An equivalence of moment closure and nonlinear variational approximation of the Fokker-Planck equation for dilute polymeric flow

Barbara Wohlmuth; Caroline Lasser; Stephan B. Lunowa

REVIEW 1 major objections 4 minor 24 references

An equivalence of moment closure and nonlinear variational approximation of the Fokker-Planck equation for dilute polymeric flow

T0 review · 1 major / 4 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read A Gaussian variational reduction of the Hookean Fokker–Planck equation reproduces the diffusive Oldroyd-B moment closure exactly, and in the absence of center-of-mass diffusion the Gaussian ansatz is an exact solution.

desk verdict The core equivalence is solid and genuinely new, but Lemma 5's well-posedness argument has a real gap that needs fixing before the global claims are trustworthy. read the letter →

arxiv 2602.04644 v2 pith:PB54B4ZH submitted 2026-02-04 math.SP math-phmath.MP

Caroline Lasser , Stephan B. Lunowa , Barbara Wohlmuth This is my paper · ORCID

classification math.SPmath-phmath.MP MSC 37L6558E3035Q8476D05

keywords Dirac–FrenkelvariationalprincipleFisher–RaometricGaussianapproximationmanifoldmomentclosureOldroyd-BmodelFokker–PlanckequationHookeanchaindilutepolymericflow

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 proves that, for the Hookean chain model of dilute polymeric fluids, two apparently unrelated reduction strategies—taking second moments of the Fokker–Planck equation and projecting the equation onto a manifold of Gaussian densities via the Dirac–Frenkel variational principle—produce the same macroscopic theory. The projected covariance C(t,x) of the Gaussian ansatz obeys exactly the classical diffusive Oldroyd-B closure equation, so the extra-stress tensor from the variational approximation coincides with the one from moment closure. In the absence of center-of-mass diffusion, the Gaussian approximation is an exact solution of the Fokker–Planck equation rather than merely an approximation. This matters because the variational route also yields an explicit a posteriori error representation, which the moment-closure route lacks, and it can be applied to nonlinear models like FENE where exact closures are unavailable. The paper thereby connects a practical rheological closure to a geometric variational structure.

What carries the argument

The Gaussian manifold of normalized centered Gaussian densities parameterized by block-diagonal covariance matrices, endowed with the Fisher–Rao weighted inner product (density-weighted L²). Its tangent space at f consists of functions (qᵀ A q − tr(A C)) f for symmetric block-diagonal A. The linear Hookean configurational operator maps any Gaussian f back into the tangent space, and the spatial operator decomposes into a tangent part plus a scalar remainder ερ that is orthogonal to the tangent space. The projection therefore acts only on the covariance, yielding a closed equation of motion for C(t,x).

What would settle it

Set d=1, N=1, choose a smooth, nonconstant covariance profile C(x) on a bounded domain with ε>0, and compute the scalar remainder ρ(C,∇C,∇²C) from Lemma 2. The paper's Proposition 4 requires its density-weighted inner products against the Gaussian tangent-space basis to vanish identically; if any such integral is nonzero, the projected covariance equation acquires an extra term and the claimed exact equivalence fails.

Watch

Extended reading notes

Core claim

For the linear Hookean chain model of dilute polymeric flow, Gaussian trial densities with covariance matrix C(t,x), evolved by the Dirac–Frenkel variational principle with Fisher–Rao weighting, satisfy the projected equation if and only if C obeys Eq. (20), which is precisely the diffusive Oldroyd-B closure. Consequently, the variational approximation and the classical second-moment closure predict identical macroscopic conformation tensors. When the center-of-mass diffusion coefficient vanishes, the Gaussian approximation is an exact solution of the Fokker–Planck equation.

Load-bearing premise

The proof relies on projecting the residual using the Fisher–Rao weighted inner product; if that projection were replaced by any other metric, the covariance dynamics would not generally reduce to the Oldroyd-B equation.

Editorial extensions

If this is right

For the Hookean chain, the variational Gaussian approximation and the classical second-moment closure produce identical conformation tensors, so numerical schemes based on the projected Gaussian dynamics inherit the known range of validity of diffusive Oldroyd-B.
With zero center-of-mass diffusion (ε=0), a Gaussian initial density is propagated exactly by the Fokker–Planck equation; the Gaussian manifold is invariant under the configurational dynamics.
The a posteriori error formula gives a computable residual, enabling error-controlled reduced simulations in high-dimensional configuration space.
The same abstract construction applies to general bead-spring chains with linear spring forces, replacing the Rouse matrix by the graph Laplacian of the chain.
For nonlinear forces such as FENE, exact closure is lost, but the same variational principle provides a systematic, closure-free reduced model on a chosen approximation manifold.

Reading between the lines

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

One consequence the paper leaves implicit: in the Hookean limit of any FENE-type model (maximum extensibility → ∞), a variational reduced scheme should reproduce Oldroyd-B; enforcing this limit numerically could serve as a consistency test for nonlinear reduced models.
The centrality of the Fisher–Rao metric suggests a testable question: do other information metrics (e.g., Wasserstein) yield different macroscopic closures for the same Gaussian ansatz, and if so, which one better matches micro–macro simulations of dilute polymer flows?
The residual ερf that is orthogonal to the tangent space could be used as an a posteriori indicator for when a single Gaussian ansatz is insufficient, motivating adaptive enrichment by multi-Gaussians within the same variational framework.
The equivalence is not specific to polymers: any Fokker–Planck equation with linear drift and constant diffusion, coupled to macroscopic fields only through second moments, will have the same Gaussian-variational/moment-closure dictionary.

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.

Desk Editor's Note

read the letter

The main result is real. Proposition 4 — that the Dirac–Frenkel Gaussian variational approximation gives exactly the diffusive Oldroyd-B closure for the Hookean chain — is clean. I checked the key computations: Lemma 2's tangent-space invariance, Lemma 3's Hermite orthogonality argument, and the projection step all hold up. The equivalence is a genuine unification, not a trivial restatement: the cited literature derives the closure via moments or Hermite spectral truncation, and I don't see this particular variational route stated anywhere. The paper is honest about its limits: it works only for the linear Hookean force, and the 'abstract error representation' is an exact a posteriori formula, not a concrete bound. That is fine; it is a useful starting point for nonlinear models.

The soft spot is Lemma 5. The proposed representation C(t) = Φ(t,0)C0Φ(t,0)^T + (1/De)∫Φ(t,s)ΛΦ(t,s)^T ds treats the evolution operator Φ as if it acted pointwise on matrix entries. For the actual advection-diffusion operator, Φ is an integral operator on L²(Ω); composing it with pointwise multiplication is not the same as solving the matrix PDE. The stress-test note's scalar heat-equation counterexample is on point: the product ansatz misses a Leibniz-rule cross term. This leaves the symmetric-positive-definite preservation and global existence claim unproved as written. The gap is auxiliary — it does not touch the equivalence derivation in Proposition 4 — but it is a genuine hole in a supporting lemma. The authors should repair or replace that argument, possibly by a maximum principle or a different well-posedness approach.

The choice of Fisher–Rao metric is motivated but not derived from physics; the paper is upfront about this, so I don't count it as a flaw. The citation pattern is normal, and the self-citation to [20] is legitimate.

Bottom line: this is a worthwhile paper with a correct central result and one fixable flaw. Send it to peer review; a good referee will catch the Lemma 5 issue, and the authors will need to address it, but the core contribution deserves to be in the literature.

Referee Report

1 major / 4 minor

Summary. The paper establishes an equivalence between the classical moment closure for the Hookean bead-spring Fokker–Planck equation and a nonlinear variational approximation on the manifold of Gaussian densities. The variational principle is a Dirac–Frenkel condition with the Fisher–Rao inner product, pointwise in x. The main result (Prop. 4) shows that the Gaussian covariance satisfies the diffusive Oldroyd-B equation (20), so the macroscopic conformation tensor from the variational approximation coincides with the classical closure. The paper also gives an abstract error representation (Prop. 1) and a well-posedness/SPD lemma (Lemma 5).

Significance. The equivalence is conceptually valuable: it provides a systematic, parameter-free derivation of Oldroyd-B from a nonlinear reduced model and offers a platform for constructing closures for nonlinear force laws. The proof of Prop. 4 is clean; the tangent-space computation (Lemma 2) and the Hermite orthogonality argument (Lemma 3) are elegant. The result is, however, explicitly tied to the Fisher–Rao projection; this is a modeling choice rather than a consequence of the Fokker–Planck dynamics. The supporting well-posedness argument contains a serious gap and needs repair before the paper can be accepted.

major comments (1)

[§4.2, Lemma 5] The solution formula C(t)=Φ(t,0)C0Φ(t,0)^T + (1/De)∫Φ(t,s)ΛΦ(t,s)^T ds is not valid for x-dependent coefficients. Here Φ(t,s) is an evolution operator on L²(Ω;R^{D×D}), not a pointwise matrix, so 'Φ C0 Φ^T' is an operator composition, not a matrix-valued function. The Leibniz verification ignores cross terms from the x-dependence of the coefficients. For example, when u=0 and M=0, Eq. (20) reduces to ∂tC=εΔxC+Λ/De; the proposed representation with Φ(t,s)=e^{(1/2)ε(t-s)Δ} gives the homogeneous term e^{(1/2)εtΔ}C0 (not e^{εtΔ}C0) when interpreted as operator composition on constants. Thus the SPD conclusion is unsupported. Since the Gaussian ansatz requires C(t,x) to be positive definite pointwise, this is a load-bearing gap in Section 4.2, even though Prop. 4's formal derivation is unaffected.

minor comments (4)

[Abstract / §1] Typographical errors: 'Dirac–Frankel' should be 'Dirac–Frenkel'; 'Oldroid-B' should be 'Oldroyd-B'.
[§4.1, Prop. 4 proof] The proof should state explicitly that the abstract evolution operator is L = −Lx − Lq, so the projected equation is ∂tC = −LxC − M C − C M^T + Λ/De. As written, applying Lemma 2 directly to A_q + A_x appears to give the opposite signs; the final Eq. (20) is correct, but the sign convention needs to be made transparent.
[§4.2, Lemma 5] The notation M∂(t)^T is ambiguous: it is neither the Hilbert-space adjoint of M∂ nor the usual matrix transpose. The authors should define it rigorously; this ambiguity is part of the problem with the claimed solution representation.
[§4.1, after Prop. 4] The statement 'if ε=0, the variational approximation even exactly solves the Hookean Fokker–Planck equation' should explicitly note that this holds when the initial density is Gaussian; for arbitrary initial data only the covariance dynamics are reproduced.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: Eq. (20) is derived from the variational principle, not assumed; only a minor non-load-bearing self-citation is present.

full rationale

The central claim is a genuine derivation rather than a restatement of its inputs. The paper defines the Gaussian ansatz f(t)=Φ(C(t),·) and imposes the Fisher-Rao weighted orthogonality condition (11). Lemma 2 explicitly computes the action of the configurational operator L_q and the spatial operator L_x on the Gaussian manifold, and Lemma 3 proves, via Hermite-polynomial orthogonality, that the spatial remainder ερf is orthogonal to the tangent space. Proposition 4 then obtains ∂tC+(u·∇x)C−εΔxC=(1/De)(Λ⊗I_d)−CM^T−MC (Eq. (20)), which is algebraically identical to the classical diffusive Oldroyd-B moment closure (7). Nowhere is Eq. (7) substituted into the variational principle; the equation of motion is obtained by setting the projected residual to zero. The covariance parameters of the Gaussian are indeed the second-moment coordinates used in the macroscopic closure, but this identification is not a circular reduction—the evolution of those coordinates is derived, not imposed. The only author-overlapping citation in the derivation chain is [20] in Lemma 2(i), where the paper notes its tangent-space statement is 'a special case of [20, Lemma 3.1]'; however, the lemma is proved in the text immediately before that remark, so the self-citation is not load-bearing. Other self-citations ([7], [8], [19]) concern numerical and Dirac-Frenkel background rather than the equivalence proof. The skepticism about Lemma 5's quadratic representation concerns well-posedness of the covariance equation and is a mathematical correctness issue, not circularity; it does not affect the derivation of the closure equivalence. There are no fitted parameters, no calibrated data, and no external closure equation used as an input. The score of 2 reflects the presence of a minor, non-load-bearing self-citation; the central derivation itself is self-contained.

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

The Gaussian covariance C(t,x) is a dynamical variable set by the variational equation, not a fitted parameter; no numerical data or constants are calibrated. The load-bearing postulates are the Hookean force law and, above all, the Fisher–Rao/Dirac–Frenkel projection — if that projection were replaced by another metric, the equivalence with the Oldroyd-B moment closure would not follow. These are explicit assumptions in the paper, but they are the price of the theorem.

assumptions (5)

domain assumption Polymer microdynamics are governed by the Fokker–Planck equation (3) with linear Hookean force F(q)=q and Rouse matrix R=tridiag(-1,2,-1).
Introduced in Sec. 2; the entire equivalence theorem is restricted to this model. FENE and nonlinear forces are explicitly excluded in Sec. 5.
ad hoc to paper The Dirac–Frenkel variational principle (11) with the inverse-density-weighted (Fisher–Rao) inner product is the correct way to project the Fokker–Planck dynamics onto the approximation manifold.
Sec. 3.1 introduces this projection as a modelling choice; the equivalence result holds only under this specific metric. The paper motivates it as diffeomorphism-invariant but does not derive it from physical principles.
domain assumption The approximation manifold consists of normalized Gaussians with block-diagonal positive definite covariance C (12), and the initial condition is a Gaussian in this manifold.
Sec. 4 postulates this ansatz; Prop 4 and Lemma 5 require C0∈Z SPD. The choice is motivated by the steady-state Gaussian (6), not derived.
standard math The non-autonomous advection-diffusion operator generates an invertible evolution operator Φ(t,s) on L²(Ω;R^{D×D}) under the stated regularity of u.
Invoked in Lemma 5 via Pazy Thm 4.8 and Engel–Nagel III.2; requires uniform bounds on ∇_x u and dissipativity of B(t). Not proved in this paper, and the subsequent representation is suspect.
standard math Hermite polynomial orthogonality and Jacobi's formula are used in Proposition 3 and Lemma 2.
Standard background for the orthogonality proof; no issue.

how reviews work

0 comments

Cite this review

Pith. "Pith review of An equivalence of moment closure and nonlinear variational approximation of the Fokker-Planck equation for dilute polymeric flow." pith.science (2026). https://pith.science/paper/PB54B4ZH

@misc{pith2026260204644,
  author       = {Pith},
  title        = {Pith review of: An equivalence of moment closure and nonlinear variational approximation of the Fokker-Planck equation for dilute polymeric flow},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PB54B4ZH}},
  note         = {Machine review of arXiv:2602.04644}
}

read the original abstract

We establish the equivalence between a classical moment closure and a nonlinear variational approximation of the Fokker-Planck equation for dilute polymeric flow in the linearized Hookean spring chain setting. The variational formulation is based on the Dirac-Frankel principle applied to a Gaussian approximation manifold endowed with the Fisher-Rao information metric. We show that the invariance of this manifold under the linear configurational dynamics yields an exact evolution for the macroscopic conformation tensor, recovering the classical diffusive Oldroyd-B closure. While the equivalence only holds in the linearized setting, the associated variational framework provides an abstract error representation. Thus it can serve in future work as a starting point for the systematic construction of reduced approximation schemes for polymeric flows with nonlinear forcing laws.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 6 canonical work pages

[20]

Lasser, C

C. Lasser, C. Lubich, Computing quantum dynamics in the semiclassical regime, Acta Numerica 29 (2020) 229–401.doi:10.1017/S0962492920000033

doi:10.1017/s0962492920000033 2020
[1]

R. B. Bird, C. F. Curtiss, R. C. Armstrong, O. Hassager, Dynamics of Polymeric Liquids, Vol. 2: Kinetic Theory, John Wiley & Sons, 1987

1987
[2]

A. N. Gorban, I. V. Karlin, P. Ilg, H. C. Öttinger, Corrections and enhancements of quasi-equilibrium states, J. Non-Newtonian Fluid Mech. 96 (2001) 203–219.doi:10.1016/S0377-0257(00)00135-X

work page doi:10.1016/s0377-0257(00)00135-x 2001
[3]

P. Ilg, I. V. Karlin, H. C. Öttinger, Canonical distribution functions in polymer dynamics. (I). Dilute so- lutions of flexible polymers, Physica A 315 (3) (2002) 367–385.doi:10.1016/S0378-4371(02)01017-8

work page doi:10.1016/s0378-4371(02)01017-8 2002
[4]

Y. Hyon, J. A. Carrillo, Q. Du, C. Liu, A maximum entropy principle based closure method for macro- micro models of polymeric materials, Kinetic and Related Models 1 (2) (2008) 171–184.doi:10.3934/ krm.2008.1.171. 13

2008
[5]

Dębiec, E

T. Dębiec, E. Süli, On a class of generalised solutions to the kinetic Hookean dumbbell model for incompressible dilute polymeric fluids: Existence and macroscopic closure, Arch. Rational Mech. Anal. 249 (43) (2025).doi:10.1007/s00205-025-02115-x

doi:10.1007/s00205-025-02115-x 2025
[6]

Hetland, E

B. Hetland, E. Jettestuen, A. Hiorth, Solving the constitutive equation of dilute polymeric flows: A general Fokker–Planck approach for linear elastic dumbbell models, Phys. Fluids 35 (9) (2023).doi: 10.1063/5.0161336

work page doi:10.1063/5.0161336 2023
[7]

Beddrich, E

J. Beddrich, E. Süli, B. Wohlmuth, Numerical simulation of the time-fractional Fokker-Planck equation and applications to polymeric fluids, J. Comput. Phys. 497 (2024) 112598.doi:10.1016/j.jcp.2023. 112598

doi:10.1016/j.jcp.2023 2024

Show all 24 references

[8]

Beddrich, S

J. Beddrich, S. B. Lunowa, B. Wohlmuth, Numerical simulation of dilute polymeric fluids with memory effects in the turbulent flow regime, J. Comput. Phys. 532 (2025) 113955.doi:10.1016/j.jcp.2025. 113955

2025 doi
[9]

Jordan, D

R. Jordan, D. Kinderlehrer, F. Otto, The variational formulation of the Fokker–Planck equation, SIAM J. Math. Anal. 29 (1) (1998) 1–17.doi:10.1137/S0036141096303359

1998 doi
[10]

Zhang, Y

H. Zhang, Y. Chen, E. Vanden-Eijnden, B. Peherstorfer, Sequential-in-time training of nonlinear parametrizations for solving time-dependent partial differential equations (2024).arXiv:2404.01145

2024 arXiv
[11]

Anderson, M

W. Anderson, M. Farazmand, Fisher information and shape-morphing modes for solving the Fokker- Planck equation in higher dimensions, Appl. Math. Comput. 467 (2024) 17, id/No 128489.doi: 10.1016/j.amc.2023.128489

2024
[12]

Bruna, B

J. Bruna, B. Peherstorfer, E. Vanden-Eijnden, Neural Galerkin schemes with active learning for high- dimensional evolution equations, J. Comput. Phys. 496 (2024) 22, id/No 112588.doi:10.1016/j.jcp. 2023.112588

2024
[13]

Y. Chen, D. Z. Huang, J. Huang, S. Reich, A. M. Stuart, Sampling via gradient flows in the space of probability measures (2024).arXiv:2310.03597

2024 arXiv
[14]

B. H. Zimm, Dynamics of polymer molecules in dilute solution: Viscoelasticity, flow birefringence and dielectric loss, J. Chem. Phys. 24 (2) (1956) 269–278.doi:10.1063/1.1742462

1956 doi
[15]

H. R. Warner Jr, Kinetic theory and rheology of dilute suspensions of finitely extendible dumbbells, Industrial Engrg. Chem. Fundamentals 11 (3) (1972) 379–387.doi:10.1021/i160043a017

1972 doi
[16]

Jedynak, Approximation of the inverse Langevin function revisited, Rheol

R. Jedynak, Approximation of the inverse Langevin function revisited, Rheol. Acta. 54 (2015) 29–39. doi:10.1007/s00397-014-0802-2

2015 doi
[17]

J. W. Barrett, E. Süli, Existence of global weak solutions to the kinetic Hookean dumbbell model for incompressible dilute polymeric fluids, Nonlinear Anal. Real World Appl. 39 (2018) 362–395

2018
[18]

Lubich, On variational approximations in quantum molecular dynamics, Math

C. Lubich, On variational approximations in quantum molecular dynamics, Math. Comp. 74 (250) (2005) 765–779.doi:10.1090/S0025-5718-04-01685-0

2005 doi
[19]

Lasser, C

C. Lasser, C. Su, Various variational approximations of quantum dynamics, J. Math. Phys. 63 (7) (2022) Paper No. 072107, 22.doi:10.1063/5.0088265

2022 doi
[21]

Arendt, Chp

W. Arendt, Chp. 1: Semigroups and evolution equations: Functional calculus, regularity and kernel estimates, in: Handbook of Differential Equations: Evolutionary Equations, Vol. 1, Elsevier/North- Holland, 2002, pp. 1–85.doi:10.1016/S1874-5717(04)80003-3. 14

2002 doi
[22]

Gilbarg, N

D. Gilbarg, N. S. Trudinger, Elliptic partial differential equations of second order, reprint of the 1998 ed. Edition, Class. Math., Springer, 2001.doi:10.1007/978-3-642-61798-0

1998 doi
[23]

Engel, R

K.-J. Engel, R. Nagel, One-parameter semigroups for linear evolution equations, Vol. 194 of Grad. Texts Math., Springer, 2000.doi:10.1007/b97696

doi
[24]

Pazy, Semigroups of linear operators and applications to partial differential equations, Vol

A. Pazy, Semigroups of linear operators and applications to partial differential equations, Vol. 44 of Appl. Math. Sci., Springer, 1983.doi:10.1007/978-1-4612-5561-1. 15

1983 doi

Pith tools

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

[1] [20]

Lasser, C

C. Lasser, C. Lubich, Computing quantum dynamics in the semiclassical regime, Acta Numerica 29 (2020) 229–401.doi:10.1017/S0962492920000033

doi:10.1017/s0962492920000033 2020

[2] [1]

R. B. Bird, C. F. Curtiss, R. C. Armstrong, O. Hassager, Dynamics of Polymeric Liquids, Vol. 2: Kinetic Theory, John Wiley & Sons, 1987

1987

[3] [2]

A. N. Gorban, I. V. Karlin, P. Ilg, H. C. Öttinger, Corrections and enhancements of quasi-equilibrium states, J. Non-Newtonian Fluid Mech. 96 (2001) 203–219.doi:10.1016/S0377-0257(00)00135-X

work page doi:10.1016/s0377-0257(00)00135-x 2001

[4] [3]

P. Ilg, I. V. Karlin, H. C. Öttinger, Canonical distribution functions in polymer dynamics. (I). Dilute so- lutions of flexible polymers, Physica A 315 (3) (2002) 367–385.doi:10.1016/S0378-4371(02)01017-8

work page doi:10.1016/s0378-4371(02)01017-8 2002

[5] [4]

Y. Hyon, J. A. Carrillo, Q. Du, C. Liu, A maximum entropy principle based closure method for macro- micro models of polymeric materials, Kinetic and Related Models 1 (2) (2008) 171–184.doi:10.3934/ krm.2008.1.171. 13

2008

[6] [5]

Dębiec, E

T. Dębiec, E. Süli, On a class of generalised solutions to the kinetic Hookean dumbbell model for incompressible dilute polymeric fluids: Existence and macroscopic closure, Arch. Rational Mech. Anal. 249 (43) (2025).doi:10.1007/s00205-025-02115-x

doi:10.1007/s00205-025-02115-x 2025

[7] [6]

Hetland, E

B. Hetland, E. Jettestuen, A. Hiorth, Solving the constitutive equation of dilute polymeric flows: A general Fokker–Planck approach for linear elastic dumbbell models, Phys. Fluids 35 (9) (2023).doi: 10.1063/5.0161336

work page doi:10.1063/5.0161336 2023

[8] [7]

Beddrich, E

J. Beddrich, E. Süli, B. Wohlmuth, Numerical simulation of the time-fractional Fokker-Planck equation and applications to polymeric fluids, J. Comput. Phys. 497 (2024) 112598.doi:10.1016/j.jcp.2023. 112598

doi:10.1016/j.jcp.2023 2024