proof of monotone convergence theorem


It is enough to prove the following

Theorem 1

Let (X,μ) be a measurable spaceMathworldPlanetmathPlanetmath and let fk:X→R∪{+∞} be a monotone increasing sequence of positive measurable functionsMathworldPlanetmath (i.e. 0≤f1≤f2≤…). Then f⁢(x)=limk→∞⁡fk⁢(x) is measurable and

limn→∞⁡∫Xfk⁢𝑑μ=∫Xf⁢(x)⁢𝑑μ.

First of all by the monotonicity of the sequence we have

f⁢(x)=supk⁡fk⁢(x)

hence we know that f is measurable. Moreover being fk≤f for all k, by the monotonicity of the integral, we immediately get

supk⁡∫Xfk⁢𝑑μ≤∫Xf⁢(x)⁢𝑑μ.

So take any simple measurable function s such that 0≤s≤f. Given also α<1 define

Ek={x∈X:fk⁢(x)≥α⁢s⁢(x)}.

The sequence Ek is an increasing sequence of measurable sets. Moreover the union of all Ek is the whole space X since limk→∞⁡fk⁢(x)=f⁢(x)≥s⁢(x)>α⁢s⁢(x). Moreover it holds

∫Xfk⁢𝑑μ≥∫Ekfk⁢𝑑μ≥α⁢∫Eks⁢𝑑μ.

Since s is a simple measurable function it is easy to check that E↦∫Es⁢𝑑μ is a measureMathworldPlanetmath and hence

supk⁡∫Xfk⁢𝑑μ≥α⁢∫Xs⁢𝑑μ.

But this last inequalityMathworldPlanetmath holds for every α<1 and for all simple measurable functions s with s≤f. Hence by the definition of Lebesgue integral

supk⁡∫Xfk⁢𝑑μ≥∫Xf⁢𝑑μ

which completesPlanetmathPlanetmathPlanetmathPlanetmath the proof.

Title proof of monotone convergence theorem
Canonical name ProofOfMonotoneConvergenceTheorem
Date of creation 2013-03-22 13:29:56
Last modified on 2013-03-22 13:29:56
Owner paolini (1187)
Last modified by paolini (1187)
Numerical id 7
Author paolini (1187)
Entry type Proof
Classification msc 28A20
Classification msc 26A42