1194 CHAPTER 34. GELFAND TRIPLES AND RELATED STUFF

≤C (∥B∥ ,T )∥u− v∥X

Thus u→ Bu(0) is continuous into V ′. If p < 2, then you do something similar.(∫ T

0∥Bu(0)−Bv(0)∥p

V ′ dt)1/p

≤(∫ T

0∥Bu(t)−Bv(t)∥p

V ′ dt)1/p

+

(∫ T

0

∥∥∥∥∫ t

0(Bu)′ (s)− (Bv)′ (s)ds

∥∥∥∥p

dt)1/p

≤ ∥B∥∥u− v∥Lp([0,T ];V )+∫ T

0

(∫ T

0

∥∥(Bu)′ (s)− (Bv)′ (s)∥∥p dt

)1/p

ds

≤ ∥B∥∥u− v∥Lp([0,T ];V )+T 1/p∫ T

0

∥∥(Bu)′ (s)− (Bv)′ (s)∥∥ds

≤ ∥B∥∥u− v∥Lp([0,T ];V )+CT 1/p∥∥(Bu)′− (Bv)′∥∥

Lp′ ([0,T ];V ′)

Thus

∥Bu(0)−Bv(0)∥V ′ T1/p ≤ ∥B∥∥u− v∥Lp([0,T ];V )+C (T )

∥∥(Bu)′− (Bv)′∥∥

Lp′ ([0,T ];V ′)

≤ C (∥B∥ ,T )∥u− v∥X .

However, one could just as easily have done this for an arbitrary s < T by repeating theargument for

Bu(t) = Bu(s)+∫ t

s(Bu)′ (r)dr

Thus this mapping is certainly continuous into V ′. The last assertion is similar. You justuse f instead of Bu and make easy modifications in the argument. It is all happening in onespace in the second case.

For u ∈ X defined above,

Bu(t) = Bu(0)+∫ t

0(Bu)′ (s)ds,

and also12⟨Bu(t) ,u(t)⟩= 1

2⟨Bu,u⟩(0)+

∫ t

0

⟨(Bu)′ (s) ,u(s)

⟩ds

This follows from a similar argument given above, (Note we write ⟨Bu,u⟩(0) instead of⟨Bu0,u0⟩ since no u0 is mentioned. One could also use the above by considering the prob-lem on [s, t] where s is not in the exceptional set where it makes a difference betweenwriting Bu(s) and B(u(s)) . Then you would get the above with 0 replaced with s and thenlet s→ 0 to finally obtain the above displayed formula. ) and

supt∈[0,T ]

⟨Bu,u⟩(t)≤C(∥∥(Bu)′

∥∥Lp′ (0,T,V ′) ,∥u∥Lp(0,T,V )

)=C (∥u∥X )

1194 CHAPTER 34. GELFAND TRIPLES AND RELATED STUFF<C((||Bl| 7) lu —vllxThus u — Bu (0) is continuous into V’. If p < 2, then you do something similar.("12% (0)—svionitear) = ([" ute) avoir)“ 7\/p< [all lu—vivcanwy +f) (ff Kew!) vy) |P'ar) as‘T< [Blu vlargozrivy +7"” | |(Bu)"(s) — Bry (6)|Las[ e1s! (6) =n)" Nas< [Bll lu — vllen(o.rjvy + CT’? || (Bu)! — Bv)' [Ley go-ry.vyThus||Bu(0)—Bv(0)|ly,T/? < Bll lu — v\|zoqo,r).v) EC (7) || Bu)’ — (Bv)'|| 10" (o.r}v1)< C(\Bl|,7) ju—vlly.However, one could just as easily have done this for an arbitrary s < T by repeating theargument forBu (t) = Bu(s) + i ' (Bu)! (r)drThus this mapping is certainly continuous into V’. The last assertion is similar. You justuse f instead of Bu and make easy modifications in the argument. It is all happening in onespace in the second case. ffFor u € X defined above,Bu (t) = Bu(0) + [ (Bu)! (s)ds,and also5 (But) u(0)) = 5 (Buu) (0)+ [ ((Bu)'(s) ,u(s))asThis follows from a similar argument given above, (Note we write (Bu,u) (0) instead of(Bug, Uo) since no up is mentioned. One could also use the above by considering the prob-lem on [s,t] where s is not in the exceptional set where it makes a difference betweenwriting Bu(s) and B(u(s)). Then you would get the above with 0 replaced with s and thenlet s + 0 to finally obtain the above displayed formula. ) andsup (Buu) (t) <C (||(Bu)' lp r.y9sllelluocrvy) =C (Mel)te[0,T] _