Skip to main content
\(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\)
Mathematics LibreTexts

7.4.E: Problems on Set Functions

  • Page ID
    24443
  • \( \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \) \( \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} \)\(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\) \(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\)

    Exercise \(\PageIndex{1}\)

    Prove Theorem 2 in detail for semirings.
    [Hint: We know that
    \[X_{n}-X_{n-1}=\bigcup_{i=1}^{m_{n}} Y_{ni} \text { (disjoint)}\]
    for some \(Y_{ni} \in \mathcal{C},\) so
    \[\overline{s}\left(X_{n}-X_{n-1}\right)=\sum_{i=1}^{m_{n}} s Y_{ni},\]
    with \(\overline{s}\) as in Theorem 1.]

    Exercise \(\PageIndex{2}\)

    Let \(s\) be additive on \(\mathcal{M},\) a ring. Prove that \(s\) is also \(\sigma\)-additive provided \(s\) is either
    (i) left continuous, or
    (ii) finite on \(\mathcal{M}\) and right-continuous at \(\emptyset;\) i.e.,
    \[\lim _{n \rightarrow \infty} s X_{n}=0\]
    when \(X_{n} \searrow \emptyset\) \(\left(X_{n} \in \mathcal{M}\right)\).
    [Hint: Let
    \[A=\bigcup_{n} A_{n} \text { (disjoint)}, \quad A, A_{n} \in \mathcal{M}.\]
    Set
    \[X_{n}=\bigcup_{k=1}^{n} A_{k}, Y_{n}=A-X_{n}.\]
    Verify that \(X_{n}, Y_{n} \in \mathcal{M}, X_{n} \nearrow A, Y_{n} \searrow \emptyset\).
    In case (i),
    \[s A=\lim s X_{n}=\sum_{k=1}^{\infty} s A_{k}.\]
    (Why?)
    For (ii), use the \(Y_{n}\).]

    Exercise \(\PageIndex{3}\)

    Let
    \[\mathcal{M}=\left\{\text {all intervals in the rational field } R \subset E^{1}\right\}.\]
    Let
    \[s X=b-a\]
    if \(a, b\) are the endpoints of \(X \in \mathcal{M}\) \((a, b \in R, a \leq b).\) Prove that
    (i) \(\mathcal{M}\) is a semiring;
    (ii) \(s\) is continuous;
    (iii) \(s\) is additive but not \(\sigma\)-additive; thus Problem 2 fails for semirings.
    [Hint: \(R\) is countable. Thus each \(X \in \mathcal{M}\) is a countable union of singletons \(\{x\}=[x, x];\) hence \(s X=0\) if \(s\) were \(\sigma\)-additive.]

    Exercise \(\PageIndex{3'}\)

    Let \(N=\) {naturals}. Let
    \[\mathcal{M}=\{\text {all finite subsets of } N \text { and their complements in } N\}.\]
    If \(X \in \mathcal{M},\) let \(s X=0\) if \(X\) is finite, and \(s X=\infty\) otherwise. Show that
    (i) \(\mathcal{M}\) is a set field;
    (ii) \(s\) is right continuous and additive, but not \(\sigma\)-additive.
    Thus Problem 2 (ii) fails if \(s\) is not finite.

    Exercise \(\PageIndex{4}\)

    Let
    \[\mathcal{C}=\left\{\text {finite and infinite intervals in } E^{1}\right\}.\]
    If \(a, b\) are the endpoints of an interval \(X\) \(\left(a, b \in E^{*}, a<b\right),\) set
    \[s X=\left\{\begin{array}{ll}{b-a,} & {a<b,} \\ {0,} & {a=b.}\end{array}\right.\]
    Show that \(s\) is \(\sigma\)-additive on \(\mathcal{C},\) a semiring.
    Let
    \[X_{n}=(n, \infty);\]
    so \(s X_{n}=\infty-n=\infty\) and \(X_{n} \searrow \emptyset.\) (Verify!) Yet
    \[\lim s X_{n}=\infty \neq s \emptyset.\]
    Does this contradict Theorem 2?

    Exercise \(\PageIndex{5}\)

    Fill in the missing proof details in Theorem 1.

    Exercise \(\PageIndex{6}\)

    Let \(s\) be additive on \(\mathcal{M}.\) Prove the following.
    (i) If \(\mathcal{M}\) is a ring or semiring, so is
    \[\mathcal{N}=\{X \in \mathcal{M}| | s X |<\infty\}\]
    if \(\mathcal{N} \neq \emptyset\).
    (ii) If \(\mathcal{M}\) is generated by a set family \(\mathcal{C},\) with \(|s|<\infty\) on \(\mathcal{C},\) then \(|s|<\infty\) on \(\mathcal{M}.\)
    [Hint: Use Problem 16 in §3.]

    Exercise \(\PageIndex{7}\)

    \(\Rightarrow\) (Lebesgue-Stieltjes set functions.) Let \(\alpha\) and \(s_{\alpha}\) be as in Example (d). Prove the following.
    (i) \(s_{\alpha} \geq 0\) on \(\mathcal{C}\) iff \(\alpha \uparrow\) on \(E^{1}\) (see Theorem 2 in Chapter 4, §5).
    (ii) \(s_{\alpha}\{p\}=s_{\alpha}[p, p]=0\) iff \(\alpha\) is continuous at \(p\).
    (iii) \(s_{\alpha}\) is additive.
    [Hint: If
    \[A=\bigcup_{i=1}^{n} A_{i} \text { (disjoint),}\]
    the intervals \(A_{i-1}, A_{i}\) must be adjacent. For two such intervals, consider all cases like
    \[(a, b] \cup(b, c),[a, b) \cup[b, c], \text { etc.}\]
    Then use induction on \(n\).]
    (iv) If \(\alpha\) is right continuous at \(a\) and \(b,\) then
    \[s_{\alpha}(a, b]=\alpha(b)-\alpha(b).\]
    If \(\alpha\) is continuous at \(a\) and \(b,\) then
    \[s_{\alpha}[a, b]=s_{\alpha}(a, b]=s_{\alpha}[a, b)=s_{\alpha}(a, b).\]
    (v) If \(\alpha \uparrow\) on \(E^{1}\), then \(s_{\alpha}\) satisfies Lemma 1 and Corollary 2 in §1 (same proof), as well as Lemma 1, Theorem 1, Corollaries 1-4, and Note 3 in §2 (everything except Corollaries 5 and 6).
    [Hint: Use (i) and (iii). For Lemma 1 in §2, take first a half-open \(B=(a, b];\) use the definition of a right-side limit along with Theorems 1 and 2 in Chapter 4, §5, to prove
    \[(\forall \varepsilon>0)(\exists c>b) \quad 0 \leq \alpha(c-)-\alpha(b+)<\varepsilon;\]
    then set \(C=(a, c).\) Similarly for \(B=[a, b),\) etc. and for the closed interval \(A \subseteq B\).]
    (vi) If \(\alpha(x)=x\) then \(s_{\alpha}=v,\) the volume (or length) function in \(E^{1}\).

    Exercise \(\PageIndex{8}\)

    Construct LS set functions (Example (d)), with \(\alpha \uparrow\) (see Problem 7(v)), so that
    (i) \(s_{\alpha}[0,1] \neq s_{\alpha}[1,2]\);
    (ii) \(s_{\alpha} E^{1}=1\) (after extending \(s_{\alpha}\) to \(\mathcal{C}_{\sigma}-sets in \(E^{1}\));
    (ii') \(s_{\alpha} E^{1}=c\) for a fixed \(c \in(0, \infty)\);
    (iii) \(s_{\alpha}\{0\}=1\) and \(s_{\alpha}[0,1]>s_{\alpha}(0,1]\).
    Describe \(s_{\alpha}\) if \(\alpha(x)=[x]\) (the integral part of \(x\)).
    [Hint: See Figure 16 in Chapter 4, §1.]

    Exercise \(\PageIndex{9}\)

    For an arbitrary \(\alpha : E^{1} \rightarrow E^{1},\) define \(\sigma_{\alpha} : \mathcal{C} \rightarrow E^{1}\) by
    \[\sigma_{\alpha}[a, b]=\sigma_{\alpha}(a, b]=\sigma_{\sigma}[a, b)=\sigma_{\alpha}(a, b)=\alpha(b)-\alpha(a)\]
    (the original Stieltjes method). Prove that \(\sigma_{\alpha}\) is additive but not \(\sigma\)-additive unless \(\alpha\) is continuous (for Theorem 2 fails).

    • Was this article helpful?