Modes of Convergence: Difference between revisions

From Optimal Transport Wiki
Jump to navigation Jump to search
No edit summary
 
(20 intermediate revisions by 2 users not shown)
Line 1: Line 1:
== Relevant Definitions<ref name="Craig">Craig, Katy. ''MATH 201A Lecture 17''. UC Santa Barbara, Fall 2020.</ref>==
== Relevant Definitions<ref name="Craig">Craig, Katy. ''MATH 201A Lecture 17''. UC Santa Barbara, Fall 2020.</ref>==
Denote our measure space as <math> (X, \mathcal{M}, \mu) </math>. Note that a property p(x) holds for almost every <math>x \in X</math> if the set <math>\{x \in X: p(x) \text{ doesn't hold }\}</math> has measure zero.
Denote our measure space as <math> (X, \mathcal{M}, \mu) </math>. Note that a property p holds for almost every <math>x \in X</math> if the set <math>\{x \in X: p \text{ doesn't hold }\}</math> has measure zero.
* A sequence of functions <math>f_n</math>  converges pointwise if <math>f_n(x) \to f(x) </math> for all <math>x \in X </math>.
* A sequence of measurable functions <math>\{f_n \}</math>  converges pointwise if <math>f_n(x) \to f(x) </math> for all <math>x \in X </math>.
* A sequence of functions <math>f_n</math> converges uniformly if <math>\sup_{x \in X} |f_n(x) - f(x)| \to 0 </math>.  
* A sequence of measureable functions <math>\{f_n \}</math> converges uniformly if <math>\sup_{x \in X} |f_n(x) - f(x)| \to 0 </math>.  
*A sequence of measurable functions <math>\{f_n \}</math> converges to <math> f</math> pointwise almost everywhere if <math> f_n (x) \to f(x)</math> for almost every <math> x </math>, or <math> \mu( \{x: f(x) \neq \lim_{n \to \infty} f(x) \}) =0</math>.
*A sequence of measurable functions <math>\{f_n \}</math> converges to <math> f</math> pointwise almost everywhere if <math> f_n (x) \to f(x)</math> for almost every <math> x </math>, or <math> \mu( \{x: f(x) \neq \lim_{n \to \infty} f(x) \}) =0</math>.
*A sequence of measurable functions <math>f_n</math> converges in <math>L^1</math> if <math>\int |f_n - f| \to 0.</math>
*A sequence of measurable functions <math>\{f_n \}</math> converges in <math>L^1</math> if <math>\int |f_n - f| \to 0.</math>
check [[Convergence in Measure]] for convergence in measure.
check [[Convergence in Measure]] for convergence in measure.


== Relevant Properties <ref name="Craig, Katy">Craig, Katy. ''MATH 201A Lecture 18''. UC Santa Barbara, Fall 2020.</ref>==
== Relevant Properties <ref name="Craig, Katy">Craig, Katy. ''MATH 201A Lecture 18''. UC Santa Barbara, Fall 2020.</ref>==
* <math>f_n \to f</math> through    uniform Convergence <math>\to </math><math> f_n \to f</math> through  pointwise convergence  <math> \to </math> <math>f_n \to f</math> pointwise a.e. convergence
* <math>f_n \to f</math> through    uniform convergence implies <math> f_n \to f</math> through  pointwise convergence, which in turn implies <math>f_n \to f</math> pointwise a.e. convergence.
* <math>f_n \to f</math> through  <math> L^1</math> convergence <math>\to </math> <math>f_n \to f</math>  through pointwise a.e convergence up to a subsequence.
* <math>f_n \to f</math> through  <math> L^1</math> convergence  implies  <math>f_n \to f</math>  through pointwise a.e convergence up to a subsequence.  
* <math>f_n \to f</math> Pointwise a.e. convergence, equipped with dominating function, implies <math>f_n \to f</math>  in  <math>L^1</math>. To see example of why we need a dominating function, read [[Dominated Convergence Theorem]], particularly applications of the theorem.
* <math>f_n \to f</math> Pointwise a.e. convergence, equipped with dominating function, implies <math>f_n \to f</math>  in  <math>L^1</math>. To see the proof and examples of why we need a dominating function, read [[Dominated Convergence Theorem]].
* [[Convergence in Measure]] lists relationships between convergence in measure and other forms of convergence.
* [[Convergence in Measure]] lists relationships between convergence in measure and other forms of convergence.
== Proof <math>f_n \to f, \text{ in }L^1</math>  implies <math>f_n \to f </math>  in measure==
Fix <math> \epsilon >0,  </math> define
<math>E_{n, \epsilon}  = \{x: |f_n(x) - f(x)| \geq \epsilon\}</math>
We can see that that
<math> \int_{E_{n,\epsilon}}\epsilon \leq  \int_{E_{n,\epsilon}}|f_n - f|\leq  \int|f_n - f|.  </math>
we can see by <math>L^1</math> convergence that <math>\lim_{n \to \infty}\int |f_n-f| = 0, </math>
meaning that
<math> \lim_{n \to \infty} \int_{E_{n,\epsilon}}\epsilon = \lim_{n \to \infty}\epsilon \mu(E_{n,\epsilon}) = 0,</math>
or
<math> \lim_{n \to \infty}  \mu(E_{n,\epsilon}) =  \lim_{n \to \infty}\mu(\{x: |f_n(x) - f(x)| \geq \epsilon\}) = 0 </math>.
Hence the proposition follows.
== Proof <math>f_n \to f</math> through  <math> L^1</math> convergence  implies  <math>f_n \to f</math>  pointwise a.e convergence up to a subsequence. ==
<math> L^1</math> convergence means <math>f_n \to f</math> in measure. Because <math>f_n \to f</math> in measure, there exists a subsequence <math>f_{n_k}</math> such that <math>f_{n_k} \to f</math> pointwise a.e.
== Example: <math>f_n \to f</math> in measure does not imply convergence in <math>L^1 </math>==
Let <math> f_n:= \frac{1}{n}1_{[0,n]} </math>. Then <math> f_n \to 0 </math> in measure, but <math> f_n </math> does not converge to 0 in <math> L^1 </math>. This is because <math> \int f_n d\lambda = 1 </math> for all <math> n </math>.
==References==

Latest revision as of 03:59, 19 December 2020

Relevant Definitions[1]

Denote our measure space as . Note that a property p holds for almost every if the set has measure zero.

  • A sequence of measurable functions converges pointwise if for all .
  • A sequence of measureable functions converges uniformly if .
  • A sequence of measurable functions converges to pointwise almost everywhere if for almost every , or .
  • A sequence of measurable functions converges in if

check Convergence in Measure for convergence in measure.

Relevant Properties [2]

  • through uniform convergence implies through pointwise convergence, which in turn implies pointwise a.e. convergence.
  • through convergence implies through pointwise a.e convergence up to a subsequence.
  • Pointwise a.e. convergence, equipped with dominating function, implies in . To see the proof and examples of why we need a dominating function, read Dominated Convergence Theorem.
  • Convergence in Measure lists relationships between convergence in measure and other forms of convergence.

Proof implies in measure

Fix define We can see that that

we can see by convergence that meaning that

or . Hence the proposition follows.

Proof through convergence implies pointwise a.e convergence up to a subsequence.

convergence means in measure. Because in measure, there exists a subsequence such that pointwise a.e.

Example: in measure does not imply convergence in

Let . Then in measure, but does not converge to 0 in . This is because for all .


References

  1. Craig, Katy. MATH 201A Lecture 17. UC Santa Barbara, Fall 2020.
  2. Craig, Katy. MATH 201A Lecture 18. UC Santa Barbara, Fall 2020.