## NA Digest Sunday, May 21, 2000 Volume 00 : Issue 21

Today's Editor:
Cleve Moler
The MathWorks, Inc.
moler@mathworks.com

### Submissions for NA Digest:

Mail to na.digest@na-net.ornl.gov.

Mail to na.help@na-net.ornl.gov.

URL for the World Wide Web: http://www.netlib.org/na-net/na_home.html
-------------------------------------------------------

From: Hans Schneider <hans@math.wisc.edu>
Date: Wed, 10 May 2000 08:27:42 -0500 (CDT)
Subject: Special Issue of Linear Algebra and its Applications

Linear Algebra and its Applications
Special Issue on
STRUCTURED AND INFINITE SYSTEMS OF LINEAR EQUATIONS

The scope of the 'Special Issue on 'Infinite Linear Systems
of Equations finitely specified' with deadline June 1st
has been slightly extended to include very large structured
systems and methods that may lead to the treatment of
infinite systems. There have already been submissions
to the previously announced special issue which
fit the extended scope. We wish to give all potential authors
the chance to submit papers fitting the broader scope and
are therefore extending the deadline for submission
to August 31st. The new, slightly adapted description of
the special issue reads as follows:

One of the traditional hunting grounds of linear algebra is
the area of finite systems of linear equations, as described by a
matrix equation \$Ax = b\$. Here \$A\$ is a known matrix,
\$b\$ a known vector of finite dimensions, and \$x\$ is an unknown
vector of finite dimensions, which is to be determined such
that the equation is either satisfied, or, if that is not possible,
approximately satisfied. Many techniques are known for finding solutions
or approximate solutions, depending on the properties of the
given data and the approximation technique choosen.

If the system of equations is not finite, i.e. \$A\$ is not a matrix
but an operator, and \$b\$ and \$x\$ are of infinite dimension, then
algebraic and numerical techniques can still be used provided the
given data are finitely specified. Operators with such a property
are often called 'structured operators', and it turns out that
one can solve such infinite equations in an exact or approximate sense using
finite methods and algorithms. Also in the case of very large
structured matrices, methods can be devised that are in principle
extendable to the infinite case.

The conjunction of linear algebra and inversion theory for finitely
specified infinite operators brings interesting connections to the
forefront: algebraic equivalents of inner-outer factorizations e.g.,
or the algebraic significance of Kalman filtering. Structured matrices
can be of many types, e.g. systems with finite displacement ranks or
time-varying systems with state spaces of finite dimensions and whose
limiting behaviour is known, e.g. as a time invariant system.

A non-limiting list of topics of interest in this area is (assuming
\$A\$ is an infinite but finitely described operator of some kind):
- inversion methods
- decomposition methods for the operator A
- complexity reduction
- equivalencies
- canonical forms
- transform techniques.

Examples of operator structure:
- systems with low displacement rank
- finitely described time-varying systems
- finitely described almost-periodic systems
- differentials of non-linear systems.

Interested authors are kindly invited to submit full papers with
significant contributions to this area to any of the three
guest editors listed below before August 31st, 2000.

Patrick Dewilde
DIMES, Delft University of Technology
POB 5031,
2600GA Delft, the Netherlands.
Fax: +31 15 262 3271
email: dewilde@DIMES.tudelft.nl

Department of Mathematics and Computer Science
Georgia State University
University Plaza
Atlanta, GA 30303, USA
Fax: +1 404 651 2246
email: volshevsky@cs.gsu.edu

Ali Sayed
Rm 44-123A Engr. IV Bldg
Dept. of Electrical Engineering
University of California
Los Angeles, CA 90095-1594, USA
Fax: +1 310 206 8495
email: sayed@biruni.icsl.ucla.edu

------------------------------

From: Mike Powell <M.J.D.Powell@damtp.cam.ac.uk>
Date: Fri, 19 May 00 12:46:07 BST
Subject: Cambridge Approximation Day

CAMBRIDGE APPROXIMATION DAY
Wednesday, June 21st, 2000

An informal meeting on Approximation will be held in DAMTP, University
of Cambridge, England on June 21st, the titles of talks (and speakers)
being as follows:

Approximation on manifolds (Jeremy Levesley, Leicester),
Multilevel approximation (Steve Hales, Leicester),
function interpolation (Anita Faul, Cambridge),
New results on the characterization of native spaces for radial
basis functions (Hans-Martin Gutmann, Cambridge), and
Scattered data interpolation from principal shift-invariant
spaces (Michael Johnson, Kuwait).

There is no registration fee, and anyone who is interested will be
welcome to attend the meeting. Further information is available at
www.damtp.cam.ac.uk/user/na/App_day.html.

M.J.D. Powell

------------------------------

From: Valerie Fraysse <Valerie.Fraysse@cerfacs.fr>
Date: Mon, 15 May 2000 09:04:14 +0200 (MET DST)
Subject: Industrial days at CERFACS on Inner-Outer iterations

Industrial days at CERFACS on Inner-Outer iterations
Numerical quality of software coupling

September 11--12, 2000
CERFACS, Toulouse, France

Keynote speaker: Prof. G. Golub, Stanford.
Invited speakers: Prof. S. Hammarling, Nag, Oxford.
Prof. A. Griewank, Dresden.

Embedded iterative solvers are ubiquitous in numerical simulation. There
is a great need by industry for better control of inner-outer iterations
These two industrial days will give an opportunity for the industrial
and academic representatives to review the state-of-the-art
and define together promising research directions for embedded solvers.

Scientific committee:
CERFACS: F. Chaitin-Chatelin, I.S. Duff, V. Fraysse, L. Giraud,
and E. Traviesas.
Industry: J.-C. Berges (CNES), Ph. Homsi (Aerospatiale),
and J.-L. Vaudescal (EDF).

Local arrangements: B. Yzel and E. Traviesas.

B. Yzel (yzel@cerfacs.fr).

This workshop is organized with the collaboration of Aerospatiale,
CNES (CCT 1) and EDF.

------------------------------

From: Willi Schoenauer <schoenauer@rz.uni-karlsruhe.de>
Date: Wed, 17 May 2000 16:38:38 +0200
Subject: Block Lecture Supercomputing in July 2000

Block Lecture Supercomputing in July 2000

Scientific Supercomputing: Architecture an Use of Shared and Distributed
Memory Supercomputers (in English)
(1637 + 1638, 2 +2 SWS)

Bock lecture in the week 31.7. - 4.8.2000

Location: Seminarraum 217 in the Computer Center of the Universitaet
Karlsruhe, 2nd floor, building 20.21
Time: Mo - Fr 8.30 - 10.00, 10.30 - 12.00 h
Mo - Tu 14.30 - 16.00 h

Schoenauer/Haefner: Exercises in Seminarraum 217 and terminal room, Mo
16.30 - 19.00 and futher terms

Background: Supercomputers are built by connecting vector pipelines or
microprocessors by a communication network. The theoretical peak performance
now is more than 1 TFLOPS (teraflops, 1E12 floating-point operations per
second). These parallel computers are used for the numerical simulation in
all technical and scientific areas: crash tests for automobiles, weather
forecast, elementary particke physics, drug design etc. The progress in
science and technology is essentially determined by this new "Computational
Science". At the Universitaet Karlsruhe and the Forschungszentrum Karlsruhe
are installed an IBM RS/6000 SP and a Siemens/Fujitsu VPP300 + 5000. At LRZ
Munic a Hitachi SR8000 with more than 1TFLOPS is available for German
scientists. You should terefore get aquainted with the usage of such
supercomputers.

Contents: The contents of the lecture and how to obtain the manuscript can
be seen at the URL

http://www.uni-karlsruhe.de/~rz03/book/

It should be mentioned that there is an open-ended (hand-written) ADDENDUM
to this basic lecture where new architectures and algorithms are published.
It is accessible bythe above URL.

Prerequisite for the exercises is the basic knowledge of UNIX and Fortran.

There is no necessity of registration for students of Karlsruhe. External
participants should send an e-mail. Information about the Computer Center
is under the URL
http://www.rz.uni-karlsruhe.de/

The next term will be 19. - 23.2.2001

Willi Schoenauer
Rechenzentrum der Universitaet Karlsruhe
D-76128 Karlsruhe, Germany
Tel: +49/721/608-6350
Fax: +49/721/32550
Email: schoenauer@rz.uni-karlsruhe.de

------------------------------

From: Svetozar Margenov <margenov@parallel.bas.bg>
Date: Fri, 19 May 2000 13:14:55 +0300 (DST)
Subject: Conference in Bulgaria on Large-Scale Scientific Computations

Third International Conference on "Large-Scale Scientific Computations"
ICLSSC, June 6-10, 2001, Sozopol, Bulgaria

The Central Laboratory on Parallel Processing, Bulgarian Academy of Sciences
in cooperation with the Division of Numerical Analysis and Statistics, Rousse
University, will organize the next ICLSSC. This third in a row scientific
meeting is planned to have two special tracks:

I. Computational Linear Algebra
II. LSSC of Engineering and Environmental Problems

Specific topics of interest (but not limited to) are the following:

1. Hierarchical, adaptive, domain decomposition and local refinement methods
2. Robust preconditioning algorithms
3. Monte-Carlo methods and algorithms
4. Numerical linear algebra
5. Methods and algorithms for inverse problems
6. Large-scale computations of environmental problems
7. High-performance algorithms for engineering problems
8. Large-scale computations of data processing
9. Parallel algorithms and performance analysis

be incorporated in the scientific program of the conference. Applications
related to the special tracks of the conference as well as in some more specific
and/or advanced topics of the LSSC and their applications will be considered.

WWW-site. http://copern.bas.bg/Conferences/SciCom01.html
E-mail: scicom01@parallel.bas.bg

------------------------------

From: Lester Ingber <ingber@ingber.com>
Date: Sun, 14 May 2000 17:47:55 -0500
Subject: Position at DRW Investments, Chicago

Programmer/Analyst in Computational Finance. DRW Investments
[www.drwtrading.com], 311 S Wacker Dr Ste 900, Chicago IL 60606.

At least 2-3 years combined experience programming in C/C++, Visual
Basic/VBA and Java, as well as financial industry experience with a
financial institution. Must have excellent background in Physics, Math,
or similar disciplines, PhD preferred. Needs practical knowledge of
methods of field theory and stochastic differential equations, as well
as an understanding of derivatives pricing models, bond futures and
modeling of financial indices. The position will be primarily dedicated
to developing and coding algorithms for automated trading. Flexible hours
in intense environment. Requires strong commitment to several ongoing
projects with shifting priorities.

See www.ingber.com for some papers on current projects. Please email
Lester Ingber ingber@drwtrading.com a resume regarding this position.

Lester Ingber <ingber@ingber.com> http://www.ingber.com/
PO Box 06440 Wacker Dr PO Sears Tower Chicago IL 60606-0440
<ingber@alumni.caltech.edu> http://www.alumni.caltech.edu/~ingber/

------------------------------

End of NA Digest

**************************
-------