A method of solution for the inverse problem for $h$-functions of planar Brownian motion

TL;DR

This paper addresses the inverse problem of reconstructing planar domains from harmonic measure distribution functions using Brownian motion and conformal invariance, providing solutions for a class of functions.

Contribution

It extends the concept of harmonic measure functions to stopping times and solves the inverse problem for a broad class of these functions using conformal invariance.

Findings

Solved the inverse problem for a class of harmonic measure functions

Extended harmonic measure functions to stopping times

Provided a method to reconstruct domains from harmonic measure functions

Abstract

Given a planar domain $D$, the harmonic measure distribution function $h_D(r)$, with base point $z$, is the harmonic measure with pole at $z$ of the parts of the boundary which are within a distance $r$ of $z$. Equivalently it is the probability Brownian motion started from $z$ first strikes the boundary within a distance $r$ from $z$. We call $h_D$ the $h$-function of $D$, this function captures geometrical aspects of the domain, such as connectivity, or curvature of the boundary. This paper is concerned with the inverse problem: given a suitable function $h$, does there exist a domain $D$ such that $h = h_D$? To answer this, we first extend the concept of a $h$-function of a domain to one of a stopping time $\tau$ . By using the conformal invariance of Brownian motion we solve the inverse problem for that of a stopping time. The associated stopping time will be the projection of a hitting time of the real line. If this projection corresponds to the hitting time of a domain $D$, then this technique solves the original inverse problem. We have found a large family of examples such that the associated stopping time is that of a hitting time.