If I have an i.i.d. random variables $X_i$ from a uniform distribution on $[0,1]$.

How would I find scaling sequences $a_n$, $b_n$ such that $a_n(M_n − b_n)$ converges in distribution to a non-trivial limit function $G$ for

$ Y_i = X_i,$ and $M_n = \max(Y_1,...,Y_n) $


1 Answer 1

  1. Answer to the earlier version of the question which mentioned $M_n$ is distributed as $ \max\{ X_1, X_2, \dots, X_n\}$:

Note that, $$\Pr(a_n(M_n - b_n)\le x) = \Pr(M_n\le b_n +x/a_n) = \prod_{i=1}^n \Pr(Y_i\le b_n +x/a_n) = \left(b_n +x/a_n \right)^n.$$ So, one possible choice would be $b_n=1$ and $a_n=n.$ In that case $n(M_n - 1)\Rightarrow Z$ where $-Z$ follows standard exponential distribution.

  1. Answer to the question which says $M_n$ is distributed as $ \max\{ 1/X_1, 1/X_2, \dots, 1/X_n\}$:

Let $Y_i = 1/X_i.$ Note that,$$\Pr(a_n(M_n - b_n)\le x) = \prod_{i=1}^n \Pr(Y_i\le b_n +x/a_n) = \prod_{i=1}^n \Pr\left(X_i \ge \left(b_n +x/a_n\right)^{-1}\right) = \left(1-\frac{a_n}{a_n b_n + x}\right)^n.$$ Now, it is well known that if $\displaystyle\lim_{n\to\infty} x_n = x,$ then $\displaystyle\lim_{n\to\infty} (1+x_n/n)^n = e^x.$ Using this fact, we can write $$\lim_{n\to\infty} \left(1-\frac{a_n}{a_n b_n + x}\right)^n = \exp\left(\displaystyle\lim_{n\to\infty}-\frac{na_n}{a_n b_n + x}\right).$$ Thus, we can take $a_n = 1/n$ and $b_n=0,$ for which the last limit will be $\exp(-1/x)$. Thus, $n^{-1}(M_n - 0)\Rightarrow Z$ where $Z$ has CDF $F(x) = \exp(-1/x) \mathbf{1}(x\ge 0).$

Note, if $b_n$ is any sequence such that $b_n/n$ converges, say to $b,$ then $n^{-1}b_n\to b$ and $n^{-1} M_n \Rightarrow Z$ implies that $n^{-1} (M_n - b_n) \Rightarrow Z-b, $ whose CDF is $F(x+b).$

  • $\begingroup$ thank you so so much I'm still abit confused is the bit were it says find 'an bn s.t an(mn-bn) converges to a non trivial limit function G' connected to the Yi=Xi and Mn=max(Y1,...Yn) or are they seperate things, Because I also have Ui = 1/Xi, and Mn = max{U1,...,Un} $\endgroup$ Jan 12, 2020 at 17:18
  • $\begingroup$ Did you mean $Y_i = 1/X_i?$ But your statement of the problem shows $Y_i = X_i$... $\endgroup$ Jan 12, 2020 at 18:32
  • $\begingroup$ Thats so amazing thank you, yes I would like to know both as you have demonstrated. I just wanted to ask one last thing is it possible to in each case to find suitable scaling sequences an, bn, where un = u/an + bn so that you get a non-trivial limit G(u) as n → ∞. $\endgroup$ Jan 14, 2020 at 18:41

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.