NA Digest Sunday, June 26, 1994 Volume 94 : Issue 26

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

Submissions for NA Digest:

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

Information about NA-NET:

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

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


From: Frits Dumortier <fd@sun4.autoctrl.rug.ac.be>
Date: Mon, 20 Jun 1994 11:18:59 +0100
Subject: Seeking PostDoc or Assistant Professorship Position

Looking for PostDoc or Assistant Professorship Position

All good things come to an end! My term at the Dept. of Control Eng.
and Automation at the University of Gent finishes by the end of 1994
and I am therefore looking for a possible PostDoc or Assistant
Professorship Position in Belgium, Europe or eventually the States.

Fields of interest for research and/or teaching assignments are:
Advanced control theory:
linear control
adaptive control
predictive control
identification
simulation
Modern computer architectures:
vector processing
parallel processing
supercomputing
Numerical Algebra:
Parallel matrix algorithms as applied
in modern control algorithms

I also have an extended practical expertise in the management of VMS and
UNIX platforms, PCs and MACs, and computer networks.

Interested institutions/organizations can contact me by E-mail (preferably)
and I will send them an extended CV and publication list (PostScript- or
LaTeX-format).

advTHANKSance,

dr.ir. Frits Dumortier

University of Gent
Dept. Control Eng. & Automation
Technologiepark 9
B-9052 Zwijnaarde
Belgium
tel.: +32-9-264 55 80
fax : +32-9-264 58 39
E-mail: Frits.Dumortier@rug.ac.be
fd@autoctrl.rug.ac.be


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

From: Flavio Sartoretto <sartoret@lewy.dmsa.unipd.it>
Date: Mon, 20 Jun 1994 11:53:02 +0200
Subject: Seeking High-dimensional Parabolic PDEs

Dear na-digest reader,
I should like to test a Monte Carlo algorithm for solving
parabolic initial mixed boundary value problems like

u_t = Lu - c u, for x in D
u(x,0)=f(x), for x in D
alpha du/dn - gamma u = h(x,t) for x on B

where D is the n-dimensional domain of the problem, B its boundary,
L is a linear elliptic differential operator, f, h known functions, d/dn is
the normal derivative.

The key point is that the dimension n must be "large" (n > 3):
If you are interested in solving such problems
or you have some reference to point out, please send me an e-mail.
Thank you in advance. Flavio Sartoretto

Prof. Flavio Sartoretto
Dept. Matematica Applicata e Informatica
Venice University
Via Torino 157
30100 Mestre (Italy)
Fax: +39-49-831995
e-mail: sartoret@lewy.dmsa.unipd.it


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

From: Gavin Pringle <gavin@maths.napier.ac.uk>
Date: Fri, 24 Jun 94 12:09:57 BST
Subject: Thesis on N-body Solver for Vortex Methods in CFD

Hi,

I have recently completed my thesis, which I think may be of interest.
Briefly stated, I have investigated the Greengard_Rokhlin O(N) Fast N-body
solver, in both 2 and 3 dimenaions, and I have ported them
onto a MIMD computer. The 3D code was then employed to speed up
a vortex simulation of a turbulent fluid flow problem (non-viscous,
incompressible). Attention has been paid to the error which arises from
the Greengard-Rokhlin solver, and the exectution time has been reduced,
by various 'tricks', without compromising the prescribed tolerance.

If you are interested, please e-mail me and I will post a postscript version
to you. If you have any comments/questions about the text, I would be
happy to communicate on this subject.

Cheers,
Gavin

Gavin Pringle email: gavin@maths.napier.ac.uk, or
gavin@uk.ac.edinburgh.festival


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

From: William F. Mitchell <mitchell@cam.nist.gov>
Date: Fri, 24 Jun 94 14:16:08 EDT
Subject: MGGHAT Version 1.1 Now Available

I am pleased to announce the availability of Version 1.1 of MGGHAT.

MGGHAT (MultiGrid Galerkin Hierarchical Adaptive Triangles) is a
FORTRAN program for the solution of general second order linear
self-adjoint elliptic partial differential equations with
Dirichlet, natural or mixed boundary conditions on 2D polygonal
domains (possibly with holes).

MGGHAT uses a finite element method with linear, quadratic or
cubic elements (user selectable) over triangles. The adaptive
refinement via newest vertex bisection and the multigrid iteration
are both based on a hierarchical basis formulation. Run time and
a posteriori graphical displays are made with gnuplot.

Several changes have been made since version 1.0. It now has
the capability to save the program state and evaluate the solution
from the saved state. This is useful for systems of equations,
time dependent problems, nonlinear equations, etc. Examples for
a system of equations and a time dependent problem are included.
All graphics are now done with gnuplot. A Tcl/Tk based menu has
been added for manipulation of the graphics.

MGGHAT is available from netlib. It can be obtained, for example,
by sending the message "send index from pdes/mgghat" (without the
quotes) to netlib@research.att.com.

MGGHAT can also be obtained from MGNET (multigrid network) by
annonymous ftp to casper.cs.yale.edu in directory mgnet/mgghat.

A hypertext version of the User's Guide can be accessed by your
favorite WWW brower at the URL
http://gams.nist.gov/reports/mgghat/userguide/userguide.html

William F. Mitchell | mitchell@cam.nist.gov
Applied and Computational Mathematics Division | na.wmitchell@na-net.ornl.gov
National Institute of Standards and Technology |
Gaithersburg, MD 20899 |


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

From: Eric Van de Velde <evdv@ama.caltech.edu>
Date: Sun, 26 Jun 1994 15:21:44 -0700
Subject: New Book on Concurrent Scientific Computing

NEW BOOK:
Concurrent Scientific Computing,
Eric F. Van de Velde,
Number 16 in Texts in Applied Mathematics,
Springer-Verlag, New York, 1994.
ISBN 0-387-94195-9 and ISBN 3-540-94195-9

This book is based on the course of the same name, which I have been
teaching at Caltech since 1988. There are 12 chapters covering
numerical linear algebra, the fast Fourier transform, classical
methods for the Poisson equation, multigrid methods, domain decomposition,
and particle methods.
With the exception of the last chapter on computer dependency,
the book has a definite slant towards multicomputers, i.e., local-memory
architectures.

The text contains enough mathematical background material to make it
self-sufficient to anyone who has taken an introductory numerical-analysis
course. However, the main emphasis is on deriving algorithms of
concurrent scientific computing.

If you have any comments on the book, please let me know at
evdv@ama.caltech.edu


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

From: John A. Crow <crow@power.amasd.anatcp.rockwell.com>
Date: Mon, 20 Jun 94 08:21:58 PDT
Subject: Question on Robin of the ``Robin problem''

The ``Robin problem'' is another name for the `boundary value problem of
the third kind' in PDEs --- problems that specify boundary conditions
of the form G(u, grad u, nu) = 0 for the unknown function u where nu is the
boundary normal. The most common form is a*u + b*(nu . grad u) = f.

Question: Who was Robin?

My own guess is that he or she is/was French. The name seems not to appear
attached to the problem prior to the 50s, but this might be based on my
own lack of information. I did a scan of the literature but came up with
nothing conclusive.

Any info on this would be appreciated.

- John A. Crow
crow@power.amasd.anatcp.rockwell.com


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

From: Jerzy Wasniewski <unijw@fysik.uni-c.dk>
Date: Mon, 20 Jun 1994 22:33:42 +0200 (MET DST)
Subject: Semi-linear Wave Equation

I am looking for some routines to solve initial value problem,
initial boundary value on semi-infinite line and initial boundary
value problems for semi-linear wave equation.

E. Kazakevicius
E.Kazakevicius@mat.dth.dk


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

From: Yousef Saad <saad@cs.umn.edu>
Date: Mon, 20 Jun 1994 19:58:18 -0500 (CDT)
Subject: SPARSKIT and Sparse Examples

Two annoucements:

1) SPARSKIT version 2 is now available - You can access it from an anonynous
FTP at cs.umn.edu. Instructions are appended to this message.
SPARSKIT is a library of routines that can be useful when working
with sparse matrices. The new version has a number of enhancements which
are too long to list. Noteworthy: more iterative solvers, more matrix
generation routines, some reordering routines,...

2) We are in the process of collecting sparse matrices from applications and
making them available also through anonymous FTP. Currently we have
matrices associated with CFD applications extracted from the examples
in the FIDAP package. We also have some matrices from an application
in a Tokamak simulation. Many (most) of these matrices are difficult
to solve with iterative methods, and a few are very difficult.
We have included for each matrix a description of the physical
problem where it originates from and the output of the INFO routine
from SPARSKIT. See the README files in the directory for details.
The matrices are located in the subdirectory /matrices in the FTP area.
They are all in the Harwell-Boeing format for compatibility with this
well-known collection.
It would be nice to make available some matrices on which `important'
numerical tests have been published. Currently, researcher on iterative
methods, are particularly interested in matrices that arise from typical
applications (e.g. CFD, electromagnetic scattering, structural
engineering, ...) If you have such matrices to contribute let me know.
We are also working on providing a contributed code for generating
matrices from CFD applications -- specifically a typical finite element
code for solving the Navier Stokes equations for a driven cavity problem.

SPARSKIT version 2 is located in an anonymous ftp area in ftp.cs.umn.edu
(128.101.230.9) -- To access it you will need to do the following:

* FTP to ftp.cs.umn.edu [128.101.230.9]
with the login name `anonymous' and your name as passwd.

* Then cd to /pub/sparse . You will find a file SPARSKIT2.tar.Z and a
directory called SPARSKIT2. You can either retrieve the
compressed tar file SPARSKIT2.tar.Z (use `binary' before
`get' when FTP'ing) or you can retrieve specific files /
modules, or all of the directory SPARSKIT2 (you may decide
to get the documentation first. - see below).

* the directory SPARSKIT2 consists of the eight subdirectories:
BLASSM, INFO, MATGEN, UNSUPP, DOC, INOUT, ORDERINGS, ITSOL,
and the files makefile and README,
Each of the subdirectories has a README file and one or
several makefiles (test programs). All the documentation is
in th DOC directory.

* If you retrieve the compressed file SPARSKIT2.tar.Z you
need to type the following in order to create the
subdirectory SPARSKIT2 in your directory:

uncompress -c SPARSKIT2.tar.Z | tar xvf -

Read carefully the README file in the main directory
SPARSKIT2. Send any comments, requests or complaints to
saad@cs.umn.edu.
Yousef Saad.

The collection of matrices is located in the subdirectory /matrices.


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

From: Ruediger Weiss <ne03@rz.uni-karlsruhe.de>
Date: Mon, 20 Jun 94 15:18:12 MESZ
Subject: Conference in Romania

First Announcement and Call for Papers

The Fourth Conference on Nonlinear Analysis,Numerical
Analysis and Applied Mathematics,April 26-30 ,1995
"OVIDIUS" University - Constanza,Romania

AIMS OF THE CONFERENCE : The Conference is organised by the
Department of Mathematics of the "OVIDIUS" University in
Constanza.It is aimed to establish exchanges and future friendships
among people working in the various fields mentioned in the title
and to point out the state-of-art of the researches in these fields.

SCIENTIFIC PROGRAM : The program will consists of about 6 invited
lectures in plenary sessions and selected presentations in three
parallel sessions : 1.Nonlinear analysis; 2.Numerical analysis;
3.Applied mathematics.

CONTRIBUTED PAPERS : You are invited to submit a paper(no more than
10 pages A4),which you may present as a lecture in approx. 15 minutes.
If you like to present a paper please submit an abstract (no more than
one page A4) to be included in a booklet of abstracts (available at
registration).

DEADLINES :
For abstracts - 15 December 1994
For full papers - 26 April 1995 (at registration).

The abstracts and papers will be sent on the address
of the Conference Secretariat.

REGISTRATION FEE : 30.- USD (you pay it arriving here,at registration).
The fee will cover the costs for the booklet of abstracts and the
printing of papers.For other informations and for sending the regis-
tration form please use the following address :

CONFERENCE SECRETARIAT
Constantin Popa
"OVIDIUS" University Constanza
Department of Mathematics,124 Mamaia Blvd.
8700 - Constanza,ROMANIA.
Phone : +40-41-614576 or +40-41-616194
Fax/Phone : +40-41-618372
Telex : 14468 UNIVO.

ORGANIZING COMMITTEE
Prof.dr.Silviu Sburlan
Prof.dr.Mirela Stefanescu
Conf.dr.Constantin Grasu.

SOCIAL PROGRAM : Constanza is one of the oldest (more than 2000
years old) and biggest towns in Romania.Because of its position on
the Black Sea shore,the town is also a resort place and it is
indeed beautiful.Some archeological places,the sea,some beautiful lakes
and interesting museums may offer wonderful trips for participants.
The Organizing Committee intends to offer possibilities of cheap
excursions to the ruins of old Histria and Callatis,on the shore
of the Black Sea and to the wine cellar of Murfatlar.

Dr. Ruediger Weiss
Rechenzentrum
Universitaet Karlsruhe
Postfach 6980
76128 Karlsruhe
Germany
Tel: +49 721 608 4034
Fax: +49 721 32550

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

From: Paul Havlak <havlak@cs.umd.edu>
Date: Tue, 21 Jun 1994 08:59:06 -0400 (EDT)
Subject: Frontiers of Massively Parallel Computation

FRONTIERS '95 -- Reminder of Deadline for Submissions: July 18, 1994

The Fifth Symposium on the Frontiers of Massively Parallel Computation
February 6-9, 1995, McLean Hilton, McLean, VA

This symposium is the fifth in a series of biennial meetings on the
design, development, and use of massively parallel computing systems.
The focus of this symposium is on research related or applicable to
systems scalable to 1,000 or more processors.

INVITED SPEAKERS:
Tom Blank, Maspar
Phil Colella, UC Berkeley
Ken Kennedy, Rice University
Burton Smith, Tera Computer

PANELS AND ORGANIZERS
Scalable I/O: Paul Messina, Caltech/JPL
Parallel C++: Carl Kesselman, Caltech
NII Application -- Radiation Treatment Planning: Robert Martino, NIH
Embedded Systems: David Schaefer, George Mason University
HPF -- Vendor Perspective: Joel Saltz, University of Maryland

Topical workshops and tutorials will be held on February 6, 1995.

The deadline for submission of papers and of workshop or tutorial
proposals is July 18, 1994. The deadline for abstracts of poster
sessions is November 1, 1994. More information is available from
http://www.cs.umd.edu:/projects/hpsl/announcements/front95.html, by
anonymous FTP at ftp.cs.umd.edu in directory pub/hpsl/frontiers95, or
by e-mail to frontiers95@umiacs.umd.edu.

Submissions and proposals should be sent to:

{Program, Tutorial, Workshop, Poster} Chair
Frontiers '95
UMIACS
University of Maryland
College Park, MD 20742-3251

For research papers, to be presented by formal presentation or
interactive demonstration, SIX (6) copies, in English, should be sent
to the program chair at the above address, so that they arrive by July
18, 1994. The text should be restricted to TEN double-spaced 8.5 by
ll inch pages. References, figures, tables, etc. may be included in a
addition to the ten pages of text. Submissions will be acknowledged
promptly by electronic mail and decisions sent by September 12 1994.
Camera-ready copy for the proceedings will be due October 31, 1994.

Each corresponding author is requested to include in a cover letter
the following information (1) complete postal address; (2) electronic
mail address; (3) phone number; (4) fax number; (5) an indication of
whether a formal presentation or researsh exhibit is desired; (6) a
statement that the manuscript is unpublished and is not under review
for another conference.

Sponsored by the IEEE Computer Society
IEEE Technical Committee on Computer Architectures
NASA Goddard Space Flight Center
UMIACS


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

From: John R. Van Rosendale <jvr@icase.edu>
Date: Tue, 21 Jun 1994 13:18:19 -0400
Subject: ICASE/LaRC workshop on Adaptive Grid Methods

ICASE/LaRC WORKSHOP ON ADAPTIVE-GRID METHODS
Radisson Hotel, Hampton, Virginia
*** NOVEMBER 7-9, 1994***
NOTE DATE CHANGE

Adaptive-grid techniques are essential to practical high resolution
fluid computations. In a three-day workshop, we plan to bring together
experts in adaptive refinement to evaluate the state-of-the-art in
adaptive-grid methods, and to discuss current research. Submissions
are invited in two areas: i) solution of proposed test cases and
ii) current research in adaptive-grid methodology. Priority will be
given to submissions which include computed solutions to one or more
of the proposed test cases.

The dates of this workshop, originally scheduled for August 26-28,
have been changed at the request of a number of the participants, who
felt the need for more time to devote to the test cases. These are a
sequence of two-dimensional compressible flow problems, chosen both
to encourage broad participation in the workshop, and to provide fair
comparisons between competing adaptive grid methodologies. All but
one of the test cases are inviscid, while four of the six are steady
state. Any adaptive-grid methods may be considered, including grid
enrichment, grid-point movement, order refinement, and hybrid methods.
As a convenience, structured and unstructured initial grids for these
test cases are available for anonymous FTP via Mosaic.

In addition to the test cases, the workshop will discuss the state-of-
the-art in adaptive-grid methods for fluid computations more generally.
The three-day format chosen should allow time for individual presenta-
tion of contributed solutions, comparison of solutions, presentations
of current research topics, and discussion on future directions in
adaptive-grid technology.

The invited speakers are:

Marsha Berger Courant Institute of Mathematical Sciences
Kim Bey NASA Langley Research Center
Phillip Colella University of California, Berkeley
Thomas Fisher University of Barcelona, Spain
Joseph Flaherty Rensselaer Polytechnic Institute
Dimitri Mavriplis ICASE, NASA Langley Research Center
Kenneth Powell University of Michigan, Ann Arbor
James Quirk ICASE, NASA Langley Research Center
Jeffrey Saltzman Los Alamos National Laboratory
Greg Spraggle Fluent, Inc.
Jean-Yves Trepanier University of Montreal, Canada
John Van Rosendale ICASE, NASA Langley Research Center
Gary Warren NASA Langley Research Center
Nigel Weatherill University College, Swansea, U.K.

Those interested in contributing should submit a four-to-six page
extended abstract for review. With the new workshop date, we are
extending the deadline for submission to August 15, 1994. Submissions
may be by hard copy, e-mail (uuencoded postscript), or fax. You will
be notified of acceptance by September 15, 1994. Final papers for
the proceedings will be due six weeks after the workshop. For more
information check the Mosaic pages:

http://k12mac.larc.nasa.gov/Wkshp.html
or
http://www.icase.edu/workshops/adapt

which contain complete descriptions of the test problems and the
proper paper format.


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

From: Lars Elde'n <laeld@math.liu.se>
Date: Wed, 22 Jun 94 09:41:25 +0200
Subject: Second Announcement Linkoping Conference

SECOND ANNOUNCEMENT AND CALL FOR PAPERS

In celebration of Ake Bjorck's 60'th birthday a conference will be
held at Linkoping University, January 9-10, 1995. The theme of the
conference will be

LEAST SQUARES METHODS: THEORY, ALGORITHMS and APPLICATIONS

The aim of the conference is to celebrate the many contributions of
Ake Bjorck to numerical analysis. This is best done by bringing
together scientists working in the field of numerical linear algebra
to discuss the latest development concerning least squares methods.
It is an interesting coincidence that in 1995 it will be 200 years
since the the least squares method was first used (according to Gauss
in his dispute with Legendre).

The following speakers have agreed to give a talk:

Iain Duff Rutherford Laboratory and CERFACS
Gene Golub Stanford University
Chris Paige McGill University
Haesun Park University of Minnesota
Mike Saunders Stanford University
Pete Stewart University of Maryland
Charles Van Loan Cornell University
Jim Varah University of British Columbia

CALL FOR PAPERS: There will be a number of contributed 20 min. talks.
No conference proceedings will be published. There are two ways to
submit an abstract:
a) send it by mail to Lars Elde'n at the address below.
b) send an abstract (ASCII file or LATEX) by email to laeld@math.liu.se
Participants and contributors should notice the following dates.

September 1, 1994 Third announcement
September 15, 1994 Deadline for abstract submission
October 15, 1994 Notification of acceptance
January 9-10, 1995 Conference

A conference dinner and birthday party will be held on the evening of
Monday, January 9.

ORGANIZING COMMITTEE: Lars Elde'n and Tommy Elfving, Linkoping, Gene
Golub, Stanford, Bo Kagstrom, Umea, and Axel Ruhe, Gothenburg.

INFORMATION: If you want to receive further information about the
conference by e-mail, please send a message to either of the addresses
below. Conference information will also be distributed by ordinary
mail.

CORRESPONDENCE should be addressed to either:

Lars Elde'n Theresia Petersson
Department of Mathematics Conference secretary
Linkoping University Same address as Lars Elde'n
S-581 83 Linkoping, Sweden
e-mail: laeld@math.liu.se thpet@math.liu.se
Phone: +46 13 28 21 83 +46 13 28 14 02
Fax: +46 13 10 07 46 +46 13 10 07 46

NOTE: January 11-13, 1995, a symposium will be held in Stockholm to
celebrate Germund Dahlquist's 70'th birthday. This will be announced
separately. For information, contact Gustaf Soderlind, e-mail
Gustaf.Soderlind@dna.lth.se.


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

From: C. J. Budd <Chris.Budd@bristol.ac.uk>
Date: Wed, 22 Jun 94 14:34:31 +0100
Subject: Singularities in Partial Differnetial Equations

First announcement
Workshop meeting on
Singularities and interfaces in Partial Differnetial Equations
Rodney Lodge, University of Bristol, UK
26th. - 28th. October, 1994

This workshop meeeting marks the start of a European HCM network on
Singularities and Interfaces in Partial Differential Equations. It aims
to bring together mathematicians in Europe interested in analytical aspects
of singularities and interfaces, their numerical computation and in their
application to physical problems. The meeting will focus on the following
topics: the theory of singularities and asymptotics in PDEs, singularities
in free boundary problems, singularities in superconductors, combustion,
phase transitions, singularities in water waves and viscous flow with
interfaces. There will be plenty of time for both formal talks and
informal interactions.

For further details contact: Chris Budd, School of Mathematics,
University Walk, Bristol, BS8 1TW, UK, Tel: 44-272-303317,
FAX: 44-272-303316, E-MAIL Chris.Budd@bristol.ac.uk.


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

From: Arieh Iserles <A.Iserles@amtp.cam.ac.uk>
Date: Thu, 23 Jun 94 11:09:14 BST
Subject: Acta Numerica 1994

The 1994 issue of Acta Numerica is now out and its contents are

Luis ALVAREZ and Jean Michel MOREL: "Formalization and computational
aspects of image analysis"..............................................1

Tony F. CHAN and Tarek P. MATHEW: "Domain decomposition algorithms"....61

S.W. ELLACOTT: "Aspects of the numerical analysis of neural networks".145

Bengt FORNBERG and David M. SLOAN: "A review of pseudospectral methods
for solving partial differential equations"...........................203

R. GLOWINSKI and J.L. LIONS: "Exact and approximate controllability
for distributed parameter systems I"..................................269

Leslie GREENGARD and Monique MOURA: "On the numerical evaluation of
electrostatic fields in composite materials"..........................379

Malcolm SABIN: "Numerical geometry of surfaces".......................411

Andrew STUART: "Numerical analysis of dynamical systems"..............467

As the na-netters might recall, the purpose of Acta Numerica is to
publish surveys of recent important developments in numerical analysis
and scientific computing. The themes and the authors are selected by a
distinguished Editorial Board (C. de Boor, F. Brezzi, J.C. Butcher,
P.G. Ciarlet, G.H. Golub, H.B. Keller, H.-O. Kreiss, K.W. Morton,
M.J.D. Powell and R. Temam). Acta Numerica is published once a year.

Acta Numerica 1994 continues the tradition that has been set in the
previous two volumes, presenting new developments on algorithms,
theory and application areas, except that the present issue is longer
than usual (572 pages).

Acta Numerica 1994 is available as a single volume for $44.95 (SIAM
members receive a 30% discount) or you can subscribe to automatically
receive your volume each year on publication. Copies of previous
volumes are also available for purchase. Further information can be
obtained and orders placed from Cambridge University Press by
contacting Amy Thomas at thomas@roslin.stanford.edu


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

From: Michael Trott <mtrott@wri.com>
Date: Mon, 20 Jun 1994 06:55:55 -0500
Subject: Positions Available at Wolfram Research

Mathematica Development

We are looking for top-quality individuals to work on innovative projects
in the areas of numerical computation, differential equations and
computational geometry. Extensive programming experience is required,
preferably in C. At least some familiarity with Mathematica is required.
Extensive experience in numerical analysis required.

Wolfram Research is headquartered in Champaign, Illinois and is an
affirmative action / equal opportunity employer. Interested applicants
should send a resume to:

Attn: Personnel,
Wolfram Research, Inc.,
100 Trade Center Drive,
Champaign,
IL 61820,

resumes@wri.com
fax: 217-398-0747


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

From: Bette Byrne <Bette.Byrne@comlab.ox.ac.uk>
Date: Fri, 24 Jun 1994 11:53:48
Subject: Fellowship Available at Oxford

CLOSING DATE: 30TH JUNE 1994
UNIVERSITY OF OXFORD
Computing Laboratory in association with the Smith Institute
SMITH RESEARCH FELLOW IN COMPUTATION

Applications are invited for a three year Research Fellowship, with the
possibility of two further years, at the Computing Laboratory, Oxford
University. The successful candidate will also be a senior member of
the Smith Institute, an innovative forum for the exchange of ideas,
problems and solutions between University researchers and industrial
scientists of Smith System Engineering Limited based in Guildford.

Applications will be welcome from Computing Scientists, Mathematicians
and Engineers specialising in any of the following areas: high
performance computing; distributed computing; parallel architectures,
systems and algorithms; system requirements, specification and
compliance; networks and communications; multimedia and synthetic
environments; computational science and engineering.

The post will involve strong interaction with Smith System Engineering,
which is engaged in a number of national and European computational
initiatives. The Fellow will have the opportunity to spend time at the
company premises in Guildford, discussing industrially relevant problems
in a stimulating business environment. Associated travel and
accommodation costs will be met by the company.

Salary will be according to age and experience, on either the RS1A or RS2
scale. (RS1A scale is currently 13,601 to 20,422 p.a. and the RS2 scale is
currently 18,855 to 25,107 p.a., both under review).

Further particulars may be obtained from: The Administrator, Oxford
University Computing Laboratory, Wolfson Building, Parks Road,
Oxford OX1 3QD. (Fax: 44-(0)865-273839, Email: sue@comlab.ox.ac.uk)

Applications in the form of a full curriculum vitae, together with the
names and addresses of three referees, should be sent to the Administrator
to arrive by 30th June 1994.

The University is an Equal Opportunity Employer.


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

From: Christopher T H Baker <cthbaker@ma.man.ac.uk>
Date: Wed, 22 Jun 94 12:41:45 BST
Subject: CFD Position in Manchester University

The University of Manchester (UK) is advertising a new chair in
Computational Fluid Dynamics established to provide a strong
interactive link between the School of Engineering and the Mathematics
Department.
The closing date for applications is 31st August.
Contact the Academic Staffing Office:
phone -- (International +44) (0)61 275 2028
fax -- (International +44) (0)61 273 5306
quoting reference 222/94.


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

From: SIAM <nelson@siam.org>
Date: Mon, 20 Jun 94 13:38:54 EST
Subject: Contents, SIAM Matrix Analysis and Applications

Contents
SIAM J. ON MATRIX ANALYSIS AND APPLICATIONS
Vol. 15, No. 4, October 1994

A Perturbation Analysis of the Generalized Sylvester
Equation (AR-LB, DR-LE) = ( C,F)
Bo Kagstrom

Uniform Stability of Markov Chains
Ilse Ipsen and Carl Meyer

An Efficient Algorithm to Compute Row and Column Counts for
Sparse Cholesky Factorization
John B. Gilbert, Esmond G. Ng, and Barry W. Peyton

Convexity and Concavity of the Perron Root and Vector of
Leslie Matrices with Applications to a Population Model
Stephen J. Kirkland and Michael Neumann

Stable Numerical Algorithms for Equilibrium Systems
Stephen A. Vavasis

Some Convergence Properties of Matrix Sets
David P. Stanford and Jose Miguel Urbano

A Hybrid Algorithm for Optimizing Eigenvalues of Symmetric Definite Pencils
Jean)Pierre A. Haeberly and Michael L. Overton

Row Sums and Inverse Row Sums for Nonnegative Matrices
Shmuel Friedland, Rohan Hemasinha, Hans Schneider, Jeffrey
Stuart, and James Weaver

Collinearity and Total Least Squares
Ricardo D. Fierro and James R. Bunch

Rank Robustness of Complex Matrices with Respect to Real Perturbations
M. A. Wicks and R. A. DeCarlo

Row Ordering for a Sparse QR Decomposition
Thomas H. Robey and Deborah L. Sulsky

Block-Triangularizations of Partitioned Matrices under
Similarity/Equivalence Transformations
Hisashi Ito, Satoru Iwata, and Kazuo Murota

Numerical Range of Matrix Polynomials
Chi-Kwong Li and Leiba Rodman

A Stable and Efficient Algorithm for the Rank-One
Modification of the Syummetric Eigenproblem
Ming Gu and Stanley C. Eisenstat

Fast Solution of Confluent Vandermonde Linear Systems
Hao Lu

Condition and Accuracy of Algorithms for Computing Schur
Coefficients of Toeplitz Matrices
I. Gohberg, I. Koltracht, and D. Xiao

Eigenvalues of Block Matrices Arising from Problems in Fluid Mechanics
K. A. Cliffe, T. J. Garratt, and A. Spence

On the Perturbation of the Cholesky Factorization
Zlatko Drmac, Matjaz Omladic, and Kresimir Veselic

Towards a Divide and Conquer Algorithm for the Real
Nonsymmetric Eigenvalue Problem
Loyce Adams and Peter Arbenz

Gaussian Elimination with Partial Pivoting can Fail in Practice
Leslie V. Foster

A Hybrid Tridiagonalization Algorithm for Symmetric Sparse Matrices
Ian A. Cavers


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

From: Richard Brualdi <brualdi@math.wisc.edu>
Date: Mon, 20 Jun 1994 13:26:53 -0500 (CDT)
Subject: Contents, Linear Algebra and its Applications

LINEAR ALGEBRA AND ITS APPLICATIONS
Contents Volume 207

Walter F. Mascarenhas (Campinas, Brazil)
The Structure of the Eigenvectors of Sparse Matrices 1

Andrew Lenard (Bloomington, Indiana)
Structure Theory of Reciprocal Pairs of Linear Transformations 21

T. Ja. Azizov (Voronezh, Russia), P. A. Binding (Calgary,
Alberta, Canada), J. Bognar (Budapest, Hungary), and
B. Najman (Zagreb, Croatia)
Nondegenerate Jordan Subspaces of Self-Adjoint Operators in
Indefinite Spaces 37

George Labahn (Waterloo, Ontario, Canada) and
Tamir Shalom (Herzlia, Israel)
Inversion of Toeplitz Structured Matrices Using Only
Standard Equations 49

Ann Sinap and Walter Van Assche (Heverlee, Belgium)
Polynomial Interpolation and Gaussian Quadrature for
Matrix-Valued Functions 71

Bingsheng He (Nanjing, People's Republic of China)
New Contraction Methods for Linear Inequalities 115

Xingping Sun (Springfield, Missouri)
Scattered Hermite Interpolation Using Radial Basis Functions 135

N. M. Missirlis and N. G. Gaitanos (Athens, Greece)
Extensions of the Ostrowski-Reich Theorem 147

Nam-Kiu Tsing (Hong Kong), Michael K. H. Fan, and
Erik I. Verriest (Atlanta, Georgia)
On Analyticity of Functions Involving Eigenvalues 159

Steve Batterson (Atlanta, Georgia)
Convergence of the Francis Shifted QR Algorithm on Normal Matrices 181

Andre C. M. Ran (Amsterdam, the Netherlands) and
Leiba Rodman (Williamsburg, Virginia)
The Rate of Convergence of Real Invariant Subspaces 197

Evgenij E. Tyrtyshnikov (Moscow, Russia)
Influence of Matrix Operations on the Distribution of Eigenvalues
and Singular Values of Toeplitz Matrices 225

Anders Barrlund (Umea, Sweden)
Perturbation Bounds for the Generalized QR Factorization 251

Haruo Yanai and Shin-ichi Mayekawa (Tokyo, Japan)
BOOK REVIEW: Linear Algebra and Linear Models, by R. B. Bapat 273

Author Index 279

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

From: SIAM <tschoban@siam.org>
Date: Thu, 23 Jun 94 11:22:41
Subject: Contents, SIAM Applied Mathematics

Contents
SIAM Journal on Applied Mathematics
Volume 54, Number 5, October 1994

Asymptotic Expansion of the Waiting Time Distribution of Two Models
of a Closed Processor-Sharing System: Heavy Usage
Roberto Barbagallo, Marina Mochi, and Francesco Zirilli

Ignition of a Rectangular Solid by an External Heat Flux
Carlos Vazquez-Espi and Amable Linan

Reconstruction of a Spherically Symmetric Speed of Sound
Joyce R. McLaughlin, Peter L. Polyakov, and Paul E. Sacks

A Uniqueness Theorem for a Lossy Inverse Problem in Reflection Seismology
Bruce J. Chaderjian

Diagonalization of Abel's Integral Operator
Oliver Knill

Localized Eigenmodes of the Induction Equation
Manuel Nunez

Multiple Supercritical Solitary Wave Solutions of the Stationary
Forced Korteweg-De Vries Equation and Their Stability
Lianger Gong and Samuel S. Shen

Chaotic Pulse Trains
N. J. Balmforth, G. R. Ierley, and E. A. Spiegel

Level Sets of Viscosity Solutions: Some Applications to Fronts and
Rendez-Vous Problems
M. Falcone, T. Giorgi, and P. Loreti

Dynamics of Interfaces in Competition-Diffusion Systems
S.-I. Ei and E. Yanagida

On the Evolution of Periodic Plane Waves in Reaction-Diffusion
Systems of (lambda)-(omega) Type
Jonathan A. Sherratt

The Existence of Spiral Waves in an Oscillatory Reaction-Diffusion System
Joseph Paullet, Bard Ermentrout, and William Troy

Stability and Bifurcations of Equilibria in a Multiple-Delayed
Differential Equation
Jacques Belair and Sue Ann Campbell

Turing Instability in Competition Models with Delay I: Linear Theory
S. Roy Choudhury

Analysis of Spatial Structure in a Predator-Prey Model with Delay
II: Nonlinear Theory
S. Roy Choudhury


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

End of NA Digest

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