Application of critical point theory to problems in partial differential equations

Speaker: 

Martin Schechter

Institution: 

UC Irvine

Time: 

Thursday, April 23, 2009 - 3:00pm

Location: 

RH 340P

For many partial differential equations and systems that arise in applications, solutions are critical points of corresponding functionals. One can solve such problems by finding the critical points. We discuss various techniques for finding them and apply the methods to specific problems. The talk can also be followed by nonspecialists and students.

Multilevel Methods for Viscoelastic Fluids Simulation

Speaker: 

Chowla Assistant Professor Chensong Zhang

Institution: 

Penn State University

Time: 

Monday, May 11, 2009 - 4:00pm

Location: 

RH 306

The link between various viscoelastic fluids models and the symmetric matrix Riccati differential equations can be a new device that brings to unified and proper ways of numerical treatments for the viscoelastic models. In this talk, we describe a few steps toward efficient numerical schemes for complex fluids simulation. First, we construct stable finite element discretizations using Eulerian--Lagrangian methods based on the Riccati formulations of the viscoelastic models. Then we develop a new multilevel time-marching scheme; together with adaptive time-stepping schemes and time parallel schemes, we can build efficient methods for complex fluids simulation. Furthermore, we discuss two robust and efficient multilevel solvers for Stokes-type systems arising at each time step in the Eulerian--Lagrangian discretization.

Finite Markov Chains and the Casino Game Craps

Speaker: 

Chris O'Grain

Institution: 

UCI

Time: 

Wednesday, March 4, 2009 - 5:00pm

Location: 

RH 594

In any casino, the craps table often has the loudest and most excited crowd. This is because when the shooter wins, everyone wins. Its also because craps offers some of the best odds in the entire casino, giving players the best chance to win money. In this talk, the rules of craps will be explained. Also, the concept of markov chains will be introduced, and a finite markov chain will be used to model the game of craps.

Elliptic and Parabolic Equations with rough coefficients in Sobolev Spaces

Speaker: 

Doyoon Kim

Institution: 

University of Southern California

Time: 

Thursday, May 7, 2009 - 3:00pm

Location: 

RH 340P

The unique solvability of second order elliptic and parabolic equations (in either divergence form or non-divergence form) in Sobolev spaces is well known if the leading coefficients are, for example, uniformly continuous.
However, in general, it is not possible to solve equations in Sobolev spaces unless the coefficients have some regularity assumptions.
In this talk we will discuss some possible classes of discontinuous coefficients with which elliptic and parabolic equations are uniquely solvable in Sobolev spaces.
Especially, as our main results, we will focus on the unique solvability of equations with coefficients only measurable in one spatial variable and having small mean oscillations in the other variables (called partially BMO coefficients).
We will also discuss some applications of our results as well as a new approach to a priori L_p estimates.
Most of the talk is based on joint work with Nicolai Krylov and with Hongjie Dong.

Image denoising/deblurring with BV and homogeneous Sobolev spaces

Speaker: 

Yunho Kim

Institution: 

UCLA

Time: 

Thursday, March 5, 2009 - 3:00pm

Location: 

RH 340P

Given a blurry image, the goal is to find the most clear image. There are many methods to solve this inverse problem in the case of cartoon images containing rather piecewise smooth objects. However, in the presence of oscillations the blurring process removes those oscillations in the images and that makes this inverse problem harder to solve. We approach this problem by minimizing a convex functional whose domain is the product of the space of functions of bounded variation and the homogeneous Sobolev space. As we will see, the homogeneous Sobolev space turns out to be a good space to capture oscillations. We will talk about the existence of a minimizer and characterization of the minimizers and PDE based numerical scheme and then briefly discuss a noisy case. If time permits, we will also talk about a medical image denoising application.

Pages

Subscribe to UCI Mathematics RSS