NA Digest Sunday, March 16, 2003 Volume 03 : Issue 11

Today's Editor:
Cleve Moler
The MathWorks, Inc.

Submissions for NA Digest:

Mail to

Information via e-mail about NA-NET: Mail to


From: Alex Pothen <>
Date: Wed, 12 Mar 2003 22:25:04 -0500 (EST)
Subject: Data Structures in Scientific Computing

I am considering a request to write a survey article on
interesting data structures in scientific computing
and on creative uses of data structures in this context.
If you have suggestions, please send me an email.
I will be happy to briefly report here on what I learn.

Alex Pothen


From: C. G. Broyden <>
Date: Thu, 13 Mar 2003 10:47:57 +0100
Subject: Seeking References for Two Linear Equations Algorithms

Dear Colleagues,
I am seeking information on two methods for solving linear
equations. The first is that of von Mises, which I came across in a paper
written in Italian in 1937. No citation was given and I have been unable to
find any reference to it either in any of the standard linear algebra texts
or by searching the Internet. I have found references to a method of von
Mises on determining eigenvalues but nothing on solving linear equations.
The other is a method of J. Morris cited in the celebrated Fox,
Huskey and Wilkinson paper as "A successive approximation process for
solving simultaneous linear equations, A.R.C.R. & M. No. 1711 (1936)".
Any information on either of these methods, and on the place of
publication for the second one, would be greatly appreciated.
Thanks in advance, C. G. Broyden

C. G. Broyden,
Universita' di Bologna,
Corso di Laurea in Scienze dell'Informazione,
via Sacchi N.3,
47023 Cesena (FO)


From: Sriram Vishwanath <>
Date: Mon, 10 Mar 2003 19:08:05 -0800 (PST)
Subject: Continuous LPs: Can you solve them?

I apologize if this is a very simple question being posted on this
forum. For the first time, I have been faced with a problem of the form

max_p(x) integral f(x) p(x) dx

such that
integral x^2 p(x) dx = P
integral p(x) dx = 1
p(x) > = 0

where f(x) is some infinitely differentiable "nice" function,
p(x) is a probability distribution with infinite support, and P is a
constant greater than zero. I am baffled by it, and have no tools that I
know to solve it.

For what class of f(x) can this be solved in closed form? Is there
a better way of solving this than discretization and then using a
discrete LP algorithm?

Thank you so much,


From: Julie Haenisch <>
Date: Wed, 12 Mar 2003 11:19:27 -0500
Subject: New Book, Analytic Theory of Global Bifurcation

Analytic Theory of Global Bifurcation
Boris Buffoni and John Toland

Rabinowitz's classical global bifurcation theory, which concerns the study
in-the-large of parameter-dependent families of nonlinear equations, uses
topological methods that address the problem of continuous parameter
dependence of solutions by showing that there are connected sets of
solutions of global extent. Even when the operators are infinitely
differentiable in all the variables and parameters, connectedness here
cannot in general be replaced by path-connectedness. However, in the context
of real-analyticity there is an alternative theory of global bifurcation due
to Dancer, which offers a much stronger notion of parameter dependence.

Cloth | 2003 | $45.00 / =A329.95 | ISBN: 0-691-11298-3
180 pp. | 6 x 9 | 5 line illus.

To read more, click here:

Julie F. Haenisch
Text Promotion Manager
Princeton University Press


From: Larry Nazareth <>
Date: Thu, 13 Mar 2003 13:33:24 -0800 (PST)
Subject: New Book on Optimization

The following book has just been published by Springer-Verlag, NY:

Algorithmic Science and the Karmarkar Revolution,'' J. L. Nazareth.
Series: CMS Books in Mathematics, Vol. 13. 256 pgs., hardcover.
Springer-Verlag, New York, 2003 (
ISBN: 0-387-95572-0

For a brief summary and cover image, please check the Springer website
and search, for example, on the ISBN. The book has five parts, each
containing three chapters. An abbreviated table of contents follows:

I: Foundations
1. The Karmarkar Revolution.
2. The Newton-Cauchy Method.
3. Euler-Newton and Lagrange-NC Methods.
II: Lessons from One Dimension
4. A Misleading Paradigm.
5. CG and the Line Search.
6. Gilding the Nelder-Mead Lily.
III: Choosing the Right Diagonal Scale
7. Historical Parallels.
8. LP from the Newton-Cauchy Perspective.
9. Diagonal Metrics and the QC Method.
IV: Linear Programming Post-Karmarkar
10. LP from the Euler-Newton Perspective.
11. Log-Barrier Transformations.
12. Karmarkar Potentials and Algorithms.
V: Algorithmic Science
13. Algorithmic Principles.
14. Multialgorithms: A New Paradigm.
15. An Emerging Discipline.


From: Wolfgang Nagel <>
Date: Mon, 10 Mar 2003 17:46:25 +0100 (CET)
Subject: Conference in Dresden on Parallel Computing

(EXTENDED DEADLINE: March 15, 2003)
Parallel Computing 2003 Conference
September 2-5, 2003
Technical University Dresden

ParCo2003 is to be held at the Technical University Dresden, Germany,
from September 2 to 5, 2003. Starting in 1983 in Berlin, this series of
international conferences is the longest running on Parallel Computing
in Europe.

The aim of the conference is to give an overview of the state of the art
and future trends in Parallel Computing for all platforms, including
grids and clusters. The conference addresses all aspects of parallel
computing, including applications, hardware and software technologies as
well as languages, tools, and development environments. Special emphasis
will be placed on the role of parallel processing in solving real life
problems in all areas, including multidisciplinary applications, practical
experiences, and evaluations.

We invite you to submit an EXTENDED ABSTRACT of not less than two A4 size
pages BY MARCH 15, 2003. Each submission will be refereed by 3 members
of the Program Committee. After acceptance, submission of full papers
will be required by July 15, 2003.

There is also a call to send proposals for MINISYMPOSIA, which will run
as a separate track during the conference. These may run for half a day
or a full day. Proposals giving the title and a short (half-page) outline
of the symposium should also be submitted to the Program Committee via
the conference e-mail address NO LATER THAN MARCH 15, 2003.
Minisymposia should cover a particular problem area. Speakers are to
be invited by the minisymposium organizers.

Important Dates:

Final Submission Deadline
(Extended Abstracts of not less than two pages): March 15, 2003
Final Submission Deadline
(Proposals for Minisymposia): March 15, 2003

Acceptance Notification: April 22, 2003

Camera Ready Papers Deadline: July 15, 2003

The official Call for Papers can be found at:

Information about the conference series is available at the official
website of ParCo:

If you have the opportunity to distribute the Final CfP via your local
mailing lists, please feel encouraged to do so.

Yours sincerely

Wolfgang E. Nagel
Head of Program Committee


From: Marc Prevost <>
Date: Mon, 10 Mar 2003 11:17:23 +0100
Subject: Seminar in Calais on Industrial Problems

Last annoucement:

The University of Littoral, announces:
(new deadline: April, 15th 2003)
The Fourth Seminar on Numerical Algorithms Applied to Industrial
May 15-16, 2003. Calais, FRANCE

This conference will be a forum for the discussion and the presentation
of new developpements in numerical matrix analysis and applications in
industrial problems.


- Numerical linear algebra.
- Large sparse and dense linear systems.
- Applications to industrial problems.


- Iain DUFF, RAL Oxfordshire, UK.
- Gerard MEURANT, CEA Bruyeres le Chalet, FRANCE.
- Yousef SAAD, Universite du Minnesota, Minneapolis, USA.
- Henk VAN DER VORST, Universite D'Utrecht, NETHELANDS.
- Paul VAN DOOREN, Universite Catholique de Louvain, BELGIUM.


From: Douglas Arnold <>
Date: Wed, 12 Mar 2003 18:48:46 -0600
Subject: IMA Short Course on Mathematical Biology

This is an opportunity for mathematicians to efficiently get a jump
start in mathematical biology (at no cost). From June 16-27, 2003 the
Institute for Mathematics and its Applications (IMA) in Minneapolis
will host a two-week intensive short course designed to efficiently
provide mathematicians the basic knowledge prerequisite to undertaking
interdisciplinary research in mathematical biology at the cellular
level. The course in Cellular Physiology will be taught by James
Keener, Professor of Mathematics and Adjunct Professor of
Bioengineering at University of Utah and author of "Mathematical
Physiology" and Alexander Mogilner, Professor and Chancellor's Fellow
at the Department of Mathematics and Center for Genetics and
Development at University of California at Davis. Prominent biologists
will be involved in several sessions as well. Participants will receive
full travel and lodging support during the workshop.

The goal of the course will be to prepare qualified participants to
start collaborative interdisciplinary research in the area of cellular
physiology, a vital part of math biology, and one in which mathematical
techniques are greatly needed and research opportunities abound.
Participants will gain an understanding the key mathematical issues in
the topic, some familiarity with the relevant literature, ideas about
problems to whose resolution they can contribute, and the basic
knowledge necessary to initiate meaningful interdisciplinary
collaborations in the field.

Details and application materials are at


From: Darrell Ross <>
Date: Thu, 13 Mar 2003 13:37:19 -0500
Subject: SIAM Conference in San Francisco on Data Mining

SIAM International Conference on Data Mining
Cathedral Hill Hotel, San Francisco, CA
May 1-3, 2003

Reminder, the deadlines Preregistration and Hotel Registration are fast

The Preregistration and Hotel Registration deadlines are on Wednesday,
April 2, 2003. SAVE and register now!

For additional information, contact SIAM Conference Department at

Darrell Ross
SIAM, Conference Program Manager
Conference Web Master


From: Laurence Yang <>
Date: Thu, 13 Mar 2003 17:31:48 -0400
Subject: Workshop in Taiwan on High Performance Computing


The 5th Workshop on High Performance Scientific and Engineering Computing with
Applications (HPSECA-03)

Kaohsiung, Taiwan, ROC
October 6-9 (Monday-Thursday), 2003

in conjunction with

Scope and Interests:

Parallel and distributed scientific and engineering computing has become a
key technology which will play an important part in determining, or at least
shaping, future research and development activities in many academic and
industrial branches. This special workshop is to bring together computer
scientists, applied mathematicians and researchers to present, discuss and
exchange idea, results, work in progress and experience of research in the
area of parallel and distributed computing for problems in science and
engineering applications.

Among the main topics (but not limited to) are:

development of advanced parallel and distributed methods,
parallel and distributed computing techniques and codes,
practical experiences using various supercomputers with software such as MPI,
PVM, and High Performance Fortran, OpenMP, etc.
Cluster and grid computing
applications to the following areas, but not limited to:
computational fluid dynamics and mechanics
material sciences
space, weather, climate systems and global changes
computational environment and energy systems
computational ocean and earth sciences
combustion system simulation
computational chemistry
computational physics
bioinformatics and computational biology
medical applications
transportation systems simulations
combinatorial and global optimization problems
structural engineering
computational electro-magnetics
computer graphics
virtual reality and multimedia
semiconductor technology, and electronic circuits and system design
dynamic systems
computational finance
data mining

Important Deadlines:

Paper submission Due May 15, 2003
Notification of Acceptance Jun 22, 2003
Final camera-ready paper Jul 01, 2003

Workshop Organizers:

Prof. Yi Pan (General Co-Chair)
Department of Computer Science
Georgia State University
Atlanta, GA 30303, USA

Prof. Robert van de Geijn (General Co-Chair)
Department of Computer Science
University of Texas at Austin
Austin, TX 78712, USA

Prof. Laurence Yang (Program Chair)
Department of Computer Science
St. Francis Xavier University
Antigonish, B2G 2W5, NS,Canada


From: Dmitri Kuzmin <>
Date: Fri, 14 Mar 2003 13:05:57 +0100 (CET)
Subject: Workshop in Dortmunda on Convection-dominated Flows

FIRST ANNOUNCEMENT of the international workshop

"High-resolution schemes for convection-dominated flows: 30 years of FCT"

Location: University of Dortmund, Germany
Date: September 29-30, 2003

Organizing committee:

Dmitri Kuzmin University of Dortmund, Germany
Rainald Loehner George Mason University, USA
Stefan Turek University of Dortmund, Germany

Background and scope:

Convection-dominated flows are notoriously difficult to treat
numerically. Solutions produced by standard discretization
techniques are corrupted by nonphysical oscillations and/or
excessive numerical diffusion. The first high-resolution
scheme to overcome these shortcomings was the now classical
Flux-Corrected-Transport (FCT) algorithm introduced 30 years
ago by Boris and Book. Their pioneering idea of blending
high- and low-order discretizations has paved the way for the
development of a whole range of high-resolution schemes which
use flux/slope limiters to prevent the formation of wiggles
in the vicinity of shocks and discontinuities while retaining
the high accuracy of approximation in regions where the
solution is sufficiently smooth. The aim of this 2-day
workshop is to provide a forum for discussion of the progress
made in the numerical simulation of convection-dominated flows
during the three decades elapsed since the birth of FCT.
Scientists from around the globe are encouraged to present
their results regarding recent trends and developments in
this challenging research field.


The tentative list of speakers includes leading experts
(D.L. Book, R. Loehner, S. Zalesak) who have laid the
foundations of the FCT methodology and demonstrated its
potential in a variety of spectacular CFD simulations.


An online registration form will be available shortly
at the homepage of the Workshop:

A small registration fee will be charged for coffee breaks and dinner.

Call for abstracts:

Please submit your contribution online or send the title
and abstract of your talk by e-mail to
no later than July 31, 2003.


Dr. Dmitri Kuzmin Phone: +49 (231) 755 3461
Institute of Applied Mathematics Fax: +49 (231) 755 5933
LS III, University of Dortmund
Vogelpothsweg 87, D-44227
Dortmund, Germany


From: Hongkai Zao <>
Date: Fri, 14 Mar 2003 18:01:59 -0800 (PST)
Subject: Southern California Applied Mathematics Symposium

The third Southern California Applied Math Symposium will be held at UC
Irvine on Saturday, May 3 this year. The purpose of this Symposium is to
bring together applied mathematicians in southern California and to
promote interactions among them. The Symposium is intended to be informal
and relaxing -- and to provide plenty of time for discussions and social
gathering. We especially encourage graduate students, postdocs, and junior
faculty to participate and to establish possible future collaborations.
Graduate students are encouraged to contribute a poster session during the
breaks. No registration fees are required for all participants.

There will be four invited speakers focusing on multiscale problems and
math biology problems. The invited speakers are Bjorn Engquist
(Princeton), Thomas Y. Hou (Caltech), Berry Merriman (UCLA), and Frederic
Wan (UCI). A tutorial lecture on the level set method will be given by
Stanley Osher (UCLA). In addition, there will be a panel discussion on
applied math education and research, and their impact on the academic or
industrial career of our applied math Ph.D. students. The panel consists
of James R. Bunch (UCSD), Tony Chan (UCLA), Ellis Cumberbatch (Claremont
Colleges), John Lowengrub (UCI), Boris Rozovsky (USC), Tony Song (JPL).
More information about the Symposium can be found in website:

Please register on line via the website above. There is also a printable
poster at the website. Please broadcast this information to people who may
be interested.

With best regards,

organizing committee:

Michael Holst (UCSD)
Thomas Y. Hou (Caltech )
Ali Nadim (Claremont College)
Stan Osher, Chair (UCLA)
Linda Petzold (UCSB)
Boris Rozovsky (USC)
Hongkai Zhao, Chair (UCI)


From: Mary Brown <>
Date: Wed, 12 Mar 2003 10:40:30 -0800
Subject: Faculty Positions at UC Davis



The Department of Computer Science, at the University of California, Davis,
invites applications for a tenure-track faculty position in areas related to
Computational Science, and the interface between Computer Science and
Computational Science. We particularly encourage applicants working on
Complex Systems, and on computational science issues in Nanocomputing. We
welcome applications from candidates who use the interaction of science and
computation to produce novel research in computer science. This new search
is in addition to ongoing searches (previously advertised) in the areas of
Information Systems, Networks, and Computer Systems. The department is
seeking candidates at the Assistant level for the Computational Science

The Computer Science Department currently has 27 faculty members, covering
all major areas of computer science. It is experiencing a period of strong
growth in the size of its faculty and expects to continue this trend. The
Davis campus is the third largest in the University of California system.
UC Davis ranks among the nation=B9s top 20 universities in research funding.

Davis is a pleasant, family-oriented community in a college town setting
with excellent public schools and a mild climate. Davis is ideally located
for many professional, cultural and recreational activities. It is just 15
miles from California=B9s capital city of Sacramento and is within easy
driving distance of the Silicon Valley, Berkeley, San Francisco, the Sierra
Nevada Mountains, and the Pacific Coast areas.

These positions require a Ph.D. or equivalent. The positions are open until
filled. For complete application instructions, please consult our webpage at

UC Davis is responsive to the concerns of dual career couples and offers a
Partner Opportunity Program. UC Davis is an affirmative action/equal
opportunity employer.

Mary Brown
Recruitment Coordinator
Department of Computer Science
University of California, Davis
(530) 752-7223
FAX: (530) 752-4767


From: Raimondas Ciegis <>
Date: Mon, 10 Mar 2003 09:52:49 +0200 (WET)
Subject: Contents, Mathematical Modelling and Analysis

Mathematical Modelling and Analysis Volume 8, Number 1, 2003
ISSN 1392-6292

K. Birgelis
Optimal control in models with conductive-radiative heat transfer
p. 1-12

T.I. Gatalskaja
Explicit solution of the Riemann boundary value problem for double
periodic functions in case of compound contour
p. 13-24

S. Guseinov, I. Volodko
Convergence order of one regularization method
p. 25-32

H. Kalis, I. Kangro
Simple methods of engineering calculation for solving heat transfer
p. 33-42

M. Meil{\=u}nas, A. U{\v s}inskas, R. Kirvaitis, R.A. Dobrovolskis
Automatic contouring of segmented human brain ischemic stroke region
on CT images

L. Popova, G. Gromyko, S. Tabakova
Numerical modelling of free thin film dynamics
p. 51-62

M. Radyna
Functionals with values in the Non-Archimedean field of Laurent series
and their applications to the equations of elasticity theory. II
p. 63-76

Yu.S. Semerich
The R-functions method in the boundary value problem for a complex
domain possessing the symmetry
p. 77-86

I. Tammeraid
Convergence acceleration and linear methods
p. 87-92


End of NA Digest