Alternative Proof of Cauchy Sequence ##\left(S_n\right) = \frac{1}{n}##

Bachelier
Messages
375
Reaction score
0
I am looking for a different proof that ##(S_n) = \frac{1}{n}## is cauchy.

The regular proof goes like this (concisely):

##\left|\frac{1}{n} - \frac{1}{m} \right| \leqslant \left|\frac{m}{nm}\right| \ (etc...) \ <\epsilon ##

but I was thinking about an alternative proof. Is my proof correct:

let ##\epsilon > 0## by Archimedian property ##\exists N \ s.t. \frac{1}{N}<\epsilon##

This is equivalent to ##\frac{1}{N}<\frac{\epsilon}{2}## "may be ommited"

Now ##\forall n, m \geqslant N## we have by ##\Delta## ineq.

##\left|\frac{1}{n} - \frac{1}{m} \right| \leqslant \left|\frac{1}{n} \right| + \left|\frac{1}{m} \right|\leqslant \frac{1}{N}+\frac{1}{N} < ε##

What do you guys think? Thanks...
 
Physics news on Phys.org
I think it's perfectly ok.
 
Bachelier said:
This is equivalent to ##\frac{1}{N}<\frac{\epsilon}{2}## "may be ommited"
I'm not sure what this means, but the rest is fine.

Another way to prove it is to note that ##(s_n)## converges to ##0## (easy proof), and use the fact that any convergent sequence is Cauchy.

Proof: if ##(x_n)## is a sequence which converges to some number ##L##, then given ##\epsilon > 0##, there is some ##N## for which ##|x_n - L| < \epsilon / 2## for all ##n \geq N##. Therefore, if ##m \geq N## and ##n \geq N##, then ##|x_n - x_m| = |x_n - L + L - x_m| \leq |x_n - L| + |x_m - L| \leq \epsilon##.
 
jbunniii said:
I'm not sure what this means, but the rest is fine.

{This is equivalent to ##\frac{1}{N}<\frac{\epsilon}{2}## "may be ommited"}

Well since ε in this case can be so small, then we can use a larger N > 1/ε (actually twice larger) to end up with ε/2+ε/2 in the end instead of 2ε.
 
Back
Top