Degenerate parabolic stochastic partial differential equations
Degenerate parabolic stochastic partial differential equations
Degenerate parabolic stochastic partial differential equations
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
4316 M. Hofmanová / Stochastic Processes and their Applications 123 (2013) 4294–4336<br />
Therefore<br />
<br />
1<br />
<br />
2<br />
≤ C E sup<br />
0≤t≤T<br />
∥u ε (t)∥ p L p (T N )<br />
1 +<br />
≤ 1 2 E sup ∥u ε (t)∥<br />
1 p L<br />
0≤t≤T<br />
p (T N ) + C +<br />
T<br />
0<br />
T<br />
0<br />
∥u ε (s)∥ p L p (T N ) ds 1<br />
2<br />
E∥u ε (s)∥ p L p (T N ) ds .<br />
<br />
<br />
T<br />
E sup ∥u ε (t)∥ p L<br />
0≤t≤T<br />
p (T N ) ≤ C 1 + E∥u 0 ∥ p L p (T N ) + E∥u ε (s)∥ p L<br />
0<br />
p (T N ) ds<br />
and the corollary follows from (27).<br />
4.3. Compactness argument<br />
□<br />
To show that there exists u : Ω × T N × [0, T ] → R, a kinetic solution to (1), one needs to<br />
verify the strong convergence of the approximate solutions u ε . This can be done by combining<br />
tightness of their laws with the pathwise uniqueness, which was proved above.<br />
First, we need to prove a better spatial regularity of the approximate solutions. Towards this<br />
end, we introduce two seminorms describing the W λ,1 -regularity of a function u ∈ L 1 (T N ). Let<br />
λ ∈ (0, 1) and define<br />
T N <br />
p λ |u(x) − u(y)|<br />
(u) =<br />
dx dy,<br />
T N |x − y| N+λ<br />
<br />
pϱ λ (u) = sup 1<br />
0 0 such that for all t ∈ [0, T ]<br />
E∥u ε (t)∥ W s,1 (T N ) ≤ C T,s,u 0<br />
<br />
1 + E∥u0 ∥ W ς,1 (T N )<br />
. (31)<br />
Proof. Proof of this statement is based on Proposition 3.2. We have<br />
E<br />
(T N ) 2 <br />
ϱ τ (x − y) f ε (x, t, ξ) f¯<br />
ε (y, t, ξ) dξ dx dy<br />
R