If $\{F_n\}$ is a decreasing sequence of compact subsets of $X$ Hausdorff and $f$ is continuous, then $f(\cap F_n) = \cap f(F_n)$

compactnessgeneral-topologysolution-verification

I have to prove the following statement:

Let $X$ and $Y$ be Hausdorff Spaces and $f:X\to Y$ a continuous function. If $\{F_n\}$ is a decreasing sequence of compact subsets of $X$, then $f(\cap F_n) = \cap f(F_n)$.

I already proved it, however, my proof doesn't use the hypothesis that $Y$ is Hausdorff, and that makes me think it might be wrong, but I don't find any mistakes. I would be really thankful if someone could verify it.

Proof. First we see that $f(\cap F_n) \subseteq \cap f(F_n)$. If $x\in f(\cap F_n)$, there must exists $a\in \cap F_n$ such that $f(a)=x$. It is clear that $a\in F_n$, for all $n\in\mathbb{N}$ and in consequence $x\in f(F_n)$, for all $n\in\mathbb{N}$, which implies $x\in\cap f(F_n)$.

Now we are going to prove that $\cap f( F_n) \subseteq f(\cap F_n)$. Let $x\in\cap f( F_n)$, this implies that $x\in f( F_n)$ for all $n\in\mathbb{N}$. From here we get for each $n\in\mathbb{N}$ there exists $y_n\in F_n$ such that $f(y_n)=x$. Since $\{F_n\}$ is a decreasing sequence, we have that $F_n \subseteq F_1$ for all $n\in\mathbb{N}$ and this implies that $y_n\in F_1$ for all $n\in\mathbb{N}$. This way, $\{y_n\}$ is a sequence in $F_1$. Because $F_1$ is compact, $\{y_n\}$ must have a convergent subsequence. Lets say this subsequence is $\{y_{n_k}\}$, and that it converges to $y$. Futhermore, since $X$ is Hausdorff, this limit is unique.

We see that $y\in\cap F_n$. Let $i\in\mathbb{N}$. There exists a $j\in\mathbb{N}$ such that $i<n_j$. Now lets consider the sequence $\{a_k^j\}$ given by $a_k^j=y_{n_{k+j}}$. It is clear that $a_k^j=y_{n_{k+j}}\in F_{n_{k+j}}\subseteq F_{n_{j}}\subseteq F_i$, which implies that the sequence $\{a_k^j\}$ is a sequence in $F_i$. Also, this sequence is just a shift of $\{y_{n_{k}}\}$, and because of this it must also converge to $y$. Because $F_i$ is compact, and $X$ is Hausdorff, $F_i$ is closed, and since $\{a_k^j\}$ is in $F_i$, we have that $y\in F_i$. As $i\in\mathbb{N}$ is arbitrary, we get that $y\in\cap F_n$.

Finally, using the continuity of $f$ and the uniqueness of the limit of $\{y_{n_k}\}$ we get that $f(y)=f(\text{lim}_{k\to\infty}y_{n_k})=\text{lim}_{k\to\infty}f(y_{n_k})=\text{lim}_{k\to\infty}x=x$, from where we get that $x=f(y)\in f(\cap F_n)$.

In conclusion, $f(\cap F_n) = \cap f(F_n)$

Best Answer

For the non-trivial backward direction I'd do a sequence-free proof because not all compact Hausdorff spaces are sequentially compact.

So let $y \in \bigcap_n f[F_n]$. And define $C_n = f^{-1}[\{y\}] \cap F_n$ which is a non-empty closed subset of $X$ (for every $n$, some $x \in F_n$ maps onto $y$ as $y \in f[F_n]$, closed as singletons are closed in $Y$ and $f$ is continuous.)

As the $F_n$ are decreasing so are the $C_n$ ($x \in C_{n+1}$ then $f(x)=y$ and $x \in F_{n+1}$, so $x \in F_n$ and still $f(x)=y$ so $x \in C_n$).

So the $C_n$ form a family of closed subsets with the FIP inside the compact set $F_1$, so there is a point $x \in \bigcap_n C_n$.

Clearly $x \in \bigcap_n F_n$ too and for any $n$, $x \in C_n$ so that $f(x)=y$ and so $x$ witnesses that $y \in f[\bigcap_n F_n]$ as required.

So it seems we can get ayway with $Y$ being merely $T_1$ instead of Hausdorff.

Related Question