Functional calculus on weighted Sobolev spaces for the Laplacian on rough domains

Nick Lindemulder, Emiel Lorist, Floris B. Roodenburg, Mark C. Veraar · 2025

Read Functional calculus on weighted Sobolev spaces for the Laplacian on rough domains by Nick Lindemulder, Emiel Lorist, Floris B. Roodenburg, Mark C. Veraar (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-2507-13478.txt.

Introduction

Nick Lindemulder and co-authors study the Laplace operator on rough domains in a 2025 math paper. They show it admits a bounded functional calculus on weighted Sobolev spaces whose weights are powers of the distance to the boundary. The domains need mild smoothness, with fixed-value or zero-flux edge terms. A trade-off drives the result: rougher edges can be offset by a larger weight power. The payoff is maximal regularity for the linked heat equation, extending well-posedness where classic methods fail.

Catalog record

CREATORNick Lindemulder, Emiel Lorist, Floris B. Roodenburg, Mark C. Veraar
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE193 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-13478

Browse the full library · About the Tower of Babel

Full text

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

Functional calculus on weighted Sobolev spaces for the Laplacian on rough domains Thanks: The first author is supported by the grant OCENW.KLEIN.358 of the Dutch Research Council (NWO). The second author is partially supported by the Academy of Finland through grant no. 336323. The third and fourth author are supported by the VICI grant VI.C.212.027 of the NWO Nick Lindemulder, Emiel Lorist, Floris B. Roodenburg, Mark C. Veraar Abstract . We study the Laplace operator on domains subject to Dirichlet or Neumann boundary conditions. We show that these operators admit a bounded H ∞ H^{\infty} -functional calculus on weighted Sobolev spaces, where the weights are powers of the distance to the boundary. Our analysis applies to bounded C 1 , λ C^{1,\lambda} -domains with λ ∈ [ 0 , 1 ] \lambda\in[0,1] , revealing a crucial trade-off: lower domain regularity can be compensated by enlarging the weight exponent. As a primary consequence, we establish maximal regularity for the corresponding heat equation. This extends the well-posedness theory for parabolic equations to domains with minimal smoothness, where classical methods are inapplicable. 2020 Mathematics Subject Classification Primary: 47A60; Secondary: 35K20, 46E35 1. Introduction This paper contributes to the extensive study of the Laplace operator on domains with minimal boundary regularity (often referred to as rough domains), see, e.g., [ 36 , 37 , 38 , 83 , 86 ] and the monographs [ 30 , 71 ] and references therein. In particular, we are interested in the H^{\infty} -functional calculus for the Laplacian on inhomogeneous weighted Sobolev spaces. The H^{\infty} -functional calculus provides a powerful framework for establishing well-posedness and regularity results for (possibly nonlinear) partial and stochastic partial differential equations ((S)PDEs). Therefore, the H^{\infty} -calculus for sectorial operators is widely studied, see for instance [ 16 , 34 , 35 , 64 ] and the references therein. Applications to PDEs and SPDEs can, e.g., be found in [ 14 , 17 , 39 , 65 , 78 , 85 ] and [ 1 , 2 , 75 , 76 ] , respectively. Given a bounded C^{2} -domain \mathcal{O}\subseteq{\mathbb{R}}^{d} , it is well known that the Laplacian with Dirichlet boundary conditions on L^{p}(\mathcal{O}) with p\in(1,\infty) and domain W^{2,p}(\mathcal{O})\cap W^{1,p}_{0}(\mathcal{O}) generates an analytic C_{0} -semigroup, has the maximal regularity property and admits a bounded H^{\infty} -functional calculus. However, if the regularity of \mathcal{O} is too low (say Lipschitz or C^{1} ), these properties fail and explicit counterexamples can be constructed, see [ 9 , 71 ] . In such counterexamples, the derivatives of the solutions to the resolvent equation \displaystyle\lambda u-\Delta u \displaystyle=f, \displaystyle u|_{\partial\mathcal{O}} \displaystyle=0, can drastically blow up near the boundary \partial\mathcal{O} . As a consequence, the canonical domain of the Dirichlet Laplacian on L^{p}(\mathcal{O}) is no longer a closed subspace of W^{2,p}(\mathcal{O}) . Moreover, if one is interested in higher-order Sobolev regularity of the solution u , higher-order regularity of \mathcal{O} is needed (see [ 25 , 58 ] ), and additional boundary conditions for the data f (compatibility conditions) need to be imposed (see [ 15 ] ). These additional boundary conditions for the data occur, in particular, in the study of mixed-order systems (see [ 18 ] ). To set up a satisfying well-posedness and regularity theory for PDE without such additional regularity or compatibility conditions, one can use a weighted function space for the solution u . In particular, one can consider spatial weights of the form w_{\gamma}^{\partial\mathcal{O}}(x):=\operatorname{dist}(x,\partial\mathcal{O})^{\gamma} for some suitable \gamma\in\mathbb{R} , which compensate the blow-up of the derivatives of the solution near \partial\mathcal{O} and relax compatibility conditions. Partial differential equations on weighted spaces have already been studied extensively, see for instance [ 19 , 20 , 21 , 47 , 52 , 56 , 57 , 73 ] for deterministic equations and [ 43 , 44 , 45 , 54 , 59 ] for stochastic equations. As stated, we are interested in the H^{\infty} -functional calculus for the Laplacian on inhomogeneous weighted Sobolev spaces of order k\in{\mathbb{N}}_{0} . This was studied in [ 67 , 69 ] for the Dirichlet and Neumann Laplacian on the half-space {\mathbb{R}}^{d}_{+} . In the present paper, we extend the results to bounded domains \mathcal{O} with minimal smoothness, while ensuring that the canonical domain of the Laplacian is a closed subspace of a weighted Sobolev space of order k+2 . Our main result for the Dirichlet Laplacian is as follows, see Theorems 6.2 and 6.4 . For the definition of the involved spaces, the reader is referred to Section 3 . Theorem 1.1 ( H^{\infty} -calculus for the Dirichlet Laplacian) . Let p\in(1,\infty) , k\in\mathbb{N}_{0} , \lambda\in[0,1] and \gamma\in(-1,2p-1)\setminus\{p-1\} . Furthermore, suppose that \lambda>1-\tfrac{\gamma+1}{p}\qquad\text{or, equivalently}\qquad\gamma>(1-\lambda)p-1 and \mathcal{O} is a bounded C^{1,\lambda} -domain. Then for all \mu\geq 0 the operator \mu-\Delta_{\operatorname{Dir}}\quad\text{ on }\quad W^{k,p}(\mathcal{O},w_{\gamma+kp}^{\partial\mathcal{O}})\quad\text{ with }\quad D(\Delta_{\operatorname{Dir}})=W^{k+2,p}_{\rm Dir}(\mathcal{O},w_{\gamma+kp}^{\partial\mathcal{O}}) has a bounded H^{\infty} -calculus of angle zero. Theorem 1.1 generalises the result in [ 69 , Theorem 6.1] , which is restricted to the case k=0 and to bounded C^{2} -domains. Theorem 1.1 allows for bounded C^{1} -domains if \gamma\in(p-1,2p-1) , while for \gamma\in(-1,p-1) we obtain that the smoothness of the domain may depend on the weight: if the power of the weight is larger, then a rougher domain is allowed. The smoothness parameter \lambda is almost optimal. Indeed, solving the Dirichlet problem in the scale of weighted Sobolev spaces with a gain of two derivatives for the solution requires the boundary of the domain to have W^{2-(\gamma+1)/p,p} -smoothness, see [ 71 , Theorem 15.6.1 applied to \ell=2-(\gamma+1)/p ] and [ 71 , Section 14.6.1] for an explicit counterexample with C^{1} -domains. Furthermore, for \gamma=p-1 the domain characterisation in Theorem 1.1 in terms of spaces with vanishing traces fails, see [ 67 , Remark 4.3] , and for this reason we omit this case. Concerning the Neumann Laplacian on bounded domains, we prove the following result, see Theorems 6.3 and 6.5 . Theorem 1.2 ( H^{\infty} -calculus for the Neumann Laplacian) . Let p\in(1,\infty) and \lambda\in(0,1] . Furthermore, suppose that either (1) k\in\mathbb{N}_{0} , \gamma\in(p-1,2p-1) , \lambda>2-\frac{\gamma+1}{p} and \mathcal{O} is a bounded C^{1,\lambda} -domain, or, (2) k\in\mathbb{N}_{1} , \gamma\in(-1,p-1) , \lambda>1-\frac{\gamma+1}{p} and \mathcal{O} is a bounded C^{2,\lambda} -domain. Then for all \mu>0 the operator \mu-\Delta_{\operatorname{Neu}}\quad\text{ on }\quad W^{k,p}(\mathcal{O},w_{\gamma+(k-1)p}^{\partial\mathcal{O}})\quad\text{ with }\quad D(\Delta_{\operatorname{Neu}})=W^{k+2,p}_{\rm Neu}(\mathcal{O},w_{\gamma+(k-1)p}^{\partial\mathcal{O}}) has a bounded H^{\infty} -calculus of angle zero. Moreover, using function spaces modulo constants gives the result for all \mu\geq 0 . Note that, compared to Theorem 1.1 , the Sobolev spaces in Theorem 1.2 have a smaller weight exponent, which is consistent with [ 67 , Theorem 1.2] . Figure 1 visualises the parameters of the spaces in Theorem 1.1 and 1.2 where we obtain a bounded H^{\infty} -calculus. Similar to the case of Dirichlet boundary conditions, we expect that the regularity of the domain in Theorem 1.2 is almost optimal as well, see [ 71 , Section 15.6] for some related results in this direction. smoothness k weight exponent \alpha \mu-\Delta_{\operatorname{Dir}} on bounded C^{1} -domain \mu-\Delta_{\operatorname{Dir}} and \mu-\Delta_{\operatorname{Neu}} on bounded C^{1,\lambda} -domain \mu-\Delta_{\operatorname{Neu}} on bounded C^{2,\lambda} -domain 1 2 3 4 -1 p-1 2p-1 3p-1 4p-1 5p-1 Figure 1. The spaces W^{k,p}(\mathcal{O},w^{\partial\mathcal{O}}_{\alpha}) where \mu-\Delta_{\operatorname{Dir}} and \mu-\Delta_{\operatorname{Neu}} as in Theorems 1.1 and 1.2 (with \alpha=\gamma+kp and \alpha=\gamma+(k-1)p , respectively) admit

Read the complete text