Mathematics University of Virginia

Real Analysis General Exam January 2022

Instructions. 2 hours. Closed book examination. Be neat in your presentation. When invoking a theorem from previous courses, name the theorem and thoroughly check its hypotheses. You must solve a significant portion of each of the three problems in order to pass the exam.

Problem 1

Let ( X,ℳ,μX, \mathscr{M}, \mu ) be a measure space. Let ( fnf_{n} ) be a sequence of nonnegative functions functions in L1(X,ℳ,μ)L^{1}(X, \mathscr{M}, \mu) and let ff be a nonnegative function in L1(X,ℳ,μ)L^{1}(X, \mathscr{M}, \mu). Suppose that

∫Xfndμ→∫Xfdμ\int_{X} f_{n} \mathrm{~d} \mu \longrightarrow \int_{X} f \mathrm{~d} \mu

and that fn→ff_{n} \rightarrow f pointwise. Prove that fnf_{n} converges to ff in L1(X,ℳ,μ)L^{1}(X, \mathscr{M}, \mu). Hint: consider gn=min⁡(f,fn)g_{n}=\min \left(f, f_{n}\right).

Problem 2

Let ( X,ℳ,μX, \mathscr{M}, \mu ) be a measure space. Let p∈[1,∞)p \in[1, \infty).
(a) Let ( fnf_{n} ) be a sequence of functions in Lp(X,ℳ,μ)L^{p}(X, \mathscr{M}, \mu) and let ff be a function in Lp(X,ℳ,μ)L^{p}(X, \mathscr{M}, \mu). Suppose that fnf_{n} converges to ff in Lp(X,ℳ,μ)L^{p}(X, \mathscr{M}, \mu). Prove that there exists a subsequence ( fnkf_{n_{k}} ) such that for μ\mu-almost all x,lim⁡k→∞fnk(x)=f(x)x, \lim _{k \rightarrow \infty} f_{n_{k}}(x)=f(x). Hint: remember the proof of completeness of LpL^{p}.
(b) Let hh be a measurable function on XX. Let

D={f∈Lp(X,ℳ,μ)∣hf∈Lp(X,ℳ,μ)}D=\left\{f \in L^{p}(X, \mathscr{M}, \mu) \mid h f \in L^{p}(X, \mathscr{M}, \mu)\right\}

Let (fn)\left(f_{n}\right) be a sequence of elements of DD, and let f,g∈Lp(X,ℳ,μ)f, g \in L^{p}(X, \mathscr{M}, \mu) be such that fnf_{n} converges to ff in LpL^{p}, and hfnh f_{n} converges to gg in LpL^{p}. Show that f∈Df \in D and g=hfg=h f.

Problem 3

For μ\mu a Borel probability measure on ℝ\mathbb{R}, we will denote by μ̂\widehat{\mu} the function ℝ→ℂ\mathbb{R} \rightarrow \mathbb{C} given by

μ̂(t)=∫ℝeitxdμ(x)\widehat{\mu}(t)=\int_{\mathbb{R}} e^{i t x} \mathrm{~d} \mu(x)

We will also adopt the notational convention sinc⁡(x)=sin⁡xx\operatorname{sinc}(x)=\frac{\sin x}{x} if x≠0x \neq 0 and sinc⁡(0)=1\operatorname{sinc}(0)=1.
(a) Show that μ̂\widehat{\mu} is a bounded continuous function.
(b) Let δ>0\delta>0. Show that

12δ∫−δδ(1−Re⁡(μ̂(t)))dt=∫ℝ(1−sinc⁡(δx))dμ(x)\frac{1}{2 \delta} \int_{-\delta}^{\delta}(1-\operatorname{Re}(\widehat{\mu}(t))) \mathrm{d} t=\int_{\mathbb{R}}(1-\operatorname{sinc}(\delta x)) \mathrm{d} \mu(x)

(c) Show that for all u∈ℝu \in \mathbb{R},

1−sinc⁡(u)≥12χ(−∞,−2)∪(2,∞)(u)1-\operatorname{sinc}(u) \geq \frac{1}{2} \chi_{(-\infty,-2) \cup(2, \infty)}(u)

and deduce that

μ({x∈ℝ||x|>2δ−1})≤1δ∫−δδ(1−Re(μ̂(t)))dt\mu\left(\left\{x \in \mathbb{R}\left||x|>2 \delta^{-1}\right\}\right) \leq \frac{1}{\delta} \int_{-\delta}^{\delta}(1-\operatorname{Re}(\widehat{\mu}(t))) \mathrm{d} t\right.

(d) Let μn\mu_{n} be a sequence of Borel probability measures on ℝ\mathbb{R}. Suppose that for all tt, the limit Φ(t)=lim⁡n→∞μn̂(t)\Phi(t)=\lim _{n \rightarrow \infty} \widehat{\mu_{n}}(t) exists and that the resulting function Φ(t)\Phi(t) is continuous at t=0t=0. Prove that for all ϵ>0\epsilon>0, there exists a compact set KK inside ℝ\mathbb{R} such that, for all n,μn(K)≥1−ϵn, \mu_{n}(K) \geq 1-\epsilon.