Wednesday, 27 December 2017

metric spaces - Correctness of Analysis argument with Cauchy sequences



Let (xn) and (yn) be Cauchy sequences in (X,d). Show that (xn) and (yn) converge to the same limit iff d(xn,yn)0



Proof



Suppose (xn)a and (yn)a, then

ϵ>0N1s.t.n>N1d(xn,a)<ϵ2ϵ>0N2s.t.n>N2d(yn,a)<ϵ2



Given ϵ>0 take N=max then \forall_{n>N} we have:
\begin{align} d(x_n, y_n) &\leq d(x_n, a) + d(a, y_n) \\ & < \dfrac{\epsilon}{2} + \dfrac{\epsilon}{2} = \epsilon \end{align}




What is the cleanest way to finish this piece of the proof? Technically I feel like I could do the following:



Consider d(d(x_n, y_n), 0):
\begin{align} d(d(x_n, y_n), 0) &< d(\epsilon, 0) < \epsilon \implies d(x_n, y_n) \to 0 \end{align}



What can I do to make this direction of the proof cleaner? Note that I only used the fact that the sequences converged and nothing special about Cauchy sequences. Additionally, what is the formal definition of d(x_n, yn) \to 0?




ADDED



\leftarrow Suppose d(x_n, y_n) \to 0 that is \forall_{\epsilon > 0} \exists_{N_1} s.t \forall_{n > N_1} \implies |d(x_n, y_n) - 0| < \epsilon. Given \epsilon > 0 take N > N_1 then by the triangle inequality we have:



\begin{align} 0 < d(x_n, y_n) \leq d(x_n, a) + d(a, y_n) < \epsilon \end{align}
We know this a exists as (X, d) is a complete metric space and (x_n) and (y_n) are Cauchy. Thus we have that d(x_n, a) < \epsilon \land d(y_n, a) < \epsilon.
Therefore, (x_n)\to a \land (y_n)\to a




How does the other direction look?


Answer



Remember that your metric d is real valued, and the real numbers are a metric space. Thus, the formal definition of d(x_n, y_n) \rightarrow 0 is the same as for convergence of a sequence in a metric space, i.e., for every \epsilon > 0 there exists an N \in \mathbb{N} such that for all n \geq N we have d_{\mathbb{R}}(d(x_n,y_n), 0) < \epsilon.



Note the d_{\mathbb{R}}! This is the metric on the real numbers (where d(x_n,y_n) and 0 live). It is defined by d_{\mathbb{R}}(x,y) = |x -y|. So a much less obfuscated way to write the definition above is the following:



d(x_n, y_n) \rightarrow 0 if for every \epsilon > 0 there exists an N \in \mathbb{N} such that for all n \geq N we have |d(x_n,y_n) - 0| < \epsilon. But this is exactly what you have shown in the part before your proposed finish! You can actually stop right there with the \epsilon. You could also add a note about how you know d(x_n,y_n) \geq 0.



The error in what you go on to write is that d(d(x_n,y_n), 0) may not be defined since d is a metric on X while d(x_n, y_n) and 0 are in \mathbb{R}.




As for the Cauchy assumption, you need that to do the converse (along with completeness of X).


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}...