Open horizons of the long-time wave turbulence justification: fluctuation and large-deviation dynamics, continuation past wave-kinetic blow-up, post-kinetic asymptotics, and optimal time dependence

Deng and Hani (arXiv:2311.10082) proved the first large-data, long-time derivation of a collisional kinetic equation, deriving the homogeneous wave kinetic equation (WKE) from the cubic nonlinear Schrödinger equation (NLS) with random initial data on large tori, valid over the full lifespan of the WKE solution. Their paper explicitly lists the problems this opens. Two particle-side counterparts have since been resolved by Deng, Hani, and Ma: the long-time extension of Lanford's theorem for the hard-sphere Boltzmann equation (arXiv:2408.07818) and the derivation of fluid equations from hard-sphere gases, completing Hilbert's sixth problem for a rarefied gas (arXiv:2503.01800). The surviving open core, on the wave side, comprises four linked questions in the setting of the source's main theorem: (1) rigorously derive the fluctuation dynamics and a large-deviation principle for the NLS spectrum around the WKE over the same long kinetic time interval; (2) determine whether the WKE admits a weak continuation past its finite-time blow-up (condensate formation) and in what sense NLS statistics approximate it beyond that time; (3) characterize the asymptotic behavior of the inhomogeneous WKE / Boltzmann-Nordheim equation and of the NLS beyond the kinetic timescale (the wave-side analogue of the just-completed particle-side fluid limits); (4) determine the optimal dependence of the justified time horizon on system size and the minimal regularity of the initial spectrum. All four remain open as of September 2026.

Background

Wave turbulence theory describes the statistical dynamics of large ensembles of weakly nonlinear dispersive waves. Its central prediction, the wave kinetic equation (WKE), was developed formally in the physics literature from the 1920s onward, but its rigorous derivation from microscopic dynamics is a long-standing mathematical challenge, parallel to the derivation of Boltzmann's equation for classical particles that was posed as part of Hilbert's sixth problem.

The setting is the cubic nonlinear Schrödinger equation (NLS) $(i\partial_t-\Delta)u+\alpha|u|^2u=0$ on the torus $\mathbb{T}^d_L=[0,L]^d$ of side length $L$ ($d\geq 3$), where $\Delta=(2\pi)^{-1}(\partial_{x_1}^2+\cdots+\partial_{x_d}^2)$ is the normalized Laplacian, with weak nonlinearity of strength $\alpha$, and random initial data whose Fourier modes are independent Gaussians with variances prescribed by a nonnegative profile $\varphi_{\mathrm{in}}$. In the kinetic limit $L\to\infty$, $\alpha\to 0$ linked by a scaling law $\alpha=L^{-\gamma}$ ($0<\gamma\leq 1$), the natural observation time is the kinetic (Van Hove) time $T_{\mathrm{kin}}=(2\alpha^2)^{-1}$, at which cumulative nonlinear interactions become visible. The WKE is the equation $\partial_\tau\varphi=\mathcal{K}(\varphi,\varphi,\varphi)$, $\varphi(0)=\varphi_{\mathrm{in}}$, for the limiting wave-action spectrum $\varphi(\tau,k)$, where the cubic collision operator $\mathcal{K}$ integrates over resonant wave-vector quadruples satisfying $k_1-k_2+k_3=k$ and $|k_1|^2-|k_2|^2+|k_3|^2=|k|^2$. It is the wave analogue of the spatially homogeneous Boltzmann equation, and its collision operator is an exact linear combination of those of the classical and quantum (Boltzmann-Nordheim) Boltzmann equations.

Prior status. For classical particles, Lanford's theorem (1975) derives Boltzmann's equation from hard-sphere dynamics, but only for a short time (a fraction of the mean free time), as later completed and refined by Gallagher, Saint-Raymond and Texier; its long-time extension was a major open problem. On the wave side, Lukkarainen and Spohn treated the nonlinear problem at equilibrium, and Deng and Hani rigorously derived the WKE from NLS up to small multiples of $T_{\mathrm{kin}}$ (Invent. Math. 2023; arXiv:2301.07063), together with propagation of chaos, the property that initially independent Fourier modes remain independent in the kinetic limit (the wave analogue of Boltzmann's Stosszahlansatz), and higher-order statistics (J. Eur. Math. Soc. 2026). Deng and Hani (arXiv:2311.10082) extended this to arbitrarily long times: for every $\tau_*<\tau_{\max}$, where $\tau_{\max}$ is the maximal existence time of the Schwartz solution of the WKE, and all large $L$, the NLS with random data has a smooth solution up to $T=\tau_* T_{\mathrm{kin}}$ with overwhelming probability, and the expected modal energies $\mathbb{E}|\hat u(t,k)|^2$ satisfy $\sup_{t\leq T}\sup_{k}\bigl|\mathbb{E}|\hat u(t,k)|^2-\varphi(t/T_{\mathrm{kin}},k)\bigr|\leq L^{-\theta}$. This is the first large-data, long-time derivation in any nonlinear (particle or wave) collisional kinetic limit. The proof propagates a self-consistent long-time cumulant ansatz (a cumulant is a joint cumulant of Fourier-mode amplitudes; its smallness quantifies decay of correlations), organized by Feynman-diagram expansions iterated in time layers.

The source lists the problems this opens (its "future horizons"). For particles: the long-time extension of Lanford's theorem, and the fluid equations beyond Boltzmann. For waves, four questions remain. (i) Fluctuations and large deviations: not just the mean spectrum but the central-limit fluctuations of the empirical spectrum around the WKE, and a large-deviation principle (LDP, i.e., precise exponential bounds on the probability of rare deviations, with an explicit rate function), over the same long time interval; the physics-level prediction exists (Guioth, Bouchet and Eyink). (ii) Behavior past blow-up: the WKE is known to develop finite-time blow-up in the form of a Dirac-mass singularity (Escobedo and Velázquez), expected to be fairly generic and corresponding physically to condensate formation; whether the WKE admits a weak continuation past blow-up, and in what modified sense the NLS statistics approximate it, is unknown. (iii) Asymptotics beyond the kinetic timescale: whether the inhomogeneous WKE or Boltzmann-Nordheim equation admits diffusive or fluid-type limits after the kinetic time, and what this implies for the NLS beyond $T_{\mathrm{kin}}$; even the derivation of the inhomogeneous equations from NLS without noise is incomplete, and the wave-side target equations are not known. (iv) Sharp quantitative reach: the justified horizon $\tau_*$ is only known to grow like $\log\log L$ in the system size $L$ when the WKE solution is globally bounded (improvable to $\log L/\log\log L$ by refining the proof), and the main theorem requires the initial spectrum $\varphi_{\mathrm{in}}$ to be Schwartz (rapidly decreasing together with all its derivatives), with the required regularity necessarily deteriorating as $\tau_*$ grows.

The particle-side problems have been resolved: Deng, Hani and Ma derived the Boltzmann equation for hard spheres for arbitrarily long times (arXiv:2408.07818; confirmed and exposited in the January 2026 Bourbaki seminar of Bodineau, Gallagher, Saint-Raymond and Simonella, arXiv:2602.04407), and then derived the compressible Euler and incompressible Navier-Stokes-Fourier equations from hard-sphere gases (arXiv:2503.01800).

Problem Statement

In the setting of the main theorem of Deng and Hani (arXiv:2311.10082) - cubic NLS on a large torus $\mathbb{T}^d_L$, $d\geq 3$, with independent-Gaussian random initial data, in the kinetic limit $L\to\infty$, $\alpha=L^{-\gamma}$ over the admissible range of scaling laws - extend the long-time kinetic justification of wave turbulence beyond the mean law and characterize its sharp reach. Four linked questions remain open; together they are the surviving core of the source's list of future horizons, the particle-side entries having been resolved separately (Deng, Hani and Ma, arXiv:2408.07818 and arXiv:2503.01800).

(1) Fluctuations and large deviations on long kinetic times. Derive, rigorously from the NLS, the effective fluctuation dynamics of the wave spectrum around the wave kinetic equation over the same long time interval covered by the source's theorem - every $\tau_*<\tau_{\max}$ in units of $T_{\mathrm{kin}}$ - including identification of the fluctuation scaling in the system size $L$ and the Gaussian (central-limit) limit of the fluctuation field of the empirical spectrum (the random finite-$L$ modal spectrum $|\hat u(t,k)|^2$, as opposed to its expectation, which is the mean law), and establish a full large-deviation principle for the spectrum dynamics with explicit speed and rate function. The source conjectures that its cumulant bounds should suffice; no derivation exists.

(2) Continuation past blow-up and condensate formation. For initial data for which the WKE solution blows up at finite time $\tau_{\max}$ with formation of a Dirac-mass singularity (condensation): determine whether the WKE admits a weak continuation past $\tau_{\max}$, and in what (possibly modified) sense the NLS statistics approximate it for times beyond the blow-up time; make this rigorous.

(3) Asymptotics beyond the kinetic timescale, wave side. Characterize the possible asymptotic behavior - diffusive, fluid-type, or other - of the inhomogeneous wave kinetic equation (the variant in which the spectrum also depends on position, $\varphi(\tau,x,k)$, with a transport term $(k\cdot\nabla_x)\varphi$) / Boltzmann-Nordheim equation after the kinetic timescale, noting that even the derivation of these inhomogeneous equations from the NLS without stochastic forcing is incomplete, and determine what these limits imply, rigorously, about the behavior of the NLS beyond the kinetic time. The particle-side analogue (Boltzmann to Euler / Navier-Stokes-Fourier) has been completed; on the wave side even the candidate limiting equations are unknown.

(4) Optimal quantitative dependence and regularity threshold. Determine the optimal growth of the justified time horizon $\tau_*$ with the system size $L$ in the case where the WKE solution is globally bounded - the proof yields $\tau_*\sim\log\log L$ and is refinable to $\log L/\log\log L$; identify the true rate, including matching lower bounds showing longer horizons cannot be justified - and determine the minimal regularity and decay assumptions on the initial spectrum $\varphi_{\mathrm{in}}$ (the main theorem requires Schwartz data; Remark 1.6 of the source), in particular how the required regularity deteriorates with the number of inductive time layers (the proof's inductive time strata) and hence with the time horizon.

The common objective of the four questions is to complete the long-time justification of wave turbulence theory at the level of the full statistics of the NLS (not only the mean law), over its entire relevant time range (up to and beyond blow-up, and beyond the kinetic timescale), and with sharp quantitative dependence on the system size.

References

1. Y. Deng and Z. Hani, 'Long time justification of wave turbulence theory', arXiv:2311.10082 (v1 2023, v2 2024). DOI: 10.48550/arXiv.2311.10082, https://arxiv.org/abs/2311.10082
2. Y. Deng, Z. Hani and X. Ma, 'Long time derivation of the Boltzmann equation from hard sphere dynamics', arXiv:2408.07818 (v1 2024, v3 2025). DOI: 10.48550/arXiv.2408.07818, https://arxiv.org/abs/2408.07818
3. Y. Deng, Z. Hani and X. Ma, 'Hilbert's sixth problem: derivation of fluid equations via Boltzmann's kinetic theory', arXiv:2503.01800 (2025). DOI: 10.48550/arXiv.2503.01800, https://arxiv.org/abs/2503.01800
4. T. Bodineau, I. Gallagher, L. Saint-Raymond and S. Simonella, 'Derivation of the Boltzmann equation from hard-sphere dynamics (after Y. Deng, Z. Hani, and X. Ma)', arXiv:2602.04407 (2026), Bourbaki Seminar, January 2026. DOI: 10.48550/arXiv.2602.04407, https://arxiv.org/abs/2602.04407
5. Y. Deng and Z. Hani, 'Full derivation of the wave kinetic equation', Inventiones Mathematicae 233 (2023), no. 2, 543-724. DOI: 10.1007/s00222-023-01189-2 (arXiv:2104.11204), https://doi.org/10.1007/s00222-023-01189-2
6. Y. Deng and Z. Hani, 'Propagation of chaos and higher order statistics in wave kinetic theory', Journal of the European Mathematical Society 28 (2026), no. 2, 673-733. DOI: 10.4171/jems/1488 (arXiv:2110.04565), https://doi.org/10.4171/jems/1488
7. Y. Deng and Z. Hani, 'Derivation of the wave kinetic equation: full range of scaling laws', arXiv:2301.07063 (2023). DOI: 10.48550/arXiv.2301.07063, https://arxiv.org/abs/2301.07063
8. O. E. Lanford, 'Time evolution of large classical systems', in J. Moser (ed.), Dynamical Systems, Theory and Applications, Lecture Notes in Physics 38, Springer, Berlin (1975), 1-111. DOI: 10.1007/3-540-07171-7_1, https://doi.org/10.1007/3-540-07171-7_1
9. I. Gallagher, L. Saint-Raymond and B. Texier, 'From Newton to Boltzmann: Hard Spheres and Short-range Potentials', EMS Zurich (2014). DOI: 10.4171/129, ISBN 9783037191293, https://doi.org/10.4171/129
10. J. Lukkarainen and H. Spohn, 'Weakly nonlinear Schrödinger equation with random initial data', Inventiones Mathematicae 183 (2011), no. 1, 79-188. DOI: 10.1007/s00222-010-0276-5, https://doi.org/10.1007/s00222-010-0276-5
11. T. Bodineau, I. Gallagher, L. Saint-Raymond and S. Simonella, 'Statistical dynamics of a hard sphere gas: fluctuating Boltzmann equation and large deviations', Annals of Mathematics (2) 198 (2023), no. 3. DOI: 10.4007/annals.2023.198.3.3, https://doi.org/10.4007/annals.2023.198.3.3
12. T. Bodineau, I. Gallagher, L. Saint-Raymond and S. Simonella, 'Long-time derivation at equilibrium of the fluctuating Boltzmann equation', Annals of Probability 52 (2024), no. 1. DOI: 10.1214/23-aop1656, https://doi.org/10.1214/23-aop1656
13. J. Guioth, F. Bouchet and G. L. Eyink, 'Path Large Deviations for the Kinetic Theory of Weak Turbulence', Journal of Statistical Physics 189 (2022), no. 2, article 20. DOI: 10.1007/s10955-022-02971-x, https://doi.org/10.1007/s10955-022-02971-x
14. M. Escobedo and J. J. L. Velázquez, 'Finite time blow-up and condensation for the bosonic Nordheim equation', Inventiones Mathematicae 200 (2015), no. 3, 761-847. DOI: 10.1007/s00222-014-0539-7, https://doi.org/10.1007/s00222-014-0539-7
15. G. Staffilani and M.-B. Tran, 'Condensation and Non-condensation Times for 4-Wave Kinetic Equations', Archive for Rational Mechanics and Analysis 250 (2026), no. 4, article 58. DOI: 10.1007/s00205-026-02217-0 (arXiv:2407.18533), https://arxiv.org/abs/2407.18533
16. G. Staffilani and M.-B. Tran, 'Formation of condensations for non-radial solutions to 3-wave kinetic equations', arXiv:2503.17066 (2025). DOI: 10.48550/arXiv.2503.17066, https://arxiv.org/abs/2503.17066
17. M. Escobedo and A. Menegaki, 'Global Well-Posedness near Rayleigh-Jeans Equilibria for the Cubic NLS Wave Kinetic Equation', arXiv:2608.19483 (2026). DOI: 10.48550/arXiv.2608.19483, https://arxiv.org/abs/2608.19483
18. M. Escobedo and J. J. L. Velázquez, 'On the onset of correlations in Wave Turbulence close to singularities', arXiv:2605.02540 (2026). DOI: 10.48550/arXiv.2605.02540, https://arxiv.org/abs/2605.02540
19. B. Douet and F. Bouchet, 'Dynamical large deviations and long-range correlations for local weak wave turbulence', arXiv:2606.12624 (2026). DOI: 10.48550/arXiv.2606.12624, https://arxiv.org/abs/2606.12624
20. R. Grande and Z. Hani, 'Rigorous Derivation of Damped–Driven Wave Turbulence Theory', Archive for Rational Mechanics and Analysis 250 (2026), no. 3, article 27. DOI: 10.1007/s00205-026-02181-9 (arXiv:2407.10711), https://arxiv.org/abs/2407.10711
21. Z. Hani, J. Shatah and H. Zhu, 'Inhomogeneous turbulence for the Wick nonlinear Schrödinger equation', arXiv:2309.12037 (v1 2023, v2 2024). DOI: 10.48550/arXiv.2309.12037, https://arxiv.org/abs/2309.12037
22. F. Xu and Y. Li, 'Large Deviations for the Nonlinear Schrödinger Equation with Randomized Quasi-Periodic Initial Data in Higher Dimensions: Subcritical Case', arXiv:2604.17253 (2026). DOI: 10.48550/arXiv.2604.17253, https://arxiv.org/abs/2604.17253
23. Y. Bruned and V. Clarisse, 'Kruskal-style algorithm for cubic Schrödinger equation molecule reduction', arXiv:2603.23298 (2026). DOI: 10.48550/arXiv.2603.23298, https://arxiv.org/abs/2603.23298
24. Y. Bruned and L. Tolomeo, 'Cancellations for dispersive PDEs with random initial data', Probability and Mathematical Physics 7 (2026), no. 3, 383-405. DOI: 10.2140/pmp.2026.7.383 (arXiv:2412.17051), https://doi.org/10.2140/pmp.2026.7.383