So I have done this exercise and the proof holds, but I really don't believe it can be correct because the proof is worth twice as much as other exercises. I am also not 100% sure if $d_s d_r=d_rdW_s=d_sdW_r=dW_r dW_s=0$.

Feel free to give any feedback or correct me, thanks a lot!

The exercise says:

Let $X_{t}$ be an Ito process, given by $X_{t} = X_{0} + \int_{0}^{t} \mu_s ds+\int_{0}^{t} \sigma_s dW_s$, where $X_0\in\mathbb{R}$, and let $Y_t$ be another Ito process given by $Y_{t} = Y_{0} + \int_{0}^{t} b_s ds+\int_{0}^{t} h_s dW_s$, where $Y_0\in\mathbb{R}$.

Show that $X_tY_t=X_0Y_0+\int_0^t X_s dYs+\int_0^t Y_s dX_s+[X,Y](t)$, where $[X,Y](t)=\int_0^t\sigma_s h_s ds$.

And my answer is...

Consider $dX_t=\mu_t dt+\sigma_t dW_t$ and $dY_t=b_t dt+h_t dW_t$ and let $LHS=X_tY_t$ and $RHS=X_0Y_0+\int_0^t X_s dYs+\int_0^t Y_s dX_s+\int_0^t\sigma_s h_s ds$.

This gives: $$LHS=X_{t} Y_{t}=\left(X_{0} +\int_{0}^{t} \mu_s ds+\int_{0}^{t} \sigma_s dW_s\right)\left(Y_{0} + \int_{0}^{t} b_s ds+\int_{0}^{t} h_s dW_s\right)=X_0 Y_0+X_0\int_0^t b_s ds+X_0\int_0^t h_s dW_s+Y_0\int_0^t \mu_s ds+\int_0^t\mu_s b_s (ds)^2+\int_0^t \mu_s h_s ds dW_s+Y_0\int_0^t \sigma_s dW_s+\int_0^t\sigma_s b_s dW_s ds+\int_0^t \sigma_s h_s (dW_s)^2=X_0 Y_0+X_0\int_0^t b_s ds+X_0\int_0^t h_s dW_s+Y_0\int_0^t \mu_s ds+Y_0\int_0^t \sigma_s dW_s+\int_0^t \sigma_s h_s ds,$$ since $dW_s ds=(ds)^2=0$ and $(dW_s)^2=ds$.

Now the right-hand side (RHS) gives:

$$RHS=X_0Y_0+\int_0^tX_sdYs+\int_0^tY_sdX_s+\int_0^t\sigma_s h_s ds= X_0Y_0+\int_0^t\left(X_{0} + \int_{0}^{s} \mu_r dr+\int_{0}^{s} \sigma_r dW_r\right)\left(b_s ds+h_s dW_s\right)+\int_0^t\left(Y_{0} + \int_{0}^{s} b_r dr+\int_{0}^{s} h_r dW_r\right)\left(\mu_s ds+\sigma_s dW_s\right)+\int_0^t\sigma_s h_s ds=$$ And using $d_s d_r=d_rdW_s=d_sdW_r=dW_r dW_s=0$, we get $LHS=RHS$.

  • $\begingroup$ Also, if it is correct, feel free to tell me so :) $\endgroup$ – jmmatias1985 Aug 10 '15 at 16:22
  • $\begingroup$ No, your calculations are not at all correct. Do you really think that $$\int_0^t \mu(s) \, ds \cdot \int_0^t b(s) \, ds = \int_0^t b(s) \mu(s) (ds)^2 = 0$$ for any $b$ and $\mu$? The first as well as the second "=" is (in general) not correct. $\endgroup$ – saz Aug 10 '15 at 17:46
  • $\begingroup$ I think I get it now, I have to use Ito's product rule, right? $\endgroup$ – jmmatias1985 Aug 10 '15 at 23:17
  • $\begingroup$ The thing is that I know how to prove it in a general case (it is in my book), but without using the formulas given and I find it really straightforward... But I wanted to try to do it using the formulas given. $\endgroup$ – jmmatias1985 Aug 11 '15 at 0:18
  • 1
    $\begingroup$ Yes, you have to use Itô's product rule (or Itô's formula). $\endgroup$ – saz Aug 11 '15 at 4:52

Ito product formula is direct from Ito lemma so you can try to derive that first. Once you are ok with the Ito product formula, you start by writing $$d(X_tY_t)=dX_tY_t+X_tdY_t+dX_tdY_t = (Y_0+\int_0^tb_sds+\int_0^th_sdW_s)(\mu_tdt+\sigma_tdW_t)+(X_0+\int_0^t\mu_sds+\int_0^t\sigma_sdW_s)(b_tdt+h_tdW_t)+(\mu_tdt+\sigma_tdW_t)(b_tdt+h_tdW_t)$$ then apply the multiplication rules ($dtdt=0...)$, integrate on each sides (not forgetting that on the RHS $X_0Y_0$ appears), and you will be done.


I really dislike writing things like "dt dt = 0". One direct way is if you know the Ito rule for processes in $R^n$. In that case, apply it to the $R^2$ process $Z_t = (X_t, Y_t)$ and the function $f(x,y) = xy$


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.