NA Digest Sunday, February 18, 1990 Volume 90 : Issue 07

Today's Editor: Cleve Moler

Today's Topics:

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


From: Cleve Moler <na.moler@na.net.stanford.edu>
Date: Sat Feb 17 08:10:18 PST 1990
Subject: Gene Golub Elected to National Academy of Engineering

Gene Golub has been elected to the National Academy of Engineering.
He's the first numerical analyst to be so honored.
Congratulations, Gene!


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

From: Cleve Moler <na.moler@na-net.stanford.edu>
Date: Sat Feb 17 09:58:52 PST 1990
Subject: Responses to "Who is Cholesky?"

Last week's query, "Who is Cholesky?", from Taketomo Mitsui
generated more response that any other note we've had in
the NA News. The next three contributions -- from Hans Oser,
Richard Heiberger and Richard Cottle -- are all of general
interest.

The LINPACK Users' Guide (Dongarra, Bunch, Moler and Stewart,
SIAM, 1979) has the following brief summary of Cottle's note
at the end of chapter 3:

Andre-Louis Cholesky (1875-1918) was a French military officer
involved in geodesy and surveying in Crete and North Africa just
before World War I. He develop[ed the method now named after
him to compute solutions to the normal equations for some least
squares data fitting problems arising in geodesy. His work was
posthumously published on his behalf in 1924 by a fellow officer,
Benoit, in the "Bulletin Geodesique".


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

From: Hans Oser <jpbm@athena.umd.edu>
Date: Wed, 14 Feb 90 20:43:27 EST
Subject: Cholesky References in Zurmuhl's Books

Taketomo Mitsui's inquiry about Cholesky led me to an obscure
reference to a French "commandant" in the Bulletin geodesique,
vol. 2, 1924. There, in a paper by Benoit ("Procede de commandant
Cholesky"), the method appears for the first time in the literature.
My source is, however, also an indirect one: a German textbook by
Rudolf Zurmuhl (Umlaut on the u), "Praktische Mathematik fur
Ingenieure und Physiker" published by Springer Verlag in 1961,
makes an explicit reference to Cholesky. Though published in
French, a reasonable assumption can be made that Cholesky was a
Russian emigre (as distinct from a Pole who would end his name
with an "i" instead of a "y").

In yet another book by Zurmuhl, MATRIZEN, also published by
Springer in 1961, he says on p. 64: "The method has been used any
number of times. For symmetric matrices it is mentioned first by
M. H. Doolittle (U.S. Coast and Geodetic Report, 1878, pp. 115-
120), later by Cholesky (op. cit.), and then with an explicit
first-time reference to machinery and general systems, by T.
Banachewicz (Bull.internat. academ. polon. sci., Ser. A, 1938,
pp. 393-404). (Zurmuhl then continues his remarks with: "The
connection between Cholesky and the Gaussian elimination
algorithm was not discovered until later.")

I hope these references are useful. At least, they point
to an original source. Benoit probably was a member of the
Geodesic Society who had to introduce Cholesky, but I am
guessing at that.

Hans J. Oser
Joint Policy Board for Mathematics
Washington, DC / USA


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

From: Richard Heiberger <V5807E%TEMPLEVM.BITNET@Forsythe.Stanford.EDU>
Date: Wed, 14 Feb 90 17:34:11 EST
Subject: Cholesky Biography in Maindonald's Book

There is a very nice biography in J. H. Maindonald's book, Statistical
Computation (Wiley, 1984).
-- Richard Heiberger
Temple University

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

From: Richard W. Cottle <cottle@sierra.STANFORD.EDU>
Date: Sun, 11 Feb 90 21:34:49 PST
Subject: Cholesky Eulogy by Benoit

The following translation of Commandant Benoit's eulogy for Cholesky
was done in 1975 and presented to Operations Research 314 on
October 15, 1975 --- the 100th anniversary of Cholesky's birth.

MAJOR CHOLESKY*

Andre-Louis Cholesky, born 15 October, 1875, in Montguyon
(Charente-Inferieure), entered l'Ecole Polytechnique at the age of
20 and, upon graduation, went into the artillery branch. Attached to
the Geodesic Section of the Geographic Service, in June 1905, he made
himself noticed at once by extraordinary intelligence, a great
facility for mathematical work, an inquiring spirit, original ideas,
sometimes even paradoxical, but always marked by a great dignity of
sentiment which he maintained with great conviction.

That was the period when the revision of the French triangulation had
just been decided in order to continue the revision of the meridian
line of Paris, to be used as the base of a new cadastral
triangulation. The problem of the adjustment of the grid preoccupied
many officers of the Section, who wished to contribute to fixing, in
the sense of speed, convenience and maximal precision, methods which
were not yet entirely agreed on. Cholesky approached this problem,
bringing in his solutions, as in everything he did, a marked
originality. He invented, for the solution of the condition equations
in the method of least squares, a very ingenious computational
procedure which immediately proved extremely useful, and which most
assuredly would have great benefits for all geodesists, if it were
published some day.**

In the midst of all this, there appeared the question of the map of
Crete, then occupied by international troops. After a proposal of
Colonel Lubanski, commanding office of the French troops in Crete, a
former geodesist, and, also, a rapid reconnaissance made in
March-April 1906 by Lieutenant Colonel Bourgeois, then Chief of the
Geodesic Section, it had been decided that the French Geographic
Service would undertake the triangulation of the French and British
sectors of the island of Crete (Departments of San Nicolo and Candie),
and the topographical survey of the French sector.

Three officers, of whom Major Lallemand, Chief of the mission, and
Lieutenant Cholesky, who had left France in November 1907, proceeded
in three months to the fundamental operations: measurement of a base
(in the Kavousi plain), determination of an astronomical latitude and
azimuth at the southern end. Then, Cholesky, alone, remained to
execute the triangulation.

The reconnaissance and the construction of the markers were pursued in
mid-winter. If one reflects that Crete, whose breadth is scarcely 40
kilometers, is covered with mountains whose altitude occasionally
exceeds 2400 meters, one appreciates how particularly hard the
operations were. At the summit of the Lassithi heights, the water
necessary for the detachment was still furnished at the end of May by
the melting of enormous drifts of snow. Nevertheless, surmounting all
the obstacles, Cholesky succeeded in finishing the work on 15 June,
1908. Unfortunately, the political circumstances did not ultimately
permit the sending of topographers to Crete to execute the surveys.

Afterwards, the geodesic activity of Cholesky underwent an interruption
of two years (September 1909 to September 1911), because of the
obligation where he found himself carrying-out his tour of duty as
Battery Commander.

Sometime after his return to the Section, he was sent to Algiers to
take direction of the precision levelling of Algeria and Tunisia,
executed by the Geodesic Section on behalf of the Governor General of
Algeria, and that of the Regency of Tunis.

In his new functions, Cholesky brought the ardor he applied to
everything. One discerns it from reading the "Report on the
Operations of the Precision Levelling of Algeria and Tunisia during
the Campaigns 1910-11, 1911-12, 1912-13."1

The primary Tunisian grid was finished in the field in the winter
1913-1914. Immediately, all the calculations were reviewed,
organized, the grid was checked and adjusted. The publication which
will be made will be the result of this effort.

At the same time, for the problem of levelling in Morocco, under
somewhat peculiar conditions, Cholesky sought to fit the means to the
purpose, and determined the methods to be used.

In May 1913, Cholesky was assigned to the Ministry of Foreign Affairs,
and put in charge of a vast service, the Topographical Service of the
Regency of Tunis. He hardly had time to show the measure of what he
would have been able to do there. In effect, the war broke out
shortly thereafter.

During the campaign, he was one of the officers who best understood
and developed the role of geodesy and topography in the organization
of artillery firing. In October 1916, his technical qualities
resulted in his being sent on a mission with the Geographical Service
of the Romanian Army. He returned in February 1918, having rendered
distinguished service there.

Five months afterward, on 31 August 1918, two months before his
comrade Levesque, also the commander of an artillery group, he fell
for his country, a sad and irretrievable loss for Geodesy.


=====================================================================

* This notice appeared in the "Bulletin geodesque" entitled
"Union geodesique et geophysique internationale, Premiere
Assemble Generale, Rome, May 1922", Section de Geodesie,
Toulouse, Privat, 1922, no.8, pp. 159-161.

Translation by Richard W. Cottle, with the assistance of Eduardo
Aguado.

**Translators note: The method was subsequently published by
Commandant (=Major) Benoit, Note sure une methode de resolution des
equation mormales provenant de l'application de la methode des
moindres carres a un systeme e'equations lineaires en nombre
inferieure a celuides inconnues. Application de la methode a la
resolution d'un systeme defini d'equations lineaires. (Procede du
Commandant Cholesky.), "Bulletin geodesique," 7:1 (1924), 67-77.

1 Cahiers du Service geographique de l'Armee, no. 35, Paris
("Imprimerie du Service geographique de l'Armee), 1913, 36 p.

2 See Rapport sur les travaux executes en 1912, Cahiers du Service
geographique de L'Armee, no. 36, Paris (Imprimerie du Service
geographique de l'Armee) 1913, 90 p., 15 pl., pp. 4-17.

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

From: Michael Mascagni <mascagni@jvncf.csc.org>
Date: Sat, 17 Feb 90 10:01:04 EST
Subject: DC Area E-mail List (Reminder)

Just a reminder that there is a newly formed e-mail list to distribute inform-
ation on scholarly talks, meetings, and other events of interest to the
"greater" Washington, DC area community involved in applied mathematics,
computer science, numerical analysis, high performance computing, and scien-
tific computing. If you would like to be on the e-mail list, or if you would
also like to contribute talk information to the list send me e-mail. DO NOT
E-MAIL ME AT MY NA-NET ADDRESS. Instead, send mail to mascagni@jvncf.csc.org
with your request. Please spread the word of this list to those who might
wish to get this information.


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

From: Paul Saylor <saylor@icase.edu>
Date: Sat, 17 Feb 90 19:19:30 EST
Subject: Midwest NA Day

There will be a Midwest NA Day on the University of Illinois,
Urbana-Champaign campus, Saturday, April 7. Invited speakers are:
Germund Dahlquist, Gene Golub and Linda Petzold. For information
and inquires either email to naday@martini.cs.uiuc.edu or send snail
mail to:

NA Day
Department of Computer Science
1304 W. Springfield
Urbana, IL 61801

The meeting will take special note of the retirement of Bill Gear from
the University of Illinois, where he has been department head of the
Computer Science Department since 1985.

Participation from young people is especially encouraged.

Plans are being made for a dinner Saturday night (April 7).


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

From: Alle-Jan van der Veen <dutentb!allejan@relay.EU.net>
Date: Mon, 12 Feb 90 18:22:06 +0100
Subject: Workshop in France on Algorithms and Parallel Architectures

Second Call For Papers

International Workshop on

ALGORITHMS AND PARALLEL VLSI ARCHITECTURES

June 10-16, 1990
Pont-a-Mousson, France


CONFERENCE COMMITTEE

Chairman:
Ed Deprettere, Delft Univ. of Technology

Members:
Patrick Dewilde, Delft Univ. of Technology
Thomas Kailath, Stanford University
Sun-Yuan Kung, Princeton University
Prabhakara Rao, Drexel University
Yves Robert, Ecole Normale Superieure de Lyon
Lothar Thiele, Saarland University


SCOPE

This international workshop will bring together senior and
junior researchers, who are interested in the fascinating interplay
between algorithm development and parallel architecture design.
The topic of interest is the design of algorithms and architectures
in signal processing, modeling, graphics, communications and control.
In order to guarantee a fruitful discussion between participants,
the workshop will provide a series of tutorial courses, plenary
introductions and poster sessions. Course lecturers and plenary
speakers are widely known experts on the design of hardware
algorithms. A subset of courses and workshops can be selected
from the full program for partial attendance.


COURSES -- LECTURERS

Schur's Algorithm and its Applications
P. Dewilde, Delft University of Technology

Array Forms of Fast Signal Processing Algorithms
T. Kailath, Stanford University, CA

Digital VLSI Neural Nets: Algorithm and Architecture
S.-Y. Kung, Princeton University, NJ

Artificial Neural Nets
L. Chua, University of California at Berkeley, CA

Multiscale Signal Processing
A. Benveniste, IRISA, Rennes

Large Scale Modeling
P. Wesseling, Delft University of Technology

Solving Large Systems of Linear Equations
Y. Robert, ENS de Lyon


POSTER SESSIONS

Papers that contribute to the poster sessions are still solicited.
Prospective authors are invited to submit four copies of a
4-page summary of their paper to the workshop secretary for review.
Authors of accepted papers will be asked to prepare a version for
publication in a post-workshop book, which will be available for
participants at a special reduced price.
Information regarding submission of the full-size paper will be
distributed by the end of April. Only a limited number of passive
attendees will be accepted, on a first-come, first-served basis.
Application forms may be obtained from the workshop secretary.
A reduced registration fee will be provided for students.


WORKSHOPS -- PLENARY SPEAKERS

Signal Estimation
R. Roy, Stanford University, CA

Adaptive Filtering
D. Slock, Philips, Brussels

Orthogonal Filtering
C.V.P. Rao, Drexel University, PA

Eigenvalues and Singular Values
G. Golub, Stanford University, CA

Modeling with Finite Elements
J.C.M. van der Vorst, Shell KSEP, Rijswijk

Computer Graphics
F. Kitson, HP Labs, Palo Alto, CA

Video Architectures
A. van Roermund, Philips, Eindhoven

Microcoded Processor Architectures
F. Catthoor, IMEC Leuven

Systolic Architectures
P. Quinton, IRISA, Rennes

Artificial Neural Nets
S.-Y. Kung, Princeton University, PA

Architectures for Communications
F. Jutand, Telecom-Paris

Architectures for Control
G. Irwin, The Queen's University, Belfast


COOPERATING ORGANIZATIONS

The workshop is the first in a series of open meetings of the Novel
Algorithms/New Architectures Project under the ESPRIT Programme, and is
sponsored by the European Association for Signal Processing
(EURASIP) and the IEEE Section Benelux.

VENUE

The workshop is to be held in the historical "Abbaye des Premontres" in
Pont-a-Mousson, France, at the bank of the Moselle River.
The abbey is located at the foot of the hills of the Lorraine forests,
and is at a close distance to the major cities Metz and Nancy,
250 km east of Paris, 80 km below Luxembourg.
For more information on the vicinity, write to: Centre Culturel des
Premontres, B.P. 125, 54705 Pont-a-Mousson Cedex, France.


TIME TABLE

Extended Summary Due Date March 15, 1990 *
Notification of Acceptance April 15, 1990
Registration Form Due Date May 20, 1990
Workshop Sessions June 10-16, 1990

* Late summary submissions may be considered, subject to capacity
constraints and chair's approval.


WORKSHOP SECRETARIAT

Alle-Jan van der Veen
Dept. of Electrical Engineering
Delft University of Technology
2628 CD Delft
The Netherlands

Email: allejan@dutentb
Fax: +31 15 623271
Tel.: +31 15 781442


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

From: Zdzislaw Jackiewicz <ATZXJ%ASUACAD.BITNET@CUNYVM.CUNY.EDU>
Date: Wed, 14 Feb 90 18:49:34 MST
Subject: Conference on Volterra and Delay Equations

General Announcement and Call for papers

The International Conference on
THE NUMERICAL SOLUTION OF VOLTERRA AND DELAY EQUATIONS

to be held May 25-28 (Friday-Monday) 1990 at Arizona State University

Organized by

Alan Feldstein and Zdzislaw Jackiewicz
Department of Mathematics
Arizona State University
Tempe, Arizona 85287

Phones: A. Feldstein Z. Jackiewicz
(602) 965-3779 (office), (602) 965-3751 (office)
(602) 966-4600 (home)

(NA.JACKIEWICZ@NA-NET.STANFORD.EDU or ATZXJ @ ASUACAD.BITNET)

The Conference will cover computational aspects of Volterra
and delay equations. They include functional differential equations
(delay, advanced and neutral), Volterra integral equations, and
Volterra integro-diffrential equations. The Conference will
emphasize the following topics:

- Convergence and order properties of numerical methods
- Stability analysis and construction of highly stable methods
- Implementation problems: software development and testing

There will be both invited lectures and contributed talks at
the meeting. In addition, there will be ample time for informal
discussions. The following researchers have accepted invitations
to speak:

C.T.H. Baker (Manchester)
A. Bellen (Trieste)
H. Brunner (Newfoundland)
K. Cooke (Pomona)
A. Iserles (Cambridge)
P. Linz (Davis)
Ch. Lubich (Innsbruck)
K. Neves (Boeing)
E. Russo (Naples)
L. Tavernini (Texas)
S. Thompson (Radford)
P.J. van der Houwen (Amsterdam)
M. Zennaro (Udine)

This Conference is supported by the National Science
Foundation and by Arizona State University. Funds are available to
assist a limited number of graduate students and recent Ph.D's who
have limited or no travel support. To apply for such assistance
please contact the conference organizers as soon as possible but
no later than April 2, 1990.

We encourage participants to submit the confernce registration
form below with the title, list of authors and their addresses,
and abstract for the talk. The dedline for receiving titles and
abstracts is April 2, 1990. The registration fee for the meeting
is $30.

A special issue of APPLIED NUMERICAL MATHEMATICS will be
devoted solely to this Conference. Papers for journal publication
will be selected from among invited and contributed papers and
will meet the usual high refereeing standards of the journal. The
deadline for submitting the manuscripts is August 1, 1990.

A group of rooms have been reserved for the participants of
the Conference at Sheraton Tempe Mission Palms Hotel which is
located within easy walking distance of the campus of Arizona
State University. The special rate for these rooms is $38 + tax
per night for each room (single or double) regardless of the
number of occupants, up to 4 persons per room. Reservations must
be made directly with the hotel by April 24, 1990. Please contact
them at:
1-800-547-8705 or (602) 894-1400
TELEFAX: (602) 894-1400 EXT. 566

SHERATON TEMPE MISSION PALMS HOTEL
60 East Fifth Street
Tempe, Arizona 85281

In order to qualify for the special, reduced rate, be sure to
mention that you are a participant of the International Conference
on Numerical Solution of Volterra and Delay Equations.

For an electronic registration form, please send mail to
na.jackiewicz@na-net.stanford.edu or atzxj @ asuacad.bitnet.


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

From: Karsten Decker <decker@iam.unibe.ch>
Date: 16 Feb 90 17:25:36+0100
Subject: Summer Course in Switzerland on Parallel Architectures

European Physical Society
Computational Physics Group

3rd Graduate Summer Course on Computational Physics

Parallel Architectures and Applications

to be held at
Cret-Berard Puidoux Switzerland
September 3 - 7, 1990

Program

R.W. Hockney, Reading Parallel Architectures
G.C. Fox, CALTECH Hypercube Architectures and Applications
K.C Bowler, Edinburgh Transputer Machines and Applications
J. Richardson, TM Connection Machine and Applications
I. Guyon, AT&T Bell Labs Neural Networks and Applications
R. Gruber, EPFL Software Development Strategies for Parallel
Architectures with Distributed Memory
Application

Fee: Sfr. 350.- for EPS members
Sfr. 400.- for others
Sfr. 250.- for students

Accommodation and boarding: Sfr. 80.- per day

Number of participants: Maximum 60

Deadline for application: End of June 1990

Address for application: Karsten M. Decker
IAM
University of Berne
Laenggassstr. 51
CH-3012 Berne
SWITZERLAND
decker@iam.unibe.ch


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

From: George Sell <sell%csfsa@umn-cs.cs.umn.edu>
Date: Mon, 12 Feb 90 18:40:14 CST
Subject: Deputy Director at University of Minnesota

Deputy Director

UNIVERSITY OF MINNESOTA

Army High Performance Computing Research Center

The University of Minnesota seeks an individual to direct the management
and oversight functions of the Infrastructure Support Program for the Army
High Performance Computing Research Center (AHPCRC). This program will
include approximately twenty-seven (27) scientists and support
personnel hired by the Minnesota Supercomputer Center, Inc. (MSC) as a
subcontractor to the Unversity. Eighteen (18) scientists will be located at
9 Army sites in the United States, and the remaining will be based at the MSC
Building in Minneapolis. The Deputy Director will report to the Director of
the AHPCRC.

A B.S. degree (or higher degree) in some area of the mathematical sciences,
physical sciences or engineering is required. Proven managerial experience
is essential, and a background in modern high performance computing is
desirable.

A qualified candidate may be appointed as an Adjunct Professor or a Tenured
Professor in an appropriate department in the Institute of Technology at the
University of Minnesota. A tenured appointment requires a Master's Degree
(or a higher degree) plus a proven and current research record.

Appointment begins March 1, 1990 or later.

Applications and nominations are due by March 1, 1990.

Nominations and applications should be sent to:

Dr. George R. Sell, Chair
AHPCRC Deputy Director Search Committee
University of Minnesota
Army High Performance Computing Research Center
6-196 EE/CS Building
200 Union Street S.E.
Minneapolis, MN 55455

telephone: 612/626-1550
fax: 612/626-1596

The University of Minnesota is an equal opportunity educator and employer
and specifically invites and encourages applications from women and minorities.


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

From: Heinz W. Engl <K310773%AEARN.BITNET@Forsythe.Stanford.EDU>
Date: Tue, 13 Feb 90 14:43:26 SET
Subject: Position at Johannes Kepler Universitaet, Austria

Position at the Johannes-Kepler-Universitaet, Linz (Austria)
(Chair for Industrial Mathematics)

Rank: Vertragsassistent (equ. to Assistant Professor)
Available: April or May 1, 1990. Employment for an initial period until
Dec. 31, 1990. If performance is satisfactory, employment should last 2-4
years.
Since this position is paid by industry (Voest-Alpine AG), the duties include
a research project for this company (general field: continuum/fluid mechanics,
inverse problems).
Qualifications expected: diploma or doctorate in mathematics, good knowledge
in applied and numerical mathematics, computing experience. working knowledge
of german desirable, but not absolutely necessary.

Applications (vita, list of publications, copies of diplomas etc., copy of
relevant pages of passport, e-mail and fax, if available) should be sent
to "Personalabteilung, Universitaet, A-4040 Linz, Austria" early enough to
arrive there by March 8 (fax: +43 (0)732 246810).

More information can be obtained from Prof. Heinz Engl at the above address,
e-mail addresses:
bitnet: k310773@aearn
na-net: na.engl@na-net.stanford.edu


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

From: Art Schoenstadt <0085P%NAVPGS.BITNET@Forsythe.Stanford.EDU>
Date: Fri, 16 Feb 90 12:05:00 PST
Subject: Position at Naval Posgraduate School

NAVAL POSTGRADUATE SCHOOL
Monterey, CA 93943

MATHEMATICS


The Mathematics Department at the Naval Postgraduate School,
Monterey, CA, expects to have several tenure-track and visiting
positions starting 1 October 1990. A strong possibility of
positions starting 1 July 1990 may exist. Duties include both
teaching and research. New faculty appointments may be made at
the Assistant or Associate Professor level. Preference will be
given at the lower rank. Applicants from all areas of applied
mathematics will be considered, but applications in numerical
analysis, discrete mathematics and differential equations will
receive special considerations.

RANK: Assistant or Associate Professor

SALARY: Competitive

MINIMUM QUALIFICATIONS: Ph.D. in Mathematics. University
experience is desired. Preference will be given to those
who can teach for the entire academic year.

EFFECTIVE: 1 October 1990 (Possibly 1 July 1990)

APPLY TO: Applicants should send a resume and three letters of
reference including information about teaching and research
experience to:
Professor Beny Neta
Code 53Nd
Department of Mathematics
Naval Postgraduate School
Monterey, CA 93943
3859P@NAVPGS.BITNET

CLOSING DATE: 1 March 1990

NPS is an affirmative action/equal employment opportunity employer.


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

End of NA Digest

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