NA Digest Sunday, January 29, 1989 Volume 89 : Issue 4

Today's Editor: Cleve Moler

Today's Topics:

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


From: K. W. Morton <DCH%VAX.OXFORD.AC.UK@Forsythe.Stanford.EDU>
Date: Mon, 23 JAN 89 15:27:14 GMT
Subject: Leslie Fox Prize

Fourth Leslie Fox Prize, September 4th, 1989

Call for Papers

Entries are invited for the fourth Leslie Fox Prize competition.
Any person who is less than 31 years old on January 1st, 1989, and has not
already won a first prize is eligible. Each entry should consist of three
copies of a paper, describing some of the candidate's research, that is
suitable for a 40 minute lecture at a numerical analysis symposium.
Whether or not the work has been published or accepted for publication is
irrelevant, but no person may submit more than one paper. Unsuccessful
candidates from previous competitions are encouraged to enter.

The entries will be considered by an Adjudicating Committee, its
members being K. W. Morton (Oxford University), J. C. Mason (Shrivenham),
and N. K. Nichols (Reading University). Particular attention will be given
to the originality and quality of each paper, and to the suitability of the
material for a 40 minute talk to a general audience of numerical analysts.
About five papers will be selected by the Committee for presentation at a
symposium that will be held at the University of Cambridge on Monday,
September 4th, 1989. Only the papers that are presented at the symposium
will be eligible for awards but, subject to this restriction, the
Adjudicating Committee may award any number of first and secondary prizes.

Entries should reach Professor K. W. Morton (Oxford University
Computing Laboratory, 8-11 Keble Road, Oxford OX1 3QD, England; e-mail
address morton@na.oxford.ac.uk) not later than April 3rd, 1989. Each
candidate should include a statement that his or her year of birth is not
earlier than 1958, and should indicate whether he or she would be available
to present his or her paper at the symposium. The Adjudicating Committee
may allow a deputy to present a paper in a case of exceptional merit. The
receipt of all entries will be acknowledged. It is unlikely that travel
funds will be available to assist candidates who attend the symposium. Any
questions on this notice should be addressed to a member of the
Adjudicating Committee.


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

From: Hans Schneider <hans@pade.math.wisc.edu>
Date: Mon, 23 Jan 89 15:52:56 cst
Subject: LAA Special Issue

Special Issue on

ITERATIONS IN LINEAR ALGEBRA AND APPLICATIONS

Contributions are invited for a special issue of Linear Algebra and its
Applications entitled " Iterations in Linear Algebra and
Applications". The issue is being dedicated jointly to three
mathematicians who have made major contributions to this field:
G. Golub, R. Varga and D. Young.

In the years following World War II much of the interest in iterative
methods was motivated by the numerical solution to partial differential
equations. This was followed by a period in which the scope and
applications of iterative methods was broadened to cover the eigenvalue
and least squares problems through the introduction of such algorithms
as the QR and the SVD. In recent years, there has been a revival of
interest in iterative methods because the introduction of vector and
parallel supercomputing and other digital technologies in science
engineering encouraged modelling and solution of problems of a very
large scale.

The scope of the issue includes the areas mentioned in the above short
account. We give further examples below of topics which we would like
to be addressed in the issue. Our list is by no means exhaustive and we
welcome all other topics which are relevant to the title:

i) Iterative methods for solving large linear systems, for example
systems which arise in Multiple Coupled 3D PDE's.

ii) Iterative methods for solving nonsymmetric systems and singular
systems.

iii) Sequential and parallel iterative algorithms for solving the
eigenvalue and the least squares problems including applications to
signal processing. Incomplete orthogonal factorization
preconditioners.

iv) Methods for determining subdominant eigenvalues of matrices.

v) New approaches to implementing classical methods such as the SOR
and SSOR together with appropriate analysis of convergence rate.

vi) Preconditioned conjugate gradient methods. To what extent do
efficient preconditioners depend on different type of computer
architecture?

vii) Substructuring and domain decomposition methods: their
parallelization and their application in structural analysis and fluid
flow.

viii) Acceleration of iteration by techniques from approximation
theory and analysis, e.g., Chebyshev semi-iterative methods and Euler
summation techniques.

ix) Efficient implementation of iterative methods (such as
multisplitting) on multiprocessor machines with shared and/or local
memory.

x) Solving nonlinear problems by linearization processes. For
example: global optimization and updating techniques.


Papers should meet the usual publication standards of Linear Algebra
and its Applications. The deadline for submission is March 1990 with
expected publication about a year later. Papers may be sent to any of
the special editors listed below.


Owe Axelsson
Department of Mathematics
University of Nijmegen
Toernooiveld 6525 ED Nijmegen
The Netherlands
e-mail: u641007@hnykun11.bitnet

John de Pillis
Department of Mathematics
University of California
Riverside, California 92521
e-mail: depillis@ucrvms.bitnet

Michael Neumann
Department of Mathematics
University of Connecticut
Storrs, Connecticut 06269-3009
e-mail: neumann@uconnvm.bitnet

Wilhelm Niethammer
Institut fuer Praktische Mathematik
Universitaet Karlsruhe
D-7500 Karlsruhe
Federal Republic of Germany
e-mail: prama@dkauni12

Robert J. Plemmons
Department of Mathematics
North Carolina State University
Raleigh, North Carolina 27695-8205
e-mail: plemmons%matple@ncsuvx.ncsu.edu


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

From: Dan Warner <WARNER@eureka.clemson.edu>
Date: Mon, 23 Jan 89 17:37 EST
Subject: Salary Surveys on Numerical Analysts

Departments of Mathematical Sciences embrace a wide range of disciplines.
My Department Head asked me to enquire whether anyone knew of a salary
survey which treated numerical analysts as group. If anyone here knows
of such a study I would appreciate hearing about it. Thanks.
Dan Warner
Dept. of Mathematical Sciences
Clemson University
Clemson, SC 29634-1907
(803) 656-5244


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

From: D Griffel <Griffel%qgb.bristol.ac.uk@nss.cs.ucl.ac.uk>
Date: Tue, 24 Jan 89 10:08:04 GMT
Subject: Differentiating Rational Approximants

Does anyone know good methods, or algorithms, for evaluating the
first couple of derivatives of rational-function interpolants?
If so, I'd be grateful for any ideas or references.

David Griffel, Maths Dept., Bristol University,

Griffel@uk.ac.bristol.qgb


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

From: Steve Stevenson <hubcap!steve@gatech.edu>
Date: 24 Jan 89 19:08:49 GMT
Subject: Request for Classic Problems and Open Questions

I want to teach a seminar this summer which would emphasize the development
of ``computational science'' through the seminal problems which have motivated
researchers throughout history. The seminar will conclude with a look
at the ``most important'' open problems. This brings up the nasty
issue of identifying said things.

To set some sort of tone, Hilbert's problems should be included since it
led to Turing's paper. The question of completeness led to Goedel's
results. Surely 3-satisfiability. The four-color conjecture.
Complementation of context sensitive languages. I would include such
things as the Dining Philosphers problem as a motivator for solutions to
sharing. I would also include problems with asynchronous sequential circuits
as a motivator for clocks. Numerical problems as well as number theory
problems also welcome.

Please submit your nominations to me via e-mail.

Problem Name:

Problem Synopsis: (keep it short)

Problem Reference: (a ready reference to statement of problem)

Solution Reference: (a ready reference a solution if it exists)

Steve (really "D. E.") Stevenson steve@hubcap.clemson.edu
fpst@hubcap.clemson.edu
Department of Computer Science, (803)656-5880.mabell
Clemson University, Clemson, SC 29634-1906


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

From: SIAM@wharton.upenn.edu
Date: Wed, 25 Jan 89 14:05 EST
Subject: SIAM Conference on Geophysical Fluid and Solid Mechanics

Call for Papers and Registration Information

SIAM Conference on Mathematical and Computational Issues in
Geophysical Fluid and Solid Mechanics
Stouffer Greenway Plaza Hotel, Houston, Texas
September 25-28, 1989

Invited Presenters:

Alain Bamberger, Institut Francais du Petrol, France; Michael M. Carroll, Rice
University; James Dieterich, U.S. Geological Survey; Jim Douglas, Jr., Purdue
University; Bjorn Engquist, University of California, Los Angeles; Paul C.
Fife, University of Utah; James M. Hyman, University of Arizona and Los Alamos
National Laboratory; Barbara L. Keyfitz, University of Houston; Andrew J.
Majda, Princeton University; Peter Ortoleva, Indiana University, Bloomington;
George Pinder, University of Vermont; Luc Tartar, Carnegie Mellon University;
Mary F. Wheeler, University of Houston; Benjamin S. White, Exxon Research and
Engineering Company.

Conference Themes:

o Systems of Conservation Laws
o Reactive Flow
o Fluid and Solid Mechanics of Geological Materials
o Partial Differential Equations of Geosciences
o Wave Propagation and Materials Response

Contributed Presentations and Poster Presentations

A description of your talk, not exceeding 100 words, must be submitted on a
SIAM abstract form. Presentations are twenty minutes in length.

Abstract Deadline: April 12, 1989

Contributed Minisymposia

Organizers are asked to provide a title, description (100-125 words), and a
tentative list of speakers for four half-hour presentations. SIAM proposal
forms and instructions are available at your request.

Minisymposium Proposal Deadline: March 22, 1989

All inquiries should be sent to:

SIAM Conference Coordinator
117 S. 17th Street, 14th Floor
Philadelphia, PA 19103
(215) 564-2929
E-Mail: SIAM@WHARTON.UPENN.EDU
FAX: (215) 564-4174


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

From: Pramath Raj Sinha <sinha@grasp.cis.upenn.edu>
Date: 25 Jan 89 21:12:20 GMT
Subject: Quadratic and Non-Linear Programming

I have to minimise a quadratic function with respect to some non-linear
constraints and am looking for a good algorithm/routine that will do it.

The main problem is that my constraints are "or" constraints which means that
depending on the values of certain variables either this constraint is valid
or another constraint is valid.

If any if you have done any such analysis before, I would appreciate some help.

I have been struggling with the IMSL routines for Non-linear Programming - is
there anyone out there who has experience with those ?

Thanks, Pramath


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

From: Richard F. Sincovec <sincovec@riacs.edu>
Date: Fri, 27 Jan 89 15:34:50 pst
Subject: Summer Positions at RIACS

RIACS

RESEARCH INSTITUTE FOR ADVANCED COMPUTER SCIENCE

SUMMER EMPLOYMENT OPPORTUNITIES

The Research Institute for Advanced Computer Science (RIACS) is
located at the NASA Ames Research Center in Mountain View, Cali-
fornia. We are a private, non-profit institute established by a
consortium of Universities to provide leadership in computer sci-
ence research in support of NASA's goals and missions. These mis-
sions require significant advances in basic computer science and
in very large scale computations.

Each summer RIACS offers several 3 month appointments to qualified
graduate students to work as Research Assistants in close colla-
boration with RIACS and NASA scientists. We also consider applica-
tions from exceptionally well-qualified undergraduates. The stu-
dents will work within one of the three RIACS Divisions:

The LEARNING SYSTEMS DIVISION seeks fundamental new
approaches to systems for pattern computation, with appli-
cations to vision, speech, robot maneuvering, and
automatic data classification.

The PARALLEL SYSTEMS DIVISION engages in studies of the
matches between large scale scientific and the algorithms
and architectures used for their solution. The
problems arise from several scientific disciplines and the
work emphasizes the use of massively parallel architec-
tures such as the Connection Machine 2.

The NETWORKED SYSTEMS DIVISION has as its goal the proto-
typing and evaluation of operating systems, networking,
workstation, and visualization technologies that enable
spatially distributed researchers and computational
resources to work in collaboration.

RIACS and NASA Ames scientists have access to a variety of power-
ful computational resources including Cray-2, Cray YMP, CM-2, Con-
vex, Alliant, Sequent, Encore, and Multiflow processors as well as
advanced workstations from Sun, Ardent, Stellar Computer and Sili-
con Graphics.

The deadline for applications is March 15, 1989. RIACS will
respond to all applicants no later than April 15, 1989. RIACS
also has visiting faculty positions available. Applicants should
send their resumes together with a brief description of the type
of work they would wish to pursue during the summer to:

RIACS, Mail Stop 230-5
NASA Ames Research Center
Moffett Field, California 94035
(415) 694-6363

(RIACS is an equal opportunity, affirmative action employer.)


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

From: Lee Dickey <ljdickey@water.waterloo.edu>
Date: 27 Jan 89 15:18:02 GMT
Subject: Polynomial Root Finders

In article <2031@water.waterloo.edu> [on the UNIX News] I asked:
>I have recently heard about a root finding algorithm called
>Kerner's Method. I would like to know more about it.
>
> Where could I find:
> (a) Kerner's original article?
> (b) performance comparisons?
> (c) someone who has experience with using it?

I found the answer to question (a):

Ein Gesamtschrittverfahren zur Berechnung
der Nullstellen of Polynomen
Immo O. Kerner
Numerische Mathematik, 8, 290-294 (1966)

What surprises me is that noone seems to have heard about it.

In the paper Kerner mentions that Newton's Method finds one root at a
time, and that Bairstow's Method finds two at a time. With Kerner's
method, each step starts with N approximations to the roots of the
polynomial of degree N, and gives N new approximations.

He states his iteration step as

X sup (m+1) = X sup (m) + J sup -1 ( A - B( X sup (m) ) )
J = ( dB over dX ) sub ( X=X(m) )

which, to me, looks a lot like the equation I use in for Newtons
method, except that here, we
look at vectors X, A, and B, and at the matrix J.

The paper is short and sweet, and includes code in Algol which looks
to me like it would be dead easy to translate into other languages.

I am surprised that more Numerical Analysts have not heard about Kerner
and his method. I think it deserves to be more widely known. If there
are better methods, they can not be easier to learn!

L. J. Dickey, Faculty of Mathematics, University of Waterloo.
ljdickey@water.UWaterloo.ca ljdickey@water.BITNET
ljdickey@water.UUCP ..!uunet!watmath!water!ljdickey
ljdickey@water.waterloo.edu


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

From: John Lewis <@atc.boeing.com:jglewis@priapus>
Date: Fri, 27 Jan 89 21:31:23 PST
Subject: Room Sharing at Sparse Matrix Symposium

Richard Hill has offered to organize an informal room sharing
clearinghouse for the SIAM Symposium on Sparse Matrices. This
can be used to reduce expenses and to stretch the number of rooms.
Participants who wish to take advantage of his offer should
communicate directly with him by email at:

16705ROH@MSU.BITNET

The obvious important characteristics he will need to know,
besides name and address, are sex and smoking/non-smoking. We presume
that believers in iterative methods will be able to coexist with
direct solvers, and vice versa.

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

End of NA Digest

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