The joint distribution of the running maximum
$ M_t = \max_{0 \leq s \leq t} W_s $
and $W_t$ is
$f_{M_t,W_t}(m,w) = \frac{2 ( 2 m – w)}{t\sqrt{2 \pi t}}e^{-\frac{(2m-w)^2}{2t}}, m \ge 0, w \leq m $ (http://en.wikipedia.org/wiki/Wiener_process#Running_maximum)
Question is,
what is running maximum here? Is it same as http://mathworld.wolfram.com/RunningMaximum.html? If this is what "running maximum" is, I am not getting what $M_t$ exactly is referring to. Can anyone explain this?
Best Answer
Let $(W_t)_{t \geq 0}$ a Wiener Process on a probability space $(\Omega,\mathcal{A},\mathbb{P})$. Then
$$M_t(\omega) := \max_{0 \leq s \leq t} W_s(\omega) \qquad (\omega \in \Omega)$$
is the running (pathwise!) maximum. This means that $M_t(\omega)$ is the maximum of the path $$[0,t] \ni s \mapsto W_s(\omega) \in \mathbb{R}$$
$\hspace{140pt}$
The mapping $[0,\infty) \times \Omega \ni (t,\omega) \mapsto M(t,\omega)=M_t(\omega)$ is again a stochastic process.