1
$\begingroup$

Let $X_0$ be a standard normal random variable and suppose that $$dX_t=-\frac{1}{2}X_tdt+dB_t.$$ $X_0$ is independent of the Brownian motion. Find the distribution of $X_t$ for $t\geq0$ and find $\text{cov}(X_t,X_s)$ for all $t,s$.

So far, I have attempted to solve the SDE using Ito's lemma, but this has seemed to imply that there exists no solution of the form $X_t=f(t,B_t)$. So I am beginning to think that maybe it is possible to describe the distribution without explicitly solving the SDE. But I'm not sure how to do this - any advice would be greatly appreciated!

$\endgroup$
1

1 Answer 1

1
$\begingroup$

From the general solution to linear SDEs, you have :

$$X_t=e^{-\frac{1}{2}t}\left(X_0+\int_0^t e^{\frac{1}{2}s}dB_t\right)$$

So what do we have here ? A deterministic term $e^{-\frac{1}{2}t}X_0$ and a martingale term (why ?) with deterministic integrand. So such stochastic integral are known as Wiener integrals and are gaussian with expectation $0$, and variance equals to the expectation of quadratic variation of the integral (please look up in this forum multiple proofs of this fact) . Using Ito's isometry we get by aggregating all of this : $$X_t \sim \mathcal{N} (e^{-\frac{1}{2}t}.X_0, 2\sinh (t)) $$

unless mistaken (I let you check the algebra) here

$\endgroup$

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.