Georgia Institute of Technology College of Sciences

The three-dimensional Maxwell-Pauli-Coulomb (MPC) equations are a system of nonlinear, coupled partial differential equations describing the time evolution of a single electron interacting with its self-generated electromagnetic field and a static (infinitly heavy) nucleus of atomic number Z. The time local (and, hence, global) well-posedness of the MPC equations for any initial data is an open problem, even when Z = 0. In this talk we present some progress towards understanding the well-posedness of the MPC equations and, in particular, how the existence of solutions depends on the stability of the one-electron atom. Our main result is that time global finite-energy weak solutions to the MPC equations exist provided Z is less than a critical charge. This is an oral comprehensive exam. All are welcome to attend.

We will talk about discrete versions of the Bethe-Sommerfeld conjecture. Namely, we study the spectra of multi-dimensional periodic Schrödinger operators on various discrete lattices with sufficiently small potentials. In particular, we provide sharp bounds on the number of gaps that may perturbatively open, we characterize those energies at which gaps may open, and we give sharp arithmetic criteria on the periods that ensure no gaps open. We will also provide examples that open the maximal number of gaps and estimate the scaling behavior of the gap lengths as the coupling constant goes to zero. This talk is based on a joint work with Svetlana Jitomirskaya and another work with Jake Fillman.

I will discuss some free probability inequalities on the circle which can be seen in two different ways, one is via random matrix approximation, and another one by itself. I will show what I believe to be the key of these new forms, namely the fact that the circle acts on itself. For instance the Poincare inequality has a certain form which reflects this aspect. I will also briefly show how a transportation inequality can be discussed and how the standard Wasserstein distance can be modified to introduce this interesting phenomena. I will end the talk with a conjecture and some supporting evidence in the classical world of functional inequalities.

We consider a class of nonlinear, degenerate drift-diffusion equations in R^d. By a scaling argument, it is widely believed that solutions are uniformly Holder continuous given L^p-bound on the drift vector field for p>d. We show the loss of such regularity in finite time for p≤d, by a series of examples with divergence free vector fields. We use a barriers argument.