2
$\begingroup$

Prove that the real line equipped with the distance $p(x,y)= |\arctan (x)-\arctan (y)|$ is an incomplete metric space.

My attempt

Duplicate but I want to ask about my specific proof and also how to prove that it is a metric space

A metric space is called complete if every cauchy sequence of points in $M$ converges in $M$.

A sequence $x=(x_1,x_2,...x_n)$ is cauchy if for all $\epsilon>0$ there exists an $N \in \mathbb{N}$ such that $m,n \geq N$ implies $d(x_n,x_m)< \epsilon$.

The $\arctan$ function is well-defined from $-\pi/2$ to $\pi/2$.

Fix $\epsilon >0$. Let $X_n=n$. $$d(n,\pi/2)<\epsilon$$

$$|\arctan(n)-\arctan(\pi/2)| < \epsilon$$

Not sure where to go from here. I think I also need to prove that this is a metric

$\endgroup$
5
  • 3
    $\begingroup$ Is the sequence $n \mapsto n$ Cauchy? Does it converge? $\endgroup$ – copper.hat May 1 '17 at 1:55
  • $\begingroup$ The domain of $\arctan$ is all of $\mathbb{R}$. Its range is $\left(-\frac{\pi}{2},\frac{\pi}{2}\right)$. $\endgroup$ – kccu May 1 '17 at 1:55
  • $\begingroup$ Note that $\arctan(x)-\arctan(y) = \arctan{\frac{x-y}{1+xy}}$. $\endgroup$ – marty cohen May 1 '17 at 4:49
  • $\begingroup$ Ah, yes, this problem comes from the most recent Putnam exam, doesn't it? $\endgroup$ – Feryll May 1 '17 at 8:03
  • $\begingroup$ $\arctan(x)$ is well-defined on all of $\mathbb{R}$, with values in $(-\frac{\pi}{2}, \frac{\pi}{2})$. The precise function is irrelevant for proving it to be a metric, it only needs to 1-1. $\endgroup$ – Henno Brandsma May 1 '17 at 8:50
4
$\begingroup$

(I). If $f:\mathbb R\to \mathbb R$ is injective then $d(x,y)=|f(x)-f(y)|$ is a metric on the set $\mathbb R:$

(i). $x\ne y\implies d(x,y)=|f(x)-f(y)|\ne 0$ because $f$ is $1$-to-$1$.

(ii). $d(x,y)=d(y,x).$ Obvious.

(iii). $d(x,z)=|f(x)-f(z)|=|(f(x)-f(y))+(f(y)-f(z)|\leq$ $\leq |f(x)-f(y)|+|f(y)-f(z)|=d(x,y)+d(y,z).$

The derivative of $\arctan x$ is $1/(1+x^2) ,$ which is always positive. So $\arctan x$ is strictly increasing, and therefore is $1$-to-$1.$

(II). We have $\lim_{n\to \infty}\arctan n=\pi /2,$ so $(\arctan n)_n$ is a Cauchy sequence, that is, $$\lim_{n\to \infty}\sup_{a>b>n}|\arctan a-\arctan b|=0.$$ So with respect to the metric $d(x,y)=|\arctan x-\arctan y|,$ the sequence $(n)_n$ is a $d$-Cauchy sequence.

But there is no $z\in \mathbb R$ such that $\lim_{n\to \infty}d(n,z)=0$, as this would require $\lim_{n\to \infty}|\arctan (n)-\arctan (z)|=0,$ which implies $\arctan z=\lim_{n\to \infty}\arctan n=\pi /2,$ and there is no such $z\in \mathbb R.$

$\endgroup$

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.