Real Analysis, 2nd Edition, G.B.

Folland
Chapter 2 Integration

Yung-Hsiang Huang∗

October 18, 2018

2.1 Measurable Functons

1. Proof.

2. Proof.

3. Proof.

4. Proof.

5. Proof.

6. Proof.

7. Proof.

8. Proof.

9. Proof.

10. Proof.

11. Proof.
∗
Department of Math., National Taiwan University. Email: [email protected]

1
 2.2 Integration of Nonnegative Functions

12. Proof.

13. Proof.

14. Proof.

15. Proof.

16. Proof.

17. Proof.

2.3 Integration of Complex Functions

18. Proof.

19. Proof.

20. Proof.

21. Proof.

22. Proof.

23. Proof.

24. Proof.

25. Proof.

26. Proof.

27. Proof.

28. Proof.

29. Proof.

30. Proof.

31. Proof.

2
 2.4 Modes of Convergence

32. Proof.

33. Proof.

34. Proof.

35. Proof.

36. Proof.

37. Proof.

38. Proof.

39. Proof.

40. Proof.

41. Proof.

42. Proof.

43. Suppose that µ(X) < ∞ and f : X × [0, 1] → C is a function such that f (·, y) is
measurable for each y ∈ [0, 1] and f (x, ·) is continuous for each x ∈ X.

(a) If 0 < , δ < 1, then E,δ = {x ∈ E : |f (x, y)−f (x, 0)| ≤  for all y < δ} is measurable.

(b) For any  > 0 there is a set E ⊂ X such that µ(E) <  and f (·, y) → f (·, 0)
uniformly on E c as y → 0.

Proof. (a) Let {rk } denotes the set of rational numbers in [0,1]. Given 0 < , δ < 1, we are
going to show
E,δ = ∩ E (rk ) := ∩ {x : |f (x, rk ) − f (x, 0)| ≤ },
rk <δ rk <δ

where the right-hand side is measurable.

(⊆) is trivial. For (⊇), given x ∈ ∩rk <δ E (rk ) and y < δ. Since for each η > 0, by the continuity
assumption, we can choose rk close to y such that

|f (x, y) − f (x, 0)| ≤ |f (x, y) − f (x, rk )| + |f (x, rk ) − f (x, 0)| ≤ η + ,

we see |f (x, y) − f (x, 0)| ≤ .

3
 (b) Fix , η > 0. For each M ∈ N, let EM = Eη, 1 which is measurable and EM ⊂ EM +1 . By
M

continuity assumption, for all x ∈ X f (x, y) → f (x, 0) as y → 0, so µ(EM ) % µ(X) < ∞ and
hence µ(X \EM ) & 0. Choose M0 such that µ(X \EM0 ) <  and on EM0 , |f (x, y)−f (x, 0)| < η
for all y < 1/M0 .

Now given  > 0 and m ∈ N, using the above procedure we can select integers M,m and
measurable sets E 1 ,M,m such that µ(X \ E 1 ,M,m ) < 2−m and |f (x, y) − f (x, 0)| ≤ m1 on
m m

E 1 ,M,m for all y < 1/M,m . Consider E = ∩m E 1 ,M,m , then µ(X \ E) ≤ m 2−m = . Given
P
m m
1 1
m ∈ N, we see for any x ∈ E, |f (x, y) − f (x, 0)| ≤ m
if y < M,m
. This means f (x, y) → f (x, 0)
uniformly on E.

Remark 1. The continuity assumption is not superfluous, see my additional exercises 2.4.1 &
2.4.2 (which follows the works of [2, 3, 1]).

44. Proof.

2.5 Product Measures

45. Proof.

46. Proof.

47. Proof.

48. Proof.

49. Proof.

50. Proof.

51. Proof.

52. Proof.

2.6 The n-dimensional Lebesgue Integral

53. Proof.

54. Proof.

4
55. Proof.

56. Proof.

57. Proof.

58. Proof.

59. Proof.

60. Proof. Z ∞ Z ∞
Γ(α)Γ(β) = tα−1 sβ−1 e−t−s dt ds
0 0

Then use change of variables t = uv, s = u − uv.

61. Proof. (a) By Fubini’s theorem and exercise 60, for each x ∈ [0, ∞)
Z x Z t Z xZ x
1 α−1 1 β−1 1
(x − t) (t − s) f (s) ds dt = (x − t)α−1 (t − s)β−1 f (s) dt ds
Γ(α) 0 Γ(β) 0 Γ(α)Γ(β) 0 s
Z x Z x−s
1
= z α−1 (x − z − s)β−1 dzf (s) ds
Γ(α)Γ(β) 0 0
Z xZ 1 Z x
1 α−1 β−1 α+β−1 1
= w (1 − w) dw(x − s) f (s) ds = (x − s)α+β−1 f (s) ds
Γ(α)Γ(β) 0 0 Γ(α + β) 0

(b)For n = 1, it’s true by Fundamental Theorem of Calculus. Let F be the antiderivative of f .

Assume it’s true for n = j − 1, then
Z x Z x
1 j−1 1
Ij (x) = (x − t) f (t) dt = (j − 1)(x − t)j−2 F (t) dt,
Γ(j) 0 (j − 1)Γ(j − 1) 0

which is the (j −1)th-order antderivative of F , and hence the j th-order ant-derivative of f .

2.7 Integration in Polar Coordinates

62. Proof.

63. Proof.
R R∞
64. Proof. Since Rn |x|a | log |x||b dx = 0 ra+n−1 | log r|b dr, we focus on which b and α := a + n − 1
R∞ R1
make 2 rα | log r|b dr and 02 rα | log r|b dr finite. Moreover, by change of variables, we know
R∞ α b
R 1 −α−2 b
R1 α
2
r | log r| dr = 0
2
t | log t| dt. So we actually focus on when is 0
2
r | log r|b dr finite ?

Since given  > 0, there is δ = δ() ∈ (0, 21 ), such that | log r| ≤ r− for all r ∈ (0, δ).

5
 (i) If α > −1 and b > 0, then
Z δ Z δ
α b
r | log r| dr ≤ rα−b dr < ∞.
0 0

(ii) If α > −1 and b ≤ 0, then

Z 1 Z 1
2 2
α b b
r | log r| dr ≤ (log 2) rα dr < ∞.
0 0

R1
(iii) If α = −1,then 0
2
r−1 | log r|b dr is infinite if b ≥ −1 and finite if b < −1.

(iv) If α < −1 and b ≥ −1, then

Z 1 Z 1
2 2
α
r | log r| dr > b
r−1 | log r|b dr = ∞.
0 0

(v) If α < −1 and b < −1, then

1
Z
2
Z δ
α b
r | log r| dr ≥ r−1 dr = ∞.
0 0

65. Proof.

References
[1] Walter, Wolfgang. A counterexample in connection with Egorov’s theorem. The American
Mathematical Monthly 84.2 (1977): 118-119.

[2] Weston, J. D. A CounterExample Concerning Egoroff ’s Theorem. Journal of the London Math-
ematical Society 1.2 (1959): 139-140.

[3] Weston, J. D. Addendum To a Note on Egoroffs Theorem. Journal of the London Mathematical
Society 1.3 (1960): 366-366.