Estimates for viscosity solutions of fully nonlinear equations near smooth boundaries

Niklas L.P. Lundström, Marcus Olofsson, Jesper Singh · 2025

Read Estimates for viscosity solutions of fully nonlinear equations near smooth boundaries by Niklas L.P. Lundström, Marcus Olofsson, Jesper Singh (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-2506-12477.txt.

Introduction

Niklas Lundström and co-authors prove estimates for viscosity solutions of fully nonlinear equations near smooth boundaries. Their paper reduces a partial differential problem to a simpler 1D one.

Decay bounds for the nonlinear case follow from like bounds for answers of one-dimensional ordinary inequalities. Their method lets ellipticity vanish near the edge and allows general, possibly large, lower-order terms. A key result is a boundary Harnack inequality for a broad class of problems near smooth edges. Combined with other bounds, quotients of vanishing answers are Hölder smooth. Phragmén-Lindelöf results in open domains close the paper.

Catalog record

CREATORNiklas L.P. Lundström, Marcus Olofsson, Jesper Singh
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE95 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-12477

Browse the full library · About the Tower of Babel

Full text

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

Estimates for viscosity solutions of fully nonlinear equations near smooth boundaries Niklas L.P. Lundström, Marcus Olofsson, Jesper Singh Abstract We reduce the problem of proving decay estimates for viscosity solutions of fully nonlinear PDEs to proving analogous estimates for solutions of one-dimensional ordinary differential inequalities. Our machinery allow the ellipticity to vanish near the boundary and permits general, possibly unbounded, lower-order terms. A key consequence is the derivation of boundary Harnack inequalities for a broad class of fully nonlinear, nonhomogeneous equations near C 1 , 1 C^{1,1} -boundaries. In combination with C 1 , α C^{1,\alpha} -estimates, we also obtain that quotients of positive vanishing solutions are Hölder continuous near C 1 , 1 C^{1,1} -boundaries.This result applies to a wide family of fully nonlinear uniformly elliptic PDEs; and for p ⁡ ( x ) p(x) -harmonic functions and planar ∞ \infty -harmonic functions near locally flat boundaries. We end by deriving some Phragmén-Lindelöf-type corollaries in unbounded domains. Mathematics Subject Classification: 35J60, 35J70, 35B40, 35B53, 35D40, 35J25, 35J92, 42B37. Keywords: boundary Harnack inequalities; quasi linear; non homogeneous; sub linear; growth estimate. 1 Department of Mathematics and Mathematical Statistics , Umeå University, SE-90187 Umeå, Sweden ; [email protected]; [email protected] 2 Odensalaskolan, Östersund, Sweden; [email protected] 1 Introduction In the first part of the paper, we develop a method that reduces the problem of obtaining decay estimates near the boundary for viscosity sub- and supersolutions of fully nonlinear PDEs to proving analogous estimates for classical sub- and supersolutions of one-dimensional problems in the form of ordinary differential inequalities (ODIs). These ODIs are only slight modifications of the original PDEs in one dimension, thereby illustrating a simple connection between the n -dimensional and the one-dimensional setting for this class of problems. We prove these results in Section 2 , in which Theorem 2.1 holds our lower decay estimates of viscosity supersolutions v of the PDE ( 2.1 ), being nonnegative on a portion of the boundary of a domain satisfying an interior sphere condition. The estimate takes the form v(x)\geq\check{h}(d(x,\partial\Omega)) , in which \check{h} is a classical solution of the corresponding one-dimensional ODI ( 2.7 ). Analogously, Theorem 2.2 gives an upper estimate u(x)\leq\hat{h}(d(x,\partial\Omega)) for viscosity subsolutions u of ( 2.1 ), being nonpositive on a portion of the boundary of a domain satisfying an exterior sphere condition, where \hat{h} solves the ODI ( 2.8 ). In the next step we conclude the following similar characterization of the boundary Harnack inequality near C^{1,1} boundaries: Suppose that u_{1} and u_{2} are viscosity solutions of ( 2.1 ), vanishing continuously on a portion of the boundary. Then, if the boundary Harnack inequality holds for the solutions of the corresponding ODIs, then it holds also for u_{1} and u_{2} . It thus suffices to prove the inequality for some suitable solutions of the one-dimensional ODIs, see Corollary 2.3 . To establish these results we only assume our main operator in ( 2.1 ) to be proper, satisfy ( 2.5 ) and/or ( 2.6 ). We believe that our work extends the existing literature in several notable ways: (i) We allow the ellipticity to vanish near the boundary – an explicit example is provided in Section 2.2.2 . (ii) We consider general unbounded lower-order terms that may depend on the spatial variable, the solution itself, and the norm of its gradient. Overall, we impose minimal restrictions on the PDE. Instead, our methodology centers on what can be proved for the associated ODIs: (iii) Rather than verifying whether a given PDE satisfies a set of technical assumptions, our approach allows one to reduce the problem to analyzing one-dimensional differential equations. Often, this reduces to classical textbook examples; for instance, see Section 2.2.1 , which includes generalizations of the well-known p -Laplace equation. The present approach allows us to explicitly track how estimates depend on the value of the solution at a given distance from the boundary. This is demonstrated by examples in Sections 2.2.3 and 2.2.4 , where we identify settings in which decay estimates and a boundary Harnack inequalities can be derived for non-homogeneous PDEs that clearly fall outside the scope of assumptions made in, for example, [ 34 ] . In the setting of harmonic functions the Boundary Harnack inequality was considered already in the 70’s. Generalizations to elliptic equations, nontangentially accessible domains, Hölder domains and uniform domains appeared in the 80’s. Concerning nonlinear PDEs, similar results were obtained for p -harmonic functions in smooth domains [ 3 ] , and in Lipschitz, and Reifenberg flat domains [ 21 , 24 , 25 , 26 ] . In the latter, the ratio u/v was also proved to be Hölder continuous, and applications of such regularity estimates was applied to free boundary problems and studies of the Martin boundary in [ 24 , 25 , 26 ] . For recent results on boundary Harnack inequalities in Lipschitz domains and beyond, as well as applications to free boundary problems, see e.g. [ 44 , 19 , 5 , 14 , 42 , 4 ] , and for nonhomogeneous equations such as the variable p -Laplace equation, we refer to [ 2 , 8 , 34 ] . For nonlocal elliptic equations in non-divergence form, see e.g. [ 41 ] and references therein. Finally, we note that our approach is inspired by our earlier work [ 33 ] on Phragmén-Lindelöf-type estimates, and that, in contrast to the aforementioned studies, our contribution mainly relies on less restrictive assumptions on the equation. In the second part of the paper (Section 3 ) we first show that the ratio u_{1}/u_{2} is Hölder continuous, where u_{1} and u_{2} are positive solutions of a fully nonlinear PDE, vanishing on a portion of the boundary of a C^{1,1} -domain, see Theorem 3.5 . In particular, by combining our results from the first part of the paper, giving bounds on the ratios u_{i}/d(x,\partial\Omega) , d(x,\partial\Omega)/u_{i} , i=1,2 , and u_{1}/u_{2} near the boundary, with known C^{1,\alpha} -estimates of u_{i} , we prove that these ratios are also Hölder continuous of order \alpha . To apply the most general available global C^{1,\alpha} -estimates in our setting – which, to the best of our knowledge, is given in [ 36 ] – we had to impose uniformly ellipticity and at most quadratic growth in the gradient, see assumption ( 3.2 ). Secondly, we consider equations for which global C^{1,\alpha} -estimates are unknown, and use reflection arguments together with local C^{1,\alpha} -estimates to establish the Hölder continuity of the ratios. In particular, we prove results for positive p(x) -harmonic functions (Theorem 3.9 ), and for planar \infty -harmonic functions (Theorem 3.14 ), vanishing on a hyperplane. A consequence of the latter is a gradient estimate on the form |Du|\approx u(x)/d(x,\partial\Omega) , see Corollary 3.15 . We end the paper with some further corollaries of our results, as well as corollaries of already known versions of these estimates for p -harmonic functions, p\in(1,\infty) in Section 3.5 . In particular, we derive some Phragmén-Lindelöf theorems as well as uniqueness (modulo normalization) of infinity- and p -harmonic functions in some unbounded domains. Notation. Points in Euclidean n -space \mathbb{R}^{n} are denoted by x=(x_{1},...,x_{n}) or x=(x^{\prime},x_{n}) , where x^{\prime}=(x_{1},...,x_{n-1})\in\mathbb{R}^{n-1} , and \mathbb{R}^{n}_{+}:=\{x\in\mathbb{R}^{n}:x_{n}>0\} . We let

Read the complete text