516 CHAPTER 19. HAUSDORFF SPACES AND MEASURES

From the picture as needed: Let µ (U \F) < ε where U is open and let K ⊆ U andµ (U \K)< ε , µ(V \ (U \F))< ε with V open and V ⊇U \F =U ∩FC so VC ⊆UC ∪F .This is possible because all sets are in S . Then VC∩K ⊆

(UC ∪F

)∩K = F∩K ⊆ F. Now

VC ∩K is compact and

µ(F \(K∩VC)) = µ

(F ∩

(KC ∪V

))= µ (F ∩V )+µ

(F ∩KC)

≤ µ (F ∩V )+µ (U \K)< µ (F ∩V )+ ε (19.6)

However,

ε > µ(V \ (U \F)) = µ

(V ∩

(U ∩FC)C)= µ

(V ∩

(UC ∪F

))≥ µ (V ∩F)

and so from 19.6, µ(F \(K∩VC

))≤ 2ε . Since K∩VC is compact, this shows 19.5. ■

19.7 Measures and Positive Linear FunctionalsThis is on the Riesz representation theorem for positive linear functionals. It is a reallymarvelous result. It produces measures on locally compact Hausdorff spaces. Thus thisdoesn’t help a lot in producing measures on infinite dimensional spaces but it works greaton Rn or closed subsets of Rn and so forth.

Definition 19.7.1 Let (Ω,τ) be a topological space. L : Cc(Ω)→ C is called apositive linear functional if L is linear, L(a f1+b f2) = aL f1+bL f2, and if L f ≥ 0 wheneverf ≥ 0.

Theorem 19.7.2 (Riesz representation theorem) Let (Ω,τ) be a locally compactHausdorff space and let L be a positive linear functional on Cc(Ω). Then there exists a σ

algebra S containing the Borel sets and a unique measure µ , defined on S , such that

µ is complete, (19.7)µ(K) < ∞ for all K compact, (19.8)

µ(F) = sup{µ(K) : K ⊆ F, K compact},

for all F open and for all F ∈S with µ(F)< ∞,

µ(F) = inf{µ(V ) : V ⊇ F, V open}

for all F ∈S , and ∫f dµ = L f for all f ∈Cc(Ω). (19.9)

The plan is to define an outer measure and then to show that it, together with the σ

algebra of sets measurable in the sense of Caratheodory, satisfies the conclusions of thetheorem. Always, K will be a compact set and V will be an open set.

Definition 19.7.3 µ(V )≡ sup{L f : f ≺V} for V open,

µ( /0) = 0, µ(E)≡ inf{µ(V ) : V ⊇ E}

for arbitrary sets E.

516 CHAPTER 19. HAUSDORFF SPACES AND MEASURESFrom the picture as needed: Let u(U \ F) < € where U is open and let K C U andw(U\ K) <e, u(V \(U \ F)) < € with V open and V DU \ F =UNF® so V© CUS UF.This is possible because all sets are in .Y%. Then VONKG (US UF) OK =FNK CF.NowV°NK is compact andu(F\(KAVS)) = w(FO(KOUV)) =H(FAV) +H (FNK®)< w(FAV)+uU(U\K) <u(FAV)+e (19.6)However,e>u(V\(U\F)) =n (Va(UnF)*) =n (Vn (UCUF)) > HVOF)and so from 19.6, (F \ (KN vo)) <2e. Since KNV® is compact, this shows 19.5.19.7. Measures and Positive Linear FunctionalsThis is on the Riesz representation theorem for positive linear functionals. It is a reallymarvelous result. It produces measures on locally compact Hausdorff spaces. Thus thisdoesn’t help a lot in producing measures on infinite dimensional spaces but it works greaton R” or closed subsets of IR” and so forth.Definition 19.7.1 Le (Q,7) be a topological space. L:C.(Q) — C is called apositive linear functional if L is linear, L(af| +bf2) =aLf\ +bL fz, and if Lf > 0 wheneverf=0.Theorem 19.7.2 (Riesz representation theorem) Let (Q,7) be a locally compactHausdorff space and let L be a positive linear functional on C,(Q). Then there exists a 0algebra S containing the Borel sets and a unique measure LL, defined on Y, such thatLt is complete, (19.7)H(K) < forall K compact, (19.8)L(F) = sup{u(K):K CF, K compact},for all F open and for all F € S with U(F) <,L(F) =inf{u(V) :V D> F, V open}forall F © Y, and[fan =Lf forall f €C.(Q). (19.9)The plan is to define an outer measure and then to show that it, together with the oalgebra of sets measurable in the sense of Caratheodory, satisfies the conclusions of thetheorem. Always, K will be a compact set and V will be an open set.Definition 19.7.3 y(V) =sup{Lf : f <V} for V open,(0) =0, u(E) =int{u(V) :V 2B}for arbitrary sets E.