Tuesday, September 28, 2010 - 15:00 , Location: Skiles 255 , Prof. Predrag Cvitanović , Physics, Georgia Institute of Technology , email@example.com , Organizer: Ronghua Pan
In the world of moderate Reynolds number, everyday turbulence of fluids flowing across planes and down pipes a velvet revolution is taking place. Experiments are almost as detailed as the numerical simulations, DNS is yielding exact numerical solutions that one dared not dream about a decade ago, and dynamical systems visualization of turbulent fluid's state space geometry is unexpectedly elegant. We shall take you on a tour of this newly breached, hitherto inaccessible territory. Mastery of fluid mechanics is no prerequisite, and perhaps a hindrance: the talk is aimed at anyone who had ever wondered why - if no cloud is ever seen twice - we know a cloud when we see one? And how do we turn that into mathematics? (Joint work with J. F. Gibson)
Tuesday, September 21, 2010 - 15:00 , Location: Skiles 255 , Professor Scott Armstrong , University of Chicago , Organizer: Ronghua Pan
We discuss how to solve a Hamilton-Jacobi-Bellman equation ``at resonance." Our characterization is in terms of invariant measures and is analogous to the Fredholm alternative in the linear case.
Tuesday, August 31, 2010 - 15:00 , Location: Skiles 114 , Professor Tai-Peng Tsai , Department of Mathematics, University of British Columbia , firstname.lastname@example.org , Organizer: Ronghua Pan
Consider a nonlinear Schrodinger equation in $R^3$ whose linear part has three or more eigenvalues satisfying some resonance conditions. Solutions which are initially small in $H^1 \cap L^1(R^3)$ and inside a neighborhood of the first excited state family are shown to converge to either a first excited state or a ground state at time infinity. An essential part of our analysis is on the linear and nonlinear estimates near nonlinear excited states, around which the linearized operators have eigenvalues with nonzero real parts and their corresponding eigenfunctions are not uniformly localized in space. This is a joint work with Kenji Nakanishi and Tuoc Van Phan.The preprint of the talk is available at http://arxiv.org/abs/1008.3581
Tuesday, April 20, 2010 - 15:00 , Location: Skiles 255 , Gianluca Crippa , University of Parma (Italy) , Organizer:
In the simplest form, our result gives a characterization of bounded,divergence-free vector fields on the plane such that the Cauchyproblem for the associated continuity equation has a unique boundedsolution (in the sense of distribution).Unlike previous results in this directions (Di Perna-Lions, Ambrosio,etc.), the proof does not rely on regularization, but rather on adimension-reduction argument which allows us to prove uniqueness usingwell-known one-dimensional results (it is indeed a variant of theclassical method of characteristics).Note that our characterization is not given in terms of functionspaces, but using a qualitative property which is completelynon-linear in character, namely a suitable weak formulation of theSard property.This is a joint work with Giovanni Alberti (University of Pisa) andStefano Bianchini (SISSA, Trieste).
Tuesday, April 13, 2010 - 15:05 , Location: Skiles 255 , Yao Li , Georgia Tech , Organizer: Zhiwu Lin
Fokker-Planck equation is a linear parabolic equation which describes the time evolution of of probability distribution of a stochastic process defined on a Euclidean space. Moreover, it is the gradient flow of free energy functional. We will present a Fokker-Planck equation which is a system of ordinary differential equations and describes the time evolution of probability distribution of a stochastic process on a graph with a finite number of vertices. It is shown that there is a strong connection but also substantial differences between the ordinary differential equations and the usual Fokker-Planck equation on Euclidean spaces. Furthermore, the ordinary differential equation is in fact a gradient flow of free energy on a Riemannian manifold whose metric is closely related to certain Wasserstein metrics. Some examples will also be discussed.
Tuesday, April 6, 2010 - 15:00 , Location: Skiles 255 , Clement Mouhot , Ecole Normale Superieure , Organizer: Wilfrid Gangbo
Landau damping is a collisionless stability result of considerable importance in plasma physics, as well as in galactic dynamics. Roughly speaking, it says that spatial waves are damped in time (very rapidly) by purely conservative mechanisms, on a time scale much lower than the effect of collisions. We shall present in this talk a recent work (joint with C. Villani) which provides the first positive mathematical result for this effect in the nonlinear regime, and qualitatively explains its robustness over extremely long time scales. Physical introduction and implications will also be discussed.
Tuesday, March 30, 2010 - 15:00 , Location: Skiles 255 , Ryan Hynd , University of California , Organizer: Michael Lacey
We discuss a non-linear eigenvalue problem where the eigenvalue has a natural control-theoretic interpretation as an optimal "long-time averaged cost." We also show how such problems arise in financial market models with small transaction costs.
Tuesday, March 9, 2010 - 15:00 , Location: Skiles 255 , Bob Pego , Carnegie Mellon University , Organizer: Chongchun Zeng
A classic story of nonlinear science started with the particle-like water wave that Russell famously chased on horseback in 1834. I will recount progress regarding the robustness of solitary waves in nonintegrable model systems such as FPU lattices, and discuss progress toward a proof (with Shu-Ming Sun) of spectral stability of small solitary waves for the 2D Euler equations for water of finite depth without surface tension.
Tuesday, March 2, 2010 - 15:05 , Location: Skiles 255 , Marius Paicu , Université Paris-Sud , Organizer: Zhiwu Lin
We consider the three dimensional Navier-Stokes equations with a large initial data and we prove the existence of a global smooth solution. The main feature of the initial data is that it varies slowly in the vertical direction and has a norm which blows up as the small parameter goes to zero. Using the language of geometrical optics, this type of initial data can be seen as the ``ill prepared" case. Using analytical-type estimates and the special structure of the nonlinear term of the equation we obtain the existence of a global smooth solution generated by this large initial data. This talk is based on a work in collaboration with J.-Y. Chemin and I. Gallagher and on a joint work with Z. Zhang.