34.2. STANDARD TECHNIQUES IN EVOLUTION EQUATIONS 1161

for fixed w ∈ V . Suppose un → u in V and fix w ∈ L∞ (D) ⊆ V . Then it follows from aneasy argument using the Vitali convergence theorem and the fact that from the estimatesabove

Ψ(un)w

is uniformly integrable that

u→−∫

DΨ(u)wdx

is continuous. For general w ∈ Lp (D) , let wn→ w in Lp (D) where each wn is in L∞ (D).Then the function

u→−∫

DΨ(u)wdx≡ ⟨Au,w⟩ (34.1.7)

is the limit of the continuous functions

u→−∫

DΨ(u)wndx

and so the function 34.1.7 is Borel measurable. Now by the Pettis theorem this showsA : V →V ′ is B (V ) measurable. This shows A is an example of an operator which satisfiessome conditions which will be considered later.

34.2 Standard Techniques In Evolution EquationsIn this section, several significant theorems are presented. Unless indicated otherwise, themeasure will be Lebesgue measure. First here is a lemma.

Lemma 34.2.1 Suppose g ∈ L1 ([a,b] ;X) where X is a Banach space. Then if∫ b

ag(t)φ (t)dt = 0

for all φ ∈C∞c (a,b) , then g(t) = 0 a.e.

Proof: Let S be a measurable subset of (a,b) and let K ⊆ S ⊆ V ⊆ (a,b) where Kis compact, V is open and m(V \K) < ε. Let K ≺ h ≺ V as in the proof of the Rieszrepresentation theorem for positive linear functionals. Enlarging K slightly and convolvingwith a mollifier, it can be assumed h ∈C∞

c (a,b) . Then∣∣∣∣∫ b

aXS (t)g(t)dt

∣∣∣∣ =

∣∣∣∣∫ b

a(XS (t)−h(t))g(t)dt

∣∣∣∣≤

∫ b

a|XS (t)−h(t)| ||g(t)||dt

≤∫

V\K||g(t)||dt.

Now let Kn ⊆ S⊆Vn with m(Vn \Kn)< 2−n. Then from the above,∣∣∣∣∫ b

aXS (t)g(t)dt

∣∣∣∣≤ ∫ b

aXVn\Kn (t) ||g(t)||dt

34.2. STANDARD TECHNIQUES IN EVOLUTION EQUATIONS 1161for fixed w € V. Suppose u, — u in V and fix w€ L’(D) CV. Then it follows from aneasy argument using the Vitali convergence theorem and the fact that from the estimatesaboveYW (un) wis uniformly integrable thatn> — [Ww wais continuous. For general w € L? (D), let w, — w in L? (D) where each wy is in L® (D).Then the functionu>— | W (u) wdx = (Au,w) (34.1.7)JDis the limit of the continuous functionsu— -| W (u) wpdxDand so the function 34.1.7 is Borel measurable. Now by the Pettis theorem this showsA:V —>V’' is @(V) measurable. This shows A is an example of an operator which satisfiessome conditions which will be considered later.34.2 Standard Techniques In Evolution EquationsIn this section, several significant theorems are presented. Unless indicated otherwise, themeasure will be Lebesgue measure. First here is a lemma.Lemma 34.2.1 Suppose g € L! ({a,b];X) where X is a Banach space. Then ifb[ s@o@a=0for all @ € Ce (a,b), then g(t) =Oae.Proof: Let S be a measurable subset of (a,b) and let K C S C V C (a,b) where Kis compact, V is open and m(V\K) < €. Let K < h < V as in the proof of the Rieszrepresentation theorem for positive linear functionals. Enlarging K slightly and convolvingwith a mollifier, it can be assumed h € CP (a,b). Thenb b[ e%osoal) = | [ (a@-rM)eoaIA[125 -nepiilgwoilatJ, lis @lletNow let K, C S CV, with m(V,, \ K,) < 27". Then from the above,IAb bi Xs(t)g(t)dt| < | DX, x, (t)|lg (t)|\ae