how to get brownian bridges maximum and minimum's joint distribution

1k Views Asked by At

I'm wondering how to get brownian bridge's maximum and minimum's joint distribution.Brownian bridge's definition is as below.

$B_t = W_t - tW_1$ for $t \in [0,1]$, W is wiener process

How to get it? In fact, I didn't get maximum's distribution of brownian bridge. I cannot catch any idea.. I tried to get this distribution via conditional probability, that fix $W_1$, but it is still hard one. Can you help me?

edit) I think that it is concerned with joint distribution of brownian motion's maximum and minimum, and W(1). Then how to attain this one?

1

There are 1 best solutions below

0
On BEST ANSWER

Let $M$ and $m$ be the Max and min of $B_t$. Let us find the joint cdf of $M$ and $-m$, defined by $$ P(\{M< a\}\cap \{-m< b\}), $$ In fact, the argument you can use is exactly the continuous version of the discrete problem here: A generalization of reflection. Namely, Brownian motion is the limit of a discrete random walk, and there is an exact formula for the number of random walks of a given length who stay between two horizontal lines, using the reflection principle.

Applying that logic, we get that $$ P(M<a,-m<b) =\frac1{P(W_1=0)}\left( \sum_{k\in \mathbb Z} P\big(W_1=2k(a+b)\big) - \sum_{k\in \mathbb Z} P\big(W_1=2k(a+b)+2a\big) \right) $$ Note that this does not quite make sense, as all the probabilities are zero, but you have to understand this as a limit. Anyways, what you get is $$ P(M<a,m<-b)=\sum_k\exp(-2k^2(a+b)^2)-\sum_k\exp\left(-2\big(k(a+b)+a\big)^2\right) $$ This gives the joint cdf of $M$ and $-m$.