4
$\begingroup$

The question Average Value of Area Closed by Brownian Motion turned out to be about the Lévy area process, which measures "signed area with multiplicity" enclosed by Brownian motion (e.g. each area gets a winding number coefficient giving the net number of times it was traversed counterclockwise). But Douglas Zare's comment suggested an version which counts area without sign and without multiplicity.

For a compact set $K \subset \mathbb{R}^2$, let $E(K)$, the "exterior" of $K$, be the unique unbounded component of $\mathbb{R}^2 \setminus K$. Let $\mathcal{A}(K) = m(\mathbb{R}^2 \setminus E(K))$ be the area (Lebesgue measure) of the region "enclosed" by $K$.

Preliminary question: let $\mathcal{K}(\mathbb{R}^2)$ be the set of all compact subsets of $\mathbb{R}^2$, with the Vietoris topology. Is $\mathcal{A} : \mathcal{K}(\mathbb{R}^2) \to [0,\infty)$ a Borel function?

Now let $B(t)$ be a standard 2-dimensional Brownian motion. Consider the process $A_t = \mathcal{A}(B([0,t]))$ which gives the area enclosed by $B$ up to time $t$. If the answer to the preliminary question is yes, then I believe $A_t$ is a stochastic process, jointly measurable and adapted to the filtration of $B$. It is nondecreasing, and I think maybe it is right continuous, but I am not quite sure.

What is known about $A_t$? Is it equal (in distribution) to some well-known process? What are its moments? Any other interesting properties?

$\endgroup$
  • $\begingroup$ math.stackexchange.com/questions/251856/… $\endgroup$ – Carlo Beenakker Apr 15 '15 at 0:51
  • $\begingroup$ @CarloBeenakker: Hmm, and I actually voted on that question. Seems that my memory doesn't last more than 2 years. Well, it got only a very partial answer there, maybe we will learn more from the MO crowd. $\endgroup$ – Nate Eldredge Apr 15 '15 at 3:00
5
$\begingroup$

In the case of the Brownian bridge in the plane (i.e. Brownian motion on the time interval $[0,1]$ conditioned to be back at the origin at time $1$), the expected area of the enclosed region is equal to $\pi/5$ as proved by Garban, cf. http://arxiv.org/abs/math/0504496 (and he also proves that the set of enclosed points where the index of the Brownian motion is nevertheless equal to $0$ has expected area $\pi/30$.

This doesn't answer the initial question, but perhaps the methods in this reference can help ...

$\endgroup$

Your Answer

By clicking "Post Your Answer", you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.