Re: brownian motion
- From: David C. Ullrich <ullrich@xxxxxxxxxxxxxxxx>
- Date: Tue, 16 May 2006 06:15:20 -0500
On 15 May 2006 17:36:42 -0700, "Fedor" <malabar_carotte@xxxxxxxxx>
wrote:
Hi all,
I have a question about brownian motion in R^n. Suppose that B_t(w)
is a brownian motion that starts from the origine and T(w) is the first
time that t->B_t(w) touches the unit sphere. What is the law of T ? I
can only do this for n=1 but I can't generalize for higher dimension
:-(
It's a unformly distributed measure on the sphere (ie
rotation-invariant). This is clear because brownian
motion is rotation-invariant, in a suitable sense.
(Which in turn is clear because a gaussian distribution
is rotation-invariant... or at least the gaussian
distributions that come up here.)
This fact is the essential reason why you can use
brownian motion to solve the Dirichlet problem
(from comments elsewhere I gather you know what
I mean by that). And then the fact that brownian
motion gives a solution to the Dirichlet problem
shows that if you start brownian motion at a point
of a ball other than the center and look at the
first point it hits the boundary the distribution
of that point is given by the Poisson kernel.
Thank you,
Fedor.
************************
David C. Ullrich
.
- Follow-Ups:
- Re: brownian motion
- From: BuddhaThu
- Re: brownian motion
- From: G. A. Edgar
- Re: brownian motion
- From: Fedor
- Re: brownian motion
- References:
- brownian motion
- From: Fedor
- brownian motion
- Prev by Date: Re: Equation for an ellipsoid
- Next by Date: Re: Increasing and decreasing functions - conflicting authors
- Previous by thread: brownian motion
- Next by thread: Re: brownian motion
- Index(es):
Relevant Pages
|