2
$\begingroup$

Assume $$\{X_i\}\stackrel{\mathrm{i.i.d.}}{\sim} \mathcal{Uniform}(0,1)$$ Find the joint p.d.f. of $$X_{(n)} \hat= \max \{X_1,X_2,\ldots,X_5\}\quad\text{ and }\quad \bar X\hat=\sum^n_{i=1}{X_i}$$

The motivation of this question: If a line of length 1 is randomly divided into five parts, computing the probability of there exist one part with length more than 1/4, i.e. find $\Pr\{X_{(n)}\ge \frac{5\bar X}{4} \}$ with the setting above.

$\endgroup$
  • 1
    $\begingroup$ It is hard to see how $X_{(n)}\ge 5\bar{X}/4$ describes the event that there is one part with length more than $1/4.$ For instance, $X_{(n)}$ could be nearly zero and the other four values could all exceed $1/2,$ leaving an initial gap of $1/2.$ $\endgroup$ – whuber May 7 at 22:08
  • $\begingroup$ @whuber It's a typo, thank you. $\endgroup$ – C.X.Neo May 8 at 7:11
  • 2
    $\begingroup$ Please edit your post to display the expression you really want to evaluate, then! $\endgroup$ – whuber May 8 at 14:11
  • 1
    $\begingroup$ @whuber this question wants to find the joint p.d.f of $X_{(n)}$ and $\bar X$ $\endgroup$ – C.X.Neo May 9 at 11:32
3
$\begingroup$

$X_1, \dotsc,X_n$ is iid from the uniform distribution on $[0,1]$. The maximum is $X_{(n)}$ which has a distribution with density $f(u)=n u^{n-1}$. Represent the joint density of the mean and the maximum as $f(m, u)=f(m \mid u) f(u)$ where $f(m \mid u)$ represents the conditional density of the mean given the maximum.

The order statistics $X_{(1)},\dotsc, X_{(n-1)}$, conditional on the maximum behaves as order statistics from an iid sample of size $n-1$ from the uniform distribution on $[0,u]$, so we can use that and the representation of the mean as $$ \bar{X}_n = \frac{u}{n}+ \frac{n-1}{n}\bar{X}_{n-1} $$ when the maximum $u$ is given. Then the distribution of $\bar{X}_{n-1}$ is a scaled version of the Irwin-Hall distribution, see also Can we make the Irwin-Hall distribution more general?. I will leave it at that, you should be able to complete.

$\endgroup$
  • 1
    $\begingroup$ Thank you, it helps :) $\endgroup$ – C.X.Neo May 8 at 7:22

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.