1
$\begingroup$

Let $D\subseteq \mathbb{E}^n$ and suppose that $\mathbf{f}:D\to \mathbb{E}^m$. Show that if $D$ is compact and if $\mathbf{f}$ is continuous on $D$ then $\mathbf{f}$ is bounded on $D$ (that is, $M=\sup\{\mathbf{f}(\mathbf{x}):\mathbf{x}\in D\}<\infty$ and there is a point $\mathbf{x_o}\in D$ such that $\mathbf{f}(\mathbf{x})=M$.

To prove that $\mathbf{f}$ is bounded, I used the Extreme Value Theorem which states if $f:D\to\mathbb{R}$ is continuous then $f$ achieves both a maximum value and a minimum value on $D$ (not sure if this also works for vector-valued functions $\mathbf{f}$ but this is what is provided to me in my book).

Since each component of $\mathbf{f} = (f_1,f_2,\cdots,f_m)$ is also continuous, I can use EVT to find maximum values, namely, $M_1,M_2, \cdots, M_m$ for each $f_1,f_2,\cdots,f_m$ respectively such that $f_i(\mathbf{x}) \leq M_i$ for all $\mathbf{x}\in D$.

Therefore, \begin{align} \|\mathbf{f}(\mathbf{x})\| &= \sqrt{f_1(\mathbf{x})^2 + \cdots + f_m(\mathbf{x})^2} \\ &\leq \sqrt{M_1^2 + \cdots + M_m^2} \\ &= M \end{align} and the function $\mathbf{f}$ must be bounded. But I am not sure if the $M$ that I found is $M=\sup\{\mathbf{f}(\mathbf{x}):\mathbf{x}\in D\}<\infty$.

How should I prove that there exists $\mathbf{x_0}$ such that $\mathbf{f}(\mathbf{x_0})=M$?

$\endgroup$
4
  • 1
    $\begingroup$ First of all, what is $E$? Second, how about "image of a compact set under a continuous function is compact" and "compact sets in $\Bbb R^n$ are bounded and closed"? $\endgroup$ May 12, 2017 at 21:05
  • $\begingroup$ @TZakrevskiy $\mathbb{E}$ is Euclidean space according to my book that has the norm operation. I think it's equivalent $\mathbb{R}$. I think that could prove all the questions but I want to prove that I can find some $\mathbf{x_0}$ such that $\mathbf{f}(\mathbf{x_0}) = M$ specifically. $\endgroup$ May 12, 2017 at 21:09
  • $\begingroup$ Your statement is wrong. Either you need $f$ is uniformly continuous or $D$ is compact. As for a proof, $x\mapsto \|f(x)\|$ is continuous. $\endgroup$
    – user251257
    May 12, 2017 at 21:41
  • $\begingroup$ The proof is flawed. $\Bbb{E}^m$ is not even an ordered set. Even if you meant $M=\sup\{|f(x)|\} $, what about each $f_i$ is negative? I would say Tzakrevskiy's approach is more simple. $\endgroup$ May 13, 2017 at 0:20

1 Answer 1

2
$\begingroup$

You can't expect for your $M$ to be the sup-norm of $\mathbf{f}$ (i.e. $\sup_{\mathbf{x} \in D} \lVert \mathbf{f}(\mathbf{x}) \rVert$). In fact, you can't expect to find any nice expression for the sup-norm of a function. The best you can do is an existence proof.

If $D$ is compact, then $\lVert \mathbf{f}(\mathbf{x}) \rVert$ achieves its maximum on $D$ since it is a continuous function on a compact domain. You can use the Extreme Value Theorem here to see this or just go through the proof of the Extreme Value Theorem again:

Let $\lVert \mathbf{f}(\mathbf{x}_1) \rVert, \lVert \mathbf{f}(\mathbf{x}_2) \rVert, \lVert \mathbf{f}(\mathbf{x}_3) \rVert, \ldots$ be a sequence that tends to $\sup_{\mathbf{x} \in D} \lVert \mathbf{f}(\mathbf{x}) \rVert$ (we don't assume this latter value is finite at the moment). Since $D$ is compact, there is a convergent subsequence $\mathbf{x}_{n_1}, \mathbf{x}_{n_2}, \mathbf{x}_{n_3} \to \mathbf{x}_0 \in D$. Then, by continuity and since a subsequence of a convergent sequence has the same limit,

$$ \lVert \mathbf{f}(\mathbf{x}_0) \rVert = \left\lVert \mathbf{f}\left( \lim_k \mathbf{x}_{n_k} \right) \right\rVert = \lim_k \lVert \mathbf{f}(\mathbf{x}_{n_k}) \rVert = \sup_{\mathbf{x} \in D} \lVert \mathbf{f}(\mathbf{x}) \rVert. $$

$\endgroup$
4
  • $\begingroup$ But does my proof show that the function is bounded and therefore there exists a sup-norm? $\endgroup$ May 13, 2017 at 16:50
  • $\begingroup$ @user3000482 Yes that's right. $\endgroup$ May 13, 2017 at 16:55
  • $\begingroup$ Can I say since $\sup_{\mathbf{x}\in D}\|\mathbf{f}(\mathbf{x})\|$ is like a cluster point, and since $\mathbf{f}$ is compact, it must contain the sup-norm in its domain, $\mathbf{f}(D)$? $\endgroup$ May 13, 2017 at 17:01
  • $\begingroup$ @user3000482 $\lVert \mathbf{f}(D) \rVert$ is compact by the EVT. Given this application of the EVT, we can say that $\sup_{x \in D} \lVert \mathbf{f}(\mathbf{x}) \rVert = \lVert \mathbf{f}(\mathbf{x}_0) \rVert$ for some $\mathbf{x}_0 \in D$. $\endgroup$ May 13, 2017 at 17:09

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .