$$\lim_{n \to \infty} b^n n ^{\alpha} = 0, \forall |b| < 1, \alpha > 0$$

I have proved it for cases when $\alpha \le 0$, but I am not sure where to go further. I have tried D'Alembert's test, but I believe that it does not work for $\alpha > 0$.

Can you give me any direction where to go ? At least, some nominal ideas.


Assume $\alpha > 0$, if you want to show that using the knowledge of series, consider the series $\sum_{n = 1}^\infty a_n$, where $a_n = b^n n^\alpha$.

Since $$\left|\frac{a_{n + 1}}{a_n}\right| = \left|\frac{b^{n + 1}(n + 1)^\alpha}{b^n n^\alpha}\right| = |b|\left(1 + \frac{1}{n}\right)^\alpha \to |b| \tag{$*$}$$ as $n \to \infty$. Since $|b| < 1$, the ratio test (or so-called D'Alembert test) states that this series converges. Therefore, the general term $a_n \to 0$ as $n \to \infty$.

Without using the convergence test of series, it's also straightforward to reach the conclusion. If $b = 0$, trivial. We thus assume $0 < |b| < 1$. By $(*)$ and the condition $|b| < 1$, you can see that for sufficiently large $n$, $|a_{n + 1}| < |a_n|$, so the sequence $\{|a_n|\}$ is decreasing eventually and bounded below by $0$, hence it converges to some limit $a \geq 0$, by the monotone bounded convergence theorem. Let $n \to \infty$ on both sides of $$|a_{n + 1}| = |b|\left(1 + \frac{1}{n}\right)^\alpha |a_n|,$$ we conclude that $$a = |b|a.$$ Since $0 < |b| < 1$, we must have $a = 0$.


An approach not using any kind of convergence criterion:

We may assume that $0<b<1$ (since, $b=0$ is trivial and convergence does not depend on sign of $b$).

Then we may write

$$ b=\frac{1}{1+r} $$

where $r>0$ is real. Thus it follows that

$$ b^n=\frac{1}{(1+r)^n}=\frac{1}{1+nr+\cdots+{n \choose k} r^k+\cdots+r^n} $$

We may assume that $n>\alpha+2$ (or just for sufficiently large $n$), then choosing $k=\lfloor{\alpha}\rfloor+1=\beta$ we have

$$ b^n=\frac{1}{1+nr+\cdots+{n \choose k} r^k+\cdots+r^n}\le \frac{1}{{n \choose \beta} r^{\beta}} = \frac{1}{\frac{n(n-1)\cdots (n-\beta+1)}{\beta !} r^{\beta}} $$

so that

$$ b^n n^{\alpha} \le n^{\alpha} \frac{1}{\frac{n(n-1)\cdots (n-\beta+1)}{\beta !} r^{\beta}} \to 0 $$

as $n\to\infty$


Because the limit of $b^n$ is $0$, since $|b| < 1$ and $n$ approaches infinity $$\lim_{n \to \infty} (b^n) (n^a) = \lim_{n \to \infty} (b^n) \cdot \lim_{n \to \infty}(n^a)$$ since limit of $b^n$ is $0$, the limit of the previous one is $0$.

  • $\begingroup$ The logic is wrong --- you can write the limit of product as product of limits only when BOTH limits exist. $\endgroup$ – Zhanxiong Dec 7 '15 at 6:24
  • $\begingroup$ that makes sense, thanks:) sorry about that $\endgroup$ – Chuxian Wang Dec 7 '15 at 14:31

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.