PROBLEMS IN MEASURE THEORY

MANJUNATH KRISHNAPUR

Note: These are largely problems I put out when teaching the course on measure and integra-
tion at IISc during Jan-Apr 2018. While a few are made up, most are taken them from various
books, for example, from Bogachev’s and Dudley’s well-known books on the subject. Also I bor-
rowed freely from Kavi Ramamurthy’s problem sets (http://www.isibang.ac.in/˜adean/
infsys/database/index.htm)

Problem 1. In each of the following cases, find σ(S) and A(S) (the sigma-algebra and algebra
generated by S).
(1) X is a set and S is the collection of all singleton subsets of X.
(2) X is a set and S is the collection of all two-element subsets of X.
S
(3) A1 , A2 , . . . are pairwise disjoint sets of X such that An = X.
n
(4) Do the previous exercise if Ai are pairwise disjoint but their union may not equal X.

Problem 2. Let F be a sigma algebra on X. Let A1 , A2 , . . . be elements of F. Show that the

following sets are also in F (first express the set in proper notation).
(1) The set of x ∈ X that belong to exactly five of the An s.
(2) The set of x ∈ X that belong to all except five of the An s.
(3) The set of x ∈ X that belong to infinitely many of the An s.
(4) The set of x ∈ X that belong to all but finitely many of the An s.

Problem 3. Decide whether the following statements are true or false and justify your answer.
(1) A finite union of σ-algebras is not necessarily a σ-algebra.
S
(2) If F1 ⊆ F2 ⊆ . . . is an increasing sequence of sigma-algebras of a set X, then F := n Fn is
also a sigma-algebra.
(3) Let F be a sigma-algebra on X and let T : X 7→ Y be a function. Then G := {T (A) : A ∈ F}
is a sigma-algebra on Y .
(4) Let F be a sigma-algebra on X and let T : Y 7→ X be a function. Then G := {T −1 (A) : A ∈
F} is a sigma-algebra on Y . Here T −1 (A) = {y ∈ Y : T (y) ∈ A}.
(5) There is no sigma-algebra with exactly 1000 elements.
1
Problem 4. Let F = σ(S) where S is a collection of subsets of X. Suppose a, b ∈ X are such that
every set in S either contains both a and b or does not contain either. Then show that the same
property holds for sets in F.

Problem 5. Consider the outer Lebesgue measure λ∗2 on R2 . Show that [a, b] × [c, d] in Lebesgue
measurable by checking the Carathéodary cut condition.

Problem 6. Suppose A ⊆ B are (Lebesgue) measurable subsets of R and λ(A) = λ(B). Then show
that any set C such that A ⊆ C ⊆ B is also measurable and that λ(C) = λ(A).

Problem 7. (*) Let (X, d) be a compact metric space. Fix α > 0 and define
 
X∞ [ 
µ∗α (A) := lim inf dia(Bj )α : Bj are open balls in X such that dia(Bj ) < δ and Bj ⊇ A .
δ↓0  
k=1 j

Show that µ∗α is an outer measure.

When X = R with the usual metric, µ∗α is the same as λ∗ (the Lebesgue outer measure) when
α = 1. Further, µ∗α is zero if α > 1 and µ∗α (A) = ∞ if α < 1 and A is any interval.

Problem 8. Let (X, F, µ) be a measure space.

(1) If An , A ∈ F and An ↑ A, show that µ(An ) ↑ µ(A).
(2) If An , A ∈ F and An ↓ A and µ(An ) < ∞ for some n, then show that µ(An ) ↓ µ(A).
(3) Show that the second conclusion may fail if µ(An ) = ∞ for all n.

Problem 9. A measure µ on (X, F) is said to be σ-finite if there exist E1 , E2 , . . . in F such that

S
X = n En and µ(En ) < ∞ for all n.
(1) Show that a σ-finite measure space has sets of arbitrarily high but finite measure.
(2) Show that a σ-finite measure has at most countably many atoms. Show that the previous
assertion is false without the σ-finiteness assumption.

Problem 10. Let F = σ(S) be a sigma algebra on X. Show that for any A ∈ F, there exists
countably many sets A1 , A2 , . . . in S such that A ∈ σ({A1 , A2 , . . .}).

Problem 11. Let F be a sigma algebra on X and assume that B ⊆ X is not in F. Show that the
smallest sigma algebra containing F and B is the collection of all sets of the form (A1 ∩ B) ∪ (A2 ∩
B c ) where A1 , A2 ∈ F.
(*) If µ is a measure on F, can you extend it to G in some way? Is the extension unique?
2
Problem 12. Let N = {1, 2, 3, . . .} and let k.N = {k, 2k, 3k, . . .}. What is the sigma algebra gener-
ated by (1) the collection k.N, k ≥ 1? (2) the collection p.N as p runs over all primes?

Problem 13. (*) Show that any convex set in Rd is (Lebesgue) measurable. Is it necessarily Borel
measurable?

Problem 14. If A ⊆ R, then there exists a Lebesgue measurable set B that contains A and such
that λ(B) = λ∗ (A). Can we choose B to be Borel measurable?

Problem 15. (*) If A ⊆ R is measurable and has positive Lebesgue measure, show that it contains
a three-term arithmetic progression, i.e., there exist a, b ∈ A such that 21 (a + b) ∈ A.

Problem 16. Let E be a measurable subset of R. Let L(E) be the supremum of all t such that there
is a surjective Lip(1) function f : E 7→ [0, t]. Show that L(E) = λ(E).

Problem 17. Let A0 ⊆ A be two algebras on X that genberate the same sigma algebra (call it F).
Now suppose we have a countably additive measure µ on A and let µ0 be the restriction of µ to A0 .
By the theorem proved in class, both µ0 and mu extend to F as measures. Are the Carathéodary
sigma algebras the same? If yes, are the extended measures equal? On the way, is the outer
measure constructed from µ and µ0 the same?
As a particular case, if we start with a measure µ on a sigma algebra F, and extend it (since F
is also an algebra), what is the extended sigma algebra? Is it F or is it larger?

Problem 18. Let µ be a Borel measure on R such that µ{x} = 0 for each x ∈ R. Show that for any
 > 0, there exists a dense open set U ⊆ R such that µ(U ) < .

Problem 19. Let X = {0, 1}N be the sequence space of zeros and ones. An element ω ∈ X is
written as ω = (ω1 , ω2 , . . .).
(1) A cylinder set is one defined by specifying the values of finitely many co-ordinates. Eg.,
{ω : ω1 = 0, ω2 = 1, ω7 = 1}. Show that the complement of a cylinder set is a finite union
of pairwise disjoint cylinder sets. Use this to describe A(S) as the collection of all finite
unions of pairwise disjoint cylinder sets.
(2) If A is a cylinder set for which exactly n co-ordinate values are specified, define µ0 (A) =
2−n . Extend in the obvious way to A(S) and show that mu0 is countably additive on A(S).
(3) Argue that there is a measure µ on σ(S) that extends µ0 .
[Note: This exercise is to make precise the notion of a infinite sequence of fair coin tosses.]
3
Problem 20. Let A and B be measurable subsets of the line.
(1) Show that λ(A + B) ≥ λ(A) + λ(B). [Note: The discrete version of this is: If A and B are
subsets of Z, then |A + B| ≥ |A| + |B| − 1, where |A| denotes the cardinality of A.]
(2) Show that there can be no reverse inequality by constructing A, B having zero measure but
such that A + B = R.

Problem 21. Let A, B be measurable subsets of the real line.

(1) If A and B have positive measure, show that A − B contains an interval.
(2) If xn + A = A for a sequence xn → 0, then show that λ(A) = 0 or λ(Ac ) = 0.

Problem 22. Let A be a bounded subset of R with positive Lebesgue measure. Given are  > 0
and a sequence δn converging to zero. Show that there exists B ⊆ A and a subseequnce δnk such
that (a) B is measurable and λ(B) > λ(A) − , (b) if x ∈ B then x ± δnk ∈ B for all k.

Problem 23. Let A be a measurable subset of R. If λ1 (A) > 0, show that (A+Q)c has zero Lebesgue
measure.

Problem 24. If K is a compact subset of Rd , the the set {x ∈ Rd : d(x, K) = 1} has zero Lebesgue
measure.

Problem 25. Let A be a measurable subset of R2 with positive Lebesgue measure. Show that there
is a measurable set B having positive measure and a number δ > 0 such that B, B + (δ, 0), B +
(0, δ), B + (δ, δ) are all subsets of A.

Problem 26. If A ⊆ R2 is measurable, then λ2 (A) = inf

P
n |Bn | where the infimum is over all
countable coverings of A by open balls B1 , B2 , . . . and |Bn | is the area. Deduce that Lebesgue
measure on R2 is rotation-invariant.
[Note: We defined Lebesgue measure using coverings by open rectangles (with sides parallel to
the axes). This exercise is to show that the shape of the basic sets does not affect the measure we
get].

Problem 27. True or false?

(1) If f : R2 7→ R2 is continuous and A ⊆ R2 is measurable, then f (A) := {f (x) : x ∈ A} is also
measurable.
4
 (2) Let µ : A 7→ [0, ∞] be a finitely additive set function on an algebra A. Then countable
additivity of µ is equivalent to continuity of µ under monotone limits (i.e., An ↑ A, A, A ∈ A
implies µ(An ) ↑ µ(A) and similarly for decreasing limits).
(3) G ⊆ F are sigma algebras on X. If µ is a σ-finite measure on F then its restriction to G is
also σ-finite.

Problem 28. Let f : R 7→ R be a function. Show that any of the following conditions implies that
f is Borel measurable.
(1) f is increasing.
(2) f is right-continuous.
(3) f is lower semi-continuous (means that f (x) = lim inf f (y) for all x).
y→x

Problem 29. Let (X, F) be a measure space and let fn : X 7→ R be a sequence of Borel measurable
functions.
P
(1) Show that supn fn , lim sup fn , lim fn (assuming limit exists), n fn (assuming that the sum
converges pointwise), are measurable (where necessary, these functions may be allowed to
take values ±∞ also).
(2) Show that the set {x ∈ X : limn fn (x) exists } is measurable in X. Same for {x : lim supn fn (x) =
+∞}, {x : lim fn (x) = +∞}.
(3) Show that the supremum of an uncountable family of measurable functions need not be
measurable (even if the supremum is finite everywhere).

Problem 30. Let (X, Σ) be a measure space and suppose f1 , f2 : X 7→ R. Then F = (f1 , f2 ) : X 7→
R2 .
(1) Show that F is Borel measurable if and only if both f1 , f2 are Borel measurable.
(2) Deduce that if f1 , f2 are measurable, then so are af1 + bf2 (for any a, b ∈ R) and f1 f2 and
f1 /f2 (for the last one, assume that f2 ≥ 0 so that f1 /f2 can be defined unambiguously as
an R-valued function).
(3) If f : R 7→ R is pointwise differentiable, then show that f 0 : R 7→ R is measurable.

Problem 31. Find a measurable function on an appropriate interval in R so that the push-forward
of Lebesgue measure of the interval is the measure µ satisfying
Rb
(1) µ(a, b] = a e−|x| dx for a < b.
Rb 1
(2) µ(a, b] = a 1+x 2 dx for a < b.
5
 (3) µ(a, b] = number of integers in (a, b].

Problem 32. Let K be the 1/3-Cantor set.

(1) Write out in detail the sketch given in class that there is a bijection T : [0, 1] 7→ K such that
T and T −1 are Borel measurable.
(2) Use T to construct a Lebesgue measurable set in R that is not Borel measurable. [Hint: Use
the existence of non-measurable sets.]

Problem 33. (*) This exercise is to show the isomorphism between ((0, 1), B(0,1) , λ1 ) and ((0, 1)2 , B(0,1)2 , λ2 ).
As pointed out in class, the essential idea is to take a number x = 0.x1 x2 ... in binary expansion
and map it to (y, z) where y = 0.x1 x3 ... and z = 0.x2 x4 .... But because of the ambiguities of binary
expansion at dyadic rationals, this is not quite a bijection between (0, 1) and (0, 1)2 . Here is how
to fix this.
(1) Find T1 : (0, 1) 7→ (0, 1)2 that is injective and such that Im(T1 ) is a Borel set in (0, 1)2 .
(2) Find T2 : (0, 1)2 7→ (0, 1) that is injective and such that Im(T2 ) is a Borel set in (0, 1).
(3) Use the idea of the proof of Schroder-Bernstein theorem to get a bijection T : (0, 1) 7→ (0, 1)2
so that T, T −1 are Borel measurable.
(4) If you base your T2 on the binary expansion idea above, it turns out that Lebesgue measure
on the two spaces are pushed forward to each other by T and T −1 .

Problem 34. (16 marks) State True or False and justify accordingly.
S
(1) If S is a collection of subsets of X, then A(S) = A(F ) where the union is over all finite
F
subsets F ⊆ S.
(2) If A is a bounded, measurable subset of R, then λ(A) = inf{λ(C) : C ⊇ A, C is closed}.
(3) For A ⊆ R2 , let Π(A) = {x : (x, y) ∈ A for some y} (a subset of R). If A is measurable, then
so is Π(A).
(4) If A is a measurable subset of Rd and 0 ≤ x ≤ λd (A), then there is a measurable subset
B ⊆ A such that λd (B) = x.

Problem 35. (10 marks) Let A = {(x, y) ∈ R2 : x > 0, y > 0, and x + y < 1}. Just from the
definition of the outer measure, calculate λ2 (A).

Problem 36. (10 marks) Let S be the collection of all intervals of the form (−a, a) with a ∈ R.
Show that if A ∈ σ(S) and x ∈ A, then −x ∈ A (in short, A = −A).

6
Problem 37. (10 marks) Let X and Y be metric spaces and let f : X 7→ Y be a continuous function.
If A is a Borel set in Y , show that f −1 (A) is a Borel set in X.

Problem 38. (10 marks) Let A and B be measurable subsets of the line with positive measure.
Show that there exists some x ∈ R such that A ∩ (B + x) is an infinite set.

Problem 39. Suppose the set of discontinuity points of f : R 7→ R has has zero Lebesgue measure.
Show that f is a measurable function (w.r.t the Lebesgue measure).

Problem 40. Suppose f : R 7→ R is such that at each point, f is either right continuous or left
continuous (or both). Is f necessarily Borel measurable?

Problem 41. Let (X, F, µ) be a measure space. If f f : X 7→ R is a bounded measurable function.

Show that there exist simple functions sn such that sn → f uniformly on X.

Problem 42. Let (X, F, µ) be a measure space and let f ∈ S+ .

(1) Show that t 7→ µ{f > t} is measurable from [0, ∞] to [0, ∞].
R R∞
(2) Show that X f dµ = 0 µ{f > t}dt.
R∞
(3) What about 0 µ{f ≥ t}dt?

Problem 43. Let f be a non-negative simple function on (X, F, µ). Define ν : F 7→ [0, ∞] be
R
defined by ν(A) = X f 1A dµ. Show that ν is a measure.

Problem 44. Let f be a non-negative simple function on (X, F, µ). Show that for any t > 0 we
have µ{f ≥ t} ≤ 1t X f dµ (Markov’s inequality).
R

Problem 45. Consider the space of Riemann integrable functions on [0, 1] with (pseudo)-metric
R1
d(f, g) = 0 |f − g|. Show that this metric space is not complete.

Problem 46. Suppose f, g are integrable functions on a measure space X, F, µ). Which of the
following are necessarily integrable? (a) f + g, (b) f − g, (c) f g, (d) f /g, (e) f ∨ g, (f) f ∧ g.

Problem 47. Given an integrable function f on (X, F, µ), show that there exist simple functions
R
sn such that X |f − sn |dµ → 0 as n → ∞.

7
 P
Problem 48. (1) If fn are non-negative measurable functions on (X, F, µ) and f = n fn , then
R P R
show that X f dµ = n n fn dµ.
R
(2) If g is a non-negative measurable function and ν(A) := A gdµ for A ∈ F, (convention: By
R R
A f dµ we just mean X f 1A dµ), then show that ν is a measure on F.

Problem 49. Suppose f : (a, b) × R 7→ R is a function such that (a) θ 7→ f (θ, x) is differentiable for
each x ∈ R, (b) x 7→ f (θ, x) is integrable for each θ ∈ (a, b), (c) |f (θ1 , x) − f (θ2 , x)| ≤ g(x)|θ1 − θ2 |
for all θ1 , θ2 ∈ (a, b) and x ∈ R and that g is integrable over R (w.r.t. Lebesgue measure). Define
R
H(θ) := R f (θ, x)dλ(x).
d
Show that H is differentiable, that dθ f (θ, x) is integrable over R for each θ, and that
Z
d d
H(θ) = f (θ, x)dλ(x).
dθ R dθ

Problem 50. Assume that fn are non-negative measurable functions that are bounded above by
an integrable function g a.e.[µ] (i.e. 0 ≤ fn ≤ g a.e.[µ]).
R R
(1) Show that lim sup X fn dµ ≤ X (lim sup fn )dµ.
n→∞ n→∞
R R R
(2) If fn ↓ f a.e.[µ], and X fn dµ is finite for some n, then show that X fn dµ ↓ X f dµ.

R
Problem 51. If f is an integrable function, show that |f |1|f |>n dµ → 0. More generally, if An is
R
any sequence of events such that µ(An ) → 0, then show that An |f |dµ → 0 as n → ∞.

Problem 52. If f : [a, b] 7→ R is a continuous function, show that its Riemann integral is equal to
its Lebesgue integral (w.r.t. Lebesgue measure on [a, b]). [Note: This is true more generally.]

Problem 53. Let An be measurable sets in (X, F, µ). Let A be the set of all x that belong to An for
infinitely many n (is A measurable?).
P
(1) If n µ(An ) < ∞, then show that µ(A) = 0.
(2) Show that µ(An ) → 0 does not necessarily imply that µ(A) = 0.

R
Problem 54. If fn are non-negative measurable functions with X fn dµ = 1 for all n and fn → f
R
a.e.[µ], then show that the set of all possible values of X f dµ is [0, 1].

Problem 55. Suppose fn , f are non-negative integrable functions on (X, F, µ) and that fn → f
R R
a.s.[µ]. Show that X fn dµ − kfn − f kL1 (µ) → X f dµ.
8
Problem 56. State whether true or false and justify. Whenever we write Lp , it is assumed that
1 ≤ p ≤ ∞.
(1) L∞ (R, B, λ1 ) is not separable (has no countable dense subset).
(2) If p1 < p2 , then Lp1 (X, F, µ) ⊇ Lp2 (X, F, µ).
(3) If p1 < p2 < p3 then Lp1 (µ) ∩ Lp3 (µ) ⊆ Lp2 (µ).
(4) If µ is a finite measure, then for any measurable f , the set {p ≥ 1 : f ∈ Lp (µ)} is an interval
of the form [1, r] or [1, r) for some r ≤ ∞.
(5) If f is a measurable function on (X, F, µ), then the set {p ≥ 1 : f ∈ Lp (µ)} is an interval
(Extra: Give exmples to show that this interval can be open or closed on the left or right).
(6) L1 (µ) ∩ L2 (µ) is complete in the norm k · k1 + k · k2 .

Problem 57. Show that L∞ (X, F, µ) is complete (This was skipped in class).

Problem 58. Let f ∈ Lp (X, F, µ) with p < ∞.

R∞
(1) Show that kf kpp = 0 ptp−1 µ{|f | > t}dt. [Note: This was shown for p = 1 and simple
non-negative f in an earlier exercise.]
(2) Show that that if f ∈ Lp (µ), then µ{|f | > t} ≤ kf kpp t−p .

Problem 59. Let f be a non-negative measurable function on (X, F, µ).

f dµ)2
R
(
(1) Show that µ{f > 0} ≥ RX
f 2 dµ .
X
R
(2) Show that no non-trivial lower bound for µ{f > 0} can be obtained in terms of X f dµ
alone.

Problem 60. Let µ be a finite measure. Then show that the following are equivalent.
(1) f ∈ L∞ (µ).
(2) f ∈ Lp (µ) for all p < ∞ and sup kf kp < ∞.
p
If these equivalent conditions hold, show that kf kp → kf k∞ .

Problem 61. (1) If f, g ∈ L2 , show that kf − gk2 + kf + gk2 = 2kf k2 + 2kgk2 .

(2) (*) If f1 , . . . , fn ∈ L2 (µ), show that (hfi , fj iL2 (µ) )1≤i,j≤n is a positive semi-definite matrix.
(3) Give examples to show that Lp (X, F, µ) is not a Hilbert space if p 6= 2.

Problem 62. Let µ be a probability measure on X. If f, g are non-negative measurable and f g ≥ 1

R
R

a.e.[µ], then X f dµ X gdµ ≥ 1.
9
Problem 63. Suppose µn is a sequence of Radon measures on Rd such that limn→∞
R
f dµn exists
(Rd ).
R
and is finite for every f ∈ Cc Then show that there is a Radon measure µ such that f dµn →
f dµ for all f ∈ Cc (Rd ).
R

Problem 64. Let 1 ≤ p ≤ ∞. Is Lp (R, B, λ1 ) separable in Lp metric?

Problem 65. Suppose fn ∈ L1 (µ) ∩ L2 (µ). If fn → g in L1 (µ) and fn → h in L2 (µ), then g = h

a.e.[µ].

Problem 66. Let (X, F, µ) and (Y, G, ν) be measure spaces such that ν = µ ◦ T −1 for some T : X 7→
Y . Then show that for any f ∈ L1 (ν), the function f ◦ T ∈ L1 (µ) and that X (f ◦ T )dµ = Y f dν.
R R

Problem 67. Let L be a positive linear functional on Cc∞ (Rd ) (endowed with sup-norm). Show
that L(f ) = X f dµ for a unique Radon measure µ on Rd .
R

Problem 68. Which of the following sets is dense in Lp ([0, 1], B, λ1 )? Consider the case p = ∞
carefully.
(1) C[0, 1].
(2) C ∞ [0, 1].
(3) The set of all polynomials.
(4) The collection of all step functions.

kf kp+1
p+1
Problem 69. If µ is a finite measure, show that kf kpp
→ kf k∞ as p → ∞.

Problem 70. Suppose 0 < p < 1. If f, g are non-negative measurable function on (X, F, µ), then
p 1/p ≥ p dµ 1/p +
R
R
R p
1/p
X (f + g) X f X g dµ .

Problem 71. Let D be a bounded open set in the complex plane and let µ be the Lebesgue measure
on D. Let H be the collection of all holomorphic functions on D that are in L2 (D). Show that H is
a closed subspace in L2 norm and hence a Hilbert space itself.

Problem 72. Suppose 1 ≤ p < q < r < ∞. If fn ∈ Lp (µ) ∩ Lr (µ) and this sequence is Cauchy in
Lp (µ) and Cauchy in Lr (µ), then fn converges in Lq (µ).

10
Problem 73. Let (X, F, µ) be a measure space. Let An ∈ F. Write out explicitly the meaning of
1An → 0 in a.e.[µ] sense, in measure and in L1 . Which of these imply the others? (Do this directly,
without invoking the general theorems proved in class). What if µ is finite?

Problem 74. Let µ be a measure that is not supported at a single point. Show that Lp (µ) norm
does not come from an inner product if p 6= 2 and does come from an inner product if p = 2.

R R
R

Problem 75. Suppose µ is a Borel measure on [0, 1] such that f gdµ = f dµ gdµ for allf, g ∈
C[0, 1]. Show that µ = δa for some a ∈ [0, 1].

Problem 76. Suppose fn , f are non-negative measurable functions such that fn → f a.e.[µ]. Show
R R R
that X fn dµ → X f dµ if and only if X |fn − f |dµ → 0.

Problem 77. If fn → f in measure µ, and |fn | ≤ g for some integrable function g, then show that
R
X |fn − f |dµ → 0 (DCT under convergence in measure only).

Problem 78. Let fn , f be measurable functions on (X, F, µ). If 1|fn −f |≥δ → 0 a.e.[µ] for every
δ > 0, then is it true that fn → f a.e.[µ]?

Problem 79. (16 marks) State True or False and justify accordingly.
(1) If f ∈ L1 (µ) ∩ L3 (µ), then f ∈ L2 (µ).
(2) Let (X, F, µ) be a measure space and let f : X 7→ R be an integrable function. Let An ∈ F
R
be a sequence of measurable sets. If µ(An ) → 0 as n → ∞, then An f dµ → 0.
R R
(3) On a measure space (X, F, µ) suppose fn → f a.e.[µ] and fn dµ → f dµ. Then there is
an integrable function g such that |fn | ≤ g a.e.[µ], for all n.
(4) Let f : R2 7→ R be defined as f (x, y) = x − y and let F be the smallest σ-algebra on R2
such that f is Borel measurable. Then F does not contain any non-empty bounded set in
the plane.

Problem 80. (10 marks) If f : R 7→ R is right-continuous, then f is Borel measurable.

Problem 81. (10 marks) Let (X, F, µ) be a measure space and let f : X 7→ R+ be a non-negative
R
measurable function. Define θ(A) = A f dµ for A ∈ F.
(1) Show that θ is a measure on F.
11
 (2) Show that g : X 7→ R is integrable w.r.t. θ if and only if gf is integrable w.r.t. µ and
R R
X gdθ = X gf dµ.

1 1 1
Problem 82. (10 marks) Suppose 1 < p1 < p2 < p3 < ∞ are such that p1 + p2 + p3 = 1. If
f1 , f2 , f3 are measurable functions on (X, F, µ) such that fi ∈ Lpi (µ) for i = 1, 2, 3, then show that
f1 f2 f3 ∈ L1 and kf1 f2 f3 k1 ≤ kf1 kp1 kf2 kp2 kf3 kp3 .

Problem 83. (10 marks) Let (X, F, µ) be a measure space and let f be a non-negative measurable
function.
R P∞
(1) If f takes only integer values, then show that f dµ = n=1 µ{f ≥ n}.
(2) In general (i.e., for non-negative measurable f ), show that f is integrable if and only if
P∞
n=1 µ{f ≥ n} converges.

f 2 dµ = f 4 dµ (all
R R R
Problem 84. If f is a measurable function on (X, F, µ) such that f dµ =
integrals assumed to exist and finite). Then f = 1A for some A ∈ F.

Problem 85. If f is a non-negative measurable function such that log f is integrable, then
R p
(1) f p−1 dµ → log f dµ as p ↓ 0.
R

(2) p1 log f p dµ → log f dµ as p ↓ 0.

R
R

Problem 86. Suppose fn → f a.e.[µ] and that sup fn2 dµ < ∞. Then |fn − f |dµ → 0.
R R
n

Problem 87. If fn → f a.e.[µ] in (X, F, µ), then show that there exist an ↑ ∞ such that an Xn → 0
a.s.[µ].

Problem 88. Let X and Y be separable metric spaces and let Z = X ×Y endowed with the product
topology.
(1) If X and Y are separable, show that BX × BY = BZ .
(2) Suppose X = Y is not separable. Then show that D = {(x, x) : x ∈ X} is in BZ but not in
BX × BY .

Problem 89. Let µ be a Radon measure on R. Suppose convergence in measure w.r.t. µ implies
convergence almost everywhere w.r.t. µ. What can you say about µ?

12
Problem 90. Let Ld denote the Lebesgue sigma algebra on Rd . Show that L2 6= L1 × L1 .

Problem 91. let f : R2 7→ R be defined by f (x, y) = sin(x)1y<x<y+2π . Show that f (x, y)dydx 6=
RR
RR
f (x, y)dxdy. Does this indicate a fatal flaw in Fubini’s theorem as presented in thousands of
books?

Problem 92. Let A be a Borel set in R2 such that its intersection with each vertical line is a finite
set. Show that for a.e. y[λ1 ], the intersection of A with the horizontal line through (0, y) has zero
Lebesgue measure (in one dimension).

Problem 93. Suppose A, B are measurable subsets of [0, 1) (treated as a group with addition mod-
ulo 1). Show that there exists some x ∈ [0, 1) such that λ(A ∩ (B + x)) ≥ λ(A)λ(B).

R
Problem 94. Let f : R 7→ R be a non-negative measurable function. Show that R f (x)dλ(x)
is equal to the area (two-dimensional Lebesgue measure) of {(x, y) : 0 ≤ y ≤ f (x)} (the region
between the graph of f and the x-axis.).

R
Problem 95. If f be a non-negative measurable function on (X, F, µ). Show that X f dµ =
R∞
0 µ{f > t}dt. [Hint: Use Fubini’s theorem on X × R+ with ...]

Problem 96. Let A ⊆ R be a measurable set with λ1 (A) = 1. Find λ2 {(x, y) ∈ R2 : (1+x2 )(y −ex ) ∈
A}.

P
Problem 97. Let fn be measurable functions on (X, F, µ) such that f = n fn converges a.e.[µ].
R P R
From Fubini’s and Tonelli’s theorem discuss when the identity f dµ = n fn dµ holds. Con-
sider the case when fn are non-negative separately. How does this compare to the conditions
required by MCT and DCT?

Problem 98. If (Xi , Fi , µi ), i = 1, 2, 3, are σ-finite measure spaces. Show that

(F1 × F2 ) × F3 = (F1 × F2 ) × F3 and (µ1 × µ2 ) × µ3 = µ1 × (µ2 × µ3 ).

This justifies writing µ1 × µ2 × µ3 etc.

Problem 99. Write R2 = R × R and R2 \ {0} = (0, ∞) × [0, 2π) with the identification (x, y) ↔ (r, θ)
if x = r cos θ and y = r sin θ (polar co-ordinates). Let λ denote the one-dimensional Lebesgue
measure on R. Let λ+ and λ0 denote its restriction to (0, ∞) and to [0, 2π).
13
 (1) Show that on R2 (or R2 \ {0}), the two measures λ × λ and λ+ × λ0 are equal.
R R R R
(2) Deduce the integration formula R R f (x, y) dx dy = (0,∞) [0,2π) f (r cos θ, r sin θ)rdθdr for
f ∈ L1 (λ2 ).

Problem 100. Let µ, ν be Borel probability measures on R. Show that (x + y)2 is integrable w.r.t.
µ × ν if and only if x2 is integrable w.r.t. µ and w.r.t. ν.

Problem 101. Let µn be a sequence of regular Borel measures on R such that limn→∞ f dµn exists
and is finite for every f ∈ Cc (R). Show that there is a regular Borel measure µ on Rd such that
f dµn → f dµ for all f ∈ Cc (Rd ).
R R

R
Problem 102. Let µn , µ be regular Borel measures on R such that limn→∞ f dµn = f dµ for every
f ∈ Cc (R). Then show that µn (a, b) → µ(a, b) for all except atmost countably many a, b.

R
Problem 103. Show that there exists probability measures µn , νn on [0, 1] such that f dµn −
R
f dνn → 0 as n → ∞, for any f ∈ C[0, 1] but such that µn [0, t] − νn [0, t] 6→ 0 for any t ∈ (0, 1).

Problem 104. Let µ, ν be finite measures on (X, F). Show that ν  µ if and only if for every  > 0,
there is a δ > 0 such that ν(A) <  for any A ∈ F with µ(A) < δ.

Problem 105. Construct a sigma-finite measure µ on (R, BR ) such that µ  λ1 and µ(a, b) = ∞ for
any a < b.

Problem 106. Let (X, F, µ) be a probability space. Let An ∈ F and let A = lim sup An .
P
(1) If n µ(An ) < ∞, then show that µ(A) = 0.
(2) If µ(An ) ≥  > 0 for all n, then show that µ(A) ≥ .

Problem 107. Let f : Rd 7→ R+ be a non-negative integrable function. Show that there exists a
radial function g : Rd 7→ R (this means that g(x) = g(y) if |x| = |y|) such that λd {f > t} = λd {g >
R R
t} for all t > 0. Deduce that f dλd = gdλd . [Note: g is called the spherical rearrangement of f .]

Problem 108. On (X, F, µ), suppose f is a non-negative integrable function. Let

X
I(λ) = λn µ{λn ≤ f < λn+1 }.
n∈Z

Show that I(λ) is well-defined (i.e., the series converges absolutely) for every λ > 1 and that
R
I(λ) ↓ f dµ as λ ↓ 1.
14
Problem 109. Let f be a non-negative measurable function on the sigma-finite measure space
R
(X, F) and let ν(A) = A f dµ. Show that ν is sigma-finite.

Problem 110. Let f be a non-negative measurable function on the sigma-finite measure space
(X, F). If G is a sub-sigma algebra of F, then show that there exists a g : X 7→ R+ that is Borel
R R
measurable w.r.t. G and such that A f dµ = A gdµ for all A ∈ G.

Problem 111. Suppose ν1  µ1 and ν2  µ2 (all sigma-finite measure on some measure spaces).
Then show that ν1 × ν2  µ1 × µ2 .

Problem 112. (16 marks) State “True” or “False” and justify accordingly.
(1) Let F be the collection of all those Borel subsets of R that have zero Lebesgue measure or
whose complement has zero Lebesgue measure. Then F is a σ-algebra.
(2) If f is an integrable function on (X, F, µ), then tµ{f > t} → 0 as t → ∞.
1
(3) Suppose An are subsets of [0, 1] such that λ(An ) ≤ 2 for all n. Then λ(lim sup An ) ≤ 21 .
(4) Suppose µ and ν are measures on (X, F) such that ν is absolutely continuous to µ. If µ is a
finite measure, then so is ν.
(5) For k ∈ N = {1, 2, 3, . . .}, let Ak = {k, 2k, 3k, . . .}. Then the sigma algebra generated by the
sets Ap as p varies over all primes, is the power set of N.

Problem 113. (16 marks)

(1) Let F be the sigma-algebra on R generated by the intervals (0, 2) and (1, 3). What is the
cardinality of F?
(2) Let µ∗ be an outer measure onX. Suppose A1 , B1 , A2 , B2 , . . . are subsets of X such that
µ∗ (An ∆Bn ) = 0 for each n ≥ 1. Show that µ∗ (∪n An ) = µ∗ (∪n Bn ).
(3) Suppose A1 , . . . , A9 are Borel subsets of [0, 1] such that each x ∈ [0, 1] belongs to at least
1
three distinct sets among A1 , . . . , A9 . Show that λ(Ak ) ≥ 3 for some k.
(4) Let A ⊆ R be a Borel set with λ1 (A) = 1. Find λ2 {(x, y) ∈ R2 : (1 + x2 )(y − ex ) ∈ A}.

Problem 114. (6 marks) Suppose fn , f, g are measurable functions on (X, F, µ).

(1) If 1|fn −f |≥δ → 0 a.e.[µ] for every δ > 0, then show that fn → f a.e.[µ].
R R
(2) Assume that f and g are integrable. If A f dµ = A gdµ for all A ∈ F, then f = g a.e.[µ].

Problem 115. (6 marks) Given f : R2 7→ R, define g : R 7→ R by g(x) = f (x, 1 − x).

15
 (1) Take the Borel sigma algebras on R and R2 . If f is measurable, then so is g.
(2) Take the Lebesgue sigma algebras on R and R2 . Show that g need not be measurable even
if f is measurable.

Problem 116. (6 marks) Consider R with the Borel sigma algebra.

(1) Let f, g : [0, 1] 7→ R be Borel measurable functions. If h(x, y) = f (x) + g(y) is integrable
over [0, 1]2 , then f and g are integrable over [0, 1].
P
(2) If f : R 7→ R is integrable, then f (x + n) converges for a.e. x (w.r.t. Lebesgue measure).
n∈Z

R
Problem 117. (5 marks) If f : R 7→ R is integrable and g(x) = f dλ, then show that g is a
[x,x+1]
continuous function.

Problem 118. (5 marks) Suppose µ is a Borel measure on R such that µ(A) = 1 whenever A is a
Borel set with λ(A) = 1. Show that µ = λ.

16