5.8: Rectifiable Arcs. Absolute Continuity
This page is a draft and is under active development.
( \newcommand{\kernel}{\mathrm{null}\,}\)
If a function f:E1→E is of bounded variation (§7) on an interval I=[a,b], we can define a real function vf on I by
vf(x)=Vf[a,x](= total variation of f on [a,x]) for x∈I;
vf is called the total variation function, or length function, generated by f on I. Note that vf↑ on I. (Why?) We now consider the case where f is also relatively continuous on I, so that the set A=f[I] is a rectifiable arc (see §7, Note 1 and Definition 2).
A function f:E1→E is (weakly) absolutely continuous on I=[a,b] iff Vf[I]<+∞ and f is relatively continuous on I.
The following are equivalent:
(i) f is (weakly) absolutely continuous on I=[a,b];
(ii) vf is finite and relatively continuous on I; and
(iii) (∀ε>0) (∃δ>0) (∀x,y∈I|0≤y−x<δ) Vf[x,y]<ε.
- Proof
-
We shall show that (ii) ⇒ (iii) ⇒ (i) ⇒ (ii).
(ii) ⇒ (iii). As I=[a,b] is compact, (ii) implies that vf is uniformly continuous on I (Theorem 4 of Chapter 4, §8). Thus
(∀ε>0) (∃δ>0) (∀x,y∈I|0≤y−x<δ)vf(y)−vf(x)<ε.
However,
vf(y)−vf(x)=Vf[a,y]−Vf[a,x]=Vf[x,y]
by additivity (Theorem 1 in §7). Thus (iii) follows.
(iii) ⇒ (i). By Corollary 3 of §7, |f(x)−f(y)|≤Vf[x,y]. Therefore, (iii) implies that
(∀ε>0) (∃δ>0) (∀x,y∈I||x−y|<δ)|f(x)−f(y)|<ε,
and so f is relatively (even uniformly) continuous on I.
Now with ε and δ as in (iii), take a partition P={t0,…,tm} of I so fine that
ti−ti−1<δ,i=1,2,…, m.
Then (∀i)Vf[ti−1,ti]<ε. Adding up these m inequalities and using the additivity of Vf, we obtain
Vf[I]=m∑i=1Vf[ti−1,ti]<mε<+∞.
Thus (i) follows, by definition.
That (i) ⇒ (ii) is given as the next theorem. ◻
If Vf[I]<+∞ and if f is relatively continuous at some p∈I (over I=[a,b]), then the same applies to the length function vf.
- Proof
-
We consider left continuity first, with a<p≤b.
Let ε>0. By assumption, there is δ>0 such that
|f(x)−f(p)|<ε2 when |x−p|<δ and x∈[a,p].
Fix any such x. Also, Vf[a,p]=supPS(f,P) over [a,p]. Thus
Vf[a,p]−ε2<k∑i=1|Δif|
for some partition
P={t0=a,…,tk−1,tk=p} of [a,p]. (Why?)
We may assume tk−1=x,x as above. (If tk−1≠x, add x to P.) Then
|Δkf|=|f(p)−f(x)|<ε2,
and hence
Vf[a,p]−ε2<k−1∑i=1|Δif|+|Δkf|<k−1∑i=1|Δif|+ε2≤Vf[a,tk−1]+ε2.
However,
Vf[a,p]=vf(p)
and
Vf[a,tk−1]=Vf[a,x]=vf(x).
Thus (1) yields
|vf(p)−vf(x)|=Vf[a,p]−Vf[a,x]<ε for x∈[a,p] with |x−p|<δ.
This shows that vf is left continuous at p.
Right continuity is proved similarly on noting that
vf(x)−vf(p)=Vf[p,b]−Vf[x,b] for p≤x<b. (Why?)
Thus vf is, indeed, relatively continuous at p. Observe that vf is also of bounded variation on I, being monotone and finite (see Theorem 3(ii) of §7).
This completes the proof of both Theorem 2 and Theorem 1. ◻
We also have the following.
If f is real and absolutely continuous on I=[a,b] (weakly), so are the nondecreasing functions g and h(f=g−h) defined in Theorem 3 of §7.
Indeed, the function g as defined there is simply vf. Thus it is relatively continuous and finite on I by Theorem 1. Hence so also is h=f−g. Both are of bounded variation (monotone!) and hence absolutely continuous (weakly).
Note 1. The proof of Theorem 1 shows that (weak) absolute continuity implies uniform continuity. The converse fails, however (see Problem 1(iv) in §7).
We now apply our theory to antiderivatives (integrals).
If F=∫f on I=[a,b] and if f is bounded (|f|≤K∈E1) on I−Q (Q countable), then F is weakly absolutely continuous on I.
(Actually, even the stronger variety of absolute continuity follows. See Chapter 7, §11, Problem 17).
- Proof
-
By definition, F=∫f is finite and relatively continuous on I, so we only have to show that VF[I]<+∞. This, however, easily follows by Problem 3 of §7 on noting that F′=f on I−S (S countable). Details are left to the reader. ◻
Our next theorem expresses arc length in the form of an integral.
If f:E1→E is continuously differentiable on I=[a,b] (§6), then vf=∫|f′| on I and
Vf[a,b]=∫ba|f′|.
- Proof
-
Let a<p<x≤b,Δx=x−p, and
Δvf=vf(x)−vf(p)=Vf[p,x].(Why?)
As a first step, we shall show that
ΔvfΔx≤sup[p,x]|f′|.
For any partition P={p=t0,…,tm=x} of [p,x], we have
S(f,P)=m∑i=1|Δif|≤m∑i=1sup[ti−1,ti]|f′|(ti−ti−1)≤sup[p,x]|f′|Δx.
Since this holds for any partition P, we have
Vf[p,x]≤sup[p,x]|f′|Δx,
which implies (2).
On the other hand,
Δvf=Vf[p,x]≥|f(x)−f(p)|=|Δf|.
Combining, we get
|ΔfΔx|≤ΔvfΔx≤sup[p,x]|f′|<+∞
since f′ is relatively continuous on [a,b], hence also uniformly continuous and bounded. (Here we assumed a<p<x≤b. However, (3) holds also if a≤x<p<b, with Δvf=−V[x,p] and Δx<0. Verify!)
Now
||f′(p)|−|f′(x)||≤|f′(p)−f′(x)|→0 as x→p,
so, taking limits as x→p, we obtain
limx→pΔvfΔx=|f′(p)|.
Thus vf is differentiable at each p in (a,b), with v′f(p)=|f′(p)|. Also, vf is relatively continuous and finite on [a,b] (by Theorem 1). Hence vf=∫|f′| on [a,b], and we obtain
∫ba|f′|=vf(b)−vf(a)=Vf[a,b], as asserted.◻
Note 2. If the range space E is En (*or Cn), f has n components
f1,f2,…,fn.
By Theorem 5 in §1, f′=(f′1,f′2,…,f′n), so
|f′|=√n∑k=1|f′k|2,
and we get
Vf[a,b]=∫ba√n∑k=1|f′k|2=∫ba√n∑k=1|f′k(t)|2dt(classical notation).
In particular, for complex functions, we have (see Chapter 4, §3, Note 5)
Vf[a,b]=∫ba√f′re(t)2+f′im(t)2dt.
In practice, formula (5) is used when a curve is given parametrically by
xk=fk(t),k=1,2,…, n,
with the fk differentiable on [a,b]. Curves in E2 are often given in nonparametric form as
y=F(x),F:E1→E1.
Here F[I] is not the desired curve but simply a set in E1. To apply (5) here, we first replace "y=F(x)" by suitable parametric equations,
x=f1(t) and y=f2(t);
i.e., we introduce a function f:E1→E, with f=(f1,f2). An obvious (but not the only) way of achieving it is to set
x=f1(t)=t and y=f2(t)=F(t)
so that f′1=1 and f′2=F′. Then formula (5) may be written as
Vf[a,b]=∫ba√1+F′(x)2dx,f(x)=(x,F(x)).
Find the length of the circle
x2+y2=r2.
Here it is convenient to use the parametric equations
x=rcost and y=rsint,
i.e., to define f:E1→E2 by
f(t)=(rcost,rsint),
or, in complex notation,
f(t)=reti.
Then the circle is obtained by letting t vary through [0,2π]. Thus (5) yields
Vf[0,2π]=∫bar√cos2t+sin2tdt=r∫ba1dt=rt|2π0=2rπ.
Note that f describes the same circle A=f[I] over I=[0,4π]. More generally, we could let t vary through any interval [a,b] with b−a≥2π. However, the length, Vf[a,b], would change (depending on b−a). This is because the circle A=f[I] is not a simple arc (see §7, Note 1), so ℓA depends on f and I, and one must be careful in selecting both appropriately.