2
$\begingroup$

Let $f$ be a real-valued stochastic process adapted to the natural filtration $\{\mathcal{F}_t\}_{t\geq0}$ of some Brownian motion $\{W_t\}_{t\geq0}$ such that

$\mathbb{E}\bigg[\int_{0}^{T}f^2(t,w)dt\bigg]<\infty $ and $\int_{0}^{T}f(t,w)dW_t=0.$

Then how one can prove that $f:=0$ for all $t\in[0,T].$

By Ito isometry, $\mathbb{E}\bigg[\int_{0}^{T}f^2(t,w)dt\bigg]=\mathbb{E}\bigg[\int_{0}^{T}f(t,w)dW_t\bigg]^2=0$. This implies $\mathbb{E}\bigg[\int_{0}^{T}f^2(t,w)dt\bigg]=0$. Then I got stuck to proceed. Anyone help?

$\endgroup$
1
$\begingroup$

If $\mu$ is a non-negative measure on a measure space $(X,\mathcal{A})$ and if $g\ge0$, then $ \int_X g(x) \, d\mu(x) = 0$ implies $\mu(g \neq 0)=0$.

Here, we consider the product measure $\mu := \lambda_T\otimes \mathbb{P}$ (where $\lambda_T $ denotes the Lebesgue measure restricted to $[0,T]$). Since, by Tonelli's theorem,

$$0=\mathbb{E} \left( \int_0^T f^2(t,\cdot) \, dt \right)= \int_{[0,T] \times \Omega } f^2(t,\omega) \, d\mu(t,\omega)$$

we find $$\mu(\{(t,\omega) \in [0,T] \times \Omega; f(t,\omega) \neq 0\})=0,$$ i.e. $f=0$ for almost all $t,\omega$.

$\endgroup$
  • $\begingroup$ @Did Thanks.... $\endgroup$ – saz Mar 13 '17 at 9:52
  • $\begingroup$ You are welcome. $\endgroup$ – Did Mar 13 '17 at 9:52

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.