NA Digest Saturday, July 1, 1995 Volume 95 : Issue 26

Today's Editor:
Cleve Moler
The MathWorks, Inc.

Submissions for NA Digest:

Mail to

Information about NA-NET:

Mail to

URL for the World Wide Web: -------------------------------------------------------

From: NA Digest <>
Date: Sun Jun 4 13:05:31 EDT 1995
Subject: NA Digest Calendar

NA Digest Calendar
Date Topic Place NA Digest #

July 3- 7 ICIAM, Int'l Cong. Indust. Appl. Math. Hamburg, Germany 94:23
July 6 Lothar Collatz Memorial Hamburg, Germany 25
July 3- 7 Computational Techniques Melbourne, Australia 10
July 3- 7 Mathematics of Neural Networks Oxford, UK 48
July 9-12 Error Bounds for Numerical Algorithms Oldenburg, Germany 05
July 9-13 Conjugate Gradient Methods Seattle, WA 10
July 9-14 Ill-posed Inverse Problems San Diego, CA 48
July 9-14 Math Methods in Geophysical Imaging San Diego, CA 51
July 10-12 Linear Algebra and Its Applications Manchester, UK 11
July 10-12 South African N.A. Symposium Scottburgh, So. Africa 52
July 10-14 Computational Mathematics Shushenskoe, Siberia 11
July 15-16 Matrix Methods for Statistics Montreal, Quebec 09
July 10-21 Nonlinear Waves Sapporo, Japan 43
July 16... Mathematics of Numerical Analysis Park City, UT 11
July 17-18 Identification and Optimization Prague, Czech. 12
July 17-19 Boundary Element Techniques Madison, WI 04
July 17-21 Modelling and Optimization Warsaw, Poland 32
July 19-21 Computer Arithmetic Bath, England 23
July 31... Rocky Mountain Numerical Analysis Salt Lake City, UT 32
July 31... Summer School Jyvaskyla, Finland 11

Aug. 7-16 Industrial Mathematics Modeling Raleigh, NC 12
Aug. 16-19 International Linear Algebra Society Atlanta, GA 07
Aug. 19-22 Operations Research Beijing, China 05
Aug. 21-23 Workshop on Conservation Laws Trondheim, Norway 44
Aug. 21 PARA95, ScaLAPACK & PVM NAG Lyngby, Denmark 23
Aug. 21-24 Applied Parallel Computing Lyngby, Denmark 15
Aug. 23-27 Biology, Ecology and Medicine Sofia, Bulgaria 09
Aug. 24-29 Differential Equations Rousse, Bulgaria 06
Aug. 27-31 Circuit Theory and Design Istanbul, Turkey 42
Aug. 28-30 European Simulation Meeting Gyor, Hungary 07
Aug. 28... Computational and Applied Mathematics Curitiba, Brazil 03
Aug. 29... Parallel Statistics and Economics Trier-Mainz, Germany 41

Sep. 5-10 Neural Networks and Neurocontrol Quintana Roo, Mexico 21
Sep. 5-14 Computing Techniques in Physics Skalsky dvur, Czech. 08
Sep. 12-15 Parallel Computing in Russia St.-Petersburg, Russia 45
Sep. 17-21 Acoustics of Submerged Structures Boston, MA 94:22
Sep. 18-22 European Conference on Numerical Math. Paris, France 18
Sep. 18-22 High Energy Physics Rio de Janeiro, Brazil 23
Sep. 20 Scottish Computational Maths Edinburgh, Scotland 20
Sep. 23 Pacific NorthWest Numerical Analysis Bellingham, WA 15
Sep. 25-28 Computer Methods and Water Resources Beirut, Lebanon 21
Sep. 26-29 Scientific Computing, Validated Numerics Wuppertal, Germany 16
Sep. 26... Cerfacs Linear Algebra Year Toulouse, France 15
Sep. 27-30 Mathematical Tools in Metrology Oxford, UK 51

Oct. 9-15 Cubature Formulae Krasnoyarsk, Russia 08
Oct. 13-14 Differential Equations Raleigh, NC 08
Oct. 13-14 Matrix Analysis Kalamazoo, MI 20
Oct. 15-19 Parallel Algorithms Wuhan, China 08
Oct. 16-17 Meshing Roundtable Albuquerque, NM 19
Oct. 16-18 1995 MATLAB Conference Cambridge, MA 12
Oct. 20-22 South-Central Student Conference Houston, TX 20
Oct. 23-26 SIAM Annual Meeting Charlotte, NC 05

Nov. 1- 4 Complementarity Problems Baltimore, MD 05
Nov. 6- 8 Innovative Time Integrators Amsterdam, Netherlands 19
Nov. 6- 9 Geometric Design Nashville, TN 04
Nov. 12-17 Semiconductor Device Modeling San Francisco, CA 20
Nov. 15-17 Simulation of Devices and Technologies Kruger, South Africa 01
Nov. 19-22 Pure and Applied Mathematics Isa Town, Bahrain 38

Dec. 8 Runge-Kutta Centenial Amsterdam, Netherlands 20
Dec. 10-14 Global Optimization Szeged, Hungary 26
Dec. 14-16 Dynamical Systems/Numerical Analysis Atlanta, GA 49
Dec. 14-20 Winter School on Iterative Methods Hong Kong 09
Dec. 16-19 Geophysical Inverse Problems Yosemite, CA 19

Jan. 19-21 Boundary Elements Kiel, Germany 20
Jan. 28-30 Discrete Algorithms Atlanta, GA 25
Feb. 12-14 Network Optimization Problems Gainesville, FL 47
Mar. 4- 6 Numerical Combustion New Orleans, LA 26
Apr. 1- 4 State of the Art in Numerical Analysis York, England 06
Apr. 9-11 Real Numbers and Computers Marseille, France 26
May 20-22 SIAM Conference on Optimization Victoria, BC, Canada 26
May 21-24 Graphics Interface Conference Toronto, Canada 18
June 13-15 Algebraic Multilevel Iteration Methods Nijmegen, Netherlands 11
June 17-20 Integral Methods in Science and Engin. Oulu, Finland 24
June 17-21 Householder XIII Symposium Pontresina,Switzerland 44
June 20-21 Mitrinovic Memorial Conference Belgrade, Serbia 26
July 8-12 Prague Mathematical Conference Prague, Czech Rep. 03
July 8-12 Quality of Numerical Software Oxford, England 19
July 15-19 Computational Mechanics Miskolc, Hungary 21
July 22-27 Summer Seminar on Plates and Shells Quebec City, Canada 26
July 27-30 Conference Honoring Mike Powell Cambridge, England 48
Aug. 25-31 Congress Theor. & Appl. Mechanics Kyoto, Japan 46
Sep. 2- 5 Nonlinear Programming Beijing, China 18
Sep. 9-13 "ECCOMAS 96" Paris, France 23
Sep. 9-14 Ill-Posed Problems Moscow, Russia 23


From: Kam Chuen Ng <>
Date: Thu, 22 Jun 1995 14:36:49 -0400
Subject: Nonlinear Generalized Eigenvalues

I have a nonlinear eigenvalues problem:


where A, B, C, D, E are sparse square matrices and w is the eigenvalues.
and E may not be invertible.

I can convert this problem into generalized eigenvalues problem by introducing

y=w x
z=w y
v=w z

(0  I  0  0 )(x)          (I  0  0  0)(x)
(0  0  I  0 )(y)          (0  I  0  0)(y)
(0  0  0  I )(z)  =  w    (0  0  I  0)(z)
(A  B  C  D )(v)          (0  0  0 -E)(v)
This problem can be solved by QZ algorithm but it is slow.

1. Is this the best method to solve the problem?
2. Is there a direct method to solve the problem?

3. Is there a method just to extract smallest eigenvalues (in magnitude)




From: Rudnei D. da Cunha <>
Date: Fri, 30 Jun 1995 17:47:10 -0300 (GMT-0300)
Subject: Release 2.0 of PIM

Dear colleagues,

We would like to announce the release of version 2.0 of PIM. Major
changes include four new iterative methods routines, support for
execution under MPI, a revised interface to the routines and the
possibility of monitoring the progress of the iterations.

You may obtain a copy of PIM 2.0 (Fortran 77) via anonymous FTP from
the following sites:, /pub/misc/netlib/pim/pim20.tar.Z, /pub/pim/pim20.tar.gz

and also via the WWW (via an appropriate browser):

We would like to ask you to advise us of any results you may obtain
using this package, including relevant bibliographical references.

Yours sincerely,

R. D. da Cunha ( and T. R. Hopkins (


From: Keith Weinman <>
Date: Fri, 23 Jun 1995 19:11:19 +1000 (EST)
Subject: ADI Relaxation Parameters

Greetings all,
I have written a 3-d finite volume code, much in the spirit of Patankar and
many others. The code is used to solve 1,2,3 D N.S. equations with a k -
epsilon model. Both the standard k-e model (Spalding etc) and a non-linear k-e
model (Speziale) have been included to account for turbulence effects. I have
included both a GMRES solver (from the Templates package by Dongarra et. al )
and a solver using the Thomas algorithm, the mechanics of both are well
documented in the literature. For the GMRES solver I may soon write a ILU
preconditioner, as I have found simple Jacobi preconditioning to be only
mildly effective, as noted in the literature. As a consequence of this work,
I have some questions upon which examination of the literature has shone no
real light [ and I stress that by no means have I read all the literature -
it is almost certain that I have missed many relevant articles].
The questions are as follows:
(1) What is the best method of selecting optimum under-relaxation factors (R)
for the "ADI" type solvers. At present I estimate the maximum eigen-value
(Lmax) of the most-recent coefficient matrix (A), and the code then makes a
new estimate based on how Lmax has varied over a sequence of previous
updates of A. If Lmax is reducing, or is maintaining a value within a limit,
R is reduced by a small fraction, otherwise R is increased. Note that
R1 <= R <= R2 where R1, R2 are limiting values. I feel certain that this
rather ad-hoc approach can be improved. If people could pass on appropriate
references or advice I would be very grateful.
(2) The next question relates to the structure of the solver. My motivation
in providing the code with the GMRES option was to enable a block solver to
be implemented easily. I realise that this is also possible with 'ADI'
schemes, but a GMRES scheme with an appropriate preconditioner seemed useful.
The discretized equations are only coupled through pressure terms and through
the coefficient matrix, in other words the dependence of, say the u-momentum
equation on the v-momentum equation is only implicitly realised through the
coefficients and structure of the coefficient matrix. It doesn't appear to
be a simple matter to extract an explicit dependence, from which a block
solver would benefit greatly (?). The only equation in which an explicit
dependance on the momentum equations is realised is the pressure correction
equation. Wilcox ( Turbulence Modelling for C.F.D ) also suggests that, due
to the relativly weak linkage between the momentum equations and 2-equation
turbulence models ( through the eddy viscosity terms (note: the dependance
of the momentum components through the advection terms is implicitly realised
if one uses HUDS )), there is no advantage in implementing a total block
solver for both momentum components and 2-equation model variables. Given
these points, is there any major advantage in solving the
following systems Ax = b where x=(U,V,W,P')^T and A'y=b' where y=(k,e)^T, as
opposed to solving the mean flow equations and turbulence model sequentially,
which is what I do at present. Once again, comments are welcome.

Keith Weinman Ph:
Department of Mechanical Engineering (07)-365 3536
University of Queensland Email:
St. Lucia
Brisbane, 4067
Queensland, Aust


From: George Karypis <>
Date: Mon, 26 Jun 1995 12:52:53 -0500 (CDT)
Subject: METIS: Unstructured Graph Partitioning Software

METIS: Unstructured Graph Partitioning and Sparse Matrix Ordering Software

We announce the release of the METIS software package for
partitioning unstructred graphs (unstructured finite element meshes)
and for producing fill reducing orderings of sparse matrices.
Release 1.0 of METIS is now available via WWW and ftp.

Application Domains:

Graph partitioning has extensive applications in many areas, including
scientific computing, VLSI design, and task scheduling. The problem is
to partition the vertices of a graph in k roughly equal parts, such that
the number of edges connecting vertices in different parts is minimized.
Graph partitioning is of particular importance in finite element
computations on parallel computers, since a good partition significantly
reduces the amount of communication, increasing the performance.
Graph partitioning algorithms are also used to compute fill reducing
orderings for sparse matrix factorization, and to increase the concurrency
that can be exploited during parallel direct factorization.

What is METIS?

METIS is a set of programs that implement various graph partitioning
algorithms that are based on the multilevel paradigm.
The advantages of METIS compared to other similar packages are:

- Provides high quality partitions!
The partitions produced by METIS are consistently 10% to 50% better than
those produced by spectral partitioning algorithms, and 5% to 15% better
than those produced by Chaco multilevel on a wide variety of graphs.

- It is extremely fast!
METIS is 10 to 40 times faster than multilevel spectral bisection, and
2 to 6 times faster than Chaco multilevel for a wide variety of a graphs.
Graphs with over 250,000 vertices can be partitioned in 256 parts, in
under a minute on scientific workstations. The run time of METIS is
comparable to (or even smaller than) the run time of geometric
partitioning algorithms that often produce much worse partitions.

- Provides low fill orderings!
The orderings produced by METIS are significantly better than those
produced by multiple minimum degree, particularly for large finite
element graphs. Furthermore, unlike multiple minimum degree, the
elimination trees produced by METIS are highly suited for parallel direct

METIS is freely distributed. Information on how to get the source code
is available on WWW at

Alternatively, METIS can be obtained via anonymous ftp from

Papers describing and analyzing the various algorithms implemented by
METIS can be retrieved via WWW from:

If you have any questions or problems obtaining METIS, send email to:

George Karypis, email
Vipin Kumar, email
Department of Computer Science
University of Minnessota
Minneapolis, MN 55455


From: Vladik Kreinovich <>
Date: Wed, 21 Jun 95 09:37:02 MDT
Subject: Interval Computations Homepage

Interval Computations Homepage

In accordance with the recommendations of the International Workshop on
Applications of Interval Computations (El Paso, TX, February 1995),
the interval computation homepage has been designed and
is now available. Its URL is

Homepage's Table of Contents:
* Interval Arithmetic
* Languages for Interval Analysis
* "Numerical Toolbox for Verified Computing" (information on the book
and software; in German and English): Pascal and C++ versions
* Interval Software
* Extended Bibliography on Interval and Related Methods
* Interval Ftp Site
* Homepages of Interval Computations Research Centers
* Personalia: Homepages of Interval Computations Researchers
* Reliable Computing (formerly Interval Computations),
an International Journal
* Applications of Interval Computations: General
* Applications of Interval Computations Presented at APIC'95
* APIC'95 Proceedings: Table of Contents
* Questions/Comments
* Thanks

We want to thank everyone who provided us with the links, ideas, and
material for this page.

Comments, suggestions, additions, and corrections will be
greatly appreciated. Please send them to one of the maintainers:
Vladik Kreinovich at or Misha Koshelev at


From: Vladik Kreinovich <>
Date: Thu, 22 Jun 95 09:43:26 MDT
Subject: Report from Workshop on Interval Computations

Researchers in Interval Computations Need Organization

(a brief write-up of the results of the discussion
held during the International Workshop on Applications of
Interval Computations, El Paso, TX, February 1995)

Interval Computations (and, more generally, computations with
automatic results verification) are applied more and more frequently.
Hundreds of researchers throughout the world design, analyze, and apply the
corresponding numerical methods. With the growth of the field, the
existing methods of communicating between the interval researchers
(like reading each other's papers, and communicating by mail and email)
are getting more and more complicated. The Workshop revealed that many
research projects are undertaken without the knowledge of similar
projects that have been and are successfully undertaken by other
research groups. This is not only true for applications areas, but
also in the mainstream research: e.g., for some time, spreadsheet
computations have been developed independently by three research
groups: in Finland, in Novosibirsk, Russia, and in Lafayette, LA.

The community is growing, and we need to coordinate and enhance our
efforts: we need to communicate, we need to meet, we need to
lobby and proselite. Currently, we have a journal, we have more or
less regular conferences; we have a mailing list, an ftp site
maintained by Baker Kearfott, and a Web page. However, maintaining
is done on an irregular basis, takes more and more time, and becomes
more and more chaotic.

A natural idea is to follow the example of other large groups and
researchers and organize ourselves. The least painful way is to
organize ourselves as a special interest group as
part of some well-established society.

The advantages of organization are as follows:

* First, being part of the society with well-established traditions
and know-how will make it much easier for us to organize and
publicize conferences, proceedings, etc.

* Being part of the well-established society with an existing
distribution network and existing lobbying and proseliting abilities
will drastically increase our visibility.

* If our conferences are routinely sponsored by a well-known society,
this will increase their prestige, make it easier to get funding for
organizing conferences and for going to them, make it more prestigious
for students and faculty to publish in the conference proceedings,
make the conferences more easily available in the libraries through
the databases maintained by the society.

* Many of us are already paying dues to different societies, so why not
use this money not only for the benefit of Mathematics or Computer
Science in general, but also for the benefit of our specific research field?

* Special interest groups usually regularly publish Bulletins with news.
Hardcopy news bulletins will definitely help in our communications.

What society should we be under? Interval Computations is a marginal
discipline between mathematics and computer science. Therefore, we
have a choice between AMS, SIAM, ACM, and IEEE Computer Society.
IEEE Computer Society may not be a good choice because it is already
a part of Electrical Engineering society, and we will be a part of
the part. AMS does not encourage groupings that much. We have,
however, a very good relationship with ACM: SIGNUM already published
our papers and algorithms. So, it seems like a good idea to organize
ourselves as a special interest subgroup within SIGNUM.

In view of that, it would be nice to know the opinion of the interval
researchers (not only those who were present at the workshop). Interval
researchers who are ACM members and who agree that organization within
ACM may be a good idea please send their suggestions, corrections,
names, and ACM numbers to Vladik Kreinovich at


From: R. Baker Kearfott <>
Date: Fri, 30 Jun 1995 11:31:40 -0500
Subject: Interval Computations Abstracts

The set of 74 extended abstracts for the

International Workshop on Applications of Interval Computations

that was held in El Paso on February 23--25, 1995 is available in
PostScript and TeX formats. Each abstract is filed individually
with a mnemonic for the author's name; five abstracts are contained
in separate directories, since they have extra PostScript files with
figures. Style files necessary for compiling the LaTeX and a cover page
for the set also are located with the abstracts.

The abstracts can be obtained via anonymous ftp to:

in the directory


or, equivalently, via World Wide Web browser to the URL:

Lists of talks, etc. are still available in the containing directory.

R. Baker Kearfott, (318) 482-5346 (fax)
(318) 482-5270 (work) (318) 981-9744 (home)
Department of Mathematics, University of Southwestern Louisiana


From: Tibor Csendes <>
Date: Thu, 22 Jun 95 11:26 MET
Subject: Workshop in Hungary on Global Optimization

Second Announcement
organized by the Austrian and Hungarian OR Societies

Program Committee
Pierre Hansen, Reiner Horst and Panos M. Pardalos

Date and place
December 10-14, 1995, Szeged, Hungary.

Registration fees

The registration fee covers the organizational costs of the workshop,
the accommodation with full board, and a short excursion or
sightseeing. The registration fee is 320 USD occupying a single room,
and 240 USD if a double room is shared. The registration fee for an
accompanying person is 300 USD occupying a single room and 220 USD
sharing a double room. To avoid excessive bank charges, CASH PAYMENT
IS PREFERRED. If you would like to transfer the registration fee,
please make sure that the fee arrives in full amount, and bring a
proper evidence of the transfer with you. For the bank transfer use
the following address:

Workshop on Global Optimization
Postabank es Takarekpenztar Rt., Szeged, Hungary
JATE: 026-00716, Sort code: 401-4131-916-01

Important dates

June 30, 1995 - deadline for the preliminary registration
September 15, 1995 - deadline for extended abstract submission
October 31, 1995 - notification of acceptance
November 30, 1995 - deadline for the final manuscripts for the proceeding
December 9, 1995 - check-in at the hotel
December 10-14, 1995 - Workshop on Global Optimization
December 15, 1995 - check-out at the hotel

Address of Organizing Committee

Tibor Csendes
Jozsef Attila University, Institute of Informatics
H-6701 Szeged, P.O. Box 652, Hungary
Phone: +36 62 310 011 (ext. 3839), Fax: +36 62 312 292
in the directory /pub/math/optimization/globopt


From: Trini Flores <>
Date: Thu, 22 Jun 95 14:14:10 EST
Subject: SIAM Conference on Numerical Combustion

Sixth International Conference on NUMERICAL COMBUSTION

March 4-6, 1996
Le Meridien New Orleans Hotel
New Orleans, Louisiana

Conducted by SIAM with the cooperation of Institut National de
Recherche en Informatique et en Automatique (INRIA)


Advances in computational algorithms and hardware continue to
have a revolutionary impact on the combustion sciences, and
permit the examination of scientific and engineering problems of
increasing complexity. The Sixth International Conference on
Numerical Combustion will focus on the integration of theory,
modeling, and numerical implementation in the study of basic
combustion physics and technological applications. The distinct
questions and challenges found in combustion and phase transition
arise from the multiplicity of length and time scales defined by
the chemical, geometric, and flow ingredients. Physically
descriptive, efficient and accurate numerical modeling
of complex phenomena is the subject of the conference.

This conference is designed to bring together computational
scientists, research and design engineers, and mathematical and
physical scientists who are interested in the interdisciplinary
area of numerical combustion.

The conference will highlight the mutual interaction of
designers, experimentalists, computational experts, and
mathematicians in the development of new ideas and methods in the
study of combustion.


The themes of the conference include, but not limited to:

o Turbulence o Kinetics o Detonation o Flames
o Pollution o Microgravity o Applications of parallel
processing o Materials synthesis
o Droplets and sprays o Ignition o Heterogeneous combustion
o Energetic materials (propellants)
o Simulation of internal engine and furnace combustion


The Transport of Combustion Products from Fires
Howard Baum
National Institute of Standards and Technology

State-of-the-Art in IC Engine Combustion Modeling
Sherif El-Tahry
General Motors Research Laboratories

Numerical Simulation of Premixed Flame Propagation in Closed Tube
Kunio Kuwahara
The Institute of Space and Astronautical Science, Japan

Coupling of Chemical Kinetics with Flow and Molecular Transport
Ulrich Maas
Konrad-Zuse-Zentrum fur Informationstechnik Berlin, Germany

Direct Simulations and Modeling of Flame/Wall Interactions
Thierry Poinsot

The Dynamics of Multi-Dimensional Detonation
D. Scott Stewart
University of Illinois, Urbana

John Buckmaster (Co-Chair)
Mitchell Smooke (Co-Chair)
D. Scott Stewart (Co-Chair)
Roland Borghi
Sebastian Candel
Robert Kee
Bernard Larrouturou
Elaine Oran
Norbert Peters
Bernd Rogg
Tadao Takeno
Jurgen Warnatz
Charles Westbrook
Forman A. Williams


August 31, 1995 - Deadline for submission of minisymposium

September 25, 1995 - Deadline for submission of contributed


Additional information regarding the conference can be accessed
in electronic format via SIAM's Gopher server: or
through the World Wide Web:


From: Michel Fortin <>
Date: Tue, 27 Jun 95 09:38:46 EDT
Subject: Summer Seminar on Plates and Shells

First announcement:
Canadian Mathematical Society, Summer Seminar 1996
Plates and shells: from mathematical models to engineering practice
July 22-27 1996, Laval University,
Quebec city, Canada


This five days seminars will bring together all groups interested to the
theory, numerical analysis and engineering use of plates and shells.
There will be a few 2-hour minicourses on various aspect of the subject.

We would also like to receive contributed papers.Theoretical and applied
papers are both welcome. Papers will be selected based on a two page
abstract. The deadline for abstracts will be January 15 1996 and
notification of acceptance will be sent by February 1996. Abstracts should
begin with the title of the paper, author's name, affiliation and email
address if available. The abstract should present a succint statement of
the results and a discussion of its significance.

Submissions should be sent to

Dep. de mathematiques et de Telephone (418) 656-5220
Universite Laval
Quebec, Canada, G1K-7P4 Fax: (418) 656-2817


From: Gradimir Milovanovic <>
Date: Tue, 27 Jun 1995 16:10:47 EDT
Subject: Mitrinovic Memorial Conference

Dragoslav S. Mitrinovic
Memorial Conference
Belgrade, June 20--21, 1996


The following institutions:

- Serbian Scientific Society, Belgrade,
- Faculty of Electrical Engineering, Belgrade,
- Faculty of Electronic Engineering, Nis,
- Institute of Mathematics SANU, Belgrade,

organize an International conference devoted to memory of the outstanding

Professor Dragoslav S. Mitrinovic

and his scientific work.

The Organizing commitee kindly ask for contribution in the following

- Approximation Theory
- Complex Analysis
- Differential, Integral and Functional Equations
- General Inequalities
- Orthogonal Polynomials and Special Functions

Other relevant mathematical subjects are also wellcomed.


The Dragoslav S. Mitrinovic Memorial Conference will take place at
Serbian Scientific Society (Safarikova 7, Belgrade, Serbia-Yugoslavia),
June 20-21, 1996. Two publications are planed to be published in ahead of
the conference with survey and shorter contributed papers (up to 8 pages).
Due to last Professor Mitrinovic passion -- Inequalities, one publication
will be entitled as "Progress in Inequalities". The other one will be
comprised from the selected papers from other topics. All the submitted
papers will be subjected to the referee process.

Manuscripts should be submitted in two hard copies (up to December 15,
1995) to the following address:

Prof. Gradimir V. Milovanovic
Faculty of Electronic Engineering, P.O. Box 73
18000 Nis, Serbia, Yugoslavia


From: Trini Flores <>
Date: Thu, 29 Jun 95 12:43:25 EST
Subject: SIAM Conference on Optimization

Fifth SIAM Conference on Optimization

Sponsored by SIAM Activity Group on Optimization

May 20-22, 1996
Victoria Conference Center
Victoria, British Columbia, Canada


About the Conference...

The field of optimization is a fascinating and lively blend
of theoretical analysis, algorithm and software development, and
scientific computing. This fifth conference will address the
most important recent developments in linear, nonlinear, and
discrete optimization. It will feature recent advances in
optimization algorithms and software, as well as important
applications of optimization in control, networks, manufacturing,
chemical engineering, and operations research. An important
emphasis of the meeting is also the increasing variety of
connections between optimization and other fields of numerical
analysis and scientific computing, such as differential
equations. The organizers have made a particular effort to
highlight some less traditional themes.

Conference Themes

o Connections between continuous and discrete optimization
o Convex analysis and applications
o Differential-algebraic equations and their connections to
o Derivative-free methods
o Industrial methodology and applications
o Nondifferentiable/structural optimization
o Semidefinite programming
o Stochastic programming

The conference will feature seven invited speakers:

Uri M. Ascher
Albert M. Erisman
Martin Groetschel
Adrian S. Lewis
Andrzej Ruszczynski
Virginia Torczon
Jochem Zowe

Electronic Access

Additional information regarding the conference,
including information on the City of Victoria, maps,
weather and travel, can be accessed in electronic
format via Gopher servers or through the World Wide
Web: (, (,
(, (

Conference Co-chairs: Andrew R. Conn and Margaret H. Wright


From: Jean-Claude Bajard <>
Date: Fri, 30 Jun 95 14:58:39 +0100
Subject: Conference in France on Real Numbers and Computers


Marseille, FRANCE
April 9-10-11 Avril 1996

The first ``real numbers and computers'' conference (St-Etienne, April 1995)
was succesfull. Many mathematicians or computer scientits are interested
by this domain. Thus we propose a second edition of this
conference in Marseille April 9-10-11 1996.

Efficient handling of real numbers in a computer is not yet solved in a
satisfying way. The "floating point" formats most often used in scientific
computing usually give sufficient results, but some reliability problems
can occur. Program portability problems could imply some rewriting costs:
some programs which work well with a machine, could become unreliable with
another one. Users (working on computer algebra, algorithmic geometry) may
need far more accurate results (even "exact results") than the ones
obtained with usual number systems. Many members of the scientific
community are concerned by this problem, they could share their knowledge
and come up with solutions. But they do not have the opportunity to meet,
they do not belong to the same scientific fields (computer science, number
theory, numerical analysis, computer algebra) and they have a different
vocabulary. The aim is to put them together during this meeting (the "rule
of the game" should be that everybody should speak a common language), in
order to establish some collaborations.

You can send 4 copies of a printed version of a full paper (not an
abstract) to:

Jean-Claude BAJARD
Laboratoire de l'Informatique de Marseille,
Centre de Math\'ematiques et d'Informatique,
Universit\'e de Provence,
39 rue Joliot-Curie,
13453 Marseille cedex 13, FRANCE.


- Algorithms and architectures for "serial" and "on line" arithmetic.
- Relations between number theory, automata theory and computer arithmetic.
- Number systems
- Floating point arithmetic
- Calculability
- Symbolic manipulation of numbers
- Algorithms for "exact" computing
- Multi-precision, interval arithmetic
- Accuracy problems in various fields and proposed solutions.

Local committee

- Jean-Paul ALLOUCHE, (LMD, CNRS, Marseille, France)
- Jean-Claude BAJARD, (LIM, Universit\'e de Provence, Marseille, France)
- Solange COUPET, (LIM, Universit\'e de Provence, Marseille, France)
- Pierre LIARDET, (LATP, Universit\'e de Provence, Marseille, France)

Program committee

- Jean-Paul ALLOUCHE, LMD, Universit\'e de Luminy, Marseille, France.
- Jean-Claude BAJARD, Laboratoire LIM, Universit\'e de Provence, France.
- Jean-Claude BERGES, CNES, Centre Spatial de Toulouse, France.
- Vasco BRATTKA, FernUniversitat, Theorische Informatik 1, Hagen, Germany.
- Jean-Marie CHESNEAUX, Lab.MASI, Universit\'e Pierre et Marie Curie, Paris, France.
- Solange COUPET, Laboratoire LIM, Universit\'e de Provence,France.
- Christiane FROUGNY, LITP, Universit\'e Pierre et Marie Curie, Paris, France.
- Peter KORNERUP, Dept of Mathematics and Computer Science, Odense, Danmark.
- David LESTER, Department of Computer Sciencem Manchester, United Kingdom.
- Pierre LIARDET, LATP URA225, Universit\'e de Provence, France.
- Maurice MARGENSTERN, Universit\'e Paris Sud, Laboratoire de Math\'ematiques, France.
- Dominique MICHELUCCI, Ecole Nationale des Mines de Saint Etienne,
SIMADE, France.
- Jean-Michel MOREAU, Ecole Nationale des Mines de Saint Etienne, SIMADE, France.
- Jean-Michel MULLER, Lab. Lip, Ecole Normale Sup\'erieure de Lyon, France.


From: Chris Petrie <>
Date: Fri, 23 Jun 1995 09:50:13 +0100 (BST)
Subject: Position at University of Newcastle upon Tyne


Available from one year from 1st September 1995, this post will
coincide with the imminent arrival of Professor Y.A. Sergeev to
take up the Chair in Engineering Mathematics following the
retirement of Professor Alan Jeffrey.

The person appointed may be either a mathematician or an engineer
with a strong mathematical background who will be expected to
contribute to the wide range of courses in Engineering
Mathematics provided for students in the Faculty of Engineering,
from foundation year to postgraduate. Opportunities for research
include involvement with other departments in the Faculty, either
in the form of collaborative projects or in the joint supervision
of research students. The Department includes engineers,
mathematicians and applied statisticians on its staff.

Salary will be at an appropriate point on the Lecturer Grade A
scale: #15,154-#19,848 per annum, according to qualifications and

No forms of application are issued. Further particulars may be
obtained from the Director of Personnel, Registrar's Office,
University of Newcastle, 1 Park Terrace, Newcastle upon Tyne, NE1
7RU with whom three complete copies of applications, including a
full cv with present salary and the names of three referees
should be lodged not later than 21st July 1995.

Informal enquiries may be addressed to the Department,

FAX: +44 - 191 - 222 - 5498

There is also a departmental World-Wide-Web page (partly developed),
The University's home page is at


From: Jeff Scroggs <>
Date: Thu, 29 Jun 1995 16:37:31 -0400
Subject: Postdoc Position at North Carolina State University

The Center for Research in Scientific Computation is preparing
to consider applications for our postdoc position on the
High-pressure Vapor Transport Reactor project originally announced
in NA Digest Volume 94, Issue 49. See
for more information about the center, and for
a brief description of the project.

If you have not applied, but plan to do so, please immediately contact

Jeff Scroggs
EMAIL: (preferred)
AT&T: (919)515-7817 work
FAX: (919)515-3798
Department of Mathematics
Box 8205
North Carolina State University
Raleigh, NC 27695

NCSU is an AA/EOE. In its commitment to diversity and equity, NCSU and the
CRSC seeks applications especially from women, minorities, and the


From: George Corliss <>
Date: Thu, 29 Jun 95 17:17 CDT
Subject: Graduate Assistantships at Marquette University

Industrial-Based Graduate Research Assistantships
Department of Mathematics, Statistics, and Computer Science
Marquette University, Milwaukee, Wisconsin, USA

The Marquette University Department of Mathematics, Statistics,
and Computer Science has openings for two Industrial-Based
Graduate Research Assistantships in industrial and applied
mathematics. Students will pursue a Master of Science degree
while working closely with industrial sponsors Johnson Controls
and SC Johnson Wax. One assistantship requires skills in
operations research, the other in statistics and database

Number of openings: two
Start date: August 15, 1995, or January 1, 1996
Annual stipend: $12,000 + tuition,
jointly funded by NSF and industrial partners
Students from groups traditionally underrepresented in the field
are especially sought.

For information:
Dr. George Corliss Dr. Karl Byleen
(414) 288-6599 (414) 288-6343
To apply:
The Graduate School
Marquette University
PO Box 1881
Milwaukee, WI 53201-1881
(414) 288-7137 Fax: 414) 288-1902

Students from groups traditionally underrepresented in the field
are especially sought.

For further information:
cd pub/corliss
get indust_gra


From: Marilyn Radcliff <>
Date: Thu, 22 Jun 1995 10:45:43 -0400 (EDT)
Subject: Contents, J. Approximation Theory

Table of Contents: J. Approx. Theory, Volume 81, Number 3, June 1995

Ding-Xuan Zhou
On smoothness characterized by Bernstein type operators

Johan Lithner and Adam P. W\'ojcik
A note on Berntein's theorems

Ding-Xuan Zhou
Construction of real-valued wavelets by symmetry

E. Kochneff
Expansions in Laguerre polynomials of negative order

Graeme J. Byrne, T. M. Mills, and Simon J. Smith
The Lebesque constant for higher order Hermite-Fej\'er interpolation on
the Chebyshev nodes

Peter K\"ohler and Geno Nikolov
Error bounds for Gauss type quadrature formulae related to spaces of
splines with equidistant knots

Gilbert Helmberg
A limit function for equidistant Fourier interpolation

Peter K\"ohler and Geno Nikolov
Error bounds for optimal definite quadrature formulae

Xie Ping Ding and E. Tarafdar
Some further generalizations of Ky Fan's best approximation theorem


Horst Alzer
On the zeroes of a polynomial

Fernando Mazzone and H\'ector Cuenya
A note on metric projections


From: SIAM <>
Date: Thu, 22 Jun 95 10:58:20 EST
Subject: Contents, SIAM Optimization

VOL.5,NO.3, AUG 1995

Nonpolyhedral Relaxations of Graph Bisection Problems
Svatopluk Poljak and Franz Rendl

Faster Simulated Annealing
Bennett Fox

Incorporating Condition Measures into the Complexity Theory
of Linear Programming
James Renegar

Global Convergence of a Long-Step Affine-Scaling Algorithm
for Degenerate Linear Programming Problems
Takashi Tsuchiya and Masakazu Muramatsu

On Eigenvalue Optimization
Alexander Shapiro and Michael K. H. Fan

Data Parallel Quadratic Programming on Box-Constrained Problems
Mike P. McKenna, Jill P. Mesirov, and Stavros A. Zenios

A Sequential Quadratic Programming Algorithm Using an
Incomplete Solution of the Subproblem
Walter Murray and Francisco J. Prieto

Local Convergence of SQP Methods in Semi-Infinite Programming
G. Gramlich, R. Hettich, and E. W. Sachs

Taylor's Formula for Ck,1 Functions
Dinh The Luc

The Linear Nonconvex Generalized Gradient and Lagrange Multipliers
Jay S. Treiman

On the Simulation and Control of Some Friction Constrained Motions
Roland Glowinski and Anthony J. Kearsley


From: SIAM <>
Date: Thu, 22 Jun 95 13:48:41 EST
Subject: Contents, SIAM Matrix Analysis

VOL. 16, NO. 4, OCTOBER 1995

On a Sturm Sequence of Polynomials for Unitary Hessenberg Matrices
Angelika Bunse-Gerstner and Chunyang He

Least Squares Sign Solvability
Bryan L. Shader

On Eigenvalue Estimates for Block Incomplete Factorization Methods
O. Axelsson and H. Lu

Diagonal Dominance in the Parallel Partition Method for Tridiagonal Systems
Chris Walshaw

Matrices with Sign Consistency of a Given Order
J. M. Pena

On a QR-like Algorithm for Some Structured Eigenvalue Problems
A. George, Kh. D. Ikramov, E. V. Matushkina, and
W.-P. Tang

The Group Inverse Associated with an Irreducible Periodic Nonnegative Matrix
Steve Kirkland

Variable Block CG Algorithms for Solving Large Sparse Symmetric Positive
Definite Linear Systems on Parallel Computers, I: General Iterative Scheme
A. A. Nikishin and A. Yu Yeremin

A Restarted GMRES Method Augmented with Eigenvectors
Ronald B. Morgan

Comments on Large Least Squares Problems Involving Kronecker Products
Hongyuan Zha

Trace and Eigenvalue Inequalities for Ordinary and Hadamard Products of
Positive Semidefinite Hermitian Matrices
Bo-Ying Wang and Fuzhen Zhang

A Basis-Kernal Representation of Orthogonal Matrices
Xiaobai Sun and Christian Bischof

On the Convergence of the Jacobi Method for Arbitrary Orderings
Walter F. Mascarenhas

Multisplitting Preconditioners Based on Incomplete Choleski Factorizations
R. Bru, C. Corral, A. Martinez, J. Mas

On the Symmetric and Unsymmetric Solution Set of Interval Systems
Gotz Alefeld and Gunter Mayer

A Domain Decomposition Method for First-Order PDEs
Lina Hemmingsson

Some Properties of Fully Semimonotone Q0-Matrices
G. S. R. Murthy and T. Parthasarathy

Stability of Linear Equations Solvers in Interior-Point Methods
Stephen J. Wright

The Algebraic Riccati Equation and Inequality for Systems with
Uncontrollable Modes on the Imaginary Axis
Carsten W. Scherer

Perturbation Bounds for the Generalized Shur Decomposition
Ji-guang Sun

Application of Vector-Valued Rational Approximations to the Matrix
Eigenvalue Problem and Connections with Krylov Subspace Methods
Avram Sidi


From: Carlos Antonio de Moura <>
Date: Thu, 22 Jun 1995 12:35:24 +0300
Subject: Contents, Computation and Applied Mathematics


Published by Birkhauser/Boston
and SBMAC - Brazilian Soc. for
Comp. and Applied Mathematics

Vol.14:1, 1995
Special Issue on
High Performance Scientific Computing

C.A. de Moura 1

Madpack: A family of abstract multigrid
or multilevel solutions
Craig C. Douglas 3

A multigrid solver for the steady state Navier-Stokes
Equations using the Pressure-Poisson formulation
David Sidlikover and Uri M. Ascher 21

A posteriori error estimates for general numerical
methods for scalar conservation laws
Bernardo Cockburn and Huiing Gau 37

SIMPAR: a parallel sparse simplex
M. Lentini, A. Reinoza, A. Teruel and A. Guill'en 49

An introduction to DIMSIMs
J.C. Butcher 59

A parallelizable characteristic scheme for two phase flow I:
Single porosity models
Jim Douglas, Jr., Felipe Pereira, and Li-Ming Yeh 73

Parallel computation of turbulent fluid flow
Paul R. Woodward, David H. Porter, B. Kevin Edgar,
Steven Anderson, and Gene Bassett 97


From: Baltzer Science Publishers <>
Date: Fri, 23 Jun 1995 14:36:41 +0200
Subject: Contents, Advances in Computational Mathematics

Advances in Computational Mathematics, Volume 3, No. IV, 1995, ISSN 1019
7168 Editors-in-Chief: John C. Mason & Charles A. Micchelli

Advances in Computational Mathematics is an interdisciplinary journal of
high quality, driven by the computational revolution and emphasising
innovation, application and practicality.
This journal is of interest to a wide audience of mathematicians,
scientists and engineers concerned with the development of mathematical
principles and practical issues in computational mathematics.

Volume 3, No. IV, 1995

pp. 309-342: R.-Q. Jia, Subdivision schemes in Lp spaces

pp. 343-352: G. Hall, A new stepsize strategy for explicit Runge-Kutta codes

pp. 353-368: K. Jetter and J. Stoeckler, A generalization of de Boor's
stability result and symmetric preconditioning

pp. 369-374: C.T. Chong, The polynomial topological complexity of
Fatou-Julia sets

pp. 375-394: J.M. Carnicer, T.N.T. Goodman and J.M. Pena, A generalization
of the variation diminishing property

pp. 393-404: J.M. Carnicer, Multivariate convexity preserving interpolation
by smooth functions

pp. 405-424: P.E. Koch, T. Lyche, M. Neamtu and L.L. Schumaker, Control
curves and knot insertion for trigonometric splines



F. Keinert, Numerical stability of biorthogonal wavelet transforms

S. Zeng, C. Vuik and P. Wesseling, Numerical solution of the incompressible
Navier--Stokes equations by Krylov subspace and multigrid methods

K. Urban, On divergence-free wavelets

P.W. Hemker, Sparse-grid finite-volume multigrid for 3D-problems

S.C. Brenner, A two-level additive Schwarz preconditioner for the
stationary Stokes equations

B. Koren and B. van Leer, Analysis of preconditioning and multigrid for
Euler flows with low-subsonic regions

A. Kunoth, Multilevel preconditioning -- Appending boundary conditions by
Lagrange multipliers

M. Griebel and P. Oswald, Tensor product type subspace splittings and
multilevel iterative methods for anisotropic problems

Submissions of articles and proposals for special issues are to be
addressed to the Editors-in-Chief:

John C. Mason
School of Computing and Mathematics, University of Huddersfield,
Queensgate, Hudersfield, HD1 3DH, United Kingdom


Charles A. Micchelli
Mathematical Sciences Department
IBM Research Center
P.O. Box 218, Yorktown Heights, NY 10598, USA E-mail: cam@yktvmz.bitnet

Requests for FREE SPECIMEN copies and orders for Advances in Computational
Mathematics are to be sent to: E-mail:


From: SIAM <>
Date: Tue, 27 Jun 95 08:56:26 EST
Subject: Contents, SIAM Numerical Analysis
Subject: Contents, SIAM Numerical Analysis

SIAM Journal on Numerical Analysis
AUGUST 1995, Volume 32, Number 4

A Fast Solver for Navier-Stokes Equations in the Laminar Regime
Using Mortar Finite Element and Boundary Element Methods
Y. Achdou and O. Pironneau

Projection Method I: Convergence and Numerical Boundary Layers
Weinan E and Jian-Guo Liu

Numerics and Hydrodynamic Stability: Toward Error Control in
Computational Fluid Dynamics
Claes Johnson, Rolf Rannacher, and Mats Boman

Convergence of Particle Methods with Random Rezoning for the Two-
Dimensional Euler and Navier-Stokes Equations
Y. Brenier and G.-H. Cottet

Particle Approximation of a Linear Convection-Diffusion Problem
with Neumann Boundary Conditions
S. Mas-Gallic

Multigrid Smoothing Factors for Red-Black Gauss-Seidel Relaxation
Applied to a Class of Elliptic Operators
Irad Yavneh

Numerical Solutions of One-Pressure Models in Multifluid Flows
Fabienne Berger and Jean-Francois Colombeau

Finite Element Approximation of Time Harmonic Waves in Periodic Structures
Gang Bao

Error Estimates on a New Nonlinear Galerkin Method Based on Two-
Grid Finite Elements
Martine Marion and Jinchao Xu

Convergence of a Second-Order Scheme for the Nonlinear Dynamical
Equations of Elastic Rods
Richard S. Falk and Jian-Ming Xu

Optimal Selection of the Bubble Function in the Stabilization of
the P1-P1 Element for the Stokes Problem
Roger Pierre

Thermal Simulation of Pipeline Flow
Philip T. Keenan

Spectral Approximation of a Boundary Condition for an Eigenvalue Problem
Anne-Sophie Bonnet-Bendhia and Nabil Gmati

Finite Element Vibration Analysis of Fluid-Solid Systems Without
Spurious Modes
A. Bermudez, R. Duran, M. A. Muschietti, R. Rodriguez, and J.

Runge-Kutta Solutions of Stiff Differential Equations Near
Stationary Points
Ch. Lubich, K. Nipp, and D. Stoffer

A Product-Decomposition Bound for Bezout Numbers
Alexander P. Morgan, Andrew J. Sommese, and Charles W. Wampler


End of NA Digest