02.12.2012 Views

Quasilinear parabolic problems with nonlinear boundary conditions

Quasilinear parabolic problems with nonlinear boundary conditions

Quasilinear parabolic problems with nonlinear boundary conditions

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Lemma 6.2.3 Let 0 < s < s0 < 1, ρ ∈ (0, ρ0], and w ∈ Xm 1, s0 ∩ C(J; C(Ω)) be a fixed<br />

K-valued function. Let further Σ ′ be as described above. Suppose that b is as in Lemma<br />

6.2.1 <strong>with</strong> r ∈ (s, 1) and bξ ∈ L∞(J × Ω × K, Rm ). Then there exists a constant C > 0<br />

not depending on T and ρ such that<br />

[b(·, ·, f) − b(·, ·, g)]X1, s ≤ Cµ(T )(|f − g|X m 1, s 0<br />

Proof. Let f, g be arbitrary functions in Σ ′ . Put<br />

+ |f − g|∞, m), f, g ∈ Σ ′ .<br />

ω(t, h, x) = b(t + h, x, f(t + h, x))−b(t + h, x, g(t + h, x))−b(t, x, f(t, x))+b(t, x, g(t, x))<br />

for t, t + h ∈ J, and x ∈ Ω. Then<br />

[b(·, ·, f)−b(·, ·, g)]X1, s<br />

�<br />

= (<br />

Ω<br />

� T<br />

0<br />

(<br />

� 1<br />

0<br />

σ −2s �<br />

1<br />

(<br />

|V (t, σ)|<br />

V (t, σ)<br />

Similarly as in the proof of Lemma 6.2.1 we may establish<br />

2 dσ<br />

|ω(t, h, x)| dh)<br />

σ<br />

|ω(t, h, x)| ≤ |bξ|∞, m |f(t + h, x) − f(t, x) − g(t + h, x) + g(t, x)| +<br />

) p<br />

2 dt dx) 1<br />

p .<br />

+ CHL|f − g|∞, m(|f(t + h, x) − f(t, x)| + |g(t + h, x) − g(t, x)| + Cb(x)|h| r )<br />

for all t, t + h ∈ J and a.a. x ∈ Ω. With<br />

Θ := (<br />

we thus obtain<br />

≤ (<br />

� T<br />

(<br />

� 1<br />

0 0<br />

� T � 1<br />

[b(·,·, f) − b(·, ·, g)]X1, s ≤<br />

�<br />

≤ C<br />

�<br />

≤ C<br />

�<br />

≤ Cµ(T )<br />

0<br />

(<br />

0<br />

σ −2s �<br />

1<br />

(<br />

|V (t, σ)|<br />

σ −2s �<br />

1<br />

(<br />

|V (t, σ)|<br />

V (t, σ)<br />

V (t, σ)<br />

|h| r 2 dσ<br />

dh)<br />

σ<br />

σ r 2 dσ<br />

dh)<br />

σ<br />

) p<br />

2 dt) 1<br />

p<br />

) p<br />

2 dt) 1<br />

p = T 1 p<br />

√ 2(r−s) ,<br />

[f − g]X m 1, s + |f − g|∞, m([f]X m 1, s + [g]X m 1, s + |Cb| Lp(Ω)Θ)<br />

[f − g]X m 1, s + |f − g|∞([f − w]X m 1, s + [g − w]X m 1, s + [w]X m 1, s<br />

[f − g]X m + |f − g|∞, 1, s m([f − w]X 0<br />

m + [g − w]X 1, s0 m 1, s0 ≤ Cµ(T )([f − g]X m 1, s 0<br />

+ |f − g|∞, m). �<br />

�<br />

�<br />

+ µ(T ))<br />

�<br />

+ 1)<br />

Lemma 6.2.3 and the trivial inequality |Bf −Bg|X ≤ Cµ(T )|f −g|∞, m yield the estimate<br />

(6.20), which we were aiming at. This completes the proof of the inequality (6.16), since<br />

we have ZT ∇ ↩→ H(1+α)/2 p (J; Lp(Ω)) ∩ C(J × Ω) and (1 + α)/2 > α.<br />

We conclude this paragraph <strong>with</strong> a result on pointwise multiplication which have been<br />

used several times in the previous sections. In Section 4.2.2 we have already seen that<br />

the space Y 0,s1<br />

1 ∩ Y 0,s2<br />

2 forms a multiplication algebra if the embedding Y 0,s1<br />

1 ∩ Y 0,s2<br />

2 ↩→<br />

C(J; C(Ω)) is valid. We will show under the above assumptions on p that this is true<br />

also for Y k1,s1<br />

1 ∩ Y 1,s2<br />

2 <strong>with</strong> k1 = 0, 1. As before, we shall only consider the seminorm<br />

terms of highest order.<br />

109

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

Saved successfully!

Ooh no, something went wrong!