2708 CHAPTER 79. INCLUDING STOCHASTIC INTEGRALS

and this reduces to

λeλ (·)(

w−∫ (·)

0e−λ (·)

ΦdW)+ eλ (·)

(w−

∫ (·)

0e−λ (·)

ΦdW)′

+1n

F(

eλ (·)w)+A

(eλ (·)w, t,ω

)= f

and this reduces to

λ

(w−

∫ (·)

0e−λ (·)

ΦdW)+

(w−

∫ (·)

0e−λ (·)

ΦdW)′

+e−λ (·) 1n

F(

eλ (·)w)+ e−λ (·)A

(eλ (·)w, t,ω

)= e−λ (·) f , w(0) = u0

which implies (w−

∫ (·)

0e−λ (·)

ΦdW)′

+λw+ e−λ (·) 1n

F(

eλ (·)w)

+e−λ (·)A(

eλ (·)w, t,ω)

= λ

∫ (·)

0e−λ (·)

ΦdW + e−λ (·) f , w(0) = u0

which is equivalent to

w(t)+∫ t

0e−λ (·) 1

nF(

eλ (·)w)

ds+∫ t

0λw+ e−λ (·)A

(eλ sw(s) ,s,ω

)ds =

∫ t

0f̂ (s)ds

where f̂ (·) = e−λ (·) f +λ∫ (·)

0 e−λ (·)ΦdW . You can consider

Ã(w, t,ω)≡ e−λ tA(

eλ tw, t,ω).

This satisfies similar conditions to A. If F were a linear Riesz map, then you would get thesame type of problem but with λ = 0 for λ large enough. It may work in other cases also.

Definition 79.4.2 Let A be given above. Then z ∈ Â(u) means that for u ∈ V , z(t,ω) ∈A(u(t,ω) , t,ω) for a.e.(t,ω), z ∈ V ′

Thus  : V →P (V ′). Now let F be the duality map from U to U ′ which satisfies

⟨Fu,u⟩= ||u||rU , ||Fu||= ||u||r−1 , r ≥max(2, p)

Thus r is at least 2. We assume that u0 ∈ L2 (Ω;H) and is F0 measurable and for eachn there exist (un,zn) a progressively measurable solution to the integral equation

un (t)−u0 +1n

∫ t

0Funds+

∫ t

0znds =

∫ t

0f ds+

∫ t

0ΦdW, zn ∈ Â(un) , f ∈ V p′ (79.4.44)

in U ′ω . This would be the case if λ I + 1

n F +A were monotone for large enough λ . Suchtheorems are now well known and versions of them are in [108]. The message here is

2708 CHAPTER 79. INCLUDING STOCHASTIC INTEGRALSand this reduces to(-) (-) ’re) (vf eeaw ) +e) (w- f e*eaw)0 0+oF (ew) +A (e)w,t,00) =fand this reduces to(-) () ’A (wf eeaw ) + (»-[ ven)0 0teA0F (ew) +e 4A (e4w,t,0) =e = Of w (0) = uowhich implies(+) ‘(vf eeaw ) awe AO lp (ew)0 n0)4e20A(MOwt,0) = a | eObdw+e% Of, w(0) = uo0which is equivalent toore eh Aw+e7 04 (ew(s),s,0)as= [ flo)aswhere f(-) =e AO f+a fe (J@dW. You can considerA(w,t,@) =e A G 'w,t,@).This satisfies similar conditions to A. If F were a linear Riesz map, then you would get thesame type of problem but with A = 0 for A large enough. It may work in other cases also.Definition 79.4.2 Let A be given above. Then z € A(u) means that for u € V, z(t,@) €A(u(t,@) ,t,@) for a.e.(t,@),zEV'Thus A: ¥ + Y(V'). Now let F be the duality map from U to U’ which satisfiesr r-1(Fu,u) = \lul|y » ||Ful| = |u|", r= max (2, p)Thus r is at least 2. We assume that uw € L? (Q;H) and is Yo measurable and for eachn there exist (uU,,Z,) a progressively measurable solution to the integral equation1 rt t t t . ,Un (t) —uy+— [ Funds-+ [ Znds =| fas+ | @dW, 27, €A(un), fe VV? (79.4.44)0 0 0 0in Y%. This would be the case if AJ + iF +A were monotone for large enough 2. Suchtheorems are now well known and versions of them are in [108]. The message here is