El e c t ro nic
Journ a l of
Pr
ob a b il i t y
Vol. 15 (2010), Paper no. 17, pages 484–525.
Journal URL
http://www.math.washington.edu/~ejpecp/
Support theorem for a stochastic Cahn-Hilliard equation
∗Lijun Bo1 Kehua Shi2,†and Yongjin Wang3
1Department of Mathematics, Xidian University, Xi’an 710071, China bolijunnk@yahoo.com.cn
2School of Mathematical Sciences, Xiamen University, Xiamen 361005, China kehuashink@gmail.com
3School of Mathematical Sciences, Nankai University, Tianjin 300071, China yjwang@nankai.edu.cn
Abstract
In this paper, we establish a Stroock-Varadhan support theorem for the global mild solution to a d (d≤3)-dimensional stochastic Cahn-Hilliard partial differential equation driven by a space- time white noise.
Key words: Stochastic Cahn-Hilliard equation, Space-time white noise, Stroock-Varadhan sup- port theorem.
AMS 2000 Subject Classification:Primary 60H15, 60H05.
Submitted to EJP on November 7, 2009, final version accepted Aril 15, 2010.
∗The research of K. Shi and Y. Wang was supported by the LPMC at Nankai University and the NSF of China (No.
10871103). The research of L. Bo was supported by the Fundamental Research Fund for the Central Universities (No.
JY10000970002).
†Corresponding author. Email: kehuashink@gmail.com
1 Introduction and main result
In this paper, we consider the following stochastic Cahn-Hilliard equation:
∂u/∂t=−∆
∆u+f(u)
+σ(u)W˙, in[0,T]×D, u(0) =ψ,
∂u/∂n=∂[∆u]/∂n=0, on[0,T]×∂D,
(1.1)
where ∆ denotes the Laplace operator, the domain D= [0,π]d (d = 1, 2, 3), and f :R→Ris a polynomial of degree 3 with positive dominant coefficient (which is due to the background of the equation from material science). Assume thatσ:R→Ris a bounded and Lipschitzian function and W˙ is a Gaussian space-time white noise on some complete probability space(Ω,F,P)satisfying
E
W˙(x,t)W˙(y,s)
=δ(|t−s|)δ(|x−y|), (t,x),(s,y)∈[0,T]×D.
Hereδ(·)is the Dirac delta function concentrated at the point zero.
The (deterministic) Cahn-Hilliard equation (i.e.,σ≡0 in (1.1)) has been extensively studied (see, e.g.,[2; 3; 4; 5; 10; 15; 18]) as a well-known model of the macro-phase separation that occurs in an isothermal binary fluid, when a spatially uniform mixture is quenched below a critical temperature at which it becomes unstable. A stochastic version of the Cahn-Hilliard equation (when σ ≡ 1 in (1.1)) was first proposed by Da Prato and Debussche [8], and the existence, uniqueness and regularity of the global mild solution were explored. In Cardon-Weber[6], the authors considered this type of stochastic equation in a general case onσ, which is equivalent to the following form:
u(t,x) = Z
D
Gt(x,y)ψ(y)dy+ Z t
0
Z
D
∆Gt−s(x,y)f(u(s,y))dyds +
Z t
0
Z
D
Gt−s(x,y)σ(u(s,y))W(dy, ds), (1.2) where Gt(·,∗) denotes the Green kernel corresponding to the operator∂ /∂t+ ∆2 with the homo- geneous Neumann’s boundary condition as in (1.1). Since Stroock and Varadhan[17]established their famous support theorem for diffusion processes, there have been many research works on this issue, for example, a variety of support theorems for 1-dimensional second-order parabolic and hy- perbolic stochastic partial differential equations (abbr. SPDEs) have been discussed in the literature (see, e.g.,[1; 7; 13; 14]). Millet and Sanz-Solé [13] characterized the support of the law of the solution to a class of hyperbolic SPDEs, which simplified the proof in[17]. In Bally et al. [1], the authors proved a support theorem for a semi-linear parabolic SPDE. Moreover, a support result for a generalized Burgers SPDE (containing a quadratic term) was established in Cardon-Weber and Millet[7]. Herein, we are attempting to establish a support theorem of the law corresponding to the solution to Equation (1.1) inC([0,T],Lp([D]))forp≥4. The main strategy used in this paper is an approximation procedure by using a space-time polygonal interpolation for the white noise, and we particularly adopt a localization argument, which was used in [7] for studying a support theorem of a Burgers-type equation. However here we need more technical estimates concerning the high-order Green kernelGt(·,∗), which sharp the estimates in[6](see Appendix).
In what follows, we introduce the main result of this paper. To do it, we define the following Cameron-Martin spaceH by
H =
¨
h(t,x) = Z t
0
Z
Qd i=1[0,xi]
h(s,y)dyds; (t,x) = (t,(x1, . . . ,xd))∈[0,T]×D,
h∈L2([0,T]×D)
« , and the corresponding norm by
khkH = s
Z T
0
Z
D
|h(s,y)|2dyds, for all h∈ H.
LetHb represent the subset of H, in which the first-order derivative h ofh∈ H is bounded. For h∈ H, consider the following skeleton equation:
S(h)(t,x) = Z
D
Gt(x,y)ψ(y)dy+ Z t
0
Z
D
∆Gt−s(x,y)f(S(h)(s,y))dyds +
Z t
0
Z
D
Gt−s(x,y)σ(S(h)(s,y))h(s,y)dyds. (1.3) Recall Equations (1.1) and (1.2). We make the following assumptions throughout the paper:
(H1). Assume thatσ:R→Ris bounded and belongs to C3(R)with bounded first to third-order partial derivatives, and
(H2). The initial functionψ∈Lp(D)forp≥4, andψis%∈]0, 1]–H¨older continuous.
Now we are at the position to state the main result of this paper.
Theorem 1.1. Under the assumptions (H1) and (H2), let u = (u(t,x))(t,x)∈[0,T]×D be the unique solution to Equation(1.2)in C([0,T],Lp(D))with p≥4and P◦u−1 denote the law (a probability measure)of the solution u. Recall the skeleton equation(1.3), and setSH ={S(h);h∈ H }. Then we have
(a) Let p>6. Then forα¯∈]0, min{12(1−d4),%
4}[andβ¯∈]0, min{2−d2,%}[, the topological support supp(P◦u−1)in Cα¯, ¯β([0,T]×D)of the lawP◦u−1is the closure ofSH.
(b) Let p ≥ 4. Then for α¯ ∈]0, min{12(1− d4),%
4}[, the topological support supp(P◦u−1) in Cα¯([0,T],Lp(D))of the lawP◦u−1is the closure ofSH.
The rest of this paper is organized as follows: In the coming section, we give a difference approxima- tion to the(d+1)–dimensional space-time white noise ˙W(x,t)and study some concrete properties of the approximating noises. In Section 3, we introduce a localization framework as in [7], and then switch to prove the support theorem by checking the conditions (C1) and (C2) below (see Section 3). Sections 4 and 5 are devoted to checking the validity of the conditions (C1) and (C2), respectively. In Section 6, we prove the continuity of the solutionS(h)to the skeleton equation (1.3) inHband finally we complete the proof of Theorem 1.1.
2 Difference approximation to white noise
In this section, we give a difference approximation to the (d+1)-dimensional space-time white noise W˙, which is a space-time polygonal interpolation for ˙W.
Letn∈Nandt∈[0,T], set tn= max
j∈{0,1,...,2n}
¦j T2−n;j T2−n≤t©
, and tn=
tn−T2−n
∨0.
Let k := (k1, . . . ,kd) ∈ Idn := {0, 1, . . . ,n−1}d. Define a partition (4j,k)j=0,1,...,2n−1,k∈Idn of OT := [0,T]×Dby
4j,k=Dk×]j T2−n,(j+1)T2−n], where Dk = Qd
j=1]kjπn−1,(kj+1)πn−1]. For xj ∈]kjπn−1,(kj+1)πn−1]with j = 1, . . . ,d, we set Dk(x) =Qd
j=1]kjπn−1,(kj +1)πn−1]. Further, for each (t,x) ∈ OT, we define the following difference approximation to ˙W by
W˙n(x,t) =
( W(4
j−1,k)
|4j−1,k| , (x,t)∈ 4j,k, j=1, . . . , 2n−1, k∈Idn, 0, (x,t)∈ 40,k, k∈Idn,
(2.1)
where 4j,k
= Tπd(nd2n)−1 is the volume of the partition4j,k for each j = 0, 1, . . . , 2n−1 and k∈Idn.
Next we suppose that
(H3). the mappings F,H,K : R → R are bounded, globally Lipschitzian and H ∈ C3(R) with bounded first to third-order derivatives.
We now consider the following equations forh∈ Hb, Xn(t,x) = Gt∗ψ(x)
+ Z t
0
Z
D
Gt−s(x,y)
F(Xn(s,y))W(dy, ds) +H(Xn(s,y))Wn(dy, ds) +
Z t
0
Z
D
Gt−s(x,y)h
K(Xn(s,y))h(s,y)
−H˙(Xn(s,y))[αn(s,y)F(Xn(s,y)) +βn(s,y)H(Xn(s,y))]i dyds +
Z t
0
Z
D
∆yGt−s(x,y)f(Xn(s,y))dyds, (2.2) and
X(t,x) = Gt∗ψ(x) + Z t
0
Z
D
Gt−s(x,y)[F+H](X(s,y))W(dy, ds) +
Z t
0
Z
D
Gt−s(x,y)K(X(s,y))h(s,y)dyds
+ Z t
0
Z
D
∆yGt−s(x,y)f(X(s,y))dyds, (2.3)
where
Gt∗ψ(x):= Z
D
Gt(x,y)ψ(y)dy, and for eachn∈N,
αn(t,x) := nd2n(Tπd)−1 Z tn
tn
Z
Dk(x)
Gt−s(x,y)dyds, βn(t,x) := nd2n(Tπd)−1
Z t
tn
Z
Dk(x)
Gt−s(x,y)dyds.
Forαn(t,x)andβn(t,x), by virtue of (A.4) in Lemma A.1, we claim that sup
(t,x)∈OT
|αn(t,x)| ≤ C nd, and (2.4)
sup
(t,x)∈OT
|βn(t,x)| ≤C nd. (2.5)
Indeed, using (A.4), we have for each t∈[0,T], sup
x∈D|αn(t,x)| ≤ C nd2nmax
k∈Idn
( 1Dk(x)
Z tn
tn
Z
Dk
|Gt−s(x,y)|dyds )
≤ C nd2n|tn−tn|
≤ C nd, and
sup
x∈D|βn(t,x)| ≤C nd2n|t−tn| ≤C nd, follows from the equality (A.19).
In the following, letF= (Ft)0≤t≤T be the natural filtration generated byW, i.e., Ft=σ{W(B×[0,s]);s∈[0,t], B∈ B(D)}.
Then for every t ∈ [0,T] and n ∈ N fixed, (W˙n(x,t))x∈D given by (2.1) is Ft-adapted. More precisely, it isFtn-adapted and which is independent of the informationFtn.
Lemma 2.1. For each fixed n∈Nand p≥1, we have sup
(t,x)∈OT
E
|W˙n(x,t)|p
≤Cpnd p22np2.
Proof. By virtue of the definition (2.1), sup
(t,x)∈OT
E
|W˙n(x,t)|p
= sup
(t,x)∈OT
E
2n−1
X
j=1
X
k∈Idn
W(4j−1,k)
|4j−1,k| 14j
−1,k(x,t)
p
≤ Cpmax
E
W(4j−1,k)
|4j−1,k|
p
; j=1, . . . , 2n−1, k∈Idn
.
Note that for each j=1, . . . , 2n−1 andk∈Idn, W(4j−1,k)
|4j−1,k| ∼N(0,|4j−1,k|−1). For any random variableZ∼N(0,σ2), it holds that
E|Z|p= 1 pπ
p
2pσ2pΓ p
2+1 2
, whereΓdenotes the Gamma function. This yields that
sup
(t,x)∈OT
E
|W˙n(x,t)|p
≤Cpmaxnp
|4j−1,k|−p; j=1, . . . , 2n−1, k∈Idno ,
and which proves the lemma.
Letn∈Nbe fixed. Forα >0 and t∈]0,T], we now define an event ¯Ωαn,t by Ω¯αn,t=
¨
ω∈Ω; sup
(s,y)∈[0,t]×D
W˙(y,s;ω)
≤αnd2n2
«
. (2.6)
For this event, we have Lemma 2.2. If chooseα >2
qlog 2 Tπd, then
nlim→∞Ph
Ω¯αn,Tic
=0.
Proof. Let Z ∼ N(0, 1) be a standard normal random variable. Then according to the definition (2.1) for ˙Wn(x,t),
Ph
Ω¯αn,Tic
= P
(j,k)∈{1,...,2maxn−1}×Idn
¨W(4j−1,k)
|4j−1,k|
«
≥αnd22n
≤ nd2nP
|Z| ≥αp
Tπdnd/2
= nd2nP
|Z|2
4 ≥ α2Tπd 4 nd
≤ nd2nexp
−α2Tπd 4 nd
E
exp
|Z|2 4
. (2.7)
Note thatE h
exp |Z|2
4
i=p
2. Then (2.7) further yields that 0≤Ph
Ω¯αn,Tic
≤ p
2ndexp
nlog 2− α2Tπd 4 nd
≤ p
2ndexp
nd
log 2−α2Tπd 4
→ 0, as n→ ∞, (2.8)
ifα >2 qlog 2
Tπd. Thus the proof of the lemma is complete.
3 Localization framework
In this section, we adopt a localization method used in[7]to deal with Equation (1.1). In addition, we will prove a key proposition, which is useful in the proof of Theorem 1.1.
Proposition 3.1. Under the assumptions (H1) and(H2), let X = (X(t,x))(t,x)∈[0,T]×D (resp. Xn) be the unique solution to Equation(2.3)(resp.(2.2))in C([0,T],Lp(D))with p≥4. Recall the skeleton equation(1.3), and setSH ={S(h);h∈ H }. Then we have
(i) Let p > 6. Then for α¯ ∈]0, min{12(1− d4),%4}[ and β¯ ∈]0, min{2− d2,%}[, the sequence Xn converges in probability to X in Cα¯, ¯β([0,T]×D).
(ii) Let p≥4. Then forα¯∈]0, min{12(1− d4),%4}[, the sequence Xn converges in probability to X in Cα¯([0,T],Lp(D)).
Next we give a sketch for the proof of the conclusion(ii)in Proposition 3.1. The similar argument can also be used to prove the part(i). For(t,x)∈ OT, set
Yn(t,x):=Xn(t,x)−X(t,x). From (2.2) and (2.3), it follows that
Yn(t,x) =
3
X
i=1
Γin(t,x) + Λn(t,x), (3.1)
where
Γ1n(t,x) := Z t
0
Z
D
Gt−s(x,y)
(F+H)(Xn(s,y))−(F+H)(X(s,y))
W(dy, ds), Γ2n(t,x) :=
Z t
0
Z
D
Gt−s(x,y)
K(Xn(s,y))−K(X(s,y))
h(s,y)dyds,
Γ3n(t,x) := Z t
0
Z
D
∆yGt−s(x,y)
f(Xn(s,y))− f(X(s,y)) dyds, and
Λn(t,x) := Z t
0
Z
D
Gt−s(x,y)H(Xn(s,y))
Wn(dy, ds)−W(dy, ds)
− Z t
0
Z
D
Gt−s(x,y)H˙(Xn(s,y))
×
αn(s,y)F(Xn(s,y)) +βn(s,y)H(Xn(s,y))
dyds. (3.2)
Introduce an auxiliaryFtn-adapted process
Xn−(t,x):=Gt−tn x,Xn(tn,·)
, for (t,x)∈ OT. (3.3)
Recall the localization argument adopted in[7]. Forγ∈(0, 1)andp≥4, define Φp,nγ(t):= sup
s∈[0,t]kYn(t,·)kp+ sup
s6=s0∈[0,t]
kYn(s,·)−Yn(s0,·)kp
|s−s0|γ , wherek · kp corresponds to the norm ofLp(D)and forδ >0,
τδn:=inf¦
t>0; Φp,nγ(t)≥δ©
∧T.
ForM> δ, define the following events At(M−δ) :=
¨
ω∈Ω; sup
s∈[0,t]kX(s,·)kp≤M−δ
«
, (3.4)
AMn(t) :=
¨
ω∈Ω; sup
s∈[0,t]kXn(s,·)kp∨ sup
s∈[0,t]kX(s,·)kp≤M
«
. (3.5)
Then fort∈]0,T],
At(M−δ)∩ {t≤τδn} ⊆AMn(t). (3.6) In fact, from the inequality|y| ≤ |x−y|+|x|, it follows that
At(M−δ)∩ {t≤τδn}
⊆
¨ sup
s∈[0,t]kX(s,·)kp≤M−δ
«
∩
¨ sup
s∈[0,t]kXn(s,·)−X(s,·)kp≤δ
«
⊆
¨ sup
s∈[0,t]kX(s,·)kp≤M
«
∩
¨ sup
s∈[0,t]kXn(s,·)kp≤M
«
= AMn(t).
Recall the event ¯Ωαn,t defined by (2.6) in Section 2 and thatα >2 qlog 2
Tπd. For each fixedδ >0 and V ∈Cγ([0,T];Lp(D)), set
kVkγ,p,τδn := sup
s∈[0,T∧τδn]kV(s,·)kp+ sup
s6=s0∈[0,T∧τδn]
kV(s,·)−V(s0,·)kp
|s−s0|γ .
Then forq≥1, P
Φp,nγ(T)≥δ
≤ P
Φp,nγ(T)≥δ, AT(M−δ)∩Ω¯αn,T +P
AcT(M−δ) +P
hΩ¯αn,Tic
= P
kYnkγ,p,τδn≥δ, AT(M−δ)∩Ω¯αn,T +P
AcT(M−δ)
+Ph
Ω¯αn,Tic
≤ δ−qE
1A
T(M−δ)∩Ω¯αn,TkYnkqγ,p,τδ n
+P
AcT(M−δ)
+Ph
Ω¯αn,Tic , However, Lemma 2.2 and Lemma 3.1 (the latter will be proved below) yield that
P
AcT(M−δ)
+Ph
Ω¯αn,Tic
→0, asM → ∞, n→ ∞. Therefore, by Lemma A.1 in[7]and (3.6), in order to prove
Φp,nγ(T)→0, in probability, asn→ ∞, (3.7) it suffices to check that there existq≥ pandθ >qα¯(we have setγ=α¯, where ¯αis the exponent presented in Proposition 3.1) such that
(C1) ∀t∈[0,T], lim
n→∞Eh
1A¯Mn(t)kYn(t,·)kqpi
=0;
(C2) ∀s<t∈[0,T], Eh
1A¯Mn(t)kYn(t,·)−Yn(s,·)kqpi
≤C|t−s|1+θ. Here the event ¯AMn(t)is defined by
A¯Mn (t) =AMn(t)∩Ω¯αn,t, t∈[0,T], (3.8) and which satisfies the order relation:
A¯Mn (t)⊂A¯Mn(r), ifr≤t.
Lemma 3.1. Let the event AT(M)be defined by(3.4). Then
M→∞lim P
AT(M)c
=0.
Proof. Note that forp≥4,β∈[p,∞[andβ∈]p, 6p/(6−p)+[(ifd=3), P
AT(M)c
=P
sup
t∈[0,T]kX(t,·)kp≥M
≤M−βE
sup
t∈[0,T]kX(t,·)kβp
. Therefore, it remains to prove
E
sup
t∈[0,T]kX(t,·)kβp
<∞. (3.9)
Define for(t,x)∈ OT,
L1(u)(t,x) := Z t
0
Z
D
Gt−s(x,y)[F+H](u(s,y))W(dy, ds);
L2(u)(t,x) := Z t
0
Z
D
Gt−s(x,y)K(u(s,y))h(s,y)dyds, and setZ=X−L(X)withL=L1+L2. Then
∂Z
∂t + ∆2Z−∆f([Z+L(X)]) =0, Z(0) =ψ,
∂Z/∂n=∂[∆Z]/∂n=0, on∂D.
(3.10)
Further, the Garsia-Rodemich-Rumsey lemma (see, e.g., Theorem B.1.1 and Theorem B.1.5 in[9]) yields that, if for anyq,δ∈]1,∞[and someγ0,γ00∈]0, 1],
(a) sup
(t,x)∈OT
E
|L(u)(t,x)|2qδ
<∞, (b) E
|L(u)(t,x)−L(u)(t0,x0)|2q
≤C
|t−t0|γ00+|x−x0|γ0q
, q>1,
then (3.9) holds. So we only need to prove (a) and (b). Note thatKis bounded andd≤3. Then in light of (A.4),
sup
(t,x)∈OT
Eh
L2(u)(t,x) 2qδ
i
≤ kKk2qδ∞ sup
(t,x)∈OT
Z t
0
Z
D
|Gt−s(x,y)h(s,y)|dyds
2qδ
≤ kKk2qδ∞ khkqδHT(1−d4)qδ<∞.
If sett>t0, then by (A.7)–(A.9) in Lemma A.2, we have forγ0∈[0, 1−d/4[andγ00∈[0, 2∧(4−d)[, E
|L2(u)(t,x)−L2(u)(t0,x0)|2q
≤ C
|t−t0|γ00+|x−x0|γ0q
.
The estimate ofL1is similar to that of L2 (or see[6]). Thus the proof the lemma is complete.
4 Auxiliary lemmas
In this section, we present a sequence of auxiliary lemmas for checking the conditions (C1) and (C2) under (H1)–(H3) given in Section 1 and 2. Throughout Sections 4–6, (H1)–(H3) are assumed to be satisfied.
The following lemma tells us that, to check(C1), it suffices to show(C1)holds withΛn(t,x)instead ofYn(t,x).
Lemma 4.1. Assume p≥4, and q≥ p if d=1, 2, and q∈]p, 6p/(6−p)+[if d =3. Then for each n∈N,
sup
t∈[0,T]
Eh
1A¯Mn(t)kYn(t,·)kqpi
≤C sup
t∈[0,T]
Eh
1A¯Mn(t)kΛn(t,·)kqpi
, (4.1)
whereΛn(t,x)is defined by(3.2).
Proof. Note that for eacht∈[0,T], Eh
kYn(t,·)kqpi
≤CE
X3
i=1
1A¯Mn(t)kΓin(t,·)kqp+1A¯Mn(t)kΛn(t,·)kqp
,
whereΓi(t,x)(i=1, 2, 3) are defined in (3.1). Therefore by (A.16) in Lemma A.3, for 1
r = 2q−2p+ 1∈[0, 1],
E h
1A¯Mn(t)kΓ1n(t,·)kqpi
≤ C Z
D
Eh
1A¯Mn(t)|Γ1n(t,x)|qi dx
≤ CE
Z t
0
Z
D
G2t−s(·,y)1A¯Mn(s)|Yn(s,y)|2dyds
q 2
q 2
≤ CE
Z t
0
(t−s)4rd−d21A¯Mn(s)kYn(s,·)k2pds
q 2
≤ C
Z t
0
(t−s)4rd−d2ds
q−2 2 Z t
0
(t−s)4rd−d2 sup
r∈[0,s]
Eh
1A¯Mn(r)kYn(r,·)kqpi ds
. Similarly, we have
Eh
1A¯Mn(t)kΓ2n(t,·)kqpi
≤ CkhkqHE
Z t
0
Z
D
G2t−s(·,y)1A¯Mn(s)|Yn(s,y)|2dyds
q 2
q 2
≤ C
Z t
0
(t−s)4rd−d2ds
q−
2 2 Z t
0
(t−s)4rd−d2 sup
r∈[0,s]
Eh
1A¯Mn(r)kYn(r,·)kqpi ds
. As forΓ3n(t,x), using (A.13) with 1
r2 =1q− 1ρ+1∈[0, 1], Eh
1A¯Mn(t)kΓ3n(t,·)kqpi
≤ CE
1A¯Mn(t)
Z t
0
Z
D
∆yG(t−s,·,y)[f(Xn(s,·))−f(X(s,·))]dyds
q
q
≤ CE
1A¯Mn(t)
Z t
0
(t−s)4rd2−d+24 kf(Xn(s,·))−f(X(s,·))kρds
q
. (4.2)
Note thatu(s,·),v(s,·)∈Lp(D), for eachs∈[0,T], we have forρ= 3p, ku(s,·)−v(s,·)kρ ≤ Cku(s,·)−v(s,·)kp,
ku2(s,·)−v2(s,·)kρ ≤ Cku(s,·)−v(s,·)kp
ku(s,·)kp+kv(s,·)kp
ku3(s,·)−v3(s,·)kρ ≤ Cku(s,·)−v(s,·)kp
×h
ku(s,·)k2p+kv(s,·)k2p+ku(s,·)kpkv(s,·)kp
i . Hence from (4.2), it follows that
Eh
1A¯Mn(t)kΓ3n(t,·)kqpi
≤ CE
Z t
0
(t−s)4rd2−d+24 1A¯Mn(s)kYn(s,·)kpds
q
≤ C
Z t
0
(t−s)4rd2−d+24 ds
q−1Z t
0
(t−s)4rd2−d+24 sup
r∈[0,s]
Eh
1A¯Mn(r)kYn(r,·)kqpi ds
. Note that the following equivalent relations holds:
d 4r − d
2+1>0 ⇔ 1 q > 1
p −1 6, d
4r2− d+2
4 +1>0 ⇔ 1 q > 1
p +1 2− 2
d.
Then the desired result follows from the Gronwall’s lemma.
Recall theFtn-adapted processX−n(t,x)defined by (3.3). Then we have Lemma 4.2. Let q≥p≥6. Then there exists a constant C:=CM>0such that
sup
(t,x)∈OT
Eh
1A¯Mn(t)|Xn(t,x)−Xn−(t,x)|qi
≤C2−nqι, (4.3)
whereι:= 12(1− d4).
Proof. Recall (2.2) and (3.3), and we get
Xn(t,x)−X−n(t,x) =
4
X
k=1
Tnk(t,x), where
Tn1(t,x) := Z t
tn
Z
D
Gt−s(x,y)F(Xn(s,y))W(dy, ds),
Tn2(t,x) := Z t
tn
Z
D
Gt−s(x,y)H(Xn(s,y))Wn(dy, ds),
Tn3(t,x) := Z t
tn
Z
D
∆yGt−s(x,y)f(Xn(s,y))dyds,
Tn4(t,x) := Z t
tn
Z
D
Gt−s(x,y)Kn(s,y)dyds,
and
Kn(s,y) = K(Xn(s,y))h(s,y)
−(FH˙)(Xn(s,y))αn(s,y)−(HH)(X˙ n(s,y))βn(s,y). From (2.4) and (2.5), it follows that forh∈ Hb,
sup
(s,y)∈OT
|Kn(s,y)| ≤C nd. (4.4)
On the other hand, using (A.4) and the boundedness ofF, E
|Tn1(t,x)|q
= E
Z t
tn
Z
D
Gt−s(x,y)F(Xn(s,y))W(dy, ds)
q
≤ CE
Z t
tn
Z
D
G2t−s(x,y)F2(Xn(s,y))dyds
q/2
≤ C
Z t
tn
Z
D
G2t−s(x,y)dyds
q/2
≤ C
Z t
tn
(t−s)−d4ds
q/2
≤ C2−12nq(1−d4). (4.5)
Further, the Hölder inequality, Lemma 2.1 and the boundedness ofH jointly imply that E
|Tn2(t,x)|q
= E
Z t
tn
Z
D
Gt−s(x,y)H(Xn(s,y))Wn(dy, ds)
q
≤ CE
Z t
tn
Z
D
|Gt−s(x,y)||W˙n(y,s)|dyds
q
≤ CE
Z t
tn
Z
D
|Gt−s(x,y)||W˙n(y,s)|qdyds
·
Z t
tn
Z
D
|Gt−s(x,y)|dyds
q−1
≤ C
Z t
tn
Z
D
|Gt−s(x,y)|dyds
q
sup
(t,x)∈OT
E
|W˙n(x,t)|q
≤ C ndq22nq2|t−tn|q
≤ C n
dq
22−2nq. (4.6)
Note thatf is a polynomial of degree 3. Then by virtue of (A.13) withκ= ∞1 −3p+1=1−3p ∈[0, 1], Eh
1A¯Mn(t)|Tn3(t,x)|qi
= E
1A¯Mn(t)
Z t
tn
Z
D
∆yGt−s(x,y)f(Xn(s,y))dyds
q
≤ E
Z t
tn
(t−s)d4κ−d+24 1A¯Mn(s)kf(Xn(s,·))kp
3ds
q
≤ E
Z t
tn
(t−s)d4(1−3p)−d+24 1A¯Mn(s)
h
kXn(s,·)kp+kXn(s,·)k2p+kXn(s,·)k3pi ds
q
≤ C
Z t
tn
(t−s)−(12+3d4p)ds
q
≤ C2−nq(
1
2−3d4p). (4.7)
Thanks to (4.4), we conclude that E
|Tn4(t,x)|q
= E
Z t
tn
Z
D
Gt−s(x,y)Kn(s,y)dyds
q
≤ C ndq
Z t
tn
Z
D
|Gt−s(x,y)|dyds
q
≤ C ndq|t−tn|q
≤ C ndq2−nq. (4.8)
Thus the estimate (4.3) follows from (4.5)–(4.8).
Remark 4.1. We can easily check that there exists a constant CM >0such that for q≥p≥4, sup
t∈[0,T]
Eh
1A¯Mn(t)kXn(t,·)−Xn−(t,·)kqpi
≤CM2−nqι, (4.9)
whereι= 12(1−d4). In fact, for the proof of(4.9), the only estimate that needs to be checked is that of Tn4. Apply(A.13)withκ= 1p− 3p+1=1−2p ∈[0, 1]to Tn4 and we can get(4.9).
Next we prove a useful lemma, which will be used frequently later. For(t,x),(s,y)∈ OT, set η(t,s,x,y)
(·,∗) := 1[0,t](·)Gt−·(x,∗)−1[0,s](·)Gs−·(y,∗), (4.10)
η(e t,s,x,y)
(·,∗) := 1[0,t](·)∆Gt−·(x,∗)−1[0,s](·)∆Gs−·(y,∗). (4.11) LetV be anF= (Ft)0≤t≤T-predictable process. For 0≤s≤t≤T and x,y ∈D, define
λkn(V)(t,s,x,y):=
Z (k+1)T
2n ∧t
kT 2n∧t
Z
D
η(t,s,x,y)
(r,z)V(r,z)Wn(dz, dr), (4.12)
withk=0, 1, . . . , 2n−1. Then we have