Sequel to the question I asked Prove that if f is continuous on $[a,\infty)$ and if $\lim\limits{x\to \infty}f(x)$ exists then $f$ is uniformly continuous and the reply I got, I want to show that $f$ is uniformly continuous on $\Bbb{R}$ if it is continuous on $\Bbb{R}$ and $\lim\limits_{x\to +\infty}f(x)$ and $\lim\limits_{x\to -\infty}f(x)$ exist

Let $\epsilon>0$ be given.

$\lim\limits_{x\to -\infty}f(x):=k$ exists at $N_\epsilon \in \Bbb{R} \implies\;\exists\,N_\epsilon<a,$ s.t. $\forall \;x< N_\epsilon,\;\;|f(x)-k|<\epsilon/2.$ The set $[N_\epsilon,a]\subset(-\infty,a] $ is a compact set and so, $f$ is uniformly continuous on $[N_\epsilon,a]$, i.e. $\exists\;\delta_1=\delta_1(\epsilon)>0\;$ s.t. $\forall\;x,y\in [N_\epsilon,a]$ with $|x-y|<\delta_1$, we have $|f(x)-f(y)|<\epsilon$.

$\lim\limits_{x\to \infty}f(x):=l$ exists at $M_\epsilon \in \Bbb{R} \implies\;\exists\,M_\epsilon>a,$ s.t. $\forall \;x> M_\epsilon,\;\;|f(x)-l|<\epsilon/2.$ The set $[a,M_\epsilon]\subset[a,\infty)$ is a compact set and so, $f$ is uniformly continuous on $[a,M_\epsilon]$, i.e. $\exists\;\delta_2=\delta_2(\epsilon)>0\;$ s.t. $\forall\;x,y\in [a,M_\epsilon]$ with $|x-y|<\delta_2$, we have $|f(x)-f(y)|<\epsilon$.

Since $f$ is continuous at $N_\epsilon \in \Bbb{R} \implies\;\exists\,\delta_3=\delta_3(N_\epsilon,\epsilon)>0$ s.t. $\forall \;x\in [N_\epsilon,a]$ with $|x-N_\epsilon|<\delta_3$ we have $|f(x)-f(N_\epsilon)|<\epsilon/2.$

Also $f$, continuous at $M_\epsilon \in \Bbb{R} \implies\;\exists\,\delta_4=\delta_4(M_\epsilon,\epsilon)>0$ s.t. $\forall \;x\in [a,M_\epsilon]$ with $|x-M_\epsilon|<\delta_4$ we have $|f(x)-f(M_\epsilon)|<\epsilon/2.$

Take $\delta=\min\{\delta_1,\delta_2,\delta_3,\delta_4\}$ s.t. $\forall\;x,y\in (-\infty,\infty)$ with $|x-y|<\delta$. WLOG, we take $x<y$.

Case 1: $x<N_\epsilon<a<y$

$|f(x)-f(y)|=|f(x)-k+k-f(y)|\leq |f(x)-k|+|f(y)-k|<\epsilon/2+\epsilon/2=\epsilon$

Case 2: $a<M_\epsilon<x<y$

$|f(x)-f(y)|=|f(x)-l+l-f(y)|\leq |f(x)-l|+|f(y)-l|<\epsilon/2+\epsilon/2=\epsilon$

Case 3: $N_\epsilon<x<y<a$

$|x-y|<\delta$, we have $|f(x)-f(y)|<\epsilon$

Case 4: $a<x<y<M_\epsilon$

$|x-y|<\delta$, we have $|f(x)-f(y)|<\epsilon$

Case 5: $N_\epsilon<x<a<y$

$|f(x)-f(y)|=|f(x)-f(N_\epsilon)+f(N_\epsilon)-f(y)|\leq |f(x)-f(N_\epsilon)|+|f(y)-f(N_\epsilon)|<\epsilon/2+\epsilon/2=\epsilon$, since $|x-N_\epsilon|<|x-y|<\delta$ as well as $|y-N_\epsilon|<|x-y|<\delta$

Case 6: $N_\epsilon<x<a<y$

$|f(x)-f(y)|=|f(x)-f(M_\epsilon)+f(M_\epsilon)-f(y)|\leq |f(x)-f(M_\epsilon)|+|f(y)-f(M_\epsilon)|<\epsilon/2+\epsilon/2=\epsilon,$ since $|x-M_\epsilon|<|x-y|<\delta$ as well as $|y-M_\epsilon|<|x-y|<\delta$

Hence, $f$ is uniformly continuous on $\Bbb{R}$.

Please, I'm I right? If no, can anyone please provide a detailed proof?


Your proof looks fine, but it is a little hard to read and consequently hard to find bugs in the proof.

It helps to give the reader an idea of what you are doing. A 'proof plan' as such. This is a subjective issue, and really depends on your audience. On advantage of a narrative is that even if you make a typographical mistake, it is clear where you are going, so it is 'self repairing' to some extent.

Here is how I would write it:

Let $\epsilon>0$. We need to find a $\delta>0$ so that if $|x-y| < \delta$ then $|f(x)-f(y)| < \epsilon$. Without any loss of generality, we can decide up front that $\delta \le 1$. This simplifies dealing with infinities.

On a (compact) set of the form $[-M,M]$, $f$ is uniformly continuous, so the idea is to choose $M$ so that $f$ varies by less than $\epsilon$ on $(M,\infty)$ and $(-\infty, -M)$. The only catch is when $x,y$ straddle $\pm M$, so some finessing is required.

Choose $M$ so that if $x,y \in [M,\infty)$ or $x,y \in (-\infty,-M]$, then $|f(x)-f(y)| < \epsilon$. This is possible because of the limit assumption on $f$.

Now partition $\mathbb{R}$ into $(-\infty, -(M+1)), [-(M+1),M+1], (M+1,\infty)$.

Note that if $|x-y| < 1$ then both $x,y$ must lie in at least one of the intervals $(-\infty, -M)$, $[-(M+1),M+1]$, or $(M,\infty)$. If $x,y$ lie in the first or last interval, then we have $|f(x)-f(y)| < \epsilon$.

Since $[-(M+1),M+1]$ is compact, we can find a $\delta_1>0$ so that $|f(x)-f(y)| < \epsilon$ if $x,y \in [-(M+1),M+1]$ and $|x-y| < \delta$.

Now if we choose $\delta = \min(1,\delta_1)$ we are finished.

  • $\begingroup$ I should add that one always makes mishtakes. $\endgroup$
    – copper.hat
    May 10 '18 at 17:58

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.