Jacobs University Bremen School of Engineering and Science Peter Oswald Fall Term 2014 Real Analysis — Problem Set 5 Issued: 2.10.14 Due: 9.10.14, in class Reading for this week: Chapters 2.2-3 in Folland [F]. Submit solutions for at least 5 problems. 5.1. Prove (for general measure spaces) the following (simple but useful) fact: If fn ∈ L1 (realor complex-valued functions), and ∞ X kfn kL1 < ∞, n=1 then P converges a.e. on X to a measurable function f (x), a) the series ∞ n=1 fn (x)P 1 b) f ∈ L with kf kL1 ≤ ∞ n=1 kfn kL1 , and c) Z ∞ Z X f dµ = fn dµ. X n=1 X Why do I (maybe not you) like this Lemma? 5.2. Given an arbitrary measure space. Suppose fn , g ∈ L+ and fn → g a.e.. We observe that Z lim fn dµ = c > 0 n→∞ X exists. What can you say about the value of R X g dµ? 5.3. Do Exercises 34 and 38 (about convergence in measure) of Chapter 2.4 in [F]. 5.4. For the counting measure on N, the comparison of the convergence types considered in class simplifies a lot. Show (using a table) which of the convergence types implies which other. For each box (but the diagonal ones) give a short argument (or reference to a theorem in [F] or else) why the implication holds and provide an example if it does not. 5.5. Suppose (X, M, µ) is a probability space, i.e., µ(X) = 1. Let f : X → R be measurable (such a function is called random variable in probability theory), and assume that |f |2 ∈ L1 (the random variable has finite second moment). R a) Show that f ∈ L1 and that E(f ) := X f dµ is finite (Hint: Use the fact that |f (x)| ≤ |f (x)|2 for all x ∈ {x ∈ X : |f (x)| ≥ 1} and that µ(X) = 1). b) Show that Z |f (x) − E(f )|2 dµ(x) V (f ) := X is finite (E(f ) and V (f ) are called expectation and variance of the random variable f , respectively. c) Prove the Chebyshev-Markov inequality µ({x ∈ X : |f (x) − E(f )| ≥ λ}) ≤ λ−2 V (f ). 5.6. Show that (assume Lebesgue measure or just work with the (improper) Riemann integral definition) that Z ∞ sin(ex ) dx → 0, n → ∞. 1 + nx2 0