2710 CHAPTER 79. INCLUDING STOCHASTIC INTEGRALS

where Mn (t) is a local martingale whose quadratic variation is

[Mn] (t)≤C∫ t

0||Φ||2L2

|un|2H ds

Then estimates give∥∥ 1

n ⟨Fun,un⟩∥∥

U ′ bounded as well as

∥un∥V ,∥zn∥V ′ (79.4.46)

One takes expectation of 79.4.45 using an appropriate localizing sequence of stoppingtimes if necessary. It follows that there is a subsequence, still denoted with n such that

un → u weakly in V (79.4.47)zn → z weakly in V ′

1n

Fun → 0 strongly in U ′

The last convergence follows from the following argument.∫ T

0

∫Ω

1n⟨Fun,w⟩dPdt

≤∫ T

0

∫Ω

1n1/r′

⟨Fun,un⟩1/r′ 1n1/r⟨Fw,w⟩1/r dPdt

≤(∫ T

0

∫Ω

1n⟨Fun,un⟩dPdt

)1/r′(∫ T

0

∫Ω

1n∥w∥r dPdt

)1/r

≤ C1

n1/r∥w∥U

and so ∣∣∣∣∣∣∣∣1nFun

∣∣∣∣∣∣∣∣U ′≤ C

n1/r

Recall

un (t)−u0 +1n

∫ t

0Funds+

∫ t

0znds

=∫ t

0f ds+

∫ t

0ΦdW, zn ∈ Â(un)

Then if φ ∈U , ⟨1n

∫ (·)

0Funds+

∫ (·)

0znds,φ

⟩U ′,U

=∫

∫ T

0

⟨(1n

∫ r

0Funds+

∫ r

0znds

),φ (r,ω)

⟩U ′,U

drdP

Then the above equals∫Ω

∫ T

0

1n

∫ r

0⟨Fun (s,ω) ,φ (r,ω)⟩dsdrdP+

∫ r

0⟨zn (s,ω) ,φ (r,ω)⟩dsdrdP

2710 CHAPTER 79. INCLUDING STOCHASTIC INTEGRALSwhere M,, (t) is a local martingale whose quadratic variation ist(Mal) <C [Uf blir dsbounded as well asThen estimates give || + (Fun, un) yl[unl y seul ys (79.4.46)One takes expectation of 79.4.45 using an appropriate localizing sequence of stoppingtimes if necessary. It follows that there is a subsequence, still denoted with n such thatUn —- uweakly in ¥ (79.4.47)Zn —> zweakly in V’1—Fun — Ostrongly in Y’nThe last convergence follows from the following argument.Teri[ [ — (Fun, w) dPdt0 JanT 1 Lr! 1 7- L bague Fene wae ww) "dPdtT 1 1/r T 1 1/r(/ [5 (Funua)aPar) (/ [loll arar)0 JQN 0 Jan<1< Cor Illaand so ;—Fuy <n yr ni/rRecall1 t tUn (t) — ug + — [ Funds + [ zpdsn Jo 0t t= [ fase | @dW, zn € A (un)0 0Then if dE Y,1 0) ()C | Funds+ [ cnds.6n JO 0 UU0 oT 1 a er-[ | (5 | Funds-+ | cals) 0 (0) ) drdPaJo n Jo JO UUThen the above equalsLf (Fun (s,@), 0 (7, w))dsdrdP+ | (Zn (s, @) ,@ (1, @)) dsdrdP