We consider the P n model to approximate the time dependent transport equation in one dimension of space. In a diffusive regime, the solution of this system is solution of a diffusion equation.We are looking for a numerical scheme having the diffusion limit property: in a diffusive regime, it has to give the solution of the limiting diffusion equation on a mesh at the diffusion scale.The numerical scheme proposed is an extension of the Godunov type scheme proposed by Gosse to solve the P 1 model without absorption term. It requires the computation of the solution of the steady state P n equations. This is made by one Monte-Carlo simulation performed outside the time loop. Using formal expansions with respect to a small parameter representing the inverse of the number of mean free path in each cell, the resulting scheme is proved to have the diffusion limit. In order to avoid the CFL constraint on the time step, we give an implicit version of the scheme which preserves the positivity of the zeroth moment.

In this paper, we first construct a model for free surface flows that takes into accountthe air entrainment by a system of four partial differential equations. We derive it bytaking averaged values of gas and fluid velocities on the cross surface flow in the Eulerequations (incompressible for the fluid and compressible for the gas). The obtained systemis conditionally hyperbolic. Then, we propose a mathematical kinetic interpretation ofthis system to finally construct a two-layer kinetic scheme in which a special treatmentfor the “missing” boundary condition is performed. Several numerical tests on closed waterpipes are performed and the impact of the loss of hyperbolicity is discussed andillustrated. Finally, we make a numerical study of the order of the kinetic method in thecase where the system is mainly non hyperbolic. This provides a useful stability resultwhen the spatial mesh size goes to zero.

In the reliability theory, the availability ofa component, characterized by non constant failure and repair rates,is obtained, at a given time, thanks to the computation of the marginal distributions of asemi-Markov process. These measures are shown to satisfy classicaltransport equations, the approximation of which can be donethanks to a finite volume method.Within a uniqueness result for the continuous solution,the convergence of the numerical scheme isthen proven in the weak measure sense,and some numerical applications, which show the efficiency and theaccuracy of the method, are given.

In this paper, we are interested in finding the optimal shapeof a magnet. The criterion to maximize is the jump of theelectromagnetic field between two different configurations.We prove existence of an optimal shape into a natural classof domains. We introduce a quasi-Newton type algorithm whichmoves the boundary. This method is very efficient to improvean initial shape. We give some numerical results.

In this paper we study a model problem describing the movement of a glacier under Glen's flow law and investigated by Colinge and Rappaz [Colinge and Rappaz, ESAIM: M2AN33 (1999) 395–406]. We establish error estimates for finite element approximation using the results of Chow [Chow, SIAM J. Numer. Analysis29 (1992) 769–780] and Liu and Barrett [Liu and Barrett, SIAM J. Numer. Analysis33 (1996) 98–106] and give an analysis of the convergence of the successive approximations used in [Colinge and Rappaz, ESAIM: M2AN33 (1999) 395–406]. Supporting numerical convergence studies are carried out and we also demonstrate the numerical performance of an a posteriori error estimator in adaptive mesh refinement computation of the problem.

We consider an initial-boundary value problem for a generalized 2D time-dependentSchrödinger equation (with variable coefficients) on a semi-infinite strip. For theCrank–Nicolson-type finite-difference scheme with approximate or discrete transparentboundary conditions (TBCs), the Strang-type splitting with respect to the potential isapplied. For the resulting method, the unconditional uniform in time L2-stability isproved. Due to the splitting, an effective direct algorithm using FFT is developed now toimplement the method with the discrete TBC for general potential. Numerical results on thetunnel effect for rectangular barriers are included together with the detailed practicalerror analysis confirming nice properties of the method.

We derive and analyze adaptive solvers for boundary value problems in which thedifferential operator depends affinely on a sequence of parameters. These methods convergeuniformly in the parameters and provide an upper bound for the maximal error. Numericalcomputations indicate that they are more efficient than similar methods that control theerror in a mean square sense.

Many inverse problems for differential equationscan be formulated as optimal control problems.It is well known that inverse problems often need tobe regularized to obtain good approximations.This work presents a systematic method to regularizeand to establish error estimates for approximations tosome control problems in high dimension, based on symplectic approximationof the Hamiltonian system for the control problem. In particularthe work derives error estimatesand constructs regularizations for numerical approximations tooptimally controlled ordinary differential equations in ${\mathbb R}^d$ ,with non smooth control.Though optimal controls in general becomenon smooth,viscosity solutions to the corresponding Hamilton-Jacobi-Bellmanequation provide good theoretical foundation, but poor computational efficiencyin high dimensions.The computational method here uses the adjoint variable and worksefficiently also for high dimensional problems with d >> 1.Controls can be discontinuous due to a lack of regularityin the Hamiltonian or due to colliding backward paths, i.e. shocks.The error analysis, for both these cases, is based on consistency with theHamilton-Jacobi-Bellman equation, in the viscosity solution sense,and a discrete Pontryagin principle:the bi-characteristic Hamiltonian ODE system is solvedwith a C 2 approximate Hamiltonian.
The error analysis leads to estimatesuseful also in high dimensions since the bounds depend on the Lipschitznorms of the Hamiltonian and the gradient of the value functionbut not on d explicitly.Applications to inverse implied volatility estimation, in mathematical finance,and to a topology optimization problem are presented.An advantage with the Pontryagin based method is that the Newton method can be applied to efficientlysolve the discrete nonlinear Hamiltonian system,with a sparse Jacobian that can be calculated explicitly.

In this paper, multiscale finite element methods (MsFEMs) and domain decomposition techniques are developed for a class of nonlinear elliptic problems with high-contrast coefficients. In the process, existing work on linear problems [Y. Efendiev, J. Galvis, R. Lazarov, S. Margenov and J. Ren, Robust two-level domain decomposition preconditioners for high-contrast anisotropic flows in multiscale media. Submitted.; Y. Efendiev, J. Galvis and X. Wu, J. Comput. Phys. 230 (2011) 937–955; J. Galvis and Y. Efendiev, SIAM Multiscale Model. Simul. 8 (2010) 1461–1483.] is extended to treat a class of nonlinear elliptic operators. The proposed method requires the solutions of (small dimension and local) nonlinear eigenvalue problems in order to systematically enrich the coarse solution space. Convergence of the method is shown to relate to the dimension of the coarse space (due to the enrichment procedure) as well as the coarse mesh size. In addition, it is shown that the coarse mesh spaces can be effectively used in two-level domain decomposition preconditioners. A number of numerical results are presented to complement the analysis.

In this paper we study a discrete Raman laser amplification model given as a Lotka-Volterra system. We show that in an ideal situation, the equations can be written as a Poisson system with boundary conditions using a global change of coordinates. We address the questions of existence and uniqueness of a solution. We deduce numerical schemes for the approximation of the solution that have good stability.

We study in this paper some numerical schemes for hyperbolic systemswith unilateral constraint. In particular, we deal with the scalar case, the isentropicgas dynamics system and the full-gas dynamics system.We prove the convergence of the scheme to an entropy solutionof the isentropicgas dynamics with unilateral constraint on the density and mass loss.We also study the non-trivial steady states of the system.

The aim of this work is to deduce the existence of solutionof a coupled problem arising in elastohydrodynamiclubrication. The lubricant pressure and concentration aremodelled by Reynolds equation, jointly with the free-boundaryElrod-Adams model in order to take into account cavitationphenomena. The bearing deformation is solution of Koitermodel for thin shells. The existence of solution to thevariational problem presents some difficulties: the coupledcharacter of the equations, the nonlinear multivaluedoperator associated to cavitation and the fact of writing theelastic and hydrodynamic equations on two different domains.In a first step, we regularize the Heaviside operator.Additional difficulty related to the differentdomains is circumvented by means of prolongation andrestriction operators, arriving to a regularized coupledproblem. This one is decoupled into elastic and hydrodynamicparts, and we prove the existence of a fixed point for theglobal operator. Estimations obtained for theregularized problem allow us to prove the existence ofsolution to the original one. Finally, a numerical method is proposed in orderto simulate a real journal-bearing device and illustrate the qualitative andquantitative properties of the solution.

The initial-boundary value problem of two-dimensionalincompressible fluid flow in stream function form is considered.A prediction-correction Legendre spectral scheme is proposed, which iseasy to be performed.The numerical solution possesses the accuracy of second-order in time and higher order in space. Thenumerical experiments show the high accuracy of this approach.

We propose a model for segmentation problemsinvolving an energy concentrated on the vertices of an unknownpolyhedral set, where the contours of the images to be recoveredhave preferred directions and focal points.We prove that such an energy is obtained as a Γ-limit offunctionals defined on sets with smooth boundary thatinvolve curvature terms of the boundary.The minimizers of the limit functional are polygons withedges either parallel to some prescribed directions or pointing to somefixed points, that can also be taken as unknown of the problem.

The penalty method when applied to the Stokes problem provides a very efficient algorithm for solving any discretization of this problem since it gives rise to a system of two equations where the unknowns are uncoupled. For a spectral or spectral element discretization of the Stokes problem, we prove a posteriori estimates that allow us to optimize the penalty parameter as a function of the discretization parameter. Numerical experiments confirm the interest of this technique.

We consider the Euler equation for an incompressible fluid on a three dimensional torus,and the construction of its solution as a power series in time. We point out some generalfacts on this subject, from convergence issues for the power series to the role ofsymmetries of the initial datum. We then turn the attention to a paper by Behr, Nečas andWu, ESAIM: M2AN 35 (2001) 229–238; here, the authors chose avery simple Fourier polynomial as an initial datum for the Euler equation and analyzed thepower series in time for the solution, determining the first 35 terms by computer algebra.Their calculations suggested for the series a finite convergence radiusτ3 in the H3 Sobolev space, with0.32 < τ3 < 0.35; they regarded this as an indicationthat the solution of the Euler equation blows up. We have repeated the calculations of E.Behr, J. Nečas and H. Wu, ESAIM: M2AN 35 (2001) 229–238,using again computer algebra; the order has been increased from 35 to 52, using thesymmetries of the initial datum to speed up computations. As forτ3, our results agree with the original computations of E.Behr, J. Nečas and H. Wu, ESAIM: M2AN 35 (2001) 229–238(yielding in fact to conjecture that 0.32 < τ3 < 0.33).Moreover, our analysis supports the following conclusions: (a) The finiteness ofτ3 is not at all an indication of a possible blow-up. (b)There is a strong indication that the solution of the Euler equation does not blow up at atime close to τ3. In fact, the solution is likely to exist, atleast, up to a time θ3 > 0.47. (c) There is a weakindication, based on Padé analysis, that the solution might blow up at a later time.

On exhibe dans cette note une paramétrix (au sens faible) de l'opérateur sous-jacent à l'équation CFIE de l'électromagnétisme. L'intérêt de cetteparamétrix est de se prêter à différentes stratégies de discrétisationet ainsi de pouvoir être utilisée comme préconditionneur de la CFIE.On montre aussi que l'opérateur sous-jacent à la CFIE satisfait une conditionInf-Sup discrète uniforme, applicable aux espaces de discrétisation usuellement rencontrésen électromagnétisme, et qui permet d'établir un résultat inédit de convergencenumérique de la CFIE.

This paper deals with modeling the passivebehavior of skeletal muscle tissue includingcertain microvibrations at the cell level. Our approach combines a continuum mechanics model with large deformation and incompressibility at the macroscale with chains of coupled nonlinear oscillators.The model verifies that an externally appliedvibration at the appropriate frequency is able to synchronize microvibrations in skeletal muscle cells.From the numerical analysis point of view, one faces here a partial differential-algebraic equation (PDAE) that after semi-discretization in space by finite elements possessesan index up to three, depending on certain physicalparameters. In this context, the consequences forthe time integration as well as possible remedies are discussed.

In this paper we present an analysis of the partial differential equations thatdescribe the Chemical Vapor Infiltration (CVI) processes. The mathematical modelrequires at least two partial differential equations, one describing thegas phase and one corresponding to the solid phase.A key difficulty in the process is the long processing times that are typicallyrequired. We address here the issue of optimization and show that we can chooseappropriate pressure and temperature to minimize these processing times.