Let $f:[a,\infty)\to\mathbb{R}$ be a continuous function.

For every $\varepsilon>0$ there exist $0<\delta_{\varepsilon}$ and $a<c_{\varepsilon}\in\mathbb{R}$ so that for every $x_{1},x_{2}>c_{\varepsilon}$ so that $|x_{1}-x_{2}|<\delta_{\varepsilon}$ it holds that $|f(x_{1})-f(x_{2})|<\varepsilon$.

Prove that $f$ is uniformly continuous in $[a,\infty]$. Hint: $[a,\infty]=[a,c_{\varepsilon/2}]\cup[c_{\varepsilon/2},\infty)$ and use the Heine-Cantor Theorem.

I get that the Heine-Cantor Theorem proves uniform continuity in $[a,[c_{\varepsilon/2}]$, but don't get how we get to that epsilon and the general idea of the proof. Could you give me a more obvious hint or where to start?

  • 1
    $\begingroup$ You need to add the condition $\lim_{\infty}f$ exists. $\endgroup$ – hamam_Abdallah May 24 '20 at 15:42

The definition that a function $f(x)$ is uniformly continuous on $[a,\infty)$ is that for any real number $\epsilon>0$, there exist a real number $\delta >0$ such that if the distance between two points in the domain is close, $|x_i-x_j|<\delta $, we have that the distance between the evaluation of the function at those two points is close ,$|f(x_i)-f(x_j)|<\epsilon$.

So given $\frac{\epsilon}{2} >0$, from the condition given to us in the question, we know that there exist $c_{\frac{\epsilon}{2}} \in (a, \infty)$ and $\delta_{\frac{\epsilon}{2}}>0$ such that if $|x_k-x_j|<\delta_{\frac{\epsilon}{2}}, x_k,x_j >c_{\frac{\epsilon}{2}}$, then $|f(x_k)-f(x_j)|<\frac{\epsilon}{2}$.

Also, on the interval $[a,c_{\frac{\epsilon}{2}}]$, we know that the function is uniformly continuous, so we have that there is an $\delta'_{\frac{\epsilon}{2}}>0$ such that if $|x_i-x_j|<\delta'_{\frac{\epsilon}{2}}, x_i,x_j \in [a,c_{\frac{\epsilon}{2}}]$, then $|f(x_i)-f(x_j)|<\frac{\epsilon}{2}$.

Now if we choose $\delta := \min(\delta'_{\frac{\epsilon}{2}},\delta_{\frac{\epsilon}{2}})$ and now we are ready to verify the condition for the function $f(X)$ to be uniformly continuous on the interval $[a, \infty)$

So we choose two arbitary numbers in the domain of the function such that their distance is less less than $\delta$, like: $x_s<x_d \in [a,\infty), |x_s-x_d|<\delta$, We have three cases for the locations of the two numbers:

if $x_s,x_d \in [a,c_{\frac{\epsilon}{2}}]$, then $|f(x_s)-f(x_d)|<\frac{\epsilon}{2}<\epsilon$

if $x_s,x_d \in [c_{\frac{\epsilon}{2}},\infty)$, then $|f(x_s)-f(x_d)|<\frac{\epsilon}{2}<\epsilon$

Now if could be that the two numbers where one belongs to $[a,c_{\frac{\epsilon}{2}}]$ and the other belongs to $[c_{\frac{\epsilon}{2}},\infty)$, but since we assume that $x_s<x_d$, we have that:

$x_s \in [a,c_{\frac{\epsilon}{2}}], x_d \in [c_{\frac{\epsilon}{2}},\infty)$, and then $|f(x_s)-f(x_d)| = |f(x_s)-f(c_{\frac{\epsilon}{2}})+f(c_{\frac{\epsilon}{2}})-f(x_d)| \leq |f(x_s)-f(c_{\frac{\epsilon}{2}})|+|f(c_{\frac{\epsilon}{2}})-f(x_d)|<\frac{\epsilon}{2}+\frac{\epsilon}{2}=\epsilon$.

Hence, in all cases for possible location that the two numbers $x_s,x_d$ that could be, we have shown the definition for uniformly continuous is satisfied.

Hence the function is uniformly continuous on the interval $[a,\infty)$.


$f: x\mapsto x^2$ is continuous at $[0,+\infty)$ but is not uniformly continuous at $ [0,+\infty)$, since

with $$u_n=n \;\ \text{ and } \; v_n=n+\frac 1n$$

we have

$$\lim_{n\to +\infty}(v_n-u_n)=0$$



$$\lim_{n\to+\infty}(2+\frac{1}{n^2})=2\ne 0$$


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.