Measure Theory/Properties of General Integrals

From testwiki
Jump to navigation Jump to search

Properties and Convergence of General Integrals

Recall the definition of the length-measure integral of a measurable function f:E,

Ef=Ef+Ef

and recall the definitions

f+=max{f,0},f=max{f,0}

In this lesson, assume

  • f,g,fn:E are measurable functions
  • fnf
  • c,d
  • |fn|g for all n+

Exercise 1. Consistency

Explain why the check for consistency, which we have done for previous generalizations of integral definitions, is trivial in this case.

Exercise 2. Basic Properties of General Integrals

Prove the basic properties of general integrals: Linearity, order-preserving, triangle inequality, the ML bound, and finite additivity.

Note: What is the positive part of f+g in terms of f+,f,g+,g?

Lebesgue's Dominated Convergence Theorem

Here resides the last (or, depending on how you count, the penultimate) of the great and famous convergence theorems of measure theory.

Template:Definition

Besides the assumptions at the top of this page, further assume that g is integrable.

Lebesgue's Dominated Convergence Theorem then states that the swaparoo follows.

limnEfn=limnfn=Ef

Exercise 3. Prove the LDCT

 Part A. 
With the observation that gfn is a nonnegative function, use Fatou's lemma.
 Part B. 
Use an earlier result to establish that f is integrable.  Then infer from Part A. that EgEfEglimEfn.
Part C.
Now apply reasoning similar to that in parts A. and B. above to g+fn to infer Eflim_Efn and conclude the proof.
Part D. 
Prove the following generalization of the LDCT.  

Let gn be a sequence of integrable functions.  Assume further that 

* gn so-to-speak "pairwise dominates" the sequence fn. Formally this means |fn|gn for each n+.

* gng on E.

* limnEgn=Elimngn=Eg

Now prove that limnEfn=Elimnfn=Ef.

The proof should merely reiterate all of the proof of the LDCT, but replacing g by gn where appropriate.  











Template:CourseCat