9.1: L-Integrals and Antiderivatives
- Last updated
- Jan 15, 2020
- Save as PDF
- Page ID
- 32388
( \newcommand{\kernel}{\mathrm{null}\,}\)
I. Lebesgue theory makes it possible to strengthen many calculus theorems. We shall start with functions on E1,f:E1→E. (A reader who has omitted the "starred" part of Chapter 8, §7, will have to set E=E∗(En,Cn) throughout.)
By L-integrals of such functions, we mean integrals with respect to Lebesgue measure m in E1. Notation:
L∫baf=L∫baf(x)dx=L∫[a,b]f
and
L∫abf=−L∫baf.
For Riemann integrals, we replace "L" by "R." We compare such integrals with antiderivatives (Chapter 5, §5), denoted
∫baf,
without the "L" or "R." Note that
L∫[a,b]f=L∫(a,b)f,
etc., since m{a}=m{b}=0 here.
Theorem 9.1.1
Let f:E1→E be L-integrable on A=[a,b]. Set
H(x)=L∫xaf,x∈A.
Then the following are true.
(i) The function f is the derivative of H at any p∈A at which f is finite and continuous. (At a and b, continuity and derivatives may be one-sided from within.)
(ii) The function H is absolutely continuous on A; hence VH[A]<∞.
- Proof
-
(i) Let p∈(a,b],q=f(p)≠±∞. Let f be left continuous at p; so, given ε>0, we can fix c∈(a,p) such that
|f(x)−q|<ε for x∈(c,p).
Then
(∀x∈(c,p))|L∫px(f−q)|≤L∫px|f−q|≤L∫px(ε)=ε⋅m[x,p]=ε(p−x).
But
L∫px(f−q)=L∫pxf−L∫pxqL∫pxq=q(p−x), and L∫pxf=L∫paf−L∫xaf=H(p)−H(x).
Thus
|H(p)−H(x)−q(p−x)|≤ε(p−x);
i.e.,
|H(p)−H(x)p−x−q|≤ε(c<x<p).
Hence
f(p)=q=limx→p−ΔHΔx=H′−(p).
If f is right continuous at p∈[a,b), a similar formula results for H′+(p). This proves clause (i).
(ii) Let ε>0 be given. Then Theorem 6 in Chapter 8, §6, yields a δ>0 such that
|L∫Xf|≤L∫X|f|<ε
whenever
mX<δ and A⊇X,X∈M.
Here we may set
X=r⋃i=1Ai (disjoint)
for some intervals
Ai=(ai,bi)⊆A
so that
mX=∑imAi=∑i(bi−ai)<δ.
Then (1) implies that
ε>L∫X|f|=∑iL∫Ai|f|≥∑i|L∫biaif|=∑i|H(bi)−H(ai)|.
Thus
∑i|H(bi)−H(ai)|<ε
whenever
∑i(bi−ai)<δ
and
A⊇⋃i(ai,bi) (disjoint).
(This is what we call "absolute continuity in the stronger sense.") By Problem 2 in Chapter 5, §8, this implies "absolute continuity" in the sense of Chapter 5, §8, hence VH[A]<∞.◻
Note 1. The converse to (i) fails: the differentiability of H at p does not imply the continuity of its derivative f at p (Problem 6 in Chapter 5, §2).
Note 2. If f is continuous on A−Q (Q countable), Theorem 1 shows that H is a primitive (antiderivative): H=∫f on A. Recall that "Q countable" implies mQ=0, but not conversely. Observe that we may always assume a,b∈Q.
We can now prove a generalized version of the so-called fundamental theorem of calculus, widely used for computing integrals via antiderivatives.
Theorem 9.1.2
If f:E1→E has a primitive F on A=[a,b], and if f is bounded on A−P for some P with mP=0, then f is L-integrable on A, and
L∫xaf=F(x)−F(a)for all x∈A.
- Proof
-
By Definition 1 of Chapter 5, §5, F is relatively continuous and finite on A=[a,b], hence bounded on A (Theorem 2 in Chapter 4, §8).
It is also differentiable, with F′=f, on A−Q for a countable set Q⊆A, with a,b∈Q. We fix this Q along with P.
As we deal with A only, we surely may redefine F and f on −A:
F(x)={F(a) if x<a,F(b) if x>b,
and f=0 on −A. Then f is bounded on −P, while F is bounded and
continuous on E1, and F′=f on −Q; so F=∫f on E1.Also, for n=1,2,… and t∈E1, set
fn(t)=n[F(t+1n)−F(t)]=F(t+1/n)−F(t)1/n.
Then
fn→F′=fon −Q;
i.e., fn→f (a.e.) on E1 (as mQ=0).
By (3), each fn is bounded and continuous (as F is). Thus by Theorem 1 of Chapter 8, §3, F and all fn are m-measurable on A (even on E1). So is f by Corollary 1 of Chapter 8, §3.
Moreover, by boundedness, F and fn are L-integrable on finite intervals. So is f. For example, let
|f|≤K<∞ on A−P;
as mP=0,
∫A|f|≤∫A(K)=K⋅mA<∞,
proving integrability. Now, as
F=∫f on any interval [t,t+1n],
Corollary 1 in Chapter 5, §4 yields
(∀t∈E1)|F(t+1n)−F(t)|≤supt∈−Q|F′(t)|1n≤Kn.
Hence
|fn(t)|=n|F(t+1n)−F(t)|≤K;
i.e., |fn|≤K for all n.
Thus f and fn satisfy Theorem 5 of Chapter 8, §6, with g=K. By Note 1 there,
limn→∞L∫xafn=L∫xaf.
In the next lemma, we show that also
limn→∞L∫xafn=F(x)−F(a),
which will complete the proof.◻
Lemma 9.1.1
Given a finite continuous F:E1→E and given fn as in (3), we have
limn→∞L∫xafn=F(x)−F(a)for all x∈E1.
- Proof
-
As before, F and fn are bounded, continuous, and L-integrable on any [a,x] or [x,a]. Fixing a, let
H(x)=L∫xaF,x∈E1.
By Theorem 1 and Note 2, H=∫F also in the sense of Chapter 5, §5, with F=H′ (derivative of H) on E1.
Hence by Definition 2 the same section,
∫xaF=H(x)−H(a)=H(x)−0=L∫xaF;
i.e.,
L∫xaF=∫xaF,
and so
L∫xafn(t)dt=n∫xaF(t+1n)dt−n∫xaF(t)dt=n∫b+1/na+1/nF(t)dt−n∫xaF(t)dt.
(We computed
∫F(t+1/n)dt
by Theorem 2 in Chapter 5, §5, with g(t)=t+1/n.) Thus by additivity,
L∫xafn=n∫x+1/na+1/nF−n∫xaF=n∫x+1/nxF−n∫a+1/naF.
But
n∫x+1/nxF=H(x+1n)−H(x)1n→H′(x)=F(x).
Similarly,
limn→∞n∫a+1/naF=F(a).
This combined with (5) proves (4), and hence Theorem 2, too.◻
We also have the following corollary.
Corollary 9.1.1
If f:E1→E∗(En,Cn) is R-integrable on A=[a,b], then
(∀x∈A)R∫xaf=L∫baf=F(x)−F(a),
provided F is primitive to f on A.
- Proof
-
This follows from Theorem 2 by Definition (c) and Theorem 2 of Chapter 8, §9.
Caution. Formulas (2) and (6) may fail if f is unbounded, or if F is not a primitive in the sense of Definition 1 of Chapter 5, §5: We need F′=f on A−Q,Q countable (mQ=0 is not enough!). Even R-integrability (which makes f bounded and a.e. continuous) does not suffice if
F≠∫f.
For examples, see Problems 2-5.
Corollary 9.1.2
If f is relatively continuous and finite on A=[a,b] and has a bounded derivative on A−Q (Q countable), then f′ is L-integrable on A and
L∫xaf′=f(x)−f(a) for x∈A.
This is simply Theorem 2 with F,f,P replaced by f,f′,Q, respectively
Corollary 9.1.3
If in Theorem 2 the primitive
F=∫f
is exact on some B⊆A, then
f(x)=ddxL∫xaf,x∈B.
(Recall that ddxF(x) is classical notation for F′(x).)
- Proof
-
By (2), this holds on B⊆A if F′=f there.◻
II. Note that under the assumptions of Theorem 2,
L∫xaf=F(x)−F(a)=∫xaf.
Thus all laws governing the primitive ∫f apply to L∫f. For example, Theorem 2 of Chapter 5, §5, yields the following corollary.
Corollary 9.1.4 (change of variable)
Let g:E1→E1 be relatively continuous on A=[a,b] and have a bounded derivative on A−Q (Q countable).
Suppose that f:E1→E (real or not) has a primitive on g[A], exact on g[A−Q], and that f is bounded on g[A−Q].
Then f is L-integrable on g[A], the function
(f∘g)g′
is L-integrable on A, and
L∫baf(g(x))g′(x)dx=L∫qpf(y)dy,
where p=g(a) and q=g(b).
For this and other applications of primitives, see Problem 9. However, often a direct approach is stronger (though not simpler), as we illustrate next.
Lemma 9.1.2 (Bonnet)
Suppose f:E1→E1 is ≥0 and monotonically decreasing on A=[a,b]. Then, if g:E1→E1 is L-integrable on A, so also is fg, and
L∫bafg=f(a)⋅L∫cag for some c∈A.
- Proof
The L-integrability of fg follows by Theorem 3 in Chapter 8, §6, as f is monotone and bounded, hence even R-integrable (Corollary 3 in Chapter 8, §9).
Using this and Lemma 1 of the same section, fix for each n a C-partition
Pn={Ani}(i=1,2,…,qn)
of A so that
(∀n)1n>¯S(f,Pn)−S_(f,Pn)=qn∑i=1wnimAni,
where we have set
wni=supf[Ani]−inff[Ani].
Consider any such P={Ai},i=1,…,q (we drop the "n" for brevity). If Ai=[ai−1,ai], then since f↓,
wi=f(ai−1)−f(ai)≥|f(x)−f(ai−1)|,x∈Ai.
Under Lebesgue measure (Problem 8 of Chapter 8, §9), we may set
Ai=[ai−1,ai](∀i)
and still get
L∫Afg=q∑i=1f(ai−1)L∫Aig(x)dx+q∑i=1L∫Ai[f(x)−f(ai−1)]g(x)dx.
(Verify!) Here a0=a and aq=b.
Now, set
G(x)=L∫xag
and rewrite the first sum (call it r or rn) as
r=q∑i=1f(ai−1)[G(ai)−G(ai−1)]=q−1∑i=1G(ai)[f(ai−1)−f(ai)]+G(b)f(aq−1),
or
r=q−1∑i=1G(ai)wi+G(b)f(aq−1),
because f(ai−1)−f(ai)=wi and G(a)=0.
Now, by Theorem 1 (with H,f replaced by G,g), G is continuous on A= [a,b]; so G attains a largest value K and a least value k on A.
As f↓ and f≥0 on A, we have
wi≥0 and f(aq−1)≥0.
Thus, replacing G(b) and G(ai) by K( or k) in (13) and noting that
q−1∑i=1wi=f(a)−f(aq−1),
we obtain
kf(a)≤r≤Kf(a);
more fully, with k=minG[A] and K=maxG[A],
(∀n)kf(a)≤rn≤Kf(a).
Next, let s (or rather sn be the second sum in (12). Noting that
wi≥|f(x)−f(ai−1)|,
suppose first that |g|≤B (bounded) on A.
Then for all n,
|sn|≤qn∑i=1L∫Ani(wniB)=Bqn∑i=1wnimAni<Bn→0 (by (11)).
But by (12),
L∫Afg=rn+sn(∀n).
As sn→0,
L∫Afg=limn→∞rn,
and so by (14),
kf(a)≤L∫Afg≤Kf(a).
By continuity, f(a)G(x) takes on the intermediate value L∫Afg at some c∈A; so
L∫Afg=f(a)G(c)=f(a)L∫cag,
since
G(x)=L∫xaf.
Thus all is proved for a bounded g.
The passage to an unbounded g is achieved by the so-called truncation method described in Problems 12 and 13. (Verify!)◻
Corollary9.1.5 (second law of the mean)
Let f:E1→E1 be monotone on A=[a,b]. Then if g:E1→E1 is L-integrable on A, so also is fg, and
L∫bafg=f(a)L∫cag+f(b)L∫bcgfor some c∈A.
- Proof
-
If, say, f↓ on A, set
h(x)=f(x)−f(b).
Then h≥0 and h↓ on A; so by Lemma 2,
∫bagh=h(a)L∫cagfor some c∈A.
As
h(a)=f(a)−f(b),
this easily implies (15).
If f↑, apply this result to −f to obtain (15) again.◻
Note 3. We may restate (15) as
(∃c∈A)L∫bafg=pL∫cag+qL∫bcg,
provided either
(i) f↑ and p≤f(a+)≤f(b−)≤q, or
(ii) f↓ and p≥f(a+)≥f(b−)≥q.
This statement slightly strengthens (15).
To prove clause (i), redefine
f(a)=p and f(b)=q.
Then still f↑; so (15) applies and yields the desired result. Similarly for (ii). For a continuous g, see also Problem 13(ii') in Chapter 8, §9, based on Stieltjes theory.
III. We now give a useful analogue to the notion of a primitive.
Definition
A map F:E1→E is called an L-primitive or an indefinite L-integral of f:E1→E, on A=[a,b] iff f is L-integrable on A and
F(x)=c+L∫xaf
for all x∈A and some fixed finite c∈E.
Notation:
F=L∫f(not F=∫f)
or
F(x)=L∫f(x)dxon A.
By (16), all L-primitives of f on A differ by finite constants only.
If E=E∗(En,Cn), one can use this concept to lift the boundedness restriction on f in Theorem 2 and the corollaries of this section. The proof will be given in §2. However, for comparison, we state the main theorems already now.
Theorem 9.1.3
Let
F=L∫fon A=[a,b]
for some f:E1→E∗(En,Cn).
Then F is differentiable, with
F′=fa.e. on A.
In classical notation,
f(x)=ddxL∫xaf(t)dtfor almost all x∈A.
A proof was sketched in Problem 6 of Chapter 8, §12. (It is brief but requires more "starred" material than used in §2.)
Theorem 9.1.4
Let F:E1→En(Cn) be differentiable on A=[a,b] (at a and b differentiability may be one sided). Let F′=f be L-integrable on A.
Then
L∫xaf=F(x)−F(a)for all x∈A.