Global solutions and Asymptotic Behavior to a Norm-preserving Non-local Parabolic Flow

Boris Shakarov

REVIEW 2 major objections 5 minor 25 references

Global solutions and Asymptotic Behavior to a Norm-preserving Non-local Parabolic Flow

T0 review · 2 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read A nonlocal heat flow that keeps its $L^2$ mass fixed is globally well-posed for every real frequency and subcritical nonlinearity, and on a ball with positive data it converges strongly to the unique ground state.

desk verdict Solid ω≥0 results and a clean Lyapunov structure, but the advertised all-ω global well-posedness rests on a false lemma. read the letter →

arxiv 2411.18532 v1 pith:TBXH6YDZ submitted 2024-11-27 math.AP

Boris Shakarov This is my paper · ORCID

classification math.AP MSC 35K5535B4035B45

keywords nonlocalparabolicflownorm-preservingL2-constraintglobalwell-posednessLyapunovfunctionalgroundstateasymptoticconvergencesubcriticalnonlinearity

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

This paper studies a heat-type equation modified by a nonlocal term chosen so that every solution keeps its $L^2$ mass exactly constant over time. The central claim is that, for every real frequency parameter $\omega$ and every subcritical nonlinearity exponent, the flow is global in time: no solution blows up, and the $H^1$ norm (size together with gradient size) stays uniformly bounded, on bounded domains and on the whole space under the stated restrictions. The paper then proves that on bounded domains the asymptotic dynamics are constrained to stationary states, and that on a ball with positive initial data and $\omega>0$ the entire trajectory converges strongly in $H^1$ to the unique positive ground state. This matters because norm-preserving flows are widely used as numerical tools to compute stationary states, and the closely related constrained gradient flow was previously known to admit growing-up solutions in the intercritical regime.

What carries the argument

The load-bearing object is the Lyapunov functional $F[u]=(\|\nabla u\|_{L^2}^2+\omega\|u\|_{L^2}^2)/\|u\|_{L^{2\sigma+2}}^{2\sigma+2}$, together with the nonlocal multiplier $\mu[u]$ that enforces the fixed-$L^2$ constraint. $F$ decreases along the flow and its explicit decay formula transfers the integrated decay of $\|\partial_t u\|_{L^2}$ into a uniform bound on $\|\nabla u\|_{L^2}$, because Gagliardo-Nirenberg interpolation bounds the denominator $\|u\|_{L^{2\sigma+2}}$ by a power of the gradient. The same functional selects the ground state as its minimizer, which is what turns subsequential convergence into full convergence on a ball.

What would settle it

Take $\omega<0$ and $u_0=\varphi$, where $\varphi$ is a Dirichlet eigenfunction of $-\Delta$ with eigenvalue $-\omega$ normalized in $L^2$: the exact solution is $u(t)=\varphi$, so $\|\nabla u(t)\|^2+\omega\|u_0\|^2\equiv 0$, contradicting the strict negativity asserted in (44). Initializing instead just inside the set $G$ and numerically integrating (1) would test whether the claimed invariance of $G$, and hence the $\omega<0$ uniform bound, actually holds.

Watch

Extended reading notes

Core claim

The discovery is that the nonlocal multiplier $\mu[u]=(\|\nabla u\|_{L^2}^2+\omega\|u\|_{L^2}^2)/\|u\|_{L^{2\sigma+2}}^{2\sigma+2}$, which enforces $\|u(t)\|_{L^2}=\|u_0\|_{L^2}$ for all $t$, makes the flow gradient-like rather than destabilizing. The quantity $F[u(t)]=(\|\nabla u(t)\|^2+\omega\|u_0\|^2)/\|u(t)\|_{L^{2\sigma+2}}^{2\sigma+2}$ is non-increasing and satisfies the exact differential identity $\frac{d}{dt}\log \frac{\sqrt{\|\nabla u\|^2+\omega\|u_0\|^2}}{\|u\|_{L^{2\sigma+2}}}=-\frac{\|\partial_t u\|_{L^2}^2}{\|\nabla u\|^2+\omega\|u_0\|^2}$. Combined with the Gagliardo-Nirenberg inequality, this identity controls the gradient uniformly for every subcritical $\sigma$, giving global existence; the same Lyapunov function, together with compactness and uniqueness of the positive stationary solution, gives strong $H^1$ convergence to the ground state on a ball.

Load-bearing premise

The load-bearing premise is Lemma 4.1's claim that a solution reaching the surface $\|\nabla u\|^2+\omega\|u_0\|^2=0$ must immediately move below it; a stationary solution that is a Dirichlet eigenfunction of $-\Delta$ with eigenvalue $-\omega$ stays on that surface for all time, so this premise fails exactly there.

Editorial extensions

If this is right

Global existence and a uniform $H^1$ bound hold for all $\omega\in\mathbb{R}$ and all subcritical $\sigma<2/(d-2)_+$ on bounded $C^2$ domains, without any smallness assumption on the initial data.
For $\omega>0$ on a bounded domain, the $\omega$-limit set is nonempty, compact, connected, and contained in the set of stationary states with the same $L^2$ norm; the same conclusion holds on $\mathbb{R}^d$ for radially symmetric data.
When the domain is a ball, $\omega>0$, and the initial datum is positive, the whole trajectory converges in $H^1$ to the unique positive stationary state $Q_{gs}$, not merely along a subsequence.
The Lyapunov identity implies $\int_0^\infty\|\partial_t u(t)\|_{L^2}^2\,dt<\infty$, so the time derivative of the solution decays to zero in $L^2$ as $t\to\infty$.
In contrast to the earlier constrained gradient flow (3), which exhibits growing-up solutions for $2/d\le\sigma<2/(d-2)_+$, the present flow is asserted to remain uniformly bounded for this whole range.

Reading between the lines

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

The paper leaves open whether the $\omega$-limit set is a singleton outside the ball/positive-data case; a natural next step would be to combine the Lyapunov identity with spectral properties of the linearized operator around the ground state to obtain full convergence and rates.
The same $L^2$-sphere constraint and the same $F$-type functional underlie normalized gradient flows used to compute ground states of Bose-Einstein condensates; the global-bound result suggests these schemes remain stable across the full subcritical range, beyond the intercritical regime where the unmodified flow grows up.
The paper notes that its Aubin-Lions/Schauder route fails on $\mathbb{R}^d$, leaving a gap for the whole space; a plausible repair is to adapt the density argument with localized compactness, which would remove the extra restrictions on $\sigma$ currently needed there.

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

Colleague,

The paper is a mix of solid work and one load-bearing gap. The headline: the abstract advertises global well-posedness for all ω∈R, but the proof for ω<0 rests on Lemma 4.1(44), which is false. A nonzero Dirichlet eigenfunction φ with -Δφ = -ωφ (ω<0) is a stationary solution: μ[φ]=0, so A(t)≡0 for all t, contradicting the claimed strict negativity in (44). That knocks out Proposition 4.3's invariance argument and leaves Theorem 1.2 unsupported for ω<0. The ω≥0 results are unaffected, and the asymptotics for ω>0 stand on their own.

What's genuinely new: local well-posedness on bounded domains for the full subcritical range via a Schauder fixed point argument (the earlier Ma–Cheng paper only treated ω=0 on compact manifolds), the treatment of R^d with two separate contraction arguments, and the ground-state convergence on balls for positive data. The Lyapunov structure (Lemma 4.1 up to the false claim, Corollary 4.2) is clean, and the asymptotic analysis in Section 5 is mostly careful. The paper is not a model of exposition—there are typos like 'M < 0' in Lemma 5.1—but the mathematics is mostly well-organized.

The central question is whether the ω<0 case can be fixed. The stationary eigenfunction sits exactly on the boundary between G and H, so the decomposition into G and H is not invariant in the claimed way. One could try to split trajectories at first hitting time of A=0, but then the H^1 bound for data in H may fail if the solution actually crosses; you'd need a different mechanism. The author should either repair this or restrict Theorems 1.1–1.2 to ω≥0. As stated, the paper overclaims.

Who is this for? People working on norm-preserving parabolic flows and ground-state computation will find the ω≥0 part useful, and the flaw itself is instructive. I'd send it to a referee, but I'd expect a major revision. My recommendation: engage with it, but do not accept the current version.

Referee Report

2 major / 5 minor

Summary. The paper studies the nonlocal parabolic equation (1), in which the nonlocal coefficient mu[u] is chosen so that the L^2 norm of solutions is preserved in time. It claims local well-posedness in H^1_0 on bounded C^2 domains for every subcritical sigma, and on R^d under additional restrictions; it then claims global well-posedness with a uniform H^1 bound for every omega in R (Theorem 1.2). The asymptotic results assert that for omega>0 the omega-limit set consists of stationary states (Theorem 1.3), and that on a ball with positive initial data the solution converges in H^1 to the unique positive ground state (Theorem 1.4). The central mechanism for the omega<0 global bound is Proposition 4.3, which relies on the invariance of the sets G and H established through Lemma 4.1.

Significance. The omega>0 results are interesting: if they are correct, they provide a sharp contrast with the grow-up behavior known for the related norm-preserving model (3), and they give a relatively clean Lyapunov-functional route to convergence to stationary states. The paper also contains useful technical work in the local well-posedness arguments, especially the Schauder fixed-point proof on bounded domains and the space-time estimates on R^d. However, the claimed global well-posedness for every omega in R, which is one of the headline results, is not established: the proof for omega<0 rests on a false statement in Lemma 4.1. The omega>=0 parts and the asymptotic analysis appear to be unaffected by this defect, but the manuscript in its present form does not prove one of its central theorems.

major comments (2)

[Lemma 4.1, eq. (44)] The assertion that A(t0)=0 implies A(t0+epsilon)<0 for some epsilon>0 is false. Let omega<0 and let phi be a nonzero Dirichlet eigenfunction of -Delta on H^1_0(Omega) with eigenvalue -omega. Then mu[phi]=0 and u(t)≡phi solves (1); moreover A(t)=||grad phi||^2+omega||phi||^2=0 for every t>=0. Hence no t0>0 admits an epsilon with A(t0+epsilon)<0. The derivation in the proof only gives A'(t0)=-2||partial_t u(t0)||^2 <=0, which is compatible with A remaining identically zero in the stationary case; the stationary case is not excluded by the hypotheses of the lemma.
[Proposition 4.3] The invariance of G and H for omega<0, and consequently the uniform H^1 bound for data in H, rests entirely on the false statement (44). The claim that (45) implies F[u0] and F[u(t)] have the same sign for all t is unjustified, because (45) is derived only while the denominator A(s)=||grad u(s)||^2+omega||u0||^2 does not vanish; once A hits zero, the representation (45) breaks down. Since Proposition 4.3 is the only argument given for global well-posedness when omega<0, Theorem 1.2 is not proved for that parameter range. The omega>=0 global result in Corollary 4.2 is not affected.

minor comments (5)

[Lemma 5.1, proof] The sentence 'sup_{t>=0} ||grad u(t)||_{L^2} <= M < 0' is nonsensical as written; M must be a positive finite constant, so it should read M < infinity.
[Proposition 3.1, after (27)] The text says 'by (28), (26) and (26)' but the second reference should be (27), since (26) alone gives the self-map property while (27) gives the contraction estimate.
[Proposition 3.8, final sentence] The sentence 'Thus, u satisfies (37)' should refer to equation (1), since (37) is the regularized equation with mu_epsilon and the limit has already been taken.
[Corollary 5.3] The set S0 is used without being defined; presumably it denotes the set of stationary states of (7) with the fixed L^2 norm, but the notation should be introduced.
[Theorem 5.4, proof] The notation Q is used inconsistently: the proof repeatedly writes 'u(t_k) -> Q' and 'E[u]=E[Q]' when the intended limit is the ground state Qgs; this should be corrected for clarity.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the Lyapunov functional and all derived bounds are obtained by direct computation from the equation, not by construction or self-reference.

full rationale

The paper's derivation chain is self-contained rather than circular. The norm preservation identity (43) follows by multiplying (1) by u and integrating, and the Lyapunov monotonicity formula (45) is derived from the equation by multiplying by ∂_t u and using the definition (2) of µ[u]; the exponential formula is a consequence of those identities, not an input. The uniform H1 bounds in Corollary 4.2 and Proposition 4.3 are obtained from monotonicity of F together with the Gagliardo-Nirenberg inequality and the subcritical condition dσ/(σ+1) < 2, with no fitted constants or data-dependent predictions. The asymptotic results use compactness, the fact that F is decreasing and bounded below, and the external, standard uniqueness theorems for positive solutions [17, 20]; those citations are not self-citations and do not smuggle in the target results. The author's own prior works [1, 2] are cited only for context on a different model (3) and on a related Ginzburg-Landau-type equation, and they are not load-bearing for any theorem here. The paper explicitly acknowledges an unresolved question about whether the ω-limit set is a singleton, which is an honest limitation rather than a circular move. Even if Lemma 4.1(44) is mathematically problematic for stationary Dirichlet eigenfunctions when ω < 0, a false or unproved strict-sign assertion is a correctness risk, not circularity, because it is not equivalent by construction to the statement being proved. Overall, no prediction or first-principles result reduces to its own inputs.

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

No free parameters are fitted and no new entities are postulated. The analysis relies on standard analytic tools plus one false dichotomy, eq. (44), that is specific to this paper and affects only the omega < 0 global bound.

assumptions (6)

standard math Gagliardo-Nirenberg inequality (Prop 2.1) holds on the domain with constant C_GN(d,sigma,Omega).
Used to turn the Lyapunov decay into a uniform H1 bound (eq. (49)).
standard math Heat semigroup smoothing estimates (Props 2.3-2.5) are valid for the Dirichlet Laplacian on bounded C2 domains and R^d.
Basis of the contraction and Schauder fixed point arguments for local well-posedness.
standard math Unique positive ground state exists for (7) on a ball and on R^d (Prop 2.6, from [17,20]).
Needed to identify the limit in Theorem 1.4 as Q_gs rather than another stationary state.
standard math Maximum principle applies to the parabolic flow so that u0 >= 0 implies u(t) > 0 for t > 0.
Used in Theorem 5.4 to ensure the subsequential limit is positive and hence the ground state.
standard math H^1_rad(R^d) embeds compactly into L^p for 2 <= p < 2d/(d-2).
Used in Lemma 5.2 to extract convergent subsequences in the radial whole-space case.
ad hoc to paper When A(t) = 0 and u(t) is not stationary, A becomes negative (Lemma 4.1 eq. (44)).
False: stationary Dirichlet eigenfunctions with -Delta u = omega u have A identically zero. This false dichotomy is load-bearing for the omega < 0 uniform bound.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Global solutions and Asymptotic Behavior to a Norm-preserving Non-local Parabolic Flow." pith.science (2026). https://pith.science/paper/TBXH6YDZ

@misc{pith2026241118532,
  author       = {Pith},
  title        = {Pith review of: Global solutions and Asymptotic Behavior to a Norm-preserving Non-local Parabolic Flow},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TBXH6YDZ}},
  note         = {Machine review of arXiv:2411.18532}
}

abstract

We consider a nonlinear parabolic model that forces solutions to stay on a $L^2$-sphere through a nonlocal term in the equation. We study the local and global well-posedness on a bounded domain and the whole Euclidean space in the energy space. Then, we consider the solutions' asymptotic behavior. We prove strong convergence to a stationary state and asymptotic convergence to the ground state in bounded domains when the initial condition is positive.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 25 canonical work pages

[1]

Antonelli, P

P. Antonelli, P. Cannarsa, and B. Shakarov. Existence an d asymptotic behavior for L2-norm preserving nonlinear heat equations. Calc. Var. Partial Diﬀer. Equ. , 63(4):1–31, 2024

work page 2024
[2]

Antonelli and B

P. Antonelli and B. Shakarov. Existence and Large Time Be havior for a Dissipative Variant of the Rotational NLS Equation. Commun. Math. Sci. , 22(6):1601–1633, 2024

work page 2024
[3]

T. Aubin. Some nonlinear problems in Riemannian geometry . Springer Monographs in Mathematics. Springer- Verlag, Berlin, 1998

work page 1998
[4]

Bao and Q

W. Bao and Q. Du. Computing the ground state solution of Bo se-Einstein condensates by a normalized gradient ﬂow. SIAM J. Sci. Comput. , 25(5):1674–1697, 2004

work page 2004
[5]

Berestycki and P

H. Berestycki and P. L. Lions. Nonlinear scalar ﬁeld equa tions, ii existence of inﬁnitely many solutions. Archive for Rational Mechanics and Analysis , 82:347–375, 1983

work page 1983
[6]

Berestycki and J

H. Berestycki and J. W ei. On least energy solutions to a se milinear elliptic equation in a strip. Discrete Contin. Dyn. Syst. , 28(3):1083–1099, 2010

work page 2010
[7]

Besse, R

C. Besse, R. Duboscq, and S. Le Coz. Gradient ﬂow approach to the calculation of stationary states on nonlinear quantum graphs. Ann. Henri Lebesgue , 5:387–428, 2022

work page 2022
[8]

Boyer and P

F. Boyer and P. Fabrie. Mathematical Tools for the Study of the Incompressible Navi er-Stokes Equations an- dRelated Models, volume 183. Springer Science & Business Media, 2012

work page 2012

Show all 25 references

[9]

Brze´ zniak, G

Z. Brze´ zniak, G. Dhariwal, and M. Mariani. 2D constrain ed Navier-Stokes equations. J. Diﬀerential Equations , 264(4):2833–2864, 2018

2018
[10]

Caﬀarelli and F

L. Caﬀarelli and F. H. Lin. Nonlocal heat ﬂows preservin g the L2 energy. Discrete Contin. Dyn. Syst. , 23(1- 2):49–64, 2009

2009
[11]

Caglioti, M

E. Caglioti, M. Pulvirenti, and F. Rousset. On a constra ined 2-D Navier-Stokes equation. Comm. Math. Phys. , 290(2):651–677, 2009

2009
[12]

Cazenave

T. Cazenave. Semilinear Schr¨ odinger equations, volume 10 of Courant Lecture Notes in Mathematics . New York University, Courant Institute of Mathematical Sciences, N ew York; American Mathematical Society, Providence, RI, 2003

2003
[13]

Cazenave and P.-L

T. Cazenave and P.-L. Lions. Orbital stability of stand ing waves for some nonlinear Schr¨ odinger equations.Comm. Math. Phys. , 85(4):549–561, 1982

1982
[14]

M. L. Chiofalo, S. Succi, and M. Tosi. Ground state of tra pped interacting Bose-Einstein condensates by an explicit imaginary-time algorithm. Phys. Rev. E , 62(5):7438, 2000

2000
[15]

Q. Du. Numerical computation of quantized vortices in t he Bose-Einstein condensate. In Recent progress in computational and applied PDEs (Zhangjiajie, 2001) , pages 157–169. Kluwer/Plenum, New York, 2002

2001
[16]

Friedman

A. Friedman. Partial Diﬀerential Equations of Parabolic Type . Dover Books on Mathematics. Dover Publications, 2013

2013
[17]

Gidas, W

B. Gidas, W. M. Ni, and L. Nirenberg. Symmetry and relate d properties via the maximum principle. Comm. Math. Phys. , 68(3):209–243, 1979

1979
[18]

Gilbarg and N

D. Gilbarg and N. S. Trudinger. Elliptic partial diﬀerential equations of second order , volume 224. Springer- Verlag, 1977

1977
[19]

J. K. Hale. Asymptotic behavior of dissipative systems . Number 25. American Mathematical Soc., 2010

2010
[20]

M. K. Kwong. Uniqueness of positive solutions of ∆ u− u+up = 0 in Rn. Arch. Rational Mech. Anal., 105(3):243– 266, 1989

1989
[21]

O. A. Ladyˇ zenskaja, V. A. Solonnikov, and N. N. Ural’ce va. Linear and quasilinear equations of parabolic type . Translations of Mathematical Monographs, Vol. 23. America n Mathematical Society, Providence, R.I., 1968. Translated from the Russian by S. Smith

1968
[22]

Ma and L

L. Ma and L. Cheng. Non-local heat ﬂows and gradient esti mates on closed manifolds. J. Evol. Equ. , 9(4):787–807, 2009

2009
[23]

Ma and L

L. Ma and L. Cheng. Global solutions to norm-preserving non-local ﬂows of porous media type. Proceedings of the Royal Society of Edinburgh: Section A Mathematics , 143(4):871––880, 2013

2013
[24]

Quittner and P

P. Quittner and P. Souplet. Superlinear parabolic problems . Birkh¨ auser Advanced Texts: Basel Textbooks. Birkh¨ auser/Springer, Cham, 2019

2019
[25]

M. Struwe. Variational methods, volume 34 of Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge . A Series of Modern Surveys in Mathematics . Springer-Verlag, Berlin, 3rd edition, 2000. Institut de math ´ematiques de Toulouse, 118, route de Narbonne, Toulouse, Fra nce ...

2000

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Antonelli, P

P. Antonelli, P. Cannarsa, and B. Shakarov. Existence an d asymptotic behavior for L2-norm preserving nonlinear heat equations. Calc. Var. Partial Diﬀer. Equ. , 63(4):1–31, 2024

work page 2024

[2] [2]

Antonelli and B

P. Antonelli and B. Shakarov. Existence and Large Time Be havior for a Dissipative Variant of the Rotational NLS Equation. Commun. Math. Sci. , 22(6):1601–1633, 2024

work page 2024

[3] [3]

T. Aubin. Some nonlinear problems in Riemannian geometry . Springer Monographs in Mathematics. Springer- Verlag, Berlin, 1998

work page 1998

[4] [4]

Bao and Q

W. Bao and Q. Du. Computing the ground state solution of Bo se-Einstein condensates by a normalized gradient ﬂow. SIAM J. Sci. Comput. , 25(5):1674–1697, 2004

work page 2004

[5] [5]

Berestycki and P

H. Berestycki and P. L. Lions. Nonlinear scalar ﬁeld equa tions, ii existence of inﬁnitely many solutions. Archive for Rational Mechanics and Analysis , 82:347–375, 1983

work page 1983

[6] [6]

Berestycki and J

H. Berestycki and J. W ei. On least energy solutions to a se milinear elliptic equation in a strip. Discrete Contin. Dyn. Syst. , 28(3):1083–1099, 2010

work page 2010

[7] [7]

Besse, R

C. Besse, R. Duboscq, and S. Le Coz. Gradient ﬂow approach to the calculation of stationary states on nonlinear quantum graphs. Ann. Henri Lebesgue , 5:387–428, 2022

work page 2022

[8] [8]

Boyer and P

F. Boyer and P. Fabrie. Mathematical Tools for the Study of the Incompressible Navi er-Stokes Equations an- dRelated Models, volume 183. Springer Science & Business Media, 2012

work page 2012