Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Let's say I have a Brownian particle of some radius $r_b$ and coefficient of diffusion $D$, freely moving about in a toroidal/doughnut-shaped chamber with inner and outer radius $R_{inner}$ and $R_{outer}$, respectively. As such, the circle that can be rotated to generate the torus has radius $R_{chamber} = \frac{1}{2}(R_{outer}-R_{inner})$, and $r_b < R_{chamber}$, though the particle is not necessarily point-like. Now, let's also say that someone places an invisible, circular reflecting boundary (of radius $R_{chamber}$) somewhere inside the toroid s.t. the Brownian particle can no longer return to a previous position by travelling in a strictly clockwise or counterclockwise fashion around the torus.

Here one should presumably be able to watch the trajectory of the Brownian particle to assign a probability for the location of the reflecting boundary. In practice though, how should one proceed? Provided the particle's diffusion coefficient $D$, how does the one's confidence in the position of the reflecting boundary change over time?

Clarification - The heart of this question is: while it's obviously possible to do so, how does one actually go about mapping the position of a reflecting barrier (of known geometry or not) using the trajectory of a Brownian particle?

share|improve this question

1 Answer 1

up vote 3 down vote accepted

If a Brownian motion hits a reflecting barrier like that, it is almost surely going to hit multiple times. In fact, it will almost surely hit in a Cantor set of times and locations, just as an unreflected one-dimensional Brownian motion almost surely returns to the value at $t=0$. This type of behavior almost surely doesn't happen if there is no barrier. The probability that the maximums of an unreflected one-dimensional Brownian motion on disjoint intervals are equal is $0$.

Let $cw(t)$ be the angle of the plane through the axis which is tangent to the particle on the clockwise side at time $t$, and let $ccw(t)$ be the analogue for the counterclockwise side. If the minimum of $cw(t)$ is achieved multiple times, then almost surely the particle has reflected off a boundary at $\min (cw(t))$. If the maximum of $ccw(t)$ is achieved multiple times, then almost surely the particle has reflected off a boundary at $\max (ccw(t))$. If neither occurs then you have restricted the possible locations of the boundary, excluding an interval $(\min(cw(t)),\max(ccw(t)))$.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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