Can I define a measurable function using base of a topology?

93 Views Asked by At

We know the more general definition of a measurable funnction: let $(X,\mathcal{X})$ and $(Y,\tau)$ a measurable and topological space, respectively. The function $f:X\to Y$ is a $\mathcal{X}$-measurable function if $f^{-1}(V) \in \mathcal{X}$, $\forall V \in \tau$.

I would like to know if it is true that we can define the measurability of the function $f$ using a base of the topology $\tau$. That is, suposse that $\beta$ is a base of $\tau$. Then, I wold like to prove that

$$f \hbox{ is } \mathcal{X}\hbox{-measurable function} \Leftrightarrow f^{-1}(B) \in \mathcal{X},\quad \forall B \in \beta $$

We know that any $V \in \tau$ can be written as

$$V = \bigcup_{A \in \mathcal{g}} A,\quad \mathcal{g} \subset \beta$$

with $g$ not necessarily enumerable. One implication of my goal is trivial. Any element of $\beta$ is open, that is, it is a element of $\tau$. Then, if $f$ is $\mathcal{X}$-measurable function, then $f^{-1}(B) \in \mathcal{X}, \forall B \in \beta$. But, because $g \in \beta$ is not necessarily enumerable, I can not conclude the other implication. That is, I can not conclude that

$$f^{-1}(V) = \bigcup_{A \in \mathcal{g}} f^{-1}(A) \in \mathcal{X}$$

I would like to know if it is possible to circumvent this or in what cases this result is valid.

1

There are 1 best solutions below

0
On BEST ANSWER

This is too long for the comment section.

You are probably after the following result. On any set $\mathrm{Z},$ denote by $\sigma(\mathscr{A})$ the minimal sigma algebra containing the subset $\mathscr{A}$ of the power set of $\mathrm{Z}$. If $\varphi$ is a function with values in $\mathrm{Z},$ we also denote by $\varphi(\mathscr{A})$ the set of $\varphi^{-1}(\mathrm{A})$ as $\mathrm{A}$ runs through $\mathscr{A}.$

Proposition. Let $f$ be any function $\mathrm{X} \to \mathrm{Y}$ and let $\mathscr{Y}$ be any subset of the power set of $\mathrm{Y}.$ Then, $f^{-1}(\sigma(\mathscr{Y})) = \sigma(f^{-1}(\mathscr{Y})).$

Proof. You probably already know that the set of preimages of a sigma algebra conforms a sigma algebra, this proves at once the $\supset.$ For the other side, consider the set $\mathscr{Y}'$ of sets $\mathrm{E}$ in $\sigma(\mathscr{Y})$ such that $f^{-1}(\mathrm{E}) \in \sigma(f^{-1}(\mathscr{Y})).$ Straightforwardly, $\mathscr{Y}'$ is a sigma field containing $\mathscr{Y},$ hence, it also contains $\mathscr{Y},$ this proves $\subset.$ Q.E.D.