SOLUTIONS OF HW6 MINGFENG ZHAO March 31, 2011 1. (Problem 32, in Page 63) Suppose µ(X) < ∞. If f and g are complex-valued measurable functions on X, define Z ρ(f, g) = X |f − g| dµ. 1 + |f − g| Then ρ is a metric on the space of measurable functions if we identify functions that are equal a.e, and fn → f with respect to this measure iff fn → f in measure. Proof. Let Fµ = {all complex-valued µ-measurable functions on X}. (1). (a). Well-defined: For any f, g ∈ Fµ , we know that 0≤ |f (x) − g(x)| ≤1 1 + |f (x) − g(x)| for all x ∈ X. So we have Z X |f − g| dµ ≤ 1 + |f − g| Z 1dµ = µ(X) < ∞. This means, 0 ≤ ρ(f, g) ≤ µ(X) < ∞. (b) Positive-definite: If ρ(f, g) = 0, that is, Z X Since 0 ≤ |f (x)−g(x)| 1+|f (x)−g(x)| |f − g| dµ = 0. 1 + |f − g| for all x ∈ X, then we have |f (x) − g(x)| = 0 a.e x ∈ X, which implies that f (x) = g(x) for all x ∈ X. In our sense, we have f = g. 1 2 MINGFENG ZHAO (c) Symmetry: For all x ∈ X, we have |f (x) − g(x)| |g(x) − f (x)| = . 1 + |f (x) − g(x)| 1 + |g(x) − f (x)| So we have Z X |f (x) − g(x)| dµ = 1 + |f (x) − g(x)| Z X |g(x) − f (x)| dµ. 1 + |g(x) − f (x)| That is, ρ(f, g) = ρ(g, f ). (d) Triangle inequality: For any f, g, h ∈ Fµ . So for all x ∈ X, we have |f (x) − g(x)| ≤ |f (x) − h(x)| + |h(x) − g(x)|. Since F (t) = t 1+t =1− |f (x) − g(x)| 1 + |f (x) − g(x)| 1 1+t is increasing on [0, ∞), then we can get for all x ∈ X, we have ≤ |f (x) − h(x)| + |h(x) − g(x)| 1 + |f (x) − h(x)| + |h(x) − g(x)| = |f (x) − h(x)| |f (x) − h(x)| + 1 + |f (x) − h(x)| + |h(x) − g(x)| 1 + |f (x) − h(x)| + |h(x) − g(x)| ≤ |f (x) − h(x)| |f (x) − h(x)| + 1 + |f (x) − h(x)| 1 + |h(x) − g(x)| So we have Z X Z |f (x) − g(x)| dµ ≤ 1 + |f (x) − g(x)| X Z = X |f (x) − h(x)| |f (x) − h(x)| + dµ 1 + |f (x) − h(x)| 1 + |h(x) − g(x)| Z |f (x) − h(x)| |f (x) − h(x)| dµ + dµ 1 + |f (x) − h(x)| X 1 + |h(x) − g(x)| That is, ρ(f, g) ≤ ρ(f, h) + ρ(h, g). Therefore, by (a),(b),(c) and (d), we know ρ is a metric on Fµ . (2). (=⇒) If fn −→ f with respect to ρ, as n → ∞. That is, Z lim n→∞ X |fn (x) − f (x)| dµ = 0. 1 + |fn (x) − f (x)| For any > 0 and fix , we have µ({x ∈ X : |fn (x) − f (x)| > }) · 1+ SOLUTIONS OF HW6 Z = {x∈X: |fn (x)−f (x)|>} Z ≤ {x∈X: |fn (x)−f (x)|>} Z ≤ X 3 dµ 1+ |fn (x) − f (x)| dµ Since |fn (x) − f (x)| > 1 + |fn (x) − f (x)| |fn (x) − f (x)| dµ. 1 + |fn (x) − f (x)| So we have 1+ µ({x ∈ X : |fn (x) − f (x)| > }) ≤ Z X |fn (x) − f (x)| dµ. 1 + |fn (x) − f (x)| Take the supremum, as n → ∞, we get lim sup µ({x ∈ X : |fn (x) − f (x)| > }) ≤ n→∞ 1+ lim n→∞ Z X |fn (x) − f (x)| dµ = 0. 1 + |fn (x) − f (x)| Therefore, by the arbitrary of > 0, we have fn → f in measure, as n → ∞. (⇐=) Assume fn −→ f in measure, as n → ∞, that is, for any > 0, we have lim µ({x ∈ X : |fn (x) − f (x)| > }) = 0. n→∞ Then we have Z ρ(fn , f ) = X |fn (x) − f (x)| dµ 1 + |fn (x) − f (x)| Z = {x∈X: Z ≤ |fn (x) − f (x)| dµ + |fn (x)−f (x)|>} 1 + |fn (x) − f (x)| Z 1dµ + {x∈X: |fn (x)−f (x)|>} Z {x∈X: |fn (x)−f (x)|≤} {x∈X: |fn (x)−f (x)|≤} dµ 1+ Z = µ({x ∈ X : |fn (x) − f (x)| > }) + {x∈X: |fn (x)−f (x)|≤} Z ≤ µ({x ∈ X : |fn (x) − f (x)| > }) + dµ X = µ({x ∈ X : |fn (x) − f (x)| > }) + µ(X). dµ 1+ |fn (x) − f (x)| dµ 1 + |fn (x) − f (x)| 4 MINGFENG ZHAO For any δ > 0, we take 0 < < δ 2µ(X) , then there exists N ∈ N such that whenever n ≥ N , we have µ({x ∈ X : |fn (x) − f (x)| > }) < δ . 2 Hence, for all n ≥ N , we have ρ(fn , f ) ≤ µ({x ∈ X : |fn (x) − f (x)| > }) + µ(X) ≤ δ δ + · µ(X) 2 2µ(X) = δ By the arbitrary of δ > 0, we know ρ(fn , f ) → ∞, as n → ∞, that is, fn −→ f with respect to ρ, as n → ∞. 2. (Problem 44, in Page 64) (Lusin’s theorem). If f : [a, b] → C is Lebesgue measurable and > 0, there is a compact set E ⊂ [a, b] such that µ(E c ) < and f |E is continuous. Proof. (1). If there exists M > 0 such that |f (x)| ≤ M a.e x ∈ [a, b], then we have f ∈ L1 ([a, b]). We extend f trivially on R by: F (x) = f (x), if x ∈ [a, b] 0, if x 6∈ [a, b]. Then we know F ∈ L1 (R), F (x) = f (x) for all x ∈ [a, b], and R R F dµ = proof of Theorem 2.26, in Page 55, there exists Gn ∈ C(R) such that Z |F (x) − Gn (x)|dµ = 0. lim n→∞ R Since Z Z |F (x) − Gn (x)|dµ ≤ [a,b] |F (x) − Gn (x)|dµ. R R [a,b] f dµ. Then by the SOLUTIONS OF HW6 5 So we have Z |F (x) − Gn (x)|dµ ≤ lim lim n→∞ Z n→∞ [a,b] |F (x) − Gn (x)|dµ = 0. R So Z |F (x) − Gn (x)|dµ = 0. lim n→∞ [a,b] Let gn (x) = Gn (x) for all x ∈ [a, b]. Since Gn is continuous on R, then gn is continuous on [a, b]. Since F (x) = f (x) for all x ∈ [a, b], so we have Z |f (x) − gn (x)|dµ = 0. lim n→∞ [a,b] By Corollary 2.32, in Page 62, we know there exists a subsequence gnk (x) = fk (x) such that fk (x) → f (x) a.e x ∈ [a, b], as k → ∞. On the other hand, since µ([a, b]) = b − a < ∞, by Egoroff’s theorem, Theorem 2.33 in Page 62, for any > 0, there exists a µ-measurable subset F ⊂ [a, b] such that µ(F ) < 2 , and fk → f uniformly on F c , as k → ∞. By the exterior regularity theorem of Lebesgue measure, Theorem 1.18, in Page 36, there exists an open subset U ⊂ [0, 1] such that F ⊂ U , and µ(U \F ) < . 2 So we have µ(U ) ≤ µ(F ) + µ(U \F ) < , and U c ⊂ F c . So fk → f uniformly on U c , as k → ∞. Let E = U c , then we have fk → f uniformly on E, as k → ∞, and µ(E c ) = µ(U ) < . So f is continuous on E. (2). If f is not bounded on [a, b]. For any n ≥ 1, we let fn (x) = f (x) · 1{x∈[a,b]: all n ≥ 1, we have |fn (x)| ≤ n for all x ∈ [a, b], and for all x ∈ [a, b], we have lim fn (x) = f (x) n→∞ |f (x)|≤n} (x). Then for 6 MINGFENG ZHAO By the result of (1), for any > 0, there exists a compact set En ⊂ [a, b] such that µ(Enc ) < 2n+1 , and fn is continuous on En . Let E0 = ∩∞ n=1 En , then we can get µ(E0c ) c µ(∪∞ n=1 En ) = ∞ X ≤ µ(Enc ) n=1 ∞ X ≤ n+1 2 n=1 = . 2 And fn is continuous on E0 for all n ≥ 1. Since µ(E0 ) ≤ µ([a, b]) < ∞, and fn (x) → f (x), as n → ∞, for all x ∈ E0 . By Egoroff’s theorem, Theorem 2.33 in Page 62, for the > 0, there exists a µ-measurable subset F ⊂ E0 such that µ(F ) < 8 , and fn → f uniformly on E0 \F , as n → ∞. Since fn is continuous on E0 \F for all n ≥ 1, by the uniformly convergence, we have f is continuous on E0 \F . By the exterior regularity theorem of Lebesgue measure, Theorem 1.18, in Page 36, there exists an open subset U ⊂ [0, 1] such that F ⊂ U , and µ(U \F ) < So we have µ(U ) ≤ µ(F ) + µ(U \F ) < 8 + 4 = 3 8 , . 4 and U c ⊂ F c . Let E = U c ∩ (E0 \F ), so E is compact, and f is continuous on E. And µ(E c ) = µ(U ∪ (E0 \F )c ) = µ(U ∪ E0c ∪ F ) ≤ µ(U ) + µ(E0c ) + µ(F ) < 3 + + 8 2 8 = . SOLUTIONS OF HW6 7 In a summary, if f : [a, b] → C is Lebesgue measurable and > 0, there is a compact set E ⊂ [a, b] such that µ(E c ) < and f |E is continuous. Department of Mathematics, University of Connecticut, 196 Auditorium Road, Unit 3009, Storrs, CT 06269-3009 E-mail address: mingfeng.zhao@uconn.edu