Skip to main content
Mathematics LibreTexts

9.1.E: Problems on L-Integrals and Antiderivatives

  • Page ID
    25152
  • \( \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \) \( \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} \)\(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\) \(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\)\(\newcommand{\AA}{\unicode[.8,0]{x212B}}\)

    Exercise \(\PageIndex{1}\)

    Fill in proof details in Theorems 1 and 2, Lemma 1, and Corollaries 1-3.

    Exercise \(\PageIndex{1'}\)

    Verify Note 2.

    Exercise \(\PageIndex{2}\)

    Let \(F\) be Cantor's function (Problem 6 in Chapter 4, §5). Let
    \[
    G=\bigcup_{k, i} G_{k t}
    \]
    \(\left(G_{k t} \text { as in that problem }\right) .\) So \([0,1]-G=P(\text { Cantor's set }) ; m P=0\) (Problem \(10 \text { in Chapter } 7, §8) .\)
    Show that \(F\) is differentiable \(\left(F^{\prime}=0\right)\) on \(G .\) By Theorems 2 and 3 of Chapter 8, §9,
    \[
    R \int_{0}^{1} F^{\prime}=L \int_{0}^{1} F^{\prime}=L \int_{G} F^{\prime}=0
    \]
    exists, yet \(F(1)-F(0)=1-0 \neq 0\).
    Does this contradict Corollary \(1 ?\) Is \(F\) a genuine antiderivative of \(f ?\) If not, find one.

    Exercise \(\PageIndex{3}\)

    Let
    \[
    F=\left\{\begin{array}{ll}{0} & {\text { on }\left[0, \frac{1}{2}\right), \text { and }} \\ {1} & {\text { on }\left[\frac{1}{2}, 1\right] .}\end{array}\right.
    \]
    Show that
    \[
    R \int_{0}^{1} F^{\prime}=0
    \]
    exists, yet
    \[
    F(1)-F(0)=1-0=1 .
    \]
    What is wrong?
    [Hint: A genuine primitive of \(\left.F^{\prime} \text { (call it } \phi\right)\) has to be relatively continuous on \([0,1] ;\) find \(\phi \text { and show that } \phi(1)-\phi(0)=0 .]\)

    Exercise \(\PageIndex{4}\)

    What is wrong with the following computations?
    (i) \(L \int_{-1}^{\frac{1}{2}} \frac{d x}{x^{2}}=-\left.\frac{1}{x}\right|_{-1} ^{\frac{1}{2}}=-1\).
    (ii) \(L \int_{-1}^{1} \frac{d x}{x}=\left.\ln |x|\right|_{-1} ^{1}=0 .\) Is there a primitive on the whole interval?
    [Hint: See hint to Problem 3.]
    (iii) How about \(L \int_{-1}^{1} \frac{|x|}{x} d x\) (cf. examples (a) and (b) of Chapter 5, §5)?

    Exercise \(\PageIndex{5}\)

    Let
    \[
    F(x)=x^{2} \cos \frac{\pi}{x^{2}}, \quad F(0)=1 .
    \]
    Prove the following:
    (i) \(F\) is differentiable on \(A=[0,1]\).
    (ii) \(f=F^{\prime}\) is bounded on any \([a, b] \subset(0,1),\) but not on \(A\).
    (iii) Let
    \[
    a_{n}=\sqrt{\frac{2}{4 n+1}} \text { and } b_{n}=\frac{1}{\sqrt{2 n}} \text { for } n=1,2, \ldots
    \]
    Show that
    \[
    A \supseteq \bigcup_{n=1}^{\infty}\left[a_{n}, b_{n}\right](\text {disjoint})
    \]
    and
    \[
    L \int_{a_{n}}^{b_{n}} f=\frac{1}{2 n} ;
    \]
    so
    \[
    L \int_{a}^{b} f \geq L \int_{\cup_{n=1}^{\infty}\left[a_{n}, b_{n}\right]} f \geq \sum_{n=1}^{\infty} \frac{1}{2 n}=\infty ,
    \]
    and \(f=F^{\prime}\) is not L-integrable on \(A\).
    What is wrong? Is there a contradiction to Theorem \(2 ?\)

    Exercise \(\PageIndex{6}\)

    Consider both
    (a) \(f(x)=\frac{\sin x}{x}, f(0)=1,\) and
    (b) \(f(x)=\frac{1-e^{-x}}{x}, f(0)=1\).
    In each case, show that \(f\) is continuous on \(A=[0,1]\) and
    \[
    R \int_{A} f \leq 1
    \]
    exists, yet it does not "work out" via primitives. What is wrong? Does a primitive exist?
    To use Corollary \(1,\) first expand \(\sin x\) and \(e^{-x}\) in a Taylor series and find the series for
    \[
    \int f
    \]
    by Theorem 3 of Chapter 5, §9.
    Find
    \[
    R \int_{A} f
    \]
    approximately, to within \(1 / 10,\) using the remainder term of the series to estimate accuracy.
    [Hint: Primitives exist, by Theorem 2 of Chapter 5, §11, even though they are none of the known "calculus functions." \(]\)

    Exercise \(\PageIndex{7}\)

    Take \(A, G_{n}=\left(a_{n}, b_{n}\right),\) and \(P(m P>0)\) as in Problem 17(iii) of Chapter 7, §8.
    Define \(F=0\) on \(P\) and
    \[
    F(x)=\left(x-a_{n}\right)^{2}\left(x-b_{n}\right)^{2} \sin \frac{1}{\left(b_{n}-a_{n}\right)\left(x-a_{n}\right)\left(x-b_{n}\right)} \quad \text { if } x \notin P .
    \]
    Prove that \(F\) has a bounded derivative \(f,\) yet \(f\) is not R-integrable on \(A ;\) so Theorem 2 applies, but Corollary 1 does not.
    [Hints: If \(p \notin P,\) compute \(F^{\prime}(p)\) as in calculus.
    If \(p \in P\) and \(x \rightarrow p+\) over \(A-P,\) then \(x\) is always in some \(\left(a_{n}, b_{n}\right), p \leq a_{n}<x .\) (Why?) Deduce that \(\Delta x=x-p>x-a_{n}\) and
    \[
    \left|\frac{\Delta F}{\Delta x}\right| \leq\left(x-a_{n}\right)(b-a)^{2} \leq|\Delta x|(b-a)^{2} ;
    \]
    so \(F_{+}^{\prime}(p)=0 .\) (What if \(x \rightarrow p+\) over \(P\) ?) Similarly, show that \(F_{-}^{\prime}=0\) on \(P\).
    Prove however that \(F^{\prime}(x)\) oscillates from 1 to \(-1\) as \(x \rightarrow a_{n}+\) or \(x \rightarrow b_{n}-\), hence also as \(x \rightarrow p \in P\) (why?); so \(F^{\prime}\) is discontinuous on all of \(P,\) with \(m P>0 .\) Now use Theorem 3 in Chapter 8, §9.]

    Exercise \(\PageIndex{8}\)

    \(\Rightarrow 8\). If
    \[
    Q \subseteq A=[a, b]
    \]
    and \(m Q=0,\) find a continuous map \(g: A \rightarrow E^{1}, g \geq 0, g \uparrow,\) with
    \[
    g^{\prime}=+\infty \quad \text { on } Q .
    \]
    [Hints: By Theorem 2 of Chapter 7, §8, fix ( \(\forall n\) ) an open \(G_{n} \supseteq Q,\) with
    \[
    m G_{n}<2^{-n} .
    \]
    Set
    \[
    g_{n}(x)=m\left(G_{n} \cap[a, x]\right)
    \]
    and
    \[
    g=\sum_{n=1}^{\infty} g_{n}
    \]
    on \(A ; \sum g_{n}\) converges uniformly on \(A .\) (Why?)
    By Problem 4 in Chapter 7, §9, and Theorem 2 of Chapter 7, §4, each \(g_{n}\) (hence g) is continuous. (Why?) If \([p, x] \subseteq G_{n},\) show that
    \[
    g_{n}(x)=g_{n}(p)+(x-p) ,
    \]
    so
    \[
    \frac{\Delta g_{n}}{\Delta x}=1
    \]
    and
    \[
    \left.\frac{\Delta g}{\Delta x}=\sum_{n=1}^{\infty} \frac{\Delta g_{n}}{\Delta x} \rightarrow \infty .\right]
    \]

    Exercise \(\PageIndex{9}\)

    (i) Prove Corollary 4.
    (ii) State and prove earlier analogues for Corollary 5 of Chapter 5, §5, and Theorems 3 and 4 from Chapter 5, §10.
    [Hint for (i): For primitives, this is Problem 3 in Chapter 5, §5. As \(g[Q]\) is countable (Problem 2 in Chapter 1, §9) and \(f\) is bounded on
    \[
    g[A]-g[Q] \subseteq g[A-Q] ,
    \]
    \(\left.f \text { satisfies Theorem 2 on } g[A], \text { with } P=g[Q], \text { while }(f \circ g) g^{\prime} \text { satisfies it on } A .\right]\)

    Exercise \(\PageIndex{10}\)

    \(\Rightarrow 10\). Show that if \(h: E^{1} \rightarrow E^{*}\) is L-integrable on \(A=[a, b],\) and
    \[
    (\forall x \in A) \quad L \int_{a}^{x} h=0 ,
    \]
    then \(h=0\) a.e. on \(A\).
    [Hints: Let \(K=A(h>0)\) and \(H=A-K,\) with, say, \(m K=\varepsilon>0 .\)
    Then by Corollary 1 in Chapter 7, §1 and Definition 2 of Chapter 7, §5,
    \[
    H \subseteq \bigcup_{n} B_{n}(d i s j o i n t)
    \]
    for some intervals \(B_{n} \subseteq A,\) with
    \[
    \sum_{n} m B_{n}<m H+\varepsilon=m H+m K=m A .
    \]
    (Why?) Set \(B=\cup_{n} B_{n} ;\) so
    \[
    \int_{B} h=\sum_{n} \int_{B_{n}} h=0
    \]
    (for \(L \int h=0\) on intervals \(B_{n}\)). Thus
    \[
    \int_{A-B} h=\int_{A} h-\int_{B} h=0 .
    \]
    But \(B \supseteq H ;\) so
    \[
    A-B \subseteq A-H=K ,
    \]
    where \(h>0,\) even though \(m(A-B)>0 .\) (Why?)
    Hence find a contradiction to Theorem \(1(\mathrm{h})\) of Chapter 8, §5. Similarly, disprove that \(m A(h<0)=\varepsilon>0 .]\)

    Exercise \(\PageIndex{11}\)

    \(\Rightarrow 11\). Let \(F \uparrow\) on \(A=[a, b],|F|<\infty,\) with derived function \(F^{\prime}=f .\) Taking Theorem 3 from Chapter 7, §10, for granted, prove that
    \[
    L \int_{a}^{x} f \leq F(x)-F(a), \quad x \in A ,
    \]
    [Hints: With \(f_{n}\) as in \((3), F\) and \(f_{n}\) are bounded on \(A\) and measurable by Theorem 1 of Chapter 8, §2. (Why?) Deduce that \(f_{n} \rightarrow f\) (a.e.) on \(A .\) Argue as in Lemma 1 using Fatou's lemma (Chapter 8, §6, Lemma 2).]

    Exercise \(\PageIndex{12}\)

    ("Truncation.") Prove that if \(g: S \rightarrow E\) is m-integrable on \(A \in \mathcal{M}\) in a measure space \((S, \mathcal{M}, m),\) then for any \(\varepsilon>0,\) there is a bounded, \(M\)-measurable and integrable on \(A \operatorname{map} g_{0}: S \rightarrow E\) such that
    \[
    \int_{A}\left|g-g_{0}\right| d m<\varepsilon .
    \]
    [Outline: Redefine \(g=0\) on a null set, to make \(g \mathcal{M}\)-measurable on \(A .\) Then for \(n=1,2, \ldots\) set
    \[
    g_{n}=\left\{\begin{array}{ll}{g} & {\text { on } A(|g|<n), \text { and }} \\ {0} & {\text { elsewhere. }}\end{array}\right.
    \]
    (The function \(g_{n}\) is called the \(n\)th truncate of \(g\).)
    \(\quad\) Each \(g_{n}\) is bounded and \(\mathcal{M}\) -measurable on \(A\) (why?), and
    \[
    \int_{A}|g| d m<\infty
    \]
    by integrability. Also, \(\left|g_{n}\right| \leq|g|\) and \(g_{n} \rightarrow g(\text {pointwise})\) on \(A .\) (Why?)
    Now use Theorem 5 from Chapter 8, §6, to show that one of the \(g_{n}\) may serve as the desired \(\left.g_{o} .\right]\)

    Exercise \(\PageIndex{13}\)

    Fill in all proof details in Lemma \(2 .\) Prove it for unbounded \(g\).
    [Hints: By Problem \(12,\) fix a bounded \(g_{o}\left(\left|g_{o}\right| \leq B\right),\) with
    \[
    L \int_{A}\left|g-g_{o}\right|<\frac{1}{2} \frac{\varepsilon}{f(a)-f(b)} .
    \]
    Verify that
    \[
    \begin{aligned}\left|s_{n}\right| \leq \sum_{i=1}^{q_{n}} \int_{A_{n i}} w_{n i}|g| & \leq \sum_{i} \int_{A_{n i}} w_{n i}\left|g_{o}\right|+\sum_{i} \int_{A_{n i}} w_{n i}\left|g-g_{o}\right| \\ & \leq B \sum_{i} w_{n i} m A_{n i}+\sum_{i} \int_{A_{n i}}[f(a)-f(b)]\left|g-g_{o}\right| \\ &<\frac{1}{n}+\int_{A}[f(a)-f(b)]\left|g-g_{o}\right|<\frac{1}{n}+\frac{1}{2} \varepsilon . \end{aligned}
    \]
    For all \(n>2 / \varepsilon,\) we get \(\left|s_{n}\right|<\frac{1}{2} \varepsilon+\frac{1}{2} \varepsilon=\varepsilon .\) Hence \(s_{n} \rightarrow 0 .\) Now finish as in the text.]

    Exercise \(\PageIndex{14}\)

    Show that Theorem 4 fails if \(F\) is not differentiable at some \(p \in A .\)
    [Hint: See Problems 2 and 3.]


    9.1.E: Problems on L-Integrals and Antiderivatives is shared under a CC BY 1.0 license and was authored, remixed, and/or curated by LibreTexts.

    • Was this article helpful?