Rice University Rice University SIAM student chapter Shell Home
Calendar of Activities
Join
More info
Contact us

Texas Student SIAM Conference 2008 at Rice University

This year's Texas Student SIAM Conference will be held at Rice University, Duncan Hall on April 11-12, 2008. The conference will start at 2:30 pm on April 11 and finish at 4:00 pm on April 12. Registration is now open. There are speakers slots as well as a poster session. If you wish to participate in either or (hopefully!) both, please email your name, institution and abstract to student.siam.texas@gmail.com by April 4, 2008. This conference is open to all and any with a mathematical leaning. Come meet your peers from across the state. Finally, we would like to give a special thanks to Shell for providing the funding to make this conference possible.

For accomodations, visit explore rice

For driving directions to Duncan Hall, please see the map below


View Larger Map

For information on parking, please visist Rice parking

Schedule on April 11, 2008

2:30-3:00 pm
Registration & Welcoming Remarks
3:00-3:30 pm
Arjun Beri (University of Houston)
Abstract: A currency option is contingent on the exchange rate whose dynamics can be modeled by a (or a system of) stochastic differential equation. We consider the INR/USD exchange rate process and fit a geometric Brownian motion as well as a stochastic volatility model to the series. The parameters are estimated using a hybrid of Method of Moments and Maximum Likelihood estimation. We price an American option using Stochastic Mesh method and demonstrate calculation of greeks.
3:30-4:00 pm
Ricardo Alonso (University of Texas)
Title: Existence of global solutions to the Cauchy problem for the inelastic Boltzmann equation with near-vacuum data
Abstract: The Cauchy problem for the inelastic Boltzmann equation is studied for small data. Existence and uniqueness of mild and weak solutions is obtained for sufficiently small data that lies in the space of functions bounded by Maxwellians. The technique used to derive the result is the well known iteration process of Kaniel & Shinbrot.
4:00-4:30 pm
Tony Kellems (Rice University)
Abstract: Realistic compartmental models of active neurons yield nonlinear dynamical systems with thousands of equations. We linearize such systems and build their balanced truncations where B permits synaptic input into each compartment and C observes only the soma potential. Using true morphologies with a broad class of synaptic inputs, we find that the full quasi-active somatic dynamics can be approximated to nearly 10 digits by reduced systems of dimension approximately two orders of magnitude smaller. This savings will permit, for the first time, one to simulate large networks of biophysically accurate cells over realistic time spans.
4:30-5:00 pm
Scott Lipton (University of Texas)
Title: T-splines and Isogeometric Analysis
Abstract: Isogeometric Analysis is a recent development in computational mechanics that bridges the gap between CAD and CAE. By equating the original geometry with the mesh, an exact geometric description is maintained throughout the design through analysis process. In this way Isogeometric Analysis overcomes challenges commonly encountered in traditional finite element analysis. T-splines [Sederberg et al. 2003], a recently proposed geometric description, relaxes constraints imposed by traditional modeling schemes and further enhances the effectiveness of the Isogeometric approach. The T-spline method allows for greater flexibility and control during the design process as well as adaptivity and local refinement during the analysis phase.
5:00-6:00 pm
Poster Session in DH 3092
6:00-
Valhalla and Food

Schedule on April 12, 2008

8:30-9:00 am
Breakfast from Brother's Taco's
9:00-9:30 am
Dr. Sean Yang (BP)
9:30-10:00 am
Pearl Flath (University of Texas)
Title: Fast algorithms for uncertainty estimation and propagation in large scale linear dynamical systems
Abstract: We consider the problem of estimating and propagating the uncertainty in the initial conditionfield of a convection-diffusion problem describing the transport of atmospheric contaminants. Estimation of the uncertainty is treated within a Bayesian framework. Standard Markov chain Monte Carlo approaches are intractable for such high-dimensional problems. Even when the data and prior uncertainty are Gaussian, and as a result the posterior estimate is Gaussian with covariance given by the inverse of the Hessian matrix of the regularized least squares objective, the computation of the exact covariance matrix is intractable due to the large size and extreme cost of forming the inverse of the Hessian. In the case of linear ill-posed inverse probems, we show that fast algorithms can be constructed that provide accurate low rank Hessian approximations of the least squares data misfit, and as a result permit estimation and propagation of the uncertainty for large-scale problems at a small multiple of the cost of solving the forward problem. Large-scale examples demonstrate the main ideas.
10:00-10:30 am
Michael Costello (Shell)
10:30-11:00 am
John Evans (University of Texas)
Title: Approximation results for the k-version of the finite element method
Abstract: The k-method is a finite element technique where spline basis functions of higher-order continuity are employed. It is a fundamental feature of the new field of isogeometric analysis. In previous works, it has been shown that using the k-method has many advantages over classical finite element technologies. In application areas such as structural dynamics, wave propagation, and turbulence, the k-method has offered almost spectral approximation properties. The Kolmogorov n-width and sup-inf were introduced as tools to assess the effectiveness of approximating functions. In this talk, I will discuss the approximation properties of the k-method with these tools. Following a review of theoretical results, I will show results of the numerical study in which the n-width and sup-inf are computed for a number of one-dimensional cases. This study sheds further light on the approximation properties of the k-method. I will finish with a comparison study of the k-method and the classical finite element method and a short discussion on polynomial approximation.
11:00-11:30 am
Amr El-Bakry (Exxon)
11:30-12:00 am
Horacio Florez (University of Texas)
Title: Geomechanics in near borehole simulations: a practical approach
Abstract: We are about to present the key ingredients to address geomechanical problems from the near borehole point of view by considering a practical approach in the sense that field data has to be taken into account. A brief discussion about what does geomechanics mean and its today challenges will be given. We are going to discuss the well geometry description based on NURBS curves and surfaces, structured mesh generation around a borehole, and how to solve the poro-elasticity governing equations with FEM by using an Object Oriented approach. Some considerations about developing software in this area will be also included.
12:00-1:30 pm
Lunch will be provided from Desert Gallery
1:30-2:15 pm
ELA Conference - Dr. Clint Dawson
Computational Science and Hurricane Storm Surge
2:30-3:00 pm
Nicola Cavallini (University of Houston)
Title: A Novel Partitioned Scheme via Operator Splitting in Blood Flow Simulations.
Abstract: n this work we present the latest application to fluid structure interaction in arterial blood flow of a novel operator splitting technique. Partitioned algorithms are based on the idea of splitting the full problem in simpler sub-problems which summed give the original full problem. The coupling at the interface between blood and arterial wall is highly nonlinear because the ratio between fluid and structure densities is closed to one; this is the reason why many partitioned schemes successfully applied in aeroelasticity fail when applied to blood flow simulations. Our partitioned scheme, based on a time discretization via operator splitting, uses the kinematic condition at the interface to couple fluid and structure velocities at each sub-step. This original use of the kinematic condition in the partitioned scheme allows to successfully solve the interaction between blood flow and wall motion in arteries.n this work we present the latest application to fluid structure interaction in arterial blood flow of a novel operator splitting technique. Partitioned algorithms are based on the idea of splitting the full problem in simpler sub-problems which summed give the original full problem. The coupling at the interface between blood and arterial wall is highly nonlinear because the ratio between fluid and structure densities is closed to one; this is the reason why many partitioned schemes successfully applied in aeroelasticity fail when applied to blood flow simulations. Our partitioned scheme, based on a time discretization via operator splitting, uses the kinematic condition at the interface to couple fluid and structure velocities at each sub-step. This original use of the kinematic condition in the partitioned scheme allows to successfully solve the interaction between blood flow and wall motion in arteries.
3:00-3:30 pm
Robert Rosenbaum (University of Houston)
Title: Correlation Propagation in Networks of Integrate-and-Fire Neurons
Abstract: Correlations between spike trains are widely believed to play a role in the encoding of neural information. We investigate the degree to which integrate-and-fire neuron models preserve spike train correlations. We show that perfect integrate-and-fire neurons reduce synchrony, but preserve spike count correlation coefficients perfectly. Leaky models, though, reduce both synchrony and correlation. These results suggest an analytical and geometric interpretation of the covariation of rate and correlation complementary to that recently pointed out by J. la Rocha, B. Doiron et al.
3:45-4:15 pm
Closing Remarks
3:45-4:15
ELA Conference - Dr. Illya Hicks (Rice University)
Min-MAX Relationships in Combinatorial Optimization
4:15-5:00 pm
ELA Conference - Student Panel on Graduate Education and Research