Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Given the sequences $a_{n},b_{n},c_{n}$ all subsets of the interval $(0,1)$, and all converges to 0. Also we have $a_{n}\leq b_{n}^{\alpha}$, and $a_{n}\leq c_{n}$, for all $n=1,2,3,...$, for some $\alpha \geq 1$ (real number).

Now, for a given $n$, we could have $$a_{n}\leq b_{n}^{\alpha} \leq c_{n}$$ or we could have $$a_{n} \leq c_{n} \leq b_{n}^{\alpha}$$

Suppose that we have the choice to pick any $\alpha\geq 1$. Is it possible to pick an $\alpha$ large enough so that the first case is true for all $n$ (or at least for a subsequence of $n$)?.

Edit: "for some $\alpha$" instead of for all.

I'm trying the following:

If there is $n_{1}, n_{2}, n_{3},...$ such that $b_{n_{i}}^{\alpha} > c_{n_{i}} $, for $i=1,2,3,.. $ ,then we can choose $\alpha_{1}, \alpha_{2}, \alpha_{3},..$ so that $ b_{n_{i}}^{\alpha_{i}} \leq c_{n_{i}}$ for all $i=1,2,3,.. $.

Now, let $ \alpha=\sup \alpha_{i}$ then $ b_{n_{i}}^{\alpha} \leq c_{n_{i}}$ for all $i=1,2,3,.. $. But the problem in this method is that the "sup" could be $\infty$!!

share|cite|improve this question
Wait... "for all $\alpha \ge 1$"? Surely that means that $a_n = 0$ for all $n$, since $0 < b_n < 1$? – huon Jun 10 '12 at 3:17
I meant we have the choice on $\alpha$. – Alpha Jun 10 '12 at 3:20

No, not for arbitrary sequences $a_n,b_n,c_n$.

Consider the case when $a_n = c_n = A 2^{-n}$ for a constant $A$ (to allow the condition $a_n \le b_n^\alpha$ to hold) and $b_n = \frac{1}{n+1}$. For the first case to hold, we must have an $\alpha$ such that $$ \left(n+1\right)^{-\alpha} = A 2^{-n}$$ for all $n$, which is impossible.

If the inequalities are strict, so $a_n < c_n$, then the above proof can be adjusted by choosing $a_n = A 2^{-n}$ and $c_n = (A + \epsilon) 2^{-n}$ for some small $\epsilon > 0$ and $b_n$ as before, then we need to find an $\alpha$ such that $$ A 2^{-n} \le \left(n+1\right)^{-\alpha} \le \left(A+\epsilon\right) 2^{-n}$$ for all $n$, which is again impossible.

share|cite|improve this answer
That's right. What if $a_{n}< c_{n}$ (strictly less than)? – Alpha Jun 10 '12 at 8:14
@Alpha, I've added that to my answer. (If this answer is useful, you should vote it up, this section of the FAQ has more information about this process towards then end :) ) – huon Jun 10 '12 at 8:30
Okay! Let me ask another question: In general, for any $a,b,c\in (0,1)$, if $a<c<b$, can we find an $\alpha$ such that $a<b^{\alpha}<c$? – Alpha Jun 10 '12 at 8:53
Yes. $\alpha = \log_b \frac{a+c}{2}$. – huon Jun 10 '12 at 9:06

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.