Past

The advent of multicore processors and other technologies like Graphical Processing Units (GPU) will considerably influence future research in High Performance Computing.

To take advantage of these architectures in dense linear algebra operations, new algorithms are

proposed that use finer granularity and minimize synchronization points.

After presenting some of these algorithms, we address the issue of pivoting and investigate randomization techniques to avoid pivoting in some cases.

In the particular case of GPUs, we show how linear algebra operations can be enhanced using

hybrid CPU-GPU calculations and mixed precision algorithms.

In order to understand the nonlinear stability of many types of time-periodic travelling waves on unbounded domains, one must overcome two main difficulties: the presence of embedded neutral eigenvalues and the time-dependence of the associated linear operator. This problem is studied in the context of time-periodic Lax shocks in systems of viscous conservation laws. Using spatial dynamics and a decomposition into separate Floquet eigenmodes, it is shown that the linear evolution for the time-dependent operator can be represented using a contour integral similar to that of the standard time-independent case. By decomposing the resulting Green's distribution, the leading order behavior associated with the embedded eigenvalues is extracted. Sharp pointwise bounds are then obtained, which are used to prove that the time-periodic Lax shocks are linearly and nonlinearly stable under the necessary conditions of spectral stability and minimal multiplicity of the translational eigenvalues. The latter conditions hold, for example, for small-oscillation time-periodic waves that emerge through a supercritical Hopf bifurcation from a family of time-independent Lax shocks of possibly large amplitude.

I show that the expansion of the real field by a total Pfaffian chain is model complete in a language with symbols for the functions in the chain, the exponential and all real constants. In particular, the expansion of the reals by all total Pfaffian functions is model complete.

The homological Hall lemma is a topological tool that has recently been used to derive Hall type theorems for systems of disjoint representatives in hypergraphs.

After outlining the general method, we.ll describe one such theorem in some detail. The main ingredients in the proof are:

1) A relation between the spectral gap of a graph and the topological connectivity of its flag complex.

2) A new graph domination parameter defined via certain vector representations of the graph.

Joint work with R. Aharoni and E. Berger

In this talk, we present an extension of the theory of rough paths to partial differential equations. This allows a robust approach to stochastic partial differential equations, and in particular we can replace Brownian motion by more general Gaussian and Markovian noise. Support theorems and large deviation statements all become easy corollaries of the corresponding statements of the driving process. This is joint work with Peter Friz in Cambridge.

It is a classical result that the Alexander polynomial of a fibered knot has to be monic. But in general the converse does not hold, i.e. the Alexander polynomial does not detect fibered knots. We will show that the collection of all twisted Alexander polynomials (which are a natural generalization of the ordinary Alexander polynomial) detect fibered 3-manifolds.

As a corollary it follows that given a 3-manifold N the product S1 x N is symplectic if and only if N is fibered.

A new general approach to optimal stopping problems in L\'evy models, regime switching L\'evy models and L\'evy models with stochastic volatility and stochastic interest rate is developed. For perpetual options, explicit solutions are found, for options with finite time horizon, time discretization is used, and explicit solutions are derived for resulting sequences of perpetual options.

The main building block is the option to abandon a monotone payoff stream. The optimal exercise boundary is found using the operator form of the Wiener-Hopf method, which is standard in analysis, and interpretation of the factors as {\em expected present value operators} (EPV-operators) under supremum and infimum processes.

Other types of options are reduced to the option to abandon a monotone stream. For regime-switching models, an additional ingredient is an efficient iteration procedure.

L\'evy models with stochastic volatility and/or stochastic interest rate are reduced to regime switching models using discretization of the state space for additional factors. The efficiency of the method for 2 factor L\'evy models with jumps and for 3-factor Heston model with stochastic interest rate is demonstrated. The method is much faster than Monte-Carlo methods and can be a viable alternative to Monte Carlo method as a general method for 2-3 factor models.

Joint work of Svetlana Boyarchenko,University of Texas at Austin and Sergei Levendorski\v{i},

University of Leicester

We discuss the valuation problem for a broad spectrum of derivatives, especially in Levy driven models. The key idea in this approach is to separate from the computational point of view the role of the two ingredients which are the payoff function and the driving process for the underlying quantity. Conditions under which valuation formulae based on Fourier and Laplace transforms hold in a general framework are analyzed. An interesting interplay between the properties of the payoff function and the driving process arises. We also derive the analytically extended characteristic function of the supremum and the infimum processes derived from a Levy process. Putting the different pieces together, we can price lookback and one-touch options in Levy driven models, as well as options on the minimum and maximum of several assets.

The large-time behavior of solutions to Burgers equation with

small viscosity is described using invariant manifolds. In particular,

a geometric explanation is provided for a phenomenon known as

metastability,which in the present context means that

solutions spend a very long time near the family of solutions known as

diffusive N-waves before finally converging to a stable self-similar

diffusion wave. More precisely, it is shown that in terms of

similarity, or scaling, variables in an algebraically weighted $L^2$

space, the self-similar diffusion waves correspond to a one-dimensional

global center manifold of stationary solutions. Through each of these

fixed points there exists a one-dimensional, global, attractive,

invariant manifold corresponding to the diffusive N-waves. Thus,

metastability corresponds to a fast transient in which solutions

approach this ``metastable" manifold of diffusive N-waves, followed by

a slow decay along this manifold, and, finally, convergence to the

self-similar diffusion wave.

(joint work with E. Hrushovski and A. Pillay)

If G is a definably compact, connected group definable in an o-minimal structure then, as is known, G/Z(G) is semisimple (no infinite normal abelian subgroup).

We show, that in every o-minimal expansion of an ordered group:

If G is a definably connected central extension of a semisimple group then it is bi-intepretable, over parameters, with the two-sorted structure (G/Z(G), Z(G)). Many corollaries follow for definably connected, definably compact G.
Here are two:

1. (G,.) is elementarily equivalent to a compact, connected real Lie group of the same dimension.

2. G can be written as an almost direct product of Z(G) and [G,G], and this last group is definable as well (note that in general [G,G] is a countable union of definable sets, thus not necessarily definable).