NA Digest Sunday, February 4, 1990 Volume 90 : Issue 05

Today's Editor: Cleve Moler

Today's Topics:

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

From: T. J. Garratt <tjg%maths.bath.ac.uk@nsfnet-relay.ac.uk>
Date: Mon, 29 Jan 90 15:30:32 GMT
Subject: Roommate Needed at Copper Mountain Conference

WANTED: Person to share room for conference:
1st - 5th April, 1990.

I am a male postgraduate studying for my PhD in Numerical Analysis
at Bath University, and will be attending the above conference.
To help with the costs of accommodation, I am looking for someone to
share a lodge room or deluxe studio.
Perhaps a student in a similar situation might be interested.

Tony Garratt,
School of Mathematical Sciences,
Univeristy of Bath,
Claverton Down, Bath.
AVON. BA2 7AY.
United Kindgom.

E-mail: tjg@uk.ac.bath.baths
(OR na.spence@edu.stanford.na-net)

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

From: Bob Ward <ward@rcwsun.EPM.ORNL.GOV>
Date: Tue, 30 Jan 90 10:42:07 EST
Subject: Liz Jessup Wins Householder Fellowship at Oak Ridge

Elizabeth R. Jessup has been selected as the winner of the first
Householder Fellowship at the Oak Ridge National Laboratory (ORNL).
Dr. Jessup, who received her doctorate degree in Computer Science in
1989 from Yale University, is currently an Assistant Professor of
Computer Science at the University of Colorado at Boulder. Her
research interests are in parallel computing and numerical linear
algebra.

Dr. Jessup will be collaborating with the researchers in ORNL's
Mathematical Sciences Section and with applied computational scientists
in various divisions at ORNL on scientific problems involving high
performance computing. Her primary interest will be on parallel
algorithms for solving large-scale eigenproblems on a
distributed-memory MIMD multiprocessor. Her fellowship appointment
will begin this summer.

Alston S. Householder was the organizer and founding Director of the
Mathematics Division (precursor of the current Mathematical Sciences
Section) at ORNL. In recognition of the seminal research contributions
of Dr. Householder to the fields of numerical analysis and scientific
computing, a distinguished postdoctoral fellowship program was
established and named in his honor. Householder Fellows will be
appointed annually for a term of one year, renewable for a second
year.

The Householder Fellowship Program is supported by the Applied
Mathematical Sciences Subprogram of the U.S. Department of Energy.

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

From: Jorge More <more@antares.mcs.anl.gov>
Date: Wed, 31 Jan 90 09:21:30 CST
Subject: Barry Smith Wins Wilkinson Fellowship at Argonne

WILKINSON FELLOWSHIP

We are pleased to announce that Barry Smith from the Courant
Institute of Mathematical Sciences is the 1990 Wilkinson fellow.
Barry is a student of Olof Widlund working on domain decomposition
algorithms for the partial differential equations of linear elasticity.
In addition to Courant, he has worked at the IBM T. J. Watson Research
Center, Los Alamos National Laboratory, and at the University of Bergen.
He will join the Mathematics and Computer Science Division of Argonne
National Laboratory in the summer.

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

From: Henry Wolkowicz <hwolkocz@orion.waterloo.edu>
Date: Mon, 29 Jan 90 15:48:23 EST
Subject: Distance of a Matrix to a Subspace

How would one find (numerically) the distance between a given real
n by n matrix A and the given subspace S, where S is the subspace
of upper triangular matrices which are themselves made up of
k by k upper triangular blocks ?

The distance is the inf of spectral norms (largest singular value).

Henry Wolkowicz; Department of Combinatorics and Optimization;
Faculty of Mathematics; University of Waterloo;
Waterloo, Ontario, Canada N2L 3G1 (519-888-4597 office; 746-6592 FAX)
{hwolkowicz@water.bitnet; na.wolkowicz@na-net.stanford.edu}
{hwolkowicz@water.uwaterloo.ca; usersunn@ualtamts.bitnet }

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

From: Ben Lotto <ben@cps3xx.egr.msu.edu>
Date: 1 Feb 90 20:10:23 GMT
Subject: Numerical Integration Program Wanted

I would like a numerical integration program that will handle a Cauchy
principal value integral of the following form:

\lim_{\epsilon\to 0}
\int_{\epsilon}^{\pi} (f(\theta - t) - f(\theta + t)) / tan(t/2) dt

(this computes the conjugate function of f) where f is a function that
has a a couple of jump discontinuities (I could probably fudge things
and get rid of this) and a log x-type singularity. In particular, I
would like the algorithm to work for the function

f(x) = log |x|, if |x| < \pi / 2
0, if |x| >= \pi / 2

-B. A. Lotto (ben@nsf1.mth.msu.edu)
Department of Mathematics/Michigan State University/East Lansing, MI 48824

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

From: Bill Anderson <XB.N64@Forsythe.Stanford.EDU>
Date: Thu, 1 Feb 90 20:41:31 PST

Last week's NA Digest included an announcement of a Summer program
programs to which I could encourage two highly qualified
undergraduates to apply? One is a math major, the other CS.

Bill Anderson
email: xa.e71@forsythe.stanford.edu

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

From: G. W. Stewart <stewart@cs.UMD.EDU>
Date: Fri, 2 Feb 90 07:47:33 -0500
Subject: Nominations Sought for Fifth Householder Prize

Alston S. Householder Award V (1990)
(Second Posting)

In recognition of the outstanding services of Alston Householder,
former Director of the Mathematics Division of the Oak Ridge National
Laboratory and Professor at the University of Tennessee, to numerical
analysis and linear algebra, it was decided at the Fourth Gatlinburg
Symposium (now renamed the Householder Symposium) in 1969 to
establish the Householder Award. This award is in the area in which
Professor Householder has worked and its natural developments, as
exemplified by the international Gatlinburg Symposia [see A. S.
Householder, The Gatlinburgs, SIAM Review 16:340-343 (1974)]. Recent
recipients of the award include James Demmel (Berkeley), Ralph Byers
(Cornell), and Nicholas Higham (Manchester).

The Householder Prize V (1990) will be awarded to the author of the
best thesis in Numerical Algebra. The term Numerical Algebra is
intended to describe those parts of mathematical research which have
both algebraic aspects and numerical content or implications. Thus
the term covers, for example, linear algebra that has numerical
applications or the algebraic aspects of ordinary differential,
partial differential, integral, and nonlinear equations.

The thesis will be assessed by an international committee consisting
of Chandler Davis (Toronto), Beresford Parlett (Berkeley), Axel Ruhe
(Gothenburg), Pete Stewart (Maryland), and Paul Van Dooren (Phillips,
Belgium).

To qualify, the thesis must be for a degree at the level of an
American Ph.D. awarded between 1 January 1987 and 31 December 1989.
An equivalent piece of work will be acceptable from those countries
where no formal thesis is normally written at that level. The
candidate's sponsor (e.g., supervisor of his research) should submit
five copies of the thesis (or equivalent) together with an appraisal
to

Professor G. W. Stewart
Department of Computer Science
University of Maryland
College Park, MD 20742
USA

by 28 February 1990. The award will be announced at the
Householder XI meeting and the candidates on the short list will

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

From: Michael Mascagni <mascagni@ncifcrf.gov>
Date: Fri, 2 Feb 90 13:04:41 EST
Subject: Washington, DC Area E-mailing List

I am happy to announce a newly formed mailing list. The list's purpose is
to distribute information 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 scientific computing. We have identified people at several sites in the
area who have agreed to serve as site contributors. We are quite biased, and
have no doubt left out several sites, group, etc. Our purpose was not to
offend, but to get things going ASAP. So if you wish to be a site contributor,
please send in a request. If you wish to be placed on the mailing list also
send us e-mail. DO NOT E-MAIL TO MY NA-NET ADDRESS. Instead, send mail to
mascagni@jvncf.csc.org with your request. As soon as we have a reasonable
number of announcements, the first mailing will go out. Until then, spread
the word, and please communicate with mascagni@jvncf.csc.org!!

Thanks for your help in this.--Michael Mascagni
(na.mascagni, but mascagni@jvncf.csc.org for this)

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

From: Jerzy Wasniewski <mfci!wasniews@uunet.UU.NET>
Date: Mon, 29 Jan 90 07:38:27 EST
Subject: Dr. Zahari Zlatev Visiting Multiflow Computer, Inc.

Dr. Zahari Zlatev
National Environmental Research Institute,
Division for Emissions and Air Pollution,
Frederiksborgvej 399,
4000 Roskilde, Denmark

visiting Multiflow Computer, Inc. Feb 14 - 16, 1990. Dr. Zlatev
will present two lectures.

1) Thursday, February 15th, 1990 - 12:00 a.m.
Multiflow Computer, Inc.
Branford, CT 06405
Tel: (203) 488-6090

SEMINAR IN COMPUTATIONAL MATHEMATICS
RUNNING LARGE AIR POLLUTION MODELS ON HIGH SPEED COMPUTERS

A b s t r a c t

The long-range transport of air pollutants ( LRTAP )
over Europe is studied, at the Air Pollution Laboratory of
the Danish Agency of Environmental Protection, by a
mathematical model based on a system of partial
differential equations ( PDE's ) . Four different
and chemical reactions (together with emission sources),
are the main components of the LRTAP . These four
processes are described by different terms in the model
(the system of PDE's). Since the space domain is very
large (including the whole of Europe together with parts of
the Atlantic Ocean, Asia and Africa), the discretization of
the system of PDE's leads to huge systems of linear
algebraic equations ( LAE's ) . In the three
dimensional case on a 32 x 32 x 9 grid the number of
LAE's that are to be solved at each time-step is more
than 10**6 when 29 chemical species are involved
in the model. Even if the model is considered as a
two-dimensional model, the number of LAE's is still
very large; more than 10**5 . This explains why one
should make some simplifications in the model description
(which are not always very well justified physically, but
lead to a model that can be handled on the computer used)
and/or one should use high-speed computers. In the latter
case, high performance can be achieved by efficiently
implementing certain kernels which perform the bulk of the
computational work. Fortunately, regular grids are to be
used during the discretization of the LRTAP model. This
leads to the solution of LAE's whose coefficient
matrices are banded and whose solution dominates the
computational load. Several such kernels for solving banded
systems of LAE's will be described. Experimental results
obtained on AMDAHL VP1100, CRAY X-MP and ALLIANT will
be presented and discussed.

2) Friday, February 16, 1990 - 11:00 a.m.
Yale University - Numerical Analysis
A. K. Watson Hall - 51 Prospect Street - room 200
New Haven, CT 06520

SEMINAR IN SCIENTIFIC COMPUTING
SOLVING GENERALLY SPARSE LINEAR SYSTEMS ON PARALLEL COMPUTERS

A b s t r a c t

Consider the system ! Ax = b !. Assume that !A! is a
large and sparse, but neither any special property of this
matrix (such as symmetry and/or positive definiteness)
nor any structure of its non-zero elements (such as
bandedness) can be exploited. For such systems direct
methods may be both time and storage consuming, while
iterative methods may not converge. A hybrid method, which
attempts to avoid the drawbacks of both direct methods and
!L! and !U! obtained by removing "small" non-zero
elements during Gaussian elimination and use them to
precondition the system. Then one of three conjugate
gradients-type methods (ORTHOMIN, GMRES and CGS) can be
used. If the iterative process does not converge, then the
criterion used in the decision whether a non-zero element
is small or not is made more stringent and new factors are
calculated and used to precondition the system. This
process can, if necessary, be repeated several times. If
after a prescribed number of trials the iterative method
is still not convergent, then a switch is made to Gaussian
elimination. Thus, with regard to the accuracy
requirements the hybrid method is not worse than Gaussian
elimination. However, even more important is the fact that
the method is often less time and storage consuming than
Gaussian elimination. This is demonstrated by many
numerical examples (including the well-known
Boeing-Harwell test-matrices).

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

From: Mikko Tarkiainen <mcsun!sunic!tut!tukki!tarkiain@uunet.uu.net>
Date: 29 Jan 90 16:41:45 GMT

Second Announcement of the

CONFERENCE ON

July 23-27, 1990 in Jyvaskyla, Finland

TOPICS OF THE MEETING. The topics covered at the conference will be:
Free boundary problems in fluid mechanics, in hydrodynamics, in
mechanics, in ground freezing and in optimal shape design, capillary
free boundaries, shape memory problems, inverse and identification
problems, control of phase transition, solidification process, etc.

PARTICIPANTS. So far, among others, the following persons are
intending to attend:
Barbu, V. (Romania), Bossavit, A. (France), Chizikalov, V.A. (USSR),
Cuvelier, C. (The Netherlands), Fage, D. (USSR), Fasano, A. (Italy),
Gets, I. (USSR), Grossman, Ch. (DDR), Haslinger, J. (Czechoslovakia),
Hoffmann, K-H (BRD), Kaliev, I. (USSR), Kenmochi, N. (Japan),
Khludnev, A.M. (USSR), Knabner, P. (BRD), Kurtze, D.A. (USA),
Magenes, E. (Italy), Maximov, A. (USSR), Meirmanov, A. (USSR),
Mittelmann, H. (USA), Myslinski, A. (Poland), Niezgodka, M. (Poland),
O'Carrol, M.J. (USA), Paolini, M. (Italy), Primicerio, M. (Italy),
Rivkind, V. (USSR), Rogers, J.C.W. (USA), Sahm, P.R. (BRD),
Schulkes, R.M.S.M.(The Netherlands), Shemetov, N. (USSR),
Shopov, P.J. (Bulgaria), Verdi, C. (Italy).

REGISTRATION. Registration forms can be ordered from the address
below. Notice that the registration must be done before March 31, 1990.
A detailed program and abstracts of the lectures will be issued to
those attending. Registration forms should be sent to Professor Pekka

CONFERENCE FEE. The conference fee, which includes attendance at the
conference, conference material, refreshments during breaks, ship
cruise on Lake P{ij{nne and conference dinner, will be $100. Participants especially from East and Southeast Europe may be given some support for the conference fee and local expenses (travel in Finland, living costs in Finland). Please inform us about required financial support in the registration form. ACCOMMODATION. Accommodation for the conference is available at the Hotel Alba on the University campus. Also, student hotels are available (2 km from the University). Please make the reservation for the accommodation, including the dates, on the accommodation registration form. If you want another hotel please inform us. If you want to stay longer in Finland before or after the conference we can help you to make reservations (hotels, summer houses, camping places, etc.) THIRD ANNOUNCEMENT including a preliminary conference program, information on preparing the paper for the conference proceedings, travel connections in Finland, etc., will be sent at the end of April 1990. Prof. Pekka Neittaanmaki University of Jyvaskyla Department of Mathematics Seminaarinkatu 15 SF-40100 Jyvaskyla, Finland email: Neittaanmaki@finjyu.bitnet tel.: (+358 41)602733 telefax: 358-41602701 telex: 28219 JYK SF Mikko Tarkiainen e-mail: mtt@jylk.jyu.fi Department of Mathematics tarkiain@tukki.jyu.fi University of Jyvaskyla, Finland phone: +358 41 292715 ------------------------------ From: Germund Dahlquist <dahlquis@nada.kth.se> Date: Fri, 2 Feb 90 12:56:51 +0100 Subject: SIAM Nordic Section meeting, June 1990 Third Annual Meeting of SIAM NORDIC SECTION June 26-27 1990 Stockholm, Sweden SIAM Nordic Section was founded in 1987. The objectives of the section are within the Nordic countries - to further the application of mathematics to industry and science - to promote basic research in mathematics leading to new methods and techniques useful to industry and science - to unite the community of researchers and graduate students in applied mathematics - to provide media for the exchange of information and ideas between mathematicians and other technical and scientific personnel. The first annual meeting was held in 1988 in Bergen, Norway, the second one in Espoo, Finland. All kinds of contributions of 25 minutes duration (including discussion) are welcome, but presentations from doctoral students and nonacademic organisations are especially invited. Please send a title of your talk and an abstract (at most one page long) before April 18, 1990. At the SIAM Nordic Section Meeting The GOLUB PRIZE will be awarded for the best contributed paper presented at the Section Meeting by a student who is from a Nordic country and has not yet finished PhD. The second Golub Prize was given to Rune Karlsson from Linkoping at the 1989 meeting in Helsinki. In addition to the contributed talks, there will be a number of talks by leading researchers from the Nordic countries. There will be a registration fee of 200 Sw.Cr. For members of the SIAM Nordic Section, 150 Sw.Cr. only. Membership can be arranged at the meeting. There will be no registration fee for graduate students from the Nordic countries. There will be a "Wine & Carrots" -party on Tuesday, June 26, at 5 p.m. The local organizer of the meeting is the Department of Numerical Analysis and Computing Science (NADA) at the Royal Institute of Technology. Housing has been arranged at a tourist class hotel, Hostel Frescati, located at the University campus, about 5 km north of Stockholm centre, while the meeting takes place at the Royal Institute of Technology. You can either have a nice (?) walk (less than 3 kms) or go by bus and subway. The same bus can also bring you downtown in about 10 minutes. Rates per night are 170 Sw Crs (about US$ 27) for a single room, 130
Sw Crs per person in a double room. The reception of the hotel is open
all the time There is an extra cost (30 Sw Crs) for linen unless you
bring linen yourself. Breakfast is not included but is served in a
Campus restaurant. If you want us to book a room for you on Hostel
Frescati, please send in the enclosed registration form as soon as
possible. Hotel prices in Stockholm are high, about 1000 Sw Crs for a
single room.

WELCOME TO THE MEETING!

Berit Gudmundson Germund Dahlquist
K T H K T H
S-100 44 Stockholm S-100 44 Stockholm
Sweden Sweden

Tel. +46 (8) 790 8077 +46 (8) 790 7142
P.S.

We like to mention that during the week June 18-22 there are two
Applied Mathematics meetings in the Nordic countries:

1) The 1990 Conference on Solution of Ordinary Differential Equations,
Helsinki, Finland (Register before April 30,1990)

Information from Prof Olavi Nevanlinna, Institute of Mathematics,
Helsinki University of Technology, 02150 Espoo 15, Finland
Email: mat-on@finhut.bitnet

2) The Householder Symposium XI Meeting in Numerical Algebra,
Tylosand,Sweden. (Deadline was Novenber 1, 1989)

Information from Prof Ake Bjorck, Dept of Mathematics,

So, if you decide to participate in one of the above meetings, you are
encouraged to extend your visit to the Nordic countries by attending
to the SIAM Nordic Section meeting. In between there is the famous
Nordic Midsummer Weekend, with midnight sun and all that + a Monday
for recovery.

By the way, there is also a great meeting in the week June 11-15:

3) 3rd International Conference on Hyperbolic Problems, Uppsala, Sweden
Information from Lena Jutestal, Dept of Scientific Computing,
Uppsala Univ, Stureg 4B, S-752 23 Uppsala, Sweden,
Email: lena@tdb.uu.se

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

From: Sven Hammarling <NAGSVEN%vax.oxford.ac.uk@nsfnet-relay.ac.uk>
Date: Mon, 29 Jan 90 18:04 GMT
Subject: NAG Floating-point Test Package

FPV is a program which attempts to test the floating-point operations + - * /
sqrt, and comparisons .LT. .GT. etc., on a systematically chosen set of
operands. The code is written with all floating-point operations in loops that
will vectorise easily. It can test that the arithmetic is rounded according to
a number of rounding rules, including all the IEEE rules. There are currently
Fortran-77 and ISO standard Pascal versions of FPV. Unlike Paranoia though, FPV
contact The Numerical Algorithms Group.

Sven Hammarling.

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

From: Andy Sherman <cs.yale.edu!topcat!sherman-andy@CS.YALE.EDU>
Date: 30 Jan 90 20:56:06 GMT
Subject: PCGPAK2 for Solving Sparse Linear Equations

SCIENTIFIC Computing Associates, Inc. is pleased to announce the
availability of PCGPAK2, its new package of subroutines for the
iterative solution of large, sparse systems of linear equations.
PCGPAK2 offers a choice of solution methods based on a collection
of preprocessing, preconditioning, and iterative techniques
that includes some of the most robust and efficient methods known.
The entire package is written in portable Fortran 77, so it can be
easily merged with the large amount of existing scientific and
engineering software that depends on solving sparse linear systems.

Four basic iterative methods are available in PCGPAK2:

--- the conjugate gradient method (CG);
--- the generalized minimal residual method (GMRES(k));
--- ORTHOMIN(k);
--- the restarted generalized conjugate residual method (GCR(k)).

All of these are Krylov subspace methods that minimize a norm of the
residual error at each step. CG is applicable only to symmetric,
positive definite systems; the others are general methods designed
mainly for systems having nonsymmetric or non-positive-definite
symmetric coefficient matrices.

PCGPAK2 includes several options that can enhance the performance of the
basic iterative methods. Among these are:

1. Incomplete factorization preconditioning --

The system is preconditioned with an approximate factorization of the
coefficient matrix generated with sparse Gaussian elimination, ignoring
some or all of the fill-in. A levelparameter is used to control the
amount of fill-in that is neglected, and a relaxation parameter is
available to fully or partially preserve the matrix row sums.

2. Reduced system preprocessing --

A preprocessing step generates a smaller, denser system that is solved
using one of the preconditioned basic iterative methods.The solution to
the full system is recovered by postprocessing the solution to the
smaller reduced system.

3. Block iteration --

All of the methods in PCGPAK2 can exploit general block structure in the
coefficient matrix. This leads to iterative methods that are extremely
robust and natural for problems with underlying block structure arising
from geometric or modeling considerations. Both constant and variable
blocksizes are supported.

PCGPAK2 is applicable to a wide range of engineering and scientific
problems that depend on the solution of large sparse systems of linear
equations. Examples of application areas include structural engineering
analysis, aerodynamic and hydrodynamic modeling, oil reservoir
simulation, ocean acoustics, simulation of VLSI circuit designs and
combustion physics. For many problems, PCGPAK2 is substantially faster
and uses far less storage than alternative banded or sparse Gaussian
elimination methods. For example, on one relatively-small nonsymmetric
system of order 3969 arising from a nine-point discretization of an
elliptic partial differential equation on the unit square,
PCGPAK2 required less than one-fourth of the time and less than
one-fifth of the storage required by the band Gaussian elimination
routines from LINPACK. For larger two-dimensional and three-dimensional
partial differential equations, the savings are far greater.

The standard Fortran version of PCGPAK2 will run on essentially any
computer. Optimized versions of PCGPAK2 are available for a number of
vector machines, including the Cray 1, Cray XMP, Cray YMP, Cray 2, IBM
3090, Convex C-1, Convex C-2, and DEC VAX 9000.

For further information, contact SCIENTIFIC at

SCIENTIFIC Computing Associates, Inc.
246 Church Street, Suite 307
New Haven, CT 06510

Tel.: (203) 777-7442
FAX: (203) 776-4074
Email: sca@yale.edu or yale!sca

PCGPAK2 is a registered trademark of SCIENTIFIC Computing Associates, Inc.
Computers mentioned may be trademarks of their respective manufacturers.

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

From: Iain Duff <duff@antares.mcs.anl.gov>
Date: Sun, 28 Jan 90 16:32:16 CST
Subject: IMA Journal of Numerical Analysis Contents

IMA JOURNAL OF NUMERICAL ANALYSIS

The contents of the current issue of the IMA Journal of Numerical
Analysis are given below.

IMA Journal of Numerical Analysis - Volume 10, Number 1

A Iserles Stability and dynamics of numerical methods
for non-linear ordinary differential
equations

M Z Liu and M N Spijker The stability of the i-methods in the
numerical solution of delay differential
equations

J Gilbert and W A Light Envelope solutions for implicit ordinary
differential equations

D Funaro Convergence analysis for pseudospectral
multidomain approximations of linear

J Solar Vortex filament method

A Bellen, A Jackiewicz, Stability analysis of Runge-Kutta methods
R Vermiglio and for Volterra integral equations of the
M Zennaro second kind

R Coquereaux, A Grossmann Iterative method for calculation of the
and B E Lautrup Weierstrass elliptic function

H Brass Optimal estimation rules for functions of
high smoothness

N Dyn, D Levin and Data dependent triangulations for piecewise
S Rippa linear interpolation

The annual subscription rate for IMAJNA is \$216 (120 pounds
outside North America and 92 pounds in UK), with a reduced rate
for members of the IMA of 38.50 pounds. There are four issues
(each of approximately 150 pages) each year. Note that it is now
possible to pay for IMA journals and IMA membership using major
credit cards.

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

From: Bob Plemmons <plemmons%matple@ncsuvx.ncsu.edu>
Date: Wed, 31 Jan 90 14:35:45 EST
Subject: SIMAX April Contents

SIAM J. on Matrix Analysis and Applications

April 1990, Vol. 11 no. 2.

1. On Perhermitian Matrices
Richard D. Hill, Ronald G. Bates, and Steven R. Waters

2. A Matrix Approach to the Design of Low-Order Regulators
L.H. Keel and S.P. Bhattacharyya

3. Some 0-1 Solutions to the Matrix Equation A(m) - A(n) = I
Chi Fai Ho

4. Sets of Positive Operators with Suprema
W.N. Anderson, Jr., T.D. Morley, and G.E. Trapp

5. Algebraic Polar Decomposition
Irving Kaplansky

6. The Laplacian Spectrum of a Graph
Robert Grone, Russell Merris, and V.S. Sunder

7. Robust Stability and Performance Analysis for State Space Systems

8. On the Singular Values of a Product of Operators

9. Points of Continuity of the Kronecker Canonical Form

10. On Rutishauser's Approach to Self-Similar Flows
D.S. Watkins and L. Elsner

11. Incremental Condition Estimation
Christian Bischof

12. A New Algorithm for Finding a Pseudoperipheral Node in a Graph
Roger G. Grimes, Daniel J. Pierce, and Horst D. Simon

The July and October issues will contain, in part, invited papers
from the Salishan, Oregon, Sparse Matrix Symposium held in 1989.

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

From: K. McKinnon <EFTM11%emas-a.edinburgh.ac.uk@nsfnet-relay.ac.uk>
Date: 01 Feb 90 10:07:38 gmt
Subject: Lectureship in Mathematics at Edinburgh University

UNIVERSITY OF EDINBURGH : DEPARTMENT OF MATHEMATICS
Lectureship in Mathematics
Particulars of Appointment

Applications are invited for a LECTURESHIP IN MATHEMATICS tenable in the
above Department. The appointment will commence on 1 October 1989 or at a date
to be decided between the department and the successful candidate.

The Department wishes to appoint an applied mathematician with strong
research interests. The ideal candidate will work in optimization theory or
numerical analysis, but strong candidates in other areas of applied
mathematics will be considered seriously. The successful candidate will have
the opportunity to interact fruitfully with the research groups in the
Department, and with other departments in the University.

There are three established chairs. The chair in Applied Mathematics is held
by D.F.Parker, whose interests include nonlinear wave propagation in solids
and optics. The other two are held by T.J.Lyons (currently Head of
Department) whose interests relate to probability theory, particularly in
analysis and geometry; and E.G.Rees whose interests are in topology and
geometry. There are five Readers, twenty four other teaching staff, two
computing officers and a number of other research workers. The interests of
the other teaching staff include optimization, numerical analysis, dynamical
systems, differential equations, analysis, probability, algebra, topology and
geometry.

The Department is responsible for teaching and research in Pure and
Applied Mathematics, and also runs (jointly with Heriot-Watt University) an
MSc course in Nonlinear Mathematics, supported by the SERC. There are separate
departments of Chemical, Electrical and Mechanical Engineering, Computer
Science, Statistics, Artificial Intelligence, Geology and Geophysics, as well
as a large Theoretical Physics group within the Department of Physics. The
Mathematics Department has strong links with the new Edinburgh-based
SERC-funded programme for the development of new techniques for design,
optimisation and control in the process engineering industries. The
Department is housed in the James Clerk Maxwell Building on the King's
Buildings site of the University, together with the combined mathematics
libraries of the University and of the Edinburgh Mathematical Society. There
are excellent computing facilities, including a 400-transputer parallel
processing facility and two Distributed Array Processors (DAPs), in the same
building. Edinburgh is an internationally recognised centre for parallel
computing.

In addition to research, duties would involve lecturing in Mathematics to
Honours and Ordinary Degree students and to postgraduate students, preparing
and attending tutorials, supervising undergraduates, examining, supervising
postgraduate students and assisting generally in the work of the Department.

The appointee is expected to join the Universities Superannuation Scheme
(USS), and to contribute 6.35% of annual salary, in which case the University
will contribute an additional sum equal to 18.55% of annual salary. The
current salary scales for lecturers A and B are 10,458 to 20,469 pounds.

The University is prepared to contribute towards removal expenses of staff
coming from other parts of the United Kingdom to Edinburgh on a first
appointment to an established post within the University, the full cost of any
reasonable vouched expenditure on removal of furniture and effects, including
insurance thereon, and the cost of fares of bringing the family to Edinburgh.
Claims in respect of travel etc from overseas will be considered on their
merits.

Applications (7 copies), including curriculum vitae and the names and
addresses of three referees, should be sent to Professor T.J.Lyons, Department
of Mathematics, Room 5320, JCMB, The King's Buildings, Edinburgh EH9 3JZ, not
later than 2nd March 1990. In the case of overseas candidates, later
applications may be considered. Such candidates need supply only one copy of
their application.

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

From: Bo Kagstrom <BOKG%SEUMDC51.BITNET@Forsythe.Stanford.EDU>
Date: Thu, 1 Feb 90 13:09 EDT
Subject: Chair in Scientific Computing at Umea

Announcement of SWEDEN's first chair as Professor in
Computer Graphics and Visualization in Scientific Computing
at the University of Umea, Sweden (Reference number: Dnr 321-189-90)

Umea university is a young university that lies at the mouth of the
river Ume, equidistant from both the capital, Stockholm, and Sweden+s
most northerly town Kiruna. Today the campus has some 3 000 employees
and 11 000 students. The university has achieved prominence in many
fields, of which bio-technology, environmental ecology and information

technology are some of those in which now intensive activity is taking
place.

Expertise in the field of information technology in its broadest sense
is rapidily growing and in certain areas such as Scientific computing

established, primarily with European and American researchers.
A couple of years ago a special action program for Information technology
- Scientific Computing was established at the faculty of Mathematical
and Natural Sciences. The program aims towards development of advanced
methods, algorithms and software in Scientific computing for different
parallel computer architectures.

The university is together with the Technical University of Lulea,
the Institute of Space Physics i Kiruna and the Industrial Development
Center in Skelleftea, founder of Supercomputer Center North (SDCN).
SDCN is one of two national centers for supercomputing in Sweden and
is connected to all swedish universities through the Swedish University
placed in Skelleftea, soon to be upgraded to a 600 J-model.

At the university we have a distributed-memory multiprocessor-system
Intel iPSC/2 hypercube with 64 nodes of which 16 nodes have a vector
facility and are about to aquire a shared-memory multiprocessor-system
with both high-performance computing power and advanced graphic
facilities for visualization.

Due to the partnership in SDCN Sweden+s first chair as professor in
Computer Graphics and Visualization in Scientific Computing is now
established at the university.

The field is very wide and interdisciplinary to its nature and candidates
for the chair can have different scientific profiles ranging from
research in tools and methods for Computer Graphics and Visualization
in Scientific Computing to graphics computing and visualization in
Scientific Computing with an emphasis on applications from biology,

biotechnology, chemistry, physics and medicine.
At the university we have applications/possible applications in for
instance biotechnology - molecular biology, chemometry, environmental

chemistry, geographical information systems, industrial design,
medicine, physical chemistry, psychology, theoretical physics and
space physics.

The professorship is placed at the department of Computing Science.
At the department there are professor chairs in numerical analysis,
computer science, and numerical analysis and parallel computing.
Since a couple of years there has been an intense development of knowledge
in the fields of parallel computing and environments and tools for
parallel computer architectures.

The university now announces a professorship in Computer Graphics and

Visualization in Scientific Computing as vacant, reference number
Dnr 321-189-90. Notice that the reference number must be mentioned on
the application!

To get started in this field as soon as possible the position can also
be a visiting professorship.

Send the application to Rektorsambetet, University of Umea, S-901 87 UMEA,

Sweden before the 30th of March 1990. Enclosed to the application should
be curriculum vitae, short summary of scientific and educational work,
and publications and ev. interest of a visiting professorship.

Questions will be answered by Professor Bo Kagstrom, Dept of Computing
Science, Umea University, S-901 87 Umea, phone +46-90165419,
email: bokg@biovax.umdc.umu.se (or na.kagstrom@na-net.stanford.edu)
or by Project coordinator Torbjorn Johansson, Supercomputer Center North,
Umea University, S-901 87 Umea, phone +46-90166585, email:
tojo@biovax.umdc.umu.se.

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

From: David Womble <dewombl@sandia.gov>
Date: 2 Feb 90 13:32:00 MST
Subject: Fellowship at Sandia National Labs

(Please distribute this announcement to colleagues and
students who do not receive the NANET distributions.)

APPLIED MATHEMATICAL SCIENCES RESEARCH FELLOWSHIP

Mathematics and Computational Science Department
Sandia National Laboratories

Sandia National Laboratories is seeking outstanding
candidates in the areas of numerical analysis, scientific
computing, or symbolic computing to fill its 1990 Applied
Mathematical Sciences Research Fellowship. The Fellowship is
supported by a special grant from the Applied Mathematical
Sciences Research Program at the U.S. Department of Energy.

The Fellowship is intended to provide an exceptional
opportunity for young researchers. Sandia's Mathematics and
Computational Science Department maintains strong programs in
theoretical computer science, analytical and computational
mathematics, computational physics and engineering, advanced
computational approaches for parallel computers, graphics, and
architectures and languages. Sandia provides a unique parallel
computing environment, including a 1024-processor NCUBE 3200
hypercube, a 1024-processor NCUBE 6400 hypercube, a Connection
Machine-2, and several large Cray supercomputers. The successful
candidate must be a U.S. citizen, must have earned a Ph.D. degree
or the equivalent, and should have a strong interest in advanced
computing research.

The fellowship appointment is for a period of one year, and
may be renewed for a second year. It includes a highly
competitive salary, moving expenses, and a generous professional
travel allowance. Applications from qualified candidates, or
nominations for the Fellowship, should be addressed to Robert
H. Banks, Division 3531-24B, Sandia National Laboratories, P.O.
Box 5800, Albuquerque, NM 87185. Applications should include a
resume, a statement of research goals, and the names of three
references. The closing date for applications is April 30, 1990.
The position will commence during 1990. Further inquiries can be
made by calling (505) 844-2248 or by sending E-mail to
RCALLEN@SANDIA.GOV.

Equal Opportunity Employer M/F/V/H
U.S. Citizenship is Required

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

End of NA Digest

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