On Young regimes for locally monotone SPDEs

TL;DR

This paper establishes the global existence of weak solutions for a class of locally monotone SPDEs using a pathwise approach based on monotone operator theory and Besov rough analysis, applicable to various operators and noise regimes.

Contribution

It introduces a novel pathwise framework for SPDEs under local monotonicity, extending Young regime analysis to a broad class of operators and noise types without probabilistic assumptions.

Findings

Proves global existence of weak solutions under local monotonicity and growth conditions.

Identifies noise regimes, including the Brownian borderline, for applicability of the method.

Applies to operators like p-Laplace, porous medium, and shear-thickening fluid operators.

Abstract

We consider the following SPDE on a Gelfand-triple $(V, H, V^*)$: $$ du(t)=A(t,u(t))dt+dI_t(u), \qquad u(0)=u_0\in H. $$ Given certain local monotonicity, continuity, coercivity and growth conditions of the operator $A:[0, T]\times V\to V^*$ and a sufficiently regular operator $I$ we establish global existence of weak solutions. In analogy to the Young regime for SDEs, no probabilistic structure is required in our analysis, which is based on a careful combination of monotone operator theory and the recently developed Besov rough analysis by Friz and Seeger. Due to the abstract nature of our approach, it applies to various examples of monotone and locally monotone operators $A$, such as the $p$-Laplace operator, the porous medium operator, and an operator that arises in the context of shear-thickening fluids; and operators $I$, including additive Young drivers $I_t(u) = Z_t-Z_0$, abstract Young integrals $I_t(u) = \int_0^t \sigma(u_s)d X_s$, and translated integrals $I_t(u) = \int_0^t b(u_s - w_s)ds$ that arise in the context of regularization by noise. In each of the latter cases, we identify corresponding noise regimes (i.e. Young regimes) that assure our abstract result to be applicable. In the case of additive drivers, we identify the Brownian setting as borderline, i.e. noises which enjoy slightly more temporal regularity are amenable to our completely pathwise analysis.