0
$\begingroup$

Pleas check my proof of the following assertion.

Let $\{E_k\}$ be a sequence of measurable sets, and $E_k \nearrow E$. Suppose $f(x)$ is nonnegative and measurable on $E$. Prove

$$\int_E f(x) \, dx = \lim_{k \to \infty} \int_{E_k} f(x) \, dx$$


Proof. First note that

$$\int_{E_k} f(x) \, dx= \int_E f(x)\chi_{E_k}(x) dx$$

and we have (using our hypothesis in the second equality)

$$\lim f(x)\chi_{E_k}(x) = f(x) \lim \chi_{E_k}(x) \stackrel{\tiny{HYP.}}{=} f(x)\chi_E(x) \tag{?}$$

Furthermore, $\left(f(x)\chi_{E_k}(x)\right)$ is monotone increasing, nonnegative, and measurable (product of two measurable functions). Therefore, by the MCT

$$\lim_k \int_{E_k} f(x) \, dx= \lim_k \int_E f(x)\chi_{E_k}(x) \, dx \stackrel{\tiny{MCT}}= \int_E f(x) \, dx$$

So the proof is complete.


I took for granted in $(?)$ that $\lim \chi_{E_k}(x) = \chi_E(x)$. So to prove that we need to show that for all $\epsilon > 0$, and any $x \in E$, there exists and $N$, such that whenever $k > N$ we have

$$ |\chi_E(x) - \chi_{E_k}(x)| < \epsilon$$

But this can be relaxed to almost any $x \in E$ for the MCT. So, let $\epsilon > 0$ and let $x \in E$. The set of $x$ on the boundary of of $E$ has measure zero, since that set has side length zero in at least one dimension (??). If $x$ is in the interior of $E$ we can, by hypothesis, produce an $E_N \subset E$ such that $x \in E_N = E \cap E_N$. Therefore, for all $k > N$ we have $x \in E_k$ which means that $\chi_E(x) = 1$ and $\chi_{E_k}(x) = 1$, which clearly satisfies the inequality above.


Something about that doesn't sit right with me, the (??) part mostly. I was thinking about setting this problem up a different way too.

$\endgroup$
0
$\begingroup$

Not sure if this answers your question, but we can show that if $E_k \uparrow \ E$ then we have that: $$ \mu\left(\bigcup_{k = 1}^{\infty}E_k\right) = \lim_{n \to \infty} \mu\left(\bigcup_{k = 1}^{n}E_k\right) $$ This is done by defining $G_1 = E_1$, $G_2 = E_2 \setminus E_1$, $G_3 = E_3 \setminus E_2$, etc, with $G_ k = E_k \setminus E_{k-1}$. Then, we see that these sets are disjoint, and $\bigcup G_k = E = \bigcup E_k$. By disjoint additivity, we have: $$ \mu(E) = \sum_{k = 1}^\infty \mu(G_k) = \lim_{n \to \infty} \sum_{k = 1}^{n}\mu = (G_k) = \lim_{n \to \infty} \mu \left(\bigcup_{k = 1}^{n}G_k\right) = \lim_{n \to \infty} \mu \left(\bigcup_{k = 1}^{n}E_k\right) $$ as we needed. Having proven the two sets have the same measure, it follows that the indicator function converges accordingly (in an A.E sense)

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.