program and abstract book - xiamen universitymath.xmu.edu.cn/conference/sfcam/abstracts.pdf ·...
TRANSCRIPT
Sino-French Conference on Computational and Applied Mathematics
1
Abstracts
Recent progress on very high order schemes for
compressible flows
Speaker: Remi Abgrall, University of Zurich
email: [email protected]
Abstract: We are interested in the development of very high order schemes
for compressible fluid dynamics governed by the Navier-Stokes equations.
These approximations use unstructured conformal meshes.
In the talk, we will describe a family of nonlinear schemes that are high
order and non-oscillatory. We will show extensions to the Navier Stokes
equations, and also how the unsteady case can be handled.
Several examples including two and three dimensional flows will be
discussed, showing the performance of the method.
On some recent advances for the indirect
controllability of coupled systems of PDE's
Speaker: Fatiha Alabau-Boussouira, University of Lorraine
email: [email protected]
Abstract: We consider controlled coupled systems of reversible hyperbolic
PDE’s. A challenging issue for such systems is to understand whether if
controllability can hold when the number of controls is strictly less than the
number of equations. If the answer is positive, one says that the equations
which are free of controls are indirectly controlled.
Such questions naturally arise in the built-in of in sensitizing controls for
scalar PDE’s, in simultaneous controllability or in engineering applications.
We present some recent advances for the indirect control of systems
coupled in cascade based on a multi-levels energy method. We also give
recent results showing the importance of the coupling operator properties
leading to positive as well as negative observability, controllability and unique
continuation results.
Sino-French Conference on Computational and Applied Mathematics
2
Geometrical constraints in the level-set method for
shape and topology optimization
Speaker: Gregoire Allaire, Ecole Polytechnique
email: [email protected]
Abstract: In the context of structural optimization via a level-set method we
propose a framework to handle geometric constraints related to a notion of
local thickness. The local thickness is calculated using the signed distance
function to the shape. We formulate global constraints using integral functions
and compute their shape derivatives. We discuss different strategies and
possible approximations to handle the geometric constraints. We implement
our approach in two and three space dimensions for a model of linearized
elasticity. As can be expected, the resulting optimized shapes are strongly
dependent on the initial guesses and on the specific treatment of the
constraints since, in particular, some topological changes may be prevented
by those constraints. This is a joint work with Francois Jouve and Georgios
Michailidis.
Inverse problems in wave propagation
Speaker: Gang Bao, Zhejiang University
email: [email protected]
Abstract: Our recent progress in mathematical analysis and computational
studies of the inverse boundary value problems in wave propagation will be
reported. Several classes of inverse problems will be studied, namely inverse
medium problems, inverse source problems, inverse obstacle problems,
inverse rough surface scattering problems, and inverse waveguide problems.
Issues on ill-posedness and its remedy for the inverse problems will be
addressed. Computational methods will be discussed for the inverse problems.
New stability results for the inverse problem will be presented. In particular, our
most recent stability result on inverse problems for the related time-domain
wave equation with possible caustics will be highlighted.
Sino-French Conference on Computational and Applied Mathematics
3
Multiscale methods and analysis for the nonlinear
Klein-Gordon equation in the nonrelativistic limit
regime
Speaker: Weizhu Bao, National University of Singapore
email: [email protected]
Abstract: In this talk, I will review our recent works on numerical methods and
analysis for solving the nonlinear Klein-Gordon (KG) equation in the
nonrelativistic limit regime, involving a small dimensionless parameter which is
inversely proportional to the speed of light. In this regime, the solution is highly
oscillating in time and the energy becomes unbounded, which bring significant
difficulty in analysis and heavy burden in numerical computation. We begin
with four frequently used finite difference time domain (FDTD) methods and
obtain their rigorous error estimates in the nonrelativistic limit regime by paying
particularly attention to how error bounds depend explicitly on mesh size and
time step as well as the small parameter.
Then we consider a numerical method by using spectral method for spatial
derivatives combined with an exponential wave integrator (EWI) in the
Gautschi-type for temporal derivatives to discretize the KG equation. Rigorious
error estimates show that the EWI spectral method show much better temporal
resolution than the FDTD methods for the KG equation in the nonrelativistic
limit regime. In order to design a multiscale method for the KG equation, we
establish error estimates of FDTD and EWI spectral methods for the nonlinear
Schrodinger equation perturbed with a wave operator. Finally, a multiscale
method is presented for discretizing the nonlinear KG equation in the
nonrelativistic limit regime based on large-small amplitude wave
decomposition. This multiscale method converges uniformly in
spatial/temporal discretization with respect to the small parameter for the
nonlinear KG equation in the nonrelativistic limite regime. Finally, applications
to several high oscillatory dispersive partial differential equations will be
discussed.
Sino-French Conference on Computational and Applied Mathematics
4
On the master equation in mean field theory
Speaker: Alain Bensoussan, University of Texas & University Hong Kong
email: [email protected]
Abstract: One of the major founders of Mean Field Games, P.L. Lions has
introduced in his lectures at College de France the concept of Master Equation.
It is obtained through a formal analogy with the set of partial differential
equations derived for the Nash equilibrium of a differential game with a large
number of players. The objective of this lecture is to explain its derivation, not
by analogy, but through its interpretation. We do that for both Mean Field Type
Control and Mean Field Games. We obtain complete solutions in the linear
quadratic case. We analyze the connection with Nash equilibrium.
When it rains on a sand beach: the Richards equations
Speaker: Christine Bernardi, Pierre and Marie Curie University
email: [email protected]
Abstract: Richards equation models the water flow in a partially saturated
underground porous medium under the surface. When it rains on the surface,
boundary conditions of Signorini type must be considered on this part of the
boundary. We first study this problem which results into a variational equality or
inequality according to the type of boundary conditions. We propose and study
a discretization by an implicit Euler's scheme in time and finite elements in
space.
The initial- and boundary-value problem for the
transport equation with low regularity data
Speaker: Franck Boyer, Aix-Marseille University
email: [email protected]
Abstract: The first part of the talk will be dedicated to well-posedness results
for the initial- and boundary-value problem for the transport equation
associated with a velocity field possessing the same Sobolev regularity as the
one required in the usual DiPerna-Lions theory. I will in particular discuss trace
theorems adapted to this framework and some applications.
In a second part, I will briefly describe some convergence results, uniform
in time, for the upwind finite volume approximation of this problem on general
unstructured grids, without any additional regularity assumptions on the data
Sino-French Conference on Computational and Applied Mathematics
5
Analysis and control of 2D turbulence
Speaker: Charles-Henri Bruneau, University of Bordeaux
email: [email protected]
Abstract: We first show what are the physical structures responsibles of the
two turbulence cascades in two-dimensions. Then we explore the ability to
control the drag coefficient of a blunt body using polymers in solution. The
numerical solutions of Navier-Stokes equations and Oldroyd-B model are
compared to soap films experiments.
Exact divergence free H(div) basis and Feymann-kac
formula based boundary element methods
Speaker: Wei Cai, University of North Carolina at Charlotte & Shanghai
Jiaotong University
email: [email protected]
Abstract: In this talk, we will discuss two results on numerical methods for
solving PDEs, one on designing high order divergence free basis for MHD
equations and another on Feymann-Kac formula based boundary element
method for elliptic PDEs. In the first, we give the construction of a new high
order hierarchical basis functions in the H(div) space and show how to use
interior bubble functions in imposing exact divergence-free conditions for MHD
problems. While for the second, we will present a communication-free domain
decomposition boundary element method for Laplace equations by combining
Feyman-Kac formula and local time of Brownian motions and local integral
equations.
Sino-French Conference on Computational and Applied Mathematics
6
A Legendre-Galerkin spectral method for optimal
control problems governed by Stokes equations
Speaker: Yanping Chen, South China Normal University
email: [email protected]
Abstract: In this work, we study the Legendre–Galerkin spectral
approximation of distributed optimal control problems governed by Stokes
equations. We show that the discretized control problems satisfy the
well-known Babuska–Brezzi conditions by choosing an appropriate pair of
discretization spaces for the velocity and the pressure. Constructing suitable
base functions of the discretization spaces leads to sparse coefficient matrices.
We first derive a priori error estimates in both H^1 and L^2 norms for the
Legendre–Galerkin approximation of the unconstrained control problems.
Then both a priori and a posteriori error estimates are obtained for control
problems with the constraints of an integral type, thanks to the higher regularity
of the optimal control. Finally, some illustrative numerical examples are
presented to demonstrate the error estimates.
Reverse time migration for reconstructing extended
obstacles in planar acoustic waveguides
Speaker: Zhiming Chen, LSEC & Institute of Mathematics and System
Sciences, CAS
email: [email protected]
Abstract: We propose a new reverse time migration method for reconstructing
extended obstacles in the planar waveguide using acoustic waves at a fixed
frequency. We prove the resolution of the reconstruction method in terms of
the aperture and the thickness of the waveguide. The resolution analysis
implies that the imaginary part of the cross-correlation imaging function is
always positive and thus may have better stability properties. Numerical
experiments are included to illustrate the powerful imaging quality and to
confirm our resolution results. This is a joint work with Guanghui Huang.
Sino-French Conference on Computational and Applied Mathematics
7
The mathematical model for the contamination
problems and related inverse problems
Speaker: Jin Cheng, Fudan University
email: [email protected]
Abstract: In this talk, we discuss the motivation for the study of the abnormal
diffusion models for the contamination problems. From the practical point of
view, several inverse problems proposed and studied Theoretic results, for
example, the uniqueness and stability are shown. The possibility of the
application of these studies is mentioned.
A new approach to elasticity problems and their finite
element discretizations
Speaker: Philippe G. Ciarlet, City University of Hong Kong
email: [email protected]
Abstract: We describe and analyze an approach to the pure Neumann problem
of three-dimensional linearized elasticity, whose novelty consists in
considering the strain tensor field as the sole unknown, instead of the
displacement vector field as is customary. This approach leads to a well-posed
minimization problem of a new type, constrained by a weak form of the
classical Saint Venant compatibility conditions, the justification of which
essentially rests on J.L. Lions lemma. Interestingly, this approach also
provides a new proof of Korn's inequality.
We also describe and analyze a natural finite element approximation of this
problem.
Sino-French Conference on Computational and Applied Mathematics
8
Mathematical models for tumor growth: construction,
validation and clinical applications
Speaker: Thierry Colin, Bordeaux Institute of Technology
email: [email protected]
Abstract: In the last few years there have been dramatic increases in the
range and quality of information available from non-invasive medical imaging
methods, so that several potentially valuable imaging measurements are now
available to quantitatively measure tumor growth, assess tumor status as well
as anatomical or functional details. Using different methods such as the CT
scan, magnetic resonance imaging (MRI), or positron emission tomography
(PET), it is now possible to evaluate and define tumor status at different levels:
physiological, molecular and cellular.
These multimodal data help the decision process of oncologists in the
definition of therapeutic protocols. At present, this decision process is mainly
based on previously acquired statistical evidence and on the practitioner
experience. The quality of the response to a treatment is decided according to
the OMS criteria by estimating the length of the two main axis of the tumor in
the largest cut. There are two blocking difficulties in this approach that we
want to attack: i) previous statistical information is not patient specific; ii) there
exist no quantitative mean of summarizing and using as predicting tools the
multimodal patient-specific data presently available thanks to CT scans, MRI,
PET scans and molecular biology data. The aim of this talk is to show how we
can provide a simulation framework based on quantitative patient-specific data
by using nonlinear model based on PDE. I will present how one can build such
model in order to describe the may features of tumor growth and how one can
expect to obtain a suitable parametrization of tumor growth by solving inverse
problem. I will present some mathematical results on the models.
The applications will concern lung and lever metastasis, meningiomas and
brain tumor.
Sino-French Conference on Computational and Applied Mathematics
9
On the control of the Korteweg-de Vries equation
Speaker: Jean-Michel Coron, Pierre and Marie Curie University
email: [email protected]
Abstract: The Korteweg-de Vries equation allows to describe approximately
long waves in water of relatively shallow depth. We study this equation on a
finite interval. The control is acting on the boundary of this interval. Depending
on these boundary conditions we get different control systems. We present
some results and methods to study the controllability and the stabilization of
these control systems.
Some Korn inequalities originating in a periodic
homogenization problem of linearized elasticity with
inclusions and contact conditions
Speaker: Alain Damlamian, University Paris-Est
email: [email protected]
Abstract: In the study of the homogenization of a contact problem in linearized
elasticity with inclusions, some new unilateral Korn inequalities are needed to
establish the coercivity of the functional to be minimized. The talk will mainly
explain how to prove various versions of the Korn inequality in this context.
The type of inequality is strongly dependent on the fact that the boundary of
each inclusion is invariant under some group of rotations or not.
Sino-French Conference on Computational and Applied Mathematics
10
An AMR method for systems arising in multifluid
dynamics
Speaker: Jean-Michel Ghidaglia, ENS Cachan
email: [email protected]
Abstract: Computing transient solutions with very strong spatial gradients is a
well known challenge in numerical analysis. Automatic Mesh Refinement
(AMR) is a way for addressing this question. It is indeed intuitively clear that, if
we can dynamically adapt the mesh in order that spatial regions where the
gradient are large are well resolved, then the computed solution will be more
accurate.
In this talk we describe the derivation of an AMR method for a non
conservative system of equations arising in CmFD (Computational multi Fluid
Dynamics). Then we shall present some numerical results on the classical
Ransom's faucet benchmark.
Eddy currents nondestructive evaluation
Speaker: Olivier Goubet, University of Picardie Jules Verne
email: [email protected]
Abstract: Eddy current nondestructive evaluation is commonly used to track a
defect inside a conductor. Eddy currents are a quasi-static approximation of
Maxwell equations. We address here a mathematical method to solve these
equations as a transmission problem at the interface of the conductor and the
coil that provides the electromagnetic source.
Sino-French Conference on Computational and Applied Mathematics
11
A second-order maximum principle preserving
continuous finite element technique for nonlinear
scalar conservation equations
Speaker: Jean Luc Guermond, Texas A&M University
email: [email protected]
Abstract: In the first part of the talk I will introduces a first-order viscosity
method for the explicit approximation of scalar conservation equations with
Lipschitz fluxes using continuous finite elements on arbitrary grids in any
space dimension. Provided the lumped mass matrix is positive definite, the
method is shown to satisfy the local maximum principle under a usual CFL
condition. The method is independent of the cell type; for instance, the mesh
can be a combination of tetrahedra, hexahedra, and prisms in three space
dimensions. An a priori convergence estimate is given provided the initial data
is BV.
In the second part of the talk I will extend the accuracy of the method to
second-order (at least). The technique is based on mass-lumping correction, a
high-order entropy viscosity method, and the Boris-Book-Zalesak flux
correction technique. The algorithm works for arbitrary meshes in any space
dimension and for all Lipschitz fluxes. The formal second-order accuracy of the
method and its convergence properties are tested on a series of linear and
nonlinear benchmark problems.
Some results on spectral element method
Speaker: Benyu Guo, Shanghai Normal University
email: [email protected]
Abstract: In this talk, we present the basic results on the generalized Jacobi
quasi-orthogonal approximation, which plays an important role in spectral
element methods. As examples of its applications, we consider the spectral
element methods for high order problems in one dimension, fourth order
problems in two dimensions and some problems defined on non-rectangular
domains. The numerical results indicate the high accuracy of suggested
algorithms.
Sino-French Conference on Computational and Applied Mathematics
12
Analysis and numerical methods in RTE based
medical imaging
Speaker: Weimin Han, University of Iowa & Xi'an Jiaotong University
email: [email protected]
Abstract: The radiative transfer equation (RTE) arises in a wide range of
applications. Recently, there has been much interest in the RTE due to its
application in biomedical imaging. It is challenging to solve RTE numerically
because of its integro-differential form, high dimension, and numerical
singularity in highly forward-peaked media. In this talk, adaptive solution
algorithms will be discussed for solving the RTE based on rigorously derived a
posteriori error estimates, and a family of differential approximations of the
RTE will be explored and analyzed. In addition, parameter identification
problems and source identification problems will be briefly discussed.
Stochastic symplectic methods for stochastic
Hamiltonian systems
Speaker: Jialin Hong, LSEC, Academy of Mathematics and Systems Science,
CAS
email: [email protected]
Abstract: The phase flow of stochastic Hamiltonian systems (SHSs)
preserves stochastic symplectic structure. A numerical method applied SHSs
is called stochastic symplectic if it preserves the structure. In this talk we
present some results on stochastic sympelctic methods for SHSs, including
stochastic generating functions for stochastic symplectic methods, and
stochastic Hamilton-Jacobi theory. We investigate the canonical form and the
stochastic symplectic structure of stochastic Schroedinger equations (SSEs),
and show that the symplectic Runge-Kutta semidiscretization of SSEs in time
preserves charge conservation law. We give a fundamental convergence
theorem, in mean-square sense, on the semidiscretization of SSEs in time,
and an application to a symplectic semidiscretization.
Sino-French Conference on Computational and Applied Mathematics
13
On the dynamical Rayleigh-Taylor instability in
compressible viscous flows without heat conductivity
Speaker: Song Jiang, Institute of Applied Physics and Computational
Mathematics, Beijing
email: [email protected]
Abstract: We investigate the instability of a smooth Rayleigh-Taylor
steady-state solution to Compressible viscous flows without heat conductivity
in the presence of a uniform gravitational field in a bounded domain with
smooth boundary. We show that the steady-state is linearly unstable by
constructing a suitable energy functional and exploiting arguments of the
modified variational method. Then, based on the constructed linearly unstable
solutions and a local well-posedness result of classical solutions to the original
nonlinear problem, we further reconstruct the initial data of linearly unstable
solutions to be the one of the original nonlinear problem and establish an
appropriate energy estimate of Gronwall-type. With the help of the established
energy estimate, we show that the steady-state is nonlinearly unstable in the
sense of Hadamard by a careful bootstrap argument. As a byproduct of our
analysis, we verify the destabilizing effect of compressibility in the linearized
problem for compressible viscous flows without heat conductivity.
Uncertainty quantification for transport equation with
uncertain coefficients in the diffusive regimes
Speaker: Shi Jin, Shanghai Jiaotong University & University of Wisconsin
email: [email protected]
Abstract: In this talk we will study generalized polynomial chaos (gPC)
approach to transport equation with uncertain cross-sections and show that
they can be made asymptotic-preserving, in the sense that in the diffusion limit
the gPC scheme for the transport equation approaches to the gPC scheme for
the diffusion equation with random diffusion coefficient. This allows the
implemention of the gPC method without numerically resolving (by space, time,
and gPC modes) the small mean free path for transport equation in the
diffusive regime.
Sino-French Conference on Computational and Applied Mathematics
14
The energy landscape of cellular systems from the
large deviation point of view
Speaker: Tiejun Li, Peking University
email: [email protected]
Abstract: The dynamics of cellular systems is of non-gradient type in general.
To understand its robustness, adaptivity and related properties with respect to
noise perturbation is a fundamental question. In this talk, we will show how to
construct the energy landscape of cellular systems by incorporating the large
deviation theory and characterize the robustness of the dynamical process.
The considered system will include the genetic switching and a simplified
budding yeast cell cycle process.
A numerical method for the quasi-incompressible
NSCH model for variable density two-phase flows with
a discrete energy law
Speaker: Ping Lin, University of Dundee
email: [email protected]
Abstract: We will present some recent work on phase-field models for
two-phase fluids with variable densities. The Quasi-Incompressible
Navier-Stokes-Cahn-Hilliard model with the gravitational force being
incorporated in the thermodynamically consistent framework will be
investigated. Under a minor reformulation of the system we show that there is
a continuous energy law underlying the system, assuming that all variables
have reasonable regularities. For the reformulated system we then design a
continuous finite element method and a special temporal scheme such that the
energy law is accurately preserved at the discrete level. Such a discrete
energy law for a variable density two-phase flow model has never been
established before with continuous finite element. Some numerical results will
be presented to demonstrate the capabilities of our numerical schemes. We
will also show an example that an energy law preserving method will perform
better for multiphase flow problems. Finally, we will extend the model to
account for the thermocapillary effects. It allows for the different properties
(densities, viscosities and heat conductivities) of each component while
maintaining thermodynamic consistency. To our knowledge such a model is
new. Numerical validation is provided too.
Sino-French Conference on Computational and Applied Mathematics
15
Modeling and simulation of Gaseous microflows
Speaker: Li-Shi Luo, Old Dominion University
email: [email protected]
Abstract: We study gaseous flows in micro-scales by using molecular
dynamics (MD), kinetic equation, and hydrodynamic equations. First, we solve
the linearized Boltzmann equation in a wide range of Knudsen number by
using an efficient high-order collocation method the singular integral equation.
Based the solution of the integral equation, we construct various approximated
solutions which can be modeled by macroscopic equations. We also simulate
molecular flows by using molecular dynamics. We propose macroscopic model
to simulate molecular and kinetic microflows. We use Couette flow in
two-dimensions as the specific example to illustrate our ideas.
A priori and a posteriori analysis for electronic
structure calculation
Speaker: Yvon Maday, Pierre and Marie Curie University
email: [email protected]
Abstract: In this talk I shall provide a priori and a posteriori estimates for the
approximation of nonlinear eigenvalue problems that are encountered in
electronic structure calculations.I shall start by reviewing the definition of the
models (Hartree Fock and Kohn-Sham) then present the complete results
(both on eigenvectors and eigenvalues) on the a priori error analysis for such
problems.
The analysis will include plane wave calculations and finite element
approximation together with DG approximations. We shall also provide a
complete analysis of a posteriori type where error due to space discretization,
incomplete iterative solution methods and models are considered and
balanced in order to get an optimal convergence.
This presentation assembles joined contributions with Eric Cances,
Rachida Chakir, Genevieve Dusson, Benjamin Stamm and Martin Vohralik.
Sino-French Conference on Computational and Applied Mathematics
16
Normal forms for semi-linear equations with
non-dense domain
Speaker: Pierre Magal, University of Bordeaux
email: [email protected]
Abstract: This presentation is devoted to bifurcation problems for some
classes of PDE arising in the context of population dynamics. The main
difficulty in such a context is to understand the dynamical properties of a PDE
with non-linear and non-local boundary conditions. A typical class of examples
is the so called age structured models. Age structured models have been well
understood in terms of existence, uniqueness, and stability of equilibria since
the 80's. Nevertheless, up to recently, the bifurcation properties of the
semiflow generated by such a system has been only poorly understood.
In this presentation, we will start with some results about existence and
smoothness of the center manifold. We will mention some Hopf’s bifurcation
results as well as some applications. We will also present some recent results
about normal form theory in such a context. As a consequence the normal
form theory, we will be in position to investigate the Bogdanov-Takens
bifurcations for a class of age-structured models.
Quasicontinuum methods for a crack model
Speaker: Pingbing Ming, LSEC, CAS
email: [email protected]
Abstract: In this talk, we study three quasicontinuum approximations of a
lattice model for crack propagation. The influence of the approximation on the
bifurcation patterns is investigated. The estimate of the modeling error is
applicable to near and beyond bifurcation points, which enables us to evaluate
the approximation over a finite range of loading and multiple mechanical
equilibria. Our results indicate that the so-called ghost forces has a nonlocal
effect on the system un study, which differs very much from that on the defect
free system. This is a joint work with Xiantao Li.
Sino-French Conference on Computational and Applied Mathematics
17
Global existence and parabolic limit of first-order
quasilinear hyperbolic systems
Speaker: Yue-Jun Peng, Blaise Pascal University
email: [email protected]
Abstract: We consider the Cauchy problem for first-order quasilinear partially
dissipative hyperbolic systems with a small parameter. Typically, the small
parameter is the relaxation time in physical models. Under stability conditions,
we show the global existence of smooth solutions in a uniform neighbourhood
of a constant equilibrium state with respect to the parameter. In a slow time,
this allows to obtain the global-in-time convergence of the systems to parabolic
equations as the parameter goes to zero. For large initial data, the
convergence is justified in a uniform local-in-time interval.
Some new results of global existence in
reaction-diffusion-advection systems
Speaker: Michel Pierre, ENS Rennes & IRMAR
email: [email protected]
Abstract: We will present several new results of global existence for
reaction-diffusion-advection systems for which preservation of positivity and
control of mass hold.It is well-known that blow up may occur infinite time even
if the total mass is uniformly bounded in time. Positive known results are
essentially of two kinds: either we assume some so-called triangular structure
of the reactive part and we
aim at classical solutions, or we look for weak global solutions. We will present
some new contributions in the following directions: anisotropic time dependent
diffusion together with not so regular advection terms; nonlinear diffusions;
Wentzell boundary conditions; electro-diffusion-reaction systems. They have
been obtained in collaboration with D. Bothe, A. Fischer, G. Rolland, G.&J.
Goldstein, M. Meyries.
Sino-French Conference on Computational and Applied Mathematics
18
Dynamic programming for mean-field type control: two
applications
Speaker: Olivier Pironneau, Pierre and Marie Curie University
email: [email protected]
Abstract: Mean-field games with an infinite number of players yield in the limit
to a new type of stochastic control problem where the coefficients are function
of the probability measure of the mean stochastic process. Assuming that a
PDF exists, the problem is standard for calculus of variations but non-standard
for dynamic programming. Using derivatives with respect to measures as in
Lions1 we shall derive the HJB equations. We shall present two numerical
applications of the method, one for a portfolio optimization and another for the
systemic risk problem studied by J. Garnier and G. Papanicolaou2. Using
freefem++, we have computed the most probable transition, among the rare
events, that passes from a state of "stability" (left hump in figure 1) to a state of
"crisis" (right hump) for a situation with many agents who optimize their utility
which is itself function of the average of the position of all agents. This is a joint
work with Mathieu Lauriere.
Figure 1: Left: Time snapshots of PDF (left) at 0,T/3,2T/3 and T.
Right:Optimization history of J : Red log( J ). Green same minus penalty part.
Blue 2 log |||| Jgradu .
Sino-French Conference on Computational and Applied Mathematics
19
Controllability of fluid flows
Speaker: Jean-Pierre Puel, University of Versailles
email: [email protected]
Abstract: This presentation will give an overview on the controllability for fluid
flows. There will be no technical proof. First of all we will describe in an
abstract situation the various concepts of controllability for evolution equations.
We will then present some problems and results concerning the controllability
of systems modeling fluid flows. First of all we will consider the Euler equation
describing the motion of an incompressible inviscid fluid. Then we will give
some results concerning the Navier-Stokes equations, modeling an
incompressible viscous fluid, and some related systems, in particular the case
of what is called Lagrangian control and which might lead to a lot of
developments. Finally, we will present a first result of controllability for the case
of a compressible fluid (in dimension 1) and some important open problems.
Exact synchronization for a coupled system of wave
equations with Dirichlet boundary controls
Speaker: Bopeng Rao, University of Strasbourg
email: [email protected]
Abstract: In this talk, the exact synchronization for a coupled system of wave
equations with Dirichlet boundary controls and some related concepts are
introduced. By means of the exact null controllability of a reduced coupled
system, under certain conditions of compatibility, the exact synchronization,
the exact synchronization by groups, and the exact null controllability and
synchronization by groups are all realized by suitable boundary controls. This
is a joint work with Tatsien Li.
Sino-French Conference on Computational and Applied Mathematics
20
Image restoration: wavelet frame approach, total
variation and beyond
Speaker: Zuowei Shen, National University of Singapore
email: [email protected]
Abstract: This talk is about the wavelet frame-based image and video
restorations. We start with some of main ideas of wavelet frame based image
restorations. Some of applications of wavelet frame based models image
analysis and restorations will be shown. Examples of such applications include
image and video inpainting, denoising, decomposition, image deblurring and
blind deblurring, segmentation, CT image reconstruction, 3D reconstruction
in electronmicroscopy, and etc. In all of these applications, spline wavelet
frames derived from the unitary extension principle are used. This allows us
to establish connections between wavelet frame base method and various
PDE based methods, that include the total variation model, nonlinear
diffusion PDE based methods, and model of Mumford-Shah. In fact, we will
show that when spline wavelet frames are used, right chosen models of a
wavelet frame method can be viewed as a discrete approximation at a given
resolution to the corresponding PDE based models. A convergence analysis in
terms of objective functionals and their approximate minimizers as image
resolution increases will be discussed.
Some aspects of finite element approximation for
Reissner-Mindlin plates
Speaker: Zhong-Ci Shi, Institute of Computational Mathematics, CAS
email: [email protected]
Abstract: The Reissner-Mindlin plate model is one of the most commonly
used models of a moderate-thick to thin elastic plate. However, a direct finite
element approximation usually yields very poor results, which is referred to
LOCKING phenomenon.
In the past two decades, many efforts have been devoted to the design of
locking free finite elements to resolve this model, most of these work focus on
triangular or rectangular elements, the latter may be extended to parallelogra
ms, but very few on quadrilaterals.
In this talk we will give an overview of the recent development of low order
quadrilateral elements and present some new results.
Sino-French Conference on Computational and Applied Mathematics
21
Viscosity solutions of path-dependent PDEs
Speaker: Nizar Touzi, Ecole polytechnique
email: [email protected]
Abstract: Partial differential equation with non-anticipative dependence on the
path arise in many applications as stochastic control and stochastic differential
games in non-Markov systems, backward stochastic differential equations,
large deviations for non-Markov diffusions, path-dependent branching
diffusions and the related simulation. We introduce an adaptation of the
Crandall-Ishii notion of viscosity solutions which bypasses the difficulty due to
the non local compacteness of the underlying space of paths. Our definition
replaces the pointwise tangency by the tangency in mean, thus admitting a
larger class of test functions. This notion of viscosity solution satisfies a
stability result, similar to the standard notion in finite-dimensional spaces.
Wellposedness results are established in the general fully nonlinear case. In
the semilinear case, the comparison result is obtained by simpler arguments,
as compared to the standard theory, by taking advantage of the larger class of
test functions.
Mathematics and swimming of aquatic organisms
Speaker: Marius Tucsnak, University of Lorraine
email: [email protected]
Abstract: We consider several mathematical models describing the motions of solids Immersed in an incompressible fluid. We begin by emphasising that, depending on the flow regime, the governing equations may exhibit a wide range of properties, leading to a rich mathematical structure. We next discuss wellposedness issues, where the major difficulty to be solved consists in tackling the presence of free boundaries. Finally we study the displacement of the solids (under the action of an exterior force or in a self-propelled manner) from a control theoretical perspective.
Sino-French Conference on Computational and Applied Mathematics
22
Analysis and computation of a polar active liquid
crystal models
Speaker: Qi Wang, University of South Carolina & Beijing Computational
Science Research Center
email: [email protected]
Abstract: Active liquid crystals encompasses a class of live and man-made
materials that consist of self-propelled molecules, nano- or micro-particles.
They can form orientation orders at sufficiently high concentration and produce
spontaneous flows, long-range spatial and temporal patterns. We will study the
polar active liquid crystal system using a continuum model and explore
mechanisms for the generation of spatial as well as temporal patterns. Stability
and special solutions in channel flows, capillary flows, and 2-D confined flows
will be analyzed. Transient flow simulation will be conducted using a finite
difference method.
Multilevel Block-Incremental unknowns for general
anisotropic elliptic equations: diagonal
preconditioning and condition analysis
Speaker: Yu-Jiang Wu, Lanzhou University
email: [email protected]
Abstract: We give, at first, estimates of condition numbers of the system
produced by block incremental unknowns (BIU) with preconditioning for
anisotropic operators. Then a diagonal preconditioner is introduced to the
BIU-preconditioned system and new reduction of condition number is obtained.
The results will also be produced when we apply to more general anisotropic
elliptic equations. Finally, some numerical experiments and results are
presented to confirm our theoretical analysis. This is a joint work with Ai-Li
Yang and Lun-Ji Song.
Sino-French Conference on Computational and Applied Mathematics
23
Parallel adaptive finite element computations of
Poisson-Nernst-Planck equations and applications to
ion channel systems
Speaker: Lin-bo Zhang, LSEC, Chinese Academy of Sciences
email: [email protected]
Abstract: In this talk I will present our collaborative efforts on developing
parallel adaptive finite element codes for solving three-dimensional coupled
electro-diffusion equations, including the Poisson-Nernst-Planck equations
(PNP) and size-modified Poisson-Nernst-Planck equations (SMPNP), using
our parallel adaptive finite element toolbox Parallel Hierarchical Grid (PHG).
By integrating these codes with the pre- and post-processing tools that our
collaborators have been working on for many years, we have developed a
software package, ichannel, for numerical simulation of ion transport through
three-dimensional ion channel systems that consist of protein and membrane,
and succeeded in carrying out finite element simulations of some channels
such as the gA channel, the VDAC channel and the PA channel. The
numerical results obtained with ichannel agree well with existing Brownian
dynamics (BD) simulation results and experimental results.
Defects of liquid crystals
Speaker: Pingwen Zhang, Peking University
email: [email protected]
Abstract: Defects in liquid crystals (LCs) are of great practical and theoretical
importance. Recently there is a growing interest in LCs materials under
topological constrain and/or external force, but the defects pattern and
dynamics are still poorly understood. We investigate three-dimensional
spherical droplet within the Landau-de Gennes model under different boundary
conditions. When the Q-tensor is uniaxial, the model degenerates to vector
model (Oseen-Frank), but Q-tensor model is superior to vector model as the
former allows biaxial in the order parameter. Using numerical simulation, a rich
variety of defects pattern are found, and the results suggest that, line
disclinations always involve biaxial, or equivalently, uniaxial only admits point
defects. Then we believe that Q-tensor model is essential to include the
disclinations line which is a common phenomenon in LCs. The mathematical
implication of this observation will be discussed in this talk.
Sino-French Conference on Computational and Applied Mathematics
24
Some inverse problems for stochastic PDEs
Speaker: Xu Zhang, Sichuan University
email: [email protected]
Abstract: In this talk, I will present some recent results on the inverse
stochastic hyperbolic/parabolic problems. I will explain that both the
formulation of stochastic inverse problems and the tools to solve them may
differ considerably from their deterministic counterpart.
Superconvergence in polynomial spectral collocation:
an unclaimed territory
Speaker: Zhimin Zhang, Beijing Computational Science Research Center&
Wayne State
email: [email protected]
Abstract: Superconvergence phenomenon is well understood for the h
-version finite element method and researchers in this field have accumulated
a vast literature during the past 40 years. However, the relevant study for the p
-version finite element method and spectral methods is lacking. We believe
that the scientific community would also benefit from the study of
superconvergence phenomenon of those methods. Recently, some efforts
have been made to expand the territory of the superconvergence. In this talk,
we summarize some recent development on superconvergence study for the
spectral collocation methods.
Sino-French Conference on Computational and Applied Mathematics
25
Radial symmetry of entire solutions of a bi-harmonic
equation with exponential nonlinearity
Speaker: Feng Zhou, East China Normal University
email: [email protected]
Abstract: We will talk about necessary and sufficient conditions for an entire
solution $u$ of a biharmonic equation with exponential nonlinearity $e^u$ to
be a radially symmetric solution. The standard tool to obtain the radial
symmetry for a system of equations is the Moving-Plane-Method (MPM). In
order to apply the MPM, we need to know the asymptotic expansions of
$u$ and $-\Delta u$ at $\infty$. The difficulties come from the fact that $e^u$ is
supercritical for $N\geq5$, $e^u \not \in L^{\frac{N}{4}} (\mathbb{R}^N)$. We
need to get the right decay rate of $u$ and $-\Delta u$ at $\infty$ in order to
start the moving plane procedure. This is a joint work with Z.M.Guo and X.
Huang.