Understanding Radon-Nikodym Derivative for Drifted Brownian Motion

105 Views Asked by At

I'm working on Drifted Brownian Motion at the moment and I'm having some difficulty with understanding Girsanov's Theorem (from Shreve):

Let $W(t) \ (0 \leq t \leq T)$ be a BM on probability space $(\Omega, F, \Pr)$. Let $\Theta(t) \ (0 \leq t \leq T)$, be an adapted process. Define

$$Z(t) = \exp \left( -\int_0^t\Theta(u)\mathrm{d}W(u) - \frac{1}{2}\int_0^t \Theta^2(u)\mathrm{d}u \right),$$

$$\widetilde{W}(t) = W(t) + \int_0^t\Theta(u)\mathrm{d}u,$$

and assume that $\displaystyle\mathbb{E}\left[\int_0^T \Theta^2(u)Z^2(u)\mathrm{d}u\right] < \infty$. Set $Z=Z(T)$. Then $\mathbb{E}[Z]=1$ and >under probability measure $\widetilde P$ given by

$$\widetilde \Pr(A) = \int_A Z(\omega)\mathrm{d}\Pr(\omega)\qquad\forall A \in F$$

the process $\widetilde W(t) \ (0 \leq t \leq T)$ is a Brownian Motion.

Say I want to apply this thorem to $W(t) + \mu t$, then, $Z(t) = e^{-\mu W(t) - \frac{1}{2}\mu^2}$. Would you then have

$$\widetilde P(A) = e^{-\mu W(T) - \frac{T}{2}\mu^2}\int_A\mathrm{d}\Pr(\omega)\qquad\forall A \in F$$

as a function for the probability under $\widetilde \Pr$? I dont really understand what you would need to do with the $\omega$ in $Z(\omega)$.

1

There are 1 best solutions below

0
On

The expression $$\widetilde{P}(A) = e^{-\mu W(T)-\frac{T}{2}\mu^2} \int_A dP(\omega)$$ doesn't make sense because $\widetilde{P}(A)$ is just a number (the probability of something), but $W(T)$ is random.

For the question about what to do with $\omega$, remember that $\omega$ is what keeps track of the randomness in the random variables. When we write $W(T)$, we really mean $W(T,\omega)$. So the correct expression would be $$\widetilde{P}(A) = \int_A e^{-\mu W(T)-\frac{T}{2}\mu^2} dP(\omega) = \int_A e^{-\mu W(T,\omega)-\frac{T}{2}\mu^2} dP(\omega).$$