Nonlocal Cahn-Hilliard equations with fractional dynamic boundary conditions

Size: px
Start display at page:

Download "Nonlocal Cahn-Hilliard equations with fractional dynamic boundary conditions"

Transcription

1 Nonlocal Cahn-Hilliard equations with fractional dynamic boundary conditions Ciprian G. Gal To cite this version: Ciprian G. Gal. Nonlocal Cahn-Hilliard equations with fractional dynamic boundary conditions. 39 pages <hal v> HAL Id: hal Submitted on 1 Nov 016 HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents whether they are published or not. The documents may come from teaching and research institutions in France or abroad or from public or private research centers. L archive ouverte pluridisciplinaire HAL est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche publiés ou non émanant des établissements d enseignement et de recherche français ou étrangers des laboratoires publics ou privés.

2 NONLOCAL CAHN-HILLIARD EQUATIONS WITH FRACTIONAL DYNAMIC BOUNDARY CONDITIONS CIPRIAN G. GAL Abstract. We consider a nonlocal version of the Cahn-Hilliard equation characterized by the presence of a fractional diffusion operator and which is subject to fractional dynamic boundary conditions. Our system generalizes the classical system in which the dynamic boundary condition was used to describe any relaxation dynamics of the orderparameter at the walls. The proposed fractional dynamic boundary condition appears to be more general in the sense that it incorporates nonlocal effects which were completely ignored in the classical approach. We aim to deduce well-posedness and regularity results as well as to establish the existence of finite dimensional attractors for this system. Contents 1. Introduction 1. Functional framework and well-posedness 4.1. A backward Euler scheme 4.. The fractional dynamic boundary condition Weak and strong solutions Finite dimensional attractors 7 5. Appendix: the Ising spin exchange model Final remarks 38 References Introduction The classical form of the Cahn-Hilliard equation [6] developed in the later 1950 s has been successfully applied in a variety of phenomena in science see also [9]. In particular it is frequently used as a model for phase separation phenomena in two-component materials which occupy a given domain R d d 3. The standard form of the classical Cahn-Hilliard equation is given by 1.1 t ϕ = µ µ = ϕ + F ϕ in 0 after we have scaled up all the relevant physical constants to one and is usually subjected to no-flux boundary conditions ν µ = 0 on 0 ν ϕ = 0 on 0. In this context one has ±1 corresponding to the pure phases of the material while ϕ 1 1 corresponds to the transition in the interface between the two material phases. A physically relevant choice for F is a singular logarithmic potential which is often approximated by a regular polynomial potential typically the double-well potential F r = r 1. We point out that 1. implies the conservation of mass for 1.1 while 1.3 implies that the diffused interface which separates the two phases of the material intersects the solid wall at a perfect contact angle of π/. This unfortunately is a strong assumption for many materials and binary systems. We refer the reader for further discussion to [7 17] and the references therein. In order to remedy this situation one needs 1991 Mathematics Subject Classification. 35R09 37L30 8C4. Key words and phrases. nonlocal Cahn-Hilliard equation fractional dynamic boundary conditions regional fractional Laplacian fractional Wentzell Laplacian well-posedness regularity global attractor. 1

3 CIPRIAN G. GAL to allow a dynamic contact angle to deviate from the stationary one of π/ and so the contact line can move also relative to the solid boundary of the domain. From this point of view it was proposed in the 1990 s see [5 13] that the dynamics near the solid wall is better described by dynamic boundary conditions of the form 1.4 t ψ ψ + ν ϕ + βψ + G ψ = 0 on 0 where ψ = ϕ in the trace sense whenever the latter exists. Generally β > 0 is a constant denotes the Laplace-Beltrami operator on and the nonlinear function G characterizes the possible preferential attraction or repulsion of one of the components by the walls. Thus in more general binary systems 1.4 must replace the condition 1.3 to account for relaxation dynamics on the boundary. We recall that various other modifications of 1.4 and 1. have also been proposed recently to account for the interaction of the two-phase material with the solid wall. See for instance [7 Remark 1.4] and the corresponding references therein. We note that an important feature of and 1.4 is that any sufficiently smooth solution ϕ ψ satisfies the energy identity d dt E totϕ t ψ t + µ x t dx + t ψ x t ds = 0 for all t 0 where the free bulk plus surface energy is given by 1.5 E tot ϕ ψ = ϕ + F ϕ dx + ψ + βψ + Gψ ds. As far as theoretical results are concerned for the classical Cahn-Hilliard systems and together with 1.4 we now have a rather complete picture on wellposedness regularity and longtime behaviour in terms of global attractors and the global asymptotic stability as time goes to infinity of trajectories associated with these systems. We refer the reader to the survey paper [7] for an extensive list of references. Although the derivation of the Cahn-Hilliard system comprising and 1.4 may appear physically sound it was only introduced as an approximation where mass currents/fluxes in and on were driven by local chemical potential functionals containing derivatives of the density up to second order; this is in agreement with the presence of gradient terms in 1.5. This derivation was performed in [5 13] starting out from a nonlocal microscopic description of the kinetics of phase changes for a lattice gas of interacting particles by ignoring any nonlocal contribution to the total free energy in the continuum limit and by restricting oneself only to the nearest-neighbour interactions. Thus the interaction between any two nearest neighbours was effectively assumed to be short-ranged. Interestingly this was already pointed out by the authors of [5 13]. Note that such approximations of first order in the free energy functional which eventually lead to the classical system were already assumed by Cahn and Hilliard in their influential paper [6] but there the interactions with the boundary walls were also ignored. Besides the interaction was once again assumed to be short ranged. Hence a rigorous derivation was required in order to develop a phase-field model which can account for both short ranged interactions as well as any long range interactions between any two distant particles and without making any kind of approximations. Such rigorous analysis was performed only recently by Giacomin and Lebowitz [1] starting out from a microscopic model for a lattice gas with long-range potentials of the form J x = J x with J C R d. It turns out that in place of the approximation 1.1 one obtains instead the following nonlocal version of the Cahn-Hilliard equation: 1.6 t ϕ = µ µ = Jx y ϕ x ϕy dy + F ϕ in 0. More precisely in this formulation the radially symmetric kernel J encodes both the local and nonlocal information in a manner in which any microscopic particle interaction takes place throughout the domain. It is interesting to point out that for 1.6 the

4 NONLOCAL CHE WITH DYNAMIC BC 3 corresponding free energy takes on the form 1.7 E nonloc ϕ = Jx y ϕx ϕy dxdy + F ϕdx. Rigorous mathematical results for the nonlocal Cahn-Hilliard equation 1.6 together with 1. have been established only recently. We refer the reader to [ ] for wellposedness results to [ ] for regularity and longtime behaviour in terms of global attractors and to [16 6 7] for issues concerning the global asymptotic stability of trajectories as time goes to infinity. However in the general formulation of Giacomin and Lebowitz [1] the interaction between particles inside the bulk domain and particles on was ignored and so any stationary boundary condition for ϕ was somewhat encoded in the properties of the interaction kernel J. This is in particular true for a strongly singular kernel J / L 1 loc R d see [1 18 3] and it covers the case when a regional fractional Laplace operator s s 0 1 appears in the formulation of the chemical potential in 1.6 of course in that case the integral in the second equation of 1.6 must be understood in a principal valued sense. We recall that the use of fractional operators is quite natural for systems that describe long ranged interactions when particles are allowed to take arbitrarily large steps up to the system size with a small finite probability for each such step in addition to any short-ranged interactions between close neighbours. In this sense normal classical diffusion in the Cahn-Hilliard equation 1.1 describes phase-transitions in systems close to equilibrium whereas anomalous diffusion in the presence of fractional Laplace operators in 1.6 seems to be inherent in phase-transitions phenomena that is far from equilibrium. Returning to our first point of this paragraph our goal is to consider the case of phase-separation phenomena that also accounts for the presence of particle interactions with the wall. We follow an idea consistent with the point of view of [13] except that now our dynamic boundary condition incorporates any of the aforementioned nonlocal effects see Appendix and may simply reduce to the classical description 1.4 under certain conditions. To give a precise formulation of the nonlocal value problem we wish to investigate we first recall the definition for the regional fractional Laplace operator as given by for u L s ux uy ux = C ds P.V. dy x y d+s = C ds lim ε 0 provided that the limit exists. Here {y y x >ε} L 1 := {u : R measurable and 1.9 C ds = ss Γ d+s π d Γ1 s ux uy dy x x y d+s ux dx < } 1 + x d+s where Γ denotes the usual Gamma function. We refer to [ 3 33] cf. also Section. below for a class of functions for which the limit in 1.8 exists. Our interest lies with the nonlocal boundary value problem: 1.10 t ϕ = µ µ = s ϕ + F ϕ in 0 subject to the boundary conditions: such that ψ = ϕ and ν µ = 0 on 0 t ψ + l ψ + C sn s ϕ + βψ + G ψ = 0 on ϕ t=0 = ϕ 0 in ψ t=0 = ψ 0 on.

5 4 CIPRIAN G. GAL In 1.1 C s N s s 1/ 1 denotes the so-called fractional normal derivative and l l 0 1 corresponds to the fractional surface diffusion operator on see Section. We observe that at least formally when s = l = 1 the boundary condition 1.1 translates to the classical boundary condition 1.4 and 1.10 is formally the classical local Cahn-Hilliard equation 1.1. A further physical justification for using the simplest model of an interacting particle systems is given in the Appendix see Section 5. It is our goal to give a comprehensive treatment of in terms of existence regularity and stability with respect to the initial data of properly defined solutions. We will also discuss and derive sufficient conditions for the nonlinearities F G so that possesses finite dimensional global and exponential attractors. Our main goal is to relax the usual framework see [7] associated with the classical Cahn-Hilliard equations subject to classical dynamic boundary conditions 1.4 in the sense that we wish to consider the more general case when is a bounded domain with Lipschitz continuous boundary. This is in particular important when the domains of the operators involved in the weak and strong formulations associated with the problem no longer enjoy explicit characterizations in terms of standard Sobolev spaces. Indeed this is not at all expected and in fact impossible in the nonlocal framework that includes fractional diffusion operators like the regional fractional Laplace s and the fractional Laplace-Beltrami boundary operator l. These features make the usual constructive arguments based on the Galerkin approximation scheme for the classical local Cahn-Hilliard equation very difficult to apply as well as place additional regularity constraints on the parameters involved in the nonlocal counter-part. To resolve these issues in the nonlocal framework we develop a time-implicit discretization scheme for an abstract initial value problem in a generic setting that includes all the problems considered in this article and much more. In fact such abstract schemes have already proved their usefulness in the treatment of other nonlocal problems see [15]. We then take advantage of several recent results of Gal and Warma [19 0] suited for nonlocal operators associated with the fractional dynamic boundary condition 1.1 to develop a suitable solution theory for problem Such results will also become crucial in the last part of this contribution where we will show that our problem admits an exponential attractor as well as a finite dimensional global attractor. The article is organized as follows. In Section we define the functional setting needed for our approach: we first consider in subsection.1 an abstract boundary value for two generic self-adjoint operators A B and solve this problem by appealing to a backward Euler scheme and by employing fixed point arguments. Then in subsection. we give a proper definition for the fractional dynamic boundary condition in 1.1 as well as define all the corresponding operators involved in the subsequent analysis. In Section 3 we define what we mean by a weak and strong solution of and then provide and prove appropriate well-posedness results. Section 4 is dedicated to the regularity of weak solutions as well as the existence of finite dimensional attractors for Section 6 contains some additional remarks and some open questions.. Functional framework and well-posedness.1. A backward Euler scheme. Let us first introduce an abstract initial-boundary value problem:.1. subject to the condition t u t + Aµ t = 0 in H A 0 T t U t + BU t + f U t = µ t in H B 0 T.3 U t=0 = U 0 in H B. Let X be a locally compact metric space and m a Radon measure on X such that suppm = X. We let H A = L X m be a given Hilbert space with inner product HA and assume that A 0 is a self-adjoint operator on H A such that 0 is an eigenvalue of A. Similarly let H = L Y m Y for some Radon measure m Y that is supported on a locally

6 NONLOCAL CHE WITH DYNAMIC BC 5 compact metric space Y. Then consider the product Hilbert space H B = H A H endowed with the natural product norm U H B = u H A + u H for all vector-valued functions U = u u H B. In. we further identify µ H A with its pair µ 0 H B. On H B we define a densely defined operator B that is positive and self-adjoint. The matrix-valued function f U = f 1 u f u is assumed to belong to Cb 1 R R. Furthermore we let V A = DA + I 1/ and V B = D B 1/ and endow them with the natural inner products of V A and V B respectively. Finally we also assume that I + A 1 and B 1 are compact operators on H A and H B respectively. Without loss of generality we also assume that the constant function 1 D B = {U V B : BU H B } since this is our case of interest. The application we have mind for the abstract problem.1-.3 in Sections 3 4 is as follows: X = is a bounded domain in R d d 3 with Lipschitz continuous boundary Y = where m is the usual Lebesgue measure on X and m Y is the standard surface measure on. Furthermore A and B are the weak Neumann Laplacian and fractional Wentzell Laplacian respectively cf. Section. below. Definition.1. Let T > 0 be given but otherwise arbitrary. Our notion of strong solution for.1-.3 is as follows: the pair U µ satisfies U L 0 T ; D B L 0 T ; V B t U L 0 T ; H B f U L 0 T ; H B µ L 0 T ; D A such that t u t + Aµ t = 0 in H A -sense t U t + BU t + f U t = µ t in the H B -sense for a.e. t 0 T. The initial condition U t=0 = U 0 is also satisfied in a strong H B -sense. The proof of the following result is based on a time-discretization scheme and compactness arguments. Theorem.. Let f 1 f be globally Lipschitz functions over R and assume there exist c f C f > 0 such that f U U c f U C f with c f C B where U H B C B U V B. Then for every U 0 V B the abstract nonlinear problem.1-.3 possesses at least one strong solution in the sense of Definition.1. Remark.1. Note that. is a system of equations for the vector-valued function U = u u. Theorem. can be also modified for the case of a single equation when one merely identifies U with u and lets H A = H B. Remark.. We also observe that as a consequence of Definition.1 we have U C w [0 T ] ; V B where the latter space stands for the space of weakly continuous functions from [0 T ] with values in V B. We point out that in fact more can be said about the solution of Definition.1. Indeed U C [0 T ] ; V B owing to the fact that by. we have d.4 dt U t V B = BU H B f U BU HB + µ BU HB which follows by testing. with BU H B. Since all the terms on the right-hand side of.4 belong to L 1 0 T we find the absolute continuity of U t V B on [0 T ]. This together with weak continuity gives the desired strong continuity. Theorem. will become crucial in the subsequent Section 3 in the existence of solutions for when we shall specialize the operators A B. Our goal in this subsection is to provide a complete proof of Theorem.. We shall employ a technique based on semi-discretization in time which can be used to construct explicit approximate solutions that will converge to a solution of the original problem under natural conditions on the physical parameters. We will divide this programme into several key steps. We first construct our timedifference scheme. To this end let n be an integer and P be a partition of the time interval [0 T ] such that P = {t 0 t 1... t n } with 0 = t 0 < t 1 <... < t m <... < t n = T where t m = mh for 0 m n and h = T/n is the step size. In applications one may also

7 6 CIPRIAN G. GAL choose variable step sizes for every subinterval t m 1 t m but in our case a fixed step size suffices for the sake of simplicity. Then for 1 m n we consider the time difference equations associated with.1-.3:.5.6 such that δ t u m = Aµ m in H A δ t U m + BU m + f U m = µ m in H B.7 U 0 = U 0 V B. Here we have set δ t u m = um u m 1 δ t U m = δ t u m δ t u m h δ t u m is defined in a similar way and we think of U m and µ m as approximations of the functions U x t m and µ x t m respectively. By analogy with Definition.1 we say that.5-.7 has a strong solution in the following sense. Definition.3. We say U m µ m is a strong solution if U m D B µ m D A for every 1 m n the initial condition U 0 = U 0 is satisfied in the strong H B -sense and the relations in.5-.6 are satisfied in the strong sense of H A and H B respectively. Step 1. Our first goal is to prove the existence of at least one solution U m µ m to the system.5-.7 for a fixed m and assuming that U m 1 is already known. To show this claim we apply the Leray-Schauder degree theory to prove the existence of a solution to the following nonlinear system:.8.9 u + haµ = g in H A hbu + U + hf U = hµ + l in H B where we have set U µ = U m µ m and g = u m 1 l = U m 1. We wish to formulate problem.8-.9 in terms of a fixed point for a nonlinear map S 1 U µ = U µ where S λ is a compact homotopy depending on the parameter λ [0 1] defined as follows: S λ : [0 1] V B V A V B V A such that S λ V w = U µ is a solution of v + haµ = λg in H A hbu + V = hw + λl hλf V in H B. In this step we may also assume without loss of generality that A = A is positive. Indeed if A 0 we could replace equation.8 by u + h A + I µ = g + hµ and then.10 by v + h A + I µ = λg + hw. Next we observe that if g H A and v H A then µ D A is the unique solution of.10. Moreover if l H B and H A w = w 0 H B and V H B which yields that f V H B owing to f Cb 1 then.11 has a unique solution U D B. Here we have endowed both D A and D B with their corresponding graph norms. Since the injections D A V A and D B V B are continuous the mapping S λ is well-defined. Lemma.4. S is continuous as a mapping from [0 1] V B V A V B V A. Proof. Let λ n λ [0 1] and V n w n V w strongly in V B V A. Write Un λn µ λn n = S λ n V n w n Un λ µ λ n = S λ Vn w n and U λ µ λ = S λ V w respectively. From it follows that Un λn Un λ µ λn n µ λ n satisfies the system.1.13 ha hb µ λn n Un λn Un λ µ λ n = λ n λ g in H A = λ n λ l h λ n λ f V n in H B. Taking the corresponding inner products of in H A H B with µ λn n µ λ n and Un λn Un λ respectively we easily deduce the estimates: µ λn µ λ.14 n C λ n λ g HA VA n

8 .15 U λn n NONLOCAL CHE WITH DYNAMIC BC 7 Un λ C λ n λ VB l HB + V n HB for some constant C > 0 independent of n λ n λ. Note that since g H A and l H B and {V n } H B is bounded we have as n that µ λn n µ λ n 0 strongly in V A as well as Un λn Un λ 0 strongly in V B. On the other hand from we see that U λ n U λ µ λ n µ λ satisfies.16 ha µ λ n µ λ = v n v in H A.17 hb Un λ U λ + V n V = h w n w hλ f V n f V in H B. As before taking the inner products of in H A H B with µ λ n µ λ and Un λ U λ respectively we immediately get.18 µ λ n µ λ VA C v n v HA C V n V HB and.19 Un λ U λ VB C V n V HB + w n w HA owing to the fact that f Cb 1 for some constant C > 0 independent of n. Thus if the right-hand sides of go to zero as n then µ λ n µ λ 0 strongly in V A - sense and Un λ U λ 0 strongly in V B. This completes the proof in view of and Lemma.5. S λ is also compact. Proof. Let {λ n V n w n } be any bounded sequence in [0 1] V B V A and consider U n µ n = S λ n V n w n as a solution of Exploiting the same arguments as in and we easily see that {U n µ n } is bounded in V B V A uniformly with respect to n N. Then taking the inner products of the equations for U n µ n with Aµ n and BU n in H A and H B respectively and using the aforementioned fact we also arrive at the fact that {U n µ n } is bounded in D B D A uniformly in n owing to f Cb 1. Since I + A 1 and B 1 are compact self-adjoint operators the injections D B V B D A V A are dense and compact. This yields in particular the compactness of S λ in V B V A. Lemma.6. The Leray Schauder degree for the mapping S 1 is for some open ball B 0 M = D Id S 1 B 0 M 0 = 1 { } U µ V B V A : U µ VB V A < M for some M > 0. In particular problem.8-.9 has at least one strong solution. Proof. For a definition of the Leray-Schauder degree please refer to [9 Chapter Section 8.3]. In order to prove the claim it suffices to show that any possible fixed point of S λ can be estimated independently of λ [0 1]. More precisely we claim that if U µ V B V A is such that S λ U µ = U µ for some λ [0 1] then there exists a constant M > 0 such that.0 U µ VB V A < M. First let us consider the linear system of equations given by u + haµ = 0 in H A and hbu + U = hµ in H B. Observing that S 0 U µ = U µ is the unique solution of such a system our first requirement is that M > 0 is large enough such that the ball B 0 M contains this unique solution. On the other hand we recall that any fixed point U µ of solves the system.1. u + haµ = λg in H A hbu + U = hµ + λl hλf U in H B.

9 8 CIPRIAN G. GAL Let us now take the inner product in H A of.1 with hµ the inner product in H B of. with U and then add the resulting relations. We deduce h U V B + U H B + h A 1/.3 µ H A = λh g µ HA + λ l U HB λh f U U HB. Since by assumption f U U c f U C f we obtain from.3 for every λ [0 1] that CB h c f U H B + h U V B + U H B + h A 1/ µ H A µ HA h g HA µ 1HA + g 1 HA µ 1 HA + ε U + C HB ε l HB C εh g H A + εh A 1/ µ µ + C g 1HA HA H A + ε U + C HB ε l HB for any ε > 0 for some constants C C ε > 0 independent of λ [0 1]. Therefore choosing a sufficiently small ε 1/ we derive the estimate:.4 U V B + h A 1/ µ C h g H H A + l µ H B + g 1HA HA. A Next taking the inner product in H B of. with V 1 we find µ 1 HA = h B 1/ U B 1/ 1 + U 1 HB λ l x 1 HB + hλ f U 1 HB. H B By virtue of.4 and the fact that f is globally Lipschitzian on R we have for every λ [0 1].5 µ 1 HA C 1 + U VB + l HB for some C > 0 independent of λ. Estimate.5 together with.4 then yields by token of elementary Young inequalities.6 U V B + µ V A C 1 + g + l HA HB for C = C h A B c f C f > 0 independent of λ. Thus if we take M > 0 large enough such that M > C1 + g H A + l H B.0 ensures that S λ U µ U µ for any U µ B 0 M for all λ [0 1]. Finally the homotopy invariance of the degree D Id S λ B 0 M 0 for λ [0 1] implies that D Id S 1 B 0 M 0 = D Id S 0 B 0 M 0 = 1 which is enough to conclude that problem.8-.9 has at least one solution U µ V B V A. The latter also immediately yields the regularity U µ D B D A owing once again to the facts that g H A l H B and f is globally Lipschitz. Remark.3. It turns out that each solution U µ of.1-.3 is also unique a fact which can be easily checked. We will not need this property here and so we leave the details to be verified by the interested reader. Step. In this step we are interested in deducing a priori estimates for solutions of.5-.7 which are uniform with respect to the step size h. In what follows C > 0 will denote a constant that is independent of h but which only depends on f and the operators A B. Since A = A 0 and 1 NullA we may assume that δ t u m 1 HA = 0; this implies that u m 1 HA = u m 1 1 = u H 0 1 A HA for every 1 m n. Next we take the inner

10 NONLOCAL CHE WITH DYNAMIC BC 9 product in H A of.5 with µ m µ m 1 HA and the inner product in H B of.6 with δ t U m. Adding the relations that we obtain we find δ t U m H B + A 1/ µ m BU m U m U m 1 H A h H B = f U m δ t U m HB. The basic identity x x y = x y + x y gives 1 BU m U m U m 1 h H B = 1 B 1/ U m B 1/ U m 1 + B 1/ U m U m 1 h H B H B while for f Cb 1 we get f U m δ t U m HB ε δ t U m H B + C ε U m H B + 1. Therefore for a sufficiently small ε > 0 from.7 we obtain after we add the relations for m = 1... r r h δ t U m H B + A 1/ µ m r + U r.8 V B + U m U m 1 V B C h H A r U m H B U 0 V B for 1 r n and h 1. We can estimate µ m 1 HA by taking the inner product in H B of.6 with V 1. Indeed we have µ m 1 HA = f U m 1 HB + B 1/ U m B 1/ 1 + δ t U m 1 HB. H B A repeated application of Young s inequality gives r.9 h µ m r 1 HA C h U m V B U 0 V B recalling once again that.8 is satisfied. We can combine the two estimates of.8 and.9. We find r r h δ t U m + µ m + U r.30 HB VA V B + U m U m 1 V B C 1 + U 0 V B + h r U m V B. Now let us recall Gronwall s lemma in discrete form. Proposition.7. Let {y r } and {a r } be two nonnegative sequences and c 0 a positive constant. If y r c 0 + r a my m for r 1 then r.31 y r c 0 exp a m for r 1. Apply now.31 to.30 to conclude that U r V B e CT 1 + U 0 V B for every 1 r n. Also going back to.30 we also obtain the following estimates that are uniform with respect to h: r r.3 h δ t U m + µ m C HB VA 0 U m U m 1 C V 0 B H B

11 10 CIPRIAN G. GAL.33 sup U r V B C 0 1 r n sup f U r V B C 0 1 r n for some constant C 0 > 0 which depends only on C T > 0 and the initial datum U 0 V B. The last inequality in.33 is immediate on account of the first one and the fact that f is globally Lipschitz. Next we need higher-order estimates for U m µ m. We take the inner product in H A of.5 with Aµ m and use Young and Cauchy-Schwarz inequalities to observe that Aµ m H A C δ t u m H A C δ t U m H B. This estimate together with.3 yields r.34 h µ m DA C 0 for all 1 r n. Analogously we take the inner product in H B of equation.6 with BU m use the fact that f Cb 1 and exploit once more to easily arrive at the estimate r.35 h U m DB C 0 for all 1 r n. Step 3. We can now complete the proof of Theorem.. To this end we need to introduce the corresponding piecewise constant interpolating functions U h µ h and also the corresponding linear interpolate functions Ũh as follows. Let P = {t 0 t 1... t n } be a partition with 0 = t 0 < t 1 <... < t m <... < t n = T where t m = mh for 0 m n and h = T/n is the step size. For the vectors {µ m } n m=0 H A {U m } n m=0 H B = H A H consider the following interpolating functions.36 µ h t = µ m U h t = U m for t [t m 1 t m ] and.37 Ũ h t = U m + t t m U m U m 1 t [t m 1 t m ]. h With these definitions we may now rewrite the entire problem.5-.7 in terms of U h µ h and Ũh as follows: such that t ũ h + Aµ h = 0 in H A 0 T t Ũ h + BU h + fu h = µ h in H B 0 T.40 Ũ h 0 = U 0 in H B. By rewriting the estimates obtained in Step in terms of the interpolations functions U h µ h and Ũh we deduce the following result. Lemma.8. We have the estimates: and U h L 0T ;V B C 0 µ h L 0T ;DA C 0 U h L 0T ;DB C 0 f U h L 0T ;H B C 0 Ũh L 0T ;V B C 0 Ũh L 0T ;DB C 0.44 t Ũ h L 0T ;H B C 0. Proof. The first of.41 and the last of.4 are immediate consequences of.33. To prove the second of.41 we observe n mh µ h L 0T ;DA A = + I µ m dt HA h m 1h n µ m DA C 0

12 NONLOCAL CHE WITH DYNAMIC BC 11 by virtue of.34. In a similar fashion one can deduce the first estimate of.4 owing to.35. Also by token of the first of.3 we have n mh n t Ũ h L 0T ;H B = δ t U m H B dt h δ t U m H B C 0. m 1h By a similar argument using.33 and.35 we obtain the other estimates in.43. We now pass to the limit in as the step size h 0 +. According to the uniform estimates there exist subsequences still labelled by U h µ h Ũ h such that as h 0 + we have the following convergence properties: U h U weakly * in L 0 T ; V B U h U weakly in L 0 T ; D B µ h µ weakly in L 0 T ; D A t Ũ h t U weakly in L 0 T ; H B Ũ h Ũ weakly * in L 0 T ; V B Ũ h Ũ weakly in L 0 T ; D B. These convergence relations and the compact embedding also allow us to conclude L 0 T ; D B H 1 0 T ; H B L 0 T ; V B.51 U h U strongly in L 0 T ; V B L 0 T ; H B. By definition we further observe Ũh U h L 0T ;H B = n = h 3 mh m 1h h t t m δ t U m H B dt n δ t U m H B = h 3 tũh L 0T ;H B. Therefore by.44 we may infer that Ũh U h L 0T ;H B C 0 h. Thus passing to the limit as h 0 and recalling we find that Ũ U in H B a.e. on 0 T. Finally also with help from.51 we can then pass to the limit in the nonlinear term of.39 to find that f U h f U strongly in L 0 T ; H B. Subsequently one may employ once again to pass to the limit in the remaining linear terms of in a standard fashion. Thus we have proved that the limit function U µ yields a strong solution in the sense of Definition.1. The proof of Theorem. is now complete. We can prove further regularity properties for the strong solution of Theorem.. Theorem.9. Let the assumptions of Theorem. be satisfied and assume U 0 D B. Then the strong solution U µ of.1-.3 has the following regularity:.5 U L 0 T ; D B W 1 0 T ; H B µ L 0 T ; D A. Proof. In order to deduce.5 we need to formally differentiate the equations of.1-.3 with respect to t > 0 and then set V := t U w = t µ. In this case V w solves the following system { t v + Aw = 0 in H.53 A 0 T t V + BV + f U V = w in H B 0 T subject to the initial condition.54 V t=0 := V 0 = µ0 BU 0 f U 0. 0

13 1 CIPRIAN G. GAL Here in we have once again identified H A w with w 0 H B and let µ 0 be the solution of the following abstract boundary value problem.55 A + I µ 0 = BU f 1 u 0 in H A where U 0 = u 0 u 0 D B f 1 is globally Lipschitz in R and there is the operator representation BU1 BU = for U D B. BU Note that for U 0 D B we have µ 0 D A H A by.55 and so in.54 V 0 H B. Moreover the matrix-valued function f U = f 1 u f u belongs to C b R R by assumption. A bound for V L 0 T ; H B is equivalent to a bound for U W 1 0 T ; H B in.5. In order to derive this bound rigorously we shall instead work with a timediscretization of.53 by connecting the problem to the one employed earlier in To this end for the same approximations U m µ m of problem.5-.6 we further define for 1 m n V m = U m U m 1 /h and w m = µ m µ m 1 /h and then set δ t V m := V m V m 1 i.e. δ t V m = δ t v m δ t v m. h For 1 m n consider the system of time-difference equations such that δ t v m = Aw m in H A δ t V m + BV m + f U m V m = w m in H B.58 V 0 = V 0 H B. Recall first that u m 1 HA = u m 1 1 = u H 0 1 A HA for every 1 m n; this yields v m 1 HA = 0. We are interested in obtaining the required uniform bound for V m L 0 T ; H B with respect to the step size h. Thus we take the inner product in H A of.56 with A 1 v m and the inner product in H B of.57 with V m. Adding the relations that we obtain we find.59 1 h + 1 h V m HB V m 1 HB + V m V m 1 HB v m VA v m 1 + v m v m 1 VA VA + BV m V m HB = f U m V m V m H B owing to the identity x x y = x y + x y. Since f Cb 1 we derive further from.59 that.60 V m H B V m 1 H B + V m V m 1 H B + h V m V B C f h V m H B. Now sum.60 over all m = 1... r; it follows that r V r H B + V m V m h V m H B V B C h r V m H B + V 0. H B In view of the Gronwall s inequality in discrete form see Proposition.7 we infer that.6 max 1 r n V r H B C 0

14 NONLOCAL CHE WITH DYNAMIC BC 13 for some constant C 0 > 0 which depends only on C T > 0 and the initial datum V 0 H B. Now returning to.61 by virtue of.6 we also derive r.63 V m V m 1 + h V m H B V B C 0 uniformly with respect to h. As usual the a posteriori analysis of the time-discretization relies on an appropriate system of equations. For a partition P = {t 0... t n } and interpolating functions of it suffices then to define new functions Ṽh t w h t which are continuous affine on each interval [t m 1 t m ] and equal to V m = U m U m 1 /h and w m = µ m µ m 1 /h respectively at t = t m also Ṽh 0 = V 0 and µ h 0 = µ 0. In particular we can define new functions V h t = t Ũ h t w h t = t µ h t where µ h t = µ m + t t m µ m µ m 1 t [t m 1 t m ] h as well as Ṽ h t = V m + t t m V m V m 1 t [t m 1 t m ]. h With these definitions we can rewrite in terms of V h w h and Ṽh as follows: such that t ṽ h + Aw h = 0 in H A 0 T t Ṽ h + BV h + f U h V h = w h in H B 0 T.66 Ṽ h 0 = V 0 in H B. Exactly as in Lemma.8 by rewriting the estimates obtained in in terms of the function V h we deduce the following estimate V h L 0T ;H B L 0T ;V B C 0. Passing to a suitable subsequence as h 0 + we can then infer t Ũ h t U weakly-* in L 0 T ; H B t Ũ h t U weakly in L 0 T ; V B. Recalling now that f U h L 0 T ; H B see.4 we can use.39 to write t Ũ h 1 + B 1/ U h B 1/ 1 + fu h 1 HB = µ h 1 HA H B H B and observe that µ h 1 HA L 0 T on account of.67 and.45. Moreover by note that µ h solves the following abstract boundary value problem.69 µ h + Aµ h = BU h 1 + f 1 u h in H A 0 T. Also by virtue of.38 if we set µ h := µ h µ h 1 HA / 1 H A we have µ h µ h HA = A 1 t ũ h µ h owing to the fact that t ũ h 1 HA = 0. Therefore µ h L 0 T ; H A provided that t ũ h L 0 T ; H A L 0 T ; VA which actually is satisfied by.67. It follows that the bound µ h L 0 T ; H A is also uniform with respect to the step size h. We can exploit once again.57 and.67 to deduce by comparison in.39 that BU h L 0 T ; H B uniformly in h > 0. Owing to the global Lipschitz continuity of f 1 from.69 we also get the uniform bound µ h L 0 T ; D A. These uniform estimates entail that as we pass along a subsequence as h 0 we additionally have the convergence properties: { Uh U weakly-* in L 0 T ; D B µ h µ weakly-* in L 0 T ; D A. Thus the final claim in the statement of the theorem easily follows. H A

15 14 CIPRIAN G. GAL Remark.4. As in Remark. one can show by similar arguments that any strong solution satisfying the assumptions of Theorem.9 also satisfies U C [0 T ] ; D B and t U C [0 T ] ; H B... The fractional dynamic boundary condition. We need to introduce some notation and preliminary results related to some auxiliary problems. We denote by L p and L p the norms on Lp and L p respectively. In the case p = and stand for the usual scalar products in L and L respectively. The norms on H s and H s are indicated by H s and H s respectively for any s R. Let now R d d 3 be a bounded domain with Lipschitz continuous boundary i.e. is of class C 01. For s 0 1 we denote by H s := {ϕ L : the fractional order Sobolev space endowed with the norm Cds ϕx ϕy u H s = x y d+s dxdy + ϕx ϕy x y d+s dxdy < } 1/ β ϕ ds having defined C ds as in 1.9 for some β > 0. It is well-known that for 0 < s 1/ the spaces H s and H s 0 = D H s coincide with equivalent norms see [ Corollary.8 and Remark.3]. We also let H s = H s H s C c and note that Hs contains H0 s as a closed subspace. By definition Hs 0 is also the smallest closed subspace of H s containing D. In order to talk about traces for H0 s in the case when s 0 1/] we need to exploit a different potential-theoretic description of H0 s using the notion of relative capacity see e.g. [33 Definition 3.]; cf. also []. Without going into too much detail let us mention that the most common property of the relative capacity is that it measures small sets more precisely than the usual Lebesgue measure on. Thus the potential description gives H0 s = {u H s ũ = 0 r.q.e. on } where ũ denotes the relative quasi-continuous version of u see e.g. [33 Theorem 4.5]. Here the property ũ = 0 r.q.e. on means that there exists a polar set P i.e whose relative capacity is zero such that ũ = 0 almost everywhere on \P see [33 Definition 3.3]. Also every u H s admits a unique up to a relatively polar set P relatively quasi-continuous version ũ : R such that ũ = u a.e. in see for instance [33 Theorem 3.7]. Therefore every ϕ H0 s is zero S-a.e. on i.e. more precisely its relative quasi-continuous version ϕ = 0 on up to a relatively polar set where S denotes the usual Lebesgue surface measure on. Thus we can more than formally say that the trace for ϕ H0 s is actually null when 0 < s 1/ so that there are no nontrivial boundary conditions for ψ in that case. Therefore from this point on we shall simply assume that 1/ < s < 1 for ϕ to possess a nontrivial trace ψ = ϕ but which is well-defined in the usual sense. Indeed when 1/ < s < 1 the trace ψ = ϕ is well defined as a bounded operator from H s L q for every ϕ H s C where q = d 1 / d s if d and s 1/ 1 and q = if d = 1 and s 1/ 1. Furthermore by standard Sobolev function theory the injection H s into L d/d s is continuous if d and s 1/ 1 and it is continuous in C when d = 1 and s 1/ 1. Clearly the continuous injection H s L is also compact see [30]. Next for l 0 1 we let H l := {ψ L ψx ψy : x y d 1+l ds x ds y < } be the fractional order Sobolev space endowed with the norm β ψ H l := ψ ds + C d 1l ψx ψy 1/ x y d 1+l ds x ds y. Similarly we have H l L p is continuous with p := d 1 d 1 l > when d > 1+l H l L is continuous provided that d < 1 + l and H l L r for any

16 NONLOCAL CHE WITH DYNAMIC BC 15 r [ when d = 1 + l. Clearly the continuous injection H l L is also compact. To this end for 1/ < s < 1 and l 0 1 we define another scale of Hilbert spaces V sl := {U = ϕ ψ : ϕ H s ψ = ϕ H l } endowed with norm U V sl = ϕ H s + ψ H l. We shall now recall an integration by parts formula that was derived in [ Theorem 3.3] for the regional fractional Laplacian 1.8 in smooth domains. For this consider the following constant C 1s t 1 1 s t 1 1 s C s := ss 1 0 t s dt where t 1 = max {t 1}. Theorem.10. Let R N be a bounded domain of class C 11 and recall that 1/ < s < 1. Define the set C s := {u : ux = fxρx s 1 + gx x for some f g C } where ρx :=distx x. For u Cs and z we define the fractional normal derivative N s u of the function u by.70 N s duz νzt uz = lim t s for z. t 0 dt Then for every u Cs and v Hs one has s u L N s u L and v s udx = C ds vx vyux uy.71 x y d+s dxdy C s vn s uds. One can also introduce a weak version for the fractional normal derivative on nonsmooth domains that are only of class C 01. Definition.11. Recall that 1/ < s < 1. i Let u W s. We say that s u L if there exists w L such that C ds vx vyux uy x y N+s dxdy = wvdx for all v Cc and so for all v H0 s by density. In that case one writes s u = w. ii Let u H s such that s u L. We say that u has a weak fractional normal derivative in L if there exists g L such that.7 v s udx = C ds vx vyux uy x y d+s dxdy gvds for all v H s. In this case g is uniquely determined by.7; we may simply write g = C s N s u and refer to g as the fractional normal derivative of u. iii Consequently the following Green identity holds:.73 v s udx = C ds C s vx vyux uy x y d+s dxdy vn s uds for all v H s whenever u W s s u L and N s u exists as an element of L. Remark.5. If is a bounded domain of class C 11 and if u C s then N s u coincides exactly with function given by definition.70. In that case N s u L and s u L ; see for instance [].

17 16 CIPRIAN G. GAL To give a proper definition of a fractional power for the Laplace-Beltrami operator acting on we let H l = H l denote the dual of the reflexive Hilbert space H l and consider the operator D l : H l H l by the following definition:.74 D l u v := C d 1l ux uyvx vy x y d 1+l ds x ds y where denotes the duality map between H l and H l. It turns out that D l is bounded and the quadratic form associated with the right-hand side of.74 defines a bilinear symmetric continuous form for u v H l. Next we can define a fractional power of the Laplace-Beltrami operator as a principal-valued integral l u = C d 1lP.V. = C d 1l lim ε 0 + ux uy x y d 1+l ds y : x y >ε ux uy x y d 1+l ds y whenever the latter exists and observe that D l u v = l u v for all u v Hl. Viewing L as a subspace of H l and denoting the part of the operator D l in L by D l again we then have the following characterization:.75 DD l = {u H l l u L } D l u = l u. We also observe that C d 1l.76 ux uyvx vy x y d 1+l ds x ds y = D l u x v x ds x for u D D l and v H l. We refer the reader for more details about this operator to [19 Section.4] the proofs of Propositions Finally we can now give the definition for the linear fractional Wentzell dynamic boundary condition associated with condition 1.1. Theorem.1. [19 Theorem.1]Assume that is of class C 01 and let s 1/ 1 and l 0 1. We define a bilinear symmetric form as follows:.77 a W U V = C ds + C d 1l a W : V sl V sl R + ux uyvx vy x y d+s dxdy + ux uyvx vy x y d 1+l ds x ds y βuvds for all U = u u V = v v V sl. Let A sl W be the closed linear self-adjoint operator on L L associated with this form a W. Then the following are satisfied: i The domain DA sl W of Asl W consists of functions U = u u V sl such that s u L D l u L N s u exists as an element of L and In this case A sl W ii Also A sl W C s N s u + D l u + βu L. has the following operator representation: A sl u s W = u C s N s u + D l u + βu u has compact resolvent and the first eigenvalue λsl W1 is positive. Finally the injection DA sl W L L is continuous provided that d = 1 and if d = 3: { s > d sd 1 4 if l d or s > d+1 4 if l s 1.. or l > d 1 4 if s 1 l < sd 1 d

18 NONLOCAL CHE WITH DYNAMIC BC 17 We call A sl W the realization of the regional fractional Laplace operator s with the fractional Wentzell type boundary condition. We conclude this subsection with a regularity result for a nonlinear elliptic boundary value problem associated with the fractional Wentzell operator A sl W. Theorem.13. [0 Section 3]Assume s 1/ 1 d/4 1 and l d 1 /4 1 if d 3. Consider the following nonlocal boundary value problem: { s.78 u + θ u = h 1 in D l u + C s N s u + βu + ϖ u = h on where h 1 h L L. i Assume that θ ϖ C R satisfy the following conditions:.79 θ x x α 0 x α 1 ϖ x x β 0 x β 1 for all x R with x x 0 for some x 0 > 0 and some positive constants on the right-hand side of.79. Then every bounded solution U = u u V sl L L of.78 satisfies the explicit estimate: u L + u L 1 C + h 1 L + h L for some constant C > 0 independent of U h 1 and h. ii Consequently any bounded solution of.78 is also a strong solution i.e. s u L N s u L and D l u L and the following estimate holds: s u dx + D l u + N s u ds Q1 + h 1 L + h L for some positive function Q independent of U h 1 and h. Remark.6. For further clarification we say that U is a bounded solution of.78 if U V sl L L and a W U V + θ u vdx + ϖ u vds = h 1 vdx + h vds for all V V sl. Finally for a domain with Lipschitz continuous boundary we define the weak Neumann Laplacian A N µ = µ with domain given by.80 D A N = { µ H 1 : µ L ν µ = 0 on }. The boundary condition in.80 is understood in the following variational sense:.81 µωdx = µ ω for all ω H 1 whenever µ L and µ H 1. By results in spectral theory it is well known that A = A 0 on H A = L and I + A 1 is compact in L. 3. Weak and strong solutions Now let us give the assumptions on the potential functions F G. H1 F G C R and there exists c F c G > 0 such that F x c F G x c G x R. H There exist c f c g > 0 and p q 1 ] such that F x p c f F x + 1 G x q c g Gx + 1 x R. H3 There exist C i > 0 i = such that F x C 1 x p/p 1 C G x C 3 x q/q 1 C 4 x R for some p q 1 ].

19 18 CIPRIAN G. GAL Remark 3.1. We observe that H1 implies that both potentials are quadratic perturbations of some strictly convex functions. Assumption H is satisfied by potentials of arbitrary polynomial growth of orders p = p/ p 1 and q = q/ q 1 respectively. In particular for the double-well potential F x = x 1 one can take p = 4/3. Let us define the weak energy space { } Y m = U = ϕ ψ V sl : F ϕ L 1 G ψ L 1 ϕ m for some given m > 0 where we have also set ϕ = 1 ϕ x dx. We equip Y m with the following metric d U 1 U = U 1 U V sl + F ϕ 1 F ϕ dx 1/ + G ψ 1 G ψ ds. Our definition of a weak solution for the boundary value problem is as follows. Definition 3.1. Let ϕ 0 ψ 0 Y m and 0 < T < + be given. We say U = ϕ ψ is a weak solution of if ψ t = ϕ t for all 0 < t < T and the functions ϕ ψ and µ satisfy ϕ ψ L 0 T ; Y m t ϕ L 0 T ; H µ L 0 T ; H 1 t ψ L 0 T ; L F ϕ L 0 T ; L 1 G ψ L 0 T ; L 1 F ϕ L 0 T ; L p G ψ L 0 T ; L q. For every V = v ṽ V sl L p L q ω H 1 a.e. t 0 T we have t ϕ t ω + µ t ω = 0 a W U t V + t ψ t ṽ + = µ v. We have ϕ0 = ϕ 0 in and ψ 0 = ψ 0 on. 1/ F ϕ t v + G ψ t ṽ Remark 3.. Setting ω 1 in 3.5 we deduce the conservation of mass ϕt 1 = ϕ 0 1 for all t 0. The initial conditions ϕ 0 = ϕ 0 ψ 0 = ψ 0 also make sense in a strong L -sense since by we have ϕ ψ C[0 T ]; L L. In 3.6 each is understood as the duality map between L p and L p and between L q and L q respectively. Here we recall once again that 1/p+1/p = 1 and 1/q +1/q = 1. We also define what we mean by a strong solution. To this end we define a strong energy space { } Z m = ϕ 0 ψ 0 DA sl W : µ 0 H 1 ϕ m and endow it with the norm 3.7 ϕ 0 ψ 0 Z m = ϕ 0 ψ 0 DA sl W + µ 0 H 1 where µ 0 is computed via the equation µ 0 = s ϕ 0 + F ϕ 0 in. In 3.7 the norm in Z m = DA sl W H1 is meant for the triplet ϕ 0 ψ 0 µ 0 but we kept the notion ϕ 0 ψ 0 Zm instead of ϕ 0 ψ 0 µ 0 Zm in order to simplify the exposition.

20 NONLOCAL CHE WITH DYNAMIC BC 19 Definition 3.. Let 0 < T < + be given. We say U = ϕ ψ is a strong solution of if it is a weak solution in the sense of Definition 3.1 and ϕ ψ and µ satisfy 3.8 ϕ ψ L 0 T ; L L L 0 T ; DA sl W 3.9 t ϕ t ψ L 0 T ; V sl µ L 0 T ; H 1 L 0 T ; D A N. In particular for the strong solution we have t ϕ = µ a.e. in 0 T while the boundary condition ν µ = 0 is understood in the sense of.81 a.e. on 0 T. Finally we have 3.10 µ = s ϕ + F ϕ a.e. in 0 T and 3.11 t ψ + D l ψ + C s N s ϕ + βψ + G ψ = 0 a.e. on 0 T. The first goal of this subsection is to prove the existence of at least one strong energy solution to problem by passing to the limit as ɛ α 0 in a regularized version of the system admitting a sufficiently smooth approximate solution ϕ ɛα ψ ɛα. The existence of such smooth solutions will be ensured by the abstract result of Theorem. which was proven by means of a backward Euler scheme. Then we will derive additional uniform estimates for the solutions ϕ ɛα ψ ɛα as ɛ α 0. Theorem 3.3. Let ϕ 0 ψ 0 Z m for s 1/ 1 d/4 1 and l d 1 /4 1 if d 3. Assume that F G satisfy the assumption H1. Then there exists at least one strong solution U = ϕ ψ L 0 T ; Z m to problem in the sense of Definition 3.. Remark 3.3. In the range provided for s 1/ 1 d/4 1 and l d 1 /4 1 we have by part ii of Theorem.1 that DA sl W L L and therefore for any ϕ ψ L 0 T ; Z m the nonlinearities F ϕ L 0 T and G ψ L 0 T are well-defined. Proof. Step 1. An approximating problem P ɛα associated with can be formulated abstractly in For the sake of notation we let f = F and g = G. We take for ɛ 0 1 smooth functions f ɛ f g ɛ g uniformly on compact intervals of R with the following properties: 3.1 f ɛ j x c fɛ g ɛ j x c gɛ for all x R for j = 0 1 and some positive constants c fɛ c gɛ which may behave like ɛ m as ɛ 0 for some m = m l N. Since F ɛ s = s 0 f ɛζdζ and G ɛ s = s 0 g ɛζdζ we may also assume 3.13 lim sup F ɛ x = F x lim sup G ɛ x = G x for all x R ɛ 0 ɛ 0 such that 3.14 F ɛ x c F and G ɛ x c G for all x R owing to assumption H1. Such constructions can be easily realized in a standard way as follows: one lets F ɛ x = F x for x ɛ 1 F ɛ x = F i ɛ 1 x ɛ 1 i for x > ɛ 1 and F ɛ x = i=1 F i ɛ 1 x + ɛ 1 i for x < ɛ 1. i=1 A similar construction applies to G ɛ see for instance [7] and the references therein. In particular we observe that 3.1 and 3.14 imply that both f ɛ g ɛ are

21 0 CIPRIAN G. GAL globally Lipschitz functions on R and that f ɛ x g ɛ x x x C 1 x + x C for all x x R for some sufficiently large constants 0 < C 1 λ sl W1 1 C > 0. We also preliminarily observe that due to H1 F and G as well as F ɛ G ɛ are merely quadratic perturbations of the convex functions F x = F x + c F x G x = G x + c G x respectively F ɛ = F ɛ + c F x G ɛ = G ɛ + c G x. We insert an artificial viscosity α 0 1 into the problem by adding the viscous term α t ϕ in the definition of the chemical potential in 1.10 i.e. we let µ = s ϕ + α t ϕ + F ϕ. More precisely our approximated problem P ɛα ɛ α 0 1 consists of finding a function ϕ ɛα ψ ɛα which has the regularity stated in such that and t ϕ ɛσ + A N µ ɛα = 0 α t ϕ ɛσ + A sl W t ψ ɛα subject to the initial conditions t ϕ ɛα L 0 T ; L ϕɛα ψ ɛα fɛ ϕɛα + = g ɛ ψɛα 3.17 ϕ ɛα 0 = ϕ 0 ψ ɛα 0 = ψ 0. In we have let B = A sl W µɛα 0 with D B = DAsl W and A = A N with domain D A = D A N. We recall that A = A 0 on H A = L and I + A 1 is compact in L provided that is of class C 01 while B = B > 0 on H B = L L and B 1 is compact by Theorem.1. We also prepare a fixed initial datum µ ɛα 0 L that solves 3.18 µ ɛα 0 + αa N µ ɛα 0 = µ 0 := s ϕ 0 + F ϕ 0 in L whenever it is assumed that ϕ 0 ψ 0 Z m DA sl W. Observe that the latter also yields that µ ɛα 0 H 1 uniformly with respect to ɛ α > 0. Indeed from 3.18 we immediately see that µ ɛα 0 + α A L N µ ɛα 0 = µ L 0 µ ɛα 0 so that µ ɛα 0 L d and αa N µ ɛα 0 L uniformly in ɛ α provided that µ 0 H 1 the latter is clearly satisfied by assumption cf. the definition of Z m. In particular 3.19 µ ɛα 0 + α A H 1 N µ ɛα 0 C ϕ L 0 ψ 0 Z m for some C > 0 independent of α ɛ. By taking a sufficiently small ɛ 1/M 0 if necessary where ϕ 0 ψ 0 L L M 0 since in that case g ɛ ψ 0 g ψ 0 we also specify a fixed set of data v ɛα 0 L such that 3.0 v ɛα 0 = D l ψ 0 + C s N s ϕ 0 + βψ 0 + g ψ 0. Then it is clear that 3.1 v ɛα 0 L Q ϕ 0 ψ 0 Z m for some function Q > 0 independent of ɛ α. Therefore the abstract results of Theorem. and Theorem.9 yield the existence of at least one smooth solution U ɛα = ϕ ɛα ψ ɛα to problem Pɛα in the sense of Definition.1 by setting V B = V sl and V A = H 1 we also note that the proof of Theorem. was given in the case α = 1 but clearly it can be easily adapted to accommodate the case α > 0. In particular we deduce the existence of a smooth function { Uɛα L 0 T ; DA sl W W 1 0 T ; L L µ ɛα L 0 T ; D A N

Analysis in weighted spaces : preliminary version

Analysis in weighted spaces : preliminary version Analysis in weighted spaces : preliminary version Frank Pacard To cite this version: Frank Pacard. Analysis in weighted spaces : preliminary version. 3rd cycle. Téhéran (Iran, 2006, pp.75.

More information

CAHN-HILLIARD-NAVIER-STOKES SYSTEMS WITH MOVING CONTACT LINES

CAHN-HILLIARD-NAVIER-STOKES SYSTEMS WITH MOVING CONTACT LINES CAHN-HILLIARD-NAVIER-STOKES SYSTEMS WITH MOVING CONTACT LINES C.G. Gal M Grasselli A Miranville To cite this version: C.G. Gal M Grasselli A Miranville. CAHN-HILLIARD-NAVIER-STOKES SYSTEMS WITH MOVING

More information

New estimates for the div-curl-grad operators and elliptic problems with L1-data in the half-space

New estimates for the div-curl-grad operators and elliptic problems with L1-data in the half-space New estimates for the div-curl-grad operators and elliptic problems with L1-data in the half-space Chérif Amrouche, Huy Hoang Nguyen To cite this version: Chérif Amrouche, Huy Hoang Nguyen. New estimates

More information

ON THE UNIQUENESS IN THE 3D NAVIER-STOKES EQUATIONS

ON THE UNIQUENESS IN THE 3D NAVIER-STOKES EQUATIONS ON THE UNIQUENESS IN THE 3D NAVIER-STOKES EQUATIONS Abdelhafid Younsi To cite this version: Abdelhafid Younsi. ON THE UNIQUENESS IN THE 3D NAVIER-STOKES EQUATIONS. 4 pages. 212. HAL Id:

More information

On some weighted fractional porous media equations

On some weighted fractional porous media equations On some weighted fractional porous media equations Gabriele Grillo Politecnico di Milano September 16 th, 2015 Anacapri Joint works with M. Muratori and F. Punzo Gabriele Grillo Weighted Fractional PME

More information

THE CAHN-HILLIARD EQUATION WITH A LOGARITHMIC POTENTIAL AND DYNAMIC BOUNDARY CONDITIONS

THE CAHN-HILLIARD EQUATION WITH A LOGARITHMIC POTENTIAL AND DYNAMIC BOUNDARY CONDITIONS THE CAHN-HILLIARD EQUATION WITH A LOGARITHMIC POTENTIAL AND DYNAMIC BOUNDARY CONDITIONS Alain Miranville Université de Poitiers, France Collaborators : L. Cherfils, G. Gilardi, G.R. Goldstein, G. Schimperna,

More information

Cutwidth and degeneracy of graphs

Cutwidth and degeneracy of graphs Cutwidth and degeneracy of graphs Benoit Kloeckner To cite this version: Benoit Kloeckner. Cutwidth and degeneracy of graphs. IF_PREPUB. 2009. HAL Id: hal-00408210 https://hal.archives-ouvertes.fr/hal-00408210v1

More information

Low frequency resolvent estimates for long range perturbations of the Euclidean Laplacian

Low frequency resolvent estimates for long range perturbations of the Euclidean Laplacian Low frequency resolvent estimates for long range perturbations of the Euclidean Laplacian Jean-Francois Bony, Dietrich Häfner To cite this version: Jean-Francois Bony, Dietrich Häfner. Low frequency resolvent

More information

On the uniform Poincaré inequality

On the uniform Poincaré inequality On the uniform Poincaré inequality Abdesslam oulkhemair, Abdelkrim Chakib To cite this version: Abdesslam oulkhemair, Abdelkrim Chakib. On the uniform Poincaré inequality. Communications in Partial Differential

More information

On Poincare-Wirtinger inequalities in spaces of functions of bounded variation

On Poincare-Wirtinger inequalities in spaces of functions of bounded variation On Poincare-Wirtinger inequalities in spaces of functions of bounded variation Maïtine Bergounioux To cite this version: Maïtine Bergounioux. On Poincare-Wirtinger inequalities in spaces of functions of

More information

Existence of Pulses for Local and Nonlocal Reaction-Diffusion Equations

Existence of Pulses for Local and Nonlocal Reaction-Diffusion Equations Existence of Pulses for Local and Nonlocal Reaction-Diffusion Equations Nathalie Eymard, Vitaly Volpert, Vitali Vougalter To cite this version: Nathalie Eymard, Vitaly Volpert, Vitali Vougalter. Existence

More information

LECTURE 1: SOURCES OF ERRORS MATHEMATICAL TOOLS A PRIORI ERROR ESTIMATES. Sergey Korotov,

LECTURE 1: SOURCES OF ERRORS MATHEMATICAL TOOLS A PRIORI ERROR ESTIMATES. Sergey Korotov, LECTURE 1: SOURCES OF ERRORS MATHEMATICAL TOOLS A PRIORI ERROR ESTIMATES Sergey Korotov, Institute of Mathematics Helsinki University of Technology, Finland Academy of Finland 1 Main Problem in Mathematical

More information

New estimates for the div-curl-grad operators and elliptic problems with L1-data in the whole space and in the half-space

New estimates for the div-curl-grad operators and elliptic problems with L1-data in the whole space and in the half-space New estimates for the div-curl-grad operators and elliptic problems with L1-data in the whole space and in the half-space Chérif Amrouche, Huy Hoang Nguyen To cite this version: Chérif Amrouche, Huy Hoang

More information

Variational Formulations

Variational Formulations Chapter 2 Variational Formulations In this chapter we will derive a variational (or weak) formulation of the elliptic boundary value problem (1.4). We will discuss all fundamental theoretical results that

More information

Finite volume method for nonlinear transmission problems

Finite volume method for nonlinear transmission problems Finite volume method for nonlinear transmission problems Franck Boyer, Florence Hubert To cite this version: Franck Boyer, Florence Hubert. Finite volume method for nonlinear transmission problems. Proceedings

More information

Some tight polynomial-exponential lower bounds for an exponential function

Some tight polynomial-exponential lower bounds for an exponential function Some tight polynomial-exponential lower bounds for an exponential function Christophe Chesneau To cite this version: Christophe Chesneau. Some tight polynomial-exponential lower bounds for an exponential

More information

Phase-field systems with nonlinear coupling and dynamic boundary conditions

Phase-field systems with nonlinear coupling and dynamic boundary conditions 1 / 46 Phase-field systems with nonlinear coupling and dynamic boundary conditions Cecilia Cavaterra Dipartimento di Matematica F. Enriques Università degli Studi di Milano cecilia.cavaterra@unimi.it VIII

More information

THE STOKES SYSTEM R.E. SHOWALTER

THE STOKES SYSTEM R.E. SHOWALTER THE STOKES SYSTEM R.E. SHOWALTER Contents 1. Stokes System 1 Stokes System 2 2. The Weak Solution of the Stokes System 3 3. The Strong Solution 4 4. The Normal Trace 6 5. The Mixed Problem 7 6. The Navier-Stokes

More information

Dissipative Systems Analysis and Control, Theory and Applications: Addendum/Erratum

Dissipative Systems Analysis and Control, Theory and Applications: Addendum/Erratum Dissipative Systems Analysis and Control, Theory and Applications: Addendum/Erratum Bernard Brogliato To cite this version: Bernard Brogliato. Dissipative Systems Analysis and Control, Theory and Applications:

More information

Linear Quadratic Zero-Sum Two-Person Differential Games

Linear Quadratic Zero-Sum Two-Person Differential Games Linear Quadratic Zero-Sum Two-Person Differential Games Pierre Bernhard To cite this version: Pierre Bernhard. Linear Quadratic Zero-Sum Two-Person Differential Games. Encyclopaedia of Systems and Control,

More information

Functional Analysis. Franck Sueur Metric spaces Definitions Completeness Compactness Separability...

Functional Analysis. Franck Sueur Metric spaces Definitions Completeness Compactness Separability... Functional Analysis Franck Sueur 2018-2019 Contents 1 Metric spaces 1 1.1 Definitions........................................ 1 1.2 Completeness...................................... 3 1.3 Compactness......................................

More information

LECTURE # 0 BASIC NOTATIONS AND CONCEPTS IN THE THEORY OF PARTIAL DIFFERENTIAL EQUATIONS (PDES)

LECTURE # 0 BASIC NOTATIONS AND CONCEPTS IN THE THEORY OF PARTIAL DIFFERENTIAL EQUATIONS (PDES) LECTURE # 0 BASIC NOTATIONS AND CONCEPTS IN THE THEORY OF PARTIAL DIFFERENTIAL EQUATIONS (PDES) RAYTCHO LAZAROV 1 Notations and Basic Functional Spaces Scalar function in R d, d 1 will be denoted by u,

More information

SPECTRAL PROPERTIES OF THE LAPLACIAN ON BOUNDED DOMAINS

SPECTRAL PROPERTIES OF THE LAPLACIAN ON BOUNDED DOMAINS SPECTRAL PROPERTIES OF THE LAPLACIAN ON BOUNDED DOMAINS TSOGTGEREL GANTUMUR Abstract. After establishing discrete spectra for a large class of elliptic operators, we present some fundamental spectral properties

More information

Existence result for the coupling problem of two scalar conservation laws with Riemann initial data

Existence result for the coupling problem of two scalar conservation laws with Riemann initial data Existence result for the coupling problem of two scalar conservation laws with Riemann initial data Benjamin Boutin, Christophe Chalons, Pierre-Arnaud Raviart To cite this version: Benjamin Boutin, Christophe

More information

On the longest path in a recursively partitionable graph

On the longest path in a recursively partitionable graph On the longest path in a recursively partitionable graph Julien Bensmail To cite this version: Julien Bensmail. On the longest path in a recursively partitionable graph. 2012. HAL Id:

More information

On Symmetric Norm Inequalities And Hermitian Block-Matrices

On Symmetric Norm Inequalities And Hermitian Block-Matrices On Symmetric Norm Inequalities And Hermitian lock-matrices Antoine Mhanna To cite this version: Antoine Mhanna On Symmetric Norm Inequalities And Hermitian lock-matrices 015 HAL Id: hal-0131860

More information

The oblique derivative problem for general elliptic systems in Lipschitz domains

The oblique derivative problem for general elliptic systems in Lipschitz domains M. MITREA The oblique derivative problem for general elliptic systems in Lipschitz domains Let M be a smooth, oriented, connected, compact, boundaryless manifold of real dimension m, and let T M and T

More information

On path partitions of the divisor graph

On path partitions of the divisor graph On path partitions of the divisor graph Paul Melotti, Eric Saias To cite this version: Paul Melotti, Eric Saias On path partitions of the divisor graph 018 HAL Id: hal-0184801 https://halarchives-ouvertesfr/hal-0184801

More information

Sobolev spaces. May 18

Sobolev spaces. May 18 Sobolev spaces May 18 2015 1 Weak derivatives The purpose of these notes is to give a very basic introduction to Sobolev spaces. More extensive treatments can e.g. be found in the classical references

More information

Confluence Algebras and Acyclicity of the Koszul Complex

Confluence Algebras and Acyclicity of the Koszul Complex Confluence Algebras and Acyclicity of the Koszul Complex Cyrille Chenavier To cite this version: Cyrille Chenavier. Confluence Algebras and Acyclicity of the Koszul Complex. Algebras and Representation

More information

Scientific Computing WS 2018/2019. Lecture 15. Jürgen Fuhrmann Lecture 15 Slide 1

Scientific Computing WS 2018/2019. Lecture 15. Jürgen Fuhrmann Lecture 15 Slide 1 Scientific Computing WS 2018/2019 Lecture 15 Jürgen Fuhrmann juergen.fuhrmann@wias-berlin.de Lecture 15 Slide 1 Lecture 15 Slide 2 Problems with strong formulation Writing the PDE with divergence and gradient

More information

Fractional Cahn-Hilliard equation

Fractional Cahn-Hilliard equation Fractional Cahn-Hilliard equation Goro Akagi Mathematical Institute, Tohoku University Abstract In this note, we review a recent joint work with G. Schimperna and A. Segatti (University of Pavia, Italy)

More information

Thermodynamic form of the equation of motion for perfect fluids of grade n

Thermodynamic form of the equation of motion for perfect fluids of grade n Thermodynamic form of the equation of motion for perfect fluids of grade n Henri Gouin To cite this version: Henri Gouin. Thermodynamic form of the equation of motion for perfect fluids of grade n. Comptes

More information

SELF-ADJOINTNESS OF SCHRÖDINGER-TYPE OPERATORS WITH SINGULAR POTENTIALS ON MANIFOLDS OF BOUNDED GEOMETRY

SELF-ADJOINTNESS OF SCHRÖDINGER-TYPE OPERATORS WITH SINGULAR POTENTIALS ON MANIFOLDS OF BOUNDED GEOMETRY Electronic Journal of Differential Equations, Vol. 2003(2003), No.??, pp. 1 8. ISSN: 1072-6691. URL: http://ejde.math.swt.edu or http://ejde.math.unt.edu ftp ejde.math.swt.edu (login: ftp) SELF-ADJOINTNESS

More information

DYNAMICAL PROPERTIES OF MONOTONE DENDRITE MAPS

DYNAMICAL PROPERTIES OF MONOTONE DENDRITE MAPS DYNAMICAL PROPERTIES OF MONOTONE DENDRITE MAPS Issam Naghmouchi To cite this version: Issam Naghmouchi. DYNAMICAL PROPERTIES OF MONOTONE DENDRITE MAPS. 2010. HAL Id: hal-00593321 https://hal.archives-ouvertes.fr/hal-00593321v2

More information

Norm Inequalities of Positive Semi-Definite Matrices

Norm Inequalities of Positive Semi-Definite Matrices Norm Inequalities of Positive Semi-Definite Matrices Antoine Mhanna To cite this version: Antoine Mhanna Norm Inequalities of Positive Semi-Definite Matrices 15 HAL Id: hal-11844 https://halinriafr/hal-11844v1

More information

HARNACK AND SHMUL YAN PRE-ORDER RELATIONS FOR HILBERT SPACE CONTRACTIONS

HARNACK AND SHMUL YAN PRE-ORDER RELATIONS FOR HILBERT SPACE CONTRACTIONS HARNACK AND SHMUL YAN PRE-ORDER RELATIONS FOR HILBERT SPACE CONTRACTIONS Catalin Badea, Laurian Suciu To cite this version: Catalin Badea, Laurian Suciu. HARNACK AND SHMUL YAN PRE-ORDER RELATIONS FOR HILBERT

More information

Traces, extensions and co-normal derivatives for elliptic systems on Lipschitz domains

Traces, extensions and co-normal derivatives for elliptic systems on Lipschitz domains Traces, extensions and co-normal derivatives for elliptic systems on Lipschitz domains Sergey E. Mikhailov Brunel University West London, Department of Mathematics, Uxbridge, UB8 3PH, UK J. Math. Analysis

More information

GLOBAL LIPSCHITZ CONTINUITY FOR MINIMA OF DEGENERATE PROBLEMS

GLOBAL LIPSCHITZ CONTINUITY FOR MINIMA OF DEGENERATE PROBLEMS GLOBAL LIPSCHITZ CONTINUITY FOR MINIMA OF DEGENERATE PROBLEMS PIERRE BOUSQUET AND LORENZO BRASCO Abstract. We consider the problem of minimizing the Lagrangian [F ( u+f u among functions on R N with given

More information

On semilinear elliptic equations with measure data

On semilinear elliptic equations with measure data On semilinear elliptic equations with measure data Andrzej Rozkosz (joint work with T. Klimsiak) Nicolaus Copernicus University (Toruń, Poland) Controlled Deterministic and Stochastic Systems Iasi, July

More information

The core of voting games: a partition approach

The core of voting games: a partition approach The core of voting games: a partition approach Aymeric Lardon To cite this version: Aymeric Lardon. The core of voting games: a partition approach. International Game Theory Review, World Scientific Publishing,

More information

An introduction to Mathematical Theory of Control

An introduction to Mathematical Theory of Control An introduction to Mathematical Theory of Control Vasile Staicu University of Aveiro UNICA, May 2018 Vasile Staicu (University of Aveiro) An introduction to Mathematical Theory of Control UNICA, May 2018

More information

About partial probabilistic information

About partial probabilistic information About partial probabilistic information Alain Chateauneuf, Caroline Ventura To cite this version: Alain Chateauneuf, Caroline Ventura. About partial probabilistic information. Documents de travail du Centre

More information

A new simple recursive algorithm for finding prime numbers using Rosser s theorem

A new simple recursive algorithm for finding prime numbers using Rosser s theorem A new simple recursive algorithm for finding prime numbers using Rosser s theorem Rédoane Daoudi To cite this version: Rédoane Daoudi. A new simple recursive algorithm for finding prime numbers using Rosser

More information

A proximal approach to the inversion of ill-conditioned matrices

A proximal approach to the inversion of ill-conditioned matrices A proximal approach to the inversion of ill-conditioned matrices Pierre Maréchal, Aude Rondepierre To cite this version: Pierre Maréchal, Aude Rondepierre. A proximal approach to the inversion of ill-conditioned

More information

A Context free language associated with interval maps

A Context free language associated with interval maps A Context free language associated with interval maps M Archana, V Kannan To cite this version: M Archana, V Kannan. A Context free language associated with interval maps. Discrete Mathematics and Theoretical

More information

i=1 α i. Given an m-times continuously

i=1 α i. Given an m-times continuously 1 Fundamentals 1.1 Classification and characteristics Let Ω R d, d N, d 2, be an open set and α = (α 1,, α d ) T N d 0, N 0 := N {0}, a multiindex with α := d i=1 α i. Given an m-times continuously differentiable

More information

ON THE REGULARITY OF SAMPLE PATHS OF SUB-ELLIPTIC DIFFUSIONS ON MANIFOLDS

ON THE REGULARITY OF SAMPLE PATHS OF SUB-ELLIPTIC DIFFUSIONS ON MANIFOLDS Bendikov, A. and Saloff-Coste, L. Osaka J. Math. 4 (5), 677 7 ON THE REGULARITY OF SAMPLE PATHS OF SUB-ELLIPTIC DIFFUSIONS ON MANIFOLDS ALEXANDER BENDIKOV and LAURENT SALOFF-COSTE (Received March 4, 4)

More information

Bodies of constant width in arbitrary dimension

Bodies of constant width in arbitrary dimension Bodies of constant width in arbitrary dimension Thomas Lachand-Robert, Edouard Oudet To cite this version: Thomas Lachand-Robert, Edouard Oudet. Bodies of constant width in arbitrary dimension. Mathematische

More information

Sobolev Spaces. Chapter Hölder spaces

Sobolev Spaces. Chapter Hölder spaces Chapter 2 Sobolev Spaces Sobolev spaces turn out often to be the proper setting in which to apply ideas of functional analysis to get information concerning partial differential equations. Here, we collect

More information

On infinite permutations

On infinite permutations On infinite permutations Dmitri G. Fon-Der-Flaass, Anna E. Frid To cite this version: Dmitri G. Fon-Der-Flaass, Anna E. Frid. On infinite permutations. Stefan Felsner. 2005 European Conference on Combinatorics,

More information

Friedrich symmetric systems

Friedrich symmetric systems viii CHAPTER 8 Friedrich symmetric systems In this chapter, we describe a theory due to Friedrich [13] for positive symmetric systems, which gives the existence and uniqueness of weak solutions of boundary

More information

OPTIMAL POTENTIALS FOR SCHRÖDINGER OPERATORS. 1. Introduction In this paper we consider optimization problems of the form. min F (V ) : V V, (1.

OPTIMAL POTENTIALS FOR SCHRÖDINGER OPERATORS. 1. Introduction In this paper we consider optimization problems of the form. min F (V ) : V V, (1. OPTIMAL POTENTIALS FOR SCHRÖDINGER OPERATORS G. BUTTAZZO, A. GEROLIN, B. RUFFINI, AND B. VELICHKOV Abstract. We consider the Schrödinger operator + V (x) on H 0 (), where is a given domain of R d. Our

More information

b-chromatic number of cacti

b-chromatic number of cacti b-chromatic number of cacti Victor Campos, Claudia Linhares Sales, Frédéric Maffray, Ana Silva To cite this version: Victor Campos, Claudia Linhares Sales, Frédéric Maffray, Ana Silva. b-chromatic number

More information

Quasi-periodic solutions of the 2D Euler equation

Quasi-periodic solutions of the 2D Euler equation Quasi-periodic solutions of the 2D Euler equation Nicolas Crouseilles, Erwan Faou To cite this version: Nicolas Crouseilles, Erwan Faou. Quasi-periodic solutions of the 2D Euler equation. Asymptotic Analysis,

More information

Exact Comparison of Quadratic Irrationals

Exact Comparison of Quadratic Irrationals Exact Comparison of Quadratic Irrationals Phuc Ngo To cite this version: Phuc Ngo. Exact Comparison of Quadratic Irrationals. [Research Report] LIGM. 20. HAL Id: hal-0069762 https://hal.archives-ouvertes.fr/hal-0069762

More information

Hook lengths and shifted parts of partitions

Hook lengths and shifted parts of partitions Hook lengths and shifted parts of partitions Guo-Niu Han To cite this version: Guo-Niu Han Hook lengths and shifted parts of partitions The Ramanujan Journal, 009, 9 p HAL Id: hal-00395690

More information

Replicator Dynamics and Correlated Equilibrium

Replicator Dynamics and Correlated Equilibrium Replicator Dynamics and Correlated Equilibrium Yannick Viossat To cite this version: Yannick Viossat. Replicator Dynamics and Correlated Equilibrium. CECO-208. 2004. HAL Id: hal-00242953

More information

On sl3 KZ equations and W3 null-vector equations

On sl3 KZ equations and W3 null-vector equations On sl3 KZ equations and W3 null-vector equations Sylvain Ribault To cite this version: Sylvain Ribault. On sl3 KZ equations and W3 null-vector equations. Conformal Field Theory, Integrable Models, and

More information

On m-accretive Schrödinger operators in L p -spaces on manifolds of bounded geometry

On m-accretive Schrödinger operators in L p -spaces on manifolds of bounded geometry On m-accretive Schrödinger operators in L p -spaces on manifolds of bounded geometry Ognjen Milatovic Department of Mathematics and Statistics University of North Florida Jacksonville, FL 32224 USA. Abstract

More information

Influence of a Rough Thin Layer on the Potential

Influence of a Rough Thin Layer on the Potential Influence of a Rough Thin Layer on the Potential Ionel Ciuperca, Ronan Perrussel, Clair Poignard To cite this version: Ionel Ciuperca, Ronan Perrussel, Clair Poignard. Influence of a Rough Thin Layer on

More information

SHARP BOUNDARY TRACE INEQUALITIES. 1. Introduction

SHARP BOUNDARY TRACE INEQUALITIES. 1. Introduction SHARP BOUNDARY TRACE INEQUALITIES GILES AUCHMUTY Abstract. This paper describes sharp inequalities for the trace of Sobolev functions on the boundary of a bounded region R N. The inequalities bound (semi-)norms

More information

A note on the acyclic 3-choosability of some planar graphs

A note on the acyclic 3-choosability of some planar graphs A note on the acyclic 3-choosability of some planar graphs Hervé Hocquard, Mickael Montassier, André Raspaud To cite this version: Hervé Hocquard, Mickael Montassier, André Raspaud. A note on the acyclic

More information

Calculus in Gauss Space

Calculus in Gauss Space Calculus in Gauss Space 1. The Gradient Operator The -dimensional Lebesgue space is the measurable space (E (E )) where E =[0 1) or E = R endowed with the Lebesgue measure, and the calculus of functions

More information

GENERATORS WITH INTERIOR DEGENERACY ON SPACES OF L 2 TYPE

GENERATORS WITH INTERIOR DEGENERACY ON SPACES OF L 2 TYPE Electronic Journal of Differential Equations, Vol. 22 (22), No. 89, pp. 3. ISSN: 72-669. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu ftp ejde.math.txstate.edu GENERATORS WITH INTERIOR

More information

2. Function spaces and approximation

2. Function spaces and approximation 2.1 2. Function spaces and approximation 2.1. The space of test functions. Notation and prerequisites are collected in Appendix A. Let Ω be an open subset of R n. The space C0 (Ω), consisting of the C

More information

A note on the computation of the fraction of smallest denominator in between two irreducible fractions

A note on the computation of the fraction of smallest denominator in between two irreducible fractions A note on the computation of the fraction of smallest denominator in between two irreducible fractions Isabelle Sivignon To cite this version: Isabelle Sivignon. A note on the computation of the fraction

More information

EXISTENCE OF SOLUTIONS FOR KIRCHHOFF TYPE EQUATIONS WITH UNBOUNDED POTENTIAL. 1. Introduction In this article, we consider the Kirchhoff type problem

EXISTENCE OF SOLUTIONS FOR KIRCHHOFF TYPE EQUATIONS WITH UNBOUNDED POTENTIAL. 1. Introduction In this article, we consider the Kirchhoff type problem Electronic Journal of Differential Equations, Vol. 207 (207), No. 84, pp. 2. ISSN: 072-669. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu EXISTENCE OF SOLUTIONS FOR KIRCHHOFF TYPE EQUATIONS

More information

A non-linear simulator written in C for orbital spacecraft rendezvous applications.

A non-linear simulator written in C for orbital spacecraft rendezvous applications. A non-linear simulator written in C for orbital spacecraft rendezvous applications. Paulo Ricardo Arantes Gilz To cite this version: Paulo Ricardo Arantes Gilz. A non-linear simulator written in C for

More information

Tropical Graph Signal Processing

Tropical Graph Signal Processing Tropical Graph Signal Processing Vincent Gripon To cite this version: Vincent Gripon. Tropical Graph Signal Processing. 2017. HAL Id: hal-01527695 https://hal.archives-ouvertes.fr/hal-01527695v2

More information

SYMMETRY RESULTS FOR PERTURBED PROBLEMS AND RELATED QUESTIONS. Massimo Grosi Filomena Pacella S. L. Yadava. 1. Introduction

SYMMETRY RESULTS FOR PERTURBED PROBLEMS AND RELATED QUESTIONS. Massimo Grosi Filomena Pacella S. L. Yadava. 1. Introduction Topological Methods in Nonlinear Analysis Journal of the Juliusz Schauder Center Volume 21, 2003, 211 226 SYMMETRY RESULTS FOR PERTURBED PROBLEMS AND RELATED QUESTIONS Massimo Grosi Filomena Pacella S.

More information

CONVERGENCE THEORY. G. ALLAIRE CMAP, Ecole Polytechnique. 1. Maximum principle. 2. Oscillating test function. 3. Two-scale convergence

CONVERGENCE THEORY. G. ALLAIRE CMAP, Ecole Polytechnique. 1. Maximum principle. 2. Oscillating test function. 3. Two-scale convergence 1 CONVERGENCE THEOR G. ALLAIRE CMAP, Ecole Polytechnique 1. Maximum principle 2. Oscillating test function 3. Two-scale convergence 4. Application to homogenization 5. General theory H-convergence) 6.

More information

Near-Potential Games: Geometry and Dynamics

Near-Potential Games: Geometry and Dynamics Near-Potential Games: Geometry and Dynamics Ozan Candogan, Asuman Ozdaglar and Pablo A. Parrilo January 29, 2012 Abstract Potential games are a special class of games for which many adaptive user dynamics

More information

Periodic solutions of differential equations with three variable in vector-valued space

Periodic solutions of differential equations with three variable in vector-valued space Periodic solutions of differential equations with three variable in vector-valued space Bahloul Rachid, Bahaj Mohamed, Sidki Omar To cite this version: Bahloul Rachid, Bahaj Mohamed, Sidki Omar. Periodic

More information

Nel s category theory based differential and integral Calculus, or Did Newton know category theory?

Nel s category theory based differential and integral Calculus, or Did Newton know category theory? Nel s category theory based differential and integral Calculus, or Did Newton know category theory? Elemer Elad Rosinger To cite this version: Elemer Elad Rosinger. Nel s category theory based differential

More information

Lower bound of the covering radius of binary irreducible Goppa codes

Lower bound of the covering radius of binary irreducible Goppa codes Lower bound of the covering radius of binary irreducible Goppa codes Sergey Bezzateev, Natalia Shekhunova To cite this version: Sergey Bezzateev, Natalia Shekhunova. Lower bound of the covering radius

More information

On a series of Ramanujan

On a series of Ramanujan On a series of Ramanujan Olivier Oloa To cite this version: Olivier Oloa. On a series of Ramanujan. Gems in Experimental Mathematics, pp.35-3,, . HAL Id: hal-55866 https://hal.archives-ouvertes.fr/hal-55866

More information

Completeness of the Tree System for Propositional Classical Logic

Completeness of the Tree System for Propositional Classical Logic Completeness of the Tree System for Propositional Classical Logic Shahid Rahman To cite this version: Shahid Rahman. Completeness of the Tree System for Propositional Classical Logic. Licence. France.

More information

On size, radius and minimum degree

On size, radius and minimum degree On size, radius and minimum degree Simon Mukwembi To cite this version: Simon Mukwembi. On size, radius and minimum degree. Discrete Mathematics and Theoretical Computer Science, DMTCS, 2014, Vol. 16 no.

More information

Integral Representation Formula, Boundary Integral Operators and Calderón projection

Integral Representation Formula, Boundary Integral Operators and Calderón projection Integral Representation Formula, Boundary Integral Operators and Calderón projection Seminar BEM on Wave Scattering Franziska Weber ETH Zürich October 22, 2010 Outline Integral Representation Formula Newton

More information

arxiv: v1 [math.oc] 12 Nov 2018

arxiv: v1 [math.oc] 12 Nov 2018 EXTERNAL OPTIMAL CONTROL OF NONLOCAL PDES HARBIR ANTIL, RATNA KHATRI, AND MAHAMADI WARMA arxiv:1811.04515v1 [math.oc] 12 Nov 2018 Abstract. Very recently Warma [35] has shown that for nonlocal PDEs associated

More information

Partial Differential Equations

Partial Differential Equations Part II Partial Differential Equations Year 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2015 Paper 4, Section II 29E Partial Differential Equations 72 (a) Show that the Cauchy problem for u(x,

More information

Approximation of the biharmonic problem using P1 finite elements

Approximation of the biharmonic problem using P1 finite elements Approximation of the biharmonic problem using P1 finite elements Robert Eymard, Raphaèle Herbin, Mohamed Rhoudaf To cite this version: Robert Eymard, Raphaèle Herbin, Mohamed Rhoudaf. Approximation of

More information

Domain Perturbation for Linear and Semi-Linear Boundary Value Problems

Domain Perturbation for Linear and Semi-Linear Boundary Value Problems CHAPTER 1 Domain Perturbation for Linear and Semi-Linear Boundary Value Problems Daniel Daners School of Mathematics and Statistics, The University of Sydney, NSW 2006, Australia E-mail: D.Daners@maths.usyd.edu.au

More information

Contents: 1. Minimization. 2. The theorem of Lions-Stampacchia for variational inequalities. 3. Γ -Convergence. 4. Duality mapping.

Contents: 1. Minimization. 2. The theorem of Lions-Stampacchia for variational inequalities. 3. Γ -Convergence. 4. Duality mapping. Minimization Contents: 1. Minimization. 2. The theorem of Lions-Stampacchia for variational inequalities. 3. Γ -Convergence. 4. Duality mapping. 1 Minimization A Topological Result. Let S be a topological

More information

Neumann and mixed problems on manifolds with boundary and bounded geometry

Neumann and mixed problems on manifolds with boundary and bounded geometry Neumann and mixed problems on manifolds with boundary and bounded geometry Nadine Grosse, Victor Nistor To cite this version: Nadine Grosse, Victor Nistor. Neumann and mixed problems on manifolds with

More information

Fact Sheet Functional Analysis

Fact Sheet Functional Analysis Fact Sheet Functional Analysis Literature: Hackbusch, W.: Theorie und Numerik elliptischer Differentialgleichungen. Teubner, 986. Knabner, P., Angermann, L.: Numerik partieller Differentialgleichungen.

More information

Soundness of the System of Semantic Trees for Classical Logic based on Fitting and Smullyan

Soundness of the System of Semantic Trees for Classical Logic based on Fitting and Smullyan Soundness of the System of Semantic Trees for Classical Logic based on Fitting and Smullyan Shahid Rahman To cite this version: Shahid Rahman. Soundness of the System of Semantic Trees for Classical Logic

More information

The Dirichlet s P rinciple. In this lecture we discuss an alternative formulation of the Dirichlet problem for the Laplace equation:

The Dirichlet s P rinciple. In this lecture we discuss an alternative formulation of the Dirichlet problem for the Laplace equation: Oct. 1 The Dirichlet s P rinciple In this lecture we discuss an alternative formulation of the Dirichlet problem for the Laplace equation: 1. Dirichlet s Principle. u = in, u = g on. ( 1 ) If we multiply

More information

Sound intensity as a function of sound insulation partition

Sound intensity as a function of sound insulation partition Sound intensity as a function of sound insulation partition S. Cvetkovic, R. Prascevic To cite this version: S. Cvetkovic, R. Prascevic. Sound intensity as a function of sound insulation partition. Journal

More information

A metric space X is a non-empty set endowed with a metric ρ (x, y):

A metric space X is a non-empty set endowed with a metric ρ (x, y): Chapter 1 Preliminaries References: Troianiello, G.M., 1987, Elliptic differential equations and obstacle problems, Plenum Press, New York. Friedman, A., 1982, Variational principles and free-boundary

More information

A non-commutative algorithm for multiplying (7 7) matrices using 250 multiplications

A non-commutative algorithm for multiplying (7 7) matrices using 250 multiplications A non-commutative algorithm for multiplying (7 7) matrices using 250 multiplications Alexandre Sedoglavic To cite this version: Alexandre Sedoglavic. A non-commutative algorithm for multiplying (7 7) matrices

More information

Widely Linear Estimation with Complex Data

Widely Linear Estimation with Complex Data Widely Linear Estimation with Complex Data Bernard Picinbono, Pascal Chevalier To cite this version: Bernard Picinbono, Pascal Chevalier. Widely Linear Estimation with Complex Data. IEEE Transactions on

More information

Posterior Covariance vs. Analysis Error Covariance in Data Assimilation

Posterior Covariance vs. Analysis Error Covariance in Data Assimilation Posterior Covariance vs. Analysis Error Covariance in Data Assimilation François-Xavier Le Dimet, Victor Shutyaev, Igor Gejadze To cite this version: François-Xavier Le Dimet, Victor Shutyaev, Igor Gejadze.

More information

Free energy estimates for the two-dimensional Keller-Segel model

Free energy estimates for the two-dimensional Keller-Segel model Free energy estimates for the two-dimensional Keller-Segel model dolbeaul@ceremade.dauphine.fr CEREMADE CNRS & Université Paris-Dauphine in collaboration with A. Blanchet (CERMICS, ENPC & Ceremade) & B.

More information

On the distributional divergence of vector fields vanishing at infinity

On the distributional divergence of vector fields vanishing at infinity Proceedings of the Royal Society of Edinburgh, 141A, 65 76, 2011 On the distributional divergence of vector fields vanishing at infinity Thierry De Pauw Institut de Recherches en Mathématiques et Physique,

More information

Math The Laplacian. 1 Green s Identities, Fundamental Solution

Math The Laplacian. 1 Green s Identities, Fundamental Solution Math. 209 The Laplacian Green s Identities, Fundamental Solution Let be a bounded open set in R n, n 2, with smooth boundary. The fact that the boundary is smooth means that at each point x the external

More information

ASYMPTOTIC BEHAVIOUR OF NONLINEAR ELLIPTIC SYSTEMS ON VARYING DOMAINS

ASYMPTOTIC BEHAVIOUR OF NONLINEAR ELLIPTIC SYSTEMS ON VARYING DOMAINS ASYMPTOTIC BEHAVIOUR OF NONLINEAR ELLIPTIC SYSTEMS ON VARYING DOMAINS Juan CASADO DIAZ ( 1 ) Adriana GARRONI ( 2 ) Abstract We consider a monotone operator of the form Au = div(a(x, Du)), with R N and

More information

Modeling and control of impact in mechanical systems: theory and experimental results

Modeling and control of impact in mechanical systems: theory and experimental results Modeling control of impact in mechanical systems: theory experimental results Antonio Tornambè To cite this version: Antonio Tornambè. Modeling control of impact in mechanical systems: theory experimental

More information