3
$\begingroup$

Let $\Omega$ be any non empty set and let $X$ be a Banach space over $\mathbb{C}$.

Let $F_b (\Omega,X)$ be a linear subspace of $F(\Omega, X)$ of all functions $f; \Omega \to X$ such that $$\{||f(x)|| : x \in \Omega\}$$ is bounded.

The vector space $F_b (\Omega , X)$ is a normed space with norm $$||f||_b =\sup \{||f(x)|| : x \in \Omega \}$$

I am trying to show $F_b (\Omega ,X)$ is a Banach space.

Proof:

Let $(f_n)$ be a cauchy sequence in $F_b (\Omega,X)$ and let $\epsilon >0$. Then there exists a $N \in \mathbb{N}$ such that $||f_n -f_m ||_b \leq \epsilon$ whenever $n ,m > N$.

For all $x \in \Omega$ we have $$||f_n(x)-f_m(x)|| \leq ||f_n -f_m||_b < \epsilon$$ whenever $n , m > N$.

Hence $(f_n (x))$ is a Cauchy sequence in $X$.

Since $X$ is complete we may define a function $f : \Omega \to X$ by $$f(x)=\lim_{n \to \infty} f_n (x)$$

I understand the proof up to this point but not any further.

The proof continues

Since $$||f_n(x)-f_m(x)|| < \epsilon$$ for all $x \in \Omega$ whenever $n,m > N$ taking the limit $ m \to \infty$ yields $$||f_n(x) -f_m(x)|| \leq \epsilon$$ provided $n > N$

The answer says we have taken the limit but it still yields $||f_n(x) -f_m(x)|| \leq \epsilon$?

Why do we take the limit $m \to \infty$? And why not $n \to \infty$?

Hence $$f(x)|| \leq \epsilon + ||f_n(x)|| \leq \epsilon + ||f_n||_b$$ provided $n >N$ for all $x \in \Omega$.

Where does $f(x)|| \leq \epsilon + ||f_n(x)|| \leq \epsilon + ||f_n||_b$ come from?

Thus $f$ is bounded and belongs to $F_b (\Omega,X)$ and $\lim_{n \to \infty} f_n =f$

I understand the definition of a Banach space and I understand that to show completeness you have to show every Cauchy sequence is convergent in the space.

So I understand what this proof is trying to do but it doesn't make sense.

What I am looking for is a clearer, simpler and a more well explained proof to the question.

$\endgroup$

2 Answers 2

2
$\begingroup$

To show completeness, take a Cauchy sequence $(f_n)_{n\in \mathbb{N}}\subseteq F_b(\Omega, X)$ and show, there exists $f\in F_b(\Omega, X)$ such that

$$ \lim_{n\rightarrow \infty} \Vert f - f_n \Vert_{b}=0.$$

As you already noted, one can construct a candidate for the limit of our sequence. Namely, we note that for fixed $x\in \Omega$ we have

$$ \Vert f_n(x) - f_m(x) \Vert_X \leq \sup_{y\in \Omega} \Vert f_n(x) - f_m(x) \Vert_X = \Vert f_n -f_m \Vert_b.$$

Hence, $(f_n(x))_{n\in \mathbb{N}}$ forms a Cauchy sequence in $X$ and so $f(x):=\lim_{n\rightarrow \infty} f_n(x)$ exists. This gives us our candidate.

First we need to check, whether $f\in F_b(\Omega, X)$. Indeed, let $x\in \Omega$, choose $n\in \mathbb{N}$ such that $\Vert f(x) - f_n(x) \Vert_X\leq 1$. Then

$$ \Vert f(x) \Vert_X \leq \Vert f(x)-f_n(x) \Vert_X + \Vert f_n(x) \Vert_X \leq 1+ \Vert f_n \Vert_b.$$

Hence,

$$ \Vert f \Vert_b \leq 1 + \sup_{n\in \mathbb{N}} \Vert f_n \Vert_b.$$

Show that the RHS is bounded (use the fact that $(f_n)_{n\in \mathbb{N}}$ is a Cauchy sequence). Now we are left to prove that $f_n \rightarrow f$. Let $\epsilon >0$. Fix $x\in \Omega$ and choose $N\in \mathbb{N}$ such that for all $n,m\geq N$ holds

$$ \Vert f_m - f_n \Vert_b \leq \frac{\epsilon}{2}.$$

Furthermore, choose $m_x\geq N$ such that

$$ \Vert f(x) -f_{m_x}(x) \Vert_X < \frac{\epsilon}{2}.$$

For all $n\geq N$ we have then

$$\Vert f(x) -f_n(x) \Vert_X \leq \Vert f(x) -f_{m_x}(x) \Vert_X + \Vert f_{m_x}(x) -f_n(x) \Vert_X \leq \frac{\epsilon}{2} + \Vert f_n - f_{m_x} \Vert_b < \epsilon. $$

Thus for all $n \geq N$ we have

$$ \Vert f -f_n \Vert_b < \epsilon. $$

This establishs $f_n \rightarrow f$ and hence the completeness of $F_b(\Omega, X)$.

$\endgroup$
4
  • $\begingroup$ where did you get $\leq 1$ from? $\endgroup$
    – Al jabra
    Apr 16, 2016 at 13:58
  • $\begingroup$ This is how I have choosen my $n$ above. It is possible, as $f_n(x) \rightarrow f(x)$. You may replace 1 by any different number you like in the argument. $\endgroup$ Apr 16, 2016 at 14:01
  • $\begingroup$ why do we have to show f is bounded? $\endgroup$
    – Al jabra
    Apr 19, 2016 at 18:30
  • $\begingroup$ Because the limit of our Cauchy sequence must lie in our space. $\endgroup$ Apr 19, 2016 at 19:46
0
$\begingroup$

You have the following typos:

  1. the display after "yields" should read $$ \|f_n(x)-f(x)\|<\epsilon $$ because you've taken the limit for $m\to\infty$ with $n$ fixed and you've used the continuity of the norm in $X$.

  2. You're missing a $\|$ bracket before $f(x)\|$ elsewhere.

Now using the inequality in 1. and the triangle inequality (in the form $\|a\|\geq\|a+b\|-\|b\|$) you get $$ \|f(x)\|<\epsilon+\|f_n(x)\|<\epsilon+\|f_n\|_b+\epsilon<B+\epsilon, $$ where $B=\sup_n\|f_n\|_b$ which is finite by your assumption that $(f_n)_{n\geq1}$ is Cauchy.

Noting that the last display yields an inequality that does not depend on $n$ (look at the two extremes) and \emph{for any $\epsilon>0$}, it follows that $$ \|f(x)\|\leq B \text{ for all } x\in\Omega, $$ and thus $\|f\|_b\leq B$, whence $f\in F_b(\Omega,X)$.

$\endgroup$

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.