Exit times from time-dependent random domains: continuity, weak convergence, and exit-time profiles Draft -currently under review at Stochastic Processes and their Applications

TL;DR

This paper investigates the continuity and convergence properties of exit times from time-dependent domains, providing theoretical results, explicit conditions, and practical examples for stochastic processes.

Contribution

It establishes a deterministic continuity theorem for exit times, characterizes the conditions for weak convergence, and introduces a functional limit theorem with concrete verification methods.

Findings

Exit-time functional is continuous under local Skorokhod J1 convergence with non-tangency condition.

Weak convergence of exit times follows from joint convergence of paths and barriers under NT.

Exit-time profile converges in Skorokhod M1 topology, with examples showing J1 convergence can fail.

Abstract

We study exit times from time-dependent domains under joint perturbations of the trajectory and the domain. Representing a moving domain by a continuous barrier $\Phi$ on space-time, we reduce the exit problem to a one-dimensional first-passage problem for the scalarised path $y(t) := \Phi(t,x(t))$. Our first main result is a deterministic continuity theorem: the exit-time functional is continuous, under local Skorokhod $J_1$ convergence of the path and local uniform convergence of the barrier, at every configuration satisfying an explicit non-tangency condition (NT). We show that NT is sharp in the sense that it characterises the continuity set of the functional. As a direct consequence, weak convergence of exit times follows from joint weak convergence of paths and barriers whenever the limiting pair satisfies NT almost surely; no independence or structural restrictions between trajectory and domain are required. Our second main result is a functional limit theorem: the exit-time profile $u\mapsto\tau(u)$, viewed as a c\`adl\`ag function of the barrier level, converges in the Skorokhod $M_1$ topology under the same hypotheses, with a concrete example showing that $J_1$ convergence can fail. Concrete verification routes for NT are provided, including a non-characteristic/It\^o criterion for diffusions, and the full framework is illustrated through a worked Donsker-type example.