1
$\begingroup$

Question:

Find an example such that $|x_{k+1}-x_k|\to 0$, but the sequence $x_k$is not Cauchy.

Any help would be appreciated.

$\endgroup$
4
  • 3
    $\begingroup$ Hint: What if $x_k$ represented the $k$th partial sum of a divergent series? Then $x_{k+1} - x_k$ would be the $(k+1)$th term of the sequence, so we are left to finding a convergent sequence whose partial sums diverge. $\endgroup$
    – user217285
    Oct 12, 2015 at 20:37
  • 1
    $\begingroup$ The sequence $x_k=\log k$ is the standard example. $\endgroup$
    – Crostul
    Oct 12, 2015 at 20:42
  • $\begingroup$ how does $\log(k+1)-\log k\to 0$? @Crostul $\endgroup$ Oct 12, 2015 at 20:44
  • 1
    $\begingroup$ @JennieDurham $\ln\frac{k+1}{k} = \ln(1+\frac{1}{k}) \sim \frac{1}{k}$. $\endgroup$
    – Clement C.
    Oct 12, 2015 at 20:46

3 Answers 3

7
$\begingroup$

The sequence of partial sums $(H_n)_{n\geqslant 1}$ of the harmonic series defined by $$ H_n = \sum_{k=1}^n \frac{1}{k} $$ verifies $H_{n+1} - H_n = 1/(n+1) \xrightarrow[n\to+\infty]{} 0$, but $$ H_{2n} - H_n = \frac{1}{n+1} + \dots + \frac{1}{2n} \geqslant \frac{n}{2n} = \frac{1}{2} $$ so $(H_n)$ is not a cauchy sequence.

$\endgroup$
2
$\begingroup$

Hint: Consider $a_n= 1 + \frac{1}{2} + \ldots + \frac{1}{n}$

Then $\{a_n\}$ is divergent, but for $p > 0$ we have $$a_{n+p} - a_n = \frac{1}{n+1} + \ldots + \frac{1}{n+p} \leq \frac{p}{n+1} \to 0$$

$\endgroup$
2
  • $\begingroup$ You proved something stronger than the required condition, by the way. You only needed to show that $|a_{k+1}-a_k|$ goes to $0$. $\endgroup$ Oct 12, 2015 at 20:50
  • $\begingroup$ Take $p =1$ and it's done. $\endgroup$ Oct 12, 2015 at 20:50
2
$\begingroup$

Hint: any sequence of that sort cannot be convergent, since any convergent sequence is Cauchy. Intuitively, any sequence diverging "slowly enough" to e.g. $\infty$ would thus do the trick.

Take for instance $x_n = \ln n$. $(x_n)$ is not Cauchy (it diverges; yet $\lvert x_{n+1}-x_n\rvert = \ln(1+\frac{1}{n}) \xrightarrow[n\to\infty]{} 0$.

$\endgroup$
1
  • $\begingroup$ Thanks, this example is far better than $H_n$, since the computation of $x_{2n} - x_n = \ln(2n) - \ln (n) = ln(2) + ln(n) - ln(2) = ln(2)$ provides an equality rather than an inequality. Plus it's easier to note that it diverges! $\endgroup$ Jan 11, 2020 at 22:31

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .