2692 CHAPTER 79. INCLUDING STOCHASTIC INTEGRALS

Then also

P(∫ T

0∥un−um∥α

V̂ ds > λ

)≤ e(m,n)

λ

and so there exists a subsequence, denoted by r such that

P(∫ T

0∥ur−ur+1∥α

V̂ ds≤ 2−r)< 2−r

Thus, by the Borel Cantelli lemma, there is a further enlarged set of measure zero, stilldenoted as N such that for ω /∈ N∫ T

0∥ur−ur+1∥α

V̂ ds≤ 2−r

for all r large enough. Hence, by the usual proof of completeness, for these ω,

{ur (·,ω)}

is Cauchy in Lα([0,T ] ,V̂

)and also ur (t,ω) converges to some u(t,ω) pointwise in V̂ for

a.e. t. In addition, from 79.1.20 these functions are a Cauchy sequence in Lα([0,T ]×Ω;V̂

)with respect to the σ algebra of progressively measurable sets. Thus from Lemma 76.3.4,it can be assumed that for ω off the set of measure zero, (t,ω)→ u(t,ω) is progressivelymeasurable. From now on, this will be the sequence or a further subsequence. For ω /∈ N,a set of measure zero and 79.1.18, there is a further subsequence for which the followingconvergences occur as r→ ∞.

ur→ u weakly in V (79.1.21)

B(ur)→ B(u) weakly in V ′ (79.1.22)

zr→ z weakly in V ′ (79.1.23)(B(

ur−∫ (·)

0ΦrdW

))′→(

B(

u−∫ (·)

0ΦdW

))′weakly in V ′ (79.1.24)

∫ (·)

0ΦrdW →

∫ (·)

0ΦdW uniformly in C ([0,T ] ;W ) (79.1.25)

Bur (t)→ Bu(t) weakly in V ′ (79.1.26)

Bu(0) = Bu0, (79.1.27)

Bu(t) = B(u(t)) a.e. t (79.1.28)

In addition to this, we can choose the subsequence such that

supr

supt ̸=s

∥∥∫ ts ΦrdW

∥∥|t− s|γ

<C (ω)< ∞ (79.1.29)

This is thanks to Corollary 79.1.1. The boundedness of the operator A, in particular thegiven estimates, imply that zr is bounded in Lp′ ([0,T ]×Ω,V ′) . Thus a subsequence canbe obtained which yields weak convergence of zr in Lp′ ([0,T ]×Ω,V ′) and then Lemma

2692 CHAPTER 79. INCLUDING STOCHASTIC INTEGRALSTe(/ \|n — Um0and so there exists a subsequence, denoted by r such thatTP (| le —upallds < >) <20Thus, by the Borel Cantelli lemma, there is a further enlarged set of measure zero, stilldenoted as N such that for @ ¢ NThen alsoe(m,n)Arfas >a) <T[lu wrsillfas s 270for all r large enough. Hence, by the usual proof of completeness, for these @,{ur (-,@)}is Cauchy in L® ({0, T| V) and also u,(t,@) converges to some u(t, @) pointwise in V fora.e. t. In addition, from 79.1.20 these functions are a Cauchy sequence in L® ({0, T] x Q; Vv)with respect to the o algebra of progressively measurable sets. Thus from Lemma 76.3.4,it can be assumed that for @ off the set of measure zero, (t,@) — u(t, @) is progressivelymeasurable. From now on, this will be the sequence or a further subsequence. For @ ¢ N,a set of measure zero and 79.1.18, there is a further subsequence for which the followingconvergences occur as r — ©,uy — u weakly in V (79.1.21)B(u,) + B(u) weakly in V’ (79.1.22)z, — z weakly in V’ (79.1.23)() ' (+) / ,Blu,.— [ o,dw)) +(Blu— [ @aw )) weakly in ¥ (79.1.24)0 0(+) (-) : ;| ®,dWw > [ @dW uniformly in C([0,7];W) (79.1.25)Jo JoBu, (t) + Bu(t) weakly in V’ (79.1.26)Bu(0) = Buo, (79.1.27)Bu(t) =B(u(t)) ae. t (79.1.28)In addition to this, we can choose the subsequence such thatIl f &-dW |sup sup ~~, < C(@) < (79.1.29)rts |t—s|”This is thanks to Corollary 79.1.1. The boundedness of the operator A, in particular thegiven estimates, imply that z, is bounded in L? ({0,7] x Q,V’). Thus a subsequence canbe obtained which yields weak convergence of z, in L” ([0,7] x Q,V’) and then Lemma