The Riemann-Stieltjes Integral

Written by Men-Gen Tsai

email: [email protected]

1. Suppose α increases on [a, b], a ≤ x0 ≤ b, α is continuous at x0 , f (x0 ) =

1, and f (x) = 0 if x 6= x0 . Prove that f ∈ R(α) and that
R
f dα = 0.
Proof: Note that L(P, f, α) = 0 for all partition P of [a, b]. Thus
Z a
f dα = 0.
b

Take a partition P such that

1 k n−1
P = {a, a + (b − a), ..., a + (b − a), ..., a + (b − a), b}
n n n
for all N 3 n > 1. Thus
n
X 2(b − a)
U (P, f, α) = Mi ∆αi ≤
i=1 n

for all N 3 n > 1. Thus

2(b − a)
0 ≤ inf U (P, f, α) ≤
n
Ra
for all n ∈ N . Thus inf U (P, f, α) = 0. Hence b f dα = 0; thus,
Ra
b f dα = 0.

Rb
2. Suppose f ≥ 0, f is continuous on [a, b], and a f (x)dx = 0. Prove that
f (x) = 0 for all x ∈ [a, b]. (Compare this with Exercise 1.)
Proof: Suppose not, then there is p ∈ [a, b] such that f (p) > 0. Since
f is continuous at x = p, for  = f (p)/2, there exist δ > 0 such that
|f (x) − f (p)| <  whenever x ∈ (x − δ, x + δ) [a, b], that is,
T

1 3
0 < f (p) < f (x) < f (p)
2 2
1
 for x ∈ Br (p) ⊂ [a, b] where r is small enough. Next, consider a
partition P of [a, b] such that
r r
P = {a, p − , p + , b}.
2 2
Thus
1 rf (p)
L(P, f ) ≥ r · f (p) = .
2 2
Thus
rf (p)
sup L(P, f ) ≥ L(P, f ) ≥ > 0,
2
a contradition since ab f (x)dx = sup L(P, f ) = 0. Hence f = 0 for all
R

x ∈ [a, b].
Note: The above conclusion holds under the condition that f is con-
tinuous. If f is not necessary continuous, then we cannot get this
conclusion. (A counter-example is shown in Exercise 6.1).

3.

4. If f (x) = 0 for all irrational x, f (x) = 1 for all rational x, prove that
f∈
/ R on [a, b] for any a < b.
Proof: Take any partition P of [a, b], say

a = x0 ≤ x1 ≤ .. ≤ xn−1 ≤ xn = b.

By P we can construct the new partition P 0 without repeated points,

and U (P, f ) = U (P 0 f ), L(P, f ) = L(P 0 , f ). Say P 0

a = y0 < y1 < ... < ym−1 < ym = b.

Note that Q is dense in R, and R − Q is also dense in R. Hence

Mi = sup f (x) = 1,
yi−1 ≤x≤yi

mi = inf f (x) = 0.
yi−1 ≤x≤yi

2
 Hence
m
0
X
U (P, f ) = U (P , f ) = Mi ∆yi = b − a,
i=1
m
L(P, f ) = L(P 0 , f ) =
X
mi ∆yi = 0.
i=1

for any partition of [a, b]. Hence

Z b
f dx = inf U (P, f ) = b − a,
a
Z b
f dx = sup L(P, f ) = 0.
a

Thus f ∈
/ R[a, b].

5. Suppose f is bounded real function on [a, b], and f 2 ∈ R on [a, b].

Does it follow that f ∈ R? Does the answer change if we assume that
f 3 ∈ R?
Proof: The first answer is NO. Define f (x) = −1 for all irrational
x ∈ [a, b], f (x) = 1 for all rational x ∈ [a, b]. Similarly, by Exercise 6.4,
f∈
/ R.
However, if we assume that f 3 ∈ R, the answer is YES. Let Φ = x1/3 ,
we apply Theorem 6.11 and get the conclusion that f ∈ R.

6.

7. Suppose f is a real funtion on (0, 1] and f ∈ R on [c, 1] for every c > 0.

Define Z 1 Z 1
f (x)dx = lim f (x)dx
0 c→0 c

if this limit exists (and is finite).

(a) If f ∈ R on [0, 1], show that this definition of the integral agrees
with the old one.

3
 (b) Construct a function f such that the above limit exists, although it
fails to exist with |f | in place of f .
Proof of (a): Take any c > 0, and note that f is bounded, say |f | ≤ M
for some M . Hence
Z c Rc
Z c
−M dx ≤ 0 f (x)dx ≤ M dx
0 Rc 0
−cM ≤ 0 f (x)dx ≤ cM
R1 R1
−cM ≤ c f (x)dx − 0 f (x)dx ≤ cM.
R1 R1
Letting c → 0 and thus f (x)dx → 0, that is, c1 f (x)dx →
R
c f (x)dx− 0
R1
0 f (x)dx as c → 0. Thus this definition of the integral agrees with the
old one.
Solution of (b): (Due to Shin-Yi Lee) Define
1 1
f (x) = n(−1)n , where <x≤ .
n+1 n
(Due to Meng-Gen Tsai). Define
sin(1/x)
f (x) = .
x
8. Suppose f ∈ R on [a, b] for every b > a where a is fixed. Define
Z ∞ Z b
f (x)dx = lim f (x)dx
a b→∞ a

if this limit exists (and is finite). In that case, we say that the integral
on the left converges. If it also converges after f has been replaced by
|f |, it is said to converge absolutely.
Assume that f (x) ≥ 0 and that f decreases monotonically on [1, ∞).
Prove that Z ∞
f (x)dx
1
converges if and only if
∞
X
f (n)
n=1

4
 converges. (This is the so-called ”integral test” for convergence of se-
ries.)
Proof: Since f is nonnegative and decreases monotonically on [1, ∞),
we have for n ≥ 2,
k
X Z k k−1
X
f (n) ≤ f (x)dx ≤ f (n).
n=2 1 n=1

We define Z y
g(x) = f (t)dt.
1
If this integral converges, then g(x) is a nondecreasing function which
tends to a limit, and so g(k) is a bounded nondecreasing sequence.
Pk
Thus denote sk = n=2 f (n), we see that sk ≤ g(k), and sk tends to a
P∞ R∞
limit. The fact that n=1 f (n) converges if 1 f (x)dx converges is now
established. If the integral diverges, then g(x) → +∞ as x → +∞.
Pk−1
Therefore g(k) → +∞. Since g(k) ≤ n=1 f (n), we conclude that
R∞ P∞
the series diverges. Hence 1 f (x)dx converges if n=1 f (n). Thus,
R∞ P∞
1 f (x)dx converges if and only if n=1 f (n) converges.

9. Show that integration by parts can sometimes be applied to the ”im-

proper” integrals defined in Exercise 7 and 8. (State appropriate hy-
potheses, formulate a theorem, and prove it.) For instance show that
Z ∞ cos x Z ∞
sin x
dx = dx.
0 1+x 0 (1 + x)2
Show that one of these integrals converges absolutely, but that the other
does not.
Proof: Recall Theorem 6.22 (integration by parts):

Suppose F and G are differentiable functions on [a, b], F 0 =

f ∈ R, and G0 = g ∈ R. Then
Z b Z b
F (x)g(x)dx = F (b)G(b) − F (a)G(a) − f (x)G(x)dx.
a a

5
10.

11.

12.

13. Define Z x+1

f (x) = sin(t2 )dt.
x
(a) Prove that |f (x)| < 1/x if x > 0.
Hint: Put t2 = u and integrate by parts, to show that f (x) is equal to

cos(x2 ) cos[(x + 1)2 ] Z (x+1)2 cos u

− − du.
2x 2(x + 1) x2 4u3/2

Replace cos u by −1.

(b) Prove that

2xf (x) = cos(x2 ) − cos[(x + 1)2 ] + r(x)

where |r(x)| < c/x and c is a constant.

(c) Find the upper and lower bound limit of xf (x), as x → ∞.
R∞
(d) Does 0 sin(t2 ) dt converges?
Proof of (a): Put t2 = u,
Z x+1
f (x) = sin(t2 )dt
x
Z (x+1)2 sin u
= du
x2 2u1/2
− cos u u=(x+1)2 Z (x+1)2 cos u
= | 2 − du
2u1/2 u=x x2 4u3/2
cos(x2 ) cos[(x + 1)2 ] Z (x+1)2 cos u
= − − du.
2x 2(x + 1) x2 4u3/2

6
 R (x+1)2 cos u
To get a bound of x2 4u3/2
du, we replace cos u by −1:
Z (x+1)2 cos u Z (x+1)2
−1
| 3/2
du| < | du|
x2 4u x 2 4u3/2
1
=
2x(x + 1)

for x > 0. Hence

cos(x2 ) cos[(x + 1)2 ] Z (x+1)2
cos u
|f (x)| ≤ | |+| |+| du|
2x 2(x + 1) x 2 4u3/2
1 1 1
< + +
2x 2(x + 1) 2x(x + 1)
1
=
x
for x > 0.
Proof of (b): By (a),

cos(x2 ) cos[(x + 1)2 ] Z (x+1)2 cos u

f (x) = − − du.
2x 2(x + 1) x2 4u3/2
Z (x+1)2
1 cos u
2xf (x) = cos(x2 )−cos[(x+1)2 ]+ cos[(x+1)2 ]−2x du.
x+1 x2 4u3/2
Let Z (x+1)2
1 2 cos u
r(x) = cos[(x + 1) ] − 2x du.
x+1 x 2 4u3/2
Note that
1 1 1
| cos[(x + 1)2 ]| ≤ < ,
x+1 x+1 x
Z (x+1)2
cos u 1 1
2x 3/2
du < 2x <
x 2 4u 2x(x + 1) x
if x > 0. Thus |r(x)| < 1/x + 1/x = c/x and c = 2 is a constant.
Note: Here, x > 0.
Proof of (c):
Proof of (d):

7
14.

15.

16. For 1 < s < ∞, define

∞
X 1
ζ(s) = .
n=1 ns
(This is Riemann’s zeta function, of great importance in the study of
the distribution of prime numbers.) Prove that
R∞ [x]
(a) ζ(s) = s 1 xs−1
dx
and that
s R∞ x−[x]
(b) ζ(s) = s−1
−s 1 xs−1
dx,
where [x] denotes the greatest integer ≤ x.
Prove that the integral in (b) converges for all s > 0.
Hint: To prove (a), compute the difference between the integral over
[1, N ] and the N th partial sum of the series that defines ζ(s).

17.

18.

19.