On conditional uniqueness of solutions to stochastic Navier-Stokes Equations

István Gyöngy, Nicolai V. Krylov · 2025

Read On conditional uniqueness of solutions to stochastic Navier-Stokes Equations by István Gyöngy, Nicolai V. Krylov (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-2503-20645.txt.

Introduction

István Gyöngy and Nicolai Krylov prove when noisy fluid flow has just one possible course in a 2025 paper, continuing their earlier work. The equations at stake add random shaking to the classic rules for moving fluids. A solution might exist yet fail to be the only one, which would break prediction. Their theorems give uniqueness under stated conditions: any two solutions in the class must match. Partner results show the solutions are also smoother than first assumed, across the whole space and in each dimension.

Catalog record

CREATORIstván Gyöngy, Nicolai V. Krylov
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE69 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-20645

Browse the full library · About the Tower of Babel

Full text

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

On conditional uniqueness of solutions to stochastic Navier-Stokes Equations István Gyöngy, Nicolai V. Krylov Abstract . This paper is a continuation of [ 26 ] . Here theorems on conditional uniqueness and regularity for solutions to stochastic Navier-Stokes equations in ℝ d \mathbb{R}^{d} are presented. 2020 Mathematics Subject Classification 35Q30, 35R60, 60H15 1. Introduction The Navier-Stokes equations, \frac{\partial}{\partial t}u=\nu\Delta-u_{(u)}-\nabla p+D_{j}\mathfrak{f}_{j}+f,\quad\text{div}\,u=0 (1.1) for t\geq 0 and x\in\mathbb{R}^{d} , with initial condition u_{t}\big|_{t=0}=u_{0} (1.2) for the evolution of the velocity and pressure fields u=(u^{1}_{t}(x),...,u^{d}_{t}(x)) and p=p_{t}(x) , are among the most important equations in fluid dynamics and are also among the most studied PDEs in the literature. Here \nu is a positive constant, u_{(u)}=u^{j}D_{j}u , D_{j}\mathfrak{f}_{j}=(\tfrac{\partial}{\partial x^{j}}\mathfrak{f}^{1}_{j},....,\tfrac{\partial}{\partial x^{j}}\mathfrak{f}^{d}_{j}),\quad f=(f^{1},...,f^{d}) are given force fields for t\geq 0 and x=(x^{1},...,x^{d})\in\mathbb{R}^{d} , and u_{0}=(u_{0}^{1},...,u_{0}^{d}) is a given velocity field on \mathbb{R}^{d} such that \text{div}\,\mathfrak{f}_{j}=\text{div}\,u_{0}=0 . By the classical results of Hopf [ 27 ] and Leray [ 37 ] for any u_{0}\in\mathcal{H} at least one weak solution to ( 1.1 )-( 1.2 ) exists in the Hopf-Leray class , \mathcal{W}:=L_{\infty}([0,\infty),\mathcal{H})\cap L_{2}([0,\infty),\mathcal{V})), where \mathcal{V}:=\{u\in W^{1}_{2}(\mathbb{R}^{d},\mathbb{R}^{d}):{\rm div}\,u=0\} and \mathcal{H} is the closure of \mathcal{V} in L_{2}(\mathbb{R}^{d},\mathbb{R}^{d}) . (See, e.g., [ 35 ] , [ 36 ] , [ 38 ] , [ 50 ] , [ 21 ] and [ 48 ] for presentations of Leray and Hopf results and for further developments.) Due to a famous theorem of Ladyzhenskaya, the uniqueness of the weak solution in this class is known when d=2 , and it is an open problem for d\geq 3 . There is, however a considerable literature on conditional uniqueness of the solution under various criteria. The most well-known criterium is the Ladyzhenskaya-Prodi-Serrin condition, which requires that for a T>0 the Hopf-Leray weak solution be in the space L_{p,q}=L_{p,q}([0,T]\times\mathbb{R}^{d},\mathbb{R}^{d}) such that \frac{d}{p}+\frac{2}{q}\leq 1\quad\text{for some $p\in(d,\infty]$ and $q\in[2,\infty)$}, (1.3) where L_{p,q} denotes the space of \mathbb{R}^{d} -valued functions v on [0,T]\times\mathbb{R}^{d} such that |v|^{q}_{L_{p,q}}:=\int_{0}^{T}|v_{t}|^{q}_{L_{p}}\,dt<\infty. Under this condition the uniqueness of the Hopf-Leray weak solution on [0,T] was proved by Prodi [ 47 ] and Serrin [ 49 ] , and smoothness of the solution was obtained by Ladyzhenskaya [ 46 ] . Our aim in the present paper is to generalise the conditional uniqueness result of Prodi and Serrin and to extend it stochastic Navier-Stokes equations and to prove also a theorem on conditional regularity of their solutions. Our condition is much weaker than the Ladyzhenskaya-Prodi-Serrin condition (see Corollary 2.3 ). We will consider the following type of stochastic Navier-Stokes equations, \displaystyle du= \displaystyle\big(\nu\Delta u-u_{(u)}+\nabla p+D_{j}\mathfrak{f}_{j}+f\big)\,dt \displaystyle+\big(\sigma\nabla u-\nabla q+h\big)\circ dw,\quad\text{div}\,u=0, (1.4) \displaystyle u_{t}\big|_{t=0}= \displaystyle u_{0},\quad (1.5) for the random velocity field u=(u_{t}^{1}(x),...,u^{d}_{t}(x)) , and pressure fields p=p_{t}(x) and q=q_{t}(x) for t\in[0,T] and x\in\mathbb{R}^{d} , where w is a Hilbert space-valued Wiener process, and \circ dw_{t} indicates that the corresponding differential is understood in the Stratonovich sense. As it was shown in [ 41 ] and [ 42 ] , equation ( 1.4 ) arises when one follows the classical scheme of Newtonian fluid dynamics, assuming that instead of a deterministic dynamics, the fluid particles satisfy the stochastic equation d\eta_{t}(x)=u_{t}(\eta_{t}(x))\,dt+\sigma_{t}(\eta_{t}(x))\circ dw_{t},\quad\eta_{0}(x)=x,\quad x\in\mathbb{R}^{d}. In [ 44 ] the authors study the Cauchy problem ( 1.4 )-( 1.5 ) with a second order elliptic differential operator instead of \Delta , and with force fields D_{j}\mathfrak{f}_{j} , f and h , which may depend also on u . They show the existence of a solution on some filtered probability space (\Omega,\mathcal{F},(\mathcal{F}_{t})_{t\geq 0},P) carrying an \mathcal{F}_{t} -Wiener process w with values in a Hilbert space. Moreover, in d=2 they show the existence and uniqueness of a unique strong solution. The solution u=(u_{t})_{t\in[0,T]} in [ 44 ] is understood as an \mathcal{H} -valued weakly continuous \mathcal{F}_{t} -adapted process such that u\in L_{2}([0,T],\mathcal{V}) (a.s.), and almost surely satisfies the equations in a weak sense, similarly to the definition of a Hopf-Leray weak solution to ( 1.1 )-( 1.2 ). We consider for d\geq 3 the type of stochastic Navier-Stokes equations as in [ 44 ] . Adapting the notion of admissible random functions from [ 28 ] and [ 26 ] , we introduce the class of admissible solutions to them, see Definitions 2.1 and 2.2 . Roughly speaking an admissible function u is function of (\omega,t,x) , such that it can be decomposed as a sum of two functions, say u^{M} and u^{B} , such that u^{M} , as a function of x\in\mathbb{R}^{d} , belongs to the Morrey space E_{r,1} for an r\in(2,d] and its Morrey norm is bounded by a constant, uniformly in (t,\omega) . The other component, u^{B} belongs to L_{2}\big([0,T],L_{\infty}(\mathbb{R}^{d},\mathbb{R}^{d})\big) (a.s.). First we show that an admissible solution has a modification which is a (strongly) continuous \mathcal{H} -valued function in t\in[0,T] for every \omega\in\Omega . This is Theorem 2.1 below. Then we present a theorem, Theorem 2.2 on conditional uniqueness, which in the special case of equation ( 1.4 ) reads as follows. There is a positive constant N depending only d , r , such that if u is an admissible solution with Morrey norm bounded by \nu/N , then every admissible solution coincides with u . Hence, see Corollary 2.3 , we get that if u is a solution such that u\in L_{p,q} almost surely for a (p,q) satisfying the Ladyzhenskaya-Prodi-Serrin condition ( 1.3 ), then the uniqueness holds in the class of admissible solutions. This is an immediate consequence of Theorem 2.2 , since it turns out that if u is a function from L_{p,q} (a.s.) such the condition ( 1.3 ) holds then u is admissible and admits a Morrey component with Morrey norm bounded by a positive constant as small as we wish. It is worth noting that the admissible functions form an essentially larger space than L_{p,q} with ( 1.3 ). Thus Theorem 2.2 generalises the result of Prodi and Serrin in the special case of deterministic Navier-Stokes equations, i.e., when \sigma=0 , q=0 and h=0 , and extends that also to the type of stochastic Navier-Stokes equations ( 1.4 )-( 1.5 ). Under the conditions of Theorem 2.2 we present also a result, Theorem 5.1 , on conditional regularity of the solution, when u_{0}\in\mathcal{V} (a.s.), \mathfrak{f}=0 , and \sigma and h satisfy appropriate differentiability conditions. We will prove further results on conditional regularity in the continuation of the present paper. Stochastic Navier-Stokes equations are also extensively studied in the literature. We mention here only a few publications concerning their theory, starting with the pioneering paper [ 7 ] , followed by [ 2 ] , [ 11 ] , [ 25 ] [ 10 ] , [ 12 ] , [ 6 ] , [ 14 ] and others on solvability of the equations in various set-ups and approaches. Uniqueness results in d=2 can be found, e.g., in [ 3 ] , [ 40 ] [ 4 ] , [ 24 ] , [ 44 ] , [ 43 ] . Existence of local L_{p} -solutions to Navier-Stokes equations with Lévy noise is studied in [ 45 ] . The existence and uniqueness of solutions in an L_{p} setting for 2D stochastic Navier-Stokes with jump noise is investigated in [ 52 ] , the local well-posedness in L_{p} -setting for 3D

Read the complete text