0
$\begingroup$

Lets say there exists a random vector $(X,Y)$ evenly distributed in the unit square $[0,1]^2$. Now lets introduce the following two new random variables $U=min(X,Y)$ and $V=max(X,Y)$, what is the value of $Cov(U,V)$.

I have an idea of what I should, use directly the definition of the covariance,

$Cov(U,V)=EUV-EUEV = EXY-EUEV=EXEY-EUEV=\frac{1}{4}-EUEV$

But the real problem is I have no idea how to get the density functions of $X$ and $Y$. If anybody has a hint or an idea it would be great.

$\endgroup$
  • $\begingroup$ What is the probability that both $X$ and $Y$ are less then a given value $c, 0 \le c \le 1$: $P(X < c, Y < c) = $?. If you can answer this, use the result to find $P(V < c)$. $\endgroup$ – Ingix Oct 16 '18 at 16:41
  • $\begingroup$ I know how to calculate $V$ and $U$ if I have $X,Y$, the problem I have is how to actually find the density function of $X and Y$ separately (should be the easiest part for some reason can't actually get it). $\endgroup$ – danielt17 Oct 16 '18 at 16:48
  • $\begingroup$ evenly distributed in the unit square gives you the density function... $\endgroup$ – karakfa Oct 16 '18 at 17:31
2
$\begingroup$

By definition, $$ U = \min(X,Y) = \begin{cases} X & ,~ \text{in the upper triangle of the unit square} \\ Y & ,~ \text{in lower triangle} \end{cases} $$ With the upper triangle region being the first integral (and the lower region being the 2nd integral), \begin{align} E[U] &= \int_{y = 0}^1 \left[ \int_{x = 0}^y 1\cdot x\, \mathrm{d}x \right]\, \mathrm{d}y + \int_{x = 0}^1 \left[ \int_{y = 0}^x 1\cdot y\, \mathrm{d}y \right]\, \mathrm{d}x \\ &= \int_{y = 0}^1 \frac{y^2}2\, \mathrm{d}y + \int_{x = 0}^1 \frac{x^2}2\, \mathrm{d}x \\ & = \frac13 \end{align} Similarly, for $V = \max(X,Y)$, we have \begin{align} E[V] &= \int_{y = 0}^1 \left[ \int_{x = 0}^y 1\cdot \color{magenta}{y}\, \mathrm{d}x \right]\, \mathrm{d}y + \int_{x = 0}^1 \left[ \int_{y = 0}^x 1\cdot \color{magenta}{x}\, \mathrm{d}y \right]\, \mathrm{d}x \\ &= \int_{y = 0}^1 y^2\, \mathrm{d}y + \int_{x = 0}^1 x^2\, \mathrm{d}x \\ & = \frac23 \end{align} Let me know if you need anything clarified.

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ thank you! now i understand my mistake $\endgroup$ – danielt17 Oct 16 '18 at 18:34

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.