How to prove convergence of integrals using almost everywhere convergence?

  • MHB
  • Thread starter Chris L T521
  • Start date
In summary, "Almost everywhere convergence" refers to a type of convergence where a sequence of functions converges to a limit almost everywhere, meaning that it converges everywhere except on a set of measure zero. It is closely related to convergence of integrals and is a weaker form of convergence than uniform convergence. To prove convergence of integrals using almost everywhere convergence, the Dominated Convergence Theorem can be used, but in certain cases, almost everywhere convergence may not imply convergence of integrals.
  • #1
Chris L T521
Gold Member
MHB
915
0
Here's this week's problem.

-----

Problem: Suppose that $\{f_n(x)\}$ and $\{g_n(x)\}$ are two sequences of measurable functions (on $\mathbb{R}$) such that $|f_n(x)|\leq g_n(x)$ for each $n=1,2,\ldots$. And suppose that $f_n$ converges to $f$ and that $g_n$ converges to $g$ almost everywhere. Show that
\[\lim_{n\to\infty}\int_{\mathbb{R}} g_n\,dm = \int_{\mathbb{R}} g\,dm\]
implies that
\[\lim_{n\to\infty}\int_{\mathbb{R}} f_n\,dm = \int_{\mathbb{R}} f\,dm\]

-----

 
Physics news on Phys.org
  • #2
This week's problem was correctly answered by PaulRS and girdav.

Here's my solution:

Proof: Let $\{g_n\}$ be a sequence of Lebesgue integrable functions such that $g_n\rightarrow g$ almost everywhere and $\displaystyle\int_{\mathbb{R}} g_n\,dm\rightarrow \int_{\mathbb{R}} g\,dm$, and let $\{f_n\}$ be a sequence of Lebesgue measurable functions such that $|f_n|\leq g_n$ and $f_n\rightarrow f$ almost everywhere. Since $|f_n|\leq g_n$, we have $|f|\leq g$. Thus, $|f_n-f|\leq |f_n| + |f|\leq g_n+g$ and $\{g_n+g-|f_n-f|\}$ is a sequence of non-negative measurable functions. Therefore, by Fatou's lemma, we have\[\int_{\mathbb{R}}\liminf(g_n+g-|f_n-f|)\,dm\leq \liminf\int_{\mathbb{R}} (g_n+g-|f_n-f|)\,dm\]
This implies that
\[\int_{\mathbb{R}} 2g\,dm \leq \int_{\mathbb{R}} 2g\,dm+\liminf\left(-\int_{\mathbb{R}}|f_n-f|\,dm\right)=\int_{\mathbb{R}} 2g\,dm -\limsup\int_{\mathbb{R}} |f_n-f|\,dm\]
Hence, we see that
\[\limsup\int_{\mathbb{R}}|f_n-f|\,dm\leq 0\leq \liminf\int_{\mathbb{R}}|f_n-f|\,dm\]
and thus we have $\displaystyle\int_{\mathbb{R}}|f_n-f|\,dm\rightarrow 0$.Therefore, we see that $\displaystyle\int_{\mathbb{R}} g_n\,dm\rightarrow \int_{\mathbb{R}} g\,dm$ implies that $\displaystyle\int_{\mathbb{R}} f_n\,dm\rightarrow\int_{\mathbb{R}} f\,dm$. Q.E.D.

and here's PaulRS's solution:

Since $|f_n| \leq g_n$ we have that $g_n - f_n \geq 0$ and $g_n + f_n \geq 0$, and so we have $L^+$ functions and we may apply Fatou's Lemma:

\[\int g dm - \int f dm = \int (g-f)dm = \int \liminf (g_n - f_n) dm \leq \liminf \int (g_n - f_n) dm = \liminf \left( \int g_n dm - \int f_n dm \right) = \int g dm - \limsup \int f_n dm \]where we have used that $\liminf g_n = g$ and $\liminf f_n = f$ almost everywhere, $\lim \int g_n dm = \int g dm$ (and so the liminf is the sum of the limit plus the liminf) and $-\limsup \left( ... \right) = \liminf (-...)$.Similarly : \[\int g dm + \int f dm = \int (g+f)dm = \int \liminf (g_n +f_n) dm \leq \liminf \int (g_n+ f_n) dm = \liminf \left( \int g_n dm + \int f_n dm \right) = \int g dm + \liminf \int f_n dm \]Now, these two inequalities translate to\[\limsup \int f_n dm \leq \int f dm \leq \liminf \int f_n dm\]which proves the assertion. $\square$
 

FAQ: How to prove convergence of integrals using almost everywhere convergence?

What is "almost everywhere convergence"?

"Almost everywhere convergence" refers to a type of convergence in which a sequence of functions converges to a limit almost everywhere, meaning that it converges everywhere except on a set of measure zero. This is a stronger form of convergence than pointwise convergence, which only requires convergence at each individual point.

How is "almost everywhere convergence" related to convergence of integrals?

Almost everywhere convergence is closely related to convergence of integrals because if a sequence of functions converges almost everywhere, then it also converges in the integral sense. This means that the integral of the limit function will be equal to the limit of the integrals of the individual functions in the sequence.

What is the difference between "almost everywhere convergence" and uniform convergence?

The main difference between almost everywhere convergence and uniform convergence is the type of convergence they require. Almost everywhere convergence only requires convergence almost everywhere, while uniform convergence requires convergence at every point in the domain. Additionally, uniform convergence is a stronger form of convergence than almost everywhere convergence, as it also implies almost everywhere convergence.

How can you prove convergence of integrals using almost everywhere convergence?

To prove convergence of integrals using almost everywhere convergence, you can use the Dominated Convergence Theorem. This theorem states that if a sequence of functions satisfies certain conditions, including almost everywhere convergence, then the integral of the limit function will be equal to the limit of the integrals of the individual functions in the sequence. The proof involves using properties of measure theory and the convergence of integrals.

Can almost everywhere convergence fail to imply convergence of integrals?

Yes, almost everywhere convergence can fail to imply convergence of integrals in certain cases. For example, if the sequence of functions is not dominated by an integrable function, then the Dominated Convergence Theorem cannot be applied and the convergence of integrals may not hold. Additionally, there may be other conditions or properties of the sequence of functions that affect the convergence of integrals, even if almost everywhere convergence is satisfied.

Similar threads

Replies
3
Views
2K
Replies
4
Views
1K
Replies
11
Views
920
Replies
1
Views
7K
Replies
4
Views
2K
Replies
9
Views
1K
Back
Top