Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Hypoellipticity and Higher Order Gaussian Bounds

Brian Street

TL;DR

The paper develops an abstract framework to derive higher-order Gaussian bounds for heat semigroups on doubling metric measure spaces under hypoellipticity-type hypotheses. A central achievement is the existence of a kernel $K_t(x,y)$ representing the semigroup with sharp bounds for $|\partial_t^{j}X\overline{Y}K_t(x,y)|$ that feature non-Euclidean Carnot–Carathéodory geometry and a nontrivial exponential decay in the scale $t$ via $\exp\left(-c\left(\frac{\rho(x,y)^{2\kappa}}{t}\right)^{\frac{1}{2\kappa-1}}\right)$. The results are localizable and symmetric in the generator and its adjoint, enabling a robust parametrix construction for general maximally subelliptic PDEs without reliance on Lie-group lifting. The paper applies the main theorem to subelliptic PDEs defined by Hörmander vector fields, establishing higher-order Gaussian bounds and two-sided parametrices, and discusses scaling, boundary value problems, and a range of concrete examples. Overall, the work provides a versatile, scale-invariant toolkit for sharp heat-kernel estimates in subelliptic settings with broad potential applications in analysis and PDEs.

Abstract

Let $(\mathfrak{M},ρ,μ)$ be a metric measure space satisfying a doubling condition, $p_0\in (1,\infty)$, and $T(t):L^{p_0}(\mathfrak{M},μ)\rightarrow L^{p_0}(\mathfrak{M},μ)$, $t\geq 0$, a strongly continuous semi-group. We provide sufficient conditions under which $T(t)$ is given by integration against an integral kernel satisfying higher-order Gaussian bounds of the form \[ \left| K_t(x,y) \right| \leq C \exp\left( -c \left( \frac{ρ(x,y)^{2κ}}{t} \right)^{\frac{1}{2κ-1}} \right) μ\left( B_ρ\left(x,ρ(x,y)+t^{1/2κ}\right) \right)^{-1}, \] where $B_ρ$ denotes the metric ball. We also provide conditions for similar bounds on ``derivatives'' of $K_t(x,y)$ and our results are localizable. If $A$ is the generator of $T(t)$ the main hypothesis is that $\partial_t -A$ and $\partial_t-A^{*}$ satisfy a hypoelliptic estimate at every scale, uniformly in the scale. We present applications to subelliptic PDEs.

Hypoellipticity and Higher Order Gaussian Bounds

TL;DR

The paper develops an abstract framework to derive higher-order Gaussian bounds for heat semigroups on doubling metric measure spaces under hypoellipticity-type hypotheses. A central achievement is the existence of a kernel representing the semigroup with sharp bounds for that feature non-Euclidean Carnot–Carathéodory geometry and a nontrivial exponential decay in the scale via . The results are localizable and symmetric in the generator and its adjoint, enabling a robust parametrix construction for general maximally subelliptic PDEs without reliance on Lie-group lifting. The paper applies the main theorem to subelliptic PDEs defined by Hörmander vector fields, establishing higher-order Gaussian bounds and two-sided parametrices, and discusses scaling, boundary value problems, and a range of concrete examples. Overall, the work provides a versatile, scale-invariant toolkit for sharp heat-kernel estimates in subelliptic settings with broad potential applications in analysis and PDEs.

Abstract

Let be a metric measure space satisfying a doubling condition, , and , , a strongly continuous semi-group. We provide sufficient conditions under which is given by integration against an integral kernel satisfying higher-order Gaussian bounds of the form where denotes the metric ball. We also provide conditions for similar bounds on ``derivatives'' of and our results are localizable. If is the generator of the main hypothesis is that and satisfy a hypoelliptic estimate at every scale, uniformly in the scale. We present applications to subelliptic PDEs.

Paper Structure

This paper contains 22 sections, 53 theorems, 199 equations.

Table of Contents

  1. Introduction
  2. The Setting and Assumptions
  3. Assumptions on the measure
  4. Hypoellipticity Assumptions
  5. The Main Result
  6. Application: Subelliptic PDEs
  7. The main theorem for subelliptic PDEs
  8. Maximal Subellipticity
  9. Parametricies
  10. Examples
  11. Proof of Theorem \ref{['Thm::SubMainThm::MainThm']}
  12. A priori subelliptic estimates at the unit scale
  13. Qualitative consequences of subellipticity
  14. Scaling
  15. Completion of the proof of Theorem \ref{['Thm::SubMainThm::MainThm']}
  16. ...and 7 more sections

Key Result

Theorem 3.1

There are open neighborhoods $U_x,U_y\subseteq (0,\infty)\times \mathfrak{M}$ of $(0,\infty)\times \mathfrak{N}_x$ and $(0,\infty)\times \mathfrak{N}_y$, respectively, such that the following holds. For $t>0$, let $U_{x,t}:=\left\{ v\in \mathfrak{M} : (t,v)\in U_x \right\}$ and similarly for $U_{y,t initially defined as operators $\mathscr{D}(A^{\infty})\rightarrow L^{p_0}\IfNoValueF{U_{x,t}}{(U_{

Theorems & Definitions (149)

  • Remark 1.1
  • Remark 2.1
  • Remark 2.2
  • Remark 2.3
  • Remark 2.4
  • Remark 2.5
  • Definition 2.6
  • Remark 2.7
  • Remark 2.8
  • Remark 2.9
  • ...and 139 more