26.12.2013 Views

Degenerate parabolic stochastic partial differential equations

Degenerate parabolic stochastic partial differential equations

Degenerate parabolic stochastic partial differential equations

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

4314 M. Hofmanová / Stochastic Processes and their Applications 123 (2013) 4294–4336<br />

dn ε 1 (x, t, ξ) = σ (x)∇u<br />

ε 2 dδ u ε (x,t)(ξ) dx dt,<br />

dn ε 2 (x, t, ξ) = ε|∇uε | 2 dδ u ε (x,t)(ξ) dx dt,<br />

and define m ε = n ε 1 + nε 2<br />

. Then, using the same approach as in Section 2, one can verify that the<br />

pair ( f ε = 1 u ε >ξ , m ε ) satisfies the kinetic formulation of (25): let ϕ ∈ Cc ∞(TN ×R), t ∈ [0, T ],<br />

then it holds true P-a.s.<br />

<br />

f ε (t), ϕ − f 0 , ϕ t <br />

− f ε (s), b ε (ξ) · ∇ϕ ds<br />

=<br />

−<br />

t<br />

0<br />

t<br />

0<br />

0<br />

<br />

f ε (s), div A(x)∇ϕ ds − ε<br />

<br />

δu ε =ξ Φ ε (u ε ) dW, ϕ + 1 2<br />

t<br />

0<br />

t<br />

0<br />

<br />

f ε (s), ϕ ds<br />

Note, that by taking limit in ε we lose this precise structure of n 2 .<br />

4.2. Energy estimates<br />

<br />

δu ε =ξ G 2 , ∂ ξ ϕ ds − m ε , ∂ ξ ϕ ([0, t)). (26)<br />

In this subsection we shall establish the so-called energy estimate that makes it possible to<br />

find uniform bounds for approximate solutions and that will later on yield a solution by invoking<br />

a compactness argument.<br />

Lemma 4.2. For all ε ∈ (0, 1), for all t ∈ [0, T ] and for all p ∈ [2, ∞), the solution u ε satisfies<br />

the inequality<br />

E∥u ε (t)∥ p L p (T N ) ≤ C 1 + E∥u 0 ∥ p L p (T N )<br />

. (27)<br />

Proof. According to Theorem 4.1, the process u ε is an L p (T N )-valued continuous<br />

semimartingale so we can apply the infinite-dimensional Itô formula [8, Theorem 4.17] for the<br />

function f (v) = ∥v∥ p L p (T N ) . If q is the conjugate exponent to p then f ′ (v) = p|v| p−2 v ∈<br />

L q (T N ) and<br />

f ′′ (v) = p(p − 1)|v| p−2 Id ∈ L L p (T N ), L q (T N ) .<br />

Therefore<br />

t <br />

∥u ε (t)∥ p L p (T N ) = ∥u 0∥ p L p (T N ) − p<br />

|u ε | p−2 u ε div B ε (u ε ) dx ds<br />

0 T<br />

N t <br />

+ p |u ε | p−2 u ε div A(x)∇u ε dx ds<br />

0 T<br />

N t <br />

+ εp |u ε | p−2 u ε u ε dx ds<br />

0 T N<br />

+ p t <br />

|u ε | p−2 u ε g ε<br />

k≥1 0 T N k (x, uε ) dx dβ k (s)<br />

+ 1 t <br />

2 p(p − 1)<br />

0<br />

T N |u ε | p−2 G 2 ε (x, uε ) dx ds. (28)

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!