A regularity theory for evolution equations with space-time anisotropic non-local operators in mixed-norm Sobolev spaces

Jae-Hwan Choi, Jaehoon Kang, Daehan Park, Jinsol Seo · 2025

Read A regularity theory for evolution equations with space-time anisotropic non-local operators in mixed-norm Sobolev spaces by Jae-Hwan Choi, Jaehoon Kang, Daehan Park, Jinsol Seo (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-2505-00984.txt.

Introduction

Jae-Hwan Choi, Jaehoon Kang, Daehan Park, and Jinsol Seo wrote this 2025 mathematics paper on how smooth the solutions of certain evolution equations are. The equations mix time-fractional change with spatial operators that act at a distance rather than point by point. Each direction in space can carry its own fractional power.

A model case adds a fractional Laplacian in x to one in y, each with its own exponent. The authors prove that solutions exist, are unique, and obey precise size estimates in mixed-norm Sobolev spaces of variable order. Standard Fourier methods fail here, since the operators' symbols have singularities. The team turns to a probability approach instead, reading the operators through Brownian motions that are slowed in each direction. The work extends regularity theory to a wider class of nonlocal equations.

Catalog record

CREATORJae-Hwan Choi, Jaehoon Kang, Daehan Park, Jinsol Seo
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE158 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-00984

Browse the full library · About the Tower of Babel

Full text

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

A regularity theory for evolution equations with space-time anisotropic non-local operators in mixed-norm Sobolev spaces Jae-Hwan Choi, Jaehoon Kang, Daehan Park, Jinsol Seo Abstract . In this article, we study the regularity of solutions to inhomogeneous time-fractional evolution equations involving anisotropic non-local operators in mixed-norm Sobolev spaces of variable order, with non-trivial initial conditions. The primary focus is on space-time non-local equations where the spatial operator is the infinitesimal generator of a vector of independent subordinate Brownian motions, making it the sum of subdimensional non-local operators. A representative example of such an operator is ( Δ x ) β 1 / 2 + ( Δ y ) β 2 / 2 (\Delta_{x})^{\beta_{1}/2}+(\Delta_{y})^{\beta_{2}/2} . We establish existence, uniqueness, and precise estimates for solutions in corresponding Sobolev spaces. Due to singularities arising in the Fourier transforms of our operators, traditional methods involving Fourier analysis are not directly applicable. Instead, we employ a probabilistic approach to derive solution estimates. Additionally, we identify the optimal initial data space using generalized real interpolation theory. 2020 Mathematics Subject Classification Primary: 26A33, 35S10, 47G20, Secondary: 30H25, 46B70, 46E35 1. Introduction 1.1. Motivations and goals Anisotropic non-local operators such as \Delta^{\beta_{1}/2}_{x}+\Delta^{\beta_{2}/2}_{y} are important in describing phenomena that exhibit distinct behaviors in different coordinate directions. Applications of anisotropic non-local operators appear frequently in various scientific fields; see, for instance, [ 7 , 8 , 22 ] . Additionally, there has been significant theoretical development and practical applications of space-time non-local operators. Examples include the derivation of space-time fractional Fokker–Planck–Kolmogorov equations within fractional kinetics frameworks [ 42 , 43 ] and the study of space-time non-local diffusion-advection equations [ 21 , 36 ] . Motivated by these applications and developments, we study the following fractional evolution equation involving anisotropic spatial non-local operators: \displaystyle\frac{1}{\Gamma(1-\alpha)}\int_{0}^{t}(t-s)^{-\alpha}\left(u(s,\vec{x})-u_{0}(\vec{x})\right)\mathrm{d}s=\int_{0}^{t}\left(\sum_{i=1}^{\ell}\phi_{i}(\Delta_{x_{i}})u(s,\vec{x})+f(s,\vec{x})\right)\mathrm{d}s,\quad(t,\vec{x})\in(0,T)\times\mathbb{R}^{d}, where \alpha\in(0,1) and the spatial dimension d is composed of \ell sub-dimensions d_{1},\dots,d_{\ell} , so that \mathbb{R}^{d}=\mathbb{R}^{d_{1}}\times\dots\times\mathbb{R}^{d_{\ell}} . Each point \vec{x}\in\mathbb{R}^{d} can thus be represented as (1.1) \vec{x}=(x_{1},\dots,x_{\ell}),\quad x_{i}=(x^{1}_{i},\dots,x^{d_{i}}_{i})\in\mathbb{R}^{d_{i}},\quad i=1,\dots,\ell. The spatial non-local operators \phi_{i}(\Delta_{x_{i}}) are defined by \displaystyle\phi_{i}(\Delta_{x_{i}})g(\vec{x}):=b_{i}\Delta_{x_{i}}g(\vec{x})+\int_{\mathbb{R}^{d_{i}}}\left(g(x_{1},\dots,x_{i-1},x_{i}+y_{i},x_{i+1},\dots,x_{\ell})-g(\vec{x})-1_{|y_{i}|\leq 1}y_{i}\cdot\nabla_{x_{i}}g(\vec{x})\right)J_{\phi_{i}}(y_{i})\,\mathrm{d}y_{i}, where b_{i}\geq 0 and \Delta_{x_{i}} denotes the standard d_{i} -dimensional Laplacian. Differentiating in time, the equation can equivalently be expressed in terms of the Caputo fractional derivative \partial^{\alpha}_{t} as follows: (1.2) \partial^{\alpha}_{t}u(t,\vec{x})=\sum_{i=1}^{\ell}\phi_{i}(\Delta_{x_{i}})u(t,\vec{x})+f(t,\vec{x}),\quad(t,\vec{x})\in(0,T)\times\mathbb{R}^{d},\quad u(0,\vec{x})=u_{0}(\vec{x}). The objectives of this article are three-fold: • Identify the optimal initial data space X (trace and extension theorem for ( 1.2 )). • Prove existence and uniqueness of solutions to ( 1.2 ) in L_{q}((0,T);L_{p}) . • Obtain maximal regularity estimates for solutions to ( 1.2 ), specifically (1.3) \displaystyle\|\partial^{\alpha}_{t}u\|_{L_{q}((0,T);L_{p})}+\left\|\sum_{i=1}^{\ell}\phi_{i}(\Delta_{x_{i}})u\right\|_{L{q}((0,T);L_{p})}\leq C\left(\|u_{0}\|_{X}+\|f\|_{L_{q}((0,T);L_{p})}\right),\quad 1<p,q<\infty. 1.2. Historical Results In this subsection, we summarize some known results from the literature concerning the fractional evolution equations, and PDEs involving anisotropic non-local operators. For a more comprehensive historical overview beyond the scope of this article, we refer the reader to the introduction of [ 4 ] . Evolution equations with time fractional derivative. The Sobolev regularity theory for fractional evolution equations initially focused on equations involving second-order differential operators. For instance, I. Kim, K.-H. Kim, and S. Lim [ 28 ] studied fractional diffusion-wave-type equations ( i.e. , \alpha\in(0,2) ) with second-order differential operators having continuous coefficients in mixed-norm Lebesgue spaces. B.-S. Han, K.-H. Kim, and D. Park [ 23 ] investigated the weighted counterpart of [ 28 ] , which was subsequently extended to higher regularity by D. Park [ 35 ] . A particularly challenging research direction has involved relaxing the continuity assumptions on coefficients, significantly advanced by H. Dong and D. Kim. Detailed unweighted results can be found in [ 10 , 11 , 13 ] , while their weighted analogues are presented in [ 12 , 15 ] . Additionally, H. Dong and Y. Liu [ 14 ] provided weighted results specifically for \alpha\in(1,2) . The regularity theory for fractional evolution equations involving non-local operators is a natural subsequent research direction. K.-H. Kim, D. Park, and J. Ryu [ 29 ] explored evolution equations with time fractional derivatives and variable-order spatial non-local operators in mixed-norm Lebesgue spaces. The assumptions regarding spatial non-local operators were further relaxed by J. Kang and D. Park [ 25 ] , who studied equations associated with infinitesimal generators of general Lévy processes. Additionally, H. Dong and Y. Liu [ 16 ] investigated fractional evolution equations involving space-dependent non-local operators. We also refer readers to [ 2 , 37 , 40 , 41 ] for alternative approaches to abstract Volterra equations. One of the important research directions to study fractional evolution equations is the trace theorem. D. Kim and K. Woo [ 27 ] provided trace theorems for fractional evolution equations involving second-order divergence and non-divergence operators. Additionally, J.-H. Choi, J. B. Lee, J. Seo, and K. Woo [ 6 ] established trace theorems for generalized time fractional equations within a generalized real interpolation framework. More references on this topic can be found in the introductions of [ 6 , 27 ] . PDEs with anisotropic non-local operators. We consider anisotropic non-local operators of the following form: (1.4) \mathcal{L}_{\beta_{1},\beta_{2}}:=\Delta^{\beta_{1}/2}_{x^{1}}+\Delta^{\beta_{2}/2}_{x^{2}},\quad(x^{1},x^{2})\in\mathbb{R}^{2},\quad\beta_{1},\beta_{2}\in(0,2). R. Mikulevičius and C. Phonsom [ 33 , 34 ] investigated the Sobolev regularity theory for parabolic PDEs involving scalable non-local operators. J.-H. Choi and I. Kim [ 5 ] extended these results specifically to the case of homogeneous parabolic PDEs. A. de Pablo, F. Quirós, and A. Rodríguez studied the well-posedness and regularity of very weak solutions to anisotropic non-local parabolic PDEs defined by operators of the form \mathcal{L}u(x)=\frac{1}{2}\int_{\mathbb{R}^{d}}(u(x+y)+u(x-y)-2u(x))\nu(\mathrm{d}y), where \nu is the Lévy measure given by \nu(A):=\int_{\mathbb{S}^{d-1}}\int_{0}^{\infty}1_{A}(r\theta)\frac{\mathrm{d}r}{r^{1+\alpha}}\mu(\mathrm{d}\theta), and \mu is a nondegenerate finite surface measure defined on \mathbb{S}^{d-1} . In particular, if d=2 and \mu(\mathrm{d}\theta):=\epsilon_{(1,0)}(\mathrm{d}\theta)+\epsilon_{(0,1)}(\mathrm{d}\theta), where \epsilon_{(1,0)} and \epsilon_{(0,1)} denote Dirac measures centered at (1,0) and (0,1) respectively, then we have \mathcal{L}=\mathcal{L}_{\alpha,\alpha} . Recently, H. Dong and J. Ryu [ 17 ] developed the weighted Sobolev regularity theory for elliptic and parabolic PDEs in C^{1,\tau} -domains associated with the operator \mathcal{L} . However, these earlier works exclusively considered operators ( 1.4 ) with \beta_{1}=\beta_{2} . J.-H. Choi, J. Kang, and D. Park [ 4 ] subsequently developed the Sobolev regularity theory for elliptic and parabolic PDEs with \mathcal{L}_{\beta_{1},\beta_{2}} for arbitrary \beta_{1},\beta_{2}\in(0,2) . Although not covered in this article, an interesting anisotropic nonlocal operator is given by Lu(x)=\int_{\mathbb{R}^{d}}\frac{u(x+y)-u(x)-\nabla u(x)\cdot y\mathbf{1}_{|y|\leq 1}}{|y_{1}|^{d+\beta_{1}}+\cdots+|y_{d}|^{d+\beta_{d}}}\mathrm{d}y. L.A. Caffarelli, R. Leitão, and J.M. Urbano developed the regularity theory for fully nonlinear integro-differential equations involving L . A version of Caffarelli-Silvestre’s extension problem [ 3 ] associated with L was explored by R. Leitão [ 31 ] . E.B. dos Santos, R. Leitão [ 18 ] studied the Hölder regularity theory for equations involving L -like operators. R. Leitão [ 32 ] also established Sobolev regularity theory for equations involving L , following the spirit of [ 9 ] . 1.3. Description of Approaches We now describe the approach employed in this article. For parabolic PDEs involving anisotropic non-local operators of the form \partial_{t}u=\sum_{i=1}^{\ell}\phi_{i}(\Delta_{x_{i}})u+f,\quad u(0)=0, the solution u admits the following representation: (1.5) \displaystyle u(t,\vec{x})=\int_{0}^{t}\int_{\mathbb{R}^{d}}p(t-s,\vec{x}-\vec{y})f(s,\vec{y})\mathrm{d}\vec{y}\mathrm{d}s=\int_{0}^{t}\mathbb{E}[f(s,\vec{x}-\vec{X}_{t-s})]\mathrm{d}s, where p(t,\vec{x}) is the transition density of the independent array of subordinate Brownian motion \vec{X}_{t} . One difficulty arises in proving the maximal regularity estimates ( 1.3 ). A natural approach to obtain ( 1.3 ) is the Calderón-Zygmund approach based on the Fourier transform. Specifically, the Fourier transform of our spatial operator

Read the complete text