\(\newcommand\notesclassalt[2]{\if\smallformat1#2\else#1\fi} \newcommand\lectureend[1]{% \noindent\makebox[\linewidth]{\rule{\paperwidth}{0.4pt}} \marginpar{\raggedleft#1}% } \newcommand\isdefinition[1]{\textbf{#1}} \newcommand\f\frac \newcommand\intd{\mathop{}\!\mathrm{d}} \newcommand\ind[1]{1_{#1}} \newcommand\seq\coloneqq \newcommand\qes{=:} \newcommand\loc{\mathrm{loc}} \newcommand\A{\mathcal A} \newcommand\R{\mathcal R} \newcommand\Q{\mathcal Q} \newcommand\B{\mathcal B} \renewcommand\P{\mathcal P} \newcommand\spt[1]{\mathrm{spt}(#1)} \newcommand\var{\mathop{\mathrm{var}}} \newcommand\diam{\mathop{\mathrm{diam}}} \newcommand\dist{\mathop{\mathrm{dist}}} \newcommand\dmo{\mathrm d} \newcommand\dm[2]{\dmo(#1,#2)} \newcommand\tb[1]{\mathop{\partial{}}{#1}} \newcommand\mb[1]{\mathop{\partial_*}{#1}} \newcommand\mbe[1]{\mb{(#1)}} \newcommand\mbb[1]{\mb{\Bigl(#1\Bigr)}} \newcommand\tbe[1]{\tb{(#1)}} \newcommand\tbb[1]{\tb{\Bigl(#1\Bigr)}} \newcommand\rb[1]{\mathop{\partial^*}{#1}} \newcommand\tc[1]{\overline{#1}} \newcommand\mc[1]{\tc{#1}^*} \newcommand\ti[1]{\mathring{#1}} \newcommand\tia[1]{#1^{\mathrm o}} \newcommand\tim[1]{\tia{(#1)}} \newcommand\tib[1]{\tia{\bigl(#1\bigr)}} \newcommand\mi[1]{\ti{#1}^*} \newcommand\mia[1]{#1^{\mathrm o*}} \newcommand\mim[1]{\mia{(#1)}} \newcommand\mib[1]{\mia{\bigl(#1\bigr)}} \newcommand\smg{\mathcal{H}} \newcommand\smo{\smg^{d-1}} \newcommand\sm[1]{\smg^{d-1}(#1)} \newcommand\smb[1]{\smg^{d-1}\Bigl(#1\Bigr)} \newcommand\smi[1]{\smg^0(#1)} \newcommand\smib[1]{\smg^0\Bigl(#1\Bigr)} \newcommand\lmr[1]{|#1|} \newcommand\lmg{\mathcal{L}} \newcommand\lmog{\lmg_*} \newcommand\lm[1]{\lmg(#1)} \newcommand\lmb[1]{\lmg\Bigl(#1\Bigr)} \newcommand\lmib[1]{\lmio\Bigl(#1\Bigr)} \newcommand\lmo[1]{\lmog(#1)} \newcommand\lmob[1]{\lmog\Bigl(#1\Bigr)} \newcommand{\mres}{\!\mathbin{\vrule height 1.6ex depth 0pt width 0.13ex\vrule height 0.13ex depth 0pt width 1.1ex}\!} \newcommand\restrictzero[2]{#1\,\mres#2} \newcommand\D{\mathbb D} \newcommand\Dq{\mathcal D} \newcommand\M{\mathcal M} \newcommand\riemannl[2]{\underline{\mathcal R}_{#1}(#2)} \newcommand\riemannstepl[2]{\underline S_{#1}(#2)} \newcommand\riemannu[2]{\overline{\mathcal R}_{#1}(#2)} \newcommand\riemannstepu[2]{\overline S_{#1}(#2)} \newcommand\riemann[1]{\mathcal R(#1)} \)

Quiz 4
Real Analysis ICTP 2025

Julian Weigt

October 1, 2025
1.

Let \(A_1,A_2,…⊂Ω\) be measurable and \(a_1,a_2,…\geq 0\). Is the function \(f=\sum _{n=1}^∞a_n\ind {A_n}\) measurable?

Solution: Yes, because for each \(k∈ℕ\) the function \(f_k=\sum _{n=1}^ka_k\ind {A_k}\) is measurable and \(f=\lim _{k→∞}f_k\).

2.

Let \(a_1,…,a_n,b_1,…,b_n∈ℝ\) and \(A_1,…,A_n,B_1,…,B_n⊂Ω\) such that

\[ \sum _{k=1}^n a_n\ind {A_n} = \sum _{k=1}^n b_n\ind {B_n} . \]
Is it true, that \(\{a_1,…,a_n\}=\{b_1,…,b_n\}\) and \(\{A_1,…,A_n\}=\{B_1,…,B_n\}\)?

Solution: No, it may happen that \(A_1=A_2=B_1=B_2\) and \(a_1+a_2=b_1+b_2\) but \(\{a_1,a_2\}\neq \{b_1,b_2\}\), for example for \(a_1=0,a_2=2,\ b_1=b_2=1\).

It similarly happen that \(a_1=a_2=b_1=b_2\) and \(A_1∪A_2=B_1∪B_2\) but \(\{A_1,A_2\}\neq \{B_1,B_2\}\).

3.

Let \(f_1,f_2:Ω→[0,∞]\) and \(f=f_1-f_2\). Does that imply \(f_1=f^+\) and \(f_2=f^-\)?

What, if for all \(x∈Ω\) we have \(f_1(x)=0\) or \(f_2(x)=0\)?

Solution: No, for example if \(f=1\) is constant, then \(f^+=1\) and \(f^-=0\) but \(f_1=2\) and \(f_2=1\) satisfy \(f=f_1-f_2\).

The statement is true however under the assumption that for all \(x∈Ω\) we have \(f_1(x)=0\) or \(f_2(x)=0\).

4.

Let \(μ\) be a measure on \(ℝ^d\). Is \(ℝ^d\) \(σ\)-finite with respect to \(μ\)?

What, if \(μ\) is a Radon measure?

Solution: No, for example for the counting measure.

If \(μ\) is a Radon measure the answer is yes, because \(ℝ^d=B(0,1)∪B(0,2)∪…\) and \(μ(B(0,n))\leqμ (\tc {B(0,1)})<∞\).

5.

Given \(f,f_1,f_2,…:Ω→ℝ\), what does it mean that \(f_n→f\) uniformly?

Solution: It means that for every \(ε>0\) exists an \(n∈ℕ\) such that for all \(k\geq n\) and \(x∈Ω\) we have \(|f_k(x)-f(x)|<ε\).

6.

Find an example of functions \(f,f_1,f_2,…:(0,1)→ℝ\) such that \(f_n→f\) pointwise, but not uniformly.

Solution: Set \(f_n(x)=x^{1/n}\). Then \(f_n→0\) pointwise but not uniformly: Let \(n∈ℕ\). Then for \(x_n=1/2^{1/n}∈(0,1)\) we have \(|f_n(x_n)-0|=1/2\).

7.

Find \(f:ℝ→ℝ\) and a closed set \(C⊂ℝ\) such that \(f:C→ℝ\) is continuous but \(f:ℝ→ℝ\) is not continuous in every \(x∈C\).

Can you also find an open set with that property?

Solution: Set \(C=[0,1]\) and \(f(x)=1\) if \(x∈C\) and \(f(x)=0\) if \(x\notin C\).

Such an example cannot be found with an open set \(U\) instead. The reason is, that if \(x∈U\) and \(x_n→x\), then there exists a \(k∈ℕ\) such that for all \(n\geq k\) we have \(x_n∈ U\). As a consequence, \(f:U→ℝ\) is continuous in \(x\) if and only if \(f:ℝ^d→ℝ\) is continuous in \(x\).

8.

What is the Lebesgue integral \(∫f\intd \lmg \) for

(i)
\(f=\ind ℝ-\ind ℝ\)
(ii)
\(f=\ind {[0,∞)}-\ind {(-∞,0)}\)

Solution:

(i)
\(0\), since the canonical form of \(f\) is \(f=0\).
(ii)
The integral is not defined because both \(∫f^+\intd \lmg =∞\) and \(∫f^-\intd \lmg =∞\).