PD09 Abstracts 45
IP1 functions. Free Boundary Regularity Yann Brenier We discuss level surfaces of solutions u to singular semilin- CNRS ear elliptic equations such as Δu = δ(u). The level surface Universit´ede Nice Sophia-Antipolis u = 0 can be interpreted as a solution to a free boundary [email protected] problem, the problem of finding the optimal shape of in- sulating material. A variant of this equation, the Prandtl- Batchelor equation, describes the wake created by a boat. IP4 We will prove that stable free boundaries are smooth in The Importance of PDEs for Modeling and Simu- three-dimensional space. Numerical methods are needed lation on Peta and Exascale Systems to explore counterexamples in higher dimensions and to establish a more robust regularity theory with numerically As scientists and engineers tackle more complex problems effective bounds. involving multiphysics and multiscales and then seek to solve these problems through computation, the partial dif- David Jerison ferential equation formulations which incorporate and more Massachusetts Institute of Technology closely approximate the complex physical or biological phe- [email protected] nomena being modeled will also be important to unlocking the approach taken on Peta and Exascale systems. Scien- tists and engineers working on problems in such areas as IP2 aerospace, biology, climate modeling, energy and other ar- Fluid Biomembranes: Modeling and Computation eas demand ever increasing compute power for their prob- lems. For Petascale and Exascale systems to be useful mas- We study two models for biomembranes. The first one is sive parallelism at the chip level is not sufficient. I will de- purely geometric since the equilibrium shapes are the min- scribe some of the challenges that will need to be considered imizers of the Willmore energy under area and volume con- in designing Petascale and eventually Exascale systems. straints. We present a novel method based on ideas from Through the combination of High Performance Comput- shape differential calculus. The second model incorporates ing (HPC) hardware coupled with novel mathematical and the effect of the inside (bulk) viscous incompressible fluid algorithmic approaches emerging from the original PDE and leads to more physical dynamics. We use a parametric formulations some efforts toward breakthroughs in science approach, which gives rise to fourth order highly nonlinear and engineering are described. While progress is being PDEs on surfaces and involves large domain deformations. made, there remain many challenges for the mathematical We discretize these PDEs in space with an adaptive fi- and computational science community to apply ultra-scale, nite element method (AFEM), with either piecewise linear multi-core systems to Big science problems with impact on or quadratic polynomials, and a semi-implicit time step- society. In conclusion, some discussion not only on the ping scheme. We employ the Taylor-Hood element for the most obvious way to use ultra-scale, multi-core HPC sys- Navier-Stokes equations together with iso-parametric ele- tems will be given but also some thoughts on incorporat- ments, the latter being crucial for the correct approxima- ing more physics in the algorithms derived from the PDE tion of curvature. We discuss several computational tools models which might allow us to better use such systems to such as space-time adaptivity and mesh smoothing. We tackle previously intractable problems. also discuss a method to execute refinement, coarsening, and smoothing of meshes on manifolds with incomplete in- Kirk E. Jordan formation about their geometry and yet preserve position IBM T.J. Watson Research and curvature accuracy. This is a new paradigm in adap- [email protected] tivity. Ricardo Nochetto IP5 Department of Mathematics Singular Limits in Thermodynamics of Viscous Flu- University of Maryland, College Park ids [email protected] We discuss some recent results concerning singular limits of the full Navier-Stokes-Fourier system, where one or several IP3 characteristic numbers become small or tend to infinity. Rearrangement, Convection and Competition The main ingredients of our approach read as follows:
Rearrangement theory is about reorganizing a given func- • a general existence theory of global-in-time weak so- tion (or map) in some specific order (monotonicity, cycle lutions for any finite energy data; monotonicity etc...). This is somewhat similar to the con- vection phenomenon in fluid mechanics, where fluid parcels are continuously reorganized in a stabler way (heavy fluid • uniform bounds independent of the value of the sin- at bottom and light fluid at top). This can also be related gular parameters based on total dissipation balance to some competition models in economy, where agents act (Second law of thermodynamics); according to their rank. In our talk, we make these analo- gies more precise by analysing the Navier-Stokes equa- • analysis of propagation of acoustic waves, in particu- tions with Boussinesq approximation of the buoyancy force lar on unbounded or “large’ spatial domains. and two distinct approximations of this model (Darcy and Boussinesq). We will see how these approximations are related to the concept, well known in optimal transport theory, of rearrangement of maps as gradient of convex Eduard Feireisl Mathematical Institute ASCR, Zitna 25, 115 67 Praha 1 Czech Republic [email protected] 46 PD09 Abstracts
IP6 berger and Rosinger but it is actually based on topological Q-curvature in Conformal Geometry processes known as convergence structures.
In this talk , I will survey some analytic results concerned Roumen Anguelov, Jan Harm Van Der Walt with the top order Q-curvature equation in conformal ge- University of Pretoria ometry. Q-curvature is the natural generalization of the [email protected], [email protected] Gauss curvature to even dimensional manifolds. Its close relation to the Pfaffian, the integrand in the Gauss-Bonnet formula, provides a direct relation between curvature and topology. The notion of Q-curvature arises naturally in CP1 conformal geometry in the context of conformally covari- Existence of a Unique Solution to Equations Mod- ant operators. In 1983, Paneitz gave the first construction eling Fluid Flow with Time-Dependent Density of the fourth order conformally covariant Paneitz opera- Data tor in the context of Lorentzian geometry in dimension four. The ambient metric construction, introduced by Fef- We consider a system of nonlinear equations modeling the ferman and Graham , provides a systematic construction flow of a barotropic, compressible fluid. The equations in general of conformally covariant operators. Each such consist of a hyperbolic equation for the velocity, an alge- operator gives rise to a semi-linear elliptic equation anal- braic equation (the equation of state) for the pressure, and ogous to the Yamabe equations which we shall call the a modified version of the conservation of mass equation. Q-curvature equation. These equations share a number of We prove the existence of a unique solution to the model’s common features. Among these we mention the following: equations with time-dependent density data and spatially- (i) the lack of compactness: the nonlinearity always oc- dependent velocity data, under periodic boundary condi- cur at the critical exponent, for which the Sobolev imbed- tions. ding is not compact; (ii) the lack of maximum principle: Diane Denny for example, it is not known whether the solution of the Texas A&M University - Corpus Christi fourth order Q-curvature equation on manifolds of dimen- [email protected] sions greater than four may touch zero. In spite of these difficulty, there has been significant progress on questions of existence, regularity and classification of entire solutions CP1 for these equations in the recent literature. In the talk, I Localization of Analytic Regularity Criteria on the will give a brief survey of the subject with emphasize on Vorticity and Balance Between the Vorticity Mag- applications to problems in conformal geometry and the nitude and Coherence of the Vorticity Direction in connection between the a class of conformal covariant op- the 3D NSE erators to that of the fractional Laplacian opeators. 2q 2q−3 1 Sun-Yung Alice Chang DaVaiga has shown that Du q ∈ L (0,T) is a regu- Princeton University larity class for the Navier-Stokes Equations for any 3 ≤ [email protected] q<∞. Beale-Kato-Majda proved the regularity for vor- ticity in time-space L1L∞-norm. Geometric conditions on the vorticity direction improve the regularity and the local- IP8 ized versions have been recently obtained by Grujic. The p q Title Not Available at Time of Publication scaling-invariant regularity class of weighted L L -type for vorticity magnitude with coherence factor as a weight will Abstract not available at time of publication. be demonstrated in a localized setting. Felix Otto Rafaela Guberovic University of Bonn Seminar for Applied Mathematics, ETH, Zurich, Germany Switzerland [email protected] University of Virginia, Charlottesville, USA [email protected]
IP9 Zoran Grujic Title Not Available at Time of Publication Department of Mathematics University of Virginia Abstract not available at time of publication. [email protected] George C. Papanicolaou Stanford University CP1 Department of Mathematics [email protected] Fully-Coupled Aeroelastic Computations of High Reynolds Number Flows
CP1 Strong (two-way) coupling of fluid and structure presents interest to vary engineering applications, particularly when Solving Nonlinear Systems of Pdes Via Real Set- the flow is turbulent and sensitive to structural motions. A Valued Maps CFD based algorithm, using large-eddy simulation (LES), The existence and regularity of solutions of nonlinear PDEs is proposed for the numerical investigation of strong aeroe- resulting from constitutive relations of fluids are problems lastic coupling. The aeroelastic response of a cranked dou- with acknowledged difficulty. We prove that the solutions ble delta wing to forced flap motion is subject of investi- of very large classes of systems of nonlinear PDE can be gation. The results show that the flapping frequency in- assimilated with real set-valued maps. Our approach is fluences the wing deformation, as well the flow field in the similar to the order completion method of Oberguggen- PD09 Abstracts 47
wake of the wing. [email protected] Marcel Ilie University of Central Florida CP2 Department of Mechanical, Material and Aerospace Analysis of An Epidemic Model with Age Depen- Engineering dency: The Case of 2009 H1N1 Influenza Outbreak [email protected] It is well known that epidemic models concerned with the infection-age are represented by PDE. In resent years, a CP1 few researchers have investigated models take account of Mixed Bottom-Friction-Kelvin-Helmholtz Destabi- not only infection-age but chronological age, and this as- lization of Source-Driven Abyssal Overflows in the sumption seems to appropriate for the case of 2009 H1N1 Ocean influenza outbreak which more likely to affect young peo- ple. Our purpose of this presentation is to develop an age- Source-driven ocean currents that flow over topographic structured model of this epidemic and to study the math- sills are important initiation sites for the abyssal compo- ematical character. nent of the thermohaline circulation. These overflows ex- hibit vigorous space and time variability over many scales Toshikazu Kuniya as they progress from a predominately gravity-driven down Waseda university slope flow to a geostrophic along slope current. Observa- Graduate school of fundamental science and engineering tions show that in the immediate vicinity of a sill, grounded [email protected] abyssal ocean overflows can possess current speeds greater than the local long internal gravity wave speed with bottom friction and down slope gravitational acceleration dominat- CP2 ing the flow evolution. It is shown that these dynamics lead Mathematical Analysis of a Free Boundary Com- to the mixed frictionally-induced and Kelvin-Helmholtz in- bustion Model stability of grounded abyssal overflows. Within the over- flow, the linearized instabilities correspond to bottom- This study concerns mathematical analysis of a free- intensified baroclinic roll waves and in the overlying wa- boundary model of solid combustion. Nonlinear transi- ter column amplifying internal gravity waves are gener- tion behaviors of small disturbances of front propagation ated. The stability characteristics are described as a func- and temperature as they evolve in time, as well as the tion of the bottom drag coefficient and slope, Froude, bulk large-time quasi-steady oscillations of these disturbances, Richardson and Reynolds numbers associated with the are studied. Previous asymptotic solutions with some dom- overflow and the fractional thickness of the abyssal current inant modes have significant errors for almost all cases. In compared to the mean depth of the overlying water col- this study, a special asymptotic expansion method is used umn. The marginal stability boundary and the boundary to improve the accuracy of the approximation. separating the parameter regimes where the most unstable mode has a finite or infinite wavenumber are determined. Jun Yu, Yi Yang When it exists, the high wavenumber cut-off is obtained. University of Vermont Conditions for the possible development of an ultra-violet [email protected], [email protected] catastrophe are determined. In the infinite Reynolds num- ber limit, an exact solution is obtained which fully includes Jianping Cai, Meili Lin the effects of mean depth variations in the overlying water Zhangzhou Normal University, P. R. China column associated with a sloping bottom. For parameter [email protected], [email protected] values characteristic of the Denmark Strait overflow, the most unstable mode has wavelength of about 19 km, a geostationary period of about 14 hours, an e-folding am- CP2 plification time of about 2 hours and a down slope phase Stability and Bifurcation Analysis to Dissipative speed of about 74 cm/s. Cavity Soliton of Lugiato-Lefever Equation in One Dimensional Bounded Interval Gordon E. Swaters Department of Mathematical and Statistical Sciences We show a mathematically rigorous analysis for bifurca- University of Alberta tion structure of a spatially uniform equilibrium in non- [email protected] linear Schr¨odingerequations with dissipation, driving and detuning terms. Numerically, it has been reported that the “ snake bifurcation occurs. We ensure that the pitchfork CP1 bifurcation happens by using the bifurcation theory with On Conserved Quantities of Some Fluid Mechanical the symmetry, and make a much finer analysis at the codi- Equations mension two bifurcation point to prove the existence of fold bifurcation of nontrivial equilibrium around there. The study of conserved quantities for PDEs related to fluid mechanics is significant from both theoretical and practical Tomoyuki Miyaji points of views. For example, conservation of energy of the Dept. of Math. and Life Sci., Euler equations is closely related to turbulence and thus Hiroshima University plays an important role in its weak solution theory. In this [email protected] talk I will present recent results giving necessary and suf- ficient conditions for the conservation of various quantities Isamu Ohnishi of the Euler, MHD and SQG equations. Hiroshima University isamu [email protected] Xinwei Yu University of Alberta Yoshio Tsutsumi 48 PD09 Abstracts
Department of Mathematics, Graduate School of Science The upper and lower solutions are verified through inte- Kyoto University gral equations and only required to be continuous. Appli- [email protected] cations of the reaction-diffusion equations include several important models in biological invasion and disease spread. CP2 Mathematical Analysis to Coupled Oscillators Sys- Haiyan Wang tem with a Conservation Law Arizona State University [email protected] We consider the 3-component reaction-diffusion system of a coupled oscillator with conservation law on an interval O = [0, 1] with homogeneous Neumann boundary condi- CP3 tion. We can prove mathematically rigorously that the A New Glimm Functional and Convergence Rate of wave instability can occur under natural and appropriate Glimm Scheme for General Systems of Hyperbolic conditions for this system. We especially notice that this Conservation Laws system has a preferable cluster size of synchronization of oscillations, which tends to smaller and smaller as e goes In this work, we introduce a new Glimm functional for to 0. It may be interesting that, if the effect by which the general systems of hyperbolic conservation laws. This new synchronized oscillation occurs is too much, then the syn- functional is consistent with the classical Glimm functional chronized cluster is vanishing and a kind of homogenization for the case when each characteristic field is either gen- happens. uinely nonlinear or linearly degenerate, so that it can be viewed as “optimal’ in some sense. With this new func- Isamu Ohnishi tional, the consistency of the Glimm scheme is proved Hiroshima University clearly for general systems. Moreover, the convergence rate isamu [email protected] of the Glimm scheme is shown to be the same as the one obtained for systems with each characteristic field being Tomoyuki Miyaji genuinely nonlinear or linearly degenerate. Dept. of Math. and Life Sci., Hiroshima University Zaihong Jiang [email protected] Joint Advanced Research Center of University of Science and Technology of China and City University of Hong Kong Ryo Kobayashi [email protected] Dept. of Math. and Life Sci.s, Hiroshima University [email protected] CP3 Stability of a New Difference Scheme With Un- Atsuko Takamatsu bounded Operator Coefficients in Hilbert Space Department of Electrical Engineering and Bioscience 2 Waseda University d u(t) ≤ The initial-value problem dt2 + A(t)u(t)=f(t)(0 atsuko [email protected] t ≤ T ),u(0) = ϕ, u(0) = ψ for hyperbolic equations in a Hilbert space H with the self-adjoint positive definite CP2 operator A(t) is considered. A new second-order accurate absolutely stable difference scheme generated by integer Bifurcation in Two-Dimensional Electrohydrody- powers of A(t) for approximately solving this problem is namics developed. The stability estimates for the solution of this Experimental work in annular electroconvection in a smec- difference scheme are established. Theoretical results are tic film reveals the birth of convection-like vortices in the supported for solving one-dimensional hyperbolic partial plane as the electric field intensity is increased. Modeling differential equation with initial conditions. this phenomenon involves two-dimensional Navier-Stokes Mehmet Emir Koksal equations coupled with simplified Maxwell equations. The Florida Institute of Technology system satisfies the prerequisites for application of O(2)- mkoksal@fit.edu equivariant bifurcation theory. A center manifold reduc- tion with coefficients computed numerically unveils inter- esting dynamics, including stationary and spatio-temporal Allaberen Ashyralyev patterns. Results are compared with those from previous Fatih University work in two-dimensional thermoconvection. aashyr@@fatih.edu.tr Dan Rusu Indiana University-Purdue University Columbus CP3 [email protected] Analytical and Numerical Methods in the Problem of Extinguishing of Vibration for Petrovsky Hyper- bolic Equations CP2 Traveling Wave Solutions for a Class of Reaction- The problem of damping vibration described by the Petro- Diffusion Equations vsky hyperbolic equation
Combining upper and lower solutions, monotone itera- m−1 2 m (−1) ytt = a Dxxy + g(t, x), 0 small vicinities of fixed points, is considered. For example conduction may not be suitable for hyperbolic heat trans- fer. In this paper, we present an anti-diffusive method t to solve the hyperbolic heat transfer equation. The solu- tion is compared with the analytical one as well as the one g(t, x)=u(t)δ(x − x0 − v(τ) dτ), obtained from a high-order TVD (Total Variation Dimin- 0 ishing) scheme. The order of accuracy of the anti-diffusive method is also investigated numerically. where δ – delta-function of Dirac, x0 – fixed point, u(t) and v(t) are two control functions. Note, that m = 1 cor- Wensheng Shen responds to the string or membrane vibration, and m =2 SUNY Brockport those ones of beam or plate. The solution of the problem of Department of Computer Science damping vibrations requires a study on the special problem [email protected] of moments, the description of the set of damper movement optimal trajectories and the solution of the minimization Leigh J. Little problem of functional arising on the trajectories. Department of Computational Science State University of New York, Brockport Igor Mikhailov [email protected] Russian Academy of Sciences mikh [email protected] CP4 Leonid A. Muravey Boundary Integral Solution of the Time Fractional Russian State Aviation Technology University Diffusion Equation in C∞ Domains [email protected] Recently, fractional derivatives have found new appli- cations in engineering, physics, finance and hydrology. CP3 In physics, fractional diffusion-type equations describe Singular Specific Heat and Second Order Phase anomalous diffusion on heterogeneous media. Transitions Our interest is to study the boundary integral solution of the fractional diffusion equation Motivated by the presence of second order phase transitions α in materials such as liquid helium, we examine specific heat ∂t Φ − ΔxΦ=0, in QT =Ω× (0,T) (1) functions containing an integrable singularity at a criti- cal temperature. Our system is non-strictly hyperbolic, with the Dirichlet boundary condition Φ = g on ΣT = damped and may fail to be genuinely nonlinear. We will Γ × (0,T) and the zero initial condition Φ(x, 0)=0,x∈ Ω α discuss breakdown of solutions, results which vary with the and where ∂t is the fractional Caputo time derivative of specific heat assumptions. A connection between steady- order 0 <α≤ 1. Boundary integral approach allows us state solutions and dipole solutions for the porous media to prove the solvability of (1) in the scale of anisotropic equation will be presented. Sobolev spaces. Katarzyna Saxton Jukka Kemppainen Loyola University, New Orleans University of Oulu [email protected] [email protected].fi CP3 CP4 What Went Wrong? Cauchy Problems for Semilinear Parabolic Equa- tions Viscosity approximations are constructed for a class of two-dimensional Riemann problems for the two-dimensinal In this paper, we consider solvability of nonlinear evolu- ”p-system” also called the nonlinear wave system. Under tion equations in Banach spaces. In particular, we study seemingly mild additional assumptions, a subsequence con- Cauchy problems for semilinear parabolic equations on un- verges and produces a global weak solution in the limit of bounded domains using fixed point methods. vanishing viscosity. However, such weak solutions provably do not exist in the generality considered. Nickolai Kosmatov University of Arkansas at Little Rock Michael Sever [email protected] Einstein Institute of Mathematics Hebrew University of Jerusalem [email protected] CP4 On Solvability of Initial-Boundary Value Problems for Multi-Weighted Parabolic Systems CP3 Anti-Diffusive Method for Hyperbolic Heat Trans- I.G. Petrovskii 1938, T. Shirota 1955 and V. Solonnikov fer 1965 introduced successively more and more general defi- nitions of parabolic systems. In spite of the fact that the The hyperbolic heat transfer equation is a model used to entries of matrix operator have different orders with respect replace the Fourier heat conduction for heat transfer of ex- to the spacial variables, the t-differentiation has a constant tremely short duration or at very low temperature. Unlike weight 2b with some natural b. Therefore all these sys- the Fourier heat conduction, in which heat energy is trans- tems can be called single-weighted. In the present work, fered by diffusion, thermal energy is transfered as wave a more general definition of multi-weighted parabolic sys- propagation at finite speed in the hyperbolic heat trans- tems is formulated and solvability of the initial-boundary fer model. Therefore methods accurate for Fourier heat value problems for the multi-weighted systems is proved in 50 PD09 Abstracts appropriate Sobolev-type spaces. ical Quantum Dots with Finite Confinement Bar- riers Alexander Kozhevnikov Professor, Department of Mathematics, University of By employing a novel model for the dielectric interface be- Haifa, 31905, Israel tween a spherical quantum dot and the surrounding matrix, [email protected] an analytical solution of the self-polarization energy of a spherical quantum dot with a finite confinement barrier is first presented. A numerical procedure is then proposed to CP4 calculate the self-polarization energy of a spherical quan- Asysmptotic Behavior of the Solutions for a Cou- tum dot with arbitrary continuous model for the dielectric pled Nonlinear Schrodinger Equations interface, which can fully recover the exact solution of the self-polarization energy. We study the solutions of a coupled nonlinear Schr¨odinger equation which has been derived in many areas of physical Shaozhong Deng modeling: University of North Carolina at Charlotte 2 2 2 2 [email protected] iut +Δu − (a|u| + b|v| )u =0,ivt +Δv − (c|u| + d|v| )v =0 where the spatial dimension n is 3. Assume that (1) a>0 CP5 (2) b = c>0 (3) d>0 are real constants. We show that the local energy of a solution is integrable in time t and A Study of the Level Sets with Point Correspon- the local L2 norm of the solution approaches zero as time dence Approach to the Stress-Strain Analysis of t approaches the infinity. Deformable Thin Elastic Membranes Jeng-Eng Lin One of the most challenging problems in the stress-strain George Mason University analysis of deformable membranes is building accurate nu- [email protected] merical discretizations for the membranes’ complex geome- tries. We propose to use the level set method with point correspondence to generate a simplified geometry of a lin- CP4 ear elastic thin membrane, perform the stress-strain anal- Periodic Solutions to a p−Calogero Class of Evolu- ysis on this new configuration which is easier to handle tion Equations numerically, and use the point correspondence mapping to predict the stress-strain distributions in the original geom- We consider spatially periodic solutions to the class of evo- etry. lution equations Jason S. Fritz 1 2 2 The Pennsylvania State University utx+uuxx = pux−(p+1) uxdx, 0 Semiconductor Gaz Sensor Boltzmann Equation. where the number of band-gaps can be made arbitrarily large. Finally, we present some numerical examples by us- The aim of this work is to derive, from a kinetic description, ing the finite element method. a drift-diffusion model for electron transport in semicon- ductor gaz sensor, by using diffusion and homogenization Keijo M. Ruotsalainen approximation. A homogenized absorption term appears Professor, University of Oulu in the fluid model. In the presence of a self-consistent spa- [email protected].fi tially oscillating electrostatic potential, the right hand side of the Boltzmann equation presents, an absorption term Sergey Nazarov added to the electron-phonon operator collision. This ab- Russian Academy of Sciences, IPME sorption term has the periodicity of the oscillating electro- St. Petersburg static potential. [email protected] Hechmi Hattab ENIT-LAMSIN - Tunis TUNISIA Jari Taskinen [email protected] Department of Mathematics University of Helsinki [email protected].fi CP5 Effect of Anisotropic Surface Energy in An Epitax- CP6 ial Growth Model On the Numerical Solution of Differential- Mullins’ general diffusion equation can be used for model- Algebraic Equations (daes) By Using Legendre ing the epitaxial growth of quantum dots. Several effects Polynomials Approximation have to be considered when defining a chemical potential. Stresses appear due to misfit in the crystal’s grid spacings, The numerical solution of differentialalgebraic equations a thin layer between the dots is apparent because of in- (DAEs) using the Legendre polynomials approximation is termolecular interactions, and the inherent anisotropy of considered in this paper. By using the theories and meth- crystals results in pyramidal shapes of the islands. In self- ods of mathematical analysis and computer algebra, a algo- assembly systems such as Ge/Si, flat films are unstable rithm of Legendre polynomials approximation method for after they excess a critical height. Nano-dots evolve and solving differential-algebraic equation systems was estab- coarsen with time. Their small slopes allow to apply a thin- lished and a Maple procedures mainproc was established. film reduction. We extend a recent model by letting the Two different problems are solved using the Legendre poly- surface energy depend on orientation. Simulations with a nomials approximation and the solutions are compared pseudospectral method show that one additional nonlinear with the exact solutions. First, we calculate the power se- term in the final evolution equation is suitable to guaran- ries of a given equation system and then transform it into tee faceting of the structures. A linear stability analysis Legendre polynomials approximation form, which gives an explains how anisotropy destabilizes flat films. The criti- arbitrary order for solving the DAE numerically. cal height is reduced while the most unstable wave number Mutlu Akar grows with increasing anisotropy parameter. Yildiz Technical University Maciek D. Korzec [email protected] WIAS Berlin [email protected] Ercan elik Ataturk University Peter L. Evans [email protected] Humboldt University in Berlin [email protected] CP6 The Discontinuous Galerkin Method for Two Di- CP5 mensional Hyperbolic Problems Motion of a Viscous Fluid Around a Rigid Body We investigate the superconvergence properties of the dis- The lecture will contain the results about the problem of continuous Galerkin method applied hyperbolic problems on triangle meshes. We show that the discontinuous finite a motion of a viscous fluid around a rotating rigid body p+2 in cases when velocity of fluid at infinity is or not parallel element solution is O(h ) superconvergent at the Legen- with rotation of body. dre points on the outflow edge for triangles having one out- flow edge. For triangles having two outflow edges the finite p+2 Sarka Necasova element error is O(h ) superconvergent at the end points Mathematical Institute of Academy of Sciences of the inflow edge. We use these results to construct sim- Czech Republic ple, efficient and asymptotically correct a posteriori error [email protected] estimates for hyperbolic problems on unstructured meshes. Mahboub Baccouch CP5 Department of Mathematics Computation of the Band-Gap Structure of the University of Nebraska at Omaha Elasticity Operator on Periodic Waveguide [email protected] In this paper we shall study the spectral problem in the lin- earized elasticity. It will be shown that an infinite waveg- CP6 uide with periodically positioned cells, the essential spec- Discrete Compactness for the p-Version of Discrete trum contains gaps. Moreover, we construct examples 52 PD09 Abstracts Differential Forms Pamukkale University, Engineering Faculty Mechanical Engineering Department The aim of this paper is to prove the discrete compactness [email protected] property for a wide class of p finite element approximations of the Maxwell eigenvalue problem in two and three space dimensions. In a very general setting, we find sufficient CP6 conditions for the p-version of the generalized discrete com- Space-Time Finite Volume Differencing Method for pactness property which is formulated in the framework of Effective Higher Order Accurate Schemes for Gen- discrete differential forms of order on a polyhedral do- eral Diffusion Equations main in Rd (0 < Vyacheslav Trofimov Faculty of Computational Mathematics Paramjeet Singh, Kapil Sharma Moscow State University Department of Mathematics [email protected] Panjab University Chandigarh, India [email protected], [email protected] I.A. Shirokov Moscow State University CP7 [email protected] On Schauder’s Fixed Point Theorem Applications to Nonlinear Variational Inequality with One Or CP6 Two Obstacle Application of the Generalized Iterative Differen- We use Schauder’s fixed point theorem to obtain existence tial Quadrature Method for Solving Lid-Driven of a solution of the elliptic unilateral problem. We consider 1,p Cavity Problem a compact convex subspace S of the space W0 (Ω). Then a mapping J is defined such that it is a continuous map in In this work, Lid-driven cavity problem was solved by using itself. Then J will have a fixed point. The nonlinear op- Generalized differential quadrature method and results are erator is of the form A(u)+F (x, u, Du) in an unbounded given by figure and tables. Good agreements are seemed domain Ω of Rn. The operator is defined from the space compared with other works. 1,p −1,p W0 (Ω) into the space W (Ω). F has growth of order Zekeriya Girgin p with respect to variable Du satisfying the sign condition. PD09 Abstracts 53 The obstacle is a measurable function with variable in R¯. viscosity solutions are allowed. We propose a finite dif- A and F satisfy Carath´eodory conditions. Similar solu- ference numerical scheme to approximate the solution of tions can be obtained for a Parabolic Inequality. The last the extended class of equations and analyze a set of initial- section gives an idea about numerical approximation of the boundary value problems for Hamilton-Jacobi equations unilateral problem. arising as the convective component of nonlinear flux lim- iter diffusion equations. Kumud S. Altmayer University of Arkansas Susana Serna kumud [email protected] Department of Mathematics Universitat Autonoma de Barcelona [email protected] CP7 Asymptotic and Numerical Solutions to Laplace Antonio Marquina Young Capillary Equation at Singularity. University of Valencia, Spain [email protected] The Laplace Young Equation is a nonlinear elliptic PDE models capillary surfaces (static liquid surfaces). An asymptotic sequences motivated by the geometry of do- CP7 main made it possible to construct an asymptotic solution Multitime Dynamic Programming in the non-power series cusp region. Also an exact-solution of an approximated equation has allowed us to obtain an This paper introduces a new type of dynamic programming accurate asymptotic solution in the circular cusp domain. PDEs for optimal control problems with performance crite- At last, a fast converging finite volume element method ria involving multiple integrals or curvilinear integrals. The was developed using asymptotic solutions. main novel feature of the multitime dynamic programming PDEs, relative to the standard Hamilton-Jacobi-Bellman Yasunori Aoki PDEs, is that they are connected to the multitime max- University of Waterloo imum principle. In other words, we present an interest- [email protected] ing and useful connection between the multitime maxi- mum principle and the multitime dynamic programming, CP7 characterizing the optimal control by means of a multi- time Hamilton-Jacobi-Bellman PDE (or system) that may The Dirichlet Problem for Willmore Surfaces of be viewed as a feedback law. In the case of performance Revolution criteria involving either multiple integrals or curvilinear in- Boundary value problems for the “Willmore equation’ are tegrals, with quadratic integrands, the new equations lead quasilinear and, in particular, frame invarint counterparts to multitime variants of the Riccati equation. Section 1 of the fourth-order linear clamped plate equation. They do introduces the multitime Hamilton-Jacobi PDEs from ge- not enjoy any form of a general comparison principle. In ometrical point of view splitting the discussions after the contrast to closed Willmore surfaces, only a few results are type of actions (multiple integrals or curvilinear integrals). known for boundary value problems. In the talk, a partic- Section 2 shows how multi-time control dynamics, based ularly symmetric situation is studied. Here, one succeeds on curvilinear integral action or on multiple integral ac- in finding strong a-priori-estimates on suitable minimising tion, determines the multitime Hamilton-Jacobi-Bellman sequences and obtains a-priori bounded classical solutions. PDEs via the maximum value function or its generating vector field. Section 3 describes the connection between Hans-Christoph Grunau multitime dynamic programming and the multitime max- Institut fur Analysis und Numerik imum principle. In the general case, the HJB PDEs do Otto-von-Guericke-Universitat, Magdeburg - Germany not have classical (smooth) solutions, but we can introduce [email protected] multitime viscosity solutions (extending the theory of P.- L. Lions and M. Crandall) or multitime minmax solutions Anna Dall’Acqua, Klaus Deckelnick (extending the theory of I. Subbotin). Otto von Guericke-Universitaet Magdeburg Constantin Udriste [email protected], [email protected] Bucharest, RO 060042, Romania [email protected] Steffen Froehlich Freie Universitaet Berlin [email protected] CP7 Over-Determined Systems of Partial Differential Friedhelm Schieweck Equations, Equilibrium Behavior of Two-Phase Otto von Guericke-Universitaet Magdeburg Granular Mixtures and Equations of Monge- [email protected] Ampere Type The equilibrium behavior of two-phase granular mixtures CP7 is described via an over-determined, quasi-linear system of Numerical Approximation of Viscosity Solutions partial differential equations. In the first part of the paper with Shocks we perform a compatibility analysis and various existential and regularity results are also presented. In the second We consider a class of Hamilton-Jacobi equations that part of the paper we demonstrate the connection of our admits a unique viscosity solution containing shocks, [Y. system to equations of Monge-Ampere type and discuss Giga, Viscosity solutions with shocks, Comm. Pure Appl. further implications. Math., vol. LV, 2002], as an extension of the classical the- ory by Crandall and Lions (1983) where only continuous Christos Varsakelis, Miltiadis Papalexandris 54 PD09 Abstracts Department of Mechanics Michael Weinstein Universite catholique de Louvain Columbia University [email protected], [email protected] [email protected] CP8 CP8 Non-Existence and Existence of Localized Solitary Generalized Analytic Functions in Axially Sym- Waves for the Two-Dimensional Long Wave-Short metric Oseen Flows Wave Interaction Equations A class of generalized analytic functions arising from the In this study we consider the two-dimensional long wave- Oseen equations in the axially symmetric case has been short wave interaction (LSI) equations which describe identified. For this class of functions, the generalized the interaction between high-frequency and low-frequency Cauchy integral formula has been obtained, and a series waves near the long wave-short wave resonance. We es- representation for the region exterior to a sphere has been tablish non-existence and existence results for the localized constructed. The velocity field has been represented in solitary waves of the LSI equations. Both the non-existence terms of two generalized analytic functions, and it has been and the existence results are based on Pohozaev-type iden- shown that for an exterior Oseen flow problem, those func- tities. We prove the existence of solitary waves by showing tions are uniquely determined provided that they both van- that the solitary waves are the minimizers of an associated ish at infinity. Also the pressure and vorticity have been variational problem. determined as real and imaginary parts of the two func- tions representing the velocity field, and the drag exerted Handan Borluk, Husnu Ata Erbay, Saadet Erbay on a solid body of revolution in the axially symmetric Os- Department of Mathematics, Isik Universty een flow has been expressed in terms of one of the involved [email protected], [email protected], generalized analytic functions. The problem of the axially [email protected] symmetric Oseen flow past a solid body of revolution has been reduced to an integral equation based on the general- ized Cauchy integral formula. The developed framework of CP8 generalized analytic functions has been illustrated in solv- Dispersion in 1-D Variational Boussinesq Model ing the problem of the Oseen flow past a solid sphere and solid bispheres. The Variational Boussinesq Model for waves over ideal fluid conserves mass, momentum, energy, and contains de- Michael Zabarankin creased dimensionality compared to the full problem. It Stevens Institute of Technology is derived from the Hamiltonian formulation via an ap- [email protected] proximation of the kinetic energy, and can provide ap- proximate dispersion characteristics. Having in mind a signalling problem, we search for optimal dispersive prop- CP9 erties of the 1-D linear model over flat bottom and, using Eigenfunctions and Eigenvalues for High Energy the finite element and spectral numerical codes, investigate Resonance for Non-Standard Sturm-Liouville Op- its robustness. erators with a Perturbation of the Coulomb Poten- tial. Ivan Lakhturov, Brenny Van Groesen University of Twente Resonances for the Hamiltonian operator result from the [email protected], separated radial Schr¨odingerequation. Perturbations e(r) [email protected] of the Coulomb potential can produce a sequence of res- onances. The high-energy resonance contributing equa- tions could be the three dimensional Schr¨odingeroperator CP8 with Coulomb potential or a non-standard Sturm-Liouville Scattering Resonances and the Dynamics of De- eigenproblem with interior singularities. We re-format the forming Bubbles in a Compressible Fluid radial Schr¨odingeroperator in R, which happens to be a non-standard Sturm-Liouville operator, and by specifying We consider the dynamics of a nearly spherical gas bub- boundary conditions that retain the resonance poles the ble in an inviscid, compressible fluid with surface tension. eigenfunctions solutions are obtained and subsequently the The time-decay of symmetric (volume) and asymmetric eigenvalues are derived. These solutions are obtained with (shape) deformation modes is governed by scattering res- minimal restriction on a non-constant perturbation e(r) of onances. These are eigenvalues of a non-self-adjoint spec- the Coulomb potential. tral problem associated with the linearization about the spherical equilibrium. We study the locations of scatter- Thomas Acho ing resonance energies and prove resonance expansions for University of the Free State, Bloemfontein, South Africa the bubble surface perturbation and fluid velocity poten- [email protected] tial. These perturbations decay exponentially with ad- vancing time. While the lifetime of decaying radial oscil- −1 lations, Tradial ∼ ( denotes the Mach number), there CP9 are asymmetric shape modes that decay much more slowly. Improved Convergence of the Allen-Cahn Equation In particular, the slowest decaying mode has a lifetime 2 to Generalized Motion by Mean Curvature Tshape ∼ exp(κ/ ), κ>0, thus occurring beyond all or- ders in . It is known that, as ε → 0, the Allen-Cahn equation ε ε −2 ε ut =Δu + ε f(u ) converges — for all times— to the Abby Shaw-Krauss generalized motion by mean curvature defined via the level- Columbia University set approach and viscosity solutions. In this talk, given a Dept of Applied Physics and Applied Mathematics rather general initial data, we include the transition lay- [email protected] PD09 Abstracts 55 ers of the solutions uε in sets moving by mean curvature. be more intricate. In some special cases, this enables to obtain an O(ε| ln ε|) estimate of the thickness (measured in space-time) of the Galimzyan G. Islamov transition layers. Udmurt State University AMS Matthieu Alfaro, Jerome Droniou [email protected] Universite Montpellier 2 [email protected], [email protected] CP9 Min-Max Game Theory and Non-Standard Differ- ential Riccati Equations under Singular Estimates Hiroshi Matano At At University of Tokyo for e B and e G in the Absence of Analyticity [email protected] We consider an abstract dynamical system, which is char- acterized by ”singular estimates”, as it arises from many CP9 concrete hyperbolic/parabolic coupled PDE models, that Existence and Asymptotic Behavior of Solutions to are subject to boundary/point control and deterministic Quasilinear Thermoelastic Systems Without Me- disturbance. For such a system, we study a min-max game chanical Dissipation theory problem with quadratic cost functional over a finite horizon. The problem is fully solved in feedback form via a We consider a class of quasilinear thermoelastic systems. Riccati operator which satisfies a non-standard differential The model considered is characterized by supercritical non- Riccati equation. Several PDE-illustrations are presented. linearity and accounts for large displacements and large strains. Existence of finite energy (weak)global solutions and exponential decay rates for the energy of weak solu- Jing Zhang, Roberto Triggiani tions will be shown. The methods of proof will involve com- University of Virginia pensated compactness along with the application of special [email protected], [email protected] PDO non-local multipliers that have been developed in the context of the proof. CP10 Albert Altarovici, Irena Lasiecka Numerical Valuation of Multi-Currency Interest University of Virginia Rate Derivatives Using a Partial Differential Equa- [email protected], [email protected] tion Approach We consider the 3D PDE arising from a multi-currency CP9 model with foreign exchange (FX) skew. Finite differences Analysis on a Reaction-Diffusion System with Non- are used to discretize the space variables. Both GMRES linear Rate of Growth with pre-conditioner solved by FFT techniques and the Alternating-Direction Implicit (ADI) methods are consid- In this talk, we present some recently qualitative results ered. We analyze the impacts of the FX skew on long- on a nonlinear reaction-diffusion system which is used as a dated exotic interest rate derivatives with cancelable and model for animal and insect dispersal. We find that under knockout features, and study the effects of the boundary certain parametric conditions the system admits one pa- conditions on the accuracy of the numerical methods. rameter Lie groups of transformations, which lead to two independent first integrals. Duy Minh Dang University of Toronto Zhaosheng Feng Computer Science Department University of Texas-Pan American [email protected] Department of Mathematics [email protected] Christina Christara University of Toronto [email protected] CP9 Pseudo-Wave Functions. Ken Jackson Dept. of Computer Science ccording to J. von Neumann a set of all having a prop- University of Toronto erty G elements of Hilbert space B is a closed subspace [email protected] PB. Here via P is denoted an orthogonal projector on this subspace. We extended this original method to fi- nite dimensional projectors P of general form. An evolu- CP10 tion of quantum mechanical system is described by self- Explicit Solutions to the Kolmogorov Backward adjoint energy operator H. We study finite dimensional Equation solutions K of equation PH − HP = K(I − P ) with minimal rank. Any solution ϕ(t) of perturbed equation To obtain the mean time to failure of a device, a model i¯hϕ (t)=(H − K)ϕ(t),t>0 is called a pseudo-wave func- involving a system of two stochastic differential equations tion if a quadratic form ((H − K)ϕ(t),ϕ(t)) is real for all is proposed. This leads to a partial differential equation t>0. Efficient Einstein-Broglie formula E =¯hν con- known as Kolmogorov’s backward equation. This equa- nects energy and frequency spectrums of object described tion is solved explicitly in important particular cases. The by wave function. For object described by pseudo-wave method of similarity solutions is used. function a correlation between these types of spectrum may Mario Lefebvre 56 PD09 Abstracts Ecole Polytechnique Montreal tions for the p(x)-Laplacian with Rough Exponents Dept of Mathematics and Industrial Engineering [email protected] In this talk, properties of local fields inside mixtures of two nonlinear power law materials are studied. This constitu- Djilali Ait Aoudia tive model is frequently used to describe several phenomena Department of Mathematics and Statistics ranging from plasticity to optical nonlinearities in dielec- University of Montreal tric media. We develop new multiscale tools to bound the [email protected] local singularity strength inside microstructured media in terms of the macroscopic applied fields. CP10 Silvia Jimenez, Robert P. Lipton Department of Mathematics On Existence of Optimal Relaxed Controls for Dis- Louisiana State University sipative Stochastic PDEs in Banach Spaces [email protected], [email protected] We study existence of optimal relaxed controls for semilin- ear stochastic PDEs in UMD type-2 Banach spaces with CP11 multiplicative noise driven by a cylindrical Wiener pro- cess. The state equation is controlled through the drift Some Applications of Minimizing Variational Prin- coefficient which satisfies a dissipative-type condition with ciples for Helmholtz Equation respect to the state-variable. The main tools are the fac- We use the variational principles of Milton et al. for the torization method for stochastic convolutions in Banach complex Helmholtz/acoustic equation to formulate a finite spaces, and certain compactness properties of the factor- element method for solving these equations that is based ization operator and the class of relaxed control policies on entirely on minimization. Various boundary conditions are Suslin metrisable control sets. addressed. Also, we apply these finite element methods to Rafael Serrano the tomography problem and use the variational principles PhD Student, University of York, UK to derive some elementary bounds on effective properties. [email protected] Russell Richins, David C. Dobson University of Utah Zdzislaw Brzezniak [email protected], [email protected] University of York, UK [email protected] Graeme W. Milton University of Utah CP11 Department of Mathematics [email protected] Homogenization of a Heat Transfer Problem with Robin Dynamic Boundary Condition MS0 In this paper, we deal with the asymptotic behaviour of the heat equation with highly oscillating coefficients in a Some Non Linear Problems Involving Integral Dif- -periodic two-constituent composite with dynamic heat fusions - Part II transfer interfacial formulation. Problems involving integral diffusion appear in questions Abdelhamid Ainouz of continuum mechanics ( in variational , or ”divergence University of Sciences and Technology Houari form” ) and in probability and optimal control, from Levi [email protected] processes, in ” non divergence” form. In the first case, to develop a non linear theory ( similar to quasilinear equa- tions in the calculus of variations, it is necessary to develop CP11 the equivalent of the De Giorgi,Nash, Moser theory for ” Geometrically Nonlinear Models in Crystal Plas- bounded measurable kernels coming from non local varia- ticity with Penalization of Elastic Deformations tional integrals in energy spaces of fractional derivatives . In the second, parallel to the theory of optimal control,the In this talk we consider models for crystal plasticity in theory parallels the Krylov Safanov Harnack inequality and the framework of the modern calculus of variations. As a the Evans Krylov theorem. We will present the main ideas particular example we study the case of one slip system and arguments of the corresponding non local theorems. in two dimensions with rigid elasticity and we investigate whether these models arise as Gamma limits of models in Luis Caffarelli which elastic deformations are penalized. Surprisingly, the University of Texas at Austin answer depends strongly on the growth conditions satisfied Department of Mathematics by the elastic and the plastic contributions to the energy. caff[email protected] In the natural setting of quadratic potentials the compact- ess of sequences with finite energy is obtained by using a MS0 version of the classical div-curl lemma. Some Non Linear Problems Involving Integral Dif- Georg Dolzmann fusions Universitaet Regensburg [email protected] Problems involving integral diffusion appear in questions of continuum mechanics ( in variational , or ”divergence form” ) and in probability and optimal control, from Levi CP11 processes, in ” non divergence” form. In the first case, to Correctors, Homogenization, and Field Fluctua- develop a non linear theory ( similar to quasilinear equa- tions in the calculus of variations, it is necessary to develop PD09 Abstracts 57 the equivalent of the De Giorgi,Nash, Moser theory for ” Istituto per le Applicazioni del Calcolo, CNR bounded measurable kernels coming from non local varia- [email protected] tional integrals in energy spaces of fractional derivatives . In the second, parallel to the theory of optimal control,the theory parallels the Krylov Safanov Harnack inequality and MS1 the Evans Krylov theorem. We will present the main ideas Convergence of Perturbed Allen Cahn Equations and arguments of the corresponding non local theorems. to Forced Mean Curvature Flow Luis Caffarelli We study perturbations of the Allen–Cahn equation and University of Texas at Austin prove the convergence to forced mean curvature flow in the Department of Mathematics sharp interface limit. We allow for perturbations that are caff[email protected] square-integrable with respect to the diffuse surface area measure. We give a suitable generalized formulation for forced mean curvature flow and apply previous results for MS1 the Allen–Cahn action functional. Finally we discuss some Wavelet Analogue of the Ginzburg-Landau Energy applications. in Variational Models of Image Processing Luca Mugnai We discuss a wavelet analogue of the classical Ginzburg- Max Planck Institute for Mathematics in the Sciences Landau energy (WGL). Functionals of this type retain cer- [email protected] tain diffuse interface features and converge, in the Γ-sense, to a weighted analogue of the TV functional on charac- teristic functions of finite-perimeter sets. However, for the MS1 same values of the diffuse interface parameter, minimiz- Existence of Weak Mean Curvature Flow with ers of WGL have sharper transitions between equillibria Non-smooth Forcing Term than those of the classical Ginzburg-Landau functional, thus making the WGL energy an advantageous tool for We prove the existence of weak solution for the surface image processing. We design a variational model for multi- evolution problem where the velocity of hypersurface is purpose image reconstruction, where WGL is used as the the sum of the mean curvature vector and a given vec- regularizing energy part and the fidelity terms are defined tor field. The vector field is in a suitable Sobolev space so by specifics of each application (such as image inpainting, the evolving hypersurface satisfies good measure-theoretic blind deconvolution, superresolution and more). properties in addition to the Brakke’s evolution law. Julia Dobrosotskaya Yoshihiro Tonegawa Department of Mathematics, UMD Hokkaido University [email protected] [email protected] Andrea L. Bertozzi MS2 UCLA Department of Mathematics [email protected] Homogenization of Multiple Integrals Under Dif- ferential Constraints Two-scale techniques are developed for sequences of maps MS1 p M {uk}k∈N ⊂ L (Ω; R ) satisfying a linear differential con- Variational Properties of the Mumford Shah Func- straint Auk = 0. These, together with Γ-convergence ar- tional with Singular Operators guments and using the unfolding operator, provide a ho- mogenization result for energies of the type We consider functionals of the Mumford and Shah type with singular operators. In such functionals the fidelity x p M Fε := f x, ,u(x) dx with u ∈ L (Ω; R ), Au =0 term of the Mumford and Shah functional is replaced by ε a functional which penalizes the L2 distance between the Ω given image and a version of the unknown function which that generalizes current results in the case where A = curl. is filtered by means of a non-invertible linear operator. De- This is joint work with Stefan Kr¨omer. pending on the type of the involved operator the resulting variational problem may have several applications in the Irene Fonseca, Stefan Kroemer field of image reconstruction: for instance image deconvo- Carnegie Mellon University lution preserving discontinuities in the case of convolution [email protected], [email protected] operators, inpainting problems in the case of suitable lo- cal operators. We study the existence of minimizers of functionals of this type in suitable classes of discontinuous MS2 functions. The discontinuity set of such functions belongs Models for Dynamic Fracture to a class of compact sets introduced by L. Rondi to study problems of crack formation in solid structures. We will We introduce mathematical models for “sharp interface” consider the cases of local operators and convolution oper- dynamic fracture that make no assumption about crack ators. geometry, etc. Existing models for static/quasi-static frac- ture, and phase-field models for dynamic brittle fracture, Massimo Fornasier are based on variational principles that cannot be directly Johann Radon Institute for Computational and Applied transferred to the dynamic setting. We show how these Mathematics (RICAM) principles can be reformulated, so that they can be incorpo- [email protected] rated in dynamics. Time permitting, progress on existence will also be discussed. Riccardo March Christopher Larsen 58 PD09 Abstracts Worcester Polytechnic Institute ancient solutions are King-Rosenau solutions. Math Dept [email protected] Panagiota Daskalopoulos Columbia University [email protected] MS2 The Time-dependent von Krmn Plate Equation as a Limit of 3d Nonlinear Elasticity MS3 Monge-Ampere Type Equations and Geometric In this talk we will discuss the asymptotic behaviour of Optics solutions of the hyperbolic equation of nonlinear elastody- namics in a three-dimensional thin domain. More precisely, This talk will describe recent work on problems in geomet- we show that when the initial data and loadings are suf- ric optics related to Monge-Ampere type equations. They ficiently small in terms of the thickness, the limit dynam- mostly concern with the construction of surfaces that re- ics corresponds to the time-dependent von K´arm´anplate flect and/or refract radiation in a prescribed manner. equation. Cristian Gutierrez Helmut Abels Department of Mathematics Max Planck Institute Temple University Leipzig [email protected] [email protected] MS3 Maria Giovanna Mora Sissa, Trieste, Italy Title Not Available at Time of Publication [email protected] Abstract not available at time of publication. Stefan Mueller YanYan Li Hausdorff Center for Mathematics Mathematics Department Bonn, Germany Rutgers University [email protected] [email protected] MS2 MS3 Simulating 2-D crack propagation with XFEM and On the Levi Monge Ampre Equation a hybrid mesh We give a survey about some notions of curvatures asso- We will present a technique for simulating quasistatic crack ciated with pseudoconvexity and the Levi form the way propagation in 2-D which combines the extended finite ele- the classical Gauss and Mean curvatures are related to the ment method (XFEM) with a general algorithm for cutting convexity and to the Hessian matrix. We shall first show triangulated domains, and introduce a simple yet general that these curvature equations contain information about and flexible quadrature rule based on the same geometric the geometric feature of a closed hypersurface. Then, we algorithm. The combination of these methods gives several shall show that the curvature operators lead to a new class advantages. First, the cutting algorithm provides a flexi- of second order fully nonlinear equations which are not ble and systematic way of determining material connectiv- elliptic at any point. However, they have the following re- ity, which is required by the XFEM enrichment functions. deeming feature: the missing ellipticity direction can be Also, our integration scheme is straightfoward to imple- recovered by suitable commutation relations. We shall use ment and accurate, without requiring a triangulation that this property to study the Dirichlet problem for graphs incorporates the new crack edges or the addition of new with prescribed Levi curvature. degrees of freedom to the system. Annamaria Montanari Casey Richardson Universita’ di Bologna, Italy UCLA [email protected] Los Angeles, CA (USA) [email protected] MS4 Jan Hegemann Critical Thresholds in Hyperbolic Relaxations University of M × ”unster Critical threshold phenomena in one dimensional 2 2 [email protected] quasi-linear hyperbolic relaxation systems are investigated. We prove global in time regularity and finite time singular- ity formation of solutions simultaneously by showing the Eftychios Sifakis, Jeffrey Hellrung, Joseph Teran critical threshold phenomena associated with the under- UCLA lying relaxation systems. Our results apply to the well- [email protected], [email protected], known isentropic Euler system with damping. This is a [email protected] joint work with Hailiang Liu. Tong Li MS3 Department of Mathematics Ancient Solutions to the Ricci Flow on Surfaces University of Iowa [email protected] We provide a classification of ancient solutions to the Ricci flow on compact surfaces. We show that the only type II PD09 Abstracts 59 Hailiang Liu duced by Liu and Yu for the study of Boltzmann profiles, Iowa State University but applied to the stationary rather than the time- evolu- USA tionary equations. This yields a simple proof by contrac- [email protected] tion mapping in weighted Hs spaces. Kevin Zumbrun MS4 Indiana University Global Well-Posedness for the Microscopic Fene USA Model with a Sharp Boundary Condition [email protected] We prove global well-posedness for the microscopic FENE model under a sharp boundary requirement. The well- MS5 posedness of the FENE model that consists of the incom- Kinetic Theory of Synchronization pressible Navier-Stokes equation and the Fokker-Planck equation has been studied intensively, mostly with the nat- A kinetic theory for phase synchronization through binary ural flux boundary condition. Recently it was illustrated interactions will be discussed. This theory describes the in [C. Liu and H. Liu, Boundary Conditions for the Mi- nonequilibrium phase transition from a nematic, ordered croscopic FENE Models, 2008, SIAM J. Appl. Math., phase to an isotropic, disordered phase that occurs in a 68(5):1304–1315] that any preassigned boundary value of system of interacting rods, that are also subjected to ran- a weighted distribution will become redundant once the dom forcing. The steady-state solution for the nonlinear non-dimensional parameter b>2. In this talk, we present and nonlocal theory is obtained by expressing the Fourier that for the well-posedness of the microscopic FENE model transform of the orientation distribution as a function of (b>2) the least boundary requirement is that the distribu- the order parameter, which in turn, is obtained in terms tion near boundary needs to approach zero faster than the of the driving strength. This exact solution is obtained distance function. This condition is strictly weaker than using iterated partitions of the integer numbers. This ki- the natural flux boundary condition. Under this condition netic theory described alignment of vibrated granular rods it is shown that there exists a unique weak solution in a or chains, as well as biological systems such as molecular weighted Sobolev space. The sharpness of this boundary motors and microtubules. requirement is shown by a construction of infinitely many solutions when the distribution approaches zero as fast as Eli Ben-Naim the distance function. Los Alamos National Laboratory [email protected] Jaemin Shin Department of Mathematics, Iowa State University [email protected] MS5 Kinetic Models for Swarming Hailiang Liu I will present a kinetic theory for swarming systems of in- Iowa State University teracting, self-propelled discrete particles. Starting from USA the the particle model, one can construct solutions to a ki- [email protected] netic equation for the single particle probability distribu- tion function using distances between measures. Moreover, MS4 I will introduce related macroscopic hydrodynamic equa- tions. General solutions include flocks of constant den- Weak Solutions and Critical Thresholds for the sity and fixed velocity and other non-trivial morphologies One-dimensional Vlasov- Poisson Equation such as compactly supported rotating mills. The kinetic We give an explicit weak solution of the one-component theory approach leads us to the identification of macro- Vlasov-Poisson equations in a single space dimension with scopic structures otherwise not recognized as solutions of smooth electron sheet initial data. Moreover, we give sharp the hydrodynamic equations, such as double mills of two conditions on whether the moments of this weak solution superimposed flows. I will also present and analyse the will blow up or not. It depends on whether or not the initial asymptotic behavior of solutions of the continuous kinetic configuration of the density and velocity crosses a critical version of flocking by Cucker and Smale, which describes threshold. Beyond the critical time, we first construct a the collective behavior of an ensemble of organisms, ani- multi-valued solution to the Euler-Poisson equation, then mals or devices. This kinetic version introduced in Ha and we give a weak solution to the Vlasov-Poisson equation. Tadmor is obtained from a particle model. The large-time behavior of the distribution in phase space is subsequently Dongming Wei studied by means of particle approximations and a stabil- Dept. of Math. ity property in distances between measures. A continuous University of Wisconsin at Madison analogue of the theorems of Cucker-Smale will be shown to [email protected] hold for the solutions on the kinetic model. More precisely, the solutions concentrate exponentially fast their velocity to their mean while in space they will converge towards a MS4 translational flocking solution. Existence of Semilinear Relaxation Shocks Jose Carrillo We establish existence with sharp rates of decay and dis- ICREA tance from the Chapman–Enskog approximation of small- Universitat Autnoma de Barcelona amplitude shock profiles of a class of semilinear relaxation [email protected] systems including discrete velocity models obtained from Boltzmann and other kinetic equations. Our method of analysis is based on the macro–micro decomposition intro- 60 PD09 Abstracts MS5 which describes initiation of combustion process in inert Self-similarity for Ballistic Aggregation Equation porous media. In the framework of the considered model initiation is associated with blow up whereas global exis- We consider ballistic aggregation equation for gases in tence of solution suggests quenching. I will present neces- which each particle is identified either by its mass and im- sary conditions for global existence of solution and condi- pulsion or by its sole impulsion. For the constant aggrega- tions ensuring blow up. tion rate we prove existence of self-similar solutions as well as convergence to the self-similarity for generic solutions. Peter Gordon For some classes of mass and/or impulsion dependent rates Department of Math Sci we are also able to estimate the large time decay of some NJIT moments of generic solutions. [email protected] Stephane Mischler CEREMADE, Universite Paris Dauphine MS6 [email protected] Uniqueness of Steady States for a Chemotaxis Model Miguel Escobedo 2 Departamento de Matem´aticas,Universidad del Pais We consider in a disc of R a class of parameter-dependent, Vasco nonlocal elliptic boundary value problems that describe the Apartado 644, Bilbao. Espa˜na steady states of some chemotaxis systems. If the appearing [email protected] parameter is less than an explicit critical value, we establish several uniqueness results for solutions that are invariant under a group of rotations. Furthermore, we discuss the as- MS5 sociated consequences for the time asymptotic behavior of Kinetic Formulations and Regularizing Effects in the solutions to the corresponding time dependent chemo- Degenerate Quasi-linear PDEs taxis systems. Our results also provide optimal constants in some Moser-Trudinger type inequalities. We quantify the regularizing effects in a general family of quasi-linear scalar PDEs, using velocity averaging of Marcello Lucia their underlying kinetic formulations. The PDEs are first Department of Mathematics and second order equations which involve nonlinear trans- CUNY, College of Staten Island port and (possibly degenerate) diffusion. In particular, [email protected] we use the Chen-Perthame characterization of solutions to degenerate problems to derive new regularity results for MS6 convection-diffusion and elliptic equations driven by de- generate, non-isotropic diffusion. Boundary Blowup-type Solutions to Elliptic Equa- tions with Hardy Potentials Eitan Tadmor University of Maryland Semilinear elliptic equations which give rise to solutions USA blowing up at the boundary are perturbed by a Hardy po- [email protected] tential involving distance to the boundary. The size of this potential affects the existence of a certain type of solu- tions (large solutions): if the potential is too small, then Terence Tao no large solution exists. The presence of the Hardy poten- Department of Mathematics, UCLA tial requires a new definition of large solutions, following [email protected] the pattern of the associated linear problem. Nonexistence and existence results for different types of solutions will be MS6 given. Considerations are based on a Phragmen-Lindeloef type theorem which enables to classify the solutions and A Constrained Capacitive System Arising in sub-solutions according to their behaviour near the bound- MEMS ary. A model describing and electrostatic deflections of an elas- Vitaly Moroz tic membrane with a volume constraint, arising in MEMS, Department of Mathematics is considered. Existence and nonexistence of a solution is Swansea University explored as well as the structure of the solution set for the [email protected] governing equation. Time-of-fight analysis, construction of limiting singular solutions, and bifurcation diagrams, based on the variational formulation of the model, are used to MS7 determine the structure of the solution set and stability of Liouville Type Theorems in the Fluid Mechanics solutions found. In this talk we discuss Liouville type theorems in the fluid Regan Beckham mechanics. For incompressible fluids suitable decay con- Department of Mathematics dition near infinity combined with sign conditions for the University of Texas at Tyler pressure integrals leads to triviality of the solution. For sta- Regan [email protected] tionary compressible fluids we can omit the conditions on the pressure to lead to similar conclusion. For time depen- MS6 dent viscous compressiblle fluids we need to assume energy inequality for weak solution to get trivilaity of solution. Ignition in Porous Media as a Blowup Problem Dongho Chae We consider a system of two reaction diffusion equations Sungkyunkwan University PD09 Abstracts 61 [email protected] [email protected] MS7 MS8 Blow-up Problems in Fluid Mechanics: a Case Sharp Estimates on Minimum Traveling Wave Study Speed of Reaction Diffusion Systems We study a family of 2D fluid equations, including 2D Euler Abstract not available at time of publication. and surface quasi-geostrophic (SQG) equations. In hypo- viscous dissipative cases, an oscillatory damping in long- Yuanwei Qi time evolution is observed, both below/above the critical- University of Central Florida ity. This suggests all-time regularity even for supercritical [email protected] cases, but with a complicated transient. For inviscid cases, the growth rates of scalar gradient in Lp are compared within the family; in particular 2D Euler grows faster than MS8 SQG. This order is reversed in L∞. Generalized Traveling Wave Solutions and Their Propagating Speeds in Almost Periodic Media Koji Ohkitani The University of Sheffield This talk is concerned with the existence, stability, and K.Ohkitani@sheffield.ac.uk propagating speeds of generalized traveling wave solutions in time dependent, specially, in time almost periodic and unique ergodic bistable and monostable reaction diffusion MS7 equations, Questions on Polymeric Equations Wenxian Shen I will consider questions related to existence and decay Department of Mathematics and Statistics of solutions to kinetic models of incompressible polymeric Auburn University flow, in the case when the space of elongations is bounded, [email protected] and the spatial domain of the polymer is either a bounded domain Ω ⊂,n=2, 3 or it is the whole space R3. Consid- eration will also be given to solutions where the probability MS8 density function is radial in the admissible elongation vec- An Introduction to Fronts in Heterogeneous Media tors. Abstract not available at time of publication. Maria E. Schonbek Univ. of California at Santa Cruz Jack Xin Dept. of Math. Department of Mathematics [email protected] UCI [email protected] MS7 MS8 Loss of Smoothness and Energy Conserving Rough Weak Solutions for the 3d Euler Equations Pulsating Front Speed-up and Diffusion Enhance- ment in Flows A basic example of shear flow was introduced by DiPerna and Majda to study the weak limit of oscillatory solutions Abstract not avialable at time of publication. of the Euler equations of incompressible ideal fluids. In par- Andrej Zlatos ticular, they proved by means of this example that weak University of Chicago limit of solutions of Euler equations may, in some cases, fail [email protected] to be a solution of Euler equations. We use this example to provide non-generic, yet nontrivial, examples concerning the loss of smoothness of solutions of the three-dimensional 1,α MS9 Euler equations, for initial data that do not belong to C . Mortar Discretization for Stokes-Darcy Systems In addition, we use this shear flow to provide explicit exam- ples of non-regular solutions of the three-dimensional Euler We consider the coupling across an interface of fluid and equations that conserve the energy, an issue which related porous media flows with Beavers-Joseph-Saffman transmis- to the Onsager conjecture. Moreover, we show by means of sion conditions. Under an adequate choice of Lagrange this shear flow example that, unlike to the two-dimensional multipliers on the interface we analyze inf-sup conditions case, the minimal regularity in the three-dimensional vor- and optimal a priori error estimates associated with the tex sheet Kelvin-Helmholtz problem need not to be the continuous and discrete formulations of this Stokes-Darcy class of real analytic solutions. system. We allow the meshes of the two regions to be non- matching across the interface. Using mortar finite element Edriss S. Titi analysis and appropriate scaled norms we show that the University of California, Irvine and constants that appear on the a priori error bounds do not Weizmann Institute of Science, Israel depend on the viscosity, permeability and ratio of mesh [email protected], [email protected] parameters. Numerical experiments are presented. Claude Bardos Juan Galvis Laboratory J.L. Lions, Universit Texas A&M University ’{e} Pierre et Marie Curie [email protected] Paris, 75013, France 62 PD09 Abstracts MS9 MS9 Mixed Finite Element Methods for a Nonlinear Domain Decomposition for Strokes-Darcy Systems Stokes-Darcy Coupled Problem In this talk we couple deterministic and stochastic Darcy’s In this paper we analyze mixed finite element methods pressure equations with log-normal permeability. We first for a nonlinear model of the coupling of fluid flow with formulate the problem as an infinity-dimensional system porous media flow. More precisely, flows are governed by and then show well-posedness using inf-sup techniques. the Stokes equations in the fluid and a nonlinear Darcy Then we formulate the finite dimensional Galerkin scheme equation in the porous medium, and the corresponding and establish well-posedness and a priori error estimates. transmission conditions are given by mass conservation, balance of normal forces, and the Beavers-Joseph-Saffman Marcus Sarkis law. The pseudostress and the velocity in the fluid, to- IMPA (Brazil) and WPI (USA) gether with the velocity, the pressure, and the gradient of [email protected] pressure in the porous medium, constitute the main un- knowns of the model. In addition, the transmission condi- Juan Galvis tions become essential, which leads to the introduction of TexasA&M the traces of the porous media pressure and the fluid veloc- [email protected] ity as the associated Lagrange multipliers. We show that the resulting variational formulation has a twofold saddle- point structure and then we prove the unique solvability MS10 of it. Also, we establish suitable hypotheses on the finite Constructing Markov Models for Barrier Options element subspaces ensuring that the associated Galerkin scheme becomes well posed. Finally, a reliable and effi- In this talk we will present an approach for modeling bar- cient residual-based a posteriori error estimator is derived. rier options which is analogous to the ”local volatility” ap- proach for modeling European options. In place of the forward equation, we employ a recent weak existence re- Gabriel Gatica sult which can handle the path dependence in the option Universidad de Concepcion payoff. We then use a minor variation on the martingale Departamento de Ingeniera Matematica problem of Stroock and Varadhan to conclude that the re- [email protected] sulting model is Markov. Ricardo Oyarzua Gerard Brunick Universidad de Concepcion Department of Mathematics [email protected] The University of Texas at Austin [email protected] Francisco Sayas Department of Mathematics MS10 University of Minnesota Portfolios and Risk Premia for the Long Run [email protected] This talk develops a method to derive optimal portfolios and risk-premia explicitly in a general diffusion model, MS9 for an investor with power utility in the limit of a long On the Coupled Problem of Navier-Stokes, Darcy horizon. The market has several risky assets and is po- and Transport Equations tentially incomplete. Investment opportunities are driven by, and partially correlated with, state variables following The study of a coupled flow and transport system in ad- an autonomous diffusion. The framework nests models of jacent surface and subsurface regions is of interest for the stochastic interest rates, return predictability, stochastic environmental problem of contaminated aquifers through volatility and correlation risk. For each relative risk aver- rivers. First, two models describing the coupled Navier- sion parameter value, the long-run portfolio, its implied Stokes and Darcy equations are presented. The two mod- risk-premia, the long-run pricing measure, and their per- els differ only by the interface condition for the balance formance on finite horizons are obtained. In the case of of forces. Weak solutions are constructed using a Galerkin a single state variable, a candidate solution is derived for approach. The numerical discretization of the coupled flow the finite horizon value function, and hence optimal port- problem uses continuous elements in the surface region folio and pricing measure. An application to a single state and discontinuous finite elements in the subsurface. Nu- variable, potentially non affine, model concludes. merical solutions obtained with the two models are com- pared as the input data (permeability, viscosity) varies. Scott Robertson Second, a transport equation is coupled with the Navier- Department of Mathematical Sciences Stokes/Darcy flow. We prove existence and uniqueness of Carnegie Mellon University a weak solution. The transport equation is solved by a [email protected] discontinuous Galerkin method. Error estimates for the coupled problem are obtained. Numerical simulations are shown for heterogeneous porous media. MS10 Optimal Investment with Random Discrete Order Beatrice Riviere Flow in Illiquid Markets Rice University Houston, Texas, USA We consider an optimal investment problem where a risky [email protected] asset is traded and observed only at a sequence of arrival times whose intensity approaches infinity as we get closer to maturity. We solve this stochastic control problem us- PD09 Abstracts 63 ing dynamic programming, by first solving the integral model oscillations. The idea is that a family of the homo- Hamilton-Jacobi-Equation and then going trough a veri- geneous Sobolev spaces could provide us with the distin- fication argument. The presentation is based on joint work guishable patterns when it comes to texture and noise in with Paul Gassiat and Huyen Pham. images. Mihai Sirbu Yunho Kim University of Texas at Austin Department of Mathematics [email protected] University of California, Irvine [email protected] MS10 Luminita A. Vese On the Optimal Stopping Problems for Levy Pro- University of California, Los Angeles cesses Department of Mathematics [email protected] Value functions of optimal stopping problems for processes with Levy jumps are known to be generalized solutions of variational inequalities. Assuming the diffusion compo- John Garnett nent of the process is nondegenerate, in my talk I will show Department of Mathematics, UCLA that the value function is a classical solution of the vari- [email protected] ational inequality in the continuation region for problems with either finite or infinite variation jumps. Moreover, MS11 the smooth-fit property is shown via the global regularity of the value function. Our global regularity results gen- Exact Reconstruction of Piecewise Constant Color eralize the results of Bensoussan and Lions (1984) which Images by a Total Variation Based Model were developed for bounded domains. On the other hand, 2,1 In this work we address the following colorization problem: until now the value function was known to be C only for How can a color image be recovered when the underly- the optimal stopping problems when the jumps have finite ing gray level function is known everywhere but only small activity. This is a joint work with my Ph.D. student Hao patches of color are available? We consider the Red-Green- Xing. Blue (RGB) total variation model proposed by Fornasier Erhan Bayraktar and subsequently studied by Fornasier and March, and we University of Michigan seek to study the faithfulness of the reconstructions it pro- Department of Mathematics vides, with particular emphasis on the possible creation of [email protected] new, spurious contours in the restored image. In the main result of the paper, using calibrations techniques, we show that the reconstruction is faithfull in the case of sufficiently Hao Xing regular piecewise constant images, if the exact information Boston University on the colors is known over a (possibly) very small but uni- [email protected] formly distributed area. This is joint work with I. Fonseca, G. Leoni, and F. Maggi. MS11 Massimiliano Morini A Variational Approach to Hyperspectral Image SISSA Fusion Trieste,Italy [email protected] There has been significant research on pan-sharpening mul- tispectral imagery with a high resolution image, but there has been little work extending the procedure to high dimen- Irene Fonseca sional hyperspectral imagery. We present a wavelet-based Carnegie-Mellon University variational method for fusing a high resolution image and Department of Mathematical Scienc a hyperspectral image with an arbitrary number of bands. [email protected] To ensure that the fused image can be used for tasks such as classication and detection, we explicitly enforce spectral Giovanni Leoni coherence in the fusion process. This procedure produces Carnegie Mellon University images with both high spatial and spectral quality. We [email protected] demonstrate this procedure on several AVIRIS and HY- DICE images. This is joint work with Michael Moeller and Francesco Maggi Todd Wittman. University of Florence, Italy [email protected]fi.it Andrea L. Bertozzi UCLA Department of Mathematics [email protected] MS11 Bidirectional Inverse Consistent Deformable Image Registration MS11 Texture - Noise Separation Model This paper presents a novel variational model for inverse consistent deformable image registration. This model de- The BV space is an excellent space to capture piecewise forms the source and target image simultaneously, and smooth components in an image. However, we still do not aligns the deformed source and deformed target images in know what differentiates texture from noise in corrupted the way that the both transformations are inverse consis- images. We will pursue a method to answer this ques- tent. The model does not computes the inverse transforms tion using homogeneous Sobolev spaces, which are good to explicitly, alternatively it finds two more deformation fields 64 PD09 Abstracts satisfying the invertibility constraints. Moreover, to im- in time of the propagation along a ”smooth” path. Thus, prove the robustness of the model to noises and the choice if time continuity is viewed as essential, the classical view of parameters, the dissimilarity measure is derived from of crack kinking along a single crack branch seems incor- the likelihood estimation of the residue image. The pro- rect and a change in crack direction is necessarily a more posed model is formulated as an energy minimization prob- complex process. lem, which involves the regularization for four deformation fields, the dissimilarity measure of the deformed source and Antonin Chambolle deformed target images, and the invertibility constraints. Centre de Math´ematiquesAppliqu´ees The experimental results on clinical data indicate the ef- Ecole´ Polytechnique, France ficiency of the proposed method, and improvements in ro- [email protected] bustness, accuracy and inverse consistency. Gilles Francfort Xiaojing Ye Universit´ede Paris Nord University of Florida Villetaneuse, France Department of Mathematics [email protected] xye@ufl.edu Jean-Jacques Marigo Yunmei Chen Universit´eP. et M. Curie University of Florida Paris, France [email protected]fl.edu [email protected] MS12 MS12 Quasistatic Evolution for Cam-Clay Plasticity Symmetry of Locally Minimizing Solutions for the 3D Ginzburg-Landau System Cam-Clay nonassociative plasticity exhibits both harden- ing and softening behavior, depending on the loading. For We shall present a result classifying nonconstant entire many initial data the classical formulation of the qua- local minimizers of the standard Ginzburg-Landau func- sistatic evolution problem has no smooth solution. We tional for maps from R3 into R3 satisfying a natural energy propose here a notion of generalized solution, based on a bound. Up to translations and rotations, such solutions of viscoplastic approximation. To study the limit of the vis- the Ginzburg-Landau system are given by an explicit solu- coplastic evolutions we rescale time, in such a way that tion equivariant under the action of the orthogonal group. the plastic strain is uniformly Lipschitz with respect to the rescaled time. The limit of these rescaled solutions, as the viscosity parameter tends to zero, is characterized Vincent Millot through an energy-dissipation balance, that can be writ- Universit´ede Paris VII ten in a natural way using the rescaled time. It turns out Paris, France that the proposed solution may be discontinuous with re- [email protected] spect to the original time, and our formulation allows to compute the amount of viscous dissipation occurring in- Adriano Pisante stantaneously at each discontinuity time. University of Rome La Sapienza Roma, Italy Gianni Dal Maso [email protected] SISSA, Trieste, Italy [email protected] MS12 Antonio DeSimone Onsager Theory and Singularities in Nematic Liq- SISSA uid Crystals 34014 Trieste, ITALY [email protected] We present a theory of orientational order in nematic liq- uid crystals which interpolates between several distinct ap- Francesco Solombrino proaches based on the director field (Oseen and Frank), SISSA order parameter tensor (Landau and de Gennes), and ori- Trieste, Italy entation probability density function (Onsager). [email protected] Ibrahim Fatkullin Department of Mathematics MS12 The University of Arizona Crack Kinking in Brittle Fracture [email protected] Smooth crack propagation in an isotropic 2d brittle mate- Valeriy Slastikov rial is widely viewed as the interplay between two separate University of Bristol criteria. Griffith’s cap on the energy release rate along Bristol, UK the crack path decides when the crack propagates, while [email protected] the Principle of Local Symmetry decides in which direction that crack propagates (the kinking direction). Other com- peting criteria yield different and even contradictory kink- MS13 ing angles. We revisit crack path in the light of energetic High Field and Diffusion Asymptotics for the meta-stability of the current state among suitable compet- Boltzmann Equation ing crack states. We then demonstrate that 2d crack kink- ing in an isotropic setting is incompatible with continuity The high field asymptotics of the fermionic Boltzmann PD09 Abstracts 65 equation is sconsidered. It consists in a singular perturba- n. tion where the driving forces balance the collision effects. The limiting equation is a nonlinear conservation law for Hailiang Liu the particle density. We show the convergence to entropy Iowa State University solutions of the limiting conservation law. Under symme- USA try conditions on the collission cross section, the nonlinear [email protected] limiting flux is parallel to the force field. We therefore in- vestigate the orthogonal direction to the electric field, and prove after a suitable rescaling that the behaviour is gov- MS13 erned by the original conservation law with an additional A Bloch Band Based Level Set Method for Com- nonlinear diffusion in the orthogonal (to the force field) di- puting the Semiclassical Limit of Schr¨odinger rection. The main tool used in the convergence proof is a Equations new estimate for the dissipation of a family of kinetic en- tropies which converge to convex entropies of the limiting A novel Bloch band based level set method is proposed for conservation law. While the entropy dissipation is usually computing the semiclassical limit of Schr¨odingerequations in periodic media. For the underlying equation subject to estimated by quantities of the type dist(f, Feq) represent- ing the distance of the distribution function to the equi- a highly oscillatory initial data, a hybrid of the WKB ap- librium set, the new estimate involves a quantity of the proximation and homogenization leads to the Bloch eigen- value problem and an associated Hamilton-Jacobi system form dist(f, Feq(u))dμ(u), where the macroscopic equi- libria depend on a velocity variable u and μ is a probability for the phase in each Bloch band, with the Bloch eigen- measure. This estimate allows to control high velocities, value be part of the Hamiltonian. We formulate a level set pass to the limit in the diffusion current and shows the description to capture multi-valued solutions to the band convergence to the entropy solution of the limiting equa- WKB system, and then evaluate total homogenized density tion. over a sample set of bands. A superposition of band den- sities is established over all bands and solution branches Naoufel Ben Abdallah when away from caustic points. The numerical approach University of Toulouse splits the solution process into several parts: i) initialize [email protected] the level set function from the band decomposition of the initial data; ii) solve the Bloch eigenvalue problem to com- Hedia Chaker pute Bloch waves; iii) evolve the band level set equation to Ecole Nationale d’Ingenieurs de Tunis compute multi-valued velocity and density on each Bloch Tunisia band; iv) evaluate the total position density over a sample [email protected] set of bands using Bloch waves and band densities obtained in step ii) and iii), respectively. Numerical examples with different number of bands are provided to demonstrate the MS13 good quality of the method. Global Existence of a Free Boundary Problem with Zhongming Wang Non-Standard Sources Dept. of Math. We present a kinetic model with free boundary in price UCSD formation. The model consists on high nonlinear parabolic [email protected] equations with nonstandard sources.Global existence re- sults and regularity for the free boundary are given. MS14 Maria Gualdani Analysis of Coupled Systems Describing Sprays Dept. of Math. Dynamics UCLA [email protected] This joint work with A. Moussa, L. He and P. Zhang is concerned with a system that couples the incompressible Navier-Stokes equations to the Vlasov-Fokker-Planck equa- MS13 tion. Such a system arises in the modeling of sprays, where Global Recovery of High Frequency Wave Fields a dense phase interacts with a disperse phase. The cou- pling arises from the Stokes drag force exerted by a phase Computation of high frequency solutions to wave equations on the other. We establish the global-in-time existence of is important in many applications, and notoriously difficult classical solutions for data close to an equilibrium and we in resolving wave oscillations. Gaussian beams are asymp- investigate their long time behavior. The proofs use energy totically valid high frequency solutions concentrated on a estimates and the hypoelliptic structure of the system. single curve through the physical domain, and superposi- tion of Gaussian beams provides a powerful tool to generate Thierry Goudon more general high frequency solutions to PDEs. An alter- INRIA, Universite Lille 1 native way to compute Gaussian beam components such as [email protected] phase, amplitude and Hessian of the phase, is to capture them in phase space by solving Liouville type equations Ayman Moussa on uniform grids. In this work we review and extend re- CMLA, ENS Cachan, France cent constructions of asymptotic high frequency wave fields [email protected] from computations in phase space. We give a new level set method of computing the Hessian and higher derivatives Lingbing He of the phase. Moreover, we prove that the kth order phase Department of Mathematical Sciences, Tsinghua space based Gaussian beam superposition converges to the University 2 k − n original wave field in L at the rate of 2 4 in dimension Beijing, China [email protected] 66 PD09 Abstracts Ping Zhang ties in constructing such solution in three dimensions. Beijing Academy of Science [email protected] Changfeng Gui University of Connecticut [email protected] MS14 Large Time Behavior of Collisionless Plasma MS15 The motion of a collisionless plasma, a high-temperature, Limiting Problems for Ginzburg-Landau Models of low-density, ionized gas, is described by the Vlasov- Superconductivity in 3D Maxwell equations. In the presence of large velocities, rel- ativistic corrections are meaningful, and when magnetic In this paper we consider the asymptotic behavior of the effects are neglected this formally becomes the relativistic Ginzburg- Landau energy in 3-d, in various energy regimes. Vlasov-Poisson system. Similarly, if one takes the classical We derive applications to superconductivity and Bose- limit as the speed of light tends to infinity, one obtains the Einstein condensation: a general expression for the first classical Vlasov-Poisson system. We study the long-time critical field, and the curvature equation for the vortex dynamics of these systems of PDE and contrast the behav- lines. ior when considering the cases of classical versus relativistic velocities and the monocharged (i.e., single species of ion) Giandomenico Orlandi versus neutral plasma situations. Department of Computer Science University of Verona Stephen Pankavich [email protected] Univertity of Texas, Arlington [email protected] MS15 Global Solutions of Phase Transitions Equations MS14 Recent Global Results for the Relativistic Boltz- We study some geometric properties of global solutions of mann Equation Allen-Cahn-type equations. For homogeneous equations, we deal with rigidity and symmetry properties. For meso- We will discuss several recent results regarding the rel- scopic equations, we constuct solutions with interface close ativistic Boltzmann equation. The talk will start with to a plane or to a sphere and we discuss their multibump a broad overview of relativistic Kinetic theory for non- features. specialists. New results to be discussed include the pre- viously open problem of stability of the Maxwellian equi- Matteo Novaga librium for the relativistic Boltzmann equation with soft University of Padova (Italy) interactions. The soft potentials are important for parti- [email protected] cles moving at relativistic speeds. We can also prove for the first time the global validity of the Newtonian Limit Enrico Valdinoci in the near Vacuum regime. Additionally we can estab- Department of Mathematics lish the rigorous connection between the Boltzmann equa- University of Rome II, Tor Vergata tion and Relativistic Euler via a Hilbert Expansion, this is [email protected] joint work with Jared Speck. Furthermore, if time permits, we will discuss the relativistic Boltzmann equation coupled with it’s internally generated electric and magnetic forces. MS15 Despite its importance, no global in time solutions have Singular Limit of a Reaction-diffusion System De- been established so far for this Lorentz invariant model. scribing Combustion in Porous Media We prove existence of the first global in time classical so- lutions. This project is joint work with Yan Guo. Abstract not available at time of publication. Robert Strain Georg S. Weiss University of Pennsylvania Graduate School of Mathematical Sciences [email protected] University of Tokyo [email protected] MS15 MS16 Quadruple Junction Solutions in the Entire Three Dimensional Space A Couple of Regularity Problems for the Mhd Equations In this talk, I will discuss the quadruple junction solutions in the entire three dimensional space to a vector-valued First, we present our work on the global regularity of clas- Allen-Cahn equation which models multiple phase separa- sical solutions of the 2D MHD equations with mixed dis- tion. The solution is the basic profile of the local structure sipation and diffusivity. Then, we present two regularity near a quadruple junction in three dimensional crystalline criteria for the 3D incompressible HMD equations. One material under the generalized Allen-Cahn model, and is is in terms of the derivative of the velocity field in one- the three dimensional counterpart of triple junction solu- direction and the other requires suitable boundedness of tion which is two dimensional. I will start with one di- the derivative of the pressure in one-direction. mensional heteroclinic solutions, and describe how we can Chongsheng Cao construct higher dimensional solutions from the lower di- Department of Mathematics mensional ones, and explain the complications and difficul- Florida International University caoc@fiu.edu PD09 Abstracts 67 MS16 being a part of the evolution problem to check that such Motion of Viscous Fluid Around a Rotating Rigid property prevails for a short time, as well as it does the Body Rayleigh-Taylor condition, depending conveniently upon the initial data. The addition of the pressure equality to Abstract not available at time of publication. the contour dynamic equations is obtained as a mathemat- ical consequence, and not as a physical assumption, from Sarka Nekasova the mere fact that we are dealing with weak solutions of Academy of Sciences of the Czech Republic Euler’s equation in the whole space. [email protected] Francisco Gancebo The University of Chicago MS16 [email protected] The Quintic NLS as the Mean Field Limit of a Bo- son Gas with Three-body Interactions MS17 In this talk we will discuss the dynamics of a boson gas with Dispersive Properties of Surface Water Waves three-body interactions in dimensions d=1,2. We prove that in the limit as the particle number N tends to infinity, I will speak on the dispersive character of waves on the the BBGKY hierarchy of k-particle marginals converges interface between vacuum and water under the influence to a limiting Gross-Pitaevskii (GP) hierarchy for which of surface tension and possibly gravity. I will begin by we prove existence and uniqueness of solutions. For fac- giving a precise account of the formulation of the surface torized initial data, the solutions of the GP hierarchy are water-wave problem and discussion of its defining features. shown to be determined by solutions of a quintic nonlinear I will describe the local smoothing effect and the gain of Schr¨odingerequation. Time permitting, we will briefly de- regularity for the nonlinear water-wave problem with some scribe our new approach for studying well-posedness of the detail of the proofs. I will examine the frequency-localized Cauchy problem for focusing and defocusing GP hierarchy. and global dispersion of the linear system and indicate how the microlocal dispersion estimates gives the Strichartz es- timates for the nonlinear problem. If time permits, I will Thomas Chen, Natasa Pavlovic speak on the application of global dispersion on the long- University of Texas at Austin time existence. This work is partly joint with Hans Chris- [email protected], [email protected] tianson and Gigliola Staffilani (MIT). Vera Mikyoung Hur MS16 University of Illinois at Urbana-Champaign Inverse Problems for Some Transport Equations [email protected] Transport equations are used often to model various types of biophysical phenomena, including the propagation of MS17 light in tissues. In one limiting case one has the traditional How Gravity Can Stabilize Kelvin-Helmholtz In- X-ray tomography, which has been the subject of tremen- stabilities and Applications dous attention both for its applications and for its theoret- ical interest. The inverse problem, in the X-ray tomogra- The mathematical properties of the motion of the inter- phy case, is a linear one. Yet it is ill-posed and presents face between two layers of fluid differ dramatically from many computational challenges. In this talk discuss nonlin- what is observed for water-waves (which is a particular ear inverse problems related to transport equations within case of the two fluids problem with an upper fluid of zero the frame-work of the so called diffuse tomography mod- density). While the water-waves equations are well-posed, els. The work presented here originated in joint research with or without surface tension, on time scales that are with F.A. Grunbaum (UCB-Mathematics) and J.R. Singer pertinent for the observation of oceanographic phenomena, (UCB-EECS). Despite many recent progresses the inverse the ”interfacial waves” equations are known to be ill-posed problem associated to diffuse tomography still presents a without surface tension, due to Kelvin-Helmotz instabili- number interesting open problems. ties. It has been shown that adding surface tension allows the solution to exist locally in time. However, the exis- Jorge P. Zubelli tence time vanishes as the surface tension goes to zero and IMPA since the physical value of surface tension is very small, [email protected] this wellposedness result does not explain why we can ob- serve many stable manifestions of interfacial waves. In this talk, we will show that it is possible to control the growth MS17 of Kelvin-Helmoltz instabilities by a careful analysis of the Interface Evolution: Water Waves in 2-D role played by gravity and by the physical characteristics (depths of the layers, amplitude, etc.) of the flow. In We study the free boundary evolution between two irro- particular, we exhibit a new condition for the existence tational, incompressible and inviscid fluids in 2-D with- of ”stable” solutions to the ”interfacial waves” equations, out surface tension. We prove local-existence in Sobolev that can be seen as a generalization of both the Rayleigh- spaces when, initially, the difference of the gradients of Taylor criterion for the stability of surface waves and the the pressure in the normal direction has the proper sign, Chandrasekhar law on the instable modes in shear flows. an assumption which is also known as the Rayleigh-Taylor high-level commands. Do not include references or cita- condition. The well-posedness of the full water wave prob- tions separately at the end of the abstract. Instead, all lem was first obtained by Wu. The methods introduced citations must be in text in the general form [Authorname, in this paper allows us to consider multiple cases: with or Title, etc] without gravity, but also a closed boundary or a periodic boundary with the fluids placed above and below it. It David Lannes is assumed that the initial interface does not touch itself, 68 PD09 Abstracts Ecole Normale Superieure MS18 France Non Conforming Space-time Domain Decomposi- [email protected] tion Algorithms for Coupling Heterogeneous Prob- lems MS17 In this talk we present non conforming space-time domain Almost Global Wellposedness of the 2-D full Water decomposition algorithms for solving evolution problems Wave Equation with discontinuous coefficients. The objective is long time computations in highly discontinuous media such as nu- We consider the problem of global in time existence and clear waste disposal simulations, or climate modeling. The uniqueness of solutions of the 2-D infinite depth full water strategy is to split the time interval into time windows wave equation. It is known that this equation has a solution and to perform, in each window, few iterations of Opti- for a time period [0,T/ ] for initial data of type Φ, where mized Schwarz Waveform Relaxation methods [D. Ben- T depends only on Φ. We show that for such data there nequin, M.J. Gander, and L. Halpern., A homographic T/ exists a unique solution for a time period [0,e ]. This best approximation problem with application to optimized is achieved by better understandings of the nature of the Schwarz waveform relaxation, Math. of Comp., pages nonlinearity of the water wave equation. 259-266, 2009]. It has been proposed in [L. Halpern and C. Japhet, Discontinuous Galerkin and Nonconforming in Sijue Wu Time Optimized Schwarz Waveform Relaxation for Het- University of Michigan at Ann Arbor erogeneous Problems, In U. Langer, M. Discacciati, D.E. [email protected] Keyes, O.B. Widlund, and W. Zulehner, editors, Decom- position Methods in Science and Engineering XVII, volume 60 of Lecture Notes in Computational Science and Engi- MS18 neering, pages 211-219. Springer, 2008] to use a discon- Domain Decomposition for Heterogeneous, Multi- tinuous Galerkin method in time as a subdomain solver scale Problems in order to have a high degree of accuracy, time-stepping approaches, and ultimately adaptive control of the time Multiscale methods are used to solve flow problems in step, see [C. Johnson, K. Eriksson, and V. Thomee, Time porous media with heterogeneous permeability. We de- discretization of parabolic problems by the discontinuous velop multiscale mortar mixed finite element discretiza- Galerkin method, RAIRO Modl. Math. Anal. Numr., 19, tions for second order elliptic equations. The continuity 1985], [Ch. Makridakis and R. Nochetto, A posteriori error of flux is imposed via a mortar finite element space on a analysis for higher order dissipative methods for evolution coarse grid scale, while the equations in the coarse elements problems, Numerische Mathematik, 2007]. We present the (or subdomains) are discretized on a fine grid scale. The analysis of the method and numerical results to illustrate mortar approximation space is based on homogenization, the performances of the method. and the method is shown to converge with respect to this microstructure. Laurence halpern Universit´eParis 13 Todd Arbogast [email protected] Dept of Math; C1200 University of Texas, Austin [email protected] Caroline Japhet Paris 13 university, France [email protected] MS18 On the Coupling of Hyperbolic and Elliptic Equa- J´er´emieSzeftel tions: Comparison Between Virtual Control Meth- Princeton University and Bordeaux 1 University ods and Extended Variational Formulations [email protected] We address the coupling of an advection equation with a diffusion-advection equation, for solutions featuring MS18 boundary layers. We consider both non-overlapping and Filtration Into Deforming Porous Media overlapping domain decompositions and we face up the het- erogeneous problem by different approaches. Starting from We review the theory of flow through deformable porous a rigorous variational analysis, the aim of this research is media. Then we describe more recent extensions to com- to investigate if other approaches may be considered to posite media, partially-saturated flow, flow through rather formulate this heterogeneous problem, for a possible ex- general poro–inelastic media, and coupling to the Stokes tension to more general couplings. In particular we will flow in an adjacent open channel. The problems will be consider the virtual control approach, based on the opti- posed and resolved in appropriate variational formulations. mal control theory, and an extend variational formulation recently proposed for geometric multiscale problems. Ralph Showalter Paola Gervasio Department of Mathematics Dipartimento di Matematica,Facolt`adi Ingegneria Oregon State University Universit`adegli Studi di Brescia, Brescia, Italy [email protected] [email protected] MS19 Alfio Quarteroni Ecole Pol. Fed. de Lausanne Convergence to Homogenized and Stochastic PDEs Alfio.Quarteroni@epfl.ch Equations with small scale structures abound in applied sciences. Many such structures cannot be modeled at a mi- PD09 Abstracts 69 croscopic level and thus require that one understand their MS19 macroscopic influence. I will consider the situation of par- Homogenization of Hamilton-Jacobi Equations in tial differential equations with random, highly oscillatory, Time-Space Random Environments potential. One is then interested in the behavior of the solutions to that equation as the frequency of oscillations We present results on the stochastic homogenization of in the micro-structure tends to infinity. Depending on spa- Hamilton-Jacobi equations in time-space random environ- tial dimension and the decorrelation properties of the ran- ments. Previous results of Souganidis, 1999, and simul- dom potential, I will show that the limit is the solution taneously by Rezakhanlou and Tarver, 2000, had been in to either a deterministic, homogenized (effective medium) the context of random fluctuations in space only. Special equation or a stochastic equation with multiplicative noise attention will be placed on the interplay between the use that should be understood in the sense of a Stratonovich of the Subadditive Ergodic Theorem and continuity esti- product. More precisely, there is a critical spatial dimen- mates for solutions of these equations that depend only on sion above which we observe convergence to a determinis- the growth of the Hamiltonian. tic solution and below which we observe convergence to a stochastic solution. In the former case, a theory of correc- Russell Schwab tors to homogenization allows one to capture the remaining Carnegie Mellon University randomness in the solution to the equation with the small [email protected] scale structure. Once properly rescaled, this corrector is shown to solve a stochastic equation with additive noise. MS20 Guillaume Bal Generalized Newton-type Methods for Energy For- Columbia University mulations in Image Processing Department of Applied Physics & Applied Mathematics [email protected] Many problems in image processing are addressed via the minimization of a cost functional. The most prominently used optimization technique is gradient-descent, often used MS19 due to its simplicity and applicability where other tech- SPDE Limits of the Random Walk Metropolis Al- niques, e.g., those coming from discrete optimization, can gorithm in High Dimensions not be applied. Yet, gradient-descent suffers from slow con- vergence, and often to just local minima which highly de- Diffusion limits of MCMC methods in high dimensions pro- pend on the initialization and the condition number of the vide a useful theoretical tool for studying efficiency. In functional Hessian. Newton-type methods, on the other particular they facilitate precise estimates of the number hand, are known to have a faster, quadratic, convergence. of steps required to explore the target measure, in station- In its classical form, the Newton method relies on the L2- arity, as a function of the dimension of the state space. type norm to define the descent direction. In this work, we However, to date such results have only been proved for generalize and reformulate this very important optimiza- target measures with a product structure, severely limit- tion method by introducing Newton-type methods based ing their applicability to real applications. I will present on more general norms. Such norms are introduced both a study diffusion limits for a class of naturally occuring in the descent computation (Newton step), and in the high dimensional measures and their associated MCMC corresponding stabilizing trust-region. This generalization algorithm. The diffusion limit to an infinite dimensional opens up new possibilities in the extraction of the New- Hilbert space valued SDE (or SPDE) will be proved. ton step, including benefits such as mathematical stability and the incorporation of smoothness constraints. We first Jonathan C. Mattingly present the derivation of the modified Newton step in the Duke University calculus of variation framework needed for image process- [email protected] ing. Then, we demonstrate the method with two common objective functionals: variational image deblurring and ge- ometric active contours for image segmentation. We show MS19 that in addition to the fast convergence, norms adapted to Unbiased Random Perturbations of Navier-Stokes the problem at hand yield different and superior results. Equation. Leah Bar A random perturbation of a deterministic Navier-Stokes University of Minnesota equation is considered in the form of an Stochastic PDE [email protected] with Wick product in the nonlinear term. The equation is solved in the space of generalized stochastic processes using the Cameron-Martin version of the Wiener chaos ex- MS20 pansion. The generalized solution is obtained as an inverse From TV-L1 Model to Continuous Graph-Cuts: of solutions to corresponding quantized equations. An in- Global & Fast Optimization Algorithms for Image teresting feature of this type of perturbation is that it pre- & Geometry Processing serves the mean dynamics: the expectation of the solution of the perturbed equation solves the underlying determin- In this talk, I will review some recent developments in istic Navier-Stokes equation. From the stand point of a global continuous optimization methods for image and ge- statistician it means that the perturbed model is unbiased. ometry processing. Starting from the TV-L1 model for The talk is based on a joint work with R. Mikulevicius. image denoising, we will show that this model has opened the path to a new paradigm to define convex formulations Boris Rozovsky for e.g. image segmentation and geometry reconstruction. Brown University Indeed, unlike state-of-the-art methods, the proposed con- Boris [email protected] vex energy minimization approach allows to solve geome- try problems independently of initial condition. Besides, this approach provides fast optimization algorithms, bor- 70 PD09 Abstracts rowed from operator splitting techniques, to solve energy and alternating direction methods. Recent results on the minimization problems related with non-linear hyperbolic Bregman methods are included. Applied to 1-type regu- PDEs. We will show some results in segmentation and sur- larization, the original Bregman method has an exact error face reconstruction. Joint work with T.F. Chan, S. Osher, cancellation property that leads to machine-precision solu- T. Goldstein (UCLA). tions. In addition, we briefly show an exact penalty prop- erty of the linearized Bregman method. These properties Xavier Bresson explain the great performance of the Bregman methods. UCLA The analysis and discussions broadly apply to polyhedral Department of Mathematics regularization functions such the (weighted) 1-norm, total [email protected] variation, piece-wise linear functions, as well as the matrix nuclear norm. MS20 Wotao Yin Coarsening in High Order, Discrete, Ill-posed Dif- Rice University fusion Equations [email protected] We study the discrete version of a family of ill-posed, non- linear diffusion equations of order 2n. The fourth order MS21 (n = 2) version of these equations constitutes our main mo- Radial Symmetry of Positive Solutions to Nonlin- tivation, as it appears prominently in image processing and ear Polyharmonic Dirichlet Problems computer vision literature. It was proposed by You and Kaveh as a model for denoising images while maintaining We extend the celebrated symmetry result of Gidas-Ni- sharp object boundaries (edges). The second order equa- Nirenberg to semilinear polyharmonic Dirichlet problems tion (n = 1) corresponds to another famous model from in the unit ball. In the proof we develop a new variant of image processing, namely Perona and Maliks anisotropic the method of moving planes relying on fine estimates for diffusion, and was studied previously. The equations pre- the Green function of the polyharmonic operator. We also sented in this talk are high order analogues of the Perona- consider minimizers for subcritical higher order Sobolev Malik equation, and like the second order model, their con- embeddings. For embeddings into weighted spaces with tinuum versions violate parabolicity and hence lack well- a radially symmetric weight function, we show that the posedness theory. As in the previous work, we follow a minimizers are at least axially symmetric. This result is recent technique by Kohn and Otto to establish rigorous sharp since we exhibit examples of minimizers which do upper bounds on the coarsening rate of the discrete ver- not have full radial symmetry. The lecture is based on joint sions of these high order equations, in any space dimension work with E. Berchio (Milan) and T. Weth (Frankfurt). and for a large class of diffusivities. However, the high order nature of our equations requires different arguments Filippo Gazzola and constructions from before. Dipartimento di Matematica - Politecnico di Milano fi[email protected] Catherine Kublik University of Michigan Tobias Weth [email protected] Institut f¨urMathematik Johann Wolfgang Goethe-Universit¨atFrankfurt - Selim Esedoglu Germany Department of Mathematics [email protected] University of Michigan [email protected] Elvise Berchio Dipartimento di Matematica MS20 Politecnico di Milano - Italy [email protected] Geodesics in Level-set Based Shape Space The equations for geodesic movement (with respect to MS21 an appropriately constructed metric) of a level-set func- tion representing a shape variable are derived. The re- ”Almost Positivity”’ in the Fourth Order Clamped lated equations for parallel transport are also constructed. Plate Equation Geodesic movement and parallel transport are used to re- A classical example for a fourth order problem in mechan- alize algoritmic concepts for non-linear optimization in the ics is the linear clamped plate equation. It is known that non-linear level-set based shape space. it does not satisfy any form of a general maximum or pos- Wolfgang Ring itivity preserving (comparison) principle. However, there Institut fuer Mathematik are positivity issues as e.g. “almost positivity’ to be dis- Universitaet Graz cussed. In particular, it is shown that the negative part [email protected] of the corresponding Green function obeys much sharper bounds than the singular positive part. For example, if n>4, the negative part is bounded while the positive part 4−n MS20 displays the singular behaviour of |x − y| . A Review of the Bregman Methods Hans-Christoph Grunau We review four Bregman methods that arise in applica- Institut fur Analysis und Numerik tions of imaging, compressed sensing, and other inverse Otto-von-Guericke-Universitat, Magdeburg - Germany problems. We explain their numerical properties and high- [email protected] light their relationships with the augmented Lagrangian Frederic Robert PD09 Abstracts 71 Laboratoire J.-A.Dieudonne Laws Universite de Nice-Sophia Antipolis [email protected] Abstract not available at time of publication. Constantine Dafermos Guido Sweers Brown Univ. Mathematisches Institut constantine [email protected] Universitat zu Koln - Germany [email protected] MS22 MS21 New Models of Chemotaxis: Analysis and Numer- ics Q-curvature Equation in Conformal and CR Ge- ometry Patlak-Keller-Segel (PKS) system is a classical PDE model of the chemotaxis. The system admits solutions that de- In this talk, I plan to outline the role this 4th order equa- velop delta-type singularities within a finite time. Even tion plays in the conformal geometry of four dimension, as though such blowing up solutions model a concentration well as the CR geometry in 3-dimension. In CR geometry, phenomenon, they are not realistic since biological cells there is a natural 4th order operator which plays an impor- do not converge to one point (while the cell density grows tant rolein the analogue of the Bochner formula governing sharply, it must remain bounded at all times). I will present the kernel of the sub-Laplacian. a new chemotaxis model, which can be viewed as a regu- Paul Yang larized PKS system. The proposed regularization is based Department of Mathematics on a basic physical principle: boundedness of the chemo- Princeton University, USA tactic convective flux, which should depend on the gradi- [email protected] ent of the chemoattractant concentration in a nonlinear way. Solutions of the new system may develop spiky struc- tures that model the concentration phenomenon. However, MS21 both cell density and chemotattractant concentration re- Fourth Order Elliptic Problems Involving Singular main bounded as supported by both our analytical results and Critical Nonlinearities and extensive numerical experiments. We study a class problems for the biharmonic operator Alina Chertock with Dirichlet boundary conditions in dimension four in- North Carolina State University volving critical growth corresponding to a Trudinger-Moser Department of Mathematics embedding. [email protected] Joao Marcos do O Alexander Kurganov Universidade Federal da Paraba Tulane University [email protected] Department of Mathematics [email protected] MS22 Xuefeng Wang On the Classical Solutions of Two Dimensional In- Tulane University viscid Rotating Shallow Water System [email protected] We prove the global existence and asymptotic behavior of the classical solutions for two dimensional inviscid Rotat- MS22 ing Shallow Water system with small initial data subject A Low Mach Number Limit from a Dispersive to the zero-relative-vorticity constraint. One of the key Navier-Stokes System ¿ to a Ghost-Effect System steps is a reformulation of the problem into a symmetric quasilinear Klein-Gordon system, whose existence of the A low Mach number limit for classical solutions of a dis- classical solutions is then proved with combination of the persive system beyond the full Navier-Stokes equations is vector field approach and the normal forms, adapting ideas investigated. We prove the convergence of this dispersive developed in [1]. The symmetric form of the system facili- system to a ghost effect system studied extensively by Sone tates the closure of energy estimates for the problem. We in the whole space case. The combined effects of large also probe the case of general initial data and reveal a lower temperature variations, thermal conduction as well as the bound for the lifespan that is almost inversely proportional dispersive structure are taken into account. to the size of initial relative vorticity. [1] Tohru Ozawa, Kimitoshi Tsutaya, and Yoshio Tsutsumi, Global existence Weiran Sun and asymptotic behavior of solutions for the Klein-Gordon Department of Mathematics equations with quadratic nonlinearity in two space dimen- University of Chicago sions, Math. Z. 222 (1996), no. 3, 341362. [email protected] Bin Cheng, Chunjing Xie University of Michigan MS23 [email protected], [email protected] Well-posedness and Mean-field Limit for Kinetic Models of Collective Motion MS22 There has been a recent interest in models for the com- Continuous Solutions of Hyperbolic Conservation plex behavior of large groups of animals, such as flocks of birds, swarms, or schools of fish. Kinetic equations, which 72 PD09 Abstracts involve the time evolution of the distribution function of lutions. An interesting feature of the model is that the individuals depending on their position and velocity, are ”trivial” steady state characterized by the uniform density tentative models for these phenomena, and raise many in- remains locally stable in the subcritical region while for teresting mathematical questions. In this work, we give sufficiently large perturbations the dynamics prefers other, a well-posedness theory for kinetic PDE models for the spatially nonuniform stady states. collective behavior of a group of individuals that interact through a pair potential and/or try to adapt their velocity Vladislav Panferov to the average velocity of their neighbours, possibly includ- California State University, Northridge ing also self-propulsion effects. The aim is to give a solid Department of Mathematics mathematical basis for the study of these models and the [email protected] range of phenomena they display. Jose Alfredo Canizo MS24 Universitat Autonoma de Barcelona Entire Solutions of the Allen Cahn Equation in the [email protected] Hyperbolic Space Qualitative properties of entire solutions of equations with MS23 double well potential in the Hyperbolic space will be dis- Accuracy and Consistency of Spectral-Lagrangian cussed, both in relation with conditions at the asymptotic Approximating Methods for the Boltzmann Equa- boundary and with conditions of stability or minimality. tion Differences and similarities with respect to the euclidean case will be discussed. We study the accuracy and consistency properties of the re- cent Lagrangian based spectral method to produced an ap- Isabeau Birindelli proximating scheme to solutions of the Boltzmann equation Sapienza Universit`adi Roma [1,2]. We develop bounds on the optimization correction in [email protected] terms of spectral accuracy and also prove consistency and conservation properties of the scheme. We also obtain er- ror estimate for the conservation routine implemented as an MS24 extended isoperimetric problem with the moment conser- A Moving Boundary Problem Involving Surface vation properties as the constraints. Combining Sobolev Tension estimates of the collisional integral with projection error and conservation correction estimates yields a Lagrange- Abstract not available at time of publication. spectral error result. Nestor Guillen Irene M. Gamba Department of Mathematics University of Texas at Austin Department of Mathematics and ICES [email protected] University of Texas [email protected] MS24 Stability in the Stefan Problem with Surface Ten- Sri Harsha Tharkabhushanam sion University of Texas at Austin [email protected] Abstract not available at time of publication. Mahir Hadzic MS23 Brown University Fractional Diffusion Limits for Kinetic Equations [email protected] We will discuss some diffusion limits for linear Boltzmann equations when the equilibrium distribution function is not MS24 a Maxwellian distribution. In particular, we will show that Domain Walls and Triple Junctions for an appropriate time scale, the small mean free path limit gives rise to fractional diffusion equations. The studies of domain walls and triple junctions arise in many applications. Their appearences , such as honeycomb Antoine Mellet structures in 2D, are deceptively simple. If the underlying University of Maryland variational energy driven such a structure is giving by the [email protected] Surface Area, it is then possible to describe some geometric and analytic properties of the structure via the classical minimal surfaces theory. In many applications, however, MS23 such a structure is driven by the so-called stress, that is, Phase Transitions for the McKean-Vlasov Equation the Bulk energy, then not much were known before. In in Finite Volume this lecture, I shall describe some recent progress on the Bulk-energy driven domain walls and multiple junctions. We consider a nonlinear parabolic equation that describes the mean-field limit for a system of N diffusions in a Fang-Hua Lin bounded domain interacting through a finite-range poten- Courant Institute of Mathematical Sciences tial V. We show that for all potentials that are not of pos- [email protected] itive type (characterized by the positivity of the Fourier transform of V) the system generically has a first-order phase transition at the critical temperature, which is man- MS25 ifested by the nonuniqueness/instability of the steady so- Global Bv Solutions to Heat Flows of Linear PD09 Abstracts 73 Growth Maps into Spheres Free Boundary In this talk I shall first present some recent developments We study evolution of simple planar curves driven by the for p-harmonic map heat flows into spheres, in particular, weighted mean curvature flow (wmc) for 1-harmonic map heat flow. The highlight of the devel- opments is the BV (bounded variation) solution theory for βV = κγ + σ the flow, which will be reviewed in detail. I shall then dis- cuss the extension of these results to heat flows for a class when the anisotropy function is piecewise linear and con- of linear growth maps. Finally, applications (and their nu- vex. The diffusion in certain direction is so strong that merical approximations) of these geometric flows to color facets are created. At the same time in other directions image denoising based on the chromaticity and brightness the equation degenerates to a first order equation. In or- decomposition will also be discussed. The results presented der to prove existence and uniqueness of solutions to the in this talk are based on recent joint works with John Bar- resulting problem we develop a theory of viscosity solutions rett of Imperial College (UK), Andreas Prohl of University to Hamilton-Jacobi equation with discontinuous Hamilto- of Tuebengin (Germany), and Minun Yoon of University of nian. In particular, we show a comparison principle and a Tennessee. stability theorem. Xiaobing H. Feng Piotr Rybka The University of Tennessee Warsaw University [email protected] Mathematics Department [email protected] MS25 Yoshikazu Giga Local and Nonlocal Weighted p-Laplacian Evolu- Graduate School of Mathematical Sciences tion Equations with Neumann Boundary Condi- University of Tokyo, Japan tions [email protected] In this talk we study existence and uniqueness for solutions Przemys law G´orka of the nonlocal diffusion equation with Neumann boundary Instituto de Matem conditions { } ’ a tica y F ’{i}sica x + y p−2 ut(t, x)= J(x−y)g |u(t, y)−u(t, x)| (u(t, y)−u(Universidadt, x)) dy de Talca Ω 2 [email protected] in ]0,T[×Ω, and for solutions of its local counterpart p−2 MS26 ut = div (g|∇u| ∇u) in]0,T[×Ω, p−2 Expanding the Prodi-Serrin Critera for Regularity |∇ | ∇ · × p g u u η in]0,T[ ∂Ω. to Weak-L Spaces We consider 1 ≤ p<∞ and g ≥ 0. We pay special atten- In this talk we discuss weakening the Prodi-Serrin criteria tion to the case in which g vanishes somewhere in Ω even for regularity of Leray-Hopf solutions to the 3d Navier- in a set of positive measure. Joint work with: F. Andreu, Stokes equation. In particular, the criteria is expanded to J. Rossi and J. Toledo include solutions which satisfy the following bound: