Pith. sign in

REVIEW 2 major objections 5 minor 28 references

An inviscid limit to an effective energy-enstrophy diffusion process

T0 review · 2 major / 5 minor · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read The stationary enstrophy-energy law of a Galerkin-Navier-Stokes diffusion with random stirring converges in the inviscid limit to an explicit cone diffusion, independent of the stirring strength, with quantitative low-mode condensation.

desk verdict A serious averaging result that likely holds; the main soft spot is the uniform L2 density bound, which is asserted as an adaptation of Bedrossian–Liss rather than proved here. read the letter →

arxiv 2602.15805 v2 pith:MDJGYE2Z submitted 2026-02-17 math.PR math-phmath.APmath.MP

Alain-Sol Sznitman , Klaus Widmayer This is my paper · ORCID
classification math.PRmath-phmath.APmath.MP MSC 60J6060J2535Q3076F55
keywords inviscidlimitenstrophy-energyprocessGalerkin-Navier-Stokesstationarydiffusionaveragingcondensationrandomstirringcone
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 a proof-of-concept inviscid limit: as epsilon -> 0, the joint law of the enstrophy |X|^2 and the Galerkin-energy |X|^2_{-1} of a stationary N-dimensional Galerkin-Navier-Stokes type diffusion with Brownian forcing and random stirring converges weakly to the law of a stationary diffusion living in an open two-dimensional cone. The limit diffusion is explicit: its generator is the original generator with each squared mode amplitude x_l^2 replaced by the conditional expectation q_l of x_l^2 under the Gaussian measure given the two conserved quadratic forms. The limit does not depend on the strength kappa of the stirring, so the same conclusion holds when the stirring strength is sent to zero along with epsilon (for suitable sequences). Combining this with the companion article's quantitative condensation bound yields an inviscid condensation bound: when the effective spectral value B1/B0 is much smaller than a threshold lambda_{l0} and l0 << N, the stationary enstrophy concentrates on the two lowest modes. A careful reader should care because this is one of the first rigorous handles on how stationary measures of randomly forced Navier-Stokes-type evolutions behave in the inviscid limit.

What carries the argument

The machinery has three parts: (i) the two quadratic forms u = |x|^2 and v = |x|^2_{-1} that the drift B and all stirring fields Z_m conserve; (ii) the functions q_l(u,v) — regular conditional expectations of x_l^2 under the Gaussian measure given u and v — which replace x_l^2 in forming the limit generator A; and (iii) the averaging argument: on each level set X_{u,v}, the diffusion generated by B + kappa D is elliptic (full-rank vector fields), has exponential convergence to equilibrium, and the singular rays u = lambda_l v are controlled by a uniform L^2 bound on stationary densities that forces the 'untamed' set to have vanishing mass.

What would settle it

Run the stationary diffusion for a concrete low-dimensional case (e.g., N=8 with lambda_3=2) at decreasing epsilon and with two very different stirring strengths, kappa=1 and kappa=10^{-6}. If the weak limits of the stationary law of (U0,V0) differ for the two kappa values, the claim of a kappa-independent limit is false. Alternatively, compute the stationary density h_eps and its L^2 norm over a range of epsilon: if sup_{0<eps<=1} ||h_eps||_2 is infinite, Proposition 3.2 fails and the averaging argument collapses.

Watch

Extended reading notes

Core claim

The central claim is Theorem 5.1: as epsilon -> 0, the stationary laws of (|X_t|^2, |X_t|^2_{-1})_{t>=0} under the generator L^eps = L + (1/eps)(B + kappa D) converge weakly to the law P of the stationary diffusion in the open cone C = {(u,v): 0 <= v <= u <= lambda_N v} with generator A in (2.10), regardless of kappa. The proof is an averaging over the fast motion B + kappa D on the level set X_{u,v}: this motion equilibrates exponentially fast away from the singular rays u = lambda_l v, and the 'untamed' set near those rays has uniformly vanishing mass. Theorem 6.1 then yields the quantitative inviscid condensation bound 2 lim E[U_0 - V_0] <= [(B_1 - B_0)/(lambda_{l0}-1) + (lambda_3/(lambda

Load-bearing premise

The load-bearing premise is the uniform L^2 bound on the stationary densities (Proposition 3.2), whose proof is only sketched as an adaptation of an existing spectral-gap argument plus the assertion that the stirring improves the energy estimate; if that uniformity failed, the set of points where the enstrophy-energy ratio nears a singular value lambda_l would not have vanishing mass, and the averaging limit might not live in the open cone.

Editorial extensions

If this is right

  • The limit law P is identical for every fixed stirring strength kappa in (0,1], and also for any sequence kappa_epsilon -> 0 for which the stationary laws converge, as noted in Remark 5.2.
  • The inviscid condensation bound (Theorem 6.1) gives an explicit quantitative estimate: when B1/B0 << lambda_{l0} and l0 << N, the stationary enstrophy is forced into the two lowest modes, with an explicit upper bound on the enstrophy-energy gap.
  • The exponential convergence of the fast motion on the level sets (Proposition B.2) is the quantitative mechanism behind the averaging limit; it transfers directly to the slow two-dimensional process.
  • Corollary 5.3 shows that exponential moments of functions of (U0,V0) converge to their values under the limit stationary measure, attaching explicit moment formulas to the cone diffusion.
  • Under the same conditions, suitably chosen vanishing stirring kappa_epsilon preserves both the weak convergence and the condensation bound (Remark 6.2 2).

Reading between the lines

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

  • The paper's proof shows the limit depends only on the conditioning functions q_l and the ergodicity of the fast motion on level sets; the same averaging principle should hold for any drift and stirring conserving the two quadratic forms, a generality the authors leave implicit.
  • The uniform L^2 bound on stationary densities is the structural soft spot: if a drift b exists that fails the 'improved energy estimate' from the stirring, the untamed set near the singular rays could retain mass and the limit might leave the open cone. This is directly testable in low-dimensional Galerkin truncations.
  • The condensation bound predicts a sharp crossover: as the effective spectral value B1/B0 crosses lambda_{l0}, the stationary mass should spread to higher modes. This threshold behavior is a quantitative prediction that can be checked in simulations of the stationary process.
  • The result identifies only the law of the enstrophy-energy pair, not the full high-dimensional law of X. A natural next step is whether the full stationary law converges to a Gaussian mixture weighted by the limit cone diffusion; the paper does not address this.
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

2 major / 5 minor

Summary. The paper studies stationary laws P^ε of an N-dimensional Galerkin-Navier-Stokes type diffusion with generator L^ε = L + ε^{-1}(B + κD), where B is a quadratic divergence-free drift and D is a stirring operator built from vector fields conserving enstrophy and energy. The main result (Theorem 5.1) states that as ε→0 the laws of the two-dimensional process W_t = (|X_t|^2, |X_t|^2_{-1}) under P^ε converge weakly to the stationary law P of a diffusion in the open cone C generated by the operator A, whose coefficients are formed from the functions q_ℓ — the conditional expectations of x_ℓ^2 given the two conserved quantities under the Gaussian measure μ. The companion paper [27] supplies well-posedness, stationarity, and condensation bounds for this cone diffusion; combining them with Theorem 5.1 yields quantitative inviscid condensation bounds on the limiting stationary distribution (Theorem 6.1). The proof proceeds by tightness, a uniform L^2 density bound for μ^ε, and an averaging argument over the fast motion on the level sets X_{u,v}, with control of the singular rays u = λ_ℓ v.

Significance. The claimed result is significant as a proof of concept: it identifies a two-dimensional effective process for the slow enstrophy-energy variables of a high-dimensional Galerkin-Navier-Stokes system with arbitrarily weak stirring, and it gives quantitative low-mode condensation in the inviscid limit. The effective coefficients q_ℓ are explicit and parameter-free; the reduction (5.4)→(5.10)→(5.12) is coherent, and the exponential/moment bounds in Section 3 are clean. The strategy of controlling the singular set via uniform density estimates and fast-equilibrium convergence is structurally plausible. However, the paper is not self-contained: the limiting process and key estimates from [27] are imported without proof or arXiv identifier, and the uniform L^2 density bound that underpins the singular-set control is only sketched. The result is therefore conditional on these two components. If the missing pieces are supplied, this would be a solid contribution.

major comments (2)
  1. [Section 3, Prop. 3.2 (eq. (3.15))] The uniform L^2 bound on the stationary densities is load-bearing: Proposition 3.3 uses it at (3.20)–(3.22), and the proof of Theorem 5.1 invokes Proposition 3.3 at (5.24) and (5.47) to make the bad-set contributions I_{2,2} and the final limsup vanish. The proof given for Prop. 3.2 is one sentence: it is said to be an adaptation of the proof of (4.4) on pp. 507–508 of [2], with the assertion that the stirring vector fields 'actually improve the energy estimate'. This is not a routine translation: the present generator contains the singular factor ε^{-1}(B + κD), the drift and stirring fields are quadratic, and the required uniformity is over all 0 < ε ≤ 1. If the claimed improvement has a hidden sign issue or an ε-dependent constant, the bound (3.15) could fail and Theorem 5.1 would collapse. I ask for a complete proof, or at least a precise statement of the adapted energy estimate and
  2. [Section 2 and Theorem 6.1 (ref. [27])] The target measure P, the functions q_ℓ, the identity (A.21), the well-posedness and uniqueness of the martingale problem for A, the stationary distribution π, and the condensation bound (2.19) are all quoted from the companion article [27], listed only as 'Available on arXiv, 2026' with no identifier. These are not auxiliary facts: Theorem 5.1 states convergence to this imported P, and Theorem 6.1 is a direct application of the imported (2.19). As the manuscript stands, the main claims cannot be independently verified without access to [27]. I recommend either including the companion results in the same submission, providing complete proofs in an appendix, or at minimum supplying a citable arXiv identifier and precise statements of the quoted theorems.
minor comments (5)
  1. [Eq. (5.23)] The decomposition is written 'I2 = I2,1 + I2,1'; the second term should be I2,2.
  2. [Eq. (5.47)] The middle term on the right-hand side should be (1 + u_max) sup_{0<ε≤1} (P^ε[U_0 ≥ u_max/2])^{1/2}, as in (5.34); the square root is missing. This does not affect the conclusion, but should be corrected.
  3. [Introduction, p. 5] 'An central result' should be 'A central result'.
  4. [Remark 5.2] The statement that 'by usual arguments' one can replace κ by a suitable κ_ε → 0 and still obtain convergence to P is not proved and no construction is given. Since this is only a robustness remark, please either supply a proof or clearly mark it as a conjecture.
  5. [Abstract] The abstract says 'random stirring (of arbitrarily small strength)'; in the main text κ is a fixed parameter in (0,1]. The independence of the limit from κ is proved, but the wording could mislead readers into thinking κ itself is sent to zero in Theorem 5.1.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: convergence proof is an explicit averaging argument; reliance on companion [27] and on an external L2-density bound is structural but not definitional.

full rationale

The derivation chain is not circular. The effective generator A in (2.10) is defined by replacing x_l^2 with the explicit conditional-expectation functions q_l, and Appendix A.21 gives q_l(u,v)=∫x_l^2 dµ_{u,v}; there are no fitted parameters, and the limit process is not defined in terms of the laws P^ε. Theorem 5.1 is proved by a direct averaging argument: the reductions (5.7)-(5.10) and the bounds (5.24), (5.26), (5.34), (5.44) control X_{s,l}^2 - q_l(W_s) via the fast-motion convergence-to-equilibrium estimate (Prop. B.2), tightness (Thm 4.1), and the untamed-set estimate (Prop. 3.3), rather than by assuming the conclusion. The final identification of any weak subsequential limit Q with the target law P does invoke the uniqueness characterization (2.16) from the companion paper [27, Prop 4.2]; this is a self-citation, but it is independent support (it concerns well-posedness/uniqueness of the cone-valued diffusion, not the inviscid convergence itself), so it is not a circular reduction. Theorem 6.1 is a direct application of the companion condensation bound (2.19) together with Corollary 5.3; it transfers a separately proved bound across the weak limit and is not a restatement of the input. The proof of Proposition 3.2 is only sketched as an adaptation of Bedrossian-Liss [2] with the assertion that the stirring fields improve the energy estimate; Proposition 3.3 and the untamed-term controls (5.24)/(5.47) depend on it, so this is a verification gap / correctness risk, but not circularity. The companion reference [27] is cited without an arXiv identifier, adding a verifiability gap. Weighing these, there is no step in which a prediction is equal by construction to a fitted input or to a self-citation chain.

Assumptions & free parameters 3 free parameters · 7 assumptions · 3 invented entities

The central theorems hold for all model inputs: a > 0, delta_l in (-1,0], spectrum (1.2), kappa in (0,1]; no parameter is fitted to data. The only hand-chosen quantity with physical effect is delta_l (it sets B_0, B_1 and hence the condensation window B_1/B_0 << lambda_l0); the stirring fields Z_m are an internal device whose full-rank property is proven as Proposition A.3; the q_l are explicit Gaussian conditional expectations; the cone limit process and its condensation bound are imported from companion [27]; the uniform L^2 density bound of Section 3 is imported from [2]. The quantitative content of (6.1) is explicit, but the spectral-gap constants in Proposition B.2 are only proven to exist, not quantified.

free parameters (3)
  • delta_l (per-pair OU variance correction) = any in (-1,0], delta_2i = delta_2i-1
    Hand-chosen model input (1.8) controlling B_0, B_1 = a*sum(lambda_l(1+delta_l)) and thus when the condensation bound (6.1) is small (requires B_1/B_0 << lambda_l0); the theorem itself is uniform in delta_l.
  • c_3(u_min,u_max,eta) spectral-gap constant = not quantified
    Existence-only constant in the mixing estimate (B.14) used to close the averaging argument; never evaluated, so the convergence rate in epsilon is not quantitative.
  • lambda_l, a, kappa = model inputs; kappa in (0,1] arbitrary, a > 0 arbitrary, spectrum (1.2)
    Fixed model inputs, not fitted; the paper proves uniformity in kappa and the limit is kappa-independent. The condensation window depends on ratios such as B_1/B_0 and lambda_l0, but the theorem holds for all these values.
assumptions (7)
  • standard math Martingale-problem well-posedness and unique stationary measure for elliptic diffusions under Lyapunov-Foster conditions (Meyn-Tweedie [24], Thms 2.1, 4.2)
    Used to obtain Proposition 1.3 (well-posedness, stationary measure, absolute continuity for L^eps) and, via [27], the analogous facts for the cone generator A; invoked in Section 1 and at (5.3)-(5.4) to identify the limit Q with P.
  • standard math Parabolic heat-kernel upper/lower bounds for uniformly elliptic operators with Lipschitz coefficients (Il'in-Kalashnikov-Oleinik [15], Thm 1 p. 67, (4.75) p. 82)
    Black box used in Appendix B to derive the Dirichlet heat-kernel estimates (B.21)-(B.22) inside local charts, which underpin the positivity (B.16) and boundedness (B.17) of the transition density needed for the mixing estimate (B.14).
  • domain assumption Conservation structure: div b = 0 = div Z_m and <x,b(x)> = <x,b(x)>_{-1} = <x,Z_m(x)> = <x,Z_m(x)>_{-1} = 0 (1.11), (1.17); the Galerkin-Navier-Stokes drift (1.12) satisfies these by (1.13)
    This is what makes the drift and stirring invisible to functions of (U,V) (1.18), confines the fast motion to X_{u,v}, and makes the averaging limit well-defined; it is the structural modeling assumption of the paper.
  • domain assumption Full-rank property of the stirring vector fields on X_{u,v} away from singular rays (Prop. A.3, (A.23))
    Ensures the fast diffusion Gamma = kappa D + B is elliptic on the level surfaces, so the convergence-to-equilibrium bound (B.14) is available; proven in Appendix A by a spanning argument.
  • ad hoc to paper Uniform-in-epsilon L^2 bound on stationary densities (Prop. 3.2, (3.15)), adapted from Bedrossian-Liss [2], with the claim that the stirring improves the relevant energy estimate
    Ad-hoc import: load-bearing for the vanishing mass of the untamed set (3.18) used to close the averaging argument; the adaptation is asserted, not demonstrated.
  • ad hoc to paper Companion results: well-posedness and uniqueness of the stationary measure for the cone diffusion, the characterization of q_l (2.9), (A.21), and the condensation bound (2.19), all from [27]
    All of Section 2 and the import into Theorem 6.1 depend on [27]; the present paper quotes these without proof, so its verifiability is tied to the companion.
  • domain assumption Ellipticity of the base operator: delta_l in (-1,0] with delta_2i = delta_2i-1 (1.8); the pairing can be dispensed with, per the text below (1.9)
    Gives the elliptic set-up used for absolute continuity and the L^2 density argument; the paper itself notes the pairing assumption is only for exposition.
invented entities (3)
  • Effective cone diffusion P with generator A (2.10)
    purpose: Candidate inviscid limit of the enstrophy-energy process (U_t,V_t); coefficients are the conditional expectations q_l.
    Constructed within the same research program (companion [27]); the only handle outside its own construction is the model-level condensation prediction (6.1), which is a theorem of the model, not an external observable.
  • Stirring vector fields Z_m = T_J, R_i (1.14)-(1.16)
    purpose: Regularize the fast motion on X_{u,v} and make it ergodic with a quantitative mixing rate; kappa may be arbitrarily small.
    Ad hoc device, chosen to preserve U and V and to satisfy the full-rank property (A.23); no direct physical counterpart is claimed.
  • q_l(u,v) — Gaussian conditional expectations of x_l^2 given U=u, V=v (0.9), (A.21)
    purpose: Coefficient functions of the effective generator A; also encode the condensation bounds via their monotonicity (Thm 2.3 of [27]).
    Parameter-free and explicitly defined, but internal to the construction; their properties (regularity, boundary values (2.6)-(2.8), monotonicity) come from [27].

how reviews work

0 comments
Cite this review

Pith. "Pith review of An inviscid limit to an effective energy-enstrophy diffusion process." pith.science (2026). https://pith.science/paper/MDJGYE2Z

@misc{pith2026260215805,
  author       = {Pith},
  title        = {Pith review of: An inviscid limit to an effective energy-enstrophy diffusion process},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MDJGYE2Z}},
  note         = {Machine review of arXiv:2602.15805}
}
abstract

In this article we consider a stationary $N$-dimensional Galerkin-Navier-Stokes type evolution with Brownian forcing and random stirring (of arbitrarily small strength that plays the role of a regularization). We show, as a ``proof of concept'', that the stationary diffusion in an open two-dimensional cone constructed in a companion article, stands as the inviscid limit of the laws of the ``enstrophy-energy'' process of the $N$-dimensional diffusion process considered here, this regardless of the strength of the stirring. With the help of the quantitative condensation bounds of the companion article, we infer quantitative inviscid condensation bounds, which for suitable forcings show an attrition of all but the lowest modes in the inviscid limit.

Figures

Figures reproduced from arXiv: 2602.15805 by the authors.

Figure 1
Figure 1. A schematic illustration of the polytope [PITH_FULL_IMAGE:figures/full_fig_p029_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 1 linked inside Pith

  1. [27]

    Sznitman and K

    A.-S. Sznitman and K. Widmayer. Effective energy-enstrophy diffusion process and condensation bound. Available on arXiv, 2026

    2026

  2. [2]

    Bedrossian and K

    J. Bedrossian and K. Liss. Quantitative spectral gaps for hypo elliptic stochastic dif- ferential equations with small noise. Probab. Math. Phys. , 2(3):477–532, 2021

    2021

  3. [1]

    M. Beck, E. Cooper, G. Lord and K. Spiliopoulos. Selection of quas i-stationary states in the stochastically forced Navier–Stokes equation on the torus. Journal of Nonlinear Science, 30:1677–1702, 2020

    2020

  4. [3]

    Bouchet and M

    F. Bouchet and M. Corvellec. Invariant measures of the 2D Euler and Vlasov equa- tions. J. Stat. Mech. , P08021, 2010

    2010

  5. [4]

    Bouchet and E

    F. Bouchet and E. Simonnet. Random Changes of flow topology in t wo-dimensional and geophysical turbulence. Phys. Rev. Lett., 102, 094504, 2009

    2009

  6. [5]

    P.G. Ciarlet. Linear and nonlinear functional analysis with applicatio ns. Society for Industrial and Applied Mathematics (SIAM) , 2nd edition, Philadelphia, 2025

    2025

  7. [6]

    Da Prato and A

    G. Da Prato and A. Debussche. Two-dimensional Navier-Stokes equations driven by a space-time white noise. J. Funct. Anal. , 196(1):180–210, 2002

    2002

  8. [7]

    Echeverria

    P. Echeverria. A criterion for invariant measures of Markov pro cesses. Z. Wahrschein- lichkeitstheorie verw. Gebiete, 61:1–16, 1982

    1982

Show all 28 references
  1. [8]

    Elesaely, D

    S. Elesaely, D. Herzog and K. Liss. Quantitative positivity of transition densities for random perturbations of Hamiltonian systems . arXiv:2509.02448

    arXiv

  2. [9]

    Freidlin and A.D

    M.I. Freidlin and A.D. Wentzell. Random perturbations of dynamical systems . 3rd edition, Springer, New York, 2012

    2012

  3. [10]

    Glatt-Holz, V

    N. Glatt-Holz, V. ˇSver´ ak and V. Vicol. On Inviscid limits for the Stochastic Navier–Stokes equations and related models. Arch Rational Mech. Anal. , 217:619– 649, 2015

    2015

  4. [11]

    Hairer and J.C

    M. Hairer and J.C. Mattingly. Ergodicity of the 2D Navier-Stokes equations with degenerate stochastic forcing. Annals of Mathematics, 164:993–1032, 2006. 36

    2006

  5. [12]

    Hairer and J.C

    M. Hairer and J.C. Mattingly. A theory of hypoellipticity and unique ergodicity for semilinear stochastic PDEs. Electronic Journal of Probability , 16, paper 23, 658–738, 2011

    2011

  6. [13]

    Hairer and T

    M. Hairer and T. Rosati. Global existence for perturbations of the 2D stochastic Navier-Stokes equations with space-time white noise. Ann. PDE , 10, No. 1, Paper No. 3, 46 p., 2024

    2024

  7. [14]

    Ikeda and S

    N. Ikeda and S. Watanabe. Stochastic Differential Equations and Diffusion Processes . 2nd edition, North-Holland; Amsterdam, Kodansha, Ltd., Tokyo, 1 989

  8. [15]

    Il’in, A.S

    A.M. Il’in, A.S. Kalashnikov and O.A. Oleinik. Linear equations of the s econd order of parabolic type. Russian Math. Surveys , 17(3):1–143, 1962

    1962

  9. [16]

    Kallenberg

    O. Kallenberg. Foundations of modern probability . 2nd edition, Springer, New York, 2002

    2002

  10. [17]

    Karatzas and S

    I. Karatzas and S. Shreve. Brownian motion and stochastic calculus . Springer, Berlin, 1988

    1988

  11. [18]

    S. Kuksin. On distribution of energy and vorticity for solutions o f 2D Navier-Stokes equation with small viscosity. Comm. Math. Phys. , 2:407–424, 2008

    2008

  12. [19]

    S. Kuksin. Weakly nonlinear stochastic CGL equations. Ann. Inst. H. Poincar´ e Probab. Statist., 49(4):1033–1056, 2013

    2013

  13. [20]

    Kuksin and A

    S. Kuksin and A. Piatnitski. Khashminskii-Whitham averaging for r andomly per- turbed KdV equation. J. Math. Pures Appl. , 89:400–428, 2008

    2008

  14. [21]

    Kuksin and A

    S. Kuksin and A. Shirikyan. Mathematics of two-dimensional turbulence . Cambridge University Press, 2012

    2012

  15. [22]

    Levin, Y

    D.A. Levin, Y. Peres, and E.L. Wilmer. Markov chains and mixing times . American Mathematical Soc., Providence, 2017

    2017

  16. [23]

    Linkmann, M

    M.M. Linkmann, M. Hohmann and B. Eckhardt. Non-universal tr ansitions to two- dimensional turbulence. J. Fluid Mech. , vol. 892, A18, 2020

    2020

  17. [24]

    Meyn and R.L

    S.P. Meyn and R.L. Tweedie. Stability of Markovian processes III . Foster-Lyapunov criteria for continuous-time processes. Adv. Appl. Prob. , 25:518–548, 1993

    1993

  18. [25]

    Nazarenko

    S. Nazarenko. Wave Turbulence. Lecture Notes in Phys., 825, Springer, New York, 2011

    2011

  19. [26]

    Stroock and S.R.S

    D.W. Stroock and S.R.S. Varadhan. Multidimensional diffusion processes . Springer, New York, 1979

    1979

  20. [28]

    F.D. Warner. Foundations of differential manifolds and Lie groups . Graduate Texts in Mathematics 94, Springer, New York, 1983. 37

    1983

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.