I am trying to do the following question. Let $E$ be an open subset of $\mathbb{R}^{n}$ containing the origin. Use the fact that if $f \in C^{1}(E)$ then for all $x,y \in N_{\delta}(0) \subset E$ (where $N_{\delta}(0)$ is a $\delta$-neighbourhood of the origin) there exists a $\xi \in N_{\delta}(0)$ such that

$$|f(x) - f(y)| \leq|| DF(\xi)|| |x-y|$$

to prove that if $f(0) = Df(0)=0$ then given any $\epsilon>0$ there exists a $\delta>0$ such that for all $x,y \in N_{\delta}(0)$ we have

$$|f(x) - f(y)| < \epsilon |x-y|.$$

Now the first inequality is obviously the so-called Mean Value Inequality (or Theorem). I don't know how to use it to prove the required statement. I could obviously take $\epsilon = \mbox{max} || DF(\xi)||$ but that doesn't help to establish a relation between $\delta$ and $\epsilon$.

Any help would be appreciated!


Hint: Let $g(t) = f(x + t(y-x)), t\in [0,1].$ Apply the mean value theorem to $g(1) - g(0).$

  • $\begingroup$ Yes I am aware of that trick - that's the standard proof of the Mean Value inequality. My question is, how to use first inequality to prove the second inequality (with the epsilon)? @zhw $\endgroup$ – Alex Nov 20 '16 at 19:53
  • $\begingroup$ You have $\nabla f (0) = 0.$ Choose $\delta > 0$ such that $|x|<\delta$ implies $|\nabla f (x)| < \epsilon.$ $\endgroup$ – zhw. Nov 20 '16 at 20:12
  • $\begingroup$ I don't see how $|x|< \delta$ implies $|Df(x)|< \epsilon$ - what do we take $\epsilon$ to be? Heuristically, the given condition implies that $f$ and its first derivative cannot be very large near 0. How to justify this rigorously? What is the bound on $\delta$ for a given $\epsilon$? @zhw $\endgroup$ – Alex Nov 20 '16 at 20:19
  • $\begingroup$ Hmm, I can't really tell what's bothering you. There's nothing tricky going on here. We have $f\in C^1$ which implies all partial derivatives of $f$ are continuous. Thus $|\nabla f(x)|=\|Df(x)\|$ is a continuous function. And we have $|\nabla f(0)|=0.$ Thus given $\epsilon>0$ there exists $\delta > 0$ … $\endgroup$ – zhw. Nov 20 '16 at 20:26
  • $\begingroup$ Ah yeah, of course. Got my wires crossed for a bit. Thanks! @zhw $\endgroup$ – Alex Nov 20 '16 at 20:31

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.