360 CHAPTER 12. THE Lp SPACES

≤(

32

)k ∥∥ fnk − fnk+1

∥∥pLp <

(32

)k

2−kp ≤ 3k

4k

Hence ∑k µ (Ek)< ∞ and so, by the Borel Cantelli lemma, Lemma 9.2.5 on Page 243, thereis a set of measure zero N such that if ω /∈ N, then

∣∣ fnk (ω)− fnk+1 (ω)∣∣≤ ( 2

3

)k/p. Since

∑k

(23

)k/p

< ∞,{XNC (ω) fnk (ω)

}∞

k=1

is a Cauchy sequence for all ω. Let it converge to f (ω), a measurable function since it is alimit of measurable functions. By Fatou’s lemma, and the Minkowski inequality, Corollary12.1.5, ∥ f − fnk∥p =

(∫ ∣∣ f − fnk

∣∣p dµ)1/p ≤

lim infm→∞

(∫ ∣∣ fnm − fnk

∣∣p dµ

)1/p

= lim infm→∞

∥∥ fnm − fnk

∥∥p ≤

lim infm→∞

m−1

∑j=k

∥∥∥ fn j+1 − fn j

∥∥∥p≤

∑i=k

∥∥ fni+1 − fni

∥∥p ≤ 2−(k−1). (12.3)

Therefore, f ∈ Lp(Ω) because ∥ f∥p ≤ ∥ f − fnk∥p +∥ fnk∥p < ∞, and limk→∞ ∥ fnk − f∥p =0. This proves b.).

This has shown fnk converges to f in Lp (Ω). It follows the original Cauchy sequencealso converges to f in Lp (Ω). This is a general fact that if a subsequence of a Cauchysequence converges, then so does the original Cauchy sequence. This is Theorem 3.2.2. ■

In working with the Lp spaces, the following inequality also known as Minkowski’sinequality is very useful. See [25]. It is similar to the Minkowski inequality for sums. Tosee this, replace the integral,

∫X with a finite summation sign and you will see the usual

Minkowski inequality or rather the version of it given in Corollary 12.1.6.

Lemma 12.1.10 Let (X ,S ,µ) and (Y,F ,λ ) be finite measure spaces and let f beµ×λ measurable. Then the following inequality is valid for p≥ 1.

∫X

(∫Y| f (x,y)|p dλ

) 1p

dµ ≥(∫

Y(∫

X| f (x,y)|dµ)pdλ

) 1p

. (12.4)

Proof: This is an application of the Fubini theorem and Holder inequality. Recall thatp−1 = p/p′. Let J (y)≡

∫X | f (x,y)|dµ. Then∫

Y

(∫X| f (x,y)|dµ

)p

dλ =∫

YJ (y)p/p′

∫X| f (x,y)|dµdλ

=∫

Y

∫X| f (x,y)|J (y)p/p′ dµdλ =

∫X

∫Y| f (x,y)|J (y)p/p′ dλdµ

≤(∫

YJ (y)p dλ

)1/p′ ∫X

(∫Y| f (x,y)|p dλ

)1/p

=

(∫Y

(∫X| f (x,y)|dµ

)p

)1/p′ ∫X

(∫Y| f (x,y)|p dλ

)1/p

360 CHAPTER 12. THE L? SPACES3\* > 3\* 3k< (5) If — Fall < (5) 2 SkHence )); Lt (Ex) < cand so, by the Borel Cantelli lemma, Lemma 9.2.5 on Page 243, thereis a set of measure zero N such that if @ ¢ N, then | fn, (@) — fr, (@)| < (2). Sincek/py (5) <0, { Ryc (@) fin, () pykis a Cauchy sequence for all @. Let it converge to f (@), a measurable function since it is alimit of measurable functions. By Fatou’s lemma, and the Minkowski inequality, Corollary12.1.5, If — fig llp = (f lf — fry? a)!” <1/plim inf (ln ~ ful” a) =lim inf || fon — Jnl, Sm—-oom1tim int > < Dlr — Sn;Therefore, f € L?(Q) because || f|lp < If — fin lo + llfnyllo < 22, and lime +l fin — fllp =0. This proves b.).This has shown f,, converges to f in L? (Q). It follows the original Cauchy sequencealso converges to f in L?(Q). This is a general fact that if a subsequence of a Cauchysequence converges, then so does the original Cauchy sequence. This is Theorem 3.2.2.In working with the L’? spaces, the following inequality also known as Minkowski’sinequality is very useful. See [25]. It is similar to the Minkowski inequality for sums. Tosee this, replace the integral, fy with a finite summation sign and you will see the usualMinkowski inequality or rather the version of it given in Corollary 12.1.6.Sir — Fn; [<2 0. (12.3)Lemma 12.1.10 Let (X,.%,u) and (Y,F,A) be finite measure spaces and let f beLL x A measurable. Then the following inequality is valid for p > 1.[(firesorar)’ ane ([cfireeniamrar)’ araProof: This is an application of the Fubini theorem and Holder inequality. Recall thatp—1= p/p’. Let J(y) = Jy |f(x,y)| du. ThenLG. flsy)law) da [ soy!" [ \rlss)lauar= [ [iro auar = [ [intel oyr!” anauJY JX x Jy(fvorar)” [ (firsorran) a(/ (/ irtesiiaw)'aa)" ([irconiran) aylA