Central limit theorem for stochastic nonlinear wave equation with pure-jump Lévy white noise

Raluca M. Balan and Guangqu Zheng · 2025

Read Central limit theorem for stochastic nonlinear wave equation with pure-jump Lévy white noise by Raluca M. Balan and Guangqu Zheng (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-18672.txt.

Introduction

A wave equation shaken by jumpy Lévy noise, think random kicks rather than smooth static, defies the usual solution formulas. Raluca Balan and Guangqu Zheng study in 2025 its random field solution when the noise multiplies the wave's own height. With Itô and Malliavin calculus they prove the solution can be differentiated in the Malliavin sense, with sharp bounds on those derivatives.

From there, spatial averages over widening intervals settle to a law of large numbers, then fluctuate in a Gaussian way, with a measured speed in Wasserstein distance. Functional and almost-sure versions follow, plus asymptotic independence of averages. The nonlinear result extends their earlier linear work without leaning on explicit expansions.

Catalog record

CREATORRaluca M. Balan and Guangqu Zheng
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE139 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-18672

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Central limit theorem for stochastic nonlinear wave equation with pure-jump Lévy white noise Raluca M. Balan and Guangqu Zheng Abstract . In this paper, we study the random field solution to the stochastic nonlinear wave equation (SNLW) with constant initial conditions and multiplicative noise σ ⁡ ( u ) ​ L ˙ \sigma(u)\dot{L} , where the nonlinearity is encoded in a Lipschitz function σ : ℝ → ℝ \sigma:\mathbb{R}\to\mathbb{R} and L ˙ \dot{L} denotes a pure-jump Lévy white noise on ℝ + × ℝ \mathbb{R}_{+}\times\mathbb{R} with finite variance. Combining tools from Itô calculus and Malliavin calculus, we are able to establish the Malliavin differentiability of the solution with sharp moment bounds for the Malliavin derivatives. As an easy consequence, we obtain the spatial ergodicity of the solution to SNLW that leads to a law of large number result for the spatial integrals of the solution over [ − R , R ] [-R,R] as R → ∞ R\to\infty . One of the main results of this paper is the obtention of the corresponding Gaussian fluctuation with rate of convergence in Wasserstein distance. To achieve this goal, we adapt the discrete Malliavin-Stein bound from Peccati, Solé, Taqqu, and Utzet ( Ann. Probab. , 2010), and further combine it with the aforementioned moment bounds of Malliavin derivatives and Itô tools. Our work substantially improves our previous results ( Trans. Amer. Math. Soc. , 2024) on the linear equation that heavily relied on the explicit chaos expansion of the solution. In current work, we also establish a functional version, an almost sure version of the central limit theorems, and the (quantitative) asymptotic independence of spatial integrals from the solution. The asymptotic independence result is established based on an observation of L. Pimentel ( Ann. Probab. , 2022) and a further adaptation of Tudor’s generalization ( Trans. Amer. Math. Soc. , 2025) to the Poisson setting. 2020 Mathematics Subject Classification Primary 60H15; secondary 60F05, 60F17, 60H07, 60G55 1. Introduction The study of stochastic partial differential equations (SPDEs) is an active research area, which has been growing tremendously over the past four decades, providing the rigorous framework for modeling random phenomena that evolve in space and time. Among various approaches in the literature, the random field approach initiated by Walsh in his Saint-Flour notes [ 49 ] has been very influential, being embraced quickly by the scientific community. 1 1 1 A random field solution is a collection of real-valued random variables indexed by space-time. Walsh’s random field approach was built on Itô’s theory of stochastic integration, and was further pushed forward by Dalang in [ 17 ] . As a result, the random field approach also bears the name of Dalang-Walsh’s approach. Two families of equations that have been extensively studied using this approach are the stochastic heat equations (SHEs) and the stochastic wave equations (SWEs). Pragmatically speaking, an SPDE often describes a complex system with two main inputs (i.e., the initial condition and the noise). We will not discuss about the initial condition here, but we will make a few comments about the noise within the random field approach. The classical example is the stochastic (nonlinear) wave equation posed on \mathbb{R}_{+}\times\mathbb{R} : (1.1) \displaystyle(\partial_{t}^{2}-\partial_{x}^{2})u(t,x)=\sigma(u(t,x))\dot{W}(t,x), where \sigma:\mathbb{R}\to\mathbb{R} is deterministic, Lipschitz continuous, and \dot{W} stands for the space-time Gaussian white noise on \mathbb{R}_{+}\times\mathbb{R} . This is the running example in Walsh’s notes [ 49 ] . The space-time Gaussian white noise \dot{W} is a generalization of the Brownian motion, and one can develop the corresponding Itô integration theory to study the above SWE ( 1.1 ). In fact, the above SWE admits a unique random field solution only in spatial dimension one, and the same comment applies to SHE (with \partial_{t}^{2}-\partial_{x}^{2} replaced by the heat operator \partial_{t}-\frac{1}{2}\partial_{x}^{2} ). When the spatial dimension is strictly bigger than one, the above SPDEs fall into the realm of singular SPDEs and thus do not admit a random-field solution. In a seminal work [ 17 ] , Dalang introduced the Gaussian noises that are white in time and spatially homogeneous such that the spectral measure of spatial correlation function satisfies certain mild integrability assumption (known as the Dalang’s condition nowadays). In Dalang’s generalized framework, one retains the martingale features necessary for Itô’s theory due to the noises being white-in-time. Another popular example of Gaussian noises is the fractional Gaussian noise, which is a generalization of the fractional Brownian motion. In this case, due to the absence of martingale structures (when the Hurst index in the temporal direction is different from 1/2 ), one has to bring in other techniques to develop a solution theory. For example, when \sigma(u)=u (i.e., we consider a linear equation), one can apply the Malliavin calculus (particularly Wiener chaos expansion) to establish a solution theory; see, e.g., [ 24 , 7 ] . For the general theory of Malliavin calculus, we refer interested readers to Nualart’s monograph [ 33 ] ; see also [ 34 ] that presents a pedagogical introduction to Malliavin calculus in both Gaussian and Poisson settings. In this paper, we consider another family of driving noises, the space-time Lévy noises that generalize the classical Lévy processes; see [ 2 ] for the corresponding Itô integration theory and its applications to SPDEs. An in-depth analysis of various solution properties of such SPDEs driven by finite-variance Lévy noise is possible using a discrete version of Malliavin calculus, with respect to the underlying Poisson random measure of the noise. More details about Malliavin calculus in the Poisson setting can be found in [ 27 , 29 , 30 , 40 ] . Among many studies of various solution properties of SPDEs, a problem that comes persistently in the recent literature is related to the asymptotic behaviour of the spatial integral of the solution, when the integration region tends to cover the whole space. Such a problem was initiated by Huang, Nualart, and Viitasaari in their paper [ 25 ] , which established the Gaussian fluctuation of the spatial averages of the SHE with space-time Gaussian white noise. Since then, there has been a rapidly growing literature on the limit theorems for these spatial statistics. This research theme also includes SWEs and spatially homogeneous Gaussian noises and/or fractional Gaussian noises. The following sample of relevant references will convince the reader of the order of magnitude of this growing area: [ 26 , 19 , 11 , 5 , 14 , 15 , 37 , 38 , 35 , 39 , 36 , 43 , 21 , 22 ] . In a nutshell, by combining Stein’s method for normal approximation with Malliavian calculus techniques, one can prove that with appropriate normalization, the spatial integral converges in distribution to the standard normal distribution, and can provide precise estimates for the rate of this central convergence, yielding a so-called qualitative central limit theorem (QCLT). From this, without a significant amount of effort, one can also usually derive a functional central limit theorem (FCLT) for the spatial integral process. In a recent article [ 9 ] , we initiated a similar program for the stochastic wave equation with pure-jump Lévy white noise \dot{L} . More precisely, we studied the hyperbolic Anderson model , in which the noise takes the

Read the complete text