Monday, 10 February 2014

real analysis - Convergence almost uniformly to zero with certain conditions.



This problem is related to Convergence in measure to zero with certain conditions.



(Let {fn}nN be a sequence of measurable functions on a measure space and f measurable.




For cn>0 such that either lim, or c_n\geq c>0 for all n, and measurable sets E_n with m(E_n)>0 consider the sequence f_n(x):=c_n\mathcal{X}_{E_n}(x). )



For the same sequence(same assumptions!) \{f_n\}_{n\in\mathbb{N}}, f_n converges almost uniformly to zero, iff c_n\to 0 as n\to \infty or m(\cup_{n\geq N}E_n)\to 0 as N\to \infty.



My approach:
(\Rightarrow) By definition of almost uniformly, we have that for all \epsilon>0 there exists A_\epsilon such that m(A_\epsilon)<\epsilon such that f_n converges to uniformly to 0 on A_\epsilon^c. Observe that we have A_\epsilon \subset \cup_{n\geq N}E_n, so \epsilon\geq m(A_\epsilon)\leq m(\cup_{n\geq N}E_n). Now if x\in A_\epsilon then clearly m(\cup_{n\geq N}E_n)\to 0 as N\to \infty and if x\notin A_\epsilon then again clearly as f_n(x)=c_n\mathcal{X}_{E_n}(x)\to 0, c_n\to 0.



(\Leftarrow) I think to prove this direction is hard.




Any help, comments and suggestions will be appreciated.


Answer



as in your other question I think it is a good idea to suppose for (\Rightarrow) that we have c_n \not\to 0 (and therefore c_n \ge c > 0). I don't understand what you mean when you write "Now if x \in A_\epsilon then clearly m(\bigcup_{n\ge N} E_n) \to 0 as ...". How can you conclude from x \in A_\epsilon something about the measure of some set which doesn't depend on x. And why A_\epsilon \subseteq \bigcup_{n\ge N} E_n?



For (\Rightarrow): Suppose c_n \ge c > 0. We want to prove m(\bigcup_{n\ge N} E_n) \to 0. For \epsilon > 0 choose A_\epsilon as you did. Now f_n \to 0 uniformly on A_\epsilon^c. Now use again (as in the other task) that f_n \ge c\chi_{E_n}. So \chi_{E_n\cap A_\epsilon^c} \to 0 uniformly. Does this tell you something?



(\Leftarrow) is IMO more direct. If c_n \to 0 use |f_n| \le c_n to conclude uniform convergence, otherwise for some \epsilon you can find a N with m(\bigcup_{n\ge N} E_n) < \epsilon. Can you show convergence outside this set?



Hope this helps, otherwise feel free to ask more.


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}...