The method of stochastic characteristics for linear second-order hypoelliptic equations

TL;DR

This paper develops a probabilistic framework for analyzing hypoelliptic SDEs and boundary value problems, establishing smoothness, boundary behavior, and inequalities for solutions using stochastic characteristics.

Contribution

It introduces new probabilistic conditions ensuring solution smoothness and boundary convergence for hypoelliptic equations, extending classical results with refined stochastic analysis.

Findings

Provided conditions for solution smoothness inside the domain.

Established boundary behavior and convergence to prescribed data.

Generalized Bony's Harnack inequality for hypoelliptic operators.

Abstract

We study hypoelliptic stochastic differential equations (SDEs) and their connection to degenerate-elliptic boundary value problems on bounded or unbounded domains. In particular, we provide probabilistic conditions that guarantee that the formal stochastic representation of a solution is smooth on the interior of the domain and continuously approaches the prescribed boundary data at a given boundary point. The main general results are proved using fine properties of the process stopped at the boundary of the domain combined with hypoellipticity of the operators associated to the SDE. The main general results are then applied to deduce properties of the associated Green's functions and to obtain a generalization of Bony's Harnack inequality. We moreover revisit the transience and recurrence dichotomy for hypoelliptic diffusions and its relationship to invariant measures.