학술

Brownian Convex Hulls in Arbitrary Dimension: Absorption and Intersection Probabilities

arXiv Math
CC BY
이 매체는 공공·자유 라이선스로 본문을 직접 표시합니다.

Abstract

For every $d\ge1$, we determine prescribed-point absorption probabilities for the convex hulls of standard Brownian motion and Brownian bridges in $\mathbb R^d$, obtaining explicit one-dimensional integral formulas involving standard special functions. For example, in dimension three the Brownian-motion absorption probability for a point at distance $r>0$ from the origin is \[ p_3^B(r)=\frac{r}{\sqrt2\,\pi^{3/2}}\int_0^\infty\left(\frac{\pi}{2\sqrt{1+v^2}}-1\right)\mathrm{e}^{-r^2(1+v^2)/4}(K_1-K_0)\!\left(\frac{r^2(1+v^2)}4\right)\,\mathrm{d}v, \] where $K_\nu$ denotes the modified Bessel function of the second kind. For fixed $d$, the non-absorption probabilities of a prescribed nonzero point by Brownian-motion and Brownian-bridge hulls over $[0,T]$ are asymptotic to explicit positive constants times $T^{-1/2}(\log T)^{d-1}$ and $T^{-1}(\log T)^{d-1}$, respectively, as $T\to\infty$.
We determine the Mellin and Laplace transforms of the first time $\tau_d$ a Brownian-motion hull absorbs a prescribed point at unit distance. In high dimensions, both $\tau_d$ and the time $\tau_d^{\mathrm{ball}}$ to absorb the entire unit ball satisfy the same central limit theorem: \[ \frac{\log\tau_d-2d}{2\sqrt d}\xrightarrow{\mathrm d}N(0,1),\qquad\frac{\log\tau_d^{\mathrm{ball}}-2d}{2\sqrt d}\xrightarrow{\mathrm d}N(0,1), \qquad d\to\infty.\] Thus both times have leading exponential scale $\mathrm{e}^{2d}$. For Brownian bridges, prescribed-point and unit-ball absorption also exhibit the same Gaussian transition, but on the shorter scale $\mathrm{e}^d$.
For two independent Brownian motions started one unit apart, we show that the non-intersection probability of their convex hulls is of order $T^{-1}(\log T)^{d-1}$ for fixed $d$ and large $T$. In high dimensions, the first intersection time has leading exponential scale $\mathrm{e}^d$, and its logarithm satisfies a central limit theorem.

관련 뉴스