Is Proving (ii) Implies (i) in Limit Theorems Trivial?

mathstatnoob
Messages
3
Reaction score
0

Homework Statement


The lemma below is from Sample-Path Analysis of Queueing Systems by El-Taha and Stidham Jr.

Lemma 2.10
Let \{x_n, n \geq 1\} be a sequence of non-negative real numbers and \{b_n, n \geq 1\} a non-decreasing sequence of real numbers such that b_n \rightarrow \infty as n \rightarrow \infty. Then the following are equivalent:

(i)b_n^{-1}x_n \rightarrow c as n \rightarrow \infty;
(ii)b_n^{-1}\max_{k \leq n}x_k \rightarrow c as n \rightarrow \infty.

The authors claim that proving statement (ii) implies statement (i) is trivial and leave no proof (they do go on to prove (i) implies (ii)). However, "trivial" is a subjective term. I'm trying to provide a proof that (ii) implies (i), but I get stuck.

2. Attempt at a solution
Let \epsilon > 0. It follows from the hypothesis that there exists an N_0 such that if n > N_0, then c-\epsilon < b_n^{-1}\max_{k \leq n}x_k < c+\epsilon.

Since b_n \rightarrow \infty as n \rightarrow \infty,there exists an N_1 such that if n > N_1, then b_n > 0 which implies b_n^{-1} > 0 which then implies b_n^{-1}x_n \leq b_n^{-1}\max_{k \leq n}x_k.

So let N=\max\{N_0,\, N_1\}. If n > N, then b_n^{-1}\max_{k \leq n}x_k < c+\epsilon. Hence b_n^{-1}x_n < c+\epsilon. Now either (a) c-\epsilon < b_n^{-1}x_n, or (b) b_n^{-1}x_n \leq c-\epsilon. If case (a) is true, then c-\epsilon < b_n^{-1}x_n < c+\epsilon whenever n > N. Regarding case (b)...

...and this is where I get stuck. My guess is case (b) leads to some contradiction, but I'm not sure how to obtain one. Any suggestions or specific references to look at? Or if there's a simpler way to show that (ii) implies (i), I'm all ears. Thanks!
 
Physics news on Phys.org
Welcome to PF, mathstatnoob!

I'm not sure, but I think breaking the proof into two cases may help, one if x_n is bounded and the other if it is unbounded.

What does 1/b_n converge to? Consequently what do \frac{x_n}{b_n} and \frac{\max_{k \leq n}x_k}{b_n} converge to in the first case?

In the second case, I think you can use the unboundedness of x_n to show (i) and (ii) are equivalent.

Hope this helps!
 
spamiam said:
Welcome to PF, mathstatnoob!

I'm not sure, but I think breaking the proof into two cases may help, one if x_n is bounded and the other if it is unbounded.

What does 1/b_n converge to? Consequently what do \frac{x_n}{b_n} and \frac{\max_{k \leq n}x_k}{b_n} converge to in the first case?

In the second case, I think you can use the unboundedness of x_n to show (i) and (ii) are equivalent.

Hope this helps!
Thanks for your input, spamiam.

So in the case when x_n is bounded, both \frac{x_n}{b_n} and \frac{\max_{k \leq n}x_k}{b_n} approach 0 as n approaches infinity. I showed this by employing the squeeze theorem. Let me think about the case when x_n is unbounded a bit more (I just wanted to quickly acknowledge your suggestion).
 
Prove $$\int\limits_0^{\sqrt2/4}\frac{1}{\sqrt{x-x^2}}\arcsin\sqrt{\frac{(x-1)\left(x-1+x\sqrt{9-16x}\right)}{1-2x}} \, \mathrm dx = \frac{\pi^2}{8}.$$ Let $$I = \int\limits_0^{\sqrt 2 / 4}\frac{1}{\sqrt{x-x^2}}\arcsin\sqrt{\frac{(x-1)\left(x-1+x\sqrt{9-16x}\right)}{1-2x}} \, \mathrm dx. \tag{1}$$ The representation integral of ##\arcsin## is $$\arcsin u = \int\limits_{0}^{1} \frac{\mathrm dt}{\sqrt{1-t^2}}, \qquad 0 \leqslant u \leqslant 1.$$ Plugging identity above into ##(1)## with ##u...
Back
Top