Weak convergence on uniformly convex Banach spaces

1k Views Asked by At

I want to prove the following:

Let X be a uniformly convex Banach space, $\{x_n\}_{n\in\mathbb{N}} \subset X, x \in X, \|x_n\| \to \|x\|$, $x_n$ converges weakly to $x$. Then $x_n$ also converges strongly to $x$.

Actually I have not really an idea how to do this.

Would be grateful for some tips!

S


I mean what I have is, that $ < f, x_n -x> \to 0$ $\forall f \in X^*$ and $\|x_n\| \to \|x\|$.

And $\forall \epsilon > 0 \exists \delta > 0 \forall x,y \in X, \|x\|,\|y\| \leq 1: \|x-y\|\geq\epsilon \Rightarrow \frac{\|x+y\|}{2} \leq 1-\delta $

or $\forall \epsilon > 0 \exists \delta > 0 \forall x,y \in X, \|x\|,\|y\| \leq 1:\frac{\|x+y\|}{2} > 1-\delta \Rightarrow \|x-y\| < \epsilon $.

And I want to show that $\|x_n-x\|\to 0$.

My first idea was that I want to have somehow $\|x_n-x\|<\epsilon$. So I need to fulfill the condition of the uniformly convexness?


EDIT - Possible Solution:

Let $\epsilon > 0$ arbitrary. As $X$ is a uniformly convex Banach space there is a $\delta > 0$ such that for every $x,y \in X$ with $\|x\|,\|y\|\leq 1: \frac{\|x+y\|}{2} > 1-\delta$ implies $\|x-y\|<\epsilon$.

Let $f\in X^*$ such that $f(x)=\langle f, x\rangle = \|x\|$ and $\|f\|_* = 1$.

Then we have $\frac{\|x_n+x\|}{2} = \frac{\|f\|_{*}\|x_n+x\|}{2} \geq \langle f,\frac{x_n+x}{2}\rangle \to \langle f,x\rangle = \|x\| = 1 > 1-\delta$.

So $\|x_n-x\| < \epsilon$.

As $\epsilon$ was arbitrary $\{x_n\}_{n\in\mathbb{N}}$ converges strongly to x.

But then I didn't use that $\|x_n\| \to \|x\|$.

1

There are 1 best solutions below

0
On

Assume that $X$ is uniformly convex, $x_n \overset{\text{weakly}}{\longrightarrow} x$ and $\| x_n \| \longrightarrow \| x\|$.

Now let $y_n = \frac{x_n}{\|x_n\|}$ and $y = \frac{x}{\|x\|}$. Then we have $\| y_n \| = 1$ and $\| y \| = 1$ and also $y_n \overset{\text{weakly}}{\longrightarrow} y$ and $\| y_n \| \longrightarrow 1$. Now, by Hahn-Banach, take $f \in X'$ with $\| f \| = 1$ and $|f(y)| = 1$. Then we get, due to weak convergence: $|f(y_n)| \longrightarrow 1$. Therefore, we obtain $\left|\frac{f(y_n) + f(y)}{2}\right| \longrightarrow 1$ and also $\left|\frac{f(y_n) + f(y)}{2}\right| \leq \| f \| \left\| \frac{y_n + y}{2} \right\| = \left\| \frac{y_n + y}{2} \right\| \leq 1$. By the sandwich principle, we have that $\left\| \frac{y_n + y}{2} \right\| \longrightarrow 1$. Now, uniform convexity of $X$ implies that $\| y_n - y \| \longrightarrow 0$, meaning $y_n \longrightarrow y$ and due to $\| x_n \| \longrightarrow \| x \| $, we obtain that $x_n \longrightarrow x$.