Computational  and Applied Mathematics Seminar

The CAM Seminar is organized  in the ISU Mathematics Department. It brings speakers from inside and outside of  ISU, raising issues and exchanging ideas on topics of current interest in the area of computational and applied mathematics.
 

Past semesters:   Fall 2003 || CM Seminar Fall 2002

The seminar meets every Monday  at 3:10 -4:00 pm in Room   124 /294 Carver.  

SCHEDULE
 

DATE	TIME & LOCATION	SPEAKER	TITLE	NOTE
01/26	3:10PM 124 Carver	Shiquan Wu   Center of Bioinformatics and Biological Statistics at ISU	Multiple genome rearrangement: Model, algorithm and program	abstract
02/3	4:10PM 408 Carver	Peter Smereka U Michigan	Quasicontinuum Monte Carlo: A computational method for surface growth calculations.	abstract Host:  J.  Evans (Joint with  Dept. Colloquium)
02/9	3:10PM 124 Carve	Masha Sosonkina Ames Lab	Parallel solution of general sparse linear systems	abstract
02/16	3:10PM  Carver  124	Sergio Guerrero University of Seville, Spain	Remarks on exact controllability of the Navier-Stokes equations	Host: O. Emanouvilov
02/23	3:10PM  Carver 294	Hassan Manouzi University of  Laval, Canada	Solving Wick-stochastic pressure equation using a mixed finite element	Host:  S. Hou
03/1	3:10PM  Carver 294	Hailiang Liu ISU	Computation of Particles Vs. Waves
03/8	3:10PM  Carver 294	Rami  Tzafriri MIT	Modeling and Optimization of Drug Delivery to Solid Tumors	Host:   H.   Levine abstract
03/22	3:10PM  Carver 294	Lina Wang University of Lund, Sweden	cancelled !
03/29	3:10PM  Carver 294	Marshall Slemrod Univ. of Wisconsin (Madison)	The Radio-Frequency Driven Plasma Sheath: Asymptotics and Analysis	Host:  H.   Liu abstract
04/12	3:10PM  Carver 294	Bob Glassey Indiana University	THE CAUCHY PROBLEM FOR THE RELATIVISTIC VLASOV-POISSON SYSTEM	Host:  H. Levine abstract
04/19	3:10PM  Carver 294	Jianhong (Jackie) Shen University of Minnesota	Mathematical Image "No" Processing	Host: H. L. Liu abstract

For more information, or you want to offer a lecture please contact Hailiang Liu at hliu@iastate.edu or Scott Hansen at shansen@iastate.edu or Paul Sacks at psacks@iastate.edu.
 

Abstract (1/26  Shiquan Wu )
 Genome rearrangement is an important problem in computational genomics. The purpose is to infer the evolutionary history for a given
collection of species. In this talk, a mathematical model is proposed for genome rearrangement by reversal. The problem is NP-hard. It turns out
that it can only be solved by approximation algorithms. Several approximation algorithms were designed to search the optimal evolutionary tree
for the problem, including grid search algorithm, greedy split algorithm, and  neighbor-perturbing algorithm. A computer program is implemented
 to run for real biological data.

Abstract (02/03 Peter Smereka )
We introduce an algorithm for treating growth on surfaces which combines important features of continuum methods and Kinetic Monte Carlo (KMC)  simulations. We treat the motion of adatoms in continuum theory, but attach them to islands one atom at a time. Our method allows us to give a realistic account of fluctuations in island shape, which is lacking in deterministic continuum  treatments and which is an important physical effect. Our method should be most  important for problems close to equilibrium where KMC becomes impractically slow.

Abstract (02/09 Masha Sosonkina)
Parallel iterative methods are a typical choice for the solution of very large sparse linear systems, which require distributed computational resources. In this talk we describe a recently developed suite of parallel iterative methods, parallel Algebraic Recursive  Multilevel Solver (pARMS). The package contains a vast
array of parallel preconditioners, with the underlying concepts ranging from the widely-used domain decomposition methods to multilevel techniques. A single framework, which encompasses all the preconditioner construction, has been made possible due to the "fully distributed sparse linear system" design principle and the adoption of a two-level interface nodes approach. Along with presenting underlying concepts of preconditioner construction, we will show the results of a few numerical experiments.

Abstract  (03/08 Rami  Tzafriri )
Discrepancy between drug potency as observed in the laboratory and clinical efficacy is a recurrent problem in pharmaceutical science and is one of the
major stumbling blocks in effective drug design. The last three decades have witnessed the development of sophisticated drug targeting controlled release
technologies for overcoming and circumventing the "physiological barriers" to drug delivery. While these technologies have already impacted the
treatment of many diseases, there is growing recognition in both academia and industry that the prevailing trial an error design of drug delivery
techniques is a serious limiting factor and mathematical modeling has been suggested as an important tool in the design of drug delivery protocols.

In this talk I will focus on the problem of drug delivery to solid tumors. Intratumoral delivery has long been advocated as a superior alternative to
systemic delivery, but progress has been hampered by the underlying complexity of this problem. We used a reaction-diffusion model to analyze
this problem and suggest an optimal protocol that is easily achieved using current technology. Encouraging preliminary animal experiments will be
presented.

Abstract  (03/29 Marshall Slemrod )
This talk considers the dynamics of a radio-frequency driven plasma consisting of ions and electrons.
The method of matched asymptotic expansions is used to derive the dynamics in bulk quasi-neutral plasma, transition, and sheath  regions.
Furthermore a constructive existence theorem is presented  for solutions of the system governing sheath dynamics.

Abstract (04/12  Bob Glassey)
The motion of a collisionless plasma is modeled by solutions to the Vlasov-Maxwell system. In the presence of very large velocities, relativistic corrections are meaningful. When magnetic effects are ignored this formally becomes the relativistic Vlasov-Poisson equation. This talk will cover the large-data global solvability of the Cauchy Problem in 3D in the case of cylindrically symmetric initial values for the phase-space density and fields.

Abstract (04/19  J. Shen)

To understand what images really are is what I called the fundamental problem of image processing (SIAM News, 2003), an area mathematicians have been continuously making contributions to in the past two decades.

Images are functions: regular, irregular, generalized (i.e. distributions), or even random fields, while image processors are operators: linear or nonlinear, and often in the nature of inverse problem solvers. Design of efficient image processors therefore  crucially depends upon how images are properly modeled: by Sobolev functions? functions with bounded variations (BV)? or something else. The answers are intimately connected to real analysis, harmonic analysis, geometric measure theory, and free-boundary problems. In this talk I am delighted to share with you some of my personal reviews and views on this topic.