Fatou’s Lemma

In mathematics, Fatou’s lemma is an inequality that demonstrates a relationship between the Lebesgue integral of a sequence of functions’ limit inferior and the limit inferior of integrals of these functions. Pierre Fatou is honoured with the name of the lemma. The definition of Fatou’s lemma, strict inequality examples, reverse Fatou’s lemma and solved problems are all covered in-depth in this article.

Table of Contents:

Fatou’s Lemma Definition

Fatou’s Lemma states that “given a measure space (Ω, F, μ), and the set X, that belongs to F. Assume that {fn} is a sequence of (F, BR≥0) – measurable non-negative functions fn: x→[0, +∞]. Now, define the function f: x→[0, +∞] by applying f(x) = lim infn→∞ fn(x), for every x that belongs to X”.

Then f is (F, BR≥0) – measurable, and

x f dμ ≤ lim infn→∞x fn dμ, in which the integrals may be infinite.”

Fatou’s Lemma holds true if its assumptions hold μ, almost everywhere

Note: BR≥0 represents the σ-algebra of Borel sets on [0, +∞].

Strict Inequality Examples

Connecting the space S with the Lebesgue measure and Borel σ-algebra, we have the following examples of strict inequality.

  • Now, let us assume the probability space. Let S = [0, 1] represent the unit interval. Hence, for every natural number, n represents:
\(\begin{array}{l}f_{n}(x)= \left\{\begin{matrix}n & for\ x\in (0, 1/n) \\0 & otherwise\\\end{matrix}\right.\end{array} \)
  • Now, let us take an example of uniform convergence. Assume that S represents the set of all real numbers, and it is defined by
\(\begin{array}{l}f_{n}(x)= \left\{\begin{matrix}1/n & for\ x\in [0, 1/n] \\0 & otherwise\\\end{matrix}\right.\end{array} \)

Reverse Fatou’s Lemma

Assume that f1, f2, and so on are a sequence of extended real-valued measurable functions defined on a measure space (S, Σ, μ). If a non-negative integrable function g on S exists such that fn ≤ g for every n, then

lim supn→∞S fn dμ ≤ ∫S lim supn→∞ fn

Also, read:

Solved Example

Example:

Assume that fn = (1/n)1[0,n] and f = 0. Prove that fn converges uniformly to 0, fn ∈ L1(R,Mλ, λ) – that is, fn is integrable with respect to the Lebesgue measure -yet ∫f dλ ≠ limn ∫ fn dλ. Explain in detail why this does not contradict the Monotone Convergence Theorem or the Dominated Convergence Theorem. And also explain Fatou’s Lemma involved?

Solution:

Given,

For each x ∈ R, |fn(x)| ≤(1/n), so if ε > 0,

|fn(x)| ≤ ε for all n ≥ 1/ε

Hence, we can say that fn→ 0 uniformly on R.

Now, for each n ∈ N, we have

∫fn dλ = (1/n) λ([0, n]) = 1, such that fn is integrable.

Hence,

0 = ∫ limn fn dλ ≠ limn ∫ fn dλ = 1

Because the sequence (fn) is not Monotone, this does not contradict the Monotone Convergence Theorem. Furthermore, fn(x) = 0 for n ≤ x for each x. When n ≥ x , the value leaps to 1/n > 0 and then drops.

We demonstrate that (fn) is not dominated by an integrable function to prove that the Dominated Convergence Theorem is not violated. Take a look at this.

|fn(x)| ≤ g(x), for all x.

Hence, for each n ∈ N, g(x) ≥ 1/n, for n − 1 < x ≤ n. So,

\(\begin{array}{l}g\geq \sum_{n}\frac{1}{n}1(n-1, n]\end{array} \)

On integrating the above function, it gives

\(\begin{array}{l}\int g d\lambda \geq \int \sum_{n}\frac{1}{n}1(n-1, n]d\lambda = \sum_{n}\frac{1}{n}\lambda ((n-1, n])= \sum_{n}\frac{1}{n}=\infty \end{array} \)

Thus, g can’t be integrable.

Fatou’s lemma must, of course, hold, as we can see.

\(\begin{array}{l}0=\int lim_{n}inf f_{n}d\lambda \leq lim_{n}inf \int f_{n}d\lambda=1\end{array} \)

Frequently Asked Questions on Fatou’s Lemma

Q1

What is Fatou’s Lemma?

Fatou’s lemma is an inequality that demonstrates a relationship between the Lebesgue integral of a sequence of functions’ limit inferior and the limit inferior of integrals of these functions

Q2

What are the two methods to prove Fatou’s lemma?

The two methods to prove Fatou’s lemma are:
Through Monotone Convergence Theorem
From first principles

Q3

Can we use Fatou’s lemma in probability theory?

Yes, we can use Fatou’s lemma in probability theory.

Comments

Leave a Comment

Your Mobile number and Email id will not be published.

*

*

close
close

Play

&

Win