Chapter 9

Differentiation and Integration

[179/19%]
Show
LaTeX
Problem 9.1.1

Calculate all Dini derivatives of χQ\chi_{\mathbb {Q}}, the characteristic function of rationals.

?
Problem 9.1.2

Let

f(x)={∣x∣ if x∈Q∣2x∣ if x∉Q f(x)= \begin{cases} \left|x\right| & \text{ if } x \in \mathbb {Q} \\ \left|2 x\right| & \text{ if } x \notin \mathbb {Q} \end{cases}

Calculate all four Dini derivatives of ff at x=0x=0 and at x=1x=1.

?
Problem 9.1.3

Calculate all Dini derivatives of xχQ(x)−xχQC(x)x \chi_{\mathbb {Q}}(x)-x \chi_{\mathbb {Q}^{C}}(x) at x=0x=0 and at x=1x=1.

?
Problem 9.1.4

Let f(x)=2−nf(x)=2^{-n} if x∈(2−n,2−n+1]x \in \left(2^{-n}, 2^{-n+1}\right] and f(0)=0f(0)=0. Calculate all four Dini derivatives of ff at x=0x=0.

?
Problem 9.1.5

Calculate all Dini derivatives of χ[0,1](x)\chi_{[0,1]}(x) at x=0x=0 and at x=1x=1.

?
Problem 9.1.6

Construct an increasing function f:[−1,1]→Rf:[-1,1] \rightarrow \mathbb {R} such that none of the four Dini derivatives at 0 are equal.

?
Problem 9.1.7

Prove that a function ff is differentiable at x0x_{0} (in the standard meaning) if and only if all 4 Dini derivatives are equal and finite, in which case f′(x0)=D+f(x0)=D+f(x0)=D−f(x0)=D−f(x0)f^{\prime }\left(x_{0}\right)= D^{+} f\left(x_{0}\right)=D_{+} f\left(x_{0}\right)=D^{-} f\left(x_{0}\right)=D_{-} f\left(x_{0}\right).

?
Problem 9.1.8

If ff is defined and finite on [−c,c][-c, c], prove the following:

?
(a)

D+(−f(x))=−D+f(x),D−(−f(x))=−D−f(x)D^{+}(-f(x))=-D_{+} f(x), D^{-}(-f(x))=-D_{-} f(x).

(b)

D+(f(−x))=−D−f(x),D+(f(−x))=−D−f(x)D^{+}(f(-x))=-D_{-} f(x), D_{+}(f(-x))=-D^{-} f(x).

(c)

D−(f(−x))=−D+f(x),D−(f(−x))=−D+f(x)D^{-}(f(-x))=-D_{+} f(x), D_{-}(f(-x))=-D^{+} f(x).

Problem 9.1.9

If ff and gg are defined and finite on [a,b][a, b], prove that for every x,D+(f+g)(x)≤D+f(x)+D+g(x)x, D^{+}(f+ g)(x) \leq D^{+} f(x)+D^{+} g(x). Give an example to show that the inequality may be strict.

?
Problem 9.1.10

Let f,g:R→Rf, g: \mathbb {R} \rightarrow \mathbb {R} be continuous and suppose that for every x∈Rx \in \mathbb {R}, D+f(x)=D+g(x)<∞D^{+} f(x)=D^{+} g(x)<\infty. Prove that ff and gg differ by a constant.

?
Problem 9.1.11

Give an example to show that the chain rule for Dini derivatives D+(f∘g)(x)=D+f(g(x))D+g(x)D^{+}(f \circ g)(x)=D^{+} f(g(x)) D^{+} g(x) does not hold.

?
Problem 9.1.12

Let AA be a subset of R\mathbb {R} of measure 0 . Prove that there exists a continuous increasing function f:R→Rf: \mathbb {R} \rightarrow \mathbb {R} for which each Dini derivative is infinite on each point of AA.

?
Problem 9.1.13

Let f:(a,b)→Rf:(a, b) \rightarrow \mathbb {R} be a continuous function. Prove that ff is nonincreasing on ( a,ba, b ) if and only if D+f≤0D^{+} f \leq 0 on ( a,ba, b ).

?
Problem 9.1.14

Let f,g:[a,b]→Rf, g:[a, b] \rightarrow \mathbb {R} be continuous functions such that D+f(t)≤0D^{+} f(t) \leq 0 and D+g(t)≤0D^{+} g(t) \leq 0 on (a,b)(a, b). If h(t)=max⁡{f(t),g(t)}h(t)=\max \left\{ f(t), g(t)\right\}, prove that D+h(t)≤0D^{+} h(t) \leq 0 on (a,b)(a, b).

?
Problem 9.1.15

If ff and gg are defined and finite on [a,b][a, b], prove the following:

?
(a)

If gg has a right derivative g′(c+)g^{\prime }(c+) at c∈[a,b]c \in [a, b], then

D+(f+g)(c)=D+f(c)+g′(c+)andD+(f+g)(c)=D+f(c)+g′(c+). D^{+}(f+g)(c)=D^{+} f(c)+g^{\prime }(c+) \quad \text{and} \quad D_{+}(f+g)(c)=D_{+} f(c)+g^{\prime }(c+).
(b)

If gg has a left derivative g′(c−)g^{\prime }(c-) at c∈[a,b]c \in [a, b], then

D−(f+g)(c)=D−f(c)+g′(c−)andD−(f+g(c))=D−f(c)+g′(c−). D^{-}(f+g)(c)=D^{-} f(c)+g^{\prime }(c-) \quad \text{and} \quad D_{-}(f+g(c))=D_{-} f(c)+g^{\prime }(c-).
Problem 9.1.16

Let ff and gg be nonnegative and continuous at cc.

?
(a)

Prove that D+(fg)(c)≤g(c)D+f(c)+f(c)D+g(c)D^{+}(f g)(c) \leq g(c) D^{+} f(c)+f(c) D^{+} g(c).

(b)

Give an example to show that the inequality in (a) may be strict.

(c)

Give an example to show that the inequality in (a) is not true without the continuity and nonnegativity assumptions.

Problem 9.1.17

Complete the proof of Proposition 9.1.4 (if ff is monotone, the four Dini derivatives of ff are measurable) by showing that if ff is monotone then the remaining three Dini derivatives of ff are measurable ( D+,D−D_{+}, D^{-}, and D−D_{-}).

?
Problem 9.1.18

Suppose that D+f(x0)D_{+} f\left(x_{0}\right) is finite. Prove that ff is continuous from the right at x0x_{0}.

?
Problem 9.1.19

Suppose ff is continuous on (a,b)(a, b). Prove that ff is increasing on (a,b)(a, b) if and only if at least one of the four Dini derivatives of ff is nonnegative on (a,b)(a, b).

?
Problem 9.1.20

Let ff be an increasing function on [a,b][a, b] and let x0∈(a,b)x_{0} \in (a, b). Prove that none of the Dini derivatives can be negative (or equal to −∞-\infty ) at x0x_{0}.

?
Problem 9.1.21

(Dini, 1878) Let f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} be a continuous function. Prove that the bounds of each of the Dini derivatives are equal to the bounds of the difference quotient

f(y)−f(x)y−x, \frac{f(y)-f(x)}{y-x},

as x,y∈[a,b]x, y \in [a, b]. Use this to conclude that if a single Dini derivative of a continuous function ff is continuous at a point x0x_{0}, then so are the other three. Show that in this case all four Dini derivatives are equal, so ff is differentiable at x0x_{0}.

?
Problem 9.1.22

Give an example to show that the conclusions of Exercise 9.1.21 need not hold if the continuity assumption on ff is dropped.

?
Problem 9.1.23

Let ff be continuous on [a,b][a, b], and let a constants MM and a set E⊂[a,b]E \subset [a, b] of positive measure satisfy D+f(x)≥MD^{+} f(x) \geq M for all x∈Ex \in E (or D−f(x)≥MD^{-} f(x) \geq M for all x∈Ex \in E ). Prove that for every ε>0\varepsilon >0 there exists a finite collection of disjoint intervals (a1,b1),…,(an,bn)\left(a_{1}, b_{1}\right), \ldots ,\left(a_{n}, b_{n}\right) in [a,b][a, b] such that ∑i=1n(bi−ai)>m(E)−ε\sum_{i=1}^{n}\left(b_{i}-a_{i}\right)> m(E)-\varepsilon and

f(bi)−f(ai)≥(M−ε)(bi−ai),i=1,2,…,n. f\left(b_{i}\right)-f\left(a_{i}\right) \geq (M-\varepsilon )\left(b_{i}-a_{i}\right), \quad i=1,2, \ldots , n .
?
Problem 9.1.24

Let ff be a real-valued function. Prove that the set of points where D+f(x)<D−f(x)D^{+} f(x)<D_{-} f(x) is countable.

?
Problem 9.1.25

Give an example showing that it is possible that the set A1A_{1} in Theorem 9.1.7 have Lebesgue measure b−ab-a. Do the same for sets A2A_{2} and A3A_{3}.

Theorem 9.1.7 (The Denjoy--Young--Saks Theorem): Let ff be a real-valued function defined on [a,b][a, b]. Then almost every point x∈[a,b]x \in [a,b] is in one of four sets: A1A_1, on which ff has a finite derivative; A2A_2, on which D+f=D−fD^{+} f=D_{-} f is finite, D−f=∞D^{-} f=\infty, and D+f=−∞D_{+} f=-\infty; A3A_3, on which D−f=D+fD^{-} f=D_{+} f is finite, D+f=∞D^{+} f=\infty, and D−f=−∞D_{-} f=-\infty; A4A_4, on which D−f=D+f=∞D^{-} f=D^{+} f=\infty and D−f=D+f=−∞D_{-} f=D_{+} f=-\infty.

?
Problem 9.1.26

Let ff be continuous on [a,b][a, b] and suppose that it has a local maximum at c∈(a,b)c \in (a, b). Prove that

D+f(c)≤D+f(c)≤0≤D−f(c)≤D−f(c). D_{+} f(c) \leq D^{+} f(c) \leq 0 \leq D_{-} f(c) \leq D^{-} f(c) .
?
Problem 9.1.27

Let f:R→Rf: \mathbb {R} \rightarrow \mathbb {R} be continuous and suppose that D+fD^{+} f is Riemann integrable over [a,b][a, b]. Prove that

∫abD+f(x)dx=f(b)−f(a) \int _{a}^{b} D^{+} f(x) d x=f(b)-f(a)
?
Problem 9.1.28

If ff is continuous on [a,b][a, b] prove that the four Dini derivatives of ff are measurable.

?
Problem 9.1.29

If ff is measurable prove that the four Dini derivatives of ff are measurable.

?
Problem 9.1.30

A function f∈C[a,b]f \in C[a, b] is called Hölder continuous of order δ\delta, notation f∈Λδf \in \Lambda_{\delta }, if there is a constant CC such that ∣f(x)−f(y)∣≤C∣x−y∣δ\left|f(x)-f(y)\right| \leq C\left|x-y\right|^{\delta } for all x,y∈[a,b]x, y \in [a, b]. Prove that the Hölder continuous functions form a set of first categories in C[a,b]C[a, b].

?
Problem 9.2.1

Let F={[p,q]:0≤p<q≤1,p,q∈Q}\mathcal{F}=\left\{ [p, q]: 0 \leq p<q \leq 1, p, q \in \mathbb {Q}\right\}. Prove that F\mathcal{F} is a Vitali cover for [0,1][0,1].

?
Problem 9.2.2

Let E⊂RE \subset \mathbb {R}, and for every x∈Ex \in E let {βn(x)}\left\{ \beta_{n}(x)\right\} be a strictly decreasing sequence of real numbers that converges to xx. Prove that F={[x,βn(x)]:x∈E}\mathcal{F}=\left\{ \left[x, \beta_{n}(x)\right]: x \in E\right\} is a Vitali cover for EE.

?
Problem 9.2.3

Let E⊂RE \subset \mathbb {R} be a union (not necessarily countable and not necessarily disjoint) of nondegenerate intervals (open, closed, half open and half closed, infinite, etc). Show that EE is measurable.

?
Problem 9.2.4

Show that the Vitali Covering Lemma extends to the case in which the covering collection consists of nondegenerate intervals that are not necessarily closed.

?
Problem 9.2.5

Show that the Vitali Covering Lemma does not extend to the case in which the covering collection contains some intervals that are not necessarily nondegenerate.

?
Problem 9.2.6

Prove that a set E⊂RE \subset \mathbb {R} has measure 0 if and only if there is a countable collection of open intervals {Ik}k=1∞\left\{ I_{k}\right\}_{k=1}^{\infty } for which each point in EE belongs to infinitely many of these intervals and ∑k=1∞ℓ(Ik)<∞\sum_{k=1}^{\infty } \ell \left(I_{k}\right)<\infty.

?
Problem 9.2.7

Prove that each function gng_{n} defined in the proof of Corollary 9.2.9 is Riemann integrable.

In the proof of Corollary 9.2.9 (which states that for ff increasing on [a,b][a,b], f′f' is integrable over [a,b][a,b] and ∫abf′≤f(b)−f(a)\int_{a}^{b} f' \leq f(b) - f(a)), ff is first extended to [a,b+1][a, b+1] by f(x)=f(b)f(x) = f(b) for x∈[b,b+1]x \in [b, b+1], and then gn(x):=f(x+1/n)−f(x)1/ng_{n}(x) := \dfrac {f(x+1/n) - f(x)}{1/n} for x∈(a,b)x \in (a, b).

?
Problem 9.2.8

Prove inequality (9.23).

Inequality (9.23): Let ff be an increasing function on [a,b][a, b], extended to the interval [b,b+1][b, b+1] by f(x)=f(b)f(x)=f(b) for x∈[b,b+1]x \in [b, b+1], and let

gn(x):=f(x+1/n)−f(x)1/n. g_{n}(x):=\frac{f(x+1 / n)-f(x)}{1 / n} .

Then

∫abgn(x)dx≤f(b)−f(a).(9.23) \int _{a}^{b} g_{n}(x) d x \leq f(b)-f(a) . \tag {9.23}
?
Problem 9.2.9

Prove Theorem 9.2.12.

Theorem 9.2.12: Let ff be a monotone function on (a,b)(a,b). Then ff has at most a countable number of jump discontinuities on (a,b)(a,b).

?
Problem 9.2.10

Prove Theorem 9.2.13.

Theorem 9.2.13: Let {xn}n=1∞\left\{ x_{n}\right\}_{n=1}^{\infty } be a countable subset of [a,b][a,b] and let {cn}n=1∞\left\{ c_{n}\right\}_{n=1}^{\infty } be a sequence of positive numbers for which the series ∑n=1∞cn\sum_{n=1}^{\infty } c_{n} converges. Given x∈[a,b]x \in [a,b], let E(x)E(x) be the subset of N\mathbb {N} consisting of those integers nn for which xn<xx_{n} < x, and let f(x):=∑n∈E(x)cnf(x) := \sum_{n \in E(x)} c_{n} (with f(x)=0f(x) = 0 if E(x)=∅E(x) = \varnothing). Then: (a) ff is a monotone-increasing function on [a,b][a,b]. (b) ff has a jump discontinuity at each xnx_{n}. (c) The jump at x=xnx = x_{n} is cnc_{n}. (d) ff is continuous at every other point of [a,b][a,b].

?
Problem 9.2.11

Let EE be a subset of R\mathbb {R} and AA a countable subset of EE. Prove or disprove: There is a monotone function on EE that is continuous only at points in E\AE \backslash A.

?
Problem 9.2.12

Let ff be an increasing function on [a,b][a, b]. Prove that

m∗({x∈(a,b):D−f(x)=+∞})=0. m^{*}\left(\left\{ x \in (a, b): D^{-} f(x)=+\infty \right\} \right)=0 .
?
Problem 9.2.13

Prove inequality (9.20).

Inequality (9.20): If ff is an increasing function on (a,b)(a, b), then

D−f(x)≤D+f(x) almost everywhere. (9.20) D^{-} f(x) \leq D_{+} f(x) \quad \text{ almost everywhere. } \tag {9.20}

This is the second half of Lemma 9.2.7, whose first half reads: Let ff be an increasing function on [a,b][a, b] and T={x∈(a,b):D+f(x)>D−f(x)}T=\left\{ x \in (a, b): D^{+} f(x)>D_{-} f(x)\right\}. Then m∗(T)=0m^{*}(T)=0.

Definition 9.1.1: For a real-valued function ff on [a,b][a, b] and x0∈(a,b)x_{0} \in (a, b), the Dini derivatives are D+f(x0)=lim sup⁡x→x0+f(x)−f(x0)x−x0D^{+} f\left(x_{0}\right)=\limsup_{x \rightarrow x_{0}+} \frac{f(x)-f\left(x_{0}\right)}{x-x_{0}}, D+f(x0)=lim inf⁡x→x0+f(x)−f(x0)x−x0D_{+} f\left(x_{0}\right)=\liminf_{x \rightarrow x_{0}+} \frac{f(x)-f\left(x_{0}\right)}{x-x_{0}}, D−f(x0)=lim sup⁡x→x0−f(x)−f(x0)x−x0D^{-} f\left(x_{0}\right)=\limsup_{x \rightarrow x_{0}-} \frac{f(x)-f\left(x_{0}\right)}{x-x_{0}}, D−f(x0)=lim inf⁡x→x0−f(x)−f(x0)x−x0D_{-} f\left(x_{0}\right)=\liminf_{x \rightarrow x_{0}-} \frac{f(x)-f\left(x_{0}\right)}{x-x_{0}}.

?
Problem 9.2.14

Let E⊂[a,b]E \subset [a, b] be a set of measure 0 . Prove that there exists a continuous increasing function ff such that f′(x)=+∞f^{\prime }(x)=+\infty for all x∈Ex \in E.

?
Problem 9.2.15

Prove or disprove: the derivative of a continuous, monotone function is differentiable a.e.

?
Problem 9.2.16

Let ff be continuously on (a,b)(a, b) and let E={x∈(a,b):D−f(x)<D+f(x)}E=\left\{ x \in (a, b): D_{-} f(x)< D^{+} f(x)\right\}. Prove that EE cannot contain a nondegenerate interval.

?
Problem 9.2.17

Let f:R→Rf: \mathbb {R} \rightarrow \mathbb {R} be monotone increasing and let g(x)=lim⁡h→0+f(x+h)g(x)=\lim_{h \rightarrow 0+} f(x+ h). Prove that the function gg has the following properties:

?
(a)

it is defined everywhere;

(b)

it is monotone increasing;

(c)

it is right continuous;

(d)

it is differentiable a.e.;

(e)

f′=g′f^{\prime }=g^{\prime } a.e.

Problem 9.2.18

Let f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} be strictly increasing and E⊂[a,b]E \subset [a, b] such that at every point x∈E,D+f(x)≤Mx \in E, D^{+} f(x) \leq M. Prove that m∗f(E)≤Mm∗(E)m^{*} f(E) \leq M m^{*}(E).

?
Problem 9.2.19

Let f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} be strictly increasing and E⊂[a,b]E \subset [a, b] such that at every point x∈E,D−f(x)≥Cx \in E, D_{-} f(x) \geq C. Prove that m∗f(E)≥Cm∗(E)m^{*} f(E) \geq C m^{*}(E).

?
Problem 9.2.20

Prove that every monotone function defined on an interval is Borel measurable.

?
Problem 9.2.21

Let ff be continuous on R\mathbb {R}. Prove or disprove: there exists an open interval on which ff is monotone.

?
Problem 9.2.22

Let ff be differentiable on R\mathbb {R}, let x0∈Rx_{0} \in \mathbb {R}, and suppose that f′(x0)>0f^{\prime }\left(x_{0}\right)>0. Prove or disprove: there exists an open interval II containing x0x_{0} such that ff is monotone on II.

?
Problem 9.2.23

(Fubini) Let fk:[a,b]→Rf_{k}:[a, b] \rightarrow \mathbb {R} be monotone increasing functions for k=1,2,…k= 1,2, \ldots and assume that the series f(x)=∑k=1∞fk(x)f(x)=\sum_{k=1}^{\infty } f_{k}(x) converges pointwise on [a,b][a, b]. Prove that f′(x)=∑k=1∞fk′(x)f^{\prime }(x)=\sum_{k=1}^{\infty } f_{k}^{\prime }(x) for almost all x∈[a,b]x \in [a, b].

?
Problem 9.2.24

Let fk:(0,1)→Rf_{k}:(0,1) \rightarrow \mathbb {R} be monotone increasing for k=1,2,…k=1,2, \ldots and suppose that lim⁡nfn(x)=1\lim_{n} f_{n}(x)=1 for a.e. x∈(0,1)x \in (0,1). Prove that lim inf⁡nfn′(x)=0\liminf_{n} f_{n}^{\prime }(x)=0 for a.e. x∈(0,1)x \in (0,1).

?
Problem 9.2.25

Give an example of functions f,g:[0,1]→Rf, g:[0,1] \rightarrow \mathbb {R} such that ff is increasing and continuous, gg is a polynomial, and

∫[0,1]fg′≠f(1)g(1)−f(0)g(0)−∫[0,1]f′g \int _{[0,1]} f g^{\prime } \neq f(1) g(1)-f(0) g(0)-\int _{[0,1]} f^{\prime } g
?
Problem 9.2.26

Prove or disprove: if f:R2→Rf: \mathbb {R}^{2} \rightarrow \mathbb {R} is increasing in each variable then it is Borel measurable.

?
Problem 9.2.27

Let ϕ\phi denote the Cantor-Lebesgue function and set ϕ(x)=0\phi (x)=0 for x<0x<0 and ϕ(x)=1\phi (x)=1 for x>1x>1. Let {[an,bn]}\left\{ \left[a_{n}, b_{n}\right]\right\} be the collection of all closed subintervals of [0,1][0,1] with rational endpoints, let

Fn(x)=ϕ(x−anbn−an), F_{n}(x)=\phi \left(\frac{x-a_{n}}{b_{n}-a_{n}}\right),

and let G=∑1∞2−nFnG=\sum_{1}^{\infty } 2^{-n} F_{n}. Prove that GG is continuous and strictly increasing on [0,1][0,1] and that G′=0G^{\prime }=0 a.e.

?
Problem 9.2.28

The purpose of this exercise is to establish the so-called Helly Selection Principle: If {fn}\left\{ f_{n}\right\} is a uniformly bounded sequence of increasing functions on [0,1][0,1] then there exists a subsequence {fnk}\left\{ f_{n_{k}}\right\} that converges pointwise at every point of [0,1][0,1] to an increasing function ff.

?
(a)

Prove that there exists a subsequence {gn}\left\{ g_{n}\right\} of {fn}\left\{ f_{n}\right\} that converges pointwise at every rational point of [0,1][0,1].

(b)

Extend ff to irrational points in [0,1][0,1] so that it is increasing on [0,1][0,1].

(c)

Prove that the subsequence {gn}\left\{ g_{n}\right\} obtained in (a) converges at every point of continuity of ff.

(d)

Use the approach as in (a) and Theorem 9.2.12 to obtain a subsequence of {gn}\left\{ g_{n}\right\} that converges at every point of [0,1][0,1].

Theorem 9.2.12: Let ff be a monotone function on (a,b)(a, b). Then ff has at most a countable number of jump discontinuities on (a,b)(a, b).

Problem 9.2.29

Let {In}\left\{ I_{n}\right\} be a sequence of subintervals of [0,1][0,1] with rational endpoints. For each InI_{n}, let EnE_{n} be the set of all continuous functions on [0,1][0,1] that are monotone on InI_{n}. Using these sets proves that there exists a continuous function on [0,1][0,1] that is not monotone on any subinterval of [0,1][0,1].

?
Problem 9.2.30

Let A⊂[0,1]A \subset [0,1] be a set of the first category, and f:A→Rf: A \rightarrow \mathbb {R} a strictly monotone function. Prove that f(A)f(A) is a set of first categories.

?
Problem 9.3.1

Let ff be differentiable on [a,b][a, b] with bounded derivative. Prove that ff is of bounded variation on [a,b][a, b].

?
Problem 9.3.2

Prove that the Dirichlet function is not of bounded variation.

?
Problem 9.3.3

Let ff be a function of bounded variation on [a,b][a, b] and [c,d]⊂[a,b][c, d] \subset [a, b]. Prove that ff is of bounded variation on [c,d][c, d].

?
Problem 9.3.4

Prove that every function of bounded variation is bounded.

?
Problem 9.3.5

Let f,g:[a,b]→Rf, g:[a, b] \rightarrow \mathbb {R} be of bounded variation. Prove that so is fgf g.

?
Problem 9.3.6

Let f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} be of bounded variation. Prove that so is ∣f∣\left|f\right|.

?
Problem 9.3.7

Let f,g:[a,b]→Rf, g:[a, b] \rightarrow \mathbb {R} be of bounded variation and suppose that 1/g1 / g is bounded on [a,b][a, b]. Prove that f/gf / g is of bounded variation.

?
Problem 9.3.8

Let f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} and suppose that ff is of bounded variation on every closed subinterval of (a,b)(a, b). Prove or disprove: ff is of bounded variation on [a,b][a, b].

?
Problem 9.3.9

Let

f(x)={x3sin⁡(π/x), for x≠00 for x=0 f(x)= \begin{cases} \sqrt[3]{x} \sin (\pi / x), & \text{ for } x \neq 0 \\ 0 & \text{ for } x=0 \end{cases}

Determine whether ff is of bounded variation on [0,1][0,1].

?
Problem 9.3.10

Let

f(x)={xαsin⁡(xβ), if 0<x≤10 if x=0. f(x)= \begin{cases} x^{\alpha } \sin \left(x^{\beta }\right), & \text{ if } 0<x \leq 1 \\ 0 & \text{ if } x=0 . \end{cases}

Prove that ff is of bounded variation on [0,1][0,1] if and only if α+β>0\alpha +\beta >0 or both α+β=0\alpha +\beta =0 and β≥0\beta \geq 0.

?
Problem 9.3.11

Let gg be a function of bounded variation on [0,1][0,1] and let ff be a continuous function on R\mathbb {R}. Prove or disprove: f∘gf \circ g is of bounded variation on [0,1][0,1].

?
Problem 9.3.12

Let gg be a function of bounded variation on [0,1][0,1] and let ff be of bounded variation on R\mathbb {R}. Prove or disprove: f∘gf \circ g is of bounded variation on [0,1][0,1].

?
Problem 9.3.13

Let ff be a function on [0,1][0,1] that is not of bounded variation. Prove that there is a point x0∈[0,1]x_{0} \in [0,1] such that ff fails to be of bounded variation on each subinterval of [0,1][0,1] that contains x0x_{0} in its interior.

?
Problem 9.3.14

Give an example of a function that is differentiable on [0,1][0,1] but not of bounded variation on [0,1][0,1].

?
Problem 9.3.15

Let f,g:[a,b]→Rf, g:[a, b] \rightarrow \mathbb {R}. Prove that Vab(f+g)≤Vab(f)+Vab(g)V_{a}^{b}(f+g) \leq V_{a}^{b}(f)+V_{a}^{b}(g).

?
Problem 9.3.16

Let f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} and let α∈R\alpha \in \mathbb {R}. Prove that Vab(αf)=∣α∣Vab(f)V_{a}^{b}(\alpha f)=\left|\alpha \right| V_{a}^{b}(f).

?
Problem 9.3.17

Let f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R}. Prove that Vab(f)=0V_{a}^{b}(f)=0 if and only if ff is constant.

?
Problem 9.3.18

Let f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} be a function of bounded variation and let x0∈(a,b)x_{0} \in (a, b). Prove that if ff is continuous from the left (right) at x0x_{0}, then so is Vax(f)V_{a}^{x}(f).

?
Problem 9.3.19

For a function f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} the length of the curve y=f(x),a≤x≤by=f(x), a \leq x \leq b is defined by

L(f)=sup⁡{∑k=1n(xk−xk−1)2+(f(xk)−f(xk−1))2}, L(f)=\sup \left\{ \sum _{k=1}^{n} \sqrt{\left(x_{k}-x_{k-1}\right)^{2}+\left(f\left(x_{k}\right)-f\left(x_{k-1}\right)\right)^{2}}\right\} ,

where the supremum is taken over all partitions P={x0,x1,…,xn}P=\left\{ x_{0}, x_{1}, \ldots , x_{n}\right\} of [a,b][a, b]. Prove that L(f)L(f) is finite if and only ff is of bounded variation on [a,b][a, b].

?
Problem 9.3.20

Let ff be differentiable on [a,b][a, b] with derivative f′f^{\prime } Riemann integrable over [a,b][a, b]. Prove that

Vab(f)=∫ab∣f′∣ V_{a}^{b}(f)=\int _{a}^{b}\left|f^{\prime }\right|

What if the integral of ∣f′∣\left|f^{\prime }\right| is a convergent improper integral?

?
Problem 9.3.21

Let ff be continuous on [a,b][a, b] and suppose that D+fD^{+} f is Riemann integrable over [a,b][a, b]. Prove that

Vab(f)=∫ab∣D+f∣ V_{a}^{b}(f)=\int _{a}^{b}\left|D^{+} f\right|

What if the integral of ∣D+f∣\left|D^{+} f\right| is a convergent improper integral?

?
Problem 9.3.22

Let f,gf, g be differentiable on [a,b][a, b] and let C>0C>0 such that ∣f′∣≤C∣g′∣\left|f^{\prime }\right| \leq C\left|g^{\prime }\right| on [a,b][a, b]. Suppose that either g′g^{\prime } is Riemann integrable over [a,b][a, b] or the integral of ∣g′∣\left|g^{\prime }\right| is a convergent improper integral. Prove that ff is of bounded variation on [a,b][a, b].

?
Problem 9.3.23

In Exercises 9.3.23-9.3.26 find the total variation: V050(ex)V_{0}^{50}\left(e^{x}\right).

?
Problem 9.3.24

In Exercises 9.3.23-9.3.26 find the total variation: V12(ln⁡x)V_{1}^{2}(\ln x).

?
Problem 9.3.25

In Exercises 9.3.23-9.3.26 find the total variation: V04π(cos⁡x)V_{0}^{4 \pi }(\cos x).

?
Problem 9.3.26

In Exercises 9.3.23-9.3.26 find the total variation: V−11(x−x3)V_{-1}^{1}\left(x-x^{3}\right).

?
Problem 9.3.27

Let f,gf, g be continuous functions of bounded variation on [0,1][0,1]. Prove that the set {(f(x),g(x)):x∈[0,1]}\left\{ (f(x), g(x)): x \in [0,1]\right\} cannot be the square [0,1]×[0,1][0,1] \times [0,1]. What if f,gf, g are merely continuous?

?
Problem 9.3.28

Let {fn}\left\{ f_{n}\right\} be a sequence of real-valued functions on [a,b][a, b] that converges pointwise on [a,b][a, b] to the real-valued function ff. Prove that

Vabf≤lim inf⁡Vabfn V_{a}^{b} f \leq \liminf V_{a}^{b} f_{n}
?
Problem 9.3.29

Find all functions ff of bounded variation on [0,1][0,1] such that

f(x)+(V0xf)1/2=1, for all x∈[0,1], and ∫01f(x)dx=13 f(x)+\left(V_{0}^{x} f\right)^{1 / 2}=1, \quad \text{ for all } x \in [0,1], \quad \text{ and } \quad \int _{0}^{1} f(x) d x=\frac{1}{3}
?
Problem 9.3.30

Let ff be a function of bounded variation on a closed and bounded interval [a,b][a, b]. Prove that there exists a sequence of partitions {Pn}\left\{ P_{n}\right\} of [a,b][a, b] for which the sequence {V(f,Pn)}\left\{ V\left(f, P_{n}\right)\right\} is increasing and converges to Vab(f)V_{a}^{b}(f).

?
Problem 9.3.31

Let f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} be a continuous function of bounded variation and let P={x0,x1,…,xn}P=\left\{ x_{0}, x_{1}, \ldots , x_{n}\right\} be a partition of [a,b][a, b]. For 1≤k≤n1 \leq k \leq n, let MkM_{k} and mkm_{k} denote the maximum and the minimum, respectively, of ff on [xk−1,xk]\left[x_{k-1}, x_{k}\right], and Ω(f,P)=∑k=1n(Mk−mk)\Omega (f, P)=\sum_{k=1}^{n}\left(M_{k}-m_{k}\right). Prove that

lim⁡∣P∣→0V(f,P)=Vabf and lim⁡∣P∣→0Ω(f,P)=Vabf. \lim _{\left|P\right| \rightarrow 0} V(f, P)=V_{a}^{b} f \quad \text{ and } \quad \lim _{\left|P\right| \rightarrow 0} \Omega (f, P)=V_{a}^{b} f .
?
Problem 9.3.32

Prove or disprove: There exists a continuous function that is not of bounded variation on any interval.

?
Problem 9.3.33

Let f(x)=sin⁡xf(x)=\sin x on [0,2π][0,2 \pi ]. Find two increasing functions hh and gg for which f=h−gf=h-g on [0,2π][0,2 \pi ].

?
Problem 9.3.34

Let f(x)=3x2−2x3f(x)=3 x^{2}-2 x^{3} on [-2,2]. Find two increasing functions hh and gg for which f=h−gf=h-g on [−2,2][-2,2].

?
Problem 9.3.35

Let f(x)=⌊x⌋−xf(x)=\left\lfloor x\right\rfloor -x on [0,2][0,2]. Find two increasing functions hh and gg for which f=h−gf=h-g on [ 0,2 ].

?
Problem 9.3.36

Let ff be a function of bounded variation on a closed and bounded interval [a,b][a, b]. Prove or disprove: ff is a measurable function.

?
Problem 9.3.37

Let ff be a function of bounded variation on a closed and bounded interval [a,b][a, b]. Prove that there exist two bounded increasing functions f1f_{1} and f2f_{2} such that f=f1−f2f=f_{1}-f_{2} and f1+f2=Vabff_{1}+f_{2}=V_{a}^{b} f.

?
Problem 9.3.38

Let ff be a function of bounded variation on a closed and bounded interval [a,b][a, b]. Prove that ∣f′(x)∣≤(Vax(f))′\left|f^{\prime }(x)\right| \leq \left(V_{a}^{x}(f)\right)^{\prime } for a.e. x∈[a,b]x \in [a, b].

?
Problem 9.3.39

Let ff be a function of bounded variation on a closed and bounded interval [a,b][a, b] and suppose that ff has the intermediate value property, i.e., if x,y∈[a,b]x, y \in [a, b] and CC is a number between f(x)f(x) and f(y)f(y) then there is z∈[x,y]z \in [x, y] such that f(z)=Cf(z)=C. Prove that ff is continuous.

?
Problem 9.3.40

Let ff be a continuous function on [a,b][a, b] and suppose that ∣f∣\left|f\right| has a bounded variation on [a,b][a, b]. Prove that ff has bounded variation on [a,b][a, b].

?
Problem 9.4.1

Prove that f(x)=xf(x)=\sqrt{x} is absolutely continuous on [0,1][0,1] but not Lipschitz.

?
Problem 9.4.2

Let f,g:[a,b]→Rf, g:[a, b] \rightarrow \mathbb {R} be absolutely continuous and let α,β∈R\alpha , \beta \in \mathbb {R}. Prove that αf+βg\alpha f+\beta g is absolutely continuous on [a,b][a, b].

?
Problem 9.4.3

Let f,g:[a,b]→Rf, g:[a, b] \rightarrow \mathbb {R} be absolutely continuous. Prove that fgf g is absolutely continuous on [a,b][a, b].

?
Problem 9.4.4

Let f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} be absolutely continuous and suppose that f(x)≠0f(x) \neq 0 for all x∈[a,b]x \in [a, b]. Prove that 1/f1 / f is absolutely continuous on [a,b][a, b].

?
Problem 9.4.5

Let gg be a strictly increasing absolutely continuous function on [a,b][a, b] and let ff be an absolutely continuous function on [g(a),g(b)][g(a), g(b)]. Prove that f∘gf \circ g is absolutely continuous on [a,b][a, b].

?
Problem 9.4.6

Prove that ff is absolutely continuous on [a,b][a, b] if and only if for every ε>0\varepsilon >0 there exists δ>0\delta >0 such that, for every collection {[ak,bk]}k=1∞\left\{ \left[a_{k}, b_{k}\right]\right\}_{k=1}^{\infty } of nonoverlapping closed intervals in [a,b][a, b],

∑k=1∞(bk−ak)<δ⇒∑k=1∞∣f(bk)−f(ak)∣<ε. \sum _{k=1}^{\infty }\left(b_{k}-a_{k}\right)<\delta \Rightarrow \quad \sum _{k=1}^{\infty }\left|f\left(b_{k}\right)-f\left(a_{k}\right)\right|<\varepsilon .
?
Problem 9.4.7

Let {Gn}\left\{ G_{n}\right\} be a decreasing sequence of open subsets of [0,1][0,1] such that m(Gn)<1/2nm\left(G_{n}\right)<1 / 2^{n}, and let f(x)=∑n=1∞m(Gn∩[0,x])f(x)=\sum_{n=1}^{\infty } m\left(G_{n} \cap [0, x]\right). Prove that:

?
(a)

f∈AC[0,1]f \in A C[0,1];

(b)

ff is Lipschitz on [0,1][0,1] if and only if there is N∈NN \in \mathbb {N} such that Gn=∅G_{n}=\emptyset for all n≥Nn \geq N.

Problem 9.4.8

Let {fn}\left\{ f_{n}\right\} be a sequence of absolutely continuous functions on [0,1][0,1] that converge to a function ff uniformly on [0,1][0,1]. Prove or disprove: ff is absolutely continuous.

?
Problem 9.4.9

Let ff be a Lipschitz function on [a,b][a, b] with constant MM and let E⊂[a,b]E \subset [a, b]. Prove that m(f(E))≤Mm(E)m(f(E)) \leq M m(E). Conclude that if ff is a Lipschitz function on [a,b][a, b] then it has the Luzin N property. Is the converse true?

Definition 9.4.20: A real-valued function defined on [a,b][a, b] has the Luzin N property, if it maps every set of measure zero to a set of measure 0.

?
Problem 9.4.10

Prove that every Lipschitz function maps measurable sets to measurable sets.

?
Problem 9.4.11

Give an example of a function without the Luzin N property.

Definition 9.4.20: A real-valued function defined on [a,b][a, b] has the Luzin N property, if it maps every set of measure zero to a set of measure 0.

?
Problem 9.4.12

(Banach-Zarecki) Prove that a function ff on the interval [a,b][a, b] is absolutely continuous if and only if it is continuous, is of bounded variation, and has the Luzin N property.

Definition 9.4.20: A real-valued function defined on [a,b][a, b] has the Luzin N property, if it maps every set of measure zero to a set of measure 0.

?
Problem 9.4.13

Prove that an absolutely continuous function maps a measurable set to a measurable set.

?
Problem 9.4.14

Prove that if the word nonoverlapping is deleted in Definition 9.4.1 the function must be Lipschitz. More precisely, let f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} be a function with the property that for every ε>0\varepsilon >0 there exists δ>0\delta >0 such that, for every collection {[ak,bk]}k=1n\left\{ \left[a_{k}, b_{k}\right]\right\}_{k=1}^{n} of closed intervals in [a,b][a, b],

∑k=1n(bk−ak)<δ⇒∑k=1n∣f(bk)−f(ak)∣<ε. \sum _{k=1}^{n}\left(b_{k}-a_{k}\right)<\delta \Rightarrow \sum _{k=1}^{n}\left|f\left(b_{k}\right)-f\left(a_{k}\right)\right|<\varepsilon .

Then ff is Lipschitz on [a,b][a, b].

Definition 9.4.1: A real-valued function ff on [a,b][a, b] is absolutely continuous on [a,b][a, b] if for every ε>0\varepsilon >0 there exists δ>0\delta >0 such that, for every collection {[ak,bk]}k=1n\left\{ \left[a_{k}, b_{k}\right]\right\}_{k=1}^{n} of closed nonoverlapping intervals in [a,b][a, b], ∑k=1n(bk−ak)<δ⇒∑k=1n∣f(bk)−f(ak)∣<ε\sum_{k=1}^{n}\left(b_{k}-a_{k}\right)<\delta \Rightarrow \sum_{k=1}^{n}\left|f\left(b_{k}\right)-f\left(a_{k}\right)\right|<\varepsilon. Here nonoverlapping means with disjoint interiors.

?
Problem 9.4.15

Let f:[0,1]→Rf:[0,1] \rightarrow \mathbb {R} be an increasing continuous function that is absolutely continuous on [ε,1][\varepsilon , 1] for each ε∈(0,1)\varepsilon \in (0,1). Prove that ff is absolutely continuous on [0,1][0,1].

?
Problem 9.4.16

Let f(x)=xf(x)=\sqrt{x} and

g(x)={x2sin⁡(1x), if x≠00, if x=0 g(x)= \begin{cases} x^{2} \sin \left(\frac{1}{x}\right), & \text{ if } x \neq 0 \\ 0, & \text{ if } x=0 \end{cases}

Prove that ff and gg are absolutely continuous on [0,1][0,1] but h:=f∘gh:=f \circ g is not even of bounded variation on [0,1][0,1].

?
Problem 9.4.17

Suppose that ff is a function on [a,b][a, b], and let Vax(f)V_{a}^{x}(f) be the total variation function for ff. Prove that VaxV_{a}^{x} is absolutely continuous on [a,b][a, b] if and only if ff is. Conclude that every absolutely continuous function is the difference of increasing absolutely continuous functions.

?
Problem 9.4.18

Prove that a function ff is Lipschitz on [a,b][a, b] if and only if it is absolutely continuous function on [a,b][a, b] with ∣f′∣\left|f^{\prime }\right| essentially bounded. (A function is essentially bounded if it is equal a.e. to a bounded function.)

?
Problem 9.4.19

Give an example of a function of bounded variation that is not absolutely continuous.

?
Problem 9.4.20

Let ff be a function of bounded variation on [a,b][a, b]. Prove that Vab(f)=∫ab∣f′∣V_{a}^{b}(f)= \int_{a}^{b}\left|f^{\prime }\right| if and only if ff is absolutely continuous on [a,b][a, b].

?
Problem 9.4.21

Let g:[a,b]→Rg:[a, b] \rightarrow \mathbb {R} be a strictly increasing, absolutely continuous function. Prove that for any measurable set E⊂[a,b]E \subset [a, b]

m(g(E))=∫Eg′(x)dx m(g(E))=\int _{E} g^{\prime }(x) d x
?
Problem 9.4.22

Let f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R}, be integrable and let g:[α,β]→Rg:[\alpha , \beta ] \rightarrow \mathbb {R} be a strictly increasing, absolutely continuous function for which g(α)=a,g(β)=bg(\alpha )=a, g(\beta )=b. Prove that

∫abf(x)dx=∫αβf(g(t))g′(t)dt \int _{a}^{b} f(x) d x=\int _{\alpha }^{\beta } f(g(t)) g^{\prime }(t) d t
?
Problem 9.4.23

Prove Corollary 9.4.14.

Corollary 9.4.14: Let f:[a,b]→Rf: [a,b] \rightarrow \mathbb {R} be a function that is differentiable a.e. with an integrable derivative. If ∫axf′=f(x)−f(a)\int_{a}^{x} f' = f(x) - f(a) for all x∈[a,b]x \in [a,b], then ff is absolutely continuous on [a,b][a,b].

?
Problem 9.4.24

Prove Corollary 9.4.16.

Corollary 9.4.16: A function ff is absolutely continuous on [a,b][a,b] if and only if it is an indefinite integral over [a,b][a,b]. (A function FF is the indefinite (Lebesgue) integral of an integrable function ff over [a,b][a,b] if F(x)=F(a)+∫axfF(x) = F(a) + \int_{a}^{x} f for all x∈[a,b]x \in [a,b].)

?
Problem 9.4.25

If f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} is continuous and real-valued, f′f^{\prime } integrable on [a,b][a, b], and ∫abf′=f(b)−f(a)\int_{a}^{b} f^{\prime }=f(b)-f(a), must ff be absolutely continuous? What if ff is monotone increasing?

?
Problem 9.4.26

Let f:R→Rf: \mathbb {R} \rightarrow \mathbb {R} be a function that is absolutely continuous on each finite interval, and suppose that both ff and f′f^{\prime } are integrable over R\mathbb {R}. Prove that:

?
(a)

∫Rf′=0\int_{\mathbb {R}} f^{\prime }=0;

(b)

f(x)→0f(x) \rightarrow 0 when ∣x∣→∞\left|x\right| \rightarrow \infty.

Problem 9.4.27

Let {fn}⊂AC[0,1]\left\{ f_{n}\right\} \subset A C[0,1] and suppose that

  1. f=∑n=1∞fnf=\sum_{n=1}^{\infty } f_{n} converges for every x∈[0,1]x \in [0,1];

  2. ∫01∑n=1∞∣fn′∣\int_{0}^{1} \sum_{n=1}^{\infty }\left|f_{n}^{\prime }\right| is finite.

Prove that f∈AC[0,1]f \in A C[0,1].

?
Problem 9.4.28

The purpose of this exercise is to establish the following result: If ff is differentiable for all x∈[a,b]x \in \left[a, b\right] and f′f^{\prime } is integrable over [a,b]\left[a, b\right] then (9.30) holds.

Formula (9.30), from Theorem 9.4.15: ∫axf′=f(x)−f(a)\int_{a}^{x} f^{\prime }=f(x)-f(a), for a≤x≤ba \leq x \leq b.

?
(a)

Let

gn(x)={f′(x),if f′(x)≤n,n,if f′(x)>n,andGn(x)=f(x)−∫axgn. g_{n}(x) = \begin{cases} f^{\prime }(x), & \text{if } f^{\prime }(x) \leq n, \\ n, & \text{if } f^{\prime }(x) > n, \end{cases} \quad \text{and} \quad G_{n}(x) = f(x) - \int _{a}^{x} g_{n}.

Prove that GnG_{n} is an increasing function, for each n∈Nn \in \mathbb {N}.

(b)

Use the result in (a) to prove that

∫abf′≤f(b)−f(a) \int _{a}^{b} f^{\prime } \leq f(b)-f(a)
(c)

Replace gng_{n} and GnG_{n} in (a) by

hn(x)={f′(x),if f′(x)≥−n,−n,if f′(x)<−n,andHn(x)=f(x)−∫axhn. h_{n}(x) = \begin{cases} f^{\prime }(x), & \text{if } f^{\prime }(x) \geq -n, \\ -n, & \text{if } f^{\prime }(x) < -n, \end{cases} \quad \text{and} \quad H_{n}(x) = f(x) - \int _{a}^{x} h_{n}.

and show that {Hn}\left\{ H_{n}\right\} is a decreasing sequence.

(d)

Use (c) to obtain

∫abf′≥f(b)−f(a) \int _{a}^{b} f^{\prime } \geq f(b)-f(a)
Problem 9.4.29

Let ff be differentiable on [a,b][a, b] with derivative f′f^{\prime } integrable over [a,b][a, b]. Prove that

Vab(f)=∫ab∣f′∣ V_{a}^{b}(f)=\int _{a}^{b}\left|f^{\prime }\right|
?
Problem 9.4.30

Let ff be a function on [a,b][a, b] for which D+fD^{+} f is integrable over [a,b][a, b]. Prove or disprove: Vab(f)=∫ab∣D+f∣V_{a}^{b}(f)=\int_{a}^{b}\left|D^{+} f\right|.

?
Problem 9.4.31

Let ff and gg be absolutely continuous functions on [a,b][a, b]. Show that

∫abfg′=f(b)g(b)−f(a)g(a)−∫abf′g \int _{a}^{b} f g^{\prime }=f(b) g(b)-f(a) g(a)-\int _{a}^{b} f^{\prime } g
?
Problem 9.4.32

Let ff be integrable over R\mathbb {R}. Prove that for a.e. x∈Rx \in \mathbb {R}

lim⁡h→0+1h∫(x−h,x+h)∣f−f(x)∣dm=0 \lim _{h \rightarrow 0^{+}} \frac{1}{h} \int _{(x-h, x+h)}\left|f-f(x)\right| d m=0
?
Problem 9.4.33

Let ff be integrable over R\mathbb {R} and suppose that ff is continuous at x=ax=a. Prove that the indefinite integral of ff is differentiable at x=ax=a.

?
Problem 9.4.34

Let AA be a measurable subset of R\mathbb {R} and suppose that there is α>0\alpha >0 such that m(A∩[a,b])≥α(b−a)m(A \cap [a, b]) \geq \alpha (b-a), whenever a<ba<b. Prove that m(AC)=0m\left(A^{C}\right)=0

?
Problem 9.4.35

Let ff be absolutely continuous and singular on [a,b][a, b]. Show that ff is constant.

?
Problem 9.4.36

When Lebesgue defined a singular function it was the following way: FsF_{s} is the singular function associated with a function of bounded variation FF, if the difference F−FsF-F_{s} is absolutely continuous and among all the functions GsG_{s} such that F−GsF-G_{s} is absolutely continuous, FsF_{s} has the smallest total variation, and Fs(a)=0F_{s}(a)=0. Prove that the decomposition F=(F−Fs)+FsF=\left(F-F_{s}\right)+F_{s} is the same as the one provided by Theorem 9.4.19.

Theorem 9.4.19: A continuous function ff of bounded variation on [a,b][a, b] can be represented in a unique way as f=g+hf=g+h where gg is absolutely continuous on [a,b][a, b] satisfying g(a)=f(a)g(a)=f(a), and hh is either a singular function or the zero function. (Definition 9.4.17: a function f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} is singular if it is continuous and of bounded variation on [a,b][a, b], its derivative vanishes almost everywhere, and it is non-constant on [a,b][a, b].)

?
Problem 9.4.37

Let AA be a measurable subset of [0,1][0,1] such that 0<m(A∩I)<m(I)0<m(A \cap I)<m(I) for every interval I⊂[0,1]I \subset [0,1] (see Exercise 5.3.47), and let F(x):=m([0,x]∩A)F(x):=m([0, x] \cap A), G(x):=m([0,x]∩A)−m([0,x]∩AC)G(x):=m([0, x] \cap A)-m\left([0, x] \cap A^{C}\right). Prove that:

?
(a)

FF is absolutely continuous and strictly increasing on [0,1][0,1], but F′=0F^{\prime }=0 on a set of positive measure.

(b)

GG is absolutely continuous on [0,1][0,1], but it is not monotone on any subinterval of [0,1][0,1].

Problem 9.4.38

Show that the condition imposed on the set AA in Exercise 9.4.37 (a) is not necessary, i.e., that there is a set that does not satisfy it, yet F(x):=m([0,x]∩A)F(x):= m([0, x] \cap A) is an absolutely continuous, strictly increasing function on [0,1][0,1] such that F′=0F^{\prime }=0 on a set of positive measures.

?
Problem 9.4.39

(Vitali) Let ff be an increasing function on a closed and bounded interval [a,b][a, b]. Prove that ff is singular if and only if for any ε,δ>0\varepsilon , \delta >0 there is a finite disjoint collection {(ak,bk)}k=1n\left\{ \left(a_{k}, b_{k}\right)\right\}_{k=1}^{n} of open intervals in ( a,ba, b ) for which

∑k=1n(bk−ak)<δ⇒∑k=1n[f(bk)−f(ak)]>f(b)−f(a)−ε. \sum _{k=1}^{n}\left(b_{k}-a_{k}\right)<\delta \Rightarrow \sum _{k=1}^{n}\left[f\left(b_{k}\right)-f\left(a_{k}\right)\right]>f(b)-f(a)-\varepsilon .
?
Problem 9.4.40

Give an example of a function f:[a,b]→Rf:[a, b] \rightarrow \mathbb {R} that is differentiable on [a,b][a, b] but its derivative is not integrable over [a,b][a, b].

?
Problem 9.4.41

Let f:R→Rf: \mathbb {R} \rightarrow \mathbb {R} be an increasing function such that f(−∞)=0f(-\infty )=0 and f(∞)=1f(\infty )=1. Prove that ff is absolutely continuous on every closed finite interval if and only if ∫Rf′=1\int_{\mathbb {R}} f^{\prime }=1.

?
Problem 9.5.1

Let (X,M,μ)(X, \mathfrak {M}, \mu ) be a measure space and ν\nu be a signed measure on (X,M)(X, \mathfrak {M}). Prove that the following are equivalent:

  1. ν\nu is absolutely continuous with respect to μ\mu.

  2. Both ν+\nu^{+}and ν−\nu^{-}are absolutely continuous with respect to μ\mu.

  3. ∣ν∣\left|\nu \right| is absolutely continuous with respect to μ\mu.

?
Problem 9.5.2

Let (X,M,μ)(X, \mathfrak {M}, \mu ) be a measure space, let {νk}\left\{ \nu_{k}\right\} be measures on ( X,MX, \mathfrak {M} ), and let {ak}\left\{ a_{k}\right\} be nonnegative real numbers. Prove that

νk≪μ for all n∈N⇒∑k=1∞akνk≪μ. \nu _{k} \ll \mu \text{ for all } n \in \mathbb {N} \Rightarrow \sum _{k=1}^{\infty } a_{k} \nu _{k} \ll \mu .

Show that the assertion remains true when {ak}\left\{ a_{k}\right\} are not necessarily nonnegative if we add the assumption that the series ∑k=1∞akνk\sum_{k=1}^{\infty } a_{k} \nu_{k} converges.

?
Problem 9.5.3

Let ( X,M,μX, \mathfrak {M}, \mu ) be a measure space, let {νk}\left\{ \nu_{k}\right\} be measured on ( X,MX, \mathfrak {M} ), and let {ak}\left\{ a_{k}\right\} be nonnegative real numbers. Prove that

νk⊥μ for all n∈N⇒∑k=1∞akνk⊥μ \nu _{k} \perp \mu \text{ for all } n \in \mathbb {N} \Rightarrow \sum _{k=1}^{\infty } a_{k} \nu _{k} \perp \mu

Show that the assertion remains true when {ak}\left\{ a_{k}\right\} are not necessarily nonnegative if we add the assumption that the series ∑k=1∞akνk\sum_{k=1}^{\infty } a_{k} \nu_{k} converges.

?
Problem 9.5.4

Let (X,M,μ)(X, \mathfrak {M}, \mu ) be a measure space and ν\nu be a signed measure on (X,M)(X, \mathfrak {M}). Suppose that ν\nu is absolutely continuous with respect to μ\mu and that μ\mu and ν\nu are mutually singular. Prove that ν\nu is the zero measure, i.e., ν(E)=0\nu (E)=0 for all E∈ME \in \mathfrak {M}.

?
Problem 9.5.5

Let ( X,M,μX, \mathfrak {M}, \mu ) be a measure space and f1,f2f_{1}, f_{2} two integrable functions on XX. For k=1,2k=1,2, let Ak={x∈X:fk(x)≠0}A_{k}=\left\{ x \in X: f_{k}(x) \neq 0\right\}, and define signed measures ν1,ν2\nu_{1}, \nu_{2} by

νk(E)=∫Efkdμ \nu _{k}(E)=\int _{E} f_{k} d \mu
?
(a)

Prove that if ν1\nu_{1} and ν2\nu_{2} are equivalent then μ(A1△A2)=0\mu \left(A_{1} \triangle A_{2}\right)=0.

(b)

Assuming that f1,f2f_{1}, f_{2} are nonnegative prove that μ(A1△A2)=0\mu \left(A_{1} \triangle A_{2}\right)=0 implies the equivalence of ν1\nu_{1} and ν2\nu_{2}.

(c)

Give an example to show that (b) is false without the nonnegativity assumption.

Problem 9.5.6

Let hh be an absolutely continuous function on [a,b][a, b] and suppose that μ\mu is a measure on Borel subsets of the interval [a,b][a, b] such that μ[a,x]=h(x)\mu [a, x]=h(x), for x∈[a,b]x \in [a, b]. Prove that for any c∈[a,b],μ({c})=0c \in [a, b], \mu (\left\{ c\right\} )=0.

?
Problem 9.5.7

Let (X,M,μ)(X, \mathfrak {M}, \mu ) be a σ\sigma-finite measure space. Prove that there exists a finite measure ν\nu on M\mathfrak {M} that is equivalent to μ\mu.

?
Problem 9.5.8

Let (X,M)(X, \mathfrak {M}) be a measurable space and μ,ν\mu , \nu finite measures on M\mathfrak {M}. Prove that the following are equivalent:

  1. μ⊥ν\mu \perp \nu;

  2. μ−(μ−ν)+=0\mu -(\mu -\nu )^{+}=0;

  3. μ+(ν−μ)+=μ+ν\mu +(\nu -\mu )^{+}=\mu +\nu.

?
Problem 9.5.9

Let {rn}\left\{ r_{n}\right\} be an ordering of rational numbers, let fn:R→Rf_{n}: \mathbb {R} \rightarrow \mathbb {R} be nonnegative, Borel-measurable functions such that fn=0f_{n}=0 outside of the interval (rn−2−n,rn+2−n)\left(r_{n}-2^{-n}, r_{n}+2^{-n}\right) and ∫Rfn=1\int_{\mathbb {R}} f_{n}=1, and let μ(E)=∫E∑nfn\mu (E)=\int_{E} \sum_{n} f_{n} for all E∈BE \in \mathcal{B}. Prove that:

?
(a)

∑nfn(x)\sum_{n} f_{n}(x) is finite a.e.

(b)

μ\mu is a σ\sigma-finite measure on B\mathcal{B}.

(c)

μ(G)=∞\mu (G)=\infty for every nonempty, open set GG.

(d)

μ≪m\mu \ll m.

Problem 9.5.10

Give an example to show that Proposition 9.5.3 fails if ν\nu is a σ\sigma-finite measure.

Proposition 9.5.3: Let (X,M,μ)(X, \mathfrak {M}, \mu ) be a measure space and ν\nu a finite measure on (X,M)(X, \mathfrak {M}). Then ν\nu is absolutely continuous with respect to μ\mu if and only if for every ε>0\varepsilon > 0 there is δ>0\delta > 0 such that for every E∈ME \in \mathfrak {M}, μ(E)<δ\mu \left(E\right) < \delta implies ν(E)<ε\nu \left(E\right) < \varepsilon.

?
Problem 9.5.11

Give an example of 2σ2 \sigma-algebras M\mathfrak {M} and N\mathfrak {N} on a set XX and 2 measures μ\mu and λ\lambda such that λ≪μ\lambda \ll \mu on ( X,MX, \mathfrak {M} ) but not on ( X,NX, \mathfrak {N} ).

?
Problem 9.5.12

Let μ\mu be a nonzero finite Borel measure on R\mathbb {R} and for every t∈Rt \in \mathbb {R}, let μt\mu_{t} be a measure defined by μt(A)=μ(A+t)\mu_{t}(A)=\mu (A+t). Suppose that each measure μt\mu_{t} is equivalent to μ\mu. Prove that μ\mu is equivalent to the Lebesgue measure on R\mathbb {R}.

?
Problem 9.5.13

Prove or disprove: there exists a measure ν\nu on Lebesgue σ\sigma-algebra M\mathcal{M} of subsets of R\mathbb {R} that is absolutely continuous with respect to Lebesgue measure and ν(A)=∞\nu (A)=\infty for any A∈MA \in \mathcal{M} with a nonempty interior.

?
Problem 9.5.14

Let ( X,M,μX, \mathfrak {M}, \mu ) be a finite measure space, let {Ek}k=1n\left\{ E_{k}\right\}_{k=1}^{n} a collection of measurable sets, and {ck}k=1n\left\{ c_{k}\right\}_{k=1}^{n} a collection of real numbers. For E∈ME \in \mathfrak {M} define

ν(E)=∑k=1nckμ(E∩Ek). \nu (E)=\sum _{k=1}^{n} c_{k} \mu \left(E \cap E_{k}\right) .
?
(a)

Prove that ν\nu is a measure on (X,M)(X, \mathfrak {M});

(b)

prove that ν\nu is absolutely continuous with respect to μ\mu;

(c)

find the Radon-Nikodym derivative dν/dμd \nu / d \mu.

Problem 9.5.15

Let μ,ν\mu , \nu, and λ\lambda be σ\sigma-finite measures on the measurable space ( X,MX, \mathfrak {M} ).

?
(a)

If ν≪μ\nu \ll \mu and ff is a measurable function on XX, prove that

∫Xfdν=∫Xfdνdμdμ \int _{X} f d \nu =\int _{X} f \frac{d \nu }{d \mu } d \mu
(b)

If ν≪μ\nu \ll \mu and λ≪μ\lambda \ll \mu prove that

d(ν+λ)dμ=dνdμ+dλdμ,a.e. [μ] \frac{d(\nu +\lambda )}{d \mu }=\frac{d \nu }{d \mu }+\frac{d \lambda }{d \mu }, \quad \text{a.e. }[\mu ]
(c)

If ν≪μ≪λ\nu \ll \mu \ll \lambda, prove that

dνdλ=dνdμdμdλ,a.e. [λ] \frac{d \nu }{d \lambda }=\frac{d \nu }{d \mu } \frac{d \mu }{d \lambda }, \quad \text{a.e. }[\lambda ]
(d)

If ν≪μ\nu \ll \mu and μ≪ν\mu \ll \nu, prove that

dνdμdμdν=1,a.e. [μ] \frac{d \nu }{d \mu } \frac{d \mu }{d \nu }=1, \quad \text{a.e. }[\mu ]
Problem 9.5.16

Let μ1,ν1\mu_{1}, \nu_{1} be measured on a measurable space ( X,SX, \mathcal{S} ) and let μ2,ν2\mu_{2}, \nu_{2} be measured on a measurable space ( Y,TY, \mathcal{T} ). Suppose that all 4 measures are σ\sigma-finite and that νk≪μk,k=1,2\nu_{k} \ll \mu_{k}, k=1,2.

?
(a)

Prove that ν1×ν2≪μ1×μ2\nu_{1} \times \nu_{2} \ll \mu_{1} \times \mu_{2}.

(b)

Find d(ν1×ν2)d(μ1×μ2)\frac{d\left(\nu_{1} \times \nu_{2}\right)}{d\left(\mu_{1} \times \mu_{2}\right)}.

Problem 9.5.17

Let (X,M)(X, \mathfrak {M}) be a measurable space and μ,νσ\mu , \nu \sigma-finite measures on M\mathfrak {M} such that ν≪μ\nu \ll \mu, let λ=μ+ν\lambda =\mu +\nu, and let f=dν/dλf=d \nu / d \lambda. Prove that 0≤f<10 \leq f<1 a.e. [μ][\mu ] and dν/dμ=f/(1−f)d \nu / d \mu =f /(1-f).

?
Problem 9.5.18

Let (X,M,μ)(X, \mathfrak {M}, \mu ) be a measure space, N\mathfrak {N} a subalgebra of M,ν\mathfrak {M}, \nu a restriction of μ\mu to N\mathfrak {N}, and ff a μ\mu-integrable function on XX. Prove that:

?
(a)

there exists a ν\nu-integrable function function gg on XX such that ∫Efdμ=∫Egdν\int_{E} f d \mu = \int_{E} g d \nu for all E∈NE \in \mathfrak {N};

(b)

gg is unique, meaning that if hh satisfies (a) then g=hg=h a.e. [ν][\nu ].

Problem 9.5.19

Let mm be Lebesgue measure on the real line and for each Lebesgue measurable subset EE of R\mathbb {R} define

μ(E)=∫E11+x2dm \mu (E)=\int _{E} \frac{1}{1+x^{2}} d m

Show that mm is absolutely continuous with respect to μ\mu, and compute the Radon-Nikodym derivative dm/dμd m / d \mu.

?
Problem 9.5.20

Let mm denote the Lebesgue measure on [0,+∞][0,+\infty ] and

μ1(E)=∑n=1∞1n3∫E∩[n,n+1]xdm,μ2(E)=∫E∩[1,+∞]1x2dm \mu _{1}(E)=\sum _{n=1}^{\infty } \frac{1}{n^{3}} \int _{E \cap [n, n+1]} x d m, \quad \mu _{2}(E)=\int _{E \cap [1,+\infty ]} \frac{1}{x^{2}} d m

for any Lebesgue measurable set E⊂[0,+∞]E \subset [0,+\infty ]. Prove or disprove:

?
(a)

m≪μ2m \ll \mu_{2};

(b)

μ2≪μ1\mu_{2} \ll \mu_{1};

(c)

μ1≪μ2\mu_{1} \ll \mu_{2}.

Find the Radon-Nikodym derivative in the cases when it exists.

Problem 9.5.21

Let ( X,MX, \mathfrak {M} ) be a measurable space and μ,ν\mu , \nu measures on M\mathfrak {M} such that μ\mu is σ\sigma-finite and ν≪μ\nu \ll \mu, let f=dν/dμf=d \nu / d \mu, and suppose that there is a constant C>0C>0 such that ν(E)>Cμ(E)\nu (E)>C \mu (E) for every set EE of finite measure. Prove that f>0f>0 for μ\mu-a.e. x∈Xx \in X.

?
Problem 9.5.22

Let ν\nu be a finite signed measure on a measurable space ( X,MX, \mathfrak {M} ). Prove that there exists a measurable function ff on XX such that ∣f(x)∣=1\left|f(x)\right|=1 for a.e. x∈Xx \in X, and

ν(E)=∫Efd∣ν∣, for any E∈M \nu (E)=\int _{E} f d\left|\nu \right|, \quad \text{ for any } E \in \mathfrak {M}
?
Problem 9.5.23

Let (X,M)(X, \mathfrak {M}) be a measurable space and let μ,ν\mu , \nu be two σ\sigma-finite measures on M\mathfrak {M} such that μ≪ν\mu \ll \nu and there is a set A∈MA \in \mathfrak {M} which is a ν\nu-atom (meaning that for any measurable subset B⊂AB \subset A we have either ν(B)=0\nu (B)=0 or ν(A\B)=\nu (A \backslash B)= 0 ). Determine the values of the Radon-Nikodym derivative dμ/dνd \mu / d \nu on AA.

?
Problem 9.5.24

Let XX be an uncountable set, and let M\mathfrak {M} consist of those sets AA such that either AA or ACA^{C} is countable. Define two measures μ,ν\mu , \nu on M\mathfrak {M} by μ(A)=ν(A)=0\mu (A)= \nu (A)=0 if AA is countable, and μ(A)=1,ν(A)=∞\mu (A)=1, \nu (A)=\infty if ACA^{C} is countable.

?
(a)

Find a measurable function f:X→[0,+∞]f: X \rightarrow [0,+\infty ] such that ν(E)=∫Efdμ\nu (E)=\int_{E} f d \mu, for any E∈ME \in \mathfrak {M}.

(b)

Prove that there does not exist a measurable function g:X→[0,+∞]g: X \rightarrow [0,+\infty ] such that μ(E)=∫Egdν\mu (E)=\int_{E} g d \nu, for any E∈ME \in \mathfrak {M}.

Problem 9.5.25

Let (X,M,μ)(X, \mathfrak {M}, \mu ) be a measure space and let λ,ν\lambda , \nu be finite signed measures on M\mathfrak {M}. Prove that λ⊥ν\lambda \perp \nu if and only if ∣λ∣⊥∣ν∣\left|\lambda \right| \perp \left|\nu \right|.

?
Problem 9.5.26

Let (X,M)(X, \mathfrak {M}) be a measurable space, let E∈ME \in \mathfrak {M}, and let ν\nu be a signed measure on M\mathfrak {M}. Prove that

∣ν∣(E)=sup⁡{∣∫Efdν∣:∣f∣≤1} \left|\nu \right|(E)=\sup \left\{ \left|\int _{E} f d \nu \right|:\left|f\right| \leq 1\right\}
?
Problem 9.5.27

Let μ\mu and ν\nu be two σ\sigma-finite positive measures on a measurable space (X,M)(X, \mathfrak {M}). Prove that there exists A,B∈MA, B \in \mathfrak {M} such that A∩B=∅,A∪B=XA \cap B=\emptyset , A \cup B=X, μ\mu and ν\nu are equivalent on ( A,A∩MA, A \cap \mathfrak {M} ) and μ⊥ν\mu \perp \nu on ( B,B∩MB, B \cap \mathfrak {M} ).

?
Problem 9.5.28

Prove Radon-Nikodym theorem in the case when ν\nu is a signed measure.

?
Problem 9.5.29

Let ( X,M,μX, \mathfrak {M}, \mu ) be a measure space and f,g:X→[0,+∞]f, g: X \rightarrow [0,+\infty ] measurable functions such that the associated measures ν(E)=∫Xfdμ\nu (E)=\int_{X} f d \mu and λ(E)=∫Xgdμ\lambda (E)= \int_{X} g d \mu are equal on M\mathfrak {M}. Prove or disprove: f=gf=g a.e. [μ][\mu ].

?
Problem 9.5.30

Give an example that the Radon-Nikodym theorem fails when ν\nu is not σ\sigma-finite.

?
Problem 9.5.31

Prove or disprove: The Lebesgue Decomposition Theorem remains true if measures μ\mu and ν\nu are not assumed to be σ\sigma-finite.

?
Problem 9.5.32

Show that the Lebesgue decomposition of a function of bounded variation is unique if the singular function is required to vanish at x=ax=a.

?
Problem 9.5.33

Let ( X,M,μX, \mathfrak {M}, \mu ) be a measure space and ff a nonnegative function that is integrable over XX with respect to μ\mu. Find the Lebesgue decomposition with respect to μ\mu of the measure ν\nu defined by ν(E)=∫Efdμ\nu (E)=\int_{E} f d \mu, for E∈ME \in \mathfrak {M}.

?
Problem 9.5.34

Let (X,M)(X, \mathfrak {M}) be a measurable space and μ,ν\mu , \nu two σ\sigma-finite measures on M\mathfrak {M}. Prove that μ⊥ν\mu \perp \nu if and only if there is no nonzero σ\sigma-finite measure λ\lambda on (X,M)(X, \mathfrak {M}) such that λ≪μ,λ≪ν\lambda \ll \mu , \lambda \ll \nu.

?
Problem 9.5.35

Let {an}\left\{ a_{n}\right\} be a fixed sequence of real numbers and let {pn}\left\{ p_{n}\right\} be a sequence of positive real numbers. Define a set function ν:M→R\nu : \mathcal{M} \rightarrow \mathbb {R} by

ν(E)=∑an∈Epn,for E∈M \nu (E)=\sum _{a_{n} \in E} p_{n}, \quad \text{for } E \in \mathcal{M}
?
(a)

Show that ν\nu is a σ\sigma-finite measure on M\mathcal{M}.

(b)

Find the Lebesgue decomposition of ν\nu with respect to the Lebesgue measure mm.

Problem 9.5.36

Let f(x)=x2−5x+6f(x)=x^{2}-5 x+6, and define a set function ν:B→R\nu : \mathcal{B} \rightarrow \mathbb {R} by

ν(E)=∫Efdm,for E∈M \nu (E)=\int _{E} f d m, \quad \text{for } E \in \mathcal{M}
?
(a)

Show that ν\nu is a σ\sigma-finite signed measure on B\mathcal{B};

(b)

Find the Hahn decomposition of R\mathbb {R} with respect to ν\nu;

(c)

Find the Jordan decomposition of ν\nu;

(d)

Find the Lebesgue decomposition of ν\nu with respect to the Lebesgue measure on M\mathcal{M}.

Problem 9.5.37

Let ( X,MX, \mathfrak {M} ) and ( Y,NY, \mathfrak {N} ) be measurable spaces, let μ,ν\mu , \nu be measures on M\mathfrak {M}, let μ′,ν′\mu^{\prime }, \nu^{\prime } be measures on N\mathfrak {N}, and suppose that μ\mu and μ′\mu^{\prime } are σ\sigma-finite. Determine the Lebesgue decomposition of μ×μ′\mu \times \mu^{\prime } with respect to ν×ν′\nu \times \nu^{\prime }, in terms of the Lebesgue decompositions of μ\mu with respect to ν\nu and μ′\mu^{\prime } with respect to ν′\nu^{\prime }.

?
Problem 9.5.38

Prove or disprove: the Lebesgue Decomposition Theorem holds when ν\nu is a signed measure.

?