Consider a metric space $(M, {\rm d})$ and $y$ fixed in $M$.

I want to prove that the function $f$ defined by $f(x)\colon={\rm d}(x,y)$ is uniformly continuous.

So I know that if this function is uniformly continuous, then for all $\epsilon>0$, there exists a $\delta>0$ such that if the ${\rm d}(x_1,x_2)<\delta$, then this implies the ${\rm d}(f(x_1),f(x_2))<\epsilon$.

So for all epsilon greater than zero, there exists a delta greater than zero such that ${\rm d}(x_1,x_2)<\delta \implies |{\rm d}(x_1,y)-{\rm d}(x_2,y)|<\epsilon$.

I can't figure out how to manipulate this equation to solve for delta!! I tried using the triangle inequality but it gets me nowhere. Please help!


Let $\epsilon > 0$, choose $\delta = \epsilon$. Take $x_1,x_2 \in M$ such that ${\rm d}(x_1,x_2) < \delta$. Then: $$|f(x_1)-f(x_2)| = |{\rm d}(x_1,y)-{\rm d}(x_2,y)| \leq {\rm d}(x_1,x_2) < \delta = \epsilon.$$

Obs.: by the triangle inequality, we have that $${\rm d}(x_1,y)\leq {\rm d}(x_1,x_2)+{\rm d}(x_2, y)\implies {\rm d}(x_1,y)-{\rm d}(x_2,y)\leq {\rm d}(x_1,x_2).$$ Swapping $x_1$ and $x_2$ you get: $ {\rm d}(x_2,y)-{\rm d}(x_1,y)\leq {\rm d}(x_2,x_1) = {\rm d}(x_1,x_2).$ This way, the inequality $|{\rm d}(x_1,y)-{\rm d}(x_2,y)| \leq {\rm d}(x_1,x_2)$ follows from the definition of absolute value.

  • $\begingroup$ why can you swap x1 and x2? $\endgroup$ – marry Jan 16 '15 at 0:18
  • $\begingroup$ It is because what I've done earlier is valid for any points $x_1,x_2$. It was a quick way to saying that you can repeat the same argument. Like when books say "analogously, yadda yadda yadda". In details, what I omitted was: $${\rm d}(x_2,y)\leq {\rm d}(x_2,x_1)+{\rm d}(x_1, y)\implies {\rm d}(x_2,y)-{\rm d}(x_1,y)\leq {\rm d}(x_2,x_1).$$ Do compare with what I've done before. $\endgroup$ – Ivo Terek Jan 16 '15 at 0:20
  • $\begingroup$ i see, that makes sense. Thanks!! how did you come up with this so quickly? $\endgroup$ – marry Jan 16 '15 at 0:21
  • $\begingroup$ Practice dealing with these problems, I reckon. I think I'm a little biased, since I'm taking a summer course about metric spaces here (: $\endgroup$ – Ivo Terek Jan 16 '15 at 0:23
  • 1
    $\begingroup$ pouco haha i just noticed your profile says universidade de sao paulo:) $\endgroup$ – marry Jan 16 '15 at 0:57

You have to show that for fixed $\;m\in X$= the metric space, the function $\;f(x):=d(m,x)\;$ is unif. continuous, so you have to show that

$$\forall\,\epsilon>0\;\exists\,\delta >0\;\;s.t.\;\;d(x,y)<\delta\implies |f(x)-f(y)|=|d(m,x)-d(m,y)|<\epsilon$$

But all you have to do then is use the triangle inequality and some symmetry:

$$d(m,x)\le d(m,y)+d(y,x)\implies |d(m,x)-d(m,y)|\le d(x,y)$$

and take now $\;\delta=\epsilon$ .

  • 1
    $\begingroup$ It might confuse OP that you changed the fixed point to $m$, and used $y$ as variable, since he was using $y$ as a fixed point. I don't know. Just my opinion, though (: $\endgroup$ – Ivo Terek Jan 16 '15 at 0:16
  • 1
    $\begingroup$ Indeed so, @IvoTerek . Yet I began answering this question before you edited it and didn't even see he did so. I think though that someone dealing with this must already be able to overcome that. Thank you. $\endgroup$ – Timbuc Jan 16 '15 at 0:20
  • $\begingroup$ thanks guys this very helpful $\endgroup$ – marry Jan 16 '15 at 0:22

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.