If $p>0$, then $ \lim_{n\to\infty}\frac{1}{n^p}=0$ using squeeze theorem for sequences.

4.5k Views Asked by At

If $p>0$, then $ \lim_{n\to\infty}\frac{1}{n^p}=0\;.$

Rudin suggests in his Principle of Mathematical Analysis to take $$n> (\frac{1}{\epsilon})^\frac{1}{p}$$ using the Archimedean property of the real number system.

This is under the assumption that we will compute the limit of the sequence based on the fact: If $\ 0 \leq x_n \leq s_n$ for $\ n \geq N$, where N is some fixed number, and if $\ s_n \rightarrow 0$, then $\ x_n \rightarrow 0.$

I don't really understand this proof, but I could try it a different way:

Letting $\ x_n = \frac{1}{n^p}$ and taking $\ s_n$ to be $\ \frac{1}{n}$, then we know that $\ x_n \leq s_n $ because $\ p>0 $. But since $ s_n = \frac{1}{n}$ goes to $0$ as $n$ approaches infinity, we know from $0 \leq x_n \leq s_n$ that $ x_n = \frac{1}{n^p}$ will also go to $0$ as $n$ approaches infinity.

Is this a valid approach?

2

There are 2 best solutions below

0
On

What you mentioned is Theorem 3.20 in Rudin's book (page 57).

If one takes $n>(1/\varepsilon)^{1/p}$, then it follows that $n^p>1/\varepsilon$ (because $p>0$) and thus $\displaystyle\frac{1}{n^p}<\varepsilon$. In particular, this argument shows that if one takes a positive integer $N>(1/\varepsilon)^{1/p}$, existence given by the archimedean property, then for any integer $n>N$, one has $$ n^p>N^p>1/\varepsilon, $$ equivalently, $$ \left|\frac{1}{n^p}-0\right|<\varepsilon\;. $$ This implies by definition of limits that $\displaystyle\lim_{n\to\infty}\frac{1}{n^p}=0$.


Rudin indeed says before the theorem that the squeeze theorem will be used:

enter image description here

But should not take his words too literally ("The proofs will all be ..."). If you take look at the (one line) proof of (e), all he says is "Take $\alpha=0$ in (d)", which is not quite using the squeeze theorem per se.

0
On

Is Rudin serious? Do you acutally prove

If $p>0$, then $ \lim_{n\to\infty}\frac{1}{n^p}=0\;.$

using the squeeze theorem?

Well, he might have been teasing the student a bit with that one. The most elementary/first principle thing we learn is that

$\tag 1 \displaystyle \lim_{n\to\infty}\frac{1}{n}=0$

and that it follows from the Archimedean property. So one 'heartbeat away' we demonstrate that

$\tag 2 \displaystyle \lim_{n\to\infty}\frac{1}{n^p}=0$

once again using the Archimedean property.

So how does Rudin justify bringing out the squeeze theorem here?

Well, just set $x_n = \frac{1}{n^p}$ and $s_n = \frac{1}{n^p}$ (you've seen in Rudin's proof that that $s_n$ converges to $0$).

I think Rudin would have made a great lawyer!