AMSC Student Seminar

If you would like to give a talk at the seminar, please insert your name and talk title in the schedule at AMSC Student Seminar Schedule

Archives
Fall 05
Spring 05
Fall 04

Date	2:30 Talk	3:00 Talk
Sept 11, 2006	Joanna Pressley - Calculating the Firing Rate of Stochastic Integrate-and-Fire Neurons. Abstract: The brain is a massively complex network of more than 100 billion neurons, each making thousands of connections with other neurons. Neurons use pulse-like electrical signals called action potentials or "spikes" to encode information. Many scientists believe the rate of spikes, or the firing rate, contains a majority of the pertinent information transferred in a spike train. This talk will focus on two methods used to calculate the firing rate for a simplified model called the Stochastic Integrate-and-Fire Neuron.	Bob Shuttleworth - Parallel Block Preconditioners for the Incompressible Navier-Stokes Equations Abstract: We consider a preconditioner derived from a block factorization of the coefficient matrix that is generated in a Newton nonlinear iteration for the incompressible Navier-Stokes equations. We discuss a computational study performed using MPSalsa, a stabilized finite element code, in which parallel versions of our preconditioners is compared with a domain decomposition preconditioner. Our results show nearly ideal convergence rates for a wide range of Reynolds numbers on two and three-dimensional problems with both enclosed and in/out flow boundary conditions on both structured and unstructured meshes.
Sept 18, 2006	Elana Fertig - Improving forecasts for chaotic physical processes: using ensembles to estimate initial conditions Abstract: Accurate estimates of the initial state are necessary for accurate forecasts of chaotic systems, like the weather. These estimates of the Initial condition are obtained from combining forecast and observation information relative to their uncertainties. Algorithms estimating the initial conditions at meteorological centers (e.g., PSAS) assume a constant forecast uncertainty. The local ensemble transform Kalman filter (LETKF) algorithm uses an ensemble of forecasts to evolve forecast uncertainties efficiently. We apply LETKF and PSAS to the NASA weather model, fvGCM. LETKFs estimate of the initial conditions has fewer errors than those of PSAS.	Jeff Heath - Global Solutions to Model-Based Clustering Abstract: The Expectation-Maximization (EM) algorithm is a very popular optimization tool in model-based clustering problems. However, while the algorithm is convenient to implement, it only produces solutions that are locally optimal. We will introduce several new algorithms designed to produce global solutions in model-based clustering problems. The building blocks for these algorithms are methods from the operations research community, namely the Cross-Entropy (CE) method and Model Reference Adaptive Search (MRAS). We find that although a single run of the new algorithms is slower than EM, they have the potential of producing significantly better global solutions to the model-based clustering problem.
Sept 25, 2006	Scheduling conflict.	Carter Price - A Case Study in Simulation: Bed Flow in a Cardiac Surgery Post-Operative Unit Abstract: Hospitals, like all institutions, have limited resources. It is essential that these resources are used in an efficient manner. This talk presents a simulation model of a 30 bed cardiac surgery post-operative unit. The simulation used real world data to determine the optimal arrangement for these beds. The model, results, and further work will be discussed.
Oct 2, 2006	Weigang Zhong - Entropy Stable Approximations of Navier-Stokes Equations with No Artificial Numerical Viscosity. Abstract: We construct a new family of entropy stable difference schemes which retain the precise entropy decay of the Navier-Stokes equations. To this end we employ the entropy conservation differences to discretize Euler convective fluxes, and centered differences to discretize the dissipative fluxes of viscosity and heat conduction. The resulting difference schemes contain no artificial numerical viscosity in the sense that their entropy dissipation is dictated solely by viscous and heat fluxes. Numerical experiments provide a remarkable evidence for the different roles of viscosity and heat conduction in forming sharp monotone profiles in the immediate neighborhoods of shocks and contacts.	Bin Chen - On the long time regularity of the Shallow Water equations Abstract: We study the long time existence of classical solutions to the Shallow Water (SW) equations with pressure gradient and rotational forcing that are both singular within certain scaling regime of the Froude and Rossby numbers. The SW dynamics is then shown to be asymptotically close to the one governed by the 2-D pressureless rotational Euler equations with subcritical initial data, which in turn yields the increasingly long time existence at this singular regime. The novelty of our approach is the use of an approximate system that is linear while still capturing both the singularity and the advection dynamics of the underlying nonlinear system. The near periodic dynamics shown here is closely related to the circular fluid motions observed in geophysical sciences.
Oct 9, 2006	Abstract:	Abstract:
Oct 16, 2006	Abstract:	Abstract:
Oct 23, 2006	Abstract:	Abstract:
Oct 30, 2006	Abstract:	Abstract: