NA Digest Monday, June 23, 1996 Volume 96 : Issue 24

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.

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

From: Pierluigi Claps <claps@unibas.it>
Date: Tue Jun 18 11:58:39 1996
Subject: Correlation Matrices of Sparse Datasets

In a multivariate time series framework the correlation matrix M is a
Gramian one and needs to be decomposed.
Decomposition is possible if M is at least positive-semidefinite. M is
actually determined from time series data and if the time series data are
very sparse, the final square matrix, needed to obtain positive-semidefinite
M can end up excluding some stations. In principle, it should be possible to
transfer the information relative to the correlations between stations A-B
and B-C, with overlapping data, to the couple A-C with nonoverlapping data.
Does anyone know about this transfer of correlation and how this can be
compatible with a final positive-definite correlation matrix?
Thank you.

Dr. Pierluigi CLAPS
Dipartimento di Ingegneria e Fisica dell'Ambiente, Universita' della Basilicata
Via della Tecnica 3, 85100, Potenza (Italy),
Tel. +39 971 474638, Fax, +39 971 56537, E-mail: CLAPS@unibas.it


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

From: Mohamed Othman <mohamed@cs.fsas.upm.edu.my>
Date: Mon, 17 Jun 1996 16:17:38 -0800
Subject: Query on How to Discretize a Poisson Equation

Dear Colleagues,

Does anyone know how to discretize the following Poisson equation?

Given a Poisson eq. as Uxx + Uyy = f(x,y) in [0x1]x[0x1] and
assuming the delta x = delta y = h.

Can we discretize the above Poisson equation at point u[i,j] using the
neighbouring points u[i+1,j+2], u[i+1,j-2], u[i-1,j+2] and u[i-1,j-2] ???...
(i.e. use the finite difference method).

Does any one can show me the way or steps of how to discretize the above
problem?. I really appreciate if someone can assist me for solving the
problem or related any articles or books how to discretize the Poisson
equation or any equation with different size of "h"?.

I would like to thank you in advance.

Your sincerely,

Mohamed Othman e-mail:mohamed@cs.fsas.upm.edu.my
Department of Computer Science tel :03-9486101 x 3502
University Pertanian Malaysia fax :03-9432508
43400 UPM Serdang Selangor
MALAYSIA


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

From: Stephen Wright <wright@mcs.anl.gov>
Date: Fri, 14 Jun 1996 17:31:00 -0500
Subject: WWW Site for Math Programming Society

The Mathematical Programming Society would like to announce its new
WWW site at

http://www.caam.rice.edu/~mathprog/

The site includes
* an online membership directory and member record updating,
* online membership application,
* information about MPS prizes
* MPS history and officers
* information about the society journals "Mathematical Programming,
Series A and B."
* issues of the Society newsletter "Optima," which are being added
as technology permits.
* link to the home page for the 1997 Math Programming symposium.

Steve Wright -- wright@mcs.anl.gov


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

From: Ake Bjorck <akbjo@math.liu.se>
Date: Sat, 15 Jun 1996 13:15:31 +0200
Subject: New Book on Least Squares Problems

New Book Available
NUMERICAL METHODS FOR LEAST SQUARES PROBLEMS
SIAM, May 1996
Ake Bjorck, Linkoping University, Sweden

408 pages, ISBN:0-89871-360-9
List Price U.S. $47.50/ SIAM Member Prize $38.00

Today applications of least squares arise in a great number of
scientific areas, such as statistics, geodetics, signal processing and
control This monograph aims at covering the full spectrum of relevant
problems and methods in least squares. It collects recent research results
and treats methods for solving very large and sparse problems with both
direct and iterative methods. It covers updating of solutions and
factorizations as well as methods for generalized and constrained least
squares problems. The 860 references provide a comprehensive survey of the
available literature on the subject.

The book should be useful for mathematicians working in numerical linear
algebra, computational scientists and engineers, statisticians, and
electrical engineers.

CONTENTS

1. Mathematical and Statistical Properties 1
2. Basic Numerical Methods 37
3. Modified Least Squares Problems 127
4. Generalized Least Squares Problems 153
5. Constrained Least Squares Problems 187
6. Direct Methods for Sparse Problems 215
7. Iterative Methods 269
8. Least Squares Problems with Special Bases 317
9. Nonlinear Least Squares Problems 339
Bibliography 359
Index 401



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

From: Shao-Po Wu <clive@isl.stanford.edu>
Date: Mon, 17 Jun 1996 10:20:33 -0700 (PDT)
Subject: Determinant Maximization and Semidefinite Programming

SDPSOL, version beta available via anonymous ftp and WWW

SDPSOL is a parser/solver for determinant maximization (MAXDET) and
semidefinite programming (SDP) problems with matrix structure.
MAXDET problems have the form

minimize c^Tx - log det G(x)
subject to G(x) > 0, F(x) > 0

where G(x)>0 and F(x)>0 are linear matrix inequality (LMI) constraints.
Two important special cases are SDP (when G(x)=1) and analytic
centering (when c=0 and F(x)=1). MAXDET (and SDP) problems arise in
control, statistics, computational geometry, and information and
communication theory.

In many cases the optimization variables have matrix structure, which
makes it tedius in practice to put the problem in the form above.
SDPSOL automates this task by allowing the user to specify (and solve)
MAXDET (or SDP) problems in a format close to its natural mathematical
description. SDPSOL parses problems expressed in the SDPSOL language,
solves them using an interior-point method, and reports the results in
a convenient form.

SDPSOL can be used stand-alone under UNIX, or from within Matlab.
The current implementation only exploits block-diagonal structure, hence
is not particularly efficient. But it is very easy to use, and does
work quite nicely on small and medium size problems (say, a hundred or
so variables).

SDPSOL is available via anonymous ftp to isl.stanford.edu in
/pub/boyd/sdpsol. It can also be accessed via WWW from URL
http://www-isl.stanford.edu/~boyd/group_index.html.

Shao-Po Wu & Stephen Boyd
clive@isl.stanford.edu, boyd@isl.stanford.edu
June 1996, Stanford CA, USA


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

From: Alison Bogardo <bogardo@siam.org>
Date: Fri, 21 Jun 96 15:04:32 EST
Subject: Nominzations for Dahlquist Prize

CALL FOR NOMINATIONS
for
THE GERMUND DAHLQUIST PRIZE

SIAM will present the award at the 1997 conference on Scientific Computing
and Differential Equations (SciCADE97) in Grado, Italy (near Trieste)
September 15-19. The award honors the contributions of Germund
Dahlquist to numerical analysis and scientific computing.

Eligibility
The prize, established in 1995, is awarded to a young scientist
(normally less than 45) for original contributions to fields
associated with Germund Dahlquist, especially the numerical solution
of differential equations and numerical methods for scientific
computing.

Description of the Award
The award is to include a certificate containing the citation and a
cash prize of $1000 plus reasonable travel costs to SciCADE97.
The recipient is expected to present a talk at the conference
and encouraged to submit a paper to SIAM Review.

Nominations
A letter of nomination, including a description of the achievements,
should be sent by November 1, 1996, to

Professor Robert Skeel
Chair, Dahlquist Prize Committee
Department of Computer Science, UIUC
1304 West Springfield Avenue
Urbana, Illinois 61801-2987

Supporting letters, or names of knowledgeable persons from whom such
letters might be solicited, are also welcome.


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

From: Are Magnus Brusat <Are.Magnus.Bruaset@si.sintef.no>
Date: Fri, 14 Jun 1996 14:36:46 +0200
Subject: Workshop on Modern Software Tools

SciTools'96
International Workshop on Modern Software Tools
for Scientific Computing

LAST CALL FOR PARTICIPATION

This workshop will be arranged by SINTEF Applied Mathematics and the
University of Oslo. It will take place September 16-18, 1996 in Oslo,
Norway.

For further information, see the workshop home page at

http://www.oslo.sintef.no/SciTools96/

or the announcement in NA Digest (vol. 96, issue 13) dated March 31. A
list of the accepted talks and corresponding abstracts is now available
from the web page. Please note that the number of participants will be
limited to about 50. Registrations are accepted on first come - first
served basis.

Deadline for registration: June 25, 1996


On behalf of the Organizing Committee

Are Magnus Bruaset
SINTEF Applied Mathematics


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

From: Jose D. P. Rolim <rolim@cui.unige.ch>
Date: Fri, 14 Jun 1996 15:37:19 +0200
Subject: Italian Conference on Algorithms and Complexity

CALL FOR PAPERS
C I A C '97
3rd Italian Conference on
Algorithms and Complexity

March 12-14, 1997
Rome, Italy

The 3rd Italian Conference on Algorithms and Complexity is organized
by the Computer Science Department of the University of Rome `La Sapienza'.

Papers are solicited describing original research contributions in
theory and applications of algorithms, data structures
and computational complexity. Typical, but not exclusive areas of interest
are complexity theory, computational complexity, dynamic algorithms,
computational geometry, graph algorithms, sequential, parallel and
distributed algorithms, data structures, implementation and application
of algorithms, approximated algorithms, analysis of algorithms.

Authors are invited to submit 11 copies of an extended abstract in
English, not exceeding ten pages (double spaced), before September 6, 1996
to the chair of the organizing committee:

Prof. D. P. Bovet
Dipartimento di Scienze dell'Informazione
Universita' degli Studi di Roma `La Sapienza'
Via Salaria 113, 00198 Roma, Italy

Authors from countries where access to copying machines is difficult
may submit a single copy of their abstract. E-mail submission in plain
LaTeX is also acceptable.

PROGRAM COMMITTEE:
G. Di Battista (Rome, chair)
H. Bodlaender (Utrecht)
D. Breslauer (Aarhus)
P. Crescenzi (Rome)
T. Hagerup (Saarbruecken)
G. Italiano (Venice)
L. Pagli (Pisa)
G. Persiano (Salerno)
R. Petreschi (Rome)
J. Rolim (Geneva)
S. Whitesides (Montreal)

ORGANIZING COMMITTEE:
D.P. Bovet (Rome, chair), G. Bongiovanni,
A. Clementi, S. De Agostino, R. Silvestri (Rome).

DATES:
Deadline for submission: September 6, 1996
Notification of acceptance: November 15, 1996
Deadline for final version: December 13, 1996

FOR MORE INFORMATION:
Phone: +39-6-4991-8352 (-8432, -8360)
Fax: +39-6-8841964
E-mail: ciac97@dsi.uniroma1.it
www: http://dsi.uniroma1.it/ciac97/ciac97.html


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

From: Marcin Paprzycki <PAPRZYCKI_M@utpb.edu>
Date: Sat, 15 Jun 1996 8:35:30 -0500 (CDT)
Subject: High Performance Algorithms for Structured Matrix Problems

SECOND CALL FOR PAPERS
High Performance Algorithms for Structured Matrix Problems

edited by: Peter Arbenz, Marcin Paprzycki, Ahmed Sameh

A volume in the series "Advances in the Theory of Computation and
Computational Mathematics" (published by ABLEX, Norwood, New Jersey).

In recent years, the pace of development of high performance algorithms
for the solution of structured linear systems (large sparse systems
assembled from relatively small dense or sparse blocks; for example
bidiagonal, tridiagonal, banded, block tridiagonal, almost block diagonal,
or arrowhead systems) has increased significantly. High performance
algorithms include those designed for vector or RISC uniprocessors, as
well as a variety of parallel architectures; shared or distributed memory,
or cluster-based that combine the characteristics of both.

The volume has three goals. First, to summarize the state of the art in
the area of high performance solution of structured linear systems as well
as of structured eigenvalue and singular-value problems. Second, to
indicate what research directions are perceived as the most important ones
for the future. Third to provide a collection of algorithms and ideas that
may enhance future development in this area.

In this volume, it is hoped that direct and iterative schemes will be
surveyed, together with in-depth review of a wide spectrum of high
performance architectures.

It is important that the contributions to this volume present the
experimental results in a more unified way than currently found in journal
articles. In other words, the authors of accepted papers may be requested
to present their experiments on equivalent linear systems and/or using
similar performance metrics.

To help in the planning of this volume we kindly ask potential
contributors to contact one of the editors at your earliest convenience.

To contribute, please send 6 hard copies of the paper (or PREFERABLY,
submit your paper electronically -- prepared in plain LaTeX or PostScript)
by

November 22, 1996

to one of the editors. The volume is expected to be published in
1997.

Peter Arbenz Marcin Paprzycki Ahmed Sameh
Inst. of Sci. Comp. Dept. of Math. & CS Dept. of Comp. Sci.
ETH Zurich UTPB Univ. of Minnesota
8092 Zurich Odessa, TX 79762 Minneapolis, MN 55455
Switzerland USA USA
arbenz@inf.ethz.ch paprzycki_m@utpb.edu sameh@cs.umn.edu


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

From: Trini Flores <flores@siam.org>
Date: Mon, 17 Jun 96 08:47:36 EST
Subject: Three Upcoming SIAM Meetings

1996 SIAM Annual Meeting
July 21-26, 1996
Hyatt Regency Crown Center Hotel
Kansas City, Missouri

Deadline for hotel reservations: JULY 1, 1996.

Deadline for advance registration: JULY 8, 1996.

World Wide Web: http://www.siam.org/meetings/an96/an96home.htm



1997 ACM-SIAM Symposium on Discrete Algorithms
January 5-7, 1997
Le Meridien Hotel, New Orleans, Louisiana

Deadline for submission of 10-page Extended Abstract: JULY 1, 1996.

World Wide Web: http://www.siam.org/meetings/da97/da97home.htm



Second SIAM Conference on Mathematical Aspects of Materials Science
May 12-14, 1997
Holiday Inn Select
Philadelphia, Pennsylvania

Conference Organizer: Robert V. Kohn, Courant Institute of
Mathematical Sciences, New York University

World Wide Web: http://www.siam.org/conf.htm


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

From: Scott Mitchell <samitch@sandia.gov>
Date: Tue, 18 Jun 1996 15:42:19 -0600
Subject: International Meshing Roundtable

The due date for paper submissions to the 5th International Meshing
Roundtable (5IMR) has been extended to Friday, June 28, 1996.
Submission instructions are unchanged and available at
http://sass577.endo.sandia.gov:80/9225/Personnel/samitch/roundtable96

This change is due to the unusually large number of authors requesting
extensions. Authors who have already submitted but wish to re-submit
with substantial modifications should contact Carl-Ollivier Gooch at
gooch@mcs.anl.gov.

Thank you,

Scott Mitchell
5IMR chair


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

From: Jeanne Butler <jeanne@tc.cornell.edu>
Date: Thu, 20 Jun 1996 11:26:46 -0400
Subject: Workshop on Advanced Topics in High Performance Computing

SPACE STILL AVAILABLE

Workshop on Advanced Topics in High Performance Computing
Cornell Theory Center
Cornell University, Ithaca, NY
August 19 - 21, 1996
Registration deadline: July 15, 1996

The Cornell Theory Center (CTC), a nationally funded high
performance computing and communications center, is offering three
days of lecture and discussion that take an in-depth look at
specific topics in high performance computing. This workshop is
intended for intermediate and expert parallel programmers who are
actively involved in research that will benefit from the topics
presented. The planned session titles are:

- Multigrid Methods
- Object-Oriented Methods for the Solution of Partial Differential
Equations
- A Parallel Partial Differential Equation Solver for Fluid
Dynamics Computations
- Special Topics in HPF Programming
- Domain Decomposition and Parallel Code Optimization
- Quantum Monte Carlo Methods for Continuum Systems
- Data Explorer for Scientific Visualization
- Iterative Methods

For more information on this workshop and access to the
registration form, please see http://www.tc.cornell.edu/Events/Advanced.Aug96/


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

From: Jan Steindorf <c0040301@ws.rz.tu-bs.de>
Date: Fri, 21 Jun 1996 11:33:11 +0100 (MESZ)
Subject: Workshop on Scientific Computing

CALL FOR PAPERS
4th Workshop on Scientific Computing
-- PARALLEL and DISTRIBUTED COMPUTING --

Date: 9 - 11 October 1996
Place: Technical University
Braunschweig

NEW DEADLINES! NEW DEADLINES! NEW DEADLINES!

Invited Speakers:

* Dr. Becker (DASA)
* Dr. Faden (DLR)
* Prof. Schweizerhof (University Karlsruhe)
* Prof. Stein (University Hannover)
* Dr. Stueben (GMD)
* Prof. Wriggers (TH Darmstadt)

The invited speakers will stress the development of effective applications
in the areas of Computational Solid Mechanics (CSM) and Computational Fluid
Dynamics (CFD) on modern computer architectures. The specific topics will
range from crash-simulations and fast methods in non-linear computational
solid and structural mechanics, aerodynamic computations on parallel
architectures, up to reports on the parallelisation of commercial codes
in the EUROPORT project.

Deadlines:

Submission of abstracts: 12 July
Notification of acceptance: 2 August
Submission of final version of full paper: 16 August

Contributions are solicited in the areas of algorithms, applications,
software tools, load-balancing and fault-tolerance, and communication.

Contributions should be sent to Parallel96@tu-bs.de. Accepted contributions
will be published in the Conference Proceedings.

Further informations can be obtained from
http://www.tu-bs.de/institute/WiR
or via e-mail from
Parallel96@tu-bs.de.

Organisers:

Prof. Dr. Hermann G. Matthies Dr. J. Sch"ule
Inst. of Scientific Computing Computing Centre

Technical University Braunschweig e-mail: Parallel96@tu-bs.de
Hans-Sommer-Str. 65 Tel. +49-531-391-3000
D-38106 Braunschweig, Germany Fax +49-531-391-3003


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

From: Matthias Bollhoefer <matthias.bollhoefer@mathematik.tu-chemnitz.de>
Date: Fri, 21 Jun 1996 15:16:39 +0200 (MESZ)
Subject: Conference of the International Linear Algebra Society

Dear colleagues,
The preliminary programme for the 6th Conference of the International Linear
Algebra Society in Chemnitz has just been put into the World Wide Web.
The programme is located at the URL:

http://www.tu-chemnitz.de/ilas/programme.html

At the same URL a PostScript-File and a DVI-File are available.
For further questions you can send an e-mail to ilas@mathematik.tu-chemnitz.de
or contact

Heike Weichelt
Fakult"at f"ur Mathematik
TU Chemnitz-Zwickau
D-09107 Chemnitz
GERMANY


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

From: Jerzy Wasniewski <unijw@unidhp1.uni-c.dk>
Date: Sat, 22 Jun 1996 09:17:19 +0200 (METDST)
Subject: PARA96, Workshop on Applied Parallel Computing

PARA96 WORKSHOP
ON
APPLIED PARALLEL SCIENTIFIC COMPUTING IN
INDUSTRIAL PROBLEMS AND OPTIMIZATION
UNI*C & IMM OF DTU, LYNGBY, DENMARK
Sunday, August 18 - Wednesday, August 21, 1996

We have finished collecting abstracts for the submitted talks. All the
abstracts are very good. They show a need of parallel computing algorithms.
All the abstracts as well for the invited talks as for the submitted talks
are implemented on the PARA96 Website

http://webhotel.uni-c.dk/para/para96.html

The Website explains also the registration and hotel reservation
procedure for the PARA96 meeting.

Best regards,
Jerzy


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

From: John R. Rice <jrr@cs.purdue.edu>
Date: Sun, 23 Jun 1996 19:30:59 -0500
Subject: Workshop on Computational Science and Engineering

FIRST IEEE WORKSHOP ON COMPUTATIONAL SCIENCE AND ENGINEERING
Purdue University, October 21--22, 1996

This workshop is to bring about 50-70 participants together
to discuss issues and current activities in education, employment
and R&D for computational science and engineering. It will have
strong representation from academia, government, and industry.
It is anticipated that subgroups of participants will produce
brief reports to be published in the IEEE Computational Science
and Engineering magazine.

There are three basic questions about a new field like CS&E:
1. Who is doing what? 2. What is succeeding and why? 3. What is
not working well and why? This workshop is to address these
basic questions across the board for CS&E.

In education, there are related questions like:

* Should we have CS&E departments or interdisciplinary programs?
* Where does one find textbooks and course materials in CS&E?
* Can CS&E programs be created entirely from existing courses?
* Where are the best current programs?
* Is internet based education particularly appropriate for CS&E?
* What about CS&E continuing education in industry?

The fundamental question for employment is how are current
graduates competing in the job market? Important related ques-
tions are:

* Do employers know about CS&E programs?
* What about the CS&E job market for 5 or 10 years ahead?
* What skills, training, and expertise are needed for CS&E work?

The fundamental question for research is: What are the main
research questions and how are they being addressed? Important
related questions are:

* Are there CS&E projects or just projects that add CS&E people?
* What are the main sources of culture clashes?
* Must one be an area expert first and a CS&E researcher second?
(area = materials, molecular biology, math analysis, etc.)
* What are the examples of CS&E successes in industry?

There are many, many questions about the status and future
of CS&E as it is still in its formative stages. This workshop
will provide many of the answers about its current status and
provide insights into its future directions of growth.

For further information contact John R. Rice, Department of
Computer Sciences, Purdue University, W. Lafayette, IN 47907,
Ph: 317-494-6007, email: rice@cs.purdue.edu or visit the web site
http://www.cse.purdue.edu/ieee.workshop.

Sponsors: IEEE Computer Society
IEEE Computational Science and Engineering magazine
IEEE Technical Committee on Supercomputer Applications


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

From: Mac Hyman <jh@beta.lanl.gov>
Date: Fri, 14 Jun 1996 09:35:08 -0600
Subject: Postdoctoral Position at Los Alamos National Laboratory

Post Doctoral Fellowship Position
Los Alamos NationalLaboratory
(COMPUTATIONAL FLUID DYNAMICS)

The Mathematical Modeling and Analysis(T-7) and the Hydrodynamics Methods
(XHM) Groups at Los Alamos National Laboratory are currently seeking
highly motivated Ph.D. graduates to participate in the Post Doctoral
Fellowship program.

Current staff members in these groups include:
Burt Wendroff, Len Margolin, Mac Hyman, Misha Shashkov, Erica Jen,
Darryl Holm, Joel Dendy, Roberto Camassa, Blair Swartz, Bill Chen,
Bill Rider, Ed Caramana, John Cerutti, Gary Dilts, Jim Painter,
Tony Scannapieco, Bob Stellingwerf, Harold Trease and Chuck Wingate.

These are highly competitive positions, however there are specific
opportunities at this time for Post Docs, working with numerical
analysists, to develop, implement and test new Lagrangian finite
difference methods for the solution of multidimensional high speed
fluid flow problems.

Experience in developing codes based on Lagrangian or continuous
rezone (ALE) methods is preferred. Familiarity with modern
multidimensional methods for advection is desirable. Extensive experience
in FORTRAN coding on a variety of computing platforms is required.

A Ph.D. completed within the last three years or soon to be completed
is required. Appointments have a duration of two years, with the
possibility of extension for a third year.

To receive fullest consideration for all fellowship possibilities, we
must receive your application as soon as possible. Interested
individuals are encouraged to send a copy of their resume by
e-mail, fax, or U.S. mail to:

Esther Vigil
Los Alamos National Laboratory
Group T-7, MS-B284
Los Alamos, NM 87545

EMAIL: evigil@lanl.gov
FAX: 505-665-5757


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

From: Julia Addy <julie@tc.cornell.edu>
Date: Fri, 14 Jun 1996 10:53:17 -0400
Subject: Position at Cornell Theory Center

CORNELL THEORY CENTER
SCIENTIFIC SOFTWARE ANALYST

Located at Cornell University, the Theory Center is a leading-edge
computational science laboratory that provides scalable, parallel computing
resources to the national research community. The Center's environment
includes the largest IBM SP2 supercomputer configuration in the world. We
are seeking a highly motivated individual with a strong scientific and
computing background to provide technical support and leadership in the
development, evaluation, installation, maintenance, and optimization of
highly specialized and complex scientific applications software on parallel
and other architectures.

Requirements: BS in Computer Science or a scientific discipline, advanced
degree strongly preferred, with five to seven years experience in a
scientific computing environment. Detailed knowledge of scientific
applications software in a high performance, parallel computing
environment. Significant experience programming in FORTRAN or C and solid
experience with UNIX operating systems fundamentals required; experience in
object-oriented programming, C++ highly
desirable. Proof of citizenship or a permanent immigration visa will be
required at time of employment.

Interested applicants should forward a cover letter and resume to Julie
Addy, Cornell University Theory Center, Job Number TB2402, Frank H.T.
Rhodes Hall, Ithaca, NY 14853-3801

AA/EOE

Julie Addy
Manager, Human Resources
Cornell Theory Center, Frank H.T. Rhodes Hall,
Cornell University, Ithaca, New York 14853-3801
phone 607-254-8712, fax 607-254-8888


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

From: Lee Seitelman <seitellh@pweh.com>
Date: Wed, 19 Jun 1996 14:42:56 -0400
Subject: Positions at National Security Agency

A contact at INFORMS (a merger of the old Operations Research Society
of America and The Institute for Management Sciences) recently advised
that there are a number of openings for OR professionals at the
NSA/CSA. They're looking for Master's and Ph.D.-level people, with
degrees in OR, EE, or CS. Interests run the spectrum of these
disciplines (specifically, modeling, simulation, computer algorithm
design and implementation,probability and statistics, engineering
analysis, operating systems and applications software, software
engineering, theory of computing, discrete mathematics and graph
theory, and mathematical programming. The jobs are all full-time:
operations research analyst, telecommunication network analyst,
computer network engineer, network performance assessment engineer,
software engineer. U.S. citizenship is required.

Resumes should be forwarded to
National Security Agency, Attn: M321(CSV), Fort Meade, MD 20755-6000.
Telephone is (301) 688-0564. FAX is (301) 688-0218.

Lee Seitelman


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

From: R. L. Boucher <RLBoucher@lbl.gov>
Date: Mon, 17 Jun 96 20:03:41 PST
Subject: Positions at Lawrence Berkeley Laboratory

CAREER OPPORTUNITIES AT BERKELEY LAB SUPERCOMPUTING SCIENCES PROGRAM
The Berkeley Lab has instituted a major new program in the
computing sciences. This program includes the National Energy Research
Scientific Computing Center; the Energy Sciences Network;
computational science, numerical mathematics, and computer science
research departments; and joint R&D projects with the scientific,
academic, and industrial communities.

We invite you to submit your qualifications to Berkeley Lab for the
Supercomputing Sciences program if you have interest in the following
opportunity:

* * * * * * * * * * * * * * * * * * * * * * * * * * * *
National Energy Research Scientific Computing Center (NERSC)
Group Leader, Future Technologies Group
Staff Scientist
Job NERS173
$3400-$9500/Mo.

DUTIES: Essential -- As Group Leader for the NERSC Future Technology
group, provide technical leadership and management of a computer
science research group which develops and implements technologies and
strategies for the efficient use of high performance computing
technology in a large scale production supercomputer center. With
input from stakeholders, develop short- and long-term Research and
Development plans and propose new technology directions for NERSC.
Create partnerships with other similar research groups to carry out
program goals, in particular foster the interaction of NERSC with the
University of California, Berkeley computer science research programs.
Provide planning and management support to DOE as required. Represent
NERSC, LBNL, and DOE in professional and industry activities, in
particular in standards committees. Prepare annual research and
development plans, technical reports, and budget documents.
Supervise scientific, technical staff; oversee program budget.

QUALIFICATIONS: Essential -- Demonstrated experience in leading and
managing computer science research and development teams. Involved in
computerscience research supporting large scale scientific
applications on high performance computing architectures, such as
PVPs, MPPs, or arrays of SMPs. Significant research experience and
demonstrated expertise in one or more aspects of computer science of
relevance to high performance computing (e.g., compilers, operating
systems, programming languages, tools, libraries). Must understand the
requirements of large scale computational science, and be able to
project a vision for scientific computing environments of the future.
Successful record of developing partnerships and collaborative
relationships with universities, research laboratories, and others to
carry out research and to deploy high performance computing
technologies. Excellent communication skills, and a demonstrated
ability to facilitate communications within institutions and with the
research community, government agencies, and industry. Demonstrated
ability to formulate and write research plans and proposals. Ph. D. in
computer science (or related field). Record of publications in
refereed journals.

POSTING DATE: June 14, 1996
CLOSING DATE: Open until filled.

For more information about Berkeley Lab and NERSC visit our WWW site
at: "http://www/lbl.gov"

To apply for the above position send your C.V. to:

Berkeley Lab Staffing Office
One Cyclotron Road
MS 938A, Box NERS172
Berkeley, CA 94720

OR

via E-Mail to:
Supercomputing-Employment@LBL.gov
1. Send as plain text (ASCII) in the body of your message.
2. Reference "NERS172" in the SUBJECT of your message.


* * * * * * * * * * * * * * * * * * * * * * * * * * * *
National Energy Research Scientific Computing Center (NERSC)
Group Leader, Future Technologies Group
Staff Scientist
Job NERS173
$3400-$9500/Mo.

DUTIES: Essential -- As Group Leader for the NERSC Future Technology
group, provide technical leadership and management of a computer
science research group which develops and implements technologies and
strategies for the efficient use of high performance computing
technology in a large scale production supercomputer center. With
input from stakeholders, develop short- and long-term Research and
Development plans and propose new technology directions for NERSC.
Create partnerships with other similar research groups to carry out
program goals, in particular foster the interaction of NERSC with the
University of California, Berkeley computer science research programs.
Provide planning and management support to DOE as required. Represent
NERSC, LBNL, and DOE in professional and industry activities, in
particular in standards committees. Prepare annual research and
development plans, technical reports, and budget documents.
Supervise scientific, technical staff; oversee program budget.

QUALIFICATIONS: Essential -- Demonstrated experience in leading and
managing computer science research and development teams. Involved in
computerscience research supporting large scale scientific
applications on high performance computing architectures, such as
PVPs, MPPs, or arrays of SMPs. Significant research experience and
demonstrated expertise in one or more aspects of computer science of
relevance to high performance computing (e.g., compilers, operating
systems, programming languages, tools, libraries). Must understand the
requirements of large scale computational science, and be able to
project a vision for scientific computing environments of the future.
Successful record of developing partnerships and collaborative
relationships with universities, research laboratories, and others to
carry out research and to deploy high performance computing
technologies. Excellent communication skills, and a demonstrated
ability to facilitate communications within institutions and with the
research community, government agencies, and industry. Demonstrated
ability to formulate and write research plans and proposals. Ph. D. in
computer science (or related field). Record of publications in
refereed journals.

POSTING DATE: June 14, 1996
CLOSING DATE: Open until filled.

For more information about Berkeley Lab and NERSC visit our WWW site
at: "http://www.lbl.gov"

To apply for the above position send your C.V. to:

Berkeley Lab Staffing Office
One Cyclotron Road
MS 938A, Box NERS173
Berkeley, CA 94720

OR

via E-Mail to:
Supercomputing-Employment@LBL.gov
1. Send as plain text (ASCII) in the body of your message.
2. Reference "NERS173" in the SUBJECT of your message.


* * * * * * * * * * * * * * * * * * * * * * * * * * * *
Scientist, Visualization Group
Staff Scientist
Job NERS174
$3400-$9500/Mo.

DUTIES: Essential -- As Staff Scientist in the NERSC Visualization
group, provide technical leadership helping to define and develop the
scientific visualization and virtual reality tools used by
computational scientists supported by a large scale production
supercomputer center. Work with NERSC users throughout the
country to analyze their data display requirements and then
research/recommend/develop/proselytize a solution for visualizing
their data. Perform independent research in visualization and VR
algorithms that address current and future problems of the NERSC
community. Possible areas of research include graphics and
visualization algorithms for MPP and cluster systems. Work closely
with scientists and other group members on scientific projects,
co-author funding proposals, and submit papers for national
conferences and journals.

QUALIFICATIONS: Essential -- Extensive experience developing
scientific visualization algorithms and scientific visualization
packages. Demonstrated collaborations with scientific users and in
development team atmosphere. Excellent communication skills, and a
demonstrated ability to facilitate communications within institutions
and with the research community, government agencies, and industry.
Demonstrated ability to formulate and write research plans and
proposals. Ph. D. in computer science (or related field) or equivalent
experience required. Record of publications in refereed journals.
Highly Desirable -- Experience in developing visualization algorithms
on SMP clusters or MPP architectures. Experience with working with
large datasets over a distributed environment. Experience in
developing applications using Web technology (VRML and JAVA).
Experience in working with visualization/VR hardware such as CAVE,
trackers, HMDs, etc.

POSTING DATE: June 14, 1996
CLOSING DATE: Open until filled.

For more information about Berkeley Lab and NERSC visit our WWW site
at: "http://www.lbl.gov"

To apply for the above position send your resume or C.V. to:

Berkeley Lab Staffing Office
One Cyclotron Road
MS 938A, Box NERS174
Berkeley, CA 94720

OR

via E-Mail to:
Supercomputing-Employment@LBL.gov
1. Send as plain text (ASCII) in the body of your message.
2. Reference "NERS174" in the SUBJECT of your message.


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

From: Paul Hammond <hammond@cambridge.scr.slb.com>
Date: Tue, 18 Jun 1996 16:45:43 +0000
Subject: Position at Schlumberger Cambridge Research

Development Engineer/Research Scientist, High Performance Computing
Schlumberger Cambridge Research (SCR)
Cambridge, England

Schlumberger is an international technical company with operations in over
100 countries, and 51,000 employees of more than 75 different
nationalities. Revenues exceed $7.6 billion in 1995. Schlumberger
Oilfield Services is the leading supplier of services and technology to the
international petroleum industry. In the Cambridge UK research centre,
approximately 75 scientists and engineers are working on the basics of new
products for these business lines.

We have an immediate opening in Cambridge for a High Performance Computing
specialist to join our Fluid Mechanics Modelling group. This permanent
position will involve both producing innovative 3D time-dependent
visualisations for the display and interpretation of results produced in
the multiphase flow Computational Fluid Dynamics group, and also assisting
in the development of efficient numerical simulation programmes on a 10
processor SGI Power Challenge.

The successful candidate will have:
* a background in computational science with an emphasis upon High
Performance Computing and scientific visualisation
* a minimum of MSc or PhD in Computer Science or in a quantitative
scientific discipline
* completed a significant project involving scientific visualisation
* the ability to work as a self-motivated member of a multidisciplinary team
* the ability to teach and encourage colleagues from other disciplines to
use new techniques.

If you are interested in this opportunity, please send your resume and a
brief note describing recent relevant experiences, in ascii, to
scr-recruit@slb.com, mentioning the reference FCP/DE/NAD24. Further
information on Schlumberger, and SCR, can be found at http://www.slb.com .

Schlumberger is committed to employee diversity.


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

From: Stefan Vandewalle <Stefan.Vandewalle@cs.kuleuven.ac.be>
Date: Fri, 21 Jun 1996 10:24:56 +0200
Subject: Position at Katholieke Universiteit Leuven, Belgium

Post-doctoral Research Fellowship at the Department of Computer Science
Katholieke Universiteit Leuven, Belgium


A post-doctoral research fellowship is available in the division of Numerical
Analysis and Applied Mathematics of the Department of Computer Science at the
Katholieke Universiteit Leuven in Belgium, in the project

"New Numerical Methods for Time-dependent Partial Differential Equations".

The division of Numerical Analysis and Applied Mathematics maintains strong
research programs in a variety of areas, including numerical methods for
differential equations, high performance scientific computing, modelling and
approximation, numerical integration, wavelets, and linear algebra (see
http://www.cs.kuleuven.ac.be/). Current research topics include iterative
methods for space-time discretizations, waveform relaxation methods, parallel
methods for time-dependent pdes, bifurcation and continuation methods for
time-periodic problems, and POD dimension reduction methods.

The successful applicant is required to have experience with numerical methods
for partial and ordinary differential equations, an interest in high
performance computing and has a strong mathematical and numerical analysis
background. He/she is expected to contribute with original research in the
area of the project. We particularly encourage applicants with an interest
in applying iterative methods to time-dependent problems (multigrid, domain
decomposition, Krylov methods), numerical methods for parabolic control and
space-time finite element methods.

The position is for one year with the possibility of extension to a second
year. The fellowship includes a competitive salary, moving expenses and a
professional travel allowance. The position can commence immediately and
applications will be accepted until the position is awarded. Interested
candidates should submit a complete resume, a statement of research objectives
and the names and addresses of three referees to

Dr. S. Vandewalle email: stefan@cs.kuleuven.ac.be
Katholieke Universiteit Leuven tel: +32-16-327081
Departement Computerwetenschappen fax: +32-16-327996
Celestijnenlaan 200A, B-3001 Heverlee www:
Belgium http://www.cs.kuleuven.ac.be/~stefan


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

From: SIAM <tschoban@siam.org>
Date: Fri, 14 Jun 96 13:37:43 EST
Subject: Contents, SIAM Numerical Analysis

SIAM Journal on Numerical Analysis
AUGUST 1996, Volume 33, Number 4
CONTENTS

Numerical Analysis for the Heat Flux in a Mixed Elliptic Problem to Obtain a
Discrete Steady-State Two-Phase Stefan Problem
Domingo Alberto Tarzia

Construction and Analysis of Fourth-Order Finite Difference Schemes for the
Acoustic Wave Equation in Nonhomogeneous Media
Gary Cohen and Patrick Joly

A Polylogarithmic Bound for an Iterative Substructuring Method for Spectral
Elements in Three Dimensions
Luca F. Pavarino and Olof B. Widlund

Uniqueness of Steady-State Solutions for Difference Equations on Overlapping
Grids
Zi-Niu Wu

A Preconditioner Based on Domain Decomposition for h-p Finite-Element
Approximation on Quasi-Uniform Meshes
Mark Ainsworth

Numerical Approximation of the One-Dimensional Vlasov-Poisson System with
Periodic Boundary Conditions
Stephen Wollman and Ercument Ozizmir

On Nonmonotone Solutions of an Integrodifferential Equation in Linear
Viscoelasticity
Olli Jokinen

An Efficient Two-Dimensional Vortex Method with Long Time Accuracy
Ibrahim Bless Ranero and Tomas Chacon Rebollo

Attractors and Error Estimates for Discretizations of Incompressible
Navier-Stokes Equations
Yin Yan

A Novel Approach to the Numerical Solution of Boundary Value Problems on
Infinite Intervals
Riccardo Fazio

Local Error Estimates for the Galerkin Method Applied to Strongly Elliptic
Integral Equations on Open Curves
Thanh Tran

A New Nonconforming Finite Element Method for the Computation of Electromagnetic
Guided Waves I: Mathematical Analysis
P. Joly, C. Poirier, J. E. Roberts, and P. Trouve

Dufort-Frankel-Type Methods for Linear and Nonlinear Schrodinger Equations
Lixin Wu

Fast Legendre-Fenchel Transform and Applications to Hamilton-Jacobi Equations
and Conservation Laws
Lucilla Corrias

A Global Convergence Theorem for a Class of Parallel Continuous Explicit
Runge-Kutta Methods and Vanishing Lag Delay Differential Equations
Christopher T. H. Baker and Christopher A. H. Paul

Trigonometric Collocation Methods with Product Integration for Boundary Integral
Equations on Closed Curves
J. Saranen and G. Vainikko

Projection Method II: Godunov-Ryabenki Analysis
Weinan E and Jian-Guo Liu

An Operator Splitting Method for the Wigner-Poisson Problem
Anton Arnold and Christian Ringhofer

Optimal Time Step Control for the Numerical Solution of Ordinary Differential
Equations
Masaki Utumi, Ryuji Takaki, and Toshio Kawai

Maximum Norm Analysis of Completely Discrete Finite Element Methods for
Parabolic Problems
C. Palencia

A Nonlinear Mixed Finite Element Method for a Degenerate Parabolic Equation
Arising in Flow in Porous Media
Todd Arbogast, Mary F. Wheeler, and Nai-Ying Zhang


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

From: SIAM <poulson@siam.org>
Date: Mon, 17 Jun 96 15:30:09 EST
Subject: Contents, SIAM Discrete Mathematics

August 1996
SIAM Journal on Discrete Mathematics
Volume 9, Number 3
349-509

Angles of Planar Triangular Graphs
Giuseppe Di Battista and Luca Vismara

Constrained Embedding Probability for Two Binary Strings
Jovan Dj. Golic

On the Complexity of a Cutting Plane Algorithm for Solving
Combinatorial Linear Programs
E. Andrew Boyd

Cohen-Macaulay Rings in Network Reliability
Jason I. Brown, Charles J. Colbourn, and David G. Wagner

Tilings of Binary Spaces
Gerard Cohen, Simon Litsyn, Alexander Vardy, and Gilles Zemor

Nonequivalent q-ary Perfect Codes
Tuvi Etzion

Orthogonal Arrays, Resilient Functions, Error-Correcting Codes, and
Linear Programming Bounds
Jurgen Bierbrauer, K. Gopalakrishnan, and D. R. Stinson

Second-Order Rigidity and Prestress Stability for Tensegrity
Frameworks
Robert Connelly and Walter Whiteley

Approximation Algorithms for the k-Clique Covering Problem
Oliver Goldschmidt, Dorit S. Hochbaum, Cor Hurkens, and Gang Yu


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

From: Ron Boisvert <boisvert@cam.nist.gov>
Date: Thu, 20 Jun 96 13:25:41 EDT
Subject: Contents, Transactions on Mathematical Software

For details, see the TOMS Web page at http://www.acm.org/toms/

Table of Contents
ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE (TOMS)
Volume 22, Number 2 (June 1996)

Adreas Griewank and David Juedes and Jean Utke
Algorithm 755. ADOL-C: A Package for the Automatic Differentiation of
Algorithms Written in C/C++,
131-167

Tobin A. Driscoll
Algorithm 756. A Matlab Toolbox for Schwarz-Christoffel Mapping
168-186

I. S. Duff and J. K. Reid
The Design of MA48, a Code for Direct Solution of Sparse Unsymmetric Linear
Systems of Equations
187-226

I. S. Duff and J. K. Reid
Exploiting Zeros on the Diagonal in the Direct
Solution of Indefinite Sparse Symmetric Systems
227-257

David T. Price
Remark on Algorithm 715
258


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

From: IEEE Computer Society <mehrlich@computer.org>
Date: Fri, 21 Jun 96 11:39:44 est
Subject: Contents, IEEE Computational Science & Engineering

Contents of IEEE COMPUTATIONAL SCIENCE & ENGINEERING magazine,
Vol. 3, No. 2, Summer 1996.
http://www.computer.org/pubs/cs&e/cs&e.htm

Editor-in-Chief: George Cybenko, Dartmouth College,
Associate EIC: Francis Sullivan, IDA Ctr. for Computing Sciences

THEME SECTION ARTICLES

As Eniac Turns 50: Perspectives on Computer Science Support for
Science and Engineering [theme section introduction]
Rudolf Eigenmann and George Cybenko 16

Is Parallelism for You?
Cherri M. Pancake 18

Taking Stock, Looking Ahead [theme essays] 38
Future Linear-Algebra Libraries --Jack Dongarra
Network Programming and CSE --Eric Grosse
The Microprocessor for Scientific Computing
in the Year 2000 --Yale N. Patt
The Scientist's Infosphere --K. Mani Chandy
How to Cooperate Across the Ocean --Yoichi Muraoka

What Should Computer Scientists Teach to Physical Scientists and
Engineers?
Gregory V. Wilson 46

Response to Wilson: Computer Scientists Should Not Teach Computational
Science
Rubin H. Landau 55

Response to Wilson: Teach Programming Principles, Not "Tools and Tips"
Steve McConnell 62


FEATURE ARTICLES

Fast Algorithms for Removing Atmospheric Effects from Satellite Images
Hassan Fallah-Adl, Joseph JaJa, Shunlin Liang, John Townshend, and
Yoram J. Kaufman 66

Rapid Design of Neural Networks for Time Series Prediction
Radu Drossu and Zoran Obradovic 78


DEPARTMENTS

> From the Editor-in-Chief 1
Large-Scope Computing: A Challenge for the 21st Century
CSE at Work 4
What Industry Needs from Academia: CSE Education in the 21st
Century [birds-of-feather session from SC '95; also on Web
site with audience responses]
Computing Prescriptions 11
A Heap of Data --I. Beichl and F. Sullivan
Conferences & Workshops 90
State of the Art in Numerical Analysis --Hans Munthe-Kaas
ACM Computer Science Conference --Chuck Koelbel
HPCN Europe --Chuck Koelbel [on Web only]
Reminder: First IEEE-CS Workshop on Computational Science
and Engineering, Oct. 21-22, 1996, Purdue
Book News & Reviews 93
_Mathematica as a Tool_, by Stephan Kaufmann, and _Mathematica
for Scientists and Engineers_, by Thomas B. Bahder;
reviewed by Sanjiva Weerawarana
_Introduction to the Numerical Solution of Markov Chains, by
W.J. Stewart; reviewed by Bernard Philippe
Product News 96 [also on Web site]
News Analysis 99
Upsizing, Downsizing, Rightsizing, Wrongsizing --Jeffrey Mohr
Interfaces 104
Hardware for High-Performance Computing: Abstract Progress,
Painful Consolidation --Norris Parker Smith
Calendar [on Web site]


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

From: Edit Kurali <kurali@math.ohio-state.edu>
Date: Fri, 21 Jun 1996 15:26:10 -0400 (EDT)
Subject: Contents, J. Approximation Theory

Table of Contents: J. Approx. Theory, Volume 85, Number 3, June 1996

Ya. I. Alber
A bound for the modulus of continuity for metric projections in a
uniformly smooth Banach space
237--249

Frank Deutsch and Sizwe Mabizela
Best interpolatory approximation in normed linear spaces
250--268

Alan Pinkus
TDI-subspaces of $C({\BBJ R}^d)$ and some density problems from neural networks
269--287

Krzysztof Przes\lawski
Centres of convex sets in $L^p$ metrics
288--296

Hovik V. Gevorgian, Hakop A. Hakopian, and Artur A. Sahakian
Bivariate Hermite interpolation and numerical curves
297--317

M. A. Al-Thagafi
Common fixed points and best approximation
318--323

Manuel Bello Hern\'andez and Andrei Mart\`inez Finkelshtein
Zero asymptotics of Laurent orthogonal polynomials
324--342

Boris Shekhtman
Another note on polynomial vs rational approximation
343--348

Author index for Volume 85
348--349


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

End of NA Digest

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