Séminaire EDP, Analyse et Applications | Metz

I will talk about the spectrum of double layer potential operators for 3D surfaces with rough features. The existence of spectrum reflects the fact that transmission problems across the surface may be ill-posed for (complex) sign-changing coefficients. The spectrum is very sensitive to the regularity sought of solutions. For $L^2$ boundary data, for domains with corners and edges, the spectrum is complex and carries an associated index theory. Through an operator-theoretic symmetrisation framework, it is also possible to recover the initial self-adjoint features of the transmission problem – corresponding to $H^{1/2}$ boundary data – in which case the spectral picture is more familiar.

Résumé à  préciser

Je vais commencer par présenter la modélisation des avalanches denses (modèle mécanique et géométrique, les hypothèses de l’écoulement de faible épaisseur, le modèle asymptotique, etc). Puis je vais lier l’analyse (statique) du déclenchement d’une avalanche avec les problèmes d’analyse limite et du spectre du 1-Laplacien (problème de Cheeger). Une deuxième partie sera consacrée à  la modélisation numérique de l’évolution dynamique d’une avalanche. Je vais terminer avec une comparaison entre le modèle et les expériences de laboratoire et quelques simulations sur des topographies réelles.

Exposé donné dans le cadre des Journées Analyse et Physique mathématique
The anisotropic Calderon problem is the inverse problem consisting in determining a metric on a compact Riemannian manifold with boundary from the Dirichlet-to-Neuman map. The resolution of the problem in a conformal class follows from a similar inverse problem on the Schrödinger equation and remains an open question in dimensions higher than 3. In previous works, we could solve this inverse problem under structural assumptions on the known metric (namely that it is conformal to a warped product with an Euclidean factor) and additional geometric assumptions on the transversal manifold. The proof of uniqueness relies on the high frequency limit in a Green identity involving pairs of complex geometrical optics solutions to the Schrödinger equation. This talk will be concerned with a description of the resolution of this nonlinear inverse problem under strong assumptions on the metric and our attempts to remove the additional transversal assumptions on the geometry by refraining from passing to the limit in the linearised problem. Unfortunately, this path only leads to partial results on the linearised problem for the time being, that is recovery of singularities of the potential in the transversal variables. This a joint work with Yaroslav Kurylev, Matti Lassas, Tony Liimatainen and Mikko Salo.

The talk discusses two works of the author linking the topological properties, i.e. “something that can be seen”, with the analytical properties of dispersion relations in waveguides. The first example is related to a quantum waveguide, i.e. to a periodic (elongated in one dimension) graph-like structure consisting of edges bearing a wave equation and nodes considered as joints. In acoustics the edges are thin pipes. The problem of this research was to estimate the number of modes that can travel (or decay) in each direction along such a waveguide. The final result is as follows. One should build a graph consisting of a closed single cell of the periodic graph. The estimation of the number of modes is a maximum degree of a linear subgraph of this graph. Thus, although the consideration is held in the algebraic way (a determinant- like dispersion equation is solved), the answer is given in the graph language. The second example is related to 2D waveguiding structures that are layered in the transversal direction. It may happen that the group velocities of all waveguide modes are lower than the biggest velocity in one of the layers. In this case, one can observe a forerunner, i.e. a pulse travelling faster than all the modes and decaying exponentially. The problem is how to find it on the dispersion diagram of the waveguide. The result is as follows. The dispersion diagram should be considered as a multivalued analytical function of, say, temporal frequency, taken on its Riemann surface. The forerunner branch then can be found on the analytical continuation of the diagram. The branch point of the diagram describes interaction between the layers.

Le résumé se trouve ici

Le résumé se trouve ici.

Dispersion is the phenomenon in which the phase velocity of a wave depends on its frequency, or alternatively when the group velocity depends on the frequency. Examples of classical nonlinear dispersive equations are the (generalized) KdV equation, the Nonlinear Schrödinger equation, and the Boussinesq equation. In addition to the well-posedness it is known blow-up effect, for critical and super-critical cases that generally depend on the exponent p > 0 present in the nonlinearity of these equations. Dispersive blow-up is a focussing type of behavior which is due to both the unbounded domain in which the problem is set and the propensity of the dispersion relation to propagate energy at different speeds. On the other hand, a damping term can prevent this blow-up effect in the dispersive equations, and it is what is studied in several works, both for the KdV and for the Schrödinger equation.

Particles are used in several different ways in computational physics. For example one can study systems of point masses, point charges, or point vortices. Another approach considers the particle system as a discretization of a continuous PDE problem; in this case one is dealing with a particle method, as an alternative to the classical discretization methods such as finite-difference, finite-element, and spectral methods. Here we consider particle methods in two areas, (1) electrostatics of solvated proteins, where the particles are nodes in a triangulation of the molecular surface, and (2) incompressible fluid dynamics, where the particles represent the flow map and carry vorticity. We discuss the challenges facing particle methods and some techniques that improve their accuracy and efficiency, including adaptive refinement, remeshing, and treecode-acceleration.

Nous étudions l’existence et la stabilité des ondes périodiques pour un modèle non linéaire, l’équation de Lugiato-Lefever, issu de l’optique. En utilisant des méthodes de la théorie des bifurcations, nous étudions les bifurcations de Turing et montrons l’existence de solutions périodiques. Cette approche permet également de conclure sur la stabilité de ces solutions vis-à -vis de perturbations périodiques dont la période est un multiple entier de la période de l’onde. En utilisant ensuite de méthodes de la théorie des opérateurs, nous montrons la stabilité de ces solutions vis-à -vis de perturbations générales, bornées.