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

Create successful ePaper yourself

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

derivative theorem to this pair of operators acting on Y to obtain<br />

1−κ<br />

1−<br />

|A α G 1−κ<br />

α y|Y ≤ C|Ay + Gy|Y , for all y ∈ D(A) ∩ D(G), (3.15)<br />

where C > 0 is a constant not depending on y. By definition, D(G 1−κ<br />

α ) = H1−κ p (R; X).<br />

This together <strong>with</strong> (3.15) implies<br />

H α p (R; X) ∩ Lp(R; DA) ↩→ H 1−κ<br />

p (R; D 1−<br />

A 1−κ<br />

α ).<br />

1<br />

− Thus, by the boundedness of DtG α in Lp(R; D 1−<br />

A 1−κ<br />

α ),<br />

1<br />

−<br />

˙v = DtG α G 1−κ<br />

α G κ<br />

α v ∈ H α+κ−1<br />

p<br />

(R; X) ∩ Lp(R; D A 1− 1−κ<br />

To determine the regularity of ˙v(0) we now use the necessity part of the second case.<br />

Let b(t) = t α+κ−2 , t ≥ 0. Then b ∈ K 1 (α + κ − 1, (α + κ − 1)π/2). Since A is R-sectorial,<br />

the operator C := A 1−(1−κ)/α is R-sectorial as well, and<br />

φ R C<br />

≤ (1 − 1−κ<br />

α )φR A<br />

≤ ( α+κ−1<br />

α<br />

)(π − α π<br />

2<br />

) = ( α+κ−1<br />

α<br />

α ).<br />

)π − ( α+κ−1<br />

2 )π,<br />

by Proposition 2.2.1 and the <strong>parabolic</strong>ity condition, combined <strong>with</strong> Remark 2.6.1. So<br />

we see that θb + φR C < π. By the necessity part of Case 2, applied to the equation<br />

˙v(t) + (b ∗ C ˙v)(t) = g(t), t ∈ [0, 1],<br />

<strong>with</strong> some g ∈ Hα+κ−1 p ([0, 1]; X), we get ˙ f(0) = ˙v(0) ∈ DC(1 − 1/p(α + κ − 1), p). It<br />

follows now, by Theorem 2.2.2, that ˙ f(0) lies in<br />

D A 1− 1−κ<br />

α<br />

�<br />

�<br />

1 1 − p(α+κ−1) , p<br />

�<br />

= DA (1 − 1−κ<br />

α )(1 − 1<br />

�<br />

= DA 1 + κ<br />

�<br />

1 1<br />

α − α − pα , p .<br />

p(α+κ−1)<br />

�<br />

), p<br />

Hence condition (iii) is satisfied.<br />

It remains to show (ii). Put J1 := [0, 1] in case J = R+ and J1 := J, otherwise.<br />

Define w1 by means of<br />

w1(t) + (a ∗ Aw1)(t) = t ˙<br />

f(0), t ∈ J1.<br />

Then, due to Theorem 3.1.2, it follows from condition (iii) that u|J1−w1 ∈ Hα+κ p (J1; X)∩<br />

Hκ p (J1; DA). We extend u|J1 − w1 to a function v ∈ ZR+ and put h(t) := v(t) + (a1 ∗<br />

Av)(t), t ∈ R+, where a1 is defined as in Case 2. As above, we see that h ∈ Hα+κ p (R+; X).<br />

By construction, we have ˙v(0) = ˙ h(0) = 0, i.e. h−ψ(·)h(0) ∈ 0H α p (R+; X), where ψ(t) =<br />

(1 + t)e−t , t ≥ 0. In fact, ψ ∈ Hα+κ p (R+), ψ(0) = 1, ψ(t) ˙ = −te−t , t ≥ 0, in particular<br />

˙ψ(0) = 0. Define now the function v1 by means of v1+a1∗Av1 = h−ψh(0), t ∈ R+. This<br />

is possible in view of Theorem 3.1.1, which also gives v1 ∈ 0H α+κ<br />

p (R+; X)∩H κ p (R+; DA).<br />

Consequently, v2 := v−v1 ∈ ZR+ , and v2 solves the equation w+a1∗Aw = ψh(0), t ∈ R+.<br />

If S(·) denotes the resolvent for (3.1), then one verifies that<br />

v2(t) = e −t S(t)h(0) + e −t (1 ∗ S)(t)h(0), t ≥ 0.<br />

42

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

Saved successfully!

Ooh no, something went wrong!