3
$\begingroup$

Is there any prototypical example of a Random process with smooth paths?

I imagine one can simple integrate each path of a Brownian motion and get a $C^{\frac32-\epsilon}$ path.

It's easy to describe paths of Brownian bridges over a circle using Fourier expansion and the normal distribution:

$$ B(w, x) = \sum_{n \in \mathbb{Z}} C_n(w) e^{inx} $$

  • $C_n \sim \mathcal{N}(0,1)$ we obtain White Noise.
  • $C_n \sim \mathcal{N}(0,\left< n \right>^{-2})$ we obtain a Brownian Bridge, paths are a.s. in $H^{\frac12-\epsilon}$
  • $C_n \sim \mathcal{N}(0,\left< n \right>^{-4})$ paths are a.s. in $H^{\frac32-\epsilon}$, so we now have at least one derivative.
  • $C_n \sim \mathcal{N}(0,\left< n \right>^{-2k})$ paths are a.s. in $H^{\frac{2k-1}{2}-\epsilon}$, so we now have at least $k$ derivatives.
  • For a $C^\infty$ maybe with $C_n \sim \mathcal{N}(0,e^{-\left< n \right>})$ paths would be a.s. in $C^\infty$?.

I can't find any literature about non-trivial random processes with smooth paths.

Do you know of any literature related with it?

I'm trying to understand something on set-valued functions, and the way an integral for such functions is defined seems related to random processes (http://econweb.ucsd.edu/~rstarr/201/Aumann%20Journal%20of%20Mathematical%20Analysis%20and%20Applications%201965.pdf).

Thanks.

$\endgroup$

2 Answers 2

3
$\begingroup$

A good place to read about this is Adler and Taylor's book Random Fields and Geometry. Regular random processes or functions are used more frequently in integral geometry.

Consider the more general Gaussian process on $S^1$ $\newcommand{\bZ}{\mathbb{Z}}$ $\newcommand{\ii}{\boldsymbol{i}}$

$$ F(\theta)=\sum_{n\in\bZ} C_n e^{n\ii \theta}, $$

where $C_n$ are independent centered Gaussian random variables. Then $F$ is a.s. smooth if $\newcommand{\vfi}{\varphi}$ $\newcommand{\bE}{\mathbb{E}}$ $\newcommand{\var}{\boldsymbol{var}}$ the covariance kernel

$$K(\theta,\vfi)=\bE\bigl[ F(\theta)\overline{F(\vfi)}\bigr]= \sum_{n\in\bZ}\var[C_n] e^{\ii n(\theta-\vfi)} $$

is smooth. This happens if

$$\sum_{n\in\bZ} n^{2s}\var[C_n]<\infty,\;\;\forall s>0. $$

For stationary processes like this one this condition is also necessary.

$\endgroup$
1
$\begingroup$

One way to get a $C^{\infty}$ path from a Brownian motion, without changing the original path too much, is to mollify it.

$\endgroup$
1
  • $\begingroup$ Hi, right, ultimately is just a multiplication on the Fourier expansion by a fast decaying function, as you multiply a normal r.v. the term would go to the variance similar as in the last example, which would ensure that any $H^k$ norm is finite, but there is any real usage of this? any literature about it? $\endgroup$ May 1, 2016 at 7:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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