Chapter 9

General Measure Spaces: Their Properties and Construction

[18/0%]
LaTeX
Problem 9.1

Define μ:2N→[0,∞)\mu : 2^{\mathbb {N}} \rightarrow [0, \infty ) by μ(E)=∑n∈E1/2n\mu (E) = \sum_{n \in E} 1 / 2^{n}. Show that this is a finite measure.

?
Problem 9.2

For each kk, let (X,M,μk)\left(X, \mathcal{M}, \mu_{k}\right) be a measure space and suppose that there is a c≥0c \geq 0 for which μk(X)≤c\mu_{k}(X) \leq c for all kk. Show that μ=∑k=1∞μk2k\mu = \sum_{k = 1}^{\infty } \frac{\mu_{k}}{2^{k}} also is a measure.

?
Problem 9.3

Let M\mathcal{M} be a σ\sigma-algebra and the set-function μ:M→[0,∞)\mu : \mathcal{M} \rightarrow [0, \infty ) be finitely additive. Prove that μ\mu is a measure if and only if whenever {Ak}k=1∞\left\{ A_{k}\right\}_{k = 1}^{\infty } is an ascending collection of sets in M\mathcal{M}, then (1) holds.

?
Problem 9.4

Let (X,M,μX, \mathcal{M}, \mu) be a finite measure space. The symmetric difference, AΔBA \Delta B, of two measurable sets AA and BB is defined by

AΔB=[A∼B]∪[B∼A]=[A∪B]∼[A∩B]. A \Delta B = [A \sim B] \cup [B \sim A] = [A \cup B] \sim [A \cap B] .

Identify two measurable sets whose symmetric difference has measure zero, and prove the following triangle inequality on M\mathcal{M} : For measurable sets A,BA, B and CC,

μ(AΔC)≤μ(AΔB)+μ(BΔC). \mu (A \Delta C) \leq \mu (A \Delta B)+\mu (B \Delta C) .

The σ\sigma-algebra M\mathcal{M}, with the metric ρ(A,B)=μ(AΔB)\rho (A, B) = \mu (A \Delta B), is called the Nikodym metric space.

?
Problem 9.5

Let M\mathcal{M} be the collection of subsets of XX that are either countable or have a countable complement in XX. Show that M\mathcal{M} is a σ\sigma-algebra. For E∈ME \in \mathcal{M}, define μ(E)=0\mu (E) = 0 if EE is countable and μ(E)=1\mu (E) = 1, if EE has a countable complement. Show that μ\mu is a measure. Is this measure space complete?

?
Problem 9.6

Let S\mathcal{S} be any collection of subsets of XX and define M\mathcal{M} to be the intersection of all σ\sigma-algebras of subsets of XX that contain S\mathcal{S}. Show that M\mathcal{M} is a σ\sigma-algebra, and it is the smallest σ\sigma-algebra that contains S\mathcal{S}, in the sense that it is contained in any σ\sigma-algebra that contains S\mathcal{S}. For X=RX = \mathbb {R}, find the smallest σ\sigma-algebra that contains all singleton sets.

?
Problem 9.7

Let M\mathcal{M} be a σ\sigma-algebra. Verify the following:

?
(i)

If μ\mu and ν\nu are measures on M\mathcal{M}, then the set-function λ\lambda defined on M\mathcal{M} by λ(E)=μ(E)+ν(E)\lambda (E) = \mu (E)+\nu (E) also is a measure.

(ii)

If μ\mu and ν\nu are measures on M\mathcal{M} and μ≥ν\mu \geq \nu, then there is a measure λ\lambda on M\mathcal{M} for which μ=ν+λ\mu = \nu +\lambda.

(iii)

If ν\nu is σ\sigma-finite, the measure λ\lambda in (ii) is unique.

(iv)

Show that in general the measure λ\lambda need not be unique but that there is always a smallest such λ\lambda.

Problem 9.8

Show that in the construction of Lebesgue measure on R\mathbb {R}, the outer-measure is unchanged if covers by open, bounded intervals are replaced by covers by general, bounded intervals.

?
Problem 9.9

Let μ∗:2X→[0,∞]\mu^{*}: 2^{X} \rightarrow [0, \infty ] be an outer-measure. Let A⊆X,{Ek}k=1∞A \subseteq X,\left\{ E_{k}\right\}_{k = 1}^{\infty } be a countable, disjoint collection of measurable sets and E=⋃k=1∞EkE = \bigcup_{k = 1}^{\infty } E_{k}. Show that

μ∗(A∩E)=∑k=1∞μ∗(A∩Ek) \mu ^{*}(A \cap E) = \sum _{k = 1}^{\infty } \mu ^{*}\left(A \cap E_{k}\right)
?
Problem 9.10

Show that any measure that is induced by an outer-measure is complete.

?
Problem 9.11

Define μ:2R→R\mu : 2^{\mathbb {R}} \rightarrow \mathbb {R} by defining μ(A)\mu (A) to be the number of integers in AA. Determine the outer-measure μ∗\mu^{*} induced by μ\mu and the σ\sigma-algebra of measurable sets.

?
Problem 9.12

Let S⊆2XS \subseteq 2^{X} and μ∗:2X→[0,∞]\mu^{*}: 2^{X} \rightarrow [0, \infty ] be the outer-measure induced by μ:S→[0,∞]\mu : S \rightarrow [0, \infty ]. Show that for a set A⊆XA \subseteq X to be measurable with respect to μ∗\mu^{*} it suffices to verify that

μ∗(E)=μ∗(A∩E)+μ∗(A∼E) for all E∈S, \mu ^{*}(E) = \mu ^{*}(A \cap E)+\mu ^{*}(A \sim E) \text{ for all } E \in \mathcal{S},

not for all of the subsets EE of XX. This was the definition of measurable set used by Lebesgue for the set-function length defined on the collection of open, bounded intervals in R\mathbb {R}.

?
Problem 9.13

Show that a non-negative set-function on a σ\sigma-algebra is a measure if and only if it is a premeasure.

?
Problem 9.14

Let Q\mathbb {Q} be the set of rational numbers and S\mathcal{S} the collection of all finite unions of intervals of the form (a,b]∩Q(a, b] \cap \mathbb {Q}, where a,b∈Qa, b \in \mathbb {Q} and a≤ba \leq b. Define μ(a,b]=∞\mu (a, b] = \infty if a<ba < b and μ(∅)=0\mu (\emptyset ) = 0. Show that S\mathcal{S} is closed with respect to relative complements and μ:S→[0,∞]\mu : \mathcal{S} \rightarrow [0, \infty ] is a premeasure. Then show that the extension of μ\mu to the smallest σ\sigma-algebra containing S\mathcal{S} is not unique.

?
Problem 9.15

Verify the assertions made in this section's first and second example.

?
Problem 9.16

Show that the three collections of sets in this section's third example are semi-rings.

?
Problem 9.17

Let S\mathcal{S} be a σ\sigma-algebra of subsets of XX and μ:S→[0,∞]\mu : \mathcal{S} \rightarrow [0, \infty ] be a measure. Let μˉ:M‾→[0,∞]\bar{\mu }: \overline{\mathcal{M}} \rightarrow [0, \infty ] be the Carathéodory measure induced by μ\mu. Show that S\mathcal{S} is a subcollection of M‾\overline{\mathcal{M}} and it may be a proper subcollection.

?
Problem 9.18

Show that Lebesgue on R\mathbb {R} is the Carathéodory extension of the premeasure Lebesgue measure on the semi-ring of Borel sets.

?