2366 CHAPTER 69. GELFAND TRIPLES

and for all t ∈ [0,T ] , ∫ t

0f ′ (s)ds ∈ H, (69.2.11)

and for a.e. t ∈ [0,T ] ,

f (t) = f (0)+∫ t

0f ′ (s)ds in H, (69.2.12)

Here f ′ is being taken in the sense of V ′ valued distributions and 1p +

1p′ = 1 and p≥ 2.

Proof: Let Ψ ∈C∞c (−T,2T ) satisfy Ψ(t) = 1 if t ∈ [−T/2,3T/2] and Ψ(t) ≥ 0. For

t ∈ R, define

f̂ (t)≡{

f (t)Ψ(t) if t ∈ [−T,2T ]0 if t /∈ [−T,2T ]

and

fn (t)≡∫ 1/n

−1/nf̂ (t− s)φ n (s)ds (69.2.13)

where φ n is a mollifier having support in (−1/n,1/n) . Then by Minkowski’s inequality∣∣∣∣∣∣ fn− f̂∣∣∣∣∣∣

Lp(R;V )=

(∫R

∣∣∣∣∣∣∣∣ f̂ (t)−∫ 1/n

−1/nf̂ (t− s)φ n (s)ds

∣∣∣∣∣∣∣∣pV

dt)1/p

=

(∫R

∣∣∣∣∣∣∣∣∫ 1/n

−1/n

(f̂ (t)− f̂ (t− s)

)φ n (s)ds

∣∣∣∣∣∣∣∣pV

dt)1/p

≤(∫

R

(∫ 1/n

−1/n

∣∣∣∣∣∣ f̂ (t)− f̂ (t− s)∣∣∣∣∣∣

Vφ n (s)ds

)p

dt)1/p

≤∫ 1/n

−1/nφ n (s)

(∫R

∣∣∣∣∣∣ f̂ (t)− f̂ (t− s)∣∣∣∣∣∣p

Vdt)1/p

ds

≤∫ 1/n

−1/nφ n (s)εds = ε

provided n is large enough. This follows from continuity of translation in Lp with Lebesguemeasure. Since ε > 0 is arbitrary, it follows fn → f̂ in Lp (R;V ) . Similarly, fn → f inL2 (R;H). This follows because p ≥ 2 and the norm in V and norm in H are related by|x|H ≤C ||x||V for some constant, C. Now

f̂ (t) =

Ψ(t) f (t) if t ∈ [0,T ] ,Ψ(t) f (2T − t) if t ∈ [T,2T ] ,Ψ(t) f (−t) if t ∈ [0,T ] ,0 if t /∈ [−T,2T ] .

An easy modification of the argument of Lemma 69.2.11 yields

f̂ ′ (t) =

Ψ′ (t) f (t)+Ψ(y) f ′ (t) if t ∈ [0,T ] ,Ψ′ (t) f (2T − t)−Ψ(t) f ′ (2T − t) if t ∈ [T,2T ] ,Ψ′ (t) f (−t)−Ψ(t) f ′ (−t) if t ∈ [−T,0] ,0 if t /∈ [−T,2T ] .

.

2366 CHAPTER 69. GELFAND TRIPLESand for allt € {0,T],t[ f'(s)ds €H, (69.2.11)0and for a.e. t € [0,T],tf(t) =f (0) + f' (s)ds in H, (69.2.12)0Here f' is being taken in the sense of V' valued distributions and ; + y = land p>2?.Proof: Let ¥ € Ce (—T,2T) satisfy Y (t) = 1 if t € [-T/2,3T/2] and Y(t) > 0. Fort ER, definea, _ f f(t) (et) ift € [-7,27]fin=| 0 if? ¢ [7,27]and /= / Fit—s)o, (s)ds (69.2.13)—I/nwhere @,, is a mollifier having support in (—1/n,1/n). Then by Minkowski’s inequality. In. P \/pAbney = Uf |[FO- [Fe 4nasl ar)pr A . P 1/p= (L\/ (Fl) —Flt—s)) 6, (s) as ar)—I/n Vpe A . P 1/p< (Lf) ||Fea-Fe—s)|, ents)as) ar)1/n p 1/p< [96 (fF - fits ar) dsI/n/ od, (s) eds =€—1/nprovided n is large enough. This follows from continuity of translation in L’? with Lebesguemeasure. Since € > 0 is arbitrary, it follows f, > f in L?(R;V). Similarly, f, > f inL’ (R;H). This follows because p > 2 and the norm in V and norm in H are related by\x|47 <C||x||y for some constant, C. NowW(t) f (0) ift € [0,7],~ J W(t) f (27-1) ift € [7,27],MO=) wo) pn) ifr e (0.7),Oift ¢ [—T,27].An easy modification of the argument of Lemma 69.2.11 yieldset Y(y) f'(@) ift € [0,7],P()= W(t) f (2T —t) —P(t) f’ (2T —2) ift € [7,27],W(t) f (—t) —¥ (0) f’ (1) ifr [-T,0],Oift ¢ [—-T,2T].