The Boundary Harnack Principle and the 3 ​ G 3G Principle in Fractal-Type Spaces

Anthony Graves-McCleary, Laurent Saloff-Coste · 2024

Read The Boundary Harnack Principle and the 3 ​ G 3G Principle in Fractal-Type Spaces by Anthony Graves-McCleary, Laurent Saloff-Coste (2024) 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-2412-18671.txt.

Introduction

Anthony Graves-McCleary and Laurent Saloff-Coste prove two linked rules for heat flow on fractal-like spaces in a 2024 paper. Fractals, such as sponge-like carpets, lack the smooth shape that standard math tools rest on. The rules at stake compare how heat spreads near a wall and how two spread paths trade off.

Their proof covers a wide class of measured spaces, not just neat shapes in flat space. It holds for carpets in three or more dimensions and for odd spaces with no clean size scale. The rules then feed into work on wave-type operators on the same spaces. Fresh cases of the trade-off rule are the main gain, the pair notes.

Catalog record

CREATORAnthony Graves-McCleary, Laurent Saloff-Coste
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE65 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-18671

Browse the full library · About the Tower of Babel

Full text

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

The Boundary Harnack Principle and the 3 ​ G 3G Principle in Fractal-Type Spaces Anthony Graves-McCleary, Laurent Saloff-Coste Abstract We prove a generalized version of the 3 ​ G 3G Principle for Green’s functions on bounded inner uniform domains in a wide class of Dirichlet spaces. In particular, our results apply to higher-dimensional fractals such as Sierpinski carpets in 𝐑 n \mathbf{R}^{n} , n ≥ 3 n\geq 3 , as well as generalized fractal-type spaces that do not have a well-defined Hausdorff dimension or walk dimension. This yields new instances of the 3 ​ G 3G Principle for these spaces. We also discuss applications to Schrödinger operators. 1 Introduction We are interested in the behavior of a Green’s function G(x,y) in a domain at points near the boundary of the domain, as well as ratios of the Green’s function across the entire domain. As a warm up and to introduce many of the concepts involved, we start with a section in Euclidean space. Then we make the leap to the abstract setting, working in domains in metric measure spaces with no a priori notion of harmonicity. It is here that we prove our main theorem 3.6 . Then in the following sections we introduce concepts that will yield many examples: inner uniform domains and Dirichlet spaces. The paper by Murugan [ 18 ] gives a rich collection of examples for which the Generalized 3G Principle, a concept linked to the main theorem and defined in 2.2 below, was not previously known to hold. 1.1 Notation Let a\vee b=\min(a,b) and a\wedge b=\max(a,b) . We let a\preceq b denote the existence of a constant C>0 such that a\leq C\cdot b . We let a\asymp b denote that both a\preceq b and a\succeq b hold. Exact constants will sometimes change line to line. On occasion we will write a:b to denote a/b . Given a metric space X , we denote by C_{c}(X) the set of all continuous, compactly supported real-valued functions on X . Given M>0 and a ball B=B(x,r) of center x and radius r>0 , MB will denote the ball B(x,Mr) of center x and radius Mr . 2 Classical Preliminaries and Motivation For this section, work in the open unit ball B in \mathbf{R}^{n} with n\geq 3 . Let \alpha(n) denote the volume of the unit ball in \mathbf{R}^{n} and let c_{n}=\frac{1}{n(n-2)\alpha(n)} . For x,y\in B , x\neq y and y\neq 0 let G_{B}(x,y)=c_{n}\left(\left\|x-y\right\|^{2-n}-\left(\left\|y\right\|\left\|x-\frac{y}{\left\|y\right\|^{2}}\right\|\right)^{2-n}\right). Then if x\neq 0 let G_{B}(x,0)=c_{n}\left(\left\|x\right\|^{2-n}-1\right). This is the classical Green’s function for B . Note that for any \xi\in\partial B and any x\in B we have \lim_{y\rightarrow\xi}G_{B}(x,y)=0 . In fact, for x\in B fixed and y\in B sufficiently close to \partial B , G_{B}(x,y) resembles the distance to the boundary d(y,\partial B) , up to constants. In other domains such as cones, the Green’s function G(x,y) may have different boundary decay, but in classical settings it can controlled by the Uniform Boundary Harnack Principle. Let D be a domain in \mathbf{R}^{n} . With an eye towards generalizing, write d(x,y)=\left\|x-y\right\| . Definition 2.1 . Let D be a domain let G be a function on D\times D . We say that D satisfies the Uniform Boundary Harnack Principle for G if there exist constants r_{0},M,C>0 such that for all \xi\in\partial D , 0<r\leq r_{0} , and for all x_{1},x_{2},y_{1},y_{2}\in D with Md(x_{j},\xi)<r\leq d(y_{j},\xi) , \frac{G(x_{1},y_{1})}{G(x_{2},y_{1})}\leq C\frac{G(x_{1},y_{2})}{G(x_{2},y_{2})}. (2.1) Note that in the inequality 2.1 , y_{1} and y_{2} can be switched. The definition of the Uniform Boundary Harnack Principle for more abstract domains is exactly the same. This is all stated in terms of G , but when G is truly the Green’s function of the domain D , this implies a similar bound for any two positive harmonic functions u,v>0 defined around \xi that vanish on \partial D . See Lierl [ 15 ] for further details. The Euclidean open unit ball B satisfies the Uniform Boundary Harnack Principle for its Green’s function G_{B} . The Uniform Boundary Harnack Principle has a rich history and has been generalized with regard to both the domain and the kind of Green’s function it concerns. It originated in papers by Ancona [ 2 ] , Dahlberg [ 11 ] and Wu [ 20 ] in the late 1970s. See Armitage and Gardiner [ 3 ] section 8.7 for a textbook presentation in a Lipschitz domain in Euclidean space. See Barlow and Murugan [ 7 ] for a more comprehensive review as well as a state-of-the-art proof on inner uniform domains in Dirichlet spaces satisfying only an elliptic Harnack inequality. Our other central concept, the 3G Principle, arises from the desire to control ratios of the Green’s function on the entire domain. This is particularly important in the study of conditional Brownian motion and Schrödinger operators, which we will discuss in section 6. In the ball, there is a constant such that for any x,y,z\in B , \frac{G_{B}(x,z)G_{B}(y,z)}{G_{B}(x,y)}\leq C\left(\left\|x-z\right\|^{2-n}+\left\|y-z\right\|^{2-n}\right). (2.2) Notably \left\|x-z\right\|^{2-n} is, up to a constant, the Green’s function of all of \mathbf{R}^{n} . Therefore this “ 3G ” ratio of Green’s functions in the ball can be controlled by Green’s functions from a larger domain. In generalizing this, we can avoid reference to any such larger domains by use of an additional benchmark Green’s function. We are now ready to state our version of the 3G Principle. Definition 2.2 . Let D be a domain and let G be a function on D\times D . Fix o\in D and let g(x)=G(x,o)\wedge 1 be the benchmark Green’s function. We say that D satisfies the Generalized 3G Principle with respect to G if there exists a constant C>0 such that for all x,y,z\in D we have \frac{G(x,z)G(y,z)}{G(x,y)}\leq C\left(\frac{g(z)}{g(x)}G(x,z)+\frac{g(z)}{g(y)}G(z,y)\right). (2.3) In practice this will be independent of choice of o\in D , up to constants. The presence of g is necessary in many cases due to boundary vanishing of the Green’s function. In applications, the expression \frac{g(z)}{g(x)}G(x,z) will be bounded above by the Green’s function of the entire space, therefore relating this back to 2.2 . See Lemma 4.2 below for more details. The 2.2 form of the 3G Principle appears for bounded Lipschitz domains in Euclidean space in Cranson, Fabes, and Zhao [ 10 ] with a view towards Schrödinger operators, which we will discuss in a later section. The Generalized 3G Principle appears in Riahi [ 19 ] . The proofs of the 3G Principle almost always make use of the Uniform Boundary Harnack Principle, which raises the question of whether they are equivalent. Hansen [ 13 ] demonstrated the equivalence between the Uniform Boundary Harnack Principle and a form of the 3G Principle (see Definition 2.4 below) in an abstract setting that, in particular, applies to bounded uniform domains in \mathbf{R}^{n} , n\geq 3 . We now state an alternate form of the Generalized 3G Principle which is routine to verify as equivalent to the original. Proposition 2.3 . Let D be a domain and let G be a function on D\times D . Fix o\in D and let g(x)=G(x,o)\wedge 1 be the benchmark Green’s function. Also let \widetilde{G}(x,y)=G(x,y)/(g(x)g(y)) . Then D satisfies the Generalized 3G Principle for G with constant C>0 if and only if for

Read the complete text