NA Digest Sunday, May 2, 1993 Volume 93 : Issue 17

Today's Editor:

Cleve Moler
The MathWorks, Inc.

Submissions for NA Digest:

Mail to

Information about NA-NET:

Mail to


From: Jack Dongarra <>
Date: Sat, 1 May 93 08:40:20 -0400
Subject: Netlib, Xnetlib, NaNet Server to be Replaced at Tennessee

Netlib, Xnetlib, NaNet Server to be Replaced at Tennessee

On Monday, May 3rd the Netlib machine will become a Sun
Sparcstation II replacing the currently overloaded
Sequent Balance 8000. (Many thanks to Sequent Computer Systems
for donating the machine which has served over 1.2 million
requests for software since 1985.)

Users should notice no change unless they are using the
explicit machine name ``''.
If you are using ``'' use ``''
instead. The Netlib and na-digest address remain the same, and

If you have a problem, let us know

The Tennessee Netlib Group


From: Peter Wark <>
Date: Tue, 27 Apr 93 11:43:25 EST
Subject: Parallel Computing with Branch and Bound

I am seeking references on the use of Parallel Computing
with Branching (and Bound) procedures. Any information
would be appreciated.

Peter Wark
Mathematics and Computing Department
University of Southern Queensland
Toowoomba Qld 4350


From: Neil Gottlieb <70431.2166@CompuServe.COM>
Date: 28 Apr 93 12:37:32 EDT
Subject: Need Simplex Info

To Whom it May Concern:

The Society for Industrial and Applied Mathematics (SIAM) suggested I
contact you as a source for helping my research available public domain
software for linear programming. Specifically, I need software for the
Dantzig simplex method. I would appreciate any suggestions for where I can
investigate and find this type of software. I am sending this message via
Compuserve and I am told the email address is:


I can also be reached directly at (212) 514-9191. Thnaks for your help,

Neil Gottlieb, APT INC.


From: Brent Lindquist <>
Date: Wed, 28 Apr 93 19:58:31 EDT
Subject: FFT's for Shared Memory Multiprocessors

I am looking for a multidimensional (2D is sufficient)
FFT code for a shared memory multiprocessor architecture,
such as the Sequent. Something in the public domain.
Preferably written in C or FORTRAN.

Brent Lindquist
Dept. Appl. Math.
SUNY - Stony Brook


From: SIAM <bogardo@siam.ora>
Date: Wed, 28 Apr 93 16:20:59 EST
Subject: Nominations for George Polya Prize


The Polya Prize

SIAM will present the award at the 1994 SIAM Annual Meeting in San Diego,
California, July 25-29. The award honors the memory of George Polya and will
be given for a notable contribution in one of the following areas:
approximation theory, complex analysis, number theory, orthogonal polynomials,
probability theory, or mathematical discovery and learning.


There are no restrictions except that the prize is broadly intended to
recognize specific work.

Description of the Award

The award will consist of an engraved medal and a $20,000 cash prize.


A letter of nomination, including description of achievement(s), should be
sent by August 1, 1993, to:

Professor Richard A. Askey
Chair, Polya Prize Committee
c/o Allison Bogardo
3600 University City Science Center
Philadelphia, PA 19104-2688
Telephone: (215) 382-9800
Fax: (215) 386-7999


From: Dave Fournier <72730.223@CompuServe.COM>
Date: 29 Apr 93 00:21:22 EDT
Subject: C++ for Numerical Analysis

Re C++ for numerical analysis

AUTODIF IS A C++ class library which we originally developed
for our own use in building nonlinear models for resource
management. It combines an array
language with the reverse mode of automatic differentiation
to provide support for nonlinear optimization, nonlinear
parameter estimation, and sensitivity
analysis. To reduce the large amount of temporary storage
required by the reverse method, the derivative calculations
for commmon vector and matrix operations are precompiled in
``adjoint code'' in a manner transparent to the user.
Information about AUTODIF can be be obtained from

Dave Fournier
Otter Research Ltd
PO Box 265, Station A
Nanaimo, B.C. V9R 5K9
Voice or FAX 604-756-0956


From: Christoph Ueberhuber <>
Date: Mon 26 Apr 1993 09:04:03
Subject: Vienna Conference on Parallel Software

Austrian Center for Parallel Computation (ACPC)
Scientific Parallel Computation Group - Technical University Vienna

Working Conference on
the Production of High Quality Parallel Software
in Numerical and Scientific Computing

June 21 - 24, 1993
Vienna, Austria

The goal of this conference is to start coordinated development
activities for high quality PARALLEL software that is portable and
efficient (on most of present parallel computer systems).
This conference is intended as a starting point of activities which
have their counterpart in the development of high quality sequential
software (Eispack, Funpack, Linpack, etc.).


User Level: Problem Solving Environments (PSEs), Expert Systems
Library Level: Linear Algebra, Integration, ODEs, PDEs,
Transformations (FFT, Wavelets)
Compiler Level: Languages, Tools, Enquiry Functions
Machine Level: Current Target Machines, Future Development,
Machine Models

Particular emphasis will be laid on high and low level standards
(standardized kernels, enquiry functions, interfaces, test beds, etc.).


The conference language will be English.
There will be morning and afternoon lectures of approximately 1/2
hour each. An attempt will be made to avoid parallel sessions.

DEADLINE for Submission of Papers is May 12, 1993.


The conference will be held at the
Hotel Schloss Wilhelminenberg
Savoyen-Strasse 2
A-1160 Wien
Austria, Phone: (+43) 1 458503-0, Fax: (+43) 1 454876.

I am looking forward to see you in Vienna in June!

Christoph W. Ueberhuber
Institute for Applied and Numerical Mathematics
Technical University Vienna
Wiedner Hauptstrasse 8-10/115
A-1040 Wien


From: Marko Laine <>
Date: Thu, 29 Apr 93 08:53:25 +0300
Subject: Helsinki Workshop on Symbolic and Numeric Computation

Symbolic and Numeric Computation
Otaniemi, Finland 24.-28.5. 1993

University of Helsinki Helsinki University of Technology
The Academy of Finland
The Rolf Nevanlinna-Institute Centre for Scientific Computing


The Helsinki Workshop on Symbolic and Numeric Computation was
organized in May 1991 for the first time. The aims of the workshop
were then and are now to focus on combined symbolic and numeric
methods for scientific and engineering problem solving with possibly
the use of high performance computing environments and visualization
techniques. We will try to provide a forum for exchanging ideas and
introducing projects that develop or use some combination of such


The workshop is divided into two independent parts:

- An AXIOM course given by Grant Keady (Mon-Tue)
- The "actual" workshop (Wed-Fri)


The Axiom course starts at 11.00 on Mon 24.5. Auditorium G
The workshop program starts at 11.00 on Wed 26.5. Auditorium G


- Mike Dewar, University of Bath
Manipulating Fortran code in AXIOM, and the AXIOM-NAG Link
Demonstration on current version of the Axiom-NAG Link
Demonstration of IRENA
- Andre J.P. Heck, CAN Expertise Centre, Amsterdam
Computer Algebra in Industry: Applications in Optics
- Jarmo Hietarinta, University of Turku
Solving huge sets of nonlinear equations
- Grant Keady, University of Western Australia
Axiom course
Fortran code production for Argument SubPrograms in
the AXIOM-NAG link
- Mike Monagan, ETH Zurich
A demonstration of the new Maple
On automatic differentiation

- Seppo Pohjolainen, Tampere University of Technology
Topics in Control theory -- Presentation and demo using
- Alexei Serebrovski, Russian Academy of Sciences
Stochastic control theory: Applications of computational
- Mika Seppala, University of Helsinki
Computational problems in the theory of Riemann surfaces and
algebraic curves
- Mark Sofroniou, University of Loughborough
An Efficient Numeric/Symbolic Environment With Code
Optimization Using Mathematica
Demonstration of the above environment
- Esko Valkeila, University of Helsinki
On the computer generation of Poisson Processes
- Ilpo Vattulainen, CSC
On pseudo-random numbers
- Pirkka Peltola, Espoo-Vantaa Institute of Technology
Solving systems of nonlinear equations by a Maple-APL-Fortran-
- Richard Eller, TMT-Team Oy
Easy and fast matrix number crunching with data from any source
- Robert Silhol, Universite de Montpellier
Some concrete computational problems in Riemann Surfaces
- Heikki Apiola, Helsinki University (of Technology)
UNIX-ESC, An overview
- Marko Laine, University of Helsinki
Combining Matlab with Maple and Fortran libraries
- Juha Haataja, CSC
A set of Symbolic-numeric tools for handling some
problems on Dirichlet polygons
- Juha Haataja, CSC
Information services for mathematicians - a review of
using computer networks
- Jussi Rahola, CSC
Visualization in mathematical research
- Jussi Rahola, CSC
What is new in Matlab 4.0

There is room for a couple of more presentations.


Demonstration on current version of the Axiom-NAG Link
Demonstration of IRENA
ESC - Environment for Scientific Computing
Sophroniou's Mathematica-based numeric-symbolic environment
Software demonstrations by sponsors


On Thursday May 27 Mike Monagan, ETH Zurich, will give a demonstration
of Maple, especially its new version 2 on a Sun workstation. This
demonstration will be in the auditorium G at 16.15. Both the
demonstration and the informal cocktail-party in the lobby of the
auditorium are sponsored by Tiedesofta.

SOFTWARE exhibitions and demonstrations by the sponsors are present.


Helsinki University of Technology
Otakaari 1 M, Auditorium G
SF-02150 Espoo


Send your name, phone, e-mail address, organization and other relevant
information to by May 17, 1993

Anne Bjorklund Fax: +358-0-4572302
CSC-Tieteellinen laskenta Oy Phone:+358-0-4572158
PO Box 405
SF-02101 Espoo

Both the Axiom-course and the workshop are free of charge. (Money from
sponsors made it possible for us to cancel the 100 mk fee mentioned in
the second announcement.)


- Tiedesofta
- Business Advantage Development
- Control Cad
- TMT-Team


Hotel reservations will be made on request.


From: Steve Smale (
Date: Fri, 30 Apr 93 09:47:51 -0700
Subject: Real Number Algorithms


The mathematical theory of real number algorithms is to be the subject
of this proposed conference. Thus numerical analysis will be central
with emphasis on geometrical, algebraic, analytic and foundational
perspectives. Investigations of efficiency will play a special role.
Practical algorithms will be the subject of theoretical analysis, but
immediate useful results will not be demanded.

It is to be hoped that the conference will give the subject of
numerical analysis a greater coherence through a focus on the
mathematical side. In particular, an aim is to strengthen the unity
of mathematics and numerical analysis, and to narrow the gap between
pure and applied mathematics. That goal is appropriate since many of
the heroes of pure mathematics, Newton, Euler, Lagrange and Gauss
among them, established the basic real number algorithms. With the
revolution of the computer and the great achievements of scientific
computation, it does service to both the pure and applied communities
to support the mathematical development of numerical analysis.

I intend to make this proposal to AMS/SIAM for a four-week summer 1995
conference on Real Number Algorithms. My first choices for the
location are Park City, Utah and Berkeley. The most likely starting
date is July 16, 1995.

The conference would be international in character with strong
representation from the most mathematically developed parts of
numerical analysis. Besides tutorials and short courses, seminars in
the following areas are contemplated:

Linear Algebra
Non-linear systems--path following
Differential equations
Linear programming problems
Algebraic questions
Information based complexity
Lower bounds
Approximation theory

-- Steve Smale


Date: 30 Apr 1993 15:30:11 -0700 (PDT)
Subject: Mathematics Modeling Workshop

Mathematics Modeling Workshop
for Graduate Students
Aug. 13-21, 1993
The Claremont Colleges will host this Workshop as a continuation of the IMA
Summer Program in Mathematical Modeling held last August. Funding has been
provided by the National Security Agency. The goals are the same as
last year:
(1) To expose 30 mathematics and statistics graduate students
to the excitement of approaching real world problems.

(2) To introduce students to the team approach to problem solving.

The Workshop will have two parts:
A. In the first part, a week long, the students will be divided
into six teams to work on problems brought by experienced modelers.
These problems are derived from industry or applied science,
and require new and fresh insight for their definition and solution.
The problem presenters are:
Don Cohen Caltech
John Collura NSA
Chuck Gartland Kent State University
Alistair Fitt Southampton University, U.K.
Mark Mathews MIT
Eric Varley Lehigh University
B. An Industrial Mathematics Workshop lasting two days will form
the second part. Claremont has hosted, for the Mathematical Sciences
Institute, Cornell University, two of these workshops previously.
The problems presented here will come directly from industry and
will receive an intensive "brainstorming" from the assembled students and
faculty, and further invited experts from industry and academia.
This year's theme is proposed to be on the mathematics of environmental

Application Procedure. Graduate students in mathematics, applied mathematics,
statistics or operations research can be nominated for this program by a
faculty member by sending a letter of recommendation. The student must
provide a transcript. The deadline for applications is June 14.
The workshop will cover local living expenses, but students will be
expected to finance their travel.

All correspondence should be sent to:
MATHEMATICS DEPT., CLAREMONT, CA 91711-3988. Phone 909-621-8080,
FAX 909-621-8390, E-Mail:


From: Ken McKinnon <>
Date: Tue, 27 Apr 93 13:18:45 BST
Subject: Positions at University of Edinburgh

The University of Edinburgh, UK.
Department of Mathematics and Statistics
2 research positions

Discrete and Non-convex optimization

A researcher is required to work on a 2.5 year project to
develop and implement methods on serial and parallel computers for
solving discrete and non-convex optimization problems. The appointee
will be based in Edinburgh Maths Department and be part of the SERC
funded ECOSSE Process Engineering Group, which is centred in the
Edinburgh Chemical Engineering Department and involves chemical and
control engineers, mathematicians and computer scientists at the
Universities of Edinburgh, Glasgow and Dundee. The methods developed
will be applied to problems in process engineering. Applicants should
have a strong background in mathematics, operational research or some
other numerate science or engineering discipline and have computing
experience. Salary will be on the scale 12638 to 20140 pounds according
to age and experience.

Parallel Linear Programming

A researcher is required to work on a 3 year project, starting on 1st
October 1993, to develop and implement methods for linear programming on
parallel computers. Initially the project will focus on the revised
simplex method together with solution techniques for unsymmetric sparse
linear systems and use facilities in the Edinburgh Parallel Computing
Centre. Applicants should have a strong background, preferably to PhD
level, in mathematics, operational research, computer science or some
other numerate discipline. Salary will be on the scale 12638 to
15563 pounds according to age and experience.

Requests for further particulars or applications, consisting of a CV and
the names of 3 referees, should be sent as soon as possible to

Dr. Ken McKinnon
Mathematics and Statistics Dept.,
King's Buildings,
Edinburgh EH9 3JZ.
Phone: (+44) 31-650-5042
Fax: (+44) 31-650-6553


Date: Tue, 27 Apr 1993 12:22:43 +0200
Subject: Contents, Computational and Applied Mathematics

Guest Editors: N. Papamichael and E.B. Saff

J. Computational and Applied Mathematics
volume 46, Number 1-2, 14 June 1993


N. Papamichael and E.B. Saff

L.H. Howell
Numerical conformal mapping of circular arc polygons

D.M. Hough, J. Levesley and S.N. Chandler-Wilde
Numerical conformal mapping via Chebyshev weighted solutions
of Symm's integral equation

T.K. DeLillo and A.R. Elcrat
A Fornberg-like conformal mapping method for slender

N. Papamichael and E.B. Saff
Local behaviour of the error in the Bergman kernel
method for numerical conformal mapping

J.M. Floryan and C. Zemach
Schwarz-Christoffel methods for conformal mapping
of regions with a periodic boundary

T.K. DeLillo and J.A. Pfaltzgraff
Extremal distance, harmonic measure and numerical
conformal mapping

M. Greenhow
A complex variable method for the floating-body
boundary-value problem

G. Starke
Fejer-Walsh points for rational functions and
their use in the ADI iterative method

E. Wegert
Iterative methods for discrete nonlinear
Riemann-Hilbert problems

R. Wegmann
Extremal problems for harmonic mappings from
the unit disc to convex regions

W.B. Gragg
Positive definite Toeplitz matrices, the
Arnoldi process for isometric operators, and
Gaussian quadrature on the unit circle

J.-P. Berrut
A formula for optimal integration in H2

G.S. Ammar, W.B. Gragg and L. Reichel
An analogue for Szego polynomials of the
Clenshaw algorithm

W.B. Jones, W.J. Thron, O. Njastad and H. Waadeland
Szego polynomial applied to frequency analysis

C. Brezinski and A.C. Matos
Least-squares orthogonal polynomials

C. Jagels and L. Reichel
On the construction of Szego polynomials

L. Baratchart and M. Zerner
On the recovery of functions from pointwise boundary
values in a Hardy-Sobolev class of the disk

R.W. Barnard, K. Pearce and R.S. Varga
An application from partial sums of ez to a problem
in several complex variables

J.M. Borwein and P.B. Borwein
Class number three Ramanujan type series for 1/pi

J.C. Mason and G.H. Elliott
Near-minimax complex approximation by four kinds
of Chebyshev polynomial expansion

A. Iserles
A constructive approach to the Schroder equation

P.A. McCoy
Near-circularity of the error curve in Chebyshev
approximation of solutions to a class of
elliptic partial differential equations in the
complex plane


From: Richard Brualdi <>
Date: Wed, 28 Apr 1993 09:05:10 -0500 (CDT)
Subject: Contents, Linear Algebra and its Applpications

Contents Volume 186

Isaiah L. Kantor (Lund, Sweden) and Ivan M. Trishin (Moscow, Russia)
Formulas for Powers and Functions of Matrices 1

D. A. Gregory (Kingston, Ontario, Canada), S. J. Kirkland,
and B. L. Shader (Minneapolis, Minnesota)
Pick's Inequality and Tournaments 15

Gongning Chen and Zhenghong Yang (Beijing, China)
Bezoutian Representation via Vandermonde Matrices 37

Hiroyuki Osaka (Okinawa, Japan)
A Series of Absolutely Indecomposable Positive Maps in Matrix Algebras 45

Javad Faghih-Habibi (New Concord, Ohio)
The Gap of the Graph of a Matrix 55

Michel Fliess (Gif-sur-Yvette, France), Jean Le(acute)vine
(Fontainebleau, France), and Pierre Rouchon (Paris, France)
Index of an Implicit Time-Varying Linear Differential
Equation: A Noncommutative Linear Algebraic Approach 59

Zhaojun Bai (Lexington, Kentucky) and James W. Demmel (Berkeley, California)
On Swapping Diagonal Blocks in Real Schur Form 73

John Zhang (Columbus, Ohio)
Relative Growth of Linear Interations and Orthogonal
Polynomials on Several Intervals 97

J. A. Ball (Blacksburg, Virginia), J. Kim (Yongbong-Dong,
Republic of Korea), L. Rodman (Williamsburg, Virginia), and
M. Verma (Montreal, Quebec, Canada)
Minimal-Degree Coprime Factorizations of Rational Matrix Functions 117

A. Frommer (Wuppertal, Germany) and G. Mayer (Karlsruhe, Germany)
Linear Systems With cap omega-Diagonally Dominant Matrices and Related Ones

Scott H. Hochwald (Jacksonville, Florida)
Natural Conditions on the Spectra of Operators

Jiu Ding (Hattiesburg, Mississippi)
A Continuation Algorithm for a Class of Linear
Complementarity Problems Using an Extrapolation Technique 199

J. A. Dias Da Silva (Lisboa, Portugal)
A Note on Preservers of Decomposability 215

Alberto Borobia (Madrid, Spain)
Diagonals of Rotation Matrices 227

You Hong and Lan Jizu (Changchun, People's Republic of China)
Decomposition of Matrices Into 2-Involutions 235

Jacob C. Engwerda (Tilburg, the Netherlands), Andre(acute) C. M. Ran
(Amsterdam, the Netherlands), and Arie L. Rijkeboer (Tilburg, the Netherlands)
Necessary and Sufficient Conditions for the Existence of
a Positive Definite Solution of the Matrix Equation X+A asterisk X_1A=Q 255


End of NA Digest