2632 CHAPTER 77. STOCHASTIC INCLUSIONS

Theorem 77.5.6 Let the conditions on A hold 77.5.25 - 77.5.28, 77.5.30 - 77.5.34. Also letB satisfy 77.3.6 and assume, if it depends on ω, it is of the form

B(ω) = k (ω)B, k (ω)≥ 0, k measurable

Let u0 be F measurable into W, and let f be product measurable into V ′, f (·,ω) ∈ V ′.Then there exists a solution to the following evolution inclusion

(B(ω)u(·,ω))′+u∗ (·,ω) = f (·,ω) in V ′

B(ω)u(0,ω) = B(ω)u0 (ω) (77.5.57)

where u∗ (·,ω)∈ A(u(·,ω) ,ω). In addition to this, (t,ω)→ u(t,ω) is product measurableinto V and (t,ω)→ u∗ (t,ω) is product measurable into V ′.

In place of the coercivity condition 77.5.28 assume the coercivity condition involvingboth B and A given in 77.5.29. Then

(B(ω)u(·,ω))′+u∗ (·,ω) = f (·,ω) in U ′ (77.5.58)B(ω)u(0,ω) = B(ω)u0 (ω) (77.5.59)

Thus the following holds in V ′

(B(ω)u(·,ω))(t)−B(ω)u0 (ω)+∫ t

0u∗ (·,ω)ds =

∫ t

0f (s,ω)ds

(Bu)′ ∈ V ′

Proof of Theorem 77.5.6: First consider the claim about replacing the coercivity con-dition. Returning to 77.5.50, one obtains by integrating up to t and adding λ

∫ t0 ⟨Buε ,uε⟩ds

to both sides,12⟨Buε ,uε⟩(t)−

12⟨Bu0,u0⟩+ ε

∫ t

0⟨Fuε ,uε⟩ds

+∫ t

0⟨u∗ε ,uε⟩ds+λ

∫ t

0⟨Buε ,uε⟩ds =

∫ t

0⟨ f ,uε⟩ds+λ

∫ t

0⟨Buε ,uε⟩ds (77.5.60)

Then from the estimate 77.5.29,

12⟨Buε ,uε⟩(t)−

12⟨Bu0,u0⟩+ ε

∫ t

0⟨Fuε ,uε⟩ds

+δ (ω)∫ t

0∥uε∥p

V ds−m(ω) =∫ t

0⟨ f ,uε⟩ds+λ

∫ t

0⟨Buε ,uε⟩ds (77.5.61)

From this, it is a routine use of Gronwall’s inequality to obtain the estimate

ε ⟨Fuε ,uε⟩U ′,U +∥uε∥V ≤C (u0, f ,λ ,ω) (77.5.62)

Then the rest of the argument is the same. You obtain the following in U ′.

B(ω)u(t,ω)−B(ω)u0 (ω)+∫ t

0u∗ (·,ω)ds =

∫ t

0f (s,ω)ds

2632 CHAPTER 77. STOCHASTIC INCLUSIONSTheorem 77.5.6 Let the conditions on A hold 77.5.25 - 77.5.28, 77.5.30 - 77.5.34. Also letB satisfy 77.3.6 and assume, if it depends on Q, it is of the formB(@) =k(@)B, k(@) > 0, k measurableLet ug be ¥ measurable into W, and let f be product measurable into V', f (-,@) € ¥'.Then there exists a solution to the following evolution inclusion(B(@)u(-,@))'+uX(,@) = f(-,@) in ¥'B(@)u(0,@) = B(@)uo(@) (77.5.57)where u* (-,@) € A (u(-,@),@). In addition to this, (t,@) — u(t, @) is product measurableinto V and (t,@) — u* (t,@) is product measurable into V’.In place of the coercivity condition 77.5.28 assume the coercivity condition involvingboth B and A given in 77.5.29. Then(B(@)u(-,@))'+u*(-,@) = f(-,@)inY”' (77.5.58)B(@)u(0,@) = B(@)uo(@) (77.5.59)Thus the following holds in V't t(B(@)u(-,0))()—B(o)w(o)+ [u'(.e)ds = | P(s,e)as(Buy) ¢ ¥V'Proof of Theorem 77.5.6: First consider the claim about replacing the coercivity con-dition. Returning to 77.5.50, one obtains by integrating up to t and adding A Jj (Bue, ue) dsto both sides,11 t5 (Bue, ue) (t) ~ 5 (Buo, uo) +e/ (Fue,ue) dst t t t+f (ug,e)ds-+2 | (Bue, ue) ds = | (Faue)ds-+A [ (Bug, Ue) ds (77.5.60)0 0 0 0Then from the estimate 77.5.29,1 1 t5 (Bue, te) (t) ~ 5 (Buosuo) +€ f (Fute,ue) ds+5 (0) [ lluellfds—m(o) -[ (Psue)ds+2 f (Bue, tte) ds (77.561)From this, it is a routine use of Gronwall’s inequality to obtain the estimateE (Fuge) ay wy + || ute || y < C(uo, f,A, @) (77.5.62)Then the rest of the argument is the same. You obtain the following in U’.B(w)u(t,@) ~B(«)uo(o)+ ['w'(.@)ds= | f(s,0)ds