Infinite energy quasi-periodic solutions to nonlinear Schr\"odinger equations on $\mathbb R$

W.-M. Wang

REVIEW 3 major objections 6 minor 4 references

Infinite energy quasi-periodic solutions to nonlinear Schr\"odinger equations on $\mathbb R$

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read The paper constructs smooth infinite-energy, space-time quasi-periodic solutions to nonlinear Schrödinger equations on the real line.

desk verdict New construction of infinite-energy space-time quasi-periodic solutions for NLS on R; the key semi-algebraic lemma is imported from the author's prior work, leaving a verification gap that needs closing. read the letter →

arxiv 1908.11627 v3 pith:IYJOVLRN submitted 2019-08-30 math.AP math-phmath.MP

W.-M. Wang This is my paper · ORCID

classification math.APmath-phmath.MP MSC 35Q5535B1537K55

keywords nonlinearSchrödingerequationquasi-periodicsolutionsinfiniteenergyspace-timeNewtonschemeAndersonlocalizationsemi-algebraicgeometrysmalldivisors

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 seeks smooth global solutions to the nonlinear Schrödinger equation $i\partial_t u = -\partial_x^2 u + |u|^{2p}u$ on $\mathbb R$ that are quasi-periodic in both time and space, with two independent frequencies in each direction. Such solutions do not decay as $|x|\to\infty$, so they carry infinite energy and mass, and they have no spatial symmetry. Earlier constructions of this type either required integrability, an exponentially decaying nonlinearity, or a compact spatial domain. The paper claims that for every integer $p\ge 1$ and for a parameter set whose measure is close to full, these infinite-energy solutions exist as convergent series with exponentially decaying Fourier coefficients.

What carries the argument

The main mechanism is a Newton scheme built on a Lyapunov-Schmidt decomposition, together with Green's function estimates for the linearized operator $F'(\theta,\varphi)=D(\theta,\varphi)+\delta^{2p}H$, where $D_{\pm}=\mathrm{diag}[\pm(n\cdot\omega+\theta)+M+(j\cdot\lambda+\varphi+m)^2]$. The auxiliary pair $(\theta,\varphi)$ cannot be bounded by the spatial scale $N$, so the proof splits into a large-$(\theta,\varphi)$ region, handled by Lipschitz eigenvalue variation, and a small-$(\theta,\varphi)$ region, handled by semi-algebraic geometry. A key new ingredient is the use of bi-rational maps $(\lambda,m,M)\leftrightarrow(\lambda,\omega)$ that approximate the diffeomorphism defined by the Q-equations, allowing the algebraic and measure estimates to be transferred back to the original parameters.

What would settle it

Compute, for a single large scale $N$ in region (i), the exceptional set $\Theta_N$ arising from the local Lipschitz eigenvalue curves after composing with the bi-rational map $M_N$; if its sectional measure in $\theta$ or $\varphi$ exceeds $e^{-N^\tau}$ or if a bad box remains at distance at least $N/10$, then the estimates (2.47)-(2.48) fail and the Newton construction cannot be continued.

Watch

Extended reading notes

Core claim

The central claim is that for any $p\ge 1$ and any non-parallel $h_1,h_2\in\mathbb Z^2$, there exists $\delta_0>0$ such that for all $0<\delta<\delta_0$ and fixed $a\in(0,\delta)^2$, there is a Cantor set $G\subset(0,2\pi)^4$ with relative measure at least $1-|\log\delta|^{-1/2}$ and a diffeomorphism $(\lambda,m,M)\mapsto(\lambda,\omega)$ satisfying $|\omega_k-(h_k\cdot\lambda+m)^2-M|=O(\delta^{2p})$, such that for every parameter point in $G$ the equation admits a space-time quasi-periodic solution of the form $u(t,x)=\sum a(n,j)e^{i(n\cdot\omega+M)t}e^{i(j\cdot\lambda+m)x}$ with exponentially decaying Fourier coefficients, close to the two-mode linear wave. The solutions are bifurcations of linear solutions with two time frequencies, and they appear to be the first smooth, global, non-localized solutions without spatial symmetry for a non-integrable equation of this type.

Load-bearing premise

The entire construction hinges on the assertion that the small-divisor estimates proved in the author's earlier standing-wave paper still hold after replacing the flat parameter space by the bi-rational change of coordinates $(\lambda,m,M)\leftrightarrow(\lambda,\omega)$ and working with two frequencies on the whole line; no proof of that transfer is given in this paper.

Editorial extensions

If this is right

The solutions have infinite energy and mass for all time, so they live outside the usual $H^1$ energy space while remaining smooth.
The existence holds for every integer $p\ge 1$, so increasing the power of the nonlinearity does not obstruct the construction.
The parameter set has relative measure at least $1-|\log\delta|^{-1/2}$, so the non-exceptional parameters form most of the phase-frequency space when $δ$ is small.
The time frequencies differ from the linear frequencies by $O(\delta^{2p})$, and the Fourier coefficients decay exponentially, giving a quantitative description of the quasi-periodic wave.
The method is stated to generalize to any dimension and any number of time frequencies independent of the space frequencies, including forced equations.

Reading between the lines

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

If the small-divisor transfer from the standing-wave setting holds as claimed, the same bi-rational Newton scheme could produce analogous infinite-energy quasi-periodic solutions for other translation-invariant PDEs on $\mathbb R$ whose linear symbol is quadratic in the spatial frequency.
The near-full measure bound suggests that at fixed small amplitude one could numerically continue from the linear two-mode wave and observe exponentially decaying Fourier coefficients, providing a concrete signature of the predicted solutions.
The bi-rational approximation step may be replaceable by a tame inverse function theorem, which would decouple the main construction from the algebraic structure of the Q-equations and widen its applicability.

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.

Desk Editor's Note

read the letter

The paper you asked about is a serious piece. It proves existence of smooth, infinite-energy, space-time quasi-periodic solutions (two frequencies each) for the non-integrable NLS on R, for any p≥1. That is new: prior constructions were either space-periodic (Bourgain, Wang) or relied on integrability (Damanik-Goldstein, Boutet de Monvel-Egorova). The method is a real extension of Bourgain's semi-algebraic set method to a non-compact, two-frequency setting, and the overall architecture—Lyapunov-Schmidt splitting, Newton iteration, Green's function estimates—is coherent.

The soft spot is exactly where the reader says it is. The central Main Lemma in Section 2.6, which gives the Green's function estimates in region (i), is not proved in this paper. In Section 2.8 the text says that after changing to the (λ,ω) coordinates, "we are now in the same setting as the Main Lemma in [W5]" and the proof there applies. But the operator here has the ±(n·ω+θ) term and a family of bi-rational coordinate changes, neither of which appears in [W5]. Properties B and V3 are assumed, not verified in detail for the actual Newton iterates; the verification in Section 3.4 is plausible but terse. If the adaptation fails, the Green's function bounds (2.47)–(2.48) collapse and so does the Newton construction.

That said, I don't see an internal contradiction or a circular argument. The proof does not assume the conclusion. The dependence on [W5] is a verification gap, not a sleight of hand. It is common in this area to cite technical lemmas from earlier papers, and the author is the one who proved [W5], so the referral is not lazy. But for a result of this significance, particularly with the two-frequency non-compact twist, the referee should insist on a self-contained proof of the Main Lemma, or at minimum a precise statement of what in [W5] is being used and why the bi-rational map preserves the required degree and determinant bounds.

Who is this for? People working in KAM theory for PDEs, quasi-periodic solutions, and Anderson localization in the nonlinear setting. It deserves a serious referee. I would condition acceptance on the author closing the gap, but the question is worth refereeing, not desk rejecting.

Referee Report

3 major / 6 minor

Summary. The paper constructs smooth, infinite-energy, space-time quasi-periodic solutions of the nonlinear Schrödinger equation i∂_t u = −∂_x^2 u + |u|^{2p}u on R for every integer p ≥ 1. The Ansatz is a two-frequency time and two-frequency space Fourier series of the form (A); the parameters are the spatial frequencies λ ∈ (0,2π)^2 and the phases (m,M), while the time frequencies ω are determined by four Q-equations. The main theorem asserts that for non-parallel h1,h2 ∈ Z^2 and sufficiently small fixed amplitudes a, there is a Cantor set G ⊂ (0,2π)^4 of measure at least (2π)^4(1−|log δ|^{−1/2}) such that the P-equations can be solved by a Newton scheme, yielding a solution close to the linear two-wave solution with exponentially decaying Fourier coefficients. The proof combines a Lyapunov–Schmidt decomposition, Green's function estimates for truncated linearized operators, and semi-algebraic set exclusions; the nonlinear analysis is carried out in Section 3.

Significance. If the main theorem is correct, it represents a notable advance: it provides global, non-decaying, spatially non-symmetric quasi-periodic solutions for a non-integrable NLS on the line, in a setting where the linearized operator is non-elliptic and the problem is non-compact. The manuscript also offers a serious extension of Bourgain's semi-algebraic method beyond the compact/lattice setting. There are genuine contributions in the paper: Lemma 3.1 gives a self-contained proof of nonvanishing of the relevant diagonal matrix elements under h1 ∦ h2; Lemma 3.3 constructs rational approximations of the inverse frequency map with degree control of the form e^{C(log N)^3}, which is the scale required by Property B; and Section 3 contains a coherent induction with explicit scales and measure losses. The main limitation is that the decisive linear estimate in region (i) of the Main Lemma is imported from the author's previous work [W5], and the adaptation to the present more general operator is asserted rather than proved. The significance of the result is therefore conditional on completing that verification.

major comments (3)

[§2.8] The proof of the Main Lemma in region (i) is not contained in this paper. After doubling (θ,φ) to (θ̃,φ̃), the text states: 'In the (λ,ω)-coordinates, we are now in the same setting as the Main Lemma in [W5]. The Lemma and the proof there apply.' According to the manuscript's own description, [W5] treated space quasi-periodic standing waves, i.e., time-independent solutions, with no term ±(n·ω+θ) in the diagonal operator (2.14)-(2.16) and no family of bi-rational maps M_N. The present operator has both features, and the covariance structure is altered by the doubling. Since estimates (2.47)-(2.48) are the input for the Newton step (3.53), the paper must either provide a proof of region (i) in the present setting or cite a version of [W5]'s Main Lemma that explicitly covers two-frequency operators with ±(n·ω+θ) and bi-rational conjugacies, and explain why Lemmas 2.3-2.4 and the Yomdin-Gromov triangulation argument remain valid after this adaptation.
[§2.6 and §3.4] Properties B and V3 are assumptions in the Main Lemma, but their verification for the actual Newton iterates is incomplete. Lemma 3.3 constructs rational maps μ_r satisfying μ_r M_r = I + O(δ_r^2) and M_r μ_r = I + O(δ_r^2), but the Main Lemma requires genuine bi-rational maps M_N with exact inverse M_N^{-1} and the determinant and degree bounds (2.38)-(2.41), as well as deg(w_N∘M_N^{-1}) ≲ e^{(log N)^3} in Property V3. It is not shown that the approximate inverse supplied by Lemma 3.3 can be replaced by an exact rational inverse without changing the Green's function estimates, nor that the degree bounds remain uniform in N after all Newton corrections. This is load-bearing because the Main Lemma is stated for TN(λ,ω;u_N,v_N) and TN(λ,m,M;u_N,v_N) after conjugation by M_N^{-1}.
[§3.3 and §2.4] The Diophantine conditions (Hiv,b2) are only verified for finite frequency ranges j,n ∈ [−A^r,A^r]^2 at each step, while Lemma B in §2.4 assumes global conditions ‖j·λ‖_T ≥ 1/(|log δ||j|^ρ) and ‖n·ω‖_T ≥ 1/(|log δ|^2|n|^ρ) for all nonzero j,n. The transition from finite-range to all-scales Diophantine control, including the dependence of ω^{(r)} on the Newton step through (3.1)-(3.2), is only sketched in the measure estimate (2.49). The authors should spell out the total measure loss after intersecting the exclusions over all scales and show explicitly that it is compatible with the claimed measure bound (1.2).

minor comments (6)

[§2.5] In the definition of a bi-rational map, "an open set I ∈ R^d" should read "an open set I ⊂ R^d".
[§2.6] In the statement of the Main Lemma, the notation is inconsistent: G_N is said to be a semi-algebraic set in (λ,ω), but is also written as G_N ⊂ I, where I is an interval in (λ,m,M); the relation Γ_N = M_N^{-1}(G_N) suggests G_N should be a subset of the image M_N(I). This should be clarified.
[§2.1] The convolution notation (u∗v)^{∗p} is used repeatedly but never defined explicitly; a combinatorial definition would help the reader verify the form of the linearized operator H in (2.10).
[§2.8] In the passage deriving sectional measures after doubling, the assertion that "the set S is independent of θ̃_1−θ̃_2" is not justified; if this independence is needed for the measure comparison, it should be proved.
[§2.3] Lemma A states the estimate (2.22) with exponents σ > τ > 0, but the proof requires additional conditions 0 < sτ < 1 and sσ > 1 relating s and the initial scales (2.18); these hypotheses should be included in the lemma statement.
[Theorem, §1] The theorem states that there is a diffeomorphism (λ,m,M) ↦ (λ,ω) on (0,2π)^4, but the map is singular on the set (h1−h2)·λ = 0, as shown in (3.26) and (3.30); the statement should either restrict the domain or explicitly include this exclusion in the definition of G.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the Newton construction does not assume the target solution; the main self-citation ([W5] Main Lemma) is an external technical tool, though its adaptation to the present bi-rational two-frequency setting is asserted rather than proved.

full rationale

The theorem is proved by a Lyapunov–Schmidt decomposition followed by a multiscale Newton scheme. The P-equations are solved by inverting truncated linearized operators, and the Q-equations determine the time frequencies; neither step fits a parameter to the desired conclusion. The measure estimate (1.2) is obtained by summing the exclusions in (Hiv,d), not by assuming the Cantor set G. The main potentially circular-looking passage is Section 2.8, where the paper says 'In the (λ,ω)-coordinates, we are now in the same setting as the Main Lemma in [W5]. The Lemma and the proof there apply.' This is a load-bearing import of the author's previous Main Lemma for the crucial region (i), and the paper explicitly does not repeat the proof. However, [W5]'s Main Lemma is an independent technical statement with its own hypotheses (Properties B and V3) and does not assume the present NLS theorem; citing it is standard mathematical practice rather than a reduction of the conclusion to its own inputs. The genuine weakness is that the adaptation to the non-compact two-frequency operator with the ±(n·ω+θ) term and the bi-rational maps M_N is asserted ('same setting') rather than verified in this paper; if that assertion fails, estimates (2.47)-(2.48) and the Newton scheme would collapse. That is a verification gap and a correctness risk, not an instance of circularity. No fitted input is renamed as a prediction, and no known result is repackaged as new. The circularity score is therefore low.

Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

The central construction rests on standard nonlinear analysis tools (Newton scheme, implicit function theorems) and on the semi-algebraic framework of Bourgain [B3], plus the author's previous work [W5] for the key Green's function estimates in region (i). No physical entities are introduced. The auxiliary variables (θ,φ) are mathematical devices with no independent physical status.

assumptions (7)

standard math Lipschitz implicit function theorem (Clarke)
Used in Section 2.4 to convert Lipschitz eigenvalue estimates into Lipschitz functions θ_i and φ_i, and in Lemma 3.3 for the inverse diffeomorphism.
standard math Inverse function theorem for C^1 maps
Used in Lemma 3.3 to assert that M_r^{-1} is well-defined on (0,2π)^4 minus the set where (h1-h2)·λ = 0.
standard math Semi-algebraic decomposition Lemma 2.2 (Bourgain)
Stated and used in Sections 2.7 and 2.8 to decompose bad parameter sets into vertical and horizontal components with measure control.
standard math Semi-algebraic variable elimination Lemmas 2.3 and 2.4 (Bourgain)
Used to eliminate the auxiliary variables (θ0,φ0) and to control the measure of bad sets in region (i).
standard math Yomdin-Gromov triangulation theorem
Cited via [BiN] and [Bu] as part of the semi-algebraic toolbox, used implicitly in the proof of the Main Lemma.
domain assumption Main Lemma of [W5] (same author) applies in region (i)
Section 2.8 imports the proof of the Main Lemma in region (i) from the author's prior work [W5] without reproducing it. This is the weakest load-bearing external input.
domain assumption Parameters h1,h2 non-parallel (h1∦h2)
Assumed in the Theorem and used in Lemma 3.1 to ensure no identically zero diagonal elements in the linearized operator, which is essential for the initial Newton steps.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Infinite energy quasi-periodic solutions to nonlinear Schr\"odinger equations on $\mathbb R$." pith.science (2026). https://pith.science/paper/IYJOVLRN

@misc{pith2026190811627,
  author       = {Pith},
  title        = {Pith review of: Infinite energy quasi-periodic solutions to nonlinear Schr\"odinger equations on $\mathbb R$},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IYJOVLRN}},
  note         = {Machine review of arXiv:1908.11627}
}

abstract

We present a set of smooth infinite energy global solutions (without spatial symmetry) to the non-integrable, nonlinear Schr\"odinger equations on $\Bbb R$. These solutions are space-time quasi-periodic with two frequencies each. Previous results [B2,1], and their generalizations [W2-4], are quasi-periodic in time, but periodic in space. This paper generalizes Bourgain's semi-algebraic set method [B3] to analyze nonlinear PDEs, in the non-compact space quasi-periodic setting on $\Bbb R$.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

4 extracted references · 4 canonical work pages

[1]

Basu, On bounding the Betti numbers and computing the Euler charac teristics of semi-algebraic sets , Disc

[Ba] S. Basu, On bounding the Betti numbers and computing the Euler charac teristics of semi-algebraic sets , Disc. Comput. Geom. 22 (1) (1999), 1-18. [BiN] G. Binyamini, D. Novikov, Complex cellular structures , Ann. of Math. 190 (2019), 145-

work page 1999
[248]

On the global well-posedness of the quadratic NLS on $L^2(\mathbb{R}) + H^1(\mathbb{T})$

[B1] J. Bourgain, Quasi-periodic solutions of Hamiltonian perturbations of 2D linear Schr¨ odinger equations, Ann. of Math. 148 (1998), 363-439. [B2] J. Bourgain, Green ’s function estimates for lattice Schr¨ odinger opera tors and applica- tions, Ann. of Math. Studies 158 (2005), Princeton University Press. [B3] J. Bourgain, Anderson localization for qua...

work page Pith review arXiv 1998
[1990]

The nonlinear Schrodinger equation on Z and R with bounded initial data: examples and conjectures

[CKSTT] J. Colliander, M. Keel, G. Staﬃlani, H. Takaoka, T. Tao, Transfer of energy to high frequencies in the cubic defocusing nonlinear Schr¨ odinge r equation , Invent. Math. 181, no. 1 (2010), 39-113. [CW] W. Craig, C. E. Wayne, Newton ’s method and periodic solutions of nonlinear equati ons, Commun. Pure Appl. Math. 46 (1993), 1409-1498. [DG] D. Dama...

work page Pith review arXiv 2010
[1997]

Anderson localization for multi-frequency quasi-periodic operators on $\mathbb{Z}^d$

[JLS] S. Jitomyrskaya, W. Liu, Y. Shi, Anderson localization for multi-frequency quasi-periodi c operators on Zd, Geom. Func. Anal. (to appear), arXiv: 1908.03805. [O1] T. Oh, On nonlinear Schr¨ odinger equations with almost periodic i nitial data , Siam. J. Math. Anal. 47(2) (2015), 1253-1270. 47 [O2] T. Oh, Global existence for the defocusing nonlinear ...

work page Pith review arXiv 2015

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Basu, On bounding the Betti numbers and computing the Euler charac teristics of semi-algebraic sets , Disc

[Ba] S. Basu, On bounding the Betti numbers and computing the Euler charac teristics of semi-algebraic sets , Disc. Comput. Geom. 22 (1) (1999), 1-18. [BiN] G. Binyamini, D. Novikov, Complex cellular structures , Ann. of Math. 190 (2019), 145-

work page 1999

[2] [248]

On the global well-posedness of the quadratic NLS on $L^2(\mathbb{R}) + H^1(\mathbb{T})$

[B1] J. Bourgain, Quasi-periodic solutions of Hamiltonian perturbations of 2D linear Schr¨ odinger equations, Ann. of Math. 148 (1998), 363-439. [B2] J. Bourgain, Green ’s function estimates for lattice Schr¨ odinger opera tors and applica- tions, Ann. of Math. Studies 158 (2005), Princeton University Press. [B3] J. Bourgain, Anderson localization for qua...

work page Pith review arXiv 1998

[3] [1990]

The nonlinear Schrodinger equation on Z and R with bounded initial data: examples and conjectures

[CKSTT] J. Colliander, M. Keel, G. Staﬃlani, H. Takaoka, T. Tao, Transfer of energy to high frequencies in the cubic defocusing nonlinear Schr¨ odinge r equation , Invent. Math. 181, no. 1 (2010), 39-113. [CW] W. Craig, C. E. Wayne, Newton ’s method and periodic solutions of nonlinear equati ons, Commun. Pure Appl. Math. 46 (1993), 1409-1498. [DG] D. Dama...

work page Pith review arXiv 2010

[4] [1997]

Anderson localization for multi-frequency quasi-periodic operators on $\mathbb{Z}^d$

[JLS] S. Jitomyrskaya, W. Liu, Y. Shi, Anderson localization for multi-frequency quasi-periodi c operators on Zd, Geom. Func. Anal. (to appear), arXiv: 1908.03805. [O1] T. Oh, On nonlinear Schr¨ odinger equations with almost periodic i nitial data , Siam. J. Math. Anal. 47(2) (2015), 1253-1270. 47 [O2] T. Oh, Global existence for the defocusing nonlinear ...

work page Pith review arXiv 2015