Small Ball Probabilities for the Stochastic Heat Equation on Compact Manifolds
Jiaming Chen · 2026
Read Small Ball Probabilities for the Stochastic Heat Equation on Compact Manifolds by Jiaming Chen (2026) 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-2601-20794.txt.
Introduction
Jiaming Chen asks how rarely a noisy heat flow stays small on a curved space. The 2026 paper studies small ball chances for the stochastic heat equation. The heat spreads on a compact smooth curved space with no edge. Random noise drives the heat at each point. A small ball chance is the odds that the whole flow stays in a tiny tube.
The noise is Gaussian and white in time. The drift term rests on time, place, and the flow itself. The shape of the space shapes how heat and noise mix. Sharp rates for these small chances guide tests of the equation. They also guide simulations that must respect rare calm events. The paper extends flat-space ideas to worlds with shape.
Catalog record
| CREATOR | Jiaming Chen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 81 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-20794
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Small Ball Probabilities for the Stochastic Heat Equation on Compact Manifolds Jiaming Chen Abstract . We consider the stochastic heat equation on a compact smooth Riemannian manifold without boundary satisfying ∂ t u ( t , x ) = 1 2 Δ M u ( t , x ) + σ ( t , x , u ) W ˙ ( t , x ) , ( t , x ) ∈ ℝ + × M , \partial_{t}u(t,x)=\frac{1}{2}\Delta_{M}u(t,x)+\sigma(t,x,u)\dot{W}(t,x),\quad(t,x)\in\mathbb{R}_{+}\times M, where W ˙ \dot{W} is a centered Gaussian noise that is white in time and colored in space. Assuming that σ \sigma is Lipschitz in u u and uniformly bounded, we estimate small ball probabilities for the solution u u when u ( 0 , x ) ≡ 0 u(0,x)\equiv 0 . 2020 Mathematics Subject Classification Primary, 60H15; Secondary, 58J65. 1. Introduction Stochastic partial differential equations (SPDEs) of parabolic type, and in particular the stochastic heat equation, form a central class of models in probability theory. In the classical Euclidean setting, the stochastic heat equation (1.1) \partial_{t}u(t,x)=\frac{1}{2}\Delta u(t,x)+\sigma(t,x,u)\dot{W}(t,x) has been studied extensively using stochastic convolution techniques and the theory of stochastic evolution equations in Hilbert spaces (see [ 29 ] and [ 11 ] for details). These foundational works establish existence, uniqueness, and regularity of solutions under various assumptions on the nonlinearity and the driving noise. More recently, fine properties of solutions, including intermittency, sample path regularity, and tail behavior, have attracted significant attention (see [ 20 , 16 , 6 , 9 , 19 , 18 , 26 ] ). In many applications, however, the underlying state space naturally carries a nontrivial geometric structure. This motivates the study of stochastic heat equations on Riemannian manifolds. Let M be a d -dimensional compact smooth Riemannian manifold without boundary. We consider the stochastic heat equation (1.2) \partial_{t}u(t,x)=\frac{1}{2}\Delta_{M}u(t,x)+\sigma(t,x,u)\dot{W}(t,x),\quad(t,x)\in\mathbb{R}_{+}\times M, with deterministic initial condition u(0,\cdot)=u_{0}:M\to\mathbb{R} , where \Delta_{M}=\mathrm{div}\circ\mathrm{grad} denotes the Laplace-Beltrami operator. A random field \{u(t,x)\}_{(t,x)\in\mathbb{R}_{+}\times M} is said to be a mild solution to ( 1.2 ) if it satisfies (1.3) u(t,x)=\int_{M}P_{t}(x,y)u_{0}(y)dy+\int_{[0,t]\times M}P_{t-s}(x,y)\sigma(s,y,u(s,y))W(dsdy), where P_{t}(x,y) denotes the heat kernel on M . Precise estimates on P_{t} can be obtained under geometric assumptions such as lower bounds on the Ricci curvature (see [ 22 ] and [ 14 ] ). These analytic properties play a crucial role in the study of SPDEs on manifolds. The stochastic integral in ( 1.3 ) is understood in the Itô-Walsh sense. The Itô-Walsh solution theory for ( 1.2 ) is well developed in spatial dimension d=1 when \dot{W} is space-time white noise. In higher dimensions, however, white noise becomes too singular, and it is well known that no function-valued solution exists for the stochastic heat equation driven by space-time white noise when d\geq 2 . To study the equation in higher dimensions, it is therefore natural to regularize the noise in the spatial variable. In the Euclidean setting \mathbb{R}^{d} , Dalang [ 13 ] extended Walsh’s theory to d\geq 2 by considering noises that are white in time and colored in space, with homogeneous spatial covariance G(x,x^{\prime})=G(x-x^{\prime}) . A necessary and sufficient condition for the existence of a random field solution is \int_{\mathbb{R}^{d}}\frac{\hat{G}(d\xi)}{1+|\xi|^{2}}<\infty, where \hat{G} denotes the Fourier transform of G . This condition, commonly referred to as Dalang’s condition, can be interpreted as a regularity requirement on the noise. In contrast to the Euclidean case, where Fourier analysis provides a natural tool via Bochner’s theorem, the construction on a compact manifold relies on the spectral decomposition of the Laplace-Beltrami operator, which plays the role of a Fourier transform in this geometric setting. The statement below is to construct an intrinsic family of Gaussian noises on compact Riemannian manifolds that are smoother than white noise and allow the study of ( 1.2 ) within the Itô-Walsh framework. Let L^{2}(M) denote the space of square-integrable functions on M , equipped with the inner product \langle\cdot,\cdot\rangle . Let \{\phi_{n}\}_{n\geq 0} be an orthonormal basis of L^{2}(M) consisting of eigenfunctions of -\Delta_{M} , with corresponding eigenvalues 0=\lambda_{0}<\lambda_{1}\leq\lambda_{2}\leq\cdots . Thus, -\Delta_{M}\phi_{n}=\lambda_{n}\phi_{n} , and for any \varphi\in L^{2}(M) there exists a unique expansion (1.4) \varphi(x)=\sum_{n\geq 0}a_{n}\phi_{n}(x). In particular, a_{0}=m_{0}^{-1/2}\int_{M}\varphi dm , where m_{0} denotes the volume of M . To define a function-valued solution to ( 1.2 ), we introduce a family of spatial Gaussian noises with covariance structure depending on parameters \alpha,\rho\geq 0 . Let (\Omega,\mathcal{F},P) be a complete probability space and let W(\varphi) and W(\psi) be centered Gaussian random variables for each \varphi and \psi in L^{2}(M) , with covariance (1.5) \mathbb{E}[W(\varphi)W(\psi)]=\langle\varphi,\psi\rangle_{\alpha,\rho}:=\rho a_{0}b_{0}+\sum_{n\neq 0}\frac{a_{n}b_{n}}{\lambda_{n}^{\alpha}}, where \{a_{n}\} and \{b_{n}\} are the coefficients of \varphi and \psi in decomposition ( 1.4 ), respectively. For \rho>0 , let \mathcal{H}^{\alpha,\rho} be the completion of L^{2}(M) with respect to \langle\cdot,\cdot\rangle_{\alpha,\rho} . Remark 1.1 . From ( 1.5 ), we have the continuous embeddings L^{2}(M)\subset\mathcal{H}^{\alpha,\rho}\subset\mathcal{H}^{\beta,\rho} for 0\leq\alpha<\beta . Moreover, the choice \rho=1 and \alpha=0 recovers spatial white noise on M . For \alpha,\rho>0 , the covariance \langle\cdot,\cdot\rangle_{\alpha,\rho} admits an integral kernel representation. Define G_{\alpha}(x,y):=\frac{1}{\Gamma(\alpha)}\int_{0}^{\infty}t^{\alpha-1}\left(P_{t}(x,y)-\frac{1}{m_{0}}\right)dt\text{ and }G_{\alpha,\rho}(x,y):=\frac{\rho}{m_{0}}+G_{\alpha}(x,y). Using the spectral expansion P_{t}(x,y)=\frac{1}{m_{0}}+\sum_{n=1}^{\infty}e^{-\lambda_{n}t}\phi_{n}(x)\phi_{n}(y) and the identity \int_{0}^{\infty}t^{\alpha-1}e^{-\lambda t}dt=\Gamma(\alpha)\lambda^{-\alpha} , we obtain G_{\alpha,\rho}(x,y)=\frac{\rho}{m_{0}}+\sum_{n=1}^{\infty}\lambda_{n}^{-\alpha}\phi_{n}(x)\phi_{n}(y). Consequently, (1.6) \langle\varphi,\psi\rangle_{\alpha,\rho}=\int_{M^{2}}\varphi(x)G_{\alpha,\rho}(x,y)\psi(y)m(dx)m(dy). Remark 1.2 . Throughout the paper, we assume that Dalang’s condition \alpha>(d-2)/2 holds. Remark 1.3 . The kernel G_{\alpha} can be viewed as an analogue of the Riesz kernel on \mathbb{R}^{d} . Although \int_{M}G_{\alpha}(x,y),m(dy)=0 and G_{\alpha} is not nonnegative, it was shown in [ 9 ] that G_{\alpha} is bounded below on M . Hence, by choosing \rho sufficiently large, the covariance function G_{\alpha,\rho} can be made nonnegative. Proposition 1.1 (Lemma 2.9 in [ 5 ] ) . For any \alpha>0 , there exists a constant C_{\alpha}>0 such that |G_{\alpha}(x,y)|\leq\begin{cases}C_{\alpha}&\alpha>d/2\\ C_{\alpha}(1+\log^{-}(d(x,y)))&\alpha=d/2\\ C_{\alpha}d(x,y)^{2\alpha-d}&\alpha<d/2\end{cases} where \log^{-}(z)=\max(0,-\log z) and d(x,y) denotes the Riemannian distance on M . We now define the noise on \mathbb{R}_{+}\times M that is white in time and colored in space. Definition 1.1 . Let \alpha>0 and define the following Hilbert space of space-time functions \mathcal{H}_{\alpha,\rho}:=L^{2}(\mathbb{R}_{+};\mathcal{H^{\alpha,\rho}}(M)) equipped with \langle\cdot,\cdot\rangle_{\mathcal{H}_{\alpha,\rho}} such that for any \varphi,\psi\in\mathcal{H}_{\alpha,\rho} , \langle\varphi,\psi\rangle_{\mathcal{H}_{\alpha,\rho}}:=\int_{0}^{\infty}\langle\varphi(t,\cdot),\psi(t,\cdot)\rangle_{\alpha,\rho}dt. On a complete probability space (\Omega,\mathcal{F},P) , let W_{\alpha,\rho} be a centered isonormal Gaussian process over \mathcal{H}_{\alpha,\rho} , i.e., \mathbb{E}[W_{\alpha,\rho}(\varphi)W_{\alpha,\rho}(\psi)]=\langle\varphi,\psi\rangle_{\mathcal{H}_{\alpha,\rho}} for each \varphi,\psi in \mathcal{H}_{\alpha,\rho} . We refer to W_{\alpha,\rho} as colored noise on M that is white in time. We impose the following assumptions on the noise coefficient \sigma . Hypothesis 1.1 . There exists a constant \mathcal{D}>0 such that for all t\geq 0 , x\in M and u,v\in\mathbb{R} , (1.7) |\sigma(t,x,u)-\sigma(t,x,v)|\leq\mathcal{D}|u-v|. Hypothesis 1.2 . There exist constants \mathcal{C}_{1} , \mathcal{C}_{2}>0 such that for all t\geq 0 , x\in M and u\in\mathbb{R} , (1.8) \mathcal{C}_{1}\leq\sigma(t,x,u)\leq\mathcal{C}_{2}. Remark 1.4 . The first assumption ensures that \sigma is Lipschitz continuous in the solution variable, while the second guarantees uniform nondegeneracy of the noise. Under Hypothesis ( 1.7 ) and Dalang’s condition, it is well known that equation ( 1.2 ) admits a unique random field solution. Small ball problems have been well studied and one can see [ 23 ] for an overview of known results on Gaussian processes and references on other processes. In summary, given a stochastic process X_{t} starting at the origin, the small ball problem concerns the asymptotic behavior of P\left(\sup_{0\leq t\leq T}|X_{t}|<\varepsilon\right)\text{ as }\varepsilon\downarrow 0. In our