Prove that the limit of the sequences corresponding to the image of two sequences converging to the same point is the same.

39 Views Asked by At

Context : I consider $(E, d_E)$ and $(F, d_f)$ two metric spaces and $A\subset E$ a dense subset of $E$ (i.e $\bar{A}=E$). The function $f$ is defined only on $A$.

I would like to prove that if I have two sequences $x_n$ and $v_m$ in $A$ that converges to the same point $x$ (not in A but in E), then the sequences $f(x_n)$ and $f(v_m)$ have the same limit provided that $f$ is uniformly continuous and that the space of arrival $(F, d_F)$ is complete.

Here is my attempt :

Consider $(x_n)$ and $(v_m)$ two sequences converging to the same point $x$. We have that

  • $\exists N_1 : n>N_1\implies d(x_n, x)\leq \delta/2$
  • $\exists N_2 : m>N_2\implies d(v_m, x)\leq \delta/2$

Thus for all $n,m>max(N_1, N_2)$ we have

$$ d(x_n, v_m)\leq d(x_n, x) + d(v_m, x)\leq\delta $$

Using the uniform continuity of $f$, we know that for $\varepsilon/3$ there exists a $\delta$ such that for all $n,m> max(N_1, N_2)$

$$ d(x_n, v_m)\leq\delta \implies d(f(x_n), f(v_m))\leq\varepsilon/3 $$

Now, we denote $y$ and $y*$ respectively the limit of $f(x_n)$ and $f(v_m)$ (the limit is well defined since the space of arrival is complete) we remark that :

  • $\exists N_3 : n>N_3\implies d(f(x_n), y)\leq\varepsilon/3$
  • $\exists N_4 : m>N_4\implies d(f(v_m), y*)\leq\varepsilon/3$

Thus, for all $\varepsilon>0$, we can find a $N=max(N_1, N_2, N_3, N_4)$ such that for all $n,m>N$ we have

$$ d(y,y*)\leq d(y, f(x_n)) + d(f(x_n),f(v_m)) + d(f(v_m), y*)\leq \varepsilon/3 + \varepsilon/3 + \varepsilon/3 = \varepsilon $$

wich shows that $d(y,y*) = 0$ so $y=y*$.

I would like to know if the proof seems correct, my intuition was to make a link between the two sequences using the uniform continuity and I tried to make this argument valid so please don't hesitate to correct me or add your proof !


EDIT

I missed some important information about the context I am very sorry.

2

There are 2 best solutions below

2
On BEST ANSWER

This new problem is equivalent to say that you can extend the function $f$ to a new continuous function $tilde f: E \rightarrow F$. Indeed if $f:A \rightarrow F$, one might choose to extend it by $\tilde f: E \rightarrow F$, setting $\displaystyle \tilde f(x):= \lim_{\substack{a \rightarrow x \\ a \in A}} f(a)$. But the well-definedness of $\tilde f$ requires to show if $a_n \rightarrow x$, $b_n \rightarrow x$, and $a_n, b_n \in A$, then $\lim f(x_n) = \lim f(v_n)$ which is exactly your problem.

Note that $A$ being dense is important in the definition since you should be able to converge to $x \in E$ by a sequence $a_n \in A$ in the first place.


Take $y_n:= f(a_n)$ and $z_n:= f(b_n)$. As $F$ is complete, cauchy sequences will converge. So let us show $y_n$ and $z_n$ are cauchy sequences. Let $\epsilon > 0$. By $\color{red}{\text{uniform continuity}}$, there is some $\delta > 0$ so that if $d(p, q) < \delta$, $d(f(p), f(q)) < \epsilon$. As $a_n \rightarrow x$, there is $N_1 \in \mathbb{N}$ so that for all $m, n \ge n$, $d(a_{\color{red}{m}}, a_{\color{red}{n}}) < \delta$. (Note that convergent sequences are Cauchy sequences.) Therefore $d(f(a_m), f(a_n)) < \epsilon$.

This shows $y_n$ is Cauchy. Similarly you can prove $z_n$ is Cauchy. Therefore there is $y \in F$ so that $y_n \rightarrow y$ and $z \in F$ so that $z_n \rightarrow z$.

Finally, lets show that $y = z$. This time we use the fact that $a_n$ and $b_n$ are both converging to the same point $x$. Therefore there is $M \in \mathbb{N}$ so that for all $n \ge M$, $a_n, b_n \in B_{\delta/2}(x)$ and therefore $d(a_n, b_n) < \delta$. Again, by uniform continuity, $d(f(a_n), f(b_n)) < \epsilon$ for all $n \ge M$.

To summarize, $d(y_n , z_n) < \epsilon$ for all $n \ge M$. Apply the triangle inequality to this, the fact that $y_n \rightarrow y$ and $z_n\rightarrow z$ to conclude that $d(y, z) < \epsilon$. As $\epsilon$ is arbitrary, you get the result.

1
On

You only need continuity. As $f$ is continuous and $x_n \rightarrow x$, $f(x_n)\rightarrow f(x)$. Similarly as $v_n \rightarrow x$, $f(v_m)\rightarrow f(x)$.


We used this fact: "If $x_n\rightarrow x$ and $f$ is continuous, then $f(x_n) \rightarrow f(x)$."

Let $\epsilon > 0$. By continuoity, there is a $\delta$ so that if $a$ is a point with $d(a, x) < \delta$, then $d(f(a), f(x)) < \epsilon$.

Use this $\delta$ in the definition of convergence. There is some natural number $N$ so that for any $n \ge N$, $d(x_n, x) < \delta$. So, by what we have just said $d(f(x_n), f(x)) < \epsilon$.

Therefore, for $\epsilon > 0$ there is some $N \in \mathbb{N}$ so that for all $n \ge N$, $d(f(x_n), f(x)) < \epsilon$. This means the sequence $\{f(x_n)\}$ converges to $f(x)$.