PDE and applications seminar | Metz

The main objective of this work is to study the stability of a linear one-dimensional thermoelastic Bresse system in a bounded domain, where the coupling is given through the first component of the Bresse model with the heat conduction of Gurtin-Pipkin type. Two kinds of coupling are considered; the first coupling is of order one with respect to space variable, and the second one is of order zero. We state the well-posedness and show the polynomial and strong stability of the systems for regular and weak solutions, respectively, where the polynomial decay rates depend on the smoothness of the initial data. Moreover, in case of coupling of order one, we prove the equivalence between the exponential stability and some new conditions on the parameters of the system. However, when the coupling is of order zero, we prove the non-exponential stability independently of the parameters of the system. Applications to the corresponding particular Timoshenko models are also given, where we prove that both couplings lead to the exponential stability if and only if some conditions on the parameters of the systems are satisfied, and both couplings guarantee the polynomial and strong stability for regular and weak solutions, respectively, independently of the parameters of the systems. The proof of the well-posedness result is based on the semigroups theory, whereas a combination of the energy method and the frequency domain approach is used for the proof of the stability results.

For the details, see the following paper:

A. Guesmia, Well-posedness and stability results for thermoelastic Bresse and Timoshenko type systems with Gurtin-Pipkin’s law through the vertical displacements, SeMa J., (2023), 1-49.

The objective of this work is to study the stability of two systems of type Rao-Nakra sandwich beam in the whole line $R$ with a frictional damping or an infinite memory acting on the Euler-Bernoulli equation. When the speeds of propagation of the two wave equations are equal, we show that the solutions do not converge to zero when time goes to infinity. In the reverse situation, we prove some $L^2 (R)$-norm and $L^1 (R)$-norm decay estimates of solutions and theirs higher order derivatives with respect to the space variable. Thanks to interpolation inequalities and Carlson inequality, these $L^2 (R)$-norm and $L^1 (R)$-norm decay estimates lead to similar ones in the $L^q (R)$-norm, for any $q\in [1,+\infty]$. In our both $L^2 (R)$-norm and $L^1 (R)$-norm decay estimates, we specify the decay rates in terms of the regularity of the initial data and the nature of the control. Applications to some Cauchy Timoshenko type systems will be also given. The proof is based on the energy method combined with the Fourier analysis (by using the transformation in the Fourier space and well chosen multipliers).

A part of these results was obtained in collaboration with Salim Messaoudi (University of Sharjah, UAE).

For the details, see the following papers:

A. Guesmia, Some $L^q (R)$-norm decay estimates ($q\in[1,+\infty]$) for two Cauchy systems of type Rao-Nakra sandwich beam with a frictional damping or an infinite memory, J. Appl. Anal. Comp., 12 (2022), 2511-2540.
A. Guesmia, On the stability of a linear Cauchy Rao-Nakra sandwich beam under frictional dampings, Taiwanese J. Math., 27 (2023), 799-811.
A. Guesmia and S. Messaoudi, Some $L^2 (R)$-norm and $L^1 (R)$-norm decay estimates for Cauchy Timoshenko type systems with a frictional damping or an infinite memory, J. Math. Anal. Appl., 527 (2023), 127385.

Ceci est un travail en commun avec Andras Vasy et Michal Wrochna.
On considère $\mathbb{R}^n$ munit d’une métrique Lorentzienne asymptotiquement Minkowski dont les géodésiques nulles sont non captées. On décrira les propriétés spectrales du carré de l’opérateur de Dirac associé et on en déduira un principe d’action spectrale à partir des puissances de $D^2$.

Il est bien connu que la reconstruction d’une donnée initiale associée à une équation parabolique à partir de mesures internes de sa solution pendant un temps $T$, sur un domaine $\omega$ appelé domaine d’observation équivaut à la question de l’observabilité, ou plus précisément à la positivité de ce qu’on appelle la constante d’observabilité associée à $\omega$. Cette constante dépend du domaine d’observation $\omega$ mais aussi de façon cruciale de l’horizon temporel $T$.

Dans cet exposé, nous nous intéressons au positionnement optimal de capteurs thermiques. Il est raisonnable de modéliser cette question par la recherche des domaines extrémaux (lorsqu’ils existent) maximisant cette constante d’observabilité. Pour être physiquement pertinent, nous imposons une restriction sur la mesure du domaine observé.

Après avoir introduit une relaxation convexe du problème d’optimisation de la forme, nous déterminons le comportement asymptotique des maximiseurs lorsque $T$ tend vers $+\infty$. En utilisant de façon cruciale un principe de la baignoire quantitatif, nous prouvons la forte convergence des maximiseurs vers la fonction caractéristique d’un ensemble mesurable que nous caractérisons précisément, et montrons en outre que cette convergence est exponentielle.

Il s’agit d’un travail en collaboration avec Idriss Mazari (univ. Paris Dauphine) et Emmanuel Trélat (Sorbonne univ.)

Il s’agit de la suite de l’exposé de la semaine dernière dont voici le résumé :

We study the Laplace and Stokes operators on a manifold with straight cylindrical ends.
We obtain useful Fredholm, regularity, and invertibility results. An important role is played by an adapted pseudodifferential calculus on manifolds with straight cylindrical ends which contains the inverses of its $L^2$-invertible, elliptic operators of non-negative order.
We also obtain the layer potentials for the elliptic operators studied, and the well-posedness of the corresponding Dirichlet problem.
Joint work with Victor Nistor (Metz) and Wolfgang L. Wendland (Stuttgart).

We study the Laplace and Stokes operators on a manifold with straight cylindrical ends.
We obtain useful Fredholm, regularity, and invertibility results. An important role is played by an adapted pseudodifferential calculus on manifolds with straight cylindrical ends which contains the inverses of its $L^2$-invertible, elliptic operators of non-negative order.
We also obtain the layer potentials for the elliptic operators studied, and the well-posedness of the corresponding Dirichlet problem.
Joint work with Victor Nistor (Metz) and Wolfgang L. Wendland (Stuttgart).

L’algorithme word2vec (Mikolov et al, 2013) permet d’associer des vecteurs à des mots.
Ces plongements de mots, sous des formes plus sophistiquées, forment le cœur des Grands Modèles de Langues (LLM) utilisés par les outils d’IA dont tout le monde a entendu parler.
Si word2vec est présenté comme un algorithme de réseaux de neurones, il peut-être décrit très simplement comme un problème d’optimisation impliquant deux matrices et rien de plus.
Dans cet exposé, nous présenterons un état des lieux de notre compréhension de cet algorithme par une approche de rétro-ingénierie, et des questions ouvertes.

D’après un travail commun avec Didier Gemmerlé, Lionel Lenôtre, Pierre Mercuriali et Saïd Toubra.

We give a light talk on optimality of shapes in geometry and physics.
First, we recollect classical geometric results that the disk has the largest area (respectively, the smallest perimeter) among all domains of a given perimeter (respectively, area).
Second, we recall that the circular drum has the lowest fundamental tone among all drums of a given area or perimeter and reinterpret the result in a quantum-mechanical language of nanostructures.
In parallel, we discuss the analogous optimality of square among all rectangles in geometry and physics.
As the main body of the talk, we present a joint work with Freitas in which we show that the disk actually stops to be the optimiser for elastically supported membranes, disproving in this way a long-standing conjecture of Bareket’s.
We also present our recent attempts to prove the same spectral-geometric properties in relativistic quantum mechanics.
It is frustrating that such an illusively simple and expected result remains unproved and apparently out of the reach of current mathematical tools.

Le but de ce groupe de travail sera de comprendre (partiellement) les différentes équations de la mécanique des fluides afin de les reconnaître au détour d’un exposé ou de la lecture d’un papier.

Dans cette troisième partie, nous discuterons de l’entropie d’un fluide.

Les notes de ce groupe de travail se trouvent ici (merci à Jérémy).

Le but de ce groupe de travail sera de comprendre (partiellement) les différentes équations de la mécanique des fluides afin de les reconnaître au détour d’un exposé ou de la lecture d’un papier.

Dans cette deuxième partie, nous discuterons de l’énergie et de l’entropie d’un fluide.

Le but de ce groupe de travail sera de comprendre (partiellement) les différentes équations de la mécanique des fluides afin de les reconnaitre au détour d’un exposé ou de la lecture d’un papier.

In this talk, we review recent results on so-called partially dissipative hyperbolic systems. Such systems model physical phenomena with degenerate dissipative terms and appear in many applications. For example, in gas dynamics where the mass is conserved during the evolution, but the momentum balance includes a diffusion (viscosity) or a damping (relaxation) term.

First, using tools from the hypocoercivity theory and precise frequency decompositions, we derive sharp stability estimates for linear systems satisfying the Kalman rank condition. This linear analysis allows us to establish new global-in-time existence and large-time behaviour results in a critical regularity framework for nonlinear systems.

Then, we interpret partially dissipative systems as hyperbolic approximations of parabolic systems, in the context of the paradox of infinite speed of propagation. In particular, we focus on a hyperbolic approximation of the multi-dimensional compressible Navier-Stokes-Fourier system and establish its hyperbolic-parabolic strong relaxation limit.