Tuesday 29 October 2013

real analysis - Show that certain lim sup additivity implies convergence.



I'm trying to prove the following:



Let $\{a_n\}$ be a bounded sequence. If for every bounded sequence $\{b_n\}$ the following holds:



$$1) \limsup_{n \to \infty} (a_n + b_n) = \limsup_{n \to \infty} a_n + \limsup_{n \to \infty} b_n$$



Show that $\{a_n\}$ is convergent.




Outline of Proof:



Here's the outline of my proof. But there's a point I get stuck below. I will be using the following theorem:



2) Let $\{a_n\}$ be a bounded sequence, let $L = \limsup_{n \to \infty} a_n$. If $\epsilon > 0$ there exists infinitely many $n$ such that $a_n > L - \epsilon$, and there exists $N_1 \in \mathbb{N}$ such that if $n \ge N_1$ then $a_n < L + \epsilon$



Since condition $1)$ holds for every bounded sequence $\{b_n\}$ we can assume it holds for a $\{b_n\}$ that's convergent. That means $\forall \epsilon > 0$ $\exists N_2 \in \mathbb{N}$ $\ni$ if $n \ge N_2$ then $L_b - \epsilon < b_n < L_b + \epsilon$ where $L_b = \lim_{n \to \infty} b_n$.



Let $\epsilon > 0$, $L_a = \limsup_{n \to \infty} a_n$, $L_b = \limsup_{n \to \infty} b_n$. Let N = $\max\{N_1, N_2\}$.




If $n \ge N$ we can easily show $a_n < L_a + \epsilon$.



This is the point where I'm stuck. For infinitely many $n$:



$$a_n + b_n > L_a + L_b - \epsilon/2$$
$$a_n > L_a + L_b - b_n - \epsilon/2$$



$\forall n \ge N$: $L_b - b_n > -\epsilon/2$.




But the only conclusion I can draw from the above is that for infinitely many $n$ (and not $\forall n \ge N$):



$$a_n > L_a - \epsilon$$


Answer



Prove that
$$
\limsup(-a_n) = -\liminf(a_n).
$$
Then with setting $b_n := -a_n$ you are done.


No comments:

Post a Comment

real analysis - How to find $lim_{hrightarrow 0}frac{sin(ha)}{h}$

How to find $\lim_{h\rightarrow 0}\frac{\sin(ha)}{h}$ without lhopital rule? I know when I use lhopital I easy get $$ \lim_{h\rightarrow 0}...