Let $B_t$ be a Brownian motion with variance 1. We know that $\int_0^1 B(t) \mathrm{d} t \sim \mathcal{N}(0,1/3)$. I am interested to know what we can say about the law of the two random variables $X = \int_{0}^1 B(t)^2 \mathrm{d}t = \langle B,B\rangle_{L^2([0,1])}$ and $Y = \int_{0}^{1} \left( B(t) - \int_0^1 B(s)\mathrm{d}s \right)^2 \mathrm{d}t = X - \left( \int_0^1 B(s)\mathrm{d}s \right)^2$.

Is there a well-known probability law hidden behind them?

Thank you for attention.


Aspects of Brownian Motion (Mansuy & Yor) give an expression for the joint Laplace transform of $B_t$ and $\int_0^t B_s^2 ds$ (section 2.1). For $\delta$-dimensional Brownian motion, $$ \mathbb{E}\left[\exp\left(-\alpha|B_t|-\frac{b^2}{2}\int_0^t |B_s|^2 ds \right) \right] $$ is equal to $$\left(\text{ch}(bt)+2\frac{\alpha}{b}\text{sh}(bt)\right)^{-\delta/2}\exp\left(-\frac{xb}{2}\frac{(1+\frac{2\alpha}{b}\coth(bt)}{(\coth(bt)+\frac{2\alpha}{b}}\right). $$

I don't recognize the generating function, even for $\delta=1$ and $\alpha=0$; and neither do Manuy & Yor give it a name.

See also Albin 1995 for some other references.


The distribution of Y has been studied in great details by Donati-Martin and Yor in their work: Fubini's theorem for double Wiener integrals and the variance of the Brownian motion path


In particular it is proved that in distribution

$Y=^{law}\int_0^1 | W_s |^2 ds$

where W is a standard complex Brownian bridge.


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.