Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

According to wikipedia a function $f\colon \mathbb{R}^n\to\mathbb{R}^n$ that is continuously, is also locally Lipschitz.

I there someone who knows a good reference which contains a proof of this statement?

share|improve this question
3  
Beware: Wikipedia reads "continuously differentiable", not "continuous differentiable in all directions" (which could be interpreted so as to make the statement false). –  D. Thomine May 10 '12 at 20:18
    
Now it just says "continuously" which makes even less sense. Please clarify. –  Nate Eldredge May 10 '12 at 21:22
add comment

2 Answers

The proof on $\mathbb{R}^n$ is fairly straightforward.

Choose some ball $B(\hat{x},\epsilon)$. The closure is compact, so the derivative $\frac{\partial f}{\partial x}$ is bounded by some $L$ on the ball. Now suppose $x,y \in B(\hat{x},\epsilon)$, then using Taylor's formula, we have: $$f(x)-f(y) = \int_o^1 \frac{\partial f (y+t(x-y))}{\partial x}(x-y)\;dt.$$ Hence we can get the bound: $$\|f(x)-f(y) \| \leq \int_o^1 \|\frac{\partial f (y+t(x-y))}{\partial x}\|\; \|(x-y)\| \;dt \leq L \|x-y\|.$$

share|improve this answer
add comment

The exact quote on wiki was:

In particular, any $C^1$ function is locally Lipschitz, as continuous functions on a locally compact space are locally bounded so its gradient is.

The logic here is we would like to show the gradient of a $C^1$-function is locally bounded on a locally compact space, thus to obtain the Lipschitz continuity. Hence your question boils down to how to prove:

A continuous function on a locally compact space is locally bounded.

While I am not one hundred percent sure which real analysis textbook has the exact result, you may refer to this Wikipedia entry, we could use Stone-Weierstrass theorem to prove: roughly speaking, since $\mathbb{R}^n$ is locally compact, for any point $x\in \mathbb{R}^n$, choose a compact neighborhood $S$ of $x$, we could use S-W theorem to prove that, for a continuous function on $S$, there exists a sequence of vector-valued polynomials $\{p_n\}$ that converge uniformly to the function of interest $f$ on $S$, and use the triangle inequality in the supreme norm, we could see $f$ is bounded on $S$. Thus the local boundedness of $f$ is proved.

To sum up: Gradient of $C^1$-function is continuous $\to$ A continuous function on a locally compact space(notice the gradient of $f$ you gave is the Jacobian matrix but the argument still applies) is locally bounded $\to$ Gradient of $C^1$-function is locally bounded $\to$ By mean value theorem $C^1$-function is locally Lipschitz.


EDIT: As Neal suggested, it suffices to argue as follows: For any $x_0\in \mathbb{R}^n$, on a compact set within some neighborhood $S \subset B(x_0,\epsilon)$, for $x,y\in S$: $$ \frac{\|f(x)-f(y)\|}{\|x-y\|} \leq \sup_{\xi\in S} \| Df\|<L(x_0) $$ again we rely on continuous functions are bounded on compact set, and the Lipschitz constant depends on $x_0$, this is the local Lipschitz property you need.

share|improve this answer
1  
You might add that being locally Lipschitz is just a slight weakening of differentiability: all you need is $|\frac{f(x)-f(y)}{x-y}|<C$ for all $x,y$ in small neighborhoods of each other. –  Neal May 10 '12 at 23:43
    
@Neal Thanks for the heads up, edited. –  Shuhao Cao May 11 '12 at 0:29
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.