ...)). The printed Eq. (52) has x - Δt( Δt q/m exp(...)(v - ...)), i.e. a Δt² q/m prefactor inserted before the rotation. This extra factor changes the characteristic shift and would corrupt the phase-space map. It should be removed.
  3. [§2.3, Eq. (53)] The two second-order terms in the B_update have reversed signs. Starting from B_3 = B_2 - Δt ∇×E_2 with E_2 = E_0 - ĵ + Δt ∇×B_0 one obtains B_3 = B_0 - Δt ∇×E_0 + Δt² ∇×ĵ - Δt² ∇×(∇×B_0). The printed Eq. (53) has -Δt² ∇×ĵ + Δt² ∇×∇×B_0. Algorithm 1 explicitly calls Eq. (53) for the B update, and the Weibel benchmark is driven by B; hence the reported growth rates cannot be reproduced from the manuscript unless the code uses a different, uncorrected-to-text formula. If these are typographical, the corrected formulas and a versioned implementation link must be supplied.
  4. [§2.3–2.4 and Algorithm 1] The single-step numerical map is not uniquely determined by the text even after fixing the algebraic typos listed above. Algorithm 1 states that ĵ(t_n) is evaluated 'using (56) to evaluate f(t_n) as well as (41)', but (56) defines f(t_{n+1}) via the full composition of all previous maps; the algorithm does not specify how the histories of E, B, and ĵ are stored and queried inside that composition. This is a reproducibility gap independent of the sign errors. The pseudo-code should be expanded so that the dependence on stored field histories is explicit.
minor comments (7)
  1. [Algorithm 1] The line 'Evaluate E(t_n) on spatial grid using (53)' should read 'Evaluate B(t_n) ... using (53)'.
  2. [§2.4, Eq. (65)] The definition of x̃ is missing the time step: it should be x̃ = x - Δt v, not x̃ = x - Δv.
  3. [Fig. 3 caption] The caption says the simulation uses '16^3 and Δt=1/10', but the text states NuFI-Ham used Δt=1/200 and NuFI-PC used Δt=1/10. The caption should distinguish the two time steps.
  4. [Fig. 4 caption] The text says the distribution is shown at t=50, 100, and 300, but panel (f) is labelled t=1000 and the caption lists times that are not all consistent with the text. Please align the caption with the actual panels.
  5. [§3.2, Figure 3 and surrounding text] The claim that entropy is 'exactly preserved by NuFI-Ham until t≈45' is stronger than the reported data, which show nonzero errors below a threshold. It would be clearer to say 'preserved up to the quadrature and round-off errors'.
  6. [References and code availability] The GitHub repository is cited only as a branch name without a commit hash or release tag. Since the printed formulas are inconsistent, a versioned artifact is essential for disambiguation. Reference [71] is also missing a DOI.
  7. [§2.4 title and notation] The section title 'Combing NuFI with a Predictor-Corrector Maxwell update' contains a typo ('Combing' → 'Combining'), and the operator order in Eq. (59) should be explained more carefully relative to Eq. (36) to avoid confusion about which sub-flow is applied first.

Circularity Check

0 steps flagged · score 1.0 of 10

No circularity: the electromagnetic NuFI update is derived from explicit Hamiltonian sub-flows and validated against external analytic growth rates and independent reference solvers; self-citations concern prior parameter choices, not the central derivation.

full rationale

The paper's derivation chain is self-contained with respect to the quantities it predicts. The NuFI-Ham one-step map is obtained by composing the analytic solutions of the three Hamiltonian sub-systems H_E, H_B, H_f (Eqs. (16)-(34)) in a fixed Lie-splitting order, giving explicit field updates (53)-(54) and a characteristic map (52). No parameter in these formulas is fitted to the Landau, Weibel, or filamentation benchmarks. The validation targets are external and quantitative: the weak Landau damping rate gamma=-0.15336, the Weibel growth rate gamma=0.03, and the filamentation growth rate gamma about 0.38 from the Bret cold-beam formula in Appendix B. The Weibel results are also compared against independent reference simulations by Cheng et al. and Kormann et al. The paper does not define its predicted quantities in terms of the stored electric/magnetic fields, nor does it rename a fitted parameter as a prediction. Self-citations appear mainly where the authors extend their own prior electrostatic NuFI work, justify the midpoint current quadrature ('following the choice from our previous work', Remark 2.3), and choose restart parameters ('explored in previous work', Section 3). These are methodological choices rather than load-bearing uniqueness claims, and the central electromagnetic extension is independently checkable against the cited Crouseilles Hamiltonian splitting and external growth rates. The conservation tests for entropy and energy are consistent with the designed measure-preserving composition of exact sub-flows, so reporting that preservation is a validation of the implementation, not a circular use of the target result. The internal sign inconsistencies and apparent typos in Eqs. (30)/(33)/(35), (52), and (53) are correctness/reproducibility concerns, not evidence that any step reduces to its own inputs. No circular step satisfying the quoted-reduction standard was found; the minor self-citations do not raise the score beyond 1.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The derivation rests on standard Hamiltonian mechanics and splitting theory; no new physical entities are introduced. The main burden is the correctness of the splitting algebra, which is where the printed equations contain errors (cf. red flags).

free parameters (2)
  • restart interval n_r^t = 100 (Weibel) / not specified (others)
    Chosen based on the authors' prior work [4,69]; affects long-time behavior and conservation, but is not fitted to the benchmark results.
  • velocity truncation boundaries (v_min, v_max) = e.g., [-5,5], [-0.5,0.5]x[-1.2,1.2], [-1,1]x[-1.2,1.2] per test
    Hand-chosen domain truncations; standard practice, not fitted to the target growth rates.
assumptions (5)
  • domain assumption The Vlasov–Maxwell system is Hamiltonian with H = H_E + H_B + H_f (eq. 15)
    Basis for the splitting; cited from Morrison [2].
  • standard math The sub-Hamiltonian flows have explicit solutions (19)–(34)
    Derived from linear transport equations; requires smoothness of fields.
  • domain assumption Lie splitting (36)/(59) is first-order accurate and the composed exact sub-flows approximate the exact Vlasov–Maxwell flow
    Standard operator-splitting result; from Crouseilles et al. [66].
  • domain assumption Gauss's laws ∇·B=0 and ∇·E=ρ are preserved by construction (Remark 2.1)
    Load-bearing: ensures the Maxwell subsystem remains consistent; not numerically verified for the electromagnetic tests.
  • domain assumption Midpoint quadrature (41) is sufficiently accurate for smooth initial data
    Relies on smoothness of f; the authors note an adaptive rule for more complex cases [4].

how reviews work

0 comments
Cite this review

Pith. "Pith review of Extending the Numerical Flow Iteration to the multi-species Vlasov-Maxwell system through Hamiltonian Splitting." pith.science (2026). https://pith.science/paper/JQFND5EQ

@misc{pith2026251111322,
  author       = {Pith},
  title        = {Pith review of: Extending the Numerical Flow Iteration to the multi-species Vlasov-Maxwell system through Hamiltonian Splitting},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JQFND5EQ}},
  note         = {Machine review of arXiv:2511.11322}
}
read the original abstract

The Numerical Flow Iteration (NuFI) method has recently been proposed as a memory-slim while accurate in phase-space method for the electro-static Vlasov--Poisson system. It stores the temporal evolution of the electric field, instead of the distribution functions, and reconstructs the solution in each time step by following the characteristics backwards in time and reconstructing the solution from the initial distribution. NuFI has been shown to be more accurate than other state-of-the-art electro-static Vlasov solvers given the same amount of degrees of freedom. In this paper, we build on the Hamiltonian structure of the full Vlasov--Maxwell system to extend NuFI to handle electro-magnetic kinetic plasma dynamics. We show that the structure-preserving properties of the NuFI time-stepping are preserved when extending to the electro-magnetic case. Furthermore we discuss how NuFI can be incorporated into existing Semi-Lagrangian codes as an efficient while accurate subcycling technique.

Figures

Figures reproduced from arXiv: 2511.11322 by the authors.

Figure 1
Figure 1. Comparison between NuFI-Ham, NuFI-PC and a reference solution (electro-static [PITH_FULL_IMAGE:figures/full_fig_p014_1.png] view at source ↗
Figure 2
Figure 2. Evolution of the energies associated to the components of the electric and magnetic [PITH_FULL_IMAGE:figures/full_fig_p015_2.png] view at source ↗
Figure 3
Figure 3. Conservation of total energy and entropy over time for the [PITH_FULL_IMAGE:figures/full_fig_p016_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Cross-sections of the velocity distribution function [PITH_FULL_IMAGE:figures/full_fig_p017_4.png]
Figure 5
Figure 5. Figure 5: Velocity distribution function f for the streaming Weibel instability close to the left boundary (x = 0.05, left gure) and in the middle of the spatial domain (right gure) at the time t = 100. 17 [PITH_FULL_IMAGE:figures/full_fig_p017_5.png]
Figure 6
Figure 6. Figure 6: In this gure we look into the velocity distribution function f for the streaming Weibel instability at x = 0.05 and time t = 100 while zooming increasingly in. The method is able to clearly resolve sub-grid structures far beyond the common capability of the linear int…
Figure 7
Figure 7. Figure 7: Evolution of electric and magnetic energy over time for a simulation of the [PITH_FULL_IMAGE:figures/full_fig_p019_7.png]
Figure 8
Figure 8. Figure 8: Cross-section of distribution functions from the simulation of the [PITH_FULL_IMAGE:figures/full_fig_p020_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

75 extracted references · 24 canonical work pages

  1. [1]

    Chen.Introduction to Plasma Physics and Controlled Fusion

    F. Chen.Introduction to Plasma Physics and Controlled Fusion. Springer Cham, 2015.isbn: 978-3-319- 22309-4.doi:10.1007/978-3-319-22309-4

    doi:10.1007/978-3-319-22309-4 2015

  2. [2]

    P. J. Morrison.The Maxwell-Vlasov equations as a continuous hamiltonian system. In:Physics Letters A 80.5 (1980), pp. 383–386.issn: 0375-9601.doi:10.1016/0375-9601(80)90776-8

    work page doi:10.1016/0375-9601(80)90776-8 1980

  3. [3]

    Verscharen, K

    D. Verscharen, K. Klein and B. Maruca.The multi-scale nature of the solar wind. In:Living reviews in solar physics16.1 (2019).doi:10.1007/s41116-019-0021-0

    doi:10.1007/s41116-019-0021-0 2019

  4. [4]

    Wilhelm and M

    R.-P. Wilhelm and M. Torrilhon.Simulation of multi-species kinetic instabilities with the Numerical Flow Iteration. In:Accepted for publication in Proceedings of the 33rd International Symposium on Rareed Gas Dynamics(2025)

    2025

  5. [5]

    Verscharen et al.A Case for Electron-Astrophysics

    D. Verscharen et al.A Case for Electron-Astrophysics. In:Experimental Astronomy54.2 (2022).doi: 10.1007/s10686-021-09761-5. 21

    work page doi:10.1007/s10686-021-09761-5 2022

  6. [6]

    Lapenta et al.Do We Need to Consider Electrons’ Kinetic Eects to Properly Model a Planetary Magnetosphere: The Case of Mercury

    G. Lapenta et al.Do We Need to Consider Electrons’ Kinetic Eects to Properly Model a Planetary Magnetosphere: The Case of Mercury. In:Journal of Geophysical Research: Space Physics127.4 (2022). e2021JA030241 2021JA030241, e2021JA030241.doi:10.1029/2021JA030241

    doi:10.1029/2021ja030241 2022

  7. [7]

    Arber and R

    T. Arber and R. Vann.A Critical Comparison of Eulerian-Grid-Based Vlasov Solvers. In:Journal of Computational Physics180.1 (2002), pp. 339–357.issn: 0021–9991.doi:10.1006/jcph.2002.7098

    arXiv 2002

  8. [8]

    Jenko et al.Electron temperature gradient driven turbulence

    F. Jenko et al.Electron temperature gradient driven turbulence. In:Physics of Plasmas7.5 (May 2000), pp. 1904–1910.issn: 1070-664X.doi:10.1063/1.874014

    doi:10.1063/1.874014 2000

Show all 75 references
  1. [9]

    G¨ orler et al.The global version of the gyrokinetic turbulence code GENE

    T. G¨ orler et al.The global version of the gyrokinetic turbulence code GENE. In:Journal of Computational Physics230.18 (2011), pp. 7053–7071.issn: 0021-9991.doi:https://doi.org/10.1016/j.jcp.2011.05.034

    2011 doi

  2. [10]

    Juno et al.Discontinuous Galerkin algorithms for fully kinetic plasmas

    J. Juno et al.Discontinuous Galerkin algorithms for fully kinetic plasmas. In:Journal of Computational Physics353 (2018), pp. 110–147.issn: 0021-9991.doi:10.1016/j.jcp.2017.10.009

    2018 doi

  3. [11]

    N. R. Mandell et al.Electromagnetic full-fgyrokinetics in the tokamak edge with discontinuous Galerkin methods. In:Journal of Plasma Physics86.1 (2020), p. 905860109.doi:10.1017/S0022377820000070

    2020 doi

  4. [12]

    Fijalkow.A numerical solution to the Vlasov equation

    E. Fijalkow.A numerical solution to the Vlasov equation. In:Computer Physics Communications116.2 (Feb. 1999), pp. 319–328.issn: 0010–4655.doi:10.1016/S0010-4655(98)00146-5

    1999 doi

  5. [13]

    Filbet, E

    F. Filbet, E. Sonnendr¨ ucker and P. Bertrand.Conservative Numerical Schemes for the Vlasov Equation. In:Journal of Computational Physics172.1 (Sept. 2001), pp. 166–187.issn: 0021–9991.doi:10.1006/jcph. 2001.6818

    2001

  6. [14]

    Wen and M

    J. Wen and M. Nazarov.An anisotropic nonlinear stabilization fornite element approximation of Vlasov–Poisson equations. In:Journal of Computational Physics536 (2025), p. 114079.issn: 0021-9991.doi:10.1016/j. jcp.2025.114079

    2025

  7. [15]

    Kiechle, E

    Y. Kiechle, E. Chudzik and C. Helzel.A positivity-preserving Active Flux method for the Vlasov-Poisson system. In:Journal of Computational Physics524 (2025), p. 113693.issn: 0021-9991.doi:10.1016/j.jcp. 2024.113693

    2025

  8. [16]

    Kormann, M

    K. Kormann, M. Nazarov and J. Wen.A structure-preservingnite element framework for the Vlasov–Maxwell system. In:Computer Methods in Applied Mechanics and Engineering446 (2025), p. 118290.issn: 0045- 7825.doi:10.1016/j.cma.2025.118290

    2025

  9. [17]

    Cheng and G

    C. Cheng and G. Knorr.The integration of the Vlasov equation in conguration space. In:Journal of Computational Physics22.3 (Nov. 1976), pp. 330–351.issn: 0021–9991.doi:10.1016/0021-9991(76)90053- X

    1976 doi

  10. [18]

    L. P. Riishøjgaard et al.The Use of Spline Interpolation in Semi-Lagrangian Transport Models. In:Monthly Weather Review126.7 (1998), pp. 2008 –2016.doi:10.1175/1520-0493(1998)126<2008:TUOSII>2.0.CO;2

    1998 doi

  11. [19]

    Kormann, K

    K. Kormann, K. Reuter and M. Rampp.A massively parallel semi-Lagrangian solver for the six-dimensional Vlasov–Poisson equation. In:The International Journal of High Performance Computing Applications33.5 (2019), pp. 924–947.doi:10.1177/1094342019834644

    2019 doi

  12. [20]

    Einkemmer.High performance computing aspects of a dimension independent semi-Lagrangian discon- tinuous Galerkin code

    L. Einkemmer.High performance computing aspects of a dimension independent semi-Lagrangian discon- tinuous Galerkin code. In:Comput. Phys. Commun.202 (2016), pp. 326–336

    2016

  13. [21]

    Crouseilles, M

    N. Crouseilles, M. Mehrenberger and E. Sonnendr¨ ucker.Conservative semi-Lagrangian schemes for Vlasov equations. In:Journal of Computational Physics229.6 (2010), pp. 1927–1953.issn: 0021-9991.doi:10. 1016/j.jcp.2009.11.007

    2010

  14. [22]

    Cottet et al.High order semi-Lagrangian particle methods for transport equations: numerical anal- ysis and implementation issues

    G.-H. Cottet et al.High order semi-Lagrangian particle methods for transport equations: numerical anal- ysis and implementation issues. In:ESAIM: Mathematical Modelling and Numerical Analysis48.4 (2014), 1029–1060.doi:10.1051/m2an/2014009

    2014

  15. [23]

    Cottet.Semi-Lagrangian particle methods for high-dimensional Vlasov–Poisson systems

    G.-H. Cottet.Semi-Lagrangian particle methods for high-dimensional Vlasov–Poisson systems. In:Journal of Computational Physics365 (2018), pp. 362–375.issn: 0021–9991.doi:10.1016/j.jcp.2018.03.042

    2018 doi

  16. [24]

    J. A. Rossmanith and D. C. Seal.A positivity-preserving high-order semi-Lagrangian discontinuous Galerkin scheme for the Vlasov–Poisson equations. In:Journal of Computational Physics230.16 (2011), pp. 6203– 6232.issn: 0021–9991.doi:doi.org/10.1016/j.jcp.2011.04.018

    2011 doi

  17. [25]

    Ganse et al.The Vlasiator 5.2 ionosphere – coupling a magnetospheric hybrid-Vlasov simulation with a height-integrated ionosphere model

    U. Ganse et al.The Vlasiator 5.2 ionosphere – coupling a magnetospheric hybrid-Vlasov simulation with a height-integrated ionosphere model. In:Geoscientic Model Development18.2 (2025), pp. 511–527.doi: 10.5194/gmd-18-511-2025. 22

    2025 doi

  18. [26]

    von Alfthan et al.Vlasiator: First global hybrid-Vlasov simulations of Earth’s foreshock and magne- tosheath

    S. von Alfthan et al.Vlasiator: First global hybrid-Vlasov simulations of Earth’s foreshock and magne- tosheath. In:Journal of Atmospheric and Solar-Terrestrial Physics120 (2014), pp. 24–35.issn: 1364-6826. doi:https://doi.org/10.1016/j.jastp.2014.08.012

    2014 doi

  19. [27]

    Schmitz and R

    H. Schmitz and R. Grauer.Kinetic Vlasov simulations of collisionless magnetic reconnection. In:Physics of Plasmas13.9 (Sept. 2006), p. 092309.issn: 1070-664X.doi:10.1063/1.2347101

    2006 doi

  20. [28]

    Allmann-Rahn et al.The muphyII code: Multiphysics plasma simulation on large HPC systems

    F. Allmann-Rahn et al.The muphyII code: Multiphysics plasma simulation on large HPC systems. In: Computer Physics Communications296 (2024), p. 109064.issn: 0010-4655.doi:10 . 1016 / j . cpc . 2023 . 109064

    2024

  21. [29]

    Einkemmer.A performance comparison of semi-Lagrangian discontinuous Galerkin and spline based Vlasov solvers in four dimensions

    L. Einkemmer.A performance comparison of semi-Lagrangian discontinuous Galerkin and spline based Vlasov solvers in four dimensions. In:Journal of Computational Physics376 (2019), pp. 937–951.issn: 0021-9991.doi:10.1016/j.jcp.2018.10.012

    2019 doi

  22. [30]

    Einkemmer and A

    L. Einkemmer and A. Moriggl.Semi-Lagrangian 4d, 5d, and 6d kinetic plasma simulation on large-scale GPU-equipped supercomputers. In:The International Journal of High Performance Computing Applica- tions37.2 (2023), pp. 180–196.doi:10.1177/10943420221137599

    2023 doi

  23. [31]

    Schild et al.A performance portable implementation of the semi-Lagrangian algorithm in six dimen- sions

    N. Schild et al.A performance portable implementation of the semi-Lagrangian algorithm in six dimen- sions. In:Comput. Phys. Commun.295 (2024), p. 108973.doi:https://doi.org/10.1016/j.cpc.2023. 108973

    2024 doi

  24. [32]

    Kormann and E

    K. Kormann and E. Sonnendr¨ ucker.Sparse Grids for the Vlasov–Poisson Equation. In:Sparse Grids and Applications - Stuttgart 2014. Ed. by J. Garcke and D. P¨ uger. Cham: Springer International Publishing, 2016, pp. 163–190.isbn: 978-3-319-28262-6

    2014

  25. [33]

    Guo and Y

    W. Guo and Y. Cheng.A Sparse Grid Discontinuous Galerkin Method for High-Dimensional Transport Equations and Its Application to Kinetic Simulations. In:SIAM Journal on Scientic Computing38.6 (2016), A3381–A3409.doi:10.1137/16M1060017

    2016 doi

  26. [34]

    Besse and E

    N. Besse and E. Sonnendr¨ ucker.Semi-Lagrangian schemes for the Vlasov equation on an unstructured mesh of phase space. In:Journal of Computational Physics191.2 (2003-11), pp. 341–376.issn: 0021–9991.doi: 10.1016/S0021-9991(03)00318-8

    2003 doi

  27. [35]

    Pfau-Kempf et al.On the Importance of Spatial and Velocity Resolution in the Hybrid-Vlasov Modeling of Collisionless Shocks

    Y. Pfau-Kempf et al.On the Importance of Spatial and Velocity Resolution in the Hybrid-Vlasov Modeling of Collisionless Shocks. In:Frontiers in PhysicsVolume 6 - 2018 (2018).issn: 2296-424X.doi:10.3389/ fphy.2018.00044

    2018

  28. [36]

    Palmroth et al.Vlasov methods in space physics and astrophysics

    M. Palmroth et al.Vlasov methods in space physics and astrophysics. In:Living Reviews in Computational Astrophysics4.1 (2018).doi:10.1007/s41115-018-0003-2

    2018 doi

  29. [37]

    Cottet and P.-A

    G.-H. Cottet and P.-A. Raviart.Particle Methods for the One-Dimensional Vlasov–Poisson Equations. In: SIAM Journal on Numerical Analysis21.1 (June 1984), pp. 52–76.issn: 0036–1429.doi:10.1137/0721003

    1984 doi

  30. [38]

    Hockney and J

    R. Hockney and J. Eastwood.Computer Simulation Using Particles. CRC Press, Mar. 2021.isbn: 978-1- 4398-2205-0

    2021

  31. [39]

    Birdsall and A

    C. Birdsall and A. Langdon.Plasma Physics via Computer Simulation. Series in Plasma Physics. Taylor & Francis, 2004.isbn: 978-1-4822-6306-0

    2004

  32. [40]

    R. P. Wilhelm and M. Kirchhart.An interpolating particle method for the Vlasov–Poisson equation. In: Journal of Computational Physics473 (2023), p. 111720.issn: 0021-9991.doi:10.1016/j.jcp.2022.111720

    2023

  33. [41]

    Evans and F

    M. Evans and F. Harlow.The Particle-In-Cell method for Hydrodynamic Calculations. In:Report LA-2139, Los Alamos Scientic laboratory of the university of California(1957)

    1957

  34. [42]

    G. H. Cottet and P. A. Raviart.On particle-in-cell methods for the Vlasov-Poisson equations. In:Transport Theory and Statistical Physics15.1-2 (1986), pp. 1–31.doi:10.1080/00411458608210442

    1986 doi

  35. [43]

    B. Wang, G. H. Miller and P. Colella.A Particle-in-cell Method with Adaptive Phase-space Remapping for Kinetic Plasmas. In:SIAM J. Sci. Comput.33.6 (Dec. 2011), pp. 3509–3537.doi:10.1137/100811805

    2011 doi

  36. [44]

    Myers, P

    A. Myers, P. Colella and B. Straalen.A 4th-Order Particle-in-Cell Method with Phase-Space Remapping for the Vlasov–Poisson Equation. In:SIAM Journal on Scientic Computing39.9 (Feb. 2016), B467–B485. issn: 1095–7197.doi:10.1137/16M105962X

    2016 doi

  37. [45]

    Ameres.Stochastic and Spectral Particle Methods for Plasma Physics

    J. Ameres.Stochastic and Spectral Particle Methods for Plasma Physics. PhD thesis. Technische Univer- sit¨ at M¨ unchen, 2018. 23

    2018

  38. [46]

    Kraus et al.GEMPIC: geometric electromagnetic particle-in-cell methods

    M. Kraus et al.GEMPIC: geometric electromagnetic particle-in-cell methods. In:Journal of Plasma Physics83.4 (2017), p. 905830401.doi:10.1017/S002237781700040X

    2017 doi

  39. [47]

    Burau et al.PIConGPU: A Fully Relativistic Particle-in-Cell Code for a GPU Cluster

    H. Burau et al.PIConGPU: A Fully Relativistic Particle-in-Cell Code for a GPU Cluster. In:IEEE Transactions on Plasma Science38.10 (2010), pp. 2831–2839.doi:10.1109/TPS.2010.2064310

    2010

  40. [48]

    Fasoulas et al.Combining particle-in-cell and direct simulation Monte Carlo for the simulation of reactive plasmaows

    S. Fasoulas et al.Combining particle-in-cell and direct simulation Monte Carlo for the simulation of reactive plasmaows. In:Physics of Fluids31.7 (July 2019), p. 072006.issn: 1070-6631.doi:10.1063/1.5097638

    2019 doi

  41. [49]

    Derouillat et al.Smilei : A collaborative, open-source, multi-purpose particle-in-cell code for plasma simulation

    J. Derouillat et al.Smilei : A collaborative, open-source, multi-purpose particle-in-cell code for plasma simulation. In:Computer Physics Communications222 (2018), pp. 351–373.issn: 0010-4655.doi:10 . 1016/j.cpc.2017.09.024

    2018

  42. [50]

    L. Fedeli et al.Pushing the Frontier in the Design of Laser-Based Electron Accelerators with Ground- breaking Mesh-Rened Particle-In-Cell Simulations on Exascale-Class Supercomputers. In:SC22: Interna- tional Conference for High Performance Computing, Networking, Storage an...

    2022 arXiv

  43. [51]

    Parodi et al.Pantera: A PIC-MCC-DSMC software for the simulation of rareed gases and plasmas

    P. Parodi et al.Pantera: A PIC-MCC-DSMC software for the simulation of rareed gases and plasmas. In:SoftwareX31 (2025), p. 102244.issn: 2352-7110.doi:10.1016/j.softx.2025.102244

    2025

  44. [52]

    Brackbill and D

    J. Brackbill and D. Forslund.An implicit method for electromagnetic plasma simulation in two dimensions. In:Journal of Computational Physics46.2 (1982), pp. 271–308.issn: 0021-9991.doi:10 . 1016 / 0021 - 9991(82)90016-X

    1982

  45. [53]

    R. J. Mason.Implicit moment particle simulation of plasmas. In:Journal of Computational Physics41.2 (1981), pp. 233–244.issn: 0021-9991.doi:10.1016/0021-9991(81)90094-2

    1981 doi

  46. [54]

    Lapenta, J

    G. Lapenta, J. Brackbill and P. Ricci.Kinetic approach to microscopic-macroscopic coupling in space and laboratory plasmas. In:Physics of Plasmas13.5 (2006).doi:10.1063/1.2173623

    2006 doi

  47. [55]

    Markidis, G

    S. Markidis, G. Lapenta and Rizwan-uddin.Multi-scale simulations of plasma with iPIC3D. In:Math- ematics and Computers in Simulation80.7 (2010). Multiscale modeling of moving interfaces in materials, pp. 1509–1519.issn: 0378-4754.doi:10.1016/j.matcom.2009.08.038

    2010 doi

  48. [56]

    Lapenta.Exactly energy conserving semi-implicit particle in cell formulation

    G. Lapenta.Exactly energy conserving semi-implicit particle in cell formulation. In:Journal of Computa- tional Physics334 (2017), pp. 349–366.issn: 0021-9991.doi:10.1016/j.jcp.2017.01.002

    2017 doi

  49. [57]

    Bacchini.RelSIM: A Relativistic Semi-implicit Method for Particle-in-cell Simulations

    F. Bacchini.RelSIM: A Relativistic Semi-implicit Method for Particle-in-cell Simulations. In:The Astro- physical Journal Supplement Series268.2 (2023).doi:10.3847/1538-4365/acefba

    2023 doi

  50. [58]

    Arshad, Y

    T. Arshad, Y. Chen and G. T´ oth.Adaptive mesh renement in semi-implicit particle-in-cell method. In: Computer Physics Communications316 (2025), p. 109806.issn: 0010-4655.doi:10 . 1016 / j . cpc . 2025 . 109806

    2025

  51. [59]

    Bacchini, J

    F. Bacchini, J. Amaya and G. Lapenta.The relativistic implicit Particle-in-Cell method. In:Journal of Physics: Conference Series1225.1 (2019), p. 012011.doi:10.1088/1742-6596/1225/1/012011

    2019 doi

  52. [60]

    Croonen et al.An Exactly Energy-conserving Electromagnetic Particle-in-cell Method in Curvilinear Coordinates

    J. Croonen et al.An Exactly Energy-conserving Electromagnetic Particle-in-cell Method in Curvilinear Coordinates. In:The Astrophysical Journal Supplement Series271.2 (2024), p. 63.doi:10.3847/1538- 4365/ad31a3

    2024 doi

  53. [61]

    Filbet and E

    F. Filbet and E. Sonnendr¨ ucker.Numerical methods for the Vlasov equation. In:Numerical Mathematics and Advanced Applications. Milano: Springer Milan, 2003, pp. 459–468.isbn: 978-88-470-2089-4

    2003

  54. [62]

    Krah et al.A Characteristic Mapping Method for Vlasov-Poisson with Extreme Resolution Properties

    P. Krah et al.A Characteristic Mapping Method for Vlasov-Poisson with Extreme Resolution Properties. In:Communications in Computational Physics35.4 (2023), 905 – 937.doi:10.4208/cicp.OA-2024-0012

    2023 doi

  55. [63]

    Kirchhart and R

    M. Kirchhart and R. P. Wilhelm.The Numerical Flow Iteration for the Vlasov–Poisson Equation. In: SIAM Journal on Scientic Computing46.3 (2024), A1972–A1997.doi:10.1137/23M154710X

    2024 doi

  56. [64]

    Wilhelm et al.Highdelity simulations of the multi-species Vlasov equation in the electro-static, collisional-less limit

    R.-P. Wilhelm et al.Highdelity simulations of the multi-species Vlasov equation in the electro-static, collisional-less limit. In:Plasma Physics and Controlled Fusion67.2 (2025), p. 025011.doi:10.1088/1361- 6587/ad9fdb

    2025 doi

  57. [65]

    J. E. Marsden and A. Weinstein.The Hamiltonian structure of the Maxwell-Vlasov equations. In:Physica D: Nonlinear Phenomena4.3 (1982), pp. 394–406.issn: 0167-2789.doi:10.1016/0167-2789(82)90043-4

    1982 doi

  58. [66]

    Crouseilles, L

    N. Crouseilles, L. Einkemmer and E. Faou.Hamiltonian splitting for the Vlasov–Maxwell equations. In: Journal of Computational Physics283 (2015), pp. 224–240.issn: 0021-9991.doi:10.1016/j.jcp.2014.11. 029. 24

    2015 doi

  59. [67]

    Hairer, G

    E. Hairer, G. Wanner and C. Lubich.Geometric Numerical Integration. Structure-Preserving Algorithms for Ordinary Dierential Equations. 2nd ed. Springer Series in Computational Mathematics 31. Springer, 2006. isbn: 3540306633

    2006

  60. [68]

    L. N. Trefethen and J. A. C. Weideman.The Exponentially Convergent Trapezoidal Rule. In:SIAM Review 56.3 (2014), pp. 385–458.doi:10.1137/130932132

    2014 doi

  61. [69]

    Wilhelm and K

    R.-P. Wilhelm and K. Kormann.Restarting the Numerical Flow Iteration through low rank tensor approx- imations. 2025.doi:10.48550/arXiv.2509.08474. arXiv:2509.08474 [math.NA].url:https://arxiv. org/abs/2509.08474

    2025 doi

  62. [70]

    Kormann.A Semi-Lagrangian Vlasov Solver in Tensor Train Format

    K. Kormann.A Semi-Lagrangian Vlasov Solver in Tensor Train Format. In:SIAM Journal on Scientic Computing37.4 (2015), B613–B632.doi:10.1137/140971270

    2015 doi

  63. [71]

    Krah et al.A Hybrid Semi-Lagrangian Flow Mapping Approach for Kinetic Electrostatic Electron Non- linear (KEEN) Waves.doi:

    P. Krah et al.A Hybrid Semi-Lagrangian Flow Mapping Approach for Kinetic Electrostatic Electron Non- linear (KEEN) Waves.doi:

  64. [72]

    Cheng et al.Discontinuous Galerkin Methods for the Vlasov–Maxwell Equations

    Y. Cheng et al.Discontinuous Galerkin Methods for the Vlasov–Maxwell Equations. In:SIAM Journal on Numerical Analysis52.2 (2014), pp. 1017–1049.doi:10.1137/130915091

    2014 doi

  65. [73]

    Bacchini and A

    F. Bacchini and A. A. Philippov.Fundamental, harmonic, and third-harmonic plasma emission from beam–plasma instabilities: arst-principles precursor for astrophysical radio bursts. In:Monthly Notices of the Royal Astronomical Society529.1 (Feb. 2024), pp. 169–177.issn: 0035-...

    2024 doi

  66. [74]

    Pezzini et al.Fully Kinetic Simulations of Proton-beam-driven Instabilities from Parker Solar Probe Observations

    L. Pezzini et al.Fully Kinetic Simulations of Proton-beam-driven Instabilities from Parker Solar Probe Observations. In:The Astrophysical Journal975.1 (2024), p. 37.doi:10.3847/1538-4357/ad7465

    2024 doi

  67. [75]

    A. Bret, L. Gremillet and M. E. Dieckmann.Multidimensional electron beam-plasma instabilities in the relativistic regime. In:Physics of Plasmas17.12 (Dec. 2010), p. 120501.issn: 1070-664X.doi:10.1063/1. 3514586. 25

    2010 doi

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.