0
$\begingroup$

Suppose we have 4 positive independent random variables, $a_1, a_2, b_1$, and $b_2$. Moreover, $a_1, a_2$ are identically distributed and have a PDF denoted by $f_A(a)$. Similarly, $b_1, b_2$ have a PDF denoted by $f_B(b)$. I need to find the joint distribution of these two random variables:

Z=max($a_1 b_1, a_2 b_2$);

Y=$a_i$, where $i=argmax_{i \in{1,2}} a_i b_i$.

Any idea how can I approach this problem?

$\endgroup$
  • $\begingroup$ Are you seeking the joint pdf of $Z$ and $Y$? $\endgroup$ – wolfies Oct 8 '16 at 7:21
  • $\begingroup$ @wolfies Yes, the joint PDF of Y & Z. $\endgroup$ – Alammouri Oct 8 '16 at 18:51
0
$\begingroup$

Here is what I got so far, which is sufficient for my problem:

$F_{Z,Y}(z,y)=\mathbb{P} \left(Z<z,Y<y \right)=\mathbb{P} \left(a_1 b_1<z,a_2 b_2<z,a_1<y,a_2<y \right)$

$=\mathbb{P} \left( a_1< \min \left(\frac{z}{b_1},y\right) ,a_2< \min \left(\frac{z}{b_2},y\right) \right)$

Exploiting the i.i.d property:

$= \left( \int\limits_{0}^{\infty} F_{A}\left(\min \left(\frac{z}{b},y\right)\right)f_{B}(b) \ db\right)^2$

$\endgroup$

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.