Quasilinear parabolic problems with nonlinear boundary conditions
Quasilinear parabolic problems with nonlinear boundary conditions Quasilinear parabolic problems with nonlinear boundary conditions
Robin type. Unique existence of solutions of these problems in certain spaces of optimal regularity is characterized in terms of regularity and compatibility conditions on the given data. The main result concerning (1.16), Theorem 3.1.4, is proven in Section 3.1. To describe it for the case J = [0, T ], let 1 < p < ∞, κ ∈ [0, 1/p), X be a Banach space of class HT , A an R-sectorial operator in X with R-angle φR A , and a a K-kernel (with angle θa) of order α ∈ (0, 2) such that α + κ /∈ {1/p, 1 + 1/p}. Let further DA denote the domain of A equipped with the graph norm of A. Assume the parabolicity condition θa + φR A < π. Then (1.16) has a unique solution u in the space Hα+κ p (J; X) ∩ Hκ p (J; DA) if and only if the function f satisfies the subsequent conditions: (i) f ∈ H α+κ p (J; X); (ii) f(0) ∈ DA(1 + κ 1 α − pα , p), if α + κ > 1/p; (iii) f(0) ˙ ∈ DA(1 + κ 1 1 α − α − pα , p), if α + κ > 1 + 1/p. Here, DA(γ, p) stands for the real interpolation space (X, DA)γ, p. In the special case a ≡ 1 (i.e. α = 1) and κ = 0, by putting g = ˙ f and u0 = f(0), we recover the main theorem on maximal Lp-regularity for the abstract evolution equation ˙u + Au = g, t ∈ J, u(0) = u0, (1.17) stating that in the above setting, unique solvability of (1.17) in the space H1 p(J; X) ∩ Lp(J; DA) is equivalent to the conditions g ∈ Lp(J; X) and u0 ∈ DA(1 − 1/p, p). We remark that the motivation for considering also the case κ > 0 comes from the problem studied in Chapter 5 which involves two independent kernels. The proof of Theorem 3.1.4 essentially relies on techniques developed in Prüss [64] using the representation of the resolvent S for (1.16) via Laplace transform, as well as on the Mikhlin theorem in the operator-valued version. With the aid of the latter result and an approximation argument, we succeed in showing Lp(R; X)-boundedness of the operator corresponding to the symbol M(ρ) = A((â(iρ)) −1 + A) −1 , ρ ∈ R \ {0}; this operator is closely related to the variation of parameters formula. After proving a rather general embedding theorem in Section 3.2, we continue the study of (1.16), now focusing on the case κ ∈ (1/p, 1 + 1/p), and establish a result corresponding to Theorem 3.1.4. This is done in Section 3.3. In Section 3.4 we collect some known results on maximal Lp-regularity of abstract problems on the halfline. Among others, we consider two abstract second order equations that play a crucial role in the treatment of problems on a strip which are respectively of the form � u − a ∗ ∂ 2 yu + a ∗ Au = f, t ∈ J, y > 0, u(t, 0) = φ(t), t ∈ J, � u − a ∗ ∂ 2 yu + a ∗ Au = f, t ∈ J, y > 0, −∂yu(t, 0) + Du(t, 0) = φ(t), t ∈ J, (1.18) where a is a K-kernel of order α ∈ (0, 2), and A and D are sectorial resp. pseudo-sectorial operators in a Banach space X with D A 1/2 ↩→ DD. The investigation of these problems is pursued in Section 3.5. We prove results characterizing unique solvability of (1.18) in the regularity class H α p (J; Lp(R+; X)) ∩ Lp(J; H 2 p(R+; X)) ∩ Lp(J; Lp(R+; DA)) in terms of regularity and compatibility conditions on the data. Besides the results concerning (1.16) and that from Section 3.4 we make here repeatedly use of the inversion of the convolution, the Dore-Venni theorem, as well as properties of real interpolation. Chapter 4 is devoted to the study of linear scalar problems of second order in the space Lp(J ×Ω), J = [0, T ] and Ω a domain in R n , with general inhomogeneous boundary 8
conditions of order ≤ 1. Sections 4.1 and 4.2 deal with the full resp. half space case. The theory from Chapter 3 is applied to find necessary and sufficient conditions on the data that characterize maximal Lp-regularity of the solutions. The strategy is to look first at problems with constant coefficients and differential operators consisting only of their principle parts, and then to use pointwise multiplication properties of the function spaces involved together with perturbation arguments to extend the results to the general case. In Section 4.3 we study the case of an arbitrary domain Ω ⊂ R n with compact C 2 -smooth boundary Γ decomposing into two disjoint closed parts ΓD and ΓN on which inhomogeneous boundary conditions of zeroth resp. first order have to be satisfied. The basic idea here is to employ the localization method to reduce the problem to related problems on R n and R n +. Proceeding this way we obtain a characterization of unique solvability of the problem with ⎧ ⎨ ⎩ v + k ∗ A(·, x, Dx)v = f, t ∈ J, x ∈ Ω, v = g, t ∈ J, x ∈ ΓD, B(t, x, Dx)v = h, t ∈ J, x ∈ ΓN, A(t, x, Dx) = −a(t, x) : ∇ 2 x + a1(t, x) · ∇x + a0(t, x), t ∈ J, x ∈ Ω, B(t, x, Dx) = b(t, x) · ∇x + b0(t, x), t ∈ J, x ∈ ΓN, (1.19) in the regularity class H α p (J; Lp(Ω)) ∩ Lp(J; H 2 p(Ω)). Here as before, k is a K-kernel of order α ∈ (0, 2). As to the regularity of the top order coefficients, we only assume a ∈ C(J × Ω, R n×n ) and that a has a limit as |x| → ∞ uniformly w.r.t. t ∈ J. In Chapter 5 we are concerned with a linear parabolic problem of second order which appears in the theory of viscoelasticity. In comparison to the problems investigated in Chapter 4, it has two new challenging features: first it is a vector-valued problem, and second it contains two independent kernels. Once more we characterize unique existence of the solution in a certain class of optimal regularity in terms of regularity and compatibility conditions on the given data. Section 5.1 gives a short account of the basic equations of linear viscoelasticity. In Section 5.2 we state the problem and discuss the assumptions on the kernels, which are stronger than in the previous chapters, since the method of proof relies heavily on H ∞ -calculus. Section 5.3 is devoted to the thorough investigation of a half space case of the problem under study, which reads ⎧ ⎪⎨ ⎪⎩ ∂tv − da ∗ (∆xv + ∂2 yv) − (db + 1 ∂tw − da ∗ ∆xw − (db + 4 3da) ∗ ∂2 yw − (db + 1 3da) ∗ (∇x∇x · v + ∂y∇xw) = fv (J × R n+1 + ) 3da) ∗ ∂y∇x · v = fw (J × R n+1 + ) −da ∗ γ∂yv − da ∗ γ∇xw = gv (J × Rn ) −(db − 2 3 da) ∗ γ∇x · v − (db + 4 3 da) ∗ γ∂yw = gw (J × R n ) v|t=0 = v0 (R n+1 + ) w|t=0 = w0 (R n+1 + ), where the unknown functions v and w are R n - resp. scalar-valued, and γ denotes the trace operator at y = 0. To solve this problem, we introduce an appropriate auxiliary function by which the system can be decoupled. Using the results from Chapter 3, the problem is further reduced to an equation on the boundary, which can be solved by means of the joint H ∞ -calculus for the pair (∂t, −∆x) in the space Lp(R+ × R n ). The essential difficulty is the estimate for the principal symbol of the problem. To be precise, 9
- Page 1 and 2: Gutachter: Quasilinear parabolic pr
- Page 3 and 4: Contents 1 Introduction 3 2 Prelimi
- Page 5 and 6: Chapter 1 Introduction The present
- Page 7 and 8: to assume that m, c are bounded fun
- Page 9: We give now an overview of the cont
- Page 13 and 14: Chapter 2 Preliminaries 2.1 Some no
- Page 15 and 16: Clearly, φA ∈ [0, π) and φA
- Page 17 and 18: Definition 2.2.3 Let X and Y be Ban
- Page 19 and 20: µ ∈ Σφα }. Let N ∈ N, Tj
- Page 21 and 22: We remark that a theorem of the Dor
- Page 23 and 24: Here f(A, ·) ∈ H0(Σ π 2 +η; B
- Page 25 and 26: Further, K ∞ (α, θa) := {a ∈
- Page 27 and 28: Using (2.19) for aω and bω yields
- Page 29 and 30: 2.7 Evolutionary integral equations
- Page 31 and 32: Example 2.8.1 For J = [0, T ] and a
- Page 33: We conclude this section by illustr
- Page 36 and 37: kernel a. The operator B is inverti
- Page 38 and 39: x := f(0) ∈ X exists and we are l
- Page 40 and 41: with two positive constants C1, C2
- Page 42 and 43: with two positive constants C1 and
- Page 44 and 45: derivative theorem to this pair of
- Page 46 and 47: 3.2 A general trace theorem Let X b
- Page 48 and 49: 3.3 More time regularity for Volter
- Page 50 and 51: Theorem 3.4.2 Suppose X is a Banach
- Page 52 and 53: Our next objective is to show neces
- Page 54 and 55: Let u1 be the restriction of v1 to
- Page 56 and 57: Proof. We begin with the necessity
- Page 59 and 60: Chapter 4 Linear Problems of Second
<strong>conditions</strong> of order ≤ 1. Sections 4.1 and 4.2 deal <strong>with</strong> the full resp. half space case.<br />
The theory from Chapter 3 is applied to find necessary and sufficient <strong>conditions</strong> on the<br />
data that characterize maximal Lp-regularity of the solutions. The strategy is to look<br />
first at <strong>problems</strong> <strong>with</strong> constant coefficients and differential operators consisting only of<br />
their principle parts, and then to use pointwise multiplication properties of the function<br />
spaces involved together <strong>with</strong> perturbation arguments to extend the results to the general<br />
case. In Section 4.3 we study the case of an arbitrary domain Ω ⊂ R n <strong>with</strong> compact<br />
C 2 -smooth <strong>boundary</strong> Γ decomposing into two disjoint closed parts ΓD and ΓN on which<br />
inhomogeneous <strong>boundary</strong> <strong>conditions</strong> of zeroth resp. first order have to be satisfied. The<br />
basic idea here is to employ the localization method to reduce the problem to related<br />
<strong>problems</strong> on R n and R n +. Proceeding this way we obtain a characterization of unique<br />
solvability of the problem<br />
<strong>with</strong><br />
⎧<br />
⎨<br />
⎩<br />
v + k ∗ A(·, x, Dx)v = f, t ∈ J, x ∈ Ω,<br />
v = g, t ∈ J, x ∈ ΓD,<br />
B(t, x, Dx)v = h, t ∈ J, x ∈ ΓN,<br />
A(t, x, Dx) = −a(t, x) : ∇ 2 x + a1(t, x) · ∇x + a0(t, x), t ∈ J, x ∈ Ω,<br />
B(t, x, Dx) = b(t, x) · ∇x + b0(t, x), t ∈ J, x ∈ ΓN,<br />
(1.19)<br />
in the regularity class H α p (J; Lp(Ω)) ∩ Lp(J; H 2 p(Ω)). Here as before, k is a K-kernel<br />
of order α ∈ (0, 2). As to the regularity of the top order coefficients, we only assume<br />
a ∈ C(J × Ω, R n×n ) and that a has a limit as |x| → ∞ uniformly w.r.t. t ∈ J.<br />
In Chapter 5 we are concerned <strong>with</strong> a linear <strong>parabolic</strong> problem of second order which<br />
appears in the theory of viscoelasticity. In comparison to the <strong>problems</strong> investigated<br />
in Chapter 4, it has two new challenging features: first it is a vector-valued problem,<br />
and second it contains two independent kernels. Once more we characterize unique<br />
existence of the solution in a certain class of optimal regularity in terms of regularity<br />
and compatibility <strong>conditions</strong> on the given data. Section 5.1 gives a short account of the<br />
basic equations of linear viscoelasticity. In Section 5.2 we state the problem and discuss<br />
the assumptions on the kernels, which are stronger than in the previous chapters, since<br />
the method of proof relies heavily on H ∞ -calculus. Section 5.3 is devoted to the thorough<br />
investigation of a half space case of the problem under study, which reads<br />
⎧<br />
⎪⎨<br />
⎪⎩<br />
∂tv − da ∗ (∆xv + ∂2 yv) − (db + 1<br />
∂tw − da ∗ ∆xw − (db + 4<br />
3da) ∗ ∂2 yw − (db + 1<br />
3da) ∗ (∇x∇x · v + ∂y∇xw) = fv (J × R n+1<br />
+ )<br />
3da) ∗ ∂y∇x · v = fw (J × R n+1<br />
+ )<br />
−da ∗ γ∂yv − da ∗ γ∇xw = gv (J × Rn )<br />
−(db − 2<br />
3 da) ∗ γ∇x · v − (db + 4<br />
3 da) ∗ γ∂yw = gw (J × R n )<br />
v|t=0 = v0 (R n+1<br />
+ )<br />
w|t=0 = w0 (R n+1<br />
+ ),<br />
where the unknown functions v and w are R n - resp. scalar-valued, and γ denotes the<br />
trace operator at y = 0. To solve this problem, we introduce an appropriate auxiliary<br />
function by which the system can be decoupled. Using the results from Chapter 3, the<br />
problem is further reduced to an equation on the <strong>boundary</strong>, which can be solved by<br />
means of the joint H ∞ -calculus for the pair (∂t, −∆x) in the space Lp(R+ × R n ). The<br />
essential difficulty is the estimate for the principal symbol of the problem. To be precise,<br />
9