14.8. NONNEGATIVE SIMPLE FUNCTIONS 335

[ f > λ ] which is what it means for f to be measurable. Also note that, in spite of the picture,in general we don’t know a good description of this set other than that it is measurable.

λ i

A B C

y = f (x)

Definition 14.7.1 Let (Ω,F , µ) be a measure space and suppose f : Ω→ [0,∞] ismeasurable. Then define∫

f dµ ≡∫

0µ ([ f > λ ])dλ =

∫∞

0µ(

f−1 (λ ,∞))

which makes sense because λ → µ ([ f > λ ]) is nonnegative and decreasing. On the rightyou have an improper Riemann integral like what was discussed above.

Note that if f ≤ g, then∫

f dµ ≤∫

gdµ because µ ([ f > λ ])≤ µ ([g > λ ]) . Next I pointout that the integral is a limit of lower sums.

Lemma 14.7.2 In the situation of the above definition,∫f dµ = sup

h>0

∑i=1

µ ([ f > hi])h

Proof: Let m(h,R) ∈N satisfy R−h < hm(h,R)≤ R. Then limR→∞ hm(h,R) = ∞ andso ∫

f dµ ≡∫

0µ ([ f > λ ])dλ ≡ sup

Msup

R

∫ R

0µ ([ f > λ ])∧Mdλ

supM

supR

∫ hm(h,R)

0µ ([ f > λ ])∧Mdλ = sup

MsupR>0

suph>0

m(h,R)

∑k=1

(µ ([ f > kh])∧M)h

because the sum is just a lower sum for the integral∫ hm(h,R)

0 µ ([ f > λ ])∧Mdλ . Hence,switching the order of the sups, this equals

supR>0

suph>0

supM

m(h,R)

∑k=1

(µ ([ f > kh])∧M)h = supR>0

suph>0

limM→∞

m(h,R)

∑k=1

(µ ([ f > kh])∧M)h

= suph>0

supR

m(R,h)

∑k=1

µ ([ f > kh])h = suph>0

∑k=1

µ ([ f > kh])h.

14.8 Nonnegative Simple FunctionsTo begin with, here is a useful lemma.

Lemma 14.8.1 If f (λ ) = 0 for all λ > a, where f is a decreasing nonnegative function,then

∫∞

0 f (λ )dλ =∫ a

0 f (λ )dλ .

14.8. NONNEGATIVE SIMPLE FUNCTIONS 335[f > A] which is what it means for f to be measurable. Also note that, in spite of the picture,in general we don’t know a good description of this set other than that it is measurable.y= f(x)Definition 14.7.1 Le (Q,.F, UW) be a measure space and suppose f :Q — [0,-] ismeasurable. Then define[fau= [u ([f >A])da = [ HU ax) aawhich makes sense because A — | ([f > A]) is nonnegative and decreasing. On the rightyou have an improper Riemann integral like what was discussed above.Note that if f < g, then [ fdu < f gdu because u ([f > A]) < u([g > A]). Next I pointout that the integral is a limit of lower sums.Lemma 14.7.2 In the situation of the above definition,[fan = supa ( [f > hil) hh>0j—Proof: Let m(h,R) € N satisfy R—h < hm(h,R) < R. Then limg_,.. hm (h, R) = c and)[fan= [ou ([f >A])daA = supsup [1 ([f > A]) AMdahm(h,R m(h,R)supsup | LL ([f > A]) AMdA = sup sup sup Ye (u (Lf > kh]) AM)hM R JO M R>0h>0 f=]because the sum is just a lower sum for the integral fon (1.8) L([f >A]) AMdA. Hence,switching the order of the sups, this equalsm(h,R) m(h,R)supsupsup )) (u([f > kh])\M)h=supsup lim )" (u([f >kh])AM)hR>0h>0 M f=] R>0n>0M>~ 5Xm(R,h) oo= sup sup y LL ([f > kh])h = sup y L([f>khl)h.h>0 R- x=] h>0k=114.8 Nonnegative Simple FunctionsTo begin with, here is a useful lemma.Lemma 14.8.1 Jf f(A) =0 forall A > a, where f is a decreasing nonnegative function,then fo f(A)dA = fo f (A) da.