Seminars and Colloquia by Series

Thursday, April 22, 2010 - 16:00 , Location: Skile 255 , Prof. Weiping Li , Oklahoma State University , Organizer: Haomin Zhou
Based on a sequence of discretized American option price processes under the multinomial model proposed by Maller, Solomon and Szimayer (2006), the sequence converges to the counterpart under the original L\'{e}vy  process in distribution for almost all time. We prove a weak convergence in this case for American put options for all time. By adapting Skorokhod representation theorem, a new sequence of approximating processes with the same laws with the multinomial tree model defined by Maller, Solomon and Szimayer (2006) is obtained. The new sequence of approximating processes satisfies Aldous' criterion for tightness. And, the sequence of filtrations generated by the new approximation converges to the filtration generated by the representative of L\'{e}vy process weakly. By using results of Coquet and Toldo (2007), we give a complete proof of the weak convergence for the approximation of American put option prices for all time.
Monday, March 15, 2010 - 11:00 , Location: Skiles 269 , Chao-Nien Chen , National Changhua University, Taiwan , Organizer: Yingfei Yi
There are many interesting patterns observed in activator-inhibitor systems. A well-known model is the FitzHugh-Nagumo system. In conjunction with calculus of variations, there is a close relation between the stability of a steady state and its relative Morse index. We give a sufficient condition in diffusivity for the existence of standing wavefronts joining with Turing patterns.
Monday, February 15, 2010 - 16:30 , Location: Skiles 255 , Vladimir Belykh , Nizhny Novgorod University , Organizer: Yingfei Yi
In this lecture, I will discuss a class of multidimensional maps with one nonlinearity, often called discrete-time Lurie systems. In the 2-D case, this class includes Lozi map and Belykh map. I will derive rigorous conditions for the multidimensional maps to have a generalized hyperbolic attractor in the sense of Bunimovich-Pesin. Then, I will show how these chaotic maps can be embedded into the flow, and I will give specific examples of three-dimensional piece-wise linear ODEs, generating this class of hyperbolic attractors.
Monday, February 1, 2010 - 11:00 , Location: 269 Skiles , Peter Kim , University of Utah , , Organizer:
We improved a computational model of leukemia development from stem cells to terminally differentiated cells by replacing the probabilistic, agent-based model of Roeder et al. (2006) with a system of deterministic, difference equations.  The model is based on the relatively recent theory that cancer originates from cancer stem cells that reside in a microenvironment, called the stem cell niche.  Depending on a stem cell’s location within the stem cell niche, the stem cell may remain quiescent or begin proliferating. This emerging theory states that leukemia (and potentially other cancers) is caused by the misregulation of the cycle ofproliferation and quiescence within the stem cell niche.Unlike the original agent-based model, which required seven hours per simulation, our model could be numerically evaluated in less than five minutes.  The results of our numerical simulations showed that our model closely replicated the average behavior of the original agent-based model. We then extended our difference equation model to a system of age-structured partial differential equations (PDEs), which also reproduced the behavior of the Roeder model.  Furthermore, the PDE model was amenable to mathematical stability analysis, which revealed three modes of behavior: stability at 0 (cancer dies out), stability at a nonzero equilibrium (a scenario akin to chronic myelogenous leukemia), and periodic oscillations (a scenario akin to accelerated myelogenous leukemia).The PDE formulation not only makes the model suitable for analysis, but also provides an effective mathematical framework for extending the model to include other aspects, such as the spatial distribution of stem cells within the niche.
Monday, November 30, 2009 - 11:00 , Location: Skiles 269 , Maria Lopez , Consejo Superior de Investigaciones Cientificas Madrid, Spain , Organizer: Yingfei Yi
We introduce a change of coordinates allowing to capture in a fixed reference frame the profile of travelling wave solutions for nonlinear parabolic equations. For nonlinearities of bistable type the asymptotic travelling wave profile becomes an equilibrium state for the augmented reaction-diffusion equation. In the new equation, the profile of the asymptotic travelling front and its propagation speed emerge simultaneously as time evolves. Several numerical experiments illustrate the effciency of the method.
Monday, November 16, 2009 - 11:00 , Location: Skiles 269 , Wei-Min Wang , Universite Paris-Sud, France , Organizer: Yingfei Yi
We present a new theory on Hamiltonian PDE. The linear theory solves an old spectral problem on boundedness of L infinity norm of the eigenfunctions of the Schroedinger operator on the 2-torus. The nonlinear theory develops Fourier geometry, eliminates the convexity condition on the (infinite dimension) Hamiltonian and is natural for PDE.
Monday, November 9, 2009 - 11:00 , Location: Skiles 269 , Dongwei Huang , Tianjin Polytechnic University, China and School of Mathematics, Georgia Tech , Organizer: Yingfei Yi
Many dynamical systems may be subject to stochastic excitations, so to find an efficient method to analyze the stochastic system is very important. As for the complexity of the stochastic systems, there are not any omnipotent methods. What I would like to present here is a brief introduction to quasi-non-integrable Hamiltonian systems and stochastic averaging method for analyzing certain stochastic dynamical systems. At the end, I will give some examples of the method.
Monday, November 2, 2009 - 11:00 , Location: Skiles 269 , Wen Huang , USTC, China and SoM, Georgia Tech , Organizer: Yingfei Yi
Stable sets and unstable sets of a dynamical system with positive entropy are investigated. It is shown that in any invertible system with positive entropy, there is a measure-theoretically ?rather big? set such that for any point from the set, the intersection of the closure of the stable set and the closure of the unstable set of the point has positive entropy. Moreover, for several kinds of specific systems, the lower bound of Hausdorff dimension of these sets is estimated. Particularly the lower bound of the Hausdorff dimension of such sets appearing in a positive entropy diffeomorphism on a smooth Riemannian manifold is given in terms of the metric entropy and of Lyapunov exponent.
Monday, October 19, 2009 - 11:00 , Location: Skiles 269 , Redouane Qesmi , York University, Canada and SoM, Georgia Tech , Organizer: Yingfei Yi
Despite advances in treatment of chronic hepatitis B virus (HBV) infection, liver transplantation remains the only hope for many patients with end-stage liver disease due to HBV. A complication with liver transplantation, however, is that the new liver is eventually reinfected in chronic HBV patients by infection in other compartments of the body. We have formulated a model to describe the dynamics of HBV after liver transplant, considering the liver and the blood of areas of infection. Analyzing the model, we observe that the system shows either a transcritical or a backward bifurcation. Explicit conditions on the model parameters are given for the backward bifurcation to be present, to be reduced, or disappear. Consequently, we investigate possible factors that are responsible for HBV/HCV infection and assess control strategies to reduce HBV/HCV reinfection and improve graft survival after liver transplantation.
Monday, September 28, 2009 - 11:00 , Location: Skiles 269 , Federico Bonetto , School of Mathematics, Georgia Tech , Organizer: Yingfei Yi

This talk continues from last week's colloquium.

Fourier's Law assert that the heat flow through a point in a solid is proportional to the temperature gradient at that point. Fourier himself thought that this law could not be derived from the mechanical properties of the elementary constituents (atoms and electrons, in modern language) of the solid. On the contrary, we now believe that such a derivation is possible and necessary. At the core of this change of opinion is the introduction of probability in the description. We now see the microscopic state of a system as a probability measure on phase space so that evolution becomes a stochastic process. Macroscopic properties are then obtained through averages. I will introduce some of the models used in this research and discuss their relevance for the physical problem and the mathematical results one can obtain.