Summer Jump-Start Program for Analysis, 2012

Song-Ying Li

1 Lecture 11: Riemann Integrals

1.1 Definition of Riemann integrals
Let f (x) be a bounded function on a bounded closed integral [a, b]. First, we
consider a partition P for [a, b] as follows:

P : a = x0 < x1 < .... < xn = b.

with the norm of the partition P defined as

||P || = max{∆xi , i = 1, 2, ...., n} and ∆xi = xi − xi−1 .

Let

mi (f ) = inf{f (x) : xi−1 ≤ x < xi } and Mi = sup{f (x) : xi−1 ≤ x < xi }.

be the infimum and supremum of f on the subinterval [xi−1 , xi ). Then we

define the lower partial sum and the upper partial sum of f with respect to the
partition P as follows:
n
X n
X
L(f, P ) = mi ∆xi ; and U (f, P ) = Mi ∆xi .
i=1 i=1

It is easy to see that L(f, P ) ≤ U (f, P ).

REMARK 1 In fact, for any two partitions P1 and P2 for [a, b], we always
have L(f, P1 ) ≤ U (f, P2 ).

Proof. Let P1 : a = x0 < x1 < ... < xn = b and P2 : a = y0 < y1 < .... <
ym = b. We define a refinement for P1 and P2 which is partition P for [a, b] with
end points of subintervals are x0 , ..., xn and y0 , ..., ym . We denote P = P1 ∪ P2 .
Then
L(f, P1 ) ≤ L(f, P ) ≤ U (f, P ) ≤ U (f, P2 ).
We define the lower and upper integrals of f on [a, b] as follows:

L(f ) = lim L(f, P ) and U (F ) = lim U (f, P ).

||P ||→0 ||P ||→0

The both are always exist.

Definition 1.1 Let f be a bounded function on [a, b]. Then we say that f is
(Riemann) integrable on [a, b] if L(f ) = U (f ).

1
1.2 Tests for Riemann-integrability
THEOREM 1.2 Let f be a bounded function on [a, b]. Then f is integrable
on [a, b] if and only if for any  > 0, there is a partition P for [a, b] such that
U (f, P ) − L(f, P ) < .

Proof. “⇒ ” Assume f R-integrable. Then, L(f ) = U (f ). For  > 0, there

is P1 and P2 such that −L(f, P1 ) + L(f ) < /2 and U (f, P2 ) − U (f ) < /2 Since
L(f ) = U (f ) and P = P1 ∪ P2 , we have U (f, P ) − L(f, P ) < .
“ ⇐ ” Since L(f ) ≥ L(f, P ), U (f ) ≤ U (f, P ). If for any  > 0, there is P
such that U (f, P ) − L(f, P ) < , then U (f ) − L(f ) < . But, U (f ) − L(f ) ≥ 0
is a fixed number and U (f ) − L(f ) <  for any  > 0, then U (f ) − L(f ) = 0.
Since U (f ) = L(f ), f is integrable.

EXAMPLE 1 If f is continuous on [a, b], then f is integrable.

Proof. For any  > 0, since f is continuous on [a, b], it is uniformly continuous
on [a, b]. So, there is δ > 0 such that if x1 , x2 ∈ [a, b] and |x1 − x2 | < δ, then
|f (x1 ) − f (x2 )| < /(b − a). Now, choose n such that b−a n < δ. Let P : x0 =
b−a
a < x1 < .... < xn =P b, where xj = xj−1 + n .PThen, Mj − mj < /(b − a).
n  n
U (f, P ) − L(f, P ) ≤ j=1 (Mj − mj )∆xj < b−a i=1 ∆xi = 

EXAMPLE 2 Let

1 x is rational
D(x) = f ((x) =
0 x irrational

D(x) is not integrable on any interval [a, b]. In fact, U (f ) = (b − a), L(f ) = 0.

Note: The Dirichlet function D(x) is the typical examples of f which is not
integrable on [a, b].

EXAMPLE 3 Let

0 x irrational
R(x) =
1/n x = m/n, (m, n) = 1, x ∈ [0, 1]

Then R(x) is integrable on [0, 1].

Exercise: Prove it on your own!

EXAMPLE 4 If f (x) is a monotone increasing function on [a, b], then it is

integrable.

Proof. Let P : a = x0 < x1 < ... < xn = b be any partition for [a, b]. Since
f is increasing, we have

mi = inf{f (x) : xi−1 ≤ x < xi } = f (xi−1 )

2
and
Mi = sup{f (x) : xi−1 ≤ x < xi } = f (xi − 0) ≤ f (xi ).
Since
M1 − m1 + M2 − m2 + · · · + Mn − mn ≤ Mn − m1 ≤ f (b) − f (a),
we have
X n
X
U (f, P ) − L(f, P ) = (Mi − mi )∆xi ≤ ||P || (Mi − mi ) ≤ (f (b) − f (a))||P ||
i=1

For  > 0, and for any partition P for [a, b] with ||P || < f (b)−f (a) , we have

U (f, P ) − L(f, P ) < (f (b) − f (a)) f (b)−f (a) = . So, f is integrable on [a, b].
• Let f be bounded on [a, b]. Let Ds (f ) be the set of all points x0 ∈ [a, b] with
f is discontinuous at x0 .
THEOREM 1.3 Let f be bounded on [a, b]. Then f is integrable on [a, b] if
and only if m(Ds ) = 0 (Ds has measure 0).
• We say that Ds has measure S∞0 if: For  > 0, Pthere is a sequence of intervals
∞
{(ai , bi )}∞
i=1 such that D s ⊂ j=1 (aj , bj ) and j=1 (b j − aj ) < 0 .

1.3 Improper Integrals

THEOREM 1.4 Let f be bounded on [a, b]. Then f is integrable on [a, b] if
and only if f + and f − are integrable on [a, b],where
 
+ f (x), f (x) ≥ 0 − −f (x), f (x) < 0
f (x) = and f (x) =
0, f (x) < 0 0, f (x) ≥ 0
Corollary 1.5 If f is integrable on [a, b], then |f | is integrable on [a, b]. Con-
verse is not true.
EXAMPLE 5

1 if x is rational
f (x) =
−1 if x is irrational
f is not integrable on [a, b] since L(f, P ) = −(b − a) and U (f, p) = (b − a), but
|f (x)| = 1 is integrable on [a, b].
Next we consider the integrality for some unbounded function on [a, b].
Definition 1.6 Let f (x) be a function on (a, b) (usually f is unbounded). As-
sume for any a < c < d < b, we have f (x) is integrable on [c, d]. We say that f
is integrable on (a, b) If
Z b Z d
f (x)dx =: lim f (x)dx (exists)
a c→a,d→b c

3
Proposition 1.7 If f is a function on (a, b) and f is integral on [c1 , c2 ] for any
a < c1 < c2 < b and if |f | is integrable on (a, b), then f is integrable on (a, b).
The converse is not true.

EXAMPLE 6 Let f (x) = x1α on (0, 1]. Then

(i) f is integrable on (0, 1] when α < 1;
(ii) f is not integrable on (0, 1] when α ≥ 1.

Proof. Since
Z 1 Z 1
1 1 1 1 1−α 1
f (x)dx = dx =  =  − (α 6= 1).
  xα 1−α α−1 α−1

we have
Z 1 
1 1 1 1−α < ∞, if α < 1;
lim+ dx = − + lim+  =
→0  x α α − 1 →0 α − 1 +∞, if α > 1

and Z 1
1
dx = +∞.
0 x

THEOREM 1.8 (Comparison test) If f (x) and g(x) are continuous on (a, b)
and |g(x)| integrable on (a, b) and |f (x) ≤ |g(x)|, x ∈ [a, b], then |f (x)| is
integrable on (a, b) and so is f (x).
R1 1
EXAMPLE 7 Prove or disprove: 0 x1/2
sin(1/x)dx converges.
1
Proof. Converges! Since x1/2
sin(1/x) is continuous on (0, 1],

1 1
sin(1/x) ≤ , x ∈ (0, 1],
x1/2 x1/2
R1 1
and we know that 0 x1/2 < +∞. By comparison test, we have
R1 1
0 x1/2
sin(1/x)dx converges.
R1 sin(x)
EXAMPLE 8 Determine if 0 x3/2
dx converges.

Solution. Since
sin(x) x 1
sin(x) ≤ x, for x ∈ [0, π/2], 0≤ 3/2
≤ 3/2 = 1/2
x x x
R1 1
R1 sin(x)
and 0 x1/2
< +∞. By comparison test, 0 x3/2
dx converges.
R1 cos(x)
EXAMPLE 9 0 x3/2
dx.

4
Solution. Diverges. Since
cos x cos 1
≥ 3/2 , x ∈ [0, 1].
x3/2 x
R1 cos 1
R1 cos x
We know 0 x3/2
dx = +∞. By comparison test, 0 x3/2
dx = +∞.
Rb Rb
Definition 1.9 (i) a f (x)dx = limb→+∞ a f (x)dx.
Rb Rb
(ii) −∞ f (x)dx = lima→−∞ a f (x)dx.
R∞ Rb
(iii) ∞ f (x)dx = lima→−∞,b→∞ a f (x)dx.
1
EXAMPLE 10 f (x) = xα , x ∈ [1, ∞).
Z ∞ 
1 finite α>1
dx =
1 xα +∞ α≤1
R∞ sin(x)
EXAMPLE 11 Determine if 0 ln(x+2) dx converges.

Solution
Z ∞ Z ∞ 0
sin(x) cos(x) ∞ 1
dx = − |x=0 + cos(x)dx
0 ln(x + 2) ln(x + 2) 0 ln(x + 2)
Z ∞
1 −1 1
= + 2 x+2
cos xdx
ln 2 0 (ln(x + 1))

We know that if p > 1 then

Z ∞ Z ∞
1 1 1
dx = d ln x
2 x (ln x)p 2 (ln x)p
1
= − (ln x)1−p |∞x=2
p−1
1
= (ln 2)1−p
p−2
< +∞.
−1 1
Since (ln(x+2))2 x+2 cos(x) is continuous on (0, ∞) and
∞
−1
Z
1
2 x+2
cos(x) dx
0 (ln(x + 2)
Z ∞
1 1
≤ dx
0 x + 2 ln(x + 2)2
= −(ln(x + 2))−1 |∞
x=0
= (ln 2)−1 < +∞

5
 R∞ −1 1
R∞ sin(x)
So 0 (ln(x+2))2 x+2 cos xdx converges and so does 0 ln(x+2) .

• Important techniques for integrations:

Rb
(1) Integration by parts: f (x)dx = f (x)g(x)|ba − A f 0 (x)g(x)dx.
(2) Substituion Φ(t) : [a, b] → [c, d] is onto, then, by letting y = Φ(x),
Z b Z d
0
f (Φ(x))Φ (x)dx = f (y)dy.
a c

THEOREM Rx 1.10 (Newton-Leibniz Theorem) If f (x) is integrable on [a, b] and

F (x) = a f (t)dt. Then F 0 (x0 ) = f (x0 ) when f is continuous at x0 .
Rx
Note: F = a f (t)dt is called an anti-derivative of f on [a, b].
R x2 −sin(x) 2
EXAMPLE 12 Let F (x) = x−cos(x)
et (t + 2)dt. Find F 0 (x).

Solution Since
R x2 −sin(x) t2 R0 2
F (x) = 0 e (t + 2d t + x−cos(x) et (t + 2)dt
R x2 −sin(x) t2 R x−cos(x) t2
= 0 e (t + 2)dt − 0 e (t + 2)dt,

we have
#0 #0
x2 −sin(x)
"Z "Z
x−cos(x)
0 t2 t2
F (x) = e (t + 2)dt − e (t + 2)dt
0 0
2
−sin(x))2 2
= e(x (x2 − sin(x) + 2)(x2 − sin x)0 − e(x−cos(x)) (x − cos x + 2)(x − cos x)0
2
−sin(x))2 2
= e(x (x2 − sin(x) + 2)(2x − cos x) − e(x−cos(x)) (x − cos x + 2)(1 + sin x).

1.4 Important inequalities

THEOREM 1.11 Let f and g are (bounded) integrable on [a, b]. Then
(i) f g is integrable on [a, b]
(ii) (Cauchy-Schwarz inequality:)
Z b !2 Z
b Z b
2
f (x)f (x)dx ≤ f (x) dx g(x)2 dx.
a a a

Proof. (i) (Exercise for integrability). (ii) We consider:

Z b
0 ≤ (f (x) + λg(x))2 dx
a
Z b Z b Z b
= f (x)dx + 2( f (x)g(x)dx)λ + ( g(x)2 dx)λ2 , λ ∈ (−∞, ∞).
a a a

6
Therefore,
Z b Z b Z b
[2 f (x)g(x)dx]2 − 4[ g(x)2 dx][ f (x)dx] ≤ 0
a a a

This implies
Z b 2 Z b Z b
f (x)g(x)dx ≤ f (x)2 dx g(x)2 dx
a a a

This completes the proof of Part (ii).

THEOREM 1.12 (Hölder’s inequality): Assume f, g are (bounded) integrable

on [a, b]. If 1 ≤ p, q < ∞ and p1 + 1q = 1, then f g is integrable on [a, b] and
Z b Z b Z b
f (x)g(x)dx ≤ ( |f (x)|p dx)1/p ( |g(x)|q )1/q .
a a a

Recall: (i) xy ≤ 21 (x2 + y 2 ).

p yq
(ii) if p1 + 1q = 1, then xy ≤ xp + q for x, y > 0.
1
ln xp + q1 ln y q 1 ln xp q
[Since xy = e p ≤ pe + 1q eln x
(ex is convex) = p1 xp + 1q y q .]
Proof. WLOG, we assume f (x) ≥ 0, g(x) ≥ 0. We divide it into two cases.
Rb Rb
Case 1: Assume that a f (x)p dx = 1 and a g(x)q dx = 1. We need to
Rb
prove: a f (x)g(x)dx ≤ 1. Since
Z b Z b
1 1
f (x)g(x)dx ≤ f (x)p + g(x)q dx
a a p q
1 b 1 b
Z Z
= f (x)p dx + g(x)q dx
p a q a
1 1
= · 1 + · 1 = 1.
p q
Case 2: General case: Let

f (x) g(x)
F (x) = R b , G(x) = R b .
p
( a f (x) dx) 1/p ( a g(x)q dx)1/q
Z b Z b Z b
p 1 Rb
Then F (x) dx = Rb dx f (x)p dx = 1 and a G(x)q dx = 1.
a a f (x) p a
Rb a
By Case 1, we have a F (x)G(x)dx ≤ 1, This implies that
Z b Z b 1/p  Z b 1/q
f (x)g(x)dx ≤ f (x)p dx g(x)q dx .
a a a

7
1.5 Exercise
1. Let R(x) be the Riemann function defined as R(x) = 0 when x is irrational
or 0, and R(x) = 1/n if x = m/n with (m, n) = 1, where n is positive integer.
R1
Prove R(x) is integrable on [0, 1] and find 0 R(x)dx.
2. Assume that f (x) is a non-negative continuous function on [a, b]. If
Rb
a
f (x)dx = 0, prove f (x) = 0 on [a, b].
3. Assume that f (x) is a real valued function on [0, 1]. Answer the following
questions
(i) If f (x)2 is integrable on [0, 1], is f (x) integrable on [0, 1]?
(ii) If f (x)3 is integrable on [0, 1], is f (x) integrable on [0, 1]?
(iii) If f is integrable and bounded, is f (x)2 integrable on [0, 1]?
4. Let f (x) and g(x) be two integrable functions on [a, b] (both functions are
bounded on [a, b]). Then one has the following Hölder inequality:
Z b Z b 1/p  Z b 1/p0
0
f (x)g(x)dx ≤ |f (x)|p dx |f (x)|p dx
a a a

where 1 ≤ p, p0 < ∞ and 1/p + 1/p0 = 1.

5. Prove
sin(x)
f (x) =
ln(2 + x)
is integrable on [0, ∞).
6. Prove Z ∞
sin(nx)
lim dx = 0
n→∞ 0 1+x
7. ProveRor disprove the Rfollowing improper
R 1 integrals converge:
∞ sin t ∞ cos t R1
(a) 0 t3/2 dt, (b) 0 t dt (c) 0 sin( t )dt, (d) 0 1t sin( 1t )dt.
1

8. Define Z x+1
f (x) = sin(t2 )dt
x

(a) Prove |f (x)| < 1/x if x > 0

(b) Prove that

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

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

R ∞ upper and lower limits of xf (x) as x → ∞.
(c) Find the
(d) Does 0 sin(t2 )dt converge ?

8
9. Mean Value Theorem for integral: Let f (x) is continuous on [a, b], and g(x)
is non-negative and integrable on [a, b]. Prove that there is x0 ∈ (a, b) so that
Z b Z b
f (x)g(x)dx = f (x0 ) g(x)dx
a a

19. Suppose f is real, continuously differentiable function on [a, b], f (a) =

f (b) = 0, and
Z b
f (x)2 dx = 1.
a
Prove that Z b
xf (x)f 0 (x)dx = −1/2
a
and that Z b Z b
[f 0 (x)]2 dx · x2 f (x)2 dx > 1/4.
a a

11. Let α be a fixed increasing function on [a, b]. For integral function u on
[a, b], define
Z b
2
kuk2 = |u(x)|2 dx.
a
Suppose that f , g and h are Riemann-Stieltjes integrable with respect to α,
prove the following triangle inequality:

kf − hk2 ≤ kf − gk2 + kg − hk2 .