Chapter 7

The Abstract Lebesgue IntegralThe general Lebesgue integral requires a measure space, (Ω,F ,µ) and, to begin with, anonnegative measurable function. I will use Lemma 1.11.2 about interchanging two supre-mums frequently. Also, I will use the observation that if {an} is an increasing sequenceof points of [0,∞] , then supn an = limn→∞ an which is obvious from the definition of sup.Lebesgue integration is a theory which depends on absolute convergence. Thus we under-stand things in terms of nonnegative functions. For complex valued functions, we considerpositive and negative parts of real and imaginary parts. Thus one typically discusses non-negative functions in statements of the main theorems.

7.1 Nonnegative Measurable Functions7.1.1 Riemann Integrals For Decreasing FunctionsFirst of all, the notation [g < f ] means {ω ∈Ω : g(ω)< f (ω)} with other variants of thisnotation being similar. Also, the convention, 0 ·∞ = 0 will be used to simplify the presen-tation whenever it is convenient to do so. The notation a∧b means the minimum of a andb.

Definition 7.1.1 Let f : [a,b]→ [0,∞] be decreasing. Note that ∞ is a possiblevalue. Define ∫ b

af (λ )dλ ≡ lim

M→∞

∫ b

aM∧ f (λ )dλ = sup

M

∫ b

aM∧ f (λ )dλ

where a∧b means the minimum of a and b. Note that for f bounded,

supM

∫ b

aM∧ f (λ )dλ =

∫ b

af (λ )dλ

where the integral on the right is the usual Riemann integral because eventually M > f .For f a nonnegative decreasing function defined on [0,∞),∫

0f dλ ≡ lim

R→∞

∫ R

0f dλ = sup

R>1

∫ R

0f dλ = sup

RsupM>0

∫ R

0f ∧Mdλ

Since decreasing bounded functions are Riemann integrable, the above definition iswell defined. See Theorem 5.1.11. Now here is an obvious property.

Lemma 7.1.2 Let f be a decreasing nonnegative function defined on an interval [a,b] .Then if [a,b] =∪m

k=1Ik where Ik ≡ [ak,bk] and the intervals Ik are non overlapping, it follows∫ b

af dλ =

m

∑k=1

∫ bk

ak

f dλ .

Proof: This follows from Theorems 5.1.7 and 5.1.11 along with the computation,∫ b

af dλ ≡ lim

M→∞

∫ b

af ∧Mdλ = lim

M→∞

m

∑k=1

∫ bk

ak

f ∧Mdλ =m

∑k=1

∫ bk

ak

f dλ

Note both sides could equal +∞. ■In all considerations below, we assume h is fairly small, certainly much smaller than R.

Thus R−h > 0.

159

Chapter 7The Abstract Lebesgue IntegralThe general Lebesgue integral requires a measure space, (Q,.¥,) and, to begin with, anonnegative measurable function. I will use Lemma 1.11.2 about interchanging two supre-mums frequently. Also, I will use the observation that if {a,} is an increasing sequenceof points of [0,°], then sup, dy = limy—..d, Which is obvious from the definition of sup.Lebesgue integration is a theory which depends on absolute convergence. Thus we under-stand things in terms of nonnegative functions. For complex valued functions, we considerpositive and negative parts of real and imaginary parts. Thus one typically discusses non-negative functions in statements of the main theorems.7.1 Nonnegative Measurable Functions7.1.1 Riemann Integrals For Decreasing FunctionsFirst of all, the notation [g < f] means {@ € Q: g(@) < f(@)} with other variants of thisnotation being similar. Also, the convention, 0 -oo = 0 will be used to simplify the presen-tation whenever it is convenient to do so. The notation a/b means the minimum of a andbDefinition 7.1.1 Lez f : [a,b] > [0,2] be decreasing. Note that ~ is a possiblevalue. Defineb b b[ fans jim | MAF (A)d2 =sup | MAf (A)dawhere aN b means the minimum of a and b. Note that for f bounded,b bsup [ Maf(ayaa = | f(A)dAwhere the integral on the right is the usual Riemann integral because eventually M > f.For f anonnegative decreasing function defined on |0,°),oo) R »R »R/ fda = lim [ fdh-=sup | fda =supsup [| fAMaa0 Re JO R>1/0 R M>0/0Since decreasing bounded functions are Riemann integrable, the above definition iswell defined. See Theorem 5.1.11. Now here is an obvious property.Lemma 7.1.2 Let f be a decreasing nonnegative function defined on an interval {a,b}.Then if [a,b] = URL Ik where Ij, = |ax, bk| and the intervals I, are non overlapping, it followsb m byi] fdd=¥) |] fda.Ja k=1" akProof: This follows from Theorems 5.1.7 and 5.1.11 along with the computation,b b m by m byda = lim | AMdA = lim \Mda = da[ft ym fF ms nt LI, 4Note both sides could equal +-co.In all considerations below, we assume / is fairly small, certainly much smaller than R.Thus R—h>0O.159