## NA Digest Sunday, January 28, 1996 Volume 96 : Issue 04

Today's Editor:
Cleve Moler
The MathWorks, Inc.
moler@mathworks.com

### Submissions for NA Digest:

Mail to na.digest@na-net.ornl.gov.

Mail to na.help@na-net.ornl.gov.

URL for the World Wide Web: http://www.netlib.org/na-net/na_home.html -------------------------------------------------------

From: Tim Davis <davis@cis.ufl.edu>
Date: Mon, 22 Jan 1996 09:31:18 -0500
Subject: Approximate Minimum Degree Algorithm

The AMD algorithm is a new variant of the minimum degree ordering method for
sparse matrices, based on a cheap-to-compute bound on the degree of nodes in
the quotient graph. We've just placed 6 variants of the AMD routine (but not
(http://www.cis.ufl.edu/~davis, also includes a tech report). The 6 codes
differ only in how they compute the degree or its upper bound, so that
different degrees can be compared in a consistent framework. These include:

Routine Method
AMDBAR our bound on the external degree. Computes an ordering of
the same quality as the best minimum degree ordering codes.
Asymptotically faster, and usually faster in practice.
AMDHAT the bound on the external degree used in MATLAB's symmmd
and colmmd routines. Usually as fast, or faster, than AMDBAR,
but tends to compute worse orderings.
AMDHAF a combination of AMDHAT and AMDEXA. Somewhat slower than
AMDBAR, can sometimes produce worse orderings than AMDBAR.
AMDATR our bound on the true degree.
AMDEXA the exact external degree, as used in MMD, for example.
AMDTRU the exact true degree, as used in MA27, for example.

For the last two methods, the leading "A" (for "Approximate") is actually a mis-
nomer, since they don't use approximations. Licensing restrictions apply.
The method is described in "An approximate minimum degree ordering algorithm,"
SIAM J. Matrix Analysis and Applications (to appear), by Patrick Amestoy,
Tim Davis, and Iain Duff. The AMD routine itself (similar to AMDBAR but
better) is intended to appear as MC47B/BD in the next release of the Harwell
Subroutine Library.

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

From: Rene Aid <Rene.Aid@imag.fr>
Date: Fri, 26 Jan 1996 18:24:22 +0100
Subject: Differential Equation in the Complex Domain

I am looking for Fortran or C sofwares that can integrate Ordinary
Differential Equations with a complex time :

dy/dz = f(z,y)

z, y and f complex valued variables.

Thanks.

Rene.Aid@imag.fr net : rene.aid@imag.fr
LMC/IMAG tel : (33) 76 57 48 66
46, Avenue Felix Viallet fax : (33) 76 57 47 54
Grenoble Cedex FRANCE

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

From: David Carrier <CarrierD@aol.com>
Date: Sun, 28 Jan 1996 07:17:26 -0500
Subject: Economic and Financial Data Available

Dear colleague;

You may wish to post the following message to NA-NET:

I have the entire U.S. National Income and Product Accounts, Input- Output
tables, state and metro area employment and income, and Flow of Funds and
other Federal Reserve data, on diskette in spreadsheet format-- I reformatted
the raw government data into tables that I could work with more easily for my
own research. If anyone is interested in avoiding this hassle and obtaining
my data, respond to this message and include your e-mail address, and I'll

The data is in delimited text files with table titles and series descriptions
included, and it can be opened and edited with any spreadsheet, statistical,
or word processing program. For the cost of a diskette and mailing I will
send you a demo diskette of quarterly current dollar and chain-weighted real
GDP, GDP by type of product, and GDP by sector for the U.S., 1959Q1 to
1995Q3. This is the BEA's controversial new data that lowers the growth rate
when compared to the old fixed-weight real GDP.

If you don't mind waiting a couple of months, I plan to set up an FTP site
and provide the GDP data there for free, along with other frequently used
data- I'll post another message when it's ready. Thanks again for your
interest.

David Carrier
NERDS
434 E. Diamond Ave.
Gaithersburg, MD 20877
carrierd@aol.com

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

From: Stephen Nash <snash@osf1.gmu.edu>
Date: Mon, 22 Jan 1996 12:33:07 -0500 (EST)
Subject: New Book on Linear and Nonlinear Programming

We would like to announce the publication of our new book:
Linear and Nonlinear Programming
Stephen G. Nash and Ariela Sofer
McGraw-Hill (1996)
ISBN: 0-07-046065-5
We have set up a web-page with information about the book:
http://bass.gmu.edu/site/ORE/books/Nash-Sofer.html

Stephen Nash (snash@gmu.edu)
Ariela Sofer (asofer@gmu.edu)

Operations Research and Engineering Department
George Mason University
Fairfax, VA 22030
(703) 993-1678

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

From: Iain Duff <I.Duff@letterbox.rl.ac.uk>
Date: Mon, 22 Jan 96 16:37:45 GMT
Subject: Workshop on Parallel Numerical Algorithms

Announcing a Euro-Par'96 Workshop
Parallel Numerical Algorithms
LIP, ENS Lyon, France. August 27-29, 1996.

Program Commitee:
Iain Duff, RAL, England and CERFACS, France , Chair
Dirk Roose, K.U.Leuven, Belgium, Vice-Chair
Marian Vajtersic, Slovak Academy, Slovakia, Vice-Chair
Jean Roman, LaBRI, Bordeaux, France, Local Chair

Deadline: February 4, 1996 (paper); February 18, 1996 (electronic)

Euro-Par is THE annual European conference in Parallel Processing. It merges
the former PARLE and CONPAR-VAPP conferences, two major events of the field.
The 1996 version will consist of a large panel of highly focused workshops on
all aspects of parallel processing, from theory to practice, from academy to
industry. They are expected to present the latest advances in their respective
domains. They will be introduced by several high level tutorials of general
interest. More than 20 workshops have been launched in parallel, ranging from
1/2 to 2 days. All accepted papers will appear in the proceedings published by
Springer-Verlag in the LNCS Series.

Workshop #07:
Parallel Numerical Algorithms

The basis for most computational work in science and engineering are robust and
efficient implementations of numerical algorithms. It is the aim of this
workshop to study the design, development, and implementation of such
algorithms on parallel architectures.

Naturally there is much work and software for the solution of linear systems,
both dense and sparse, but we are interested to attract papers and discussion
in other areas of linear algebra (for example, eigenvalue and SVD
calculations) and in other areas of parallel numerical computation, for example
solving nonlinear systems, differential and integral equations, quadrature,
mathematical transformations (e.g. FFT, wavelet transforms), and
number-theoretical methods. We welcome contributions that reflect
actual parallel computational models.

Topics of interest include:

Dense and sparse systems of algebraic equations
Dense and sparse eigenproblems
Matrix operations
Mathematical transformations (e.g. FFT, wavelet transforms)
SVD computations
Optimization techniques
Number-theoretical computing
Differential equations

Further information is available at URL http://www.ens-lyon.fr/LIP/europar96.
It includes the list of all workshops. Please send all information requests
and comments to europar96@lip.ens-lyon.fr. Register today on the Euro-Par'96
mailing list by sending us a mail! See below for additional information.

Valerie Roger, Euro-Par'96 Secretary
LIP, ENS Lyon, 46, Allee d'Italie, F-69364 Lyon Cedex 07, France
Phone: (~33)72728037; Fax: (~33)72728080;
Internet: europar96@lip.ens-lyon.fr

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

From: Alfio Quarteroni <alfqua@ipmma1.mate.polimi.it>
Date: Wed, 24 Jan 1996 10:12:09 +0100
Subject: School and Conference in Trieste on Numerical PDEs

School On Numerical Simulation Of Partial Differential Equations:
Methods, Algorithms, Applications

Miramare - Trieste, Italy
9 - 27 September 1996

The International Centre for Theoretical Physics, with the co-sponsorship
of the European Commission, will organize a School on Numerical Simulation
of Partial Differential Equations: Methods, Algorithms, Applications from 9
to 27 September 1996. It will be directed by A. Quarteroni (Politecnico di
Milano, Italy), R. Duran (Universidad de Buenos Aires, Argentina), and A.
Valli (Universita di Trento, Italy).
A variety of modern numerical methods, such as finite elements, finite
volumes and spectral methods will be introduced and analysed. Their
implementation by efficient algebraic solvers, like conjugate gradient or
multigrid iterations with preconditioning, will be discussed. Applications
to the classical, linear elliptic, parabolic and hyperbolic equations will
be illustrated, as well as to some selected problems arising from the
analysis of fluids and structures. Examples of implementation will be
carried out in a laboratory session with the help of a research software
library. Lectures will be complemented by tutorial activity (laboratory
sessions).
For the programme of the School see: http://www.crs4.it/HTML/trieste.html

The third week of the School will be devoted to a
'Conference on Recent Advances in Numerical Methods for Partial
Differential Equations'.

The lectures will cover several innovative contributions to the field of
numerical solution of elliptic, parabolic and hyperbolic partial
differential equations. Among others, the following topics will be
addressed: shock capturing methods for conservation laws, numerical
modelling of turbulent flows, adaptivity and a-posteriori error control,
wavelets and post-processing, domain decomposition methods for parallel
computing .

Invited speakers will include (titles indicated when available): A. Bellen
(Waveform Relaxation - Multigrid methods for semidiscretized parabolic
equations), F. Brezzi (Spurious eigenvalues for fluid mechanics and
electromagnetic problems), C. Canuto (Application of Wavelets to the
Numerical Solution of Partial Differential Equations), B. Cockburn, J.
Douglas, M. Dryja (Additive Schwarz methods for the mortar finite element
elliptic problems), I. Duff, R. Duran (Numerical approximation of parabolic
problems with blowing up solutions), M. Falcone, L. Franca (Applications of
the Residual-Free Bubble Finite Element Method), L. Gastaldi (Domain
decomposition methods for advection-diffusion equations), S. Jaffard
(Wavelet methods in evolution equations), J.A. Kuznestov ,P. Le Tallec
(Domain Decomposition Techniques for Solving Multimodel Problems), Y.
Maday, D. Marini (Domain decomposition methods with non-matching grids), O.
Pironneau (Algorithms for Parallel Computing for Navier-Stokes equations),
R. Rannacher (Adaptive Error Control in Finite Element Methods), A. Russo
(Finite Element Methods and Residual-Free Bubbles), J.E. Santos (Numerical
Simulation of Waves in Dispersive media), Rolf Stenberg (Stabilized Finite
Element Methods for Plates and Shells),P. Sweby (On The Importance of the
and a posteriori error control for Stefan problems).

International Centre for Theoretical Physics
School on Numerical Simulation of Partial Differential Equations: Methods,
Algorithms, Applications
P.O. Box 586
I - 34100 Trieste
Italy

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

From: Yuan Ya-xiang <yyx@indigo12.cc.ac.cn>
Date: Wed, 24 Jan 1996 11:22:19 +0000
Subject: Nonlinear Programming Conference in Beijing

International Conference on Nonlinear Programming
Beijing, Sept 2-5, 1996

The conference highlights recent advances in the theoretical,
computational and practical aspects of nonlinear programming. The
aim of the conference is to gather nonlinear programming experts
to exchanges ideas and to discuss future developments and trends
of nonlinear programming. In part, this meeting is intended to honor
the many contributions of Professor M.J.D. Powell to Optimization.

INVITED SPEAKERS: (Not all confirmed)

R. Byrd (University of Colorado at Boulder, USA)
T. Coleman (Cornell University, USA)
A.R. Conn (IBM Watson Research Center, USA)
W.C. Davidon (Haverford College, USA)
J.Y. Han (Chinese Academy of Sciences, CHINA)
J. More (Argonne National Lab, USA)
L. Nazareth (Washington State University, USA)
J. Nocedal (Northwestern University, USA)
M.J.D. Powell (University of Cambridge, UK)
R.B. Schnabel (University of Colorado at Boulder, USA)
K. Tanabe (Institute Statistical Mathematics, JAPAN)
R. Tapia (Rice University, USA)
Ph. Toint (University of Namur, BELGIUM)
H. Wolkowitz (University of Waterloo, CANADA)
M.H. Wright (AT&T Bell Laboratories, USA)
Y. Yuan (Chinese Academy of Sciences, CHINA)

ORGANIZING COMMITTEE:

A.R. Conn (IBM Watson Research Center, USA)
J. Nocedal (Northwestern University, USA)
Ph. Toint (University of Namur, BELGIUM)
Y. Yuan (Chinese Academy of Sciences, CHINA)

Chinese National Science Foundation,
Institute of Computational Mathematics and Scientific/Engineering Computing

Up to date information of the conference can be obtained by WWW:
http://lsec.cc.ac.cn/home2/yyx/www/conf96.html

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

From: Bruce Kellogg <kellogg@ipst.umd.edu>
Date: Fri, 26 Jan 1996 15:29:26 -0500 (EST)
Subject: Babuska Conference on Finite Element Methods

BABUSKA CONFERENCE ON FINITE ELEMENT METHODS
March 21 - March 24, 1996

This conference was previously announced on NA-NET. There
will be 19 invited speakers, as well as the opportunity
for a limited number of contributed talks.

We would appreciate advanced notice of attendance at the
conference. There will be a $20 registration fee; checks should be made out to the University of Maryland. Please contact: Alice Ashton IPST University of Maryland College Park, Md 20742 301-405-4889 voice 301-314-9363 fax ashton@ipst.umd.edu For further information, see the conference web page at: http://www.glue.umd.edu/~kellogg/ivo/ ------------------------------ From: Venkat Sastry <SASTRY@rmcs.cranfield.ac.uk> Date: Fri, 26 Jan 1996 17:55:40 GMT Subject: Short Courses at Royal Miltary College Shrivehmam The following short courses are available during 1996. Additional information can be obtained from the address below: NUMERICAL/COMPUTING COURSES for scientists, engineers technical managers and software developers Learn to compute fast and effective solutions from leading researchers. The one week short courses may also be included in a part-time MSc degree. EXPERT SYSTEMS (19-23 Feb) - Introduction to expert system architectures, Knowledge acquisition and representation rechniques, Inference mechanism Practicals using CLIPS and fuzzy tool box. FINITE & BOUNDARY ELEMENTS (4-8 Mar) - Theory and applications including solving stress and crack problems, electro painting. NEURAL NETWORKS (11-15 Mar) - Typical architectures and learning algorithms, Networks for recognition of patterns and speech, even with degraded data. Practicals using Neural Net Toolbox (Matlab) and Neural Professional II NUMERICAL PDE's (25-29 Mar) - Finite difference, finite volume, grid genertion, advance methods for shock solution, CFD. COMPUTER GRAPHICS (15-19 Apr) - solid modelling and visualisation; Graphics libraries PHIGS, (open)GL; Case studies using Iris explorer, PV-WAVE. CAD Packages. MATHEMATICAL MODELLING (29 Apr-3 May) - Stability, asymptotics, conservation and case studies. GUIs for FORTRAN PROGRAMS (10 - 11 June) - front-end GUIs for suites of FORTRAN programs using Tcl/Tk. The Royal Military College of Science is set in open countryside with its own 11 hole golf course. Easy access via the M4, M40 or railway to Swindon. Contact: Department of Applied Mathematics and Operational Research Cranfield University, RMCS Shrivenham, Swindon, Wilts. SN6 8LA Tel: 01793 785316 Fax:01793 784196 e-mail: amor@rmcs.cran.ac.uk ------------------------------ From: Jack Dongarra <dongarra@cs.utk.edu> Date: Sat, 27 Jan 1996 13:17:55 -0500 Subject: Workshop on Parallel BLAS and Related Interfaces We are organizing a workshop on February 19th and 20th, 1996 in Nashville to begin a forum for parallel and sparse BLAS. This will be a follow-up to the discussions that took place on November 13th and 14th when 52 people attended a Workshop on Parallel BLAS and Related Interfaces at the University of Tennessee, and the Birds Of a Feather session that took place at the San Diego Supercomputer 95 meeting in December. The existing BLAS have proven to be very effective in assisting portable, efficient software for sequential and some of the current class of high-performance computers. We would like to investigate the possibility of extending the currently accepted standards to provide greater coverage of sparse matrices and provide additional facilities for parallel computing. The purposes of this meeting is to review the findings of the Workshop in November and the BOF at the Supercomputer 95 conference; and to begin a forum to develop calling sequences for a set of low-level computational kernels for the parallel and sequential settings, as well as standards for basic matrix data structures, both dense and sparse. For background information see: http://www.netlib.org/utk/papers/sblas-meeting.html. For additional information see: http://www.netlib.org/utk/papers/blast-forum.html. Workshop Format: We will start the meeting at 12:00 noon on Monday, February 19th and plan to finish around 4:00 pm on Tuesday, February 20th. There will be a$75.00 registration fee, payable at the meeting,
to cover the meeting room, reception, and refreshments during the breaks.
Included in the registration fee will be a lunch on Tuesday.

If you are planning to attend please let me know (dongarra@cs.utk.edu)
as I will be sending a follow-up message with further details.

Jack

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

From: Lawrence Cowsar <cowsar@research.att.com>
Date: Fri, 26 Jan 1996 07:20:59 EST
Subject: Position at AT&T Bell Laboratories

The Scientific Computing Research Department at AT&T Bell Laboratories
(Murray Hill, New Jersey) anticipates an opening to be filled in 1996.
The department is seeking candidates displaying innovative research in
the area of numerical methods for partial differential equations or in
the intersection of the areas of partial differential equations and
optimization such as optimal design or parameter estimation.
Applicants should have a broad understanding of both analysis and
computer science as well as some experience in realistic scientific
computation. Preference will be given to candidates who anticipate
receiving their Ph.D. by the spring of 1996 or have received their
Ph.D. within the last 3 years.

The members of the Scientific Computing Research Department do research
in the core areas of numerical analysis, as well as computational
geometry, simulation, modeling languages, and Internet technologies.
The department is part of the Information Sciences Research Division
which includes groups working on a broad research program in the
computer and mathematical sciences as well as applications. Further
background information can be obtained via the following URL:

http://netlib.att.com/csrc/sci-comp.html

The interested applicant should submit a curriculum vitae including a
description of their computational experience to:

Lawrence Cowsar (cowsar@research.att.com)
AT&T Bell Laboratories, Rm. 2C-464
600 Mountain Avenue
Murray Hill, NJ 07974-0636.

Electronic submission in Postscript, TeX, or LaTeX is preferred.

All inquiries should be directed to the same address.

AT&T is an equal opportunity employer.

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

From: Herman te Riele <Herman.te.Riele@cwi.nl>
Date: Fri, 26 Jan 1996 18:15:28 +0100
Subject: Position at CWI Amsterdam/Utrecht University

Position at CWI Amsterdam/Utrecht University

Since 1993, the Department of Numerical Mathematics of the
Centre for Mathematics and Computer Science (CWI) Amsterdam,
and the Mathematical Institute of Utrecht University are
cooperating in research of parallel numerical algorithms
for finding eigenvalues and -vectors of large sparse matrices
coming from plasma- and astrophysics.

In this project a position is vacant for a

JUNIOR RESEARCHER

His/her task will be to develop and analyze parallel preconditioning
techniques for iterative methods for solving large systems of equations
and eigenproblems which arise (a.o.) in plasma- and astrophysics.
These techniques will be based on suitable chosen domain decomposition
methods for linear systems of equations. For practical implementation
The project is intended to be concluded after four years with a
Doctoral Thesis and an Academic Promotion. It will be carried out
partly at CWI Amsterdam and partly at Utrecht University.

Candidates, not older than 26 years, have finished their Master's Degree
with specialization in numerical mathematics. Experience with parallel
numerical algorithms and implementation is an advantage.

Appointment will be in temporary service with NWO, the Netherlands
Organization of Scientific Research, for at most four years.
Salary will be Dfl. 2.078 in the first year, rising to Dfl. 3.710
in the fourth year. Women and candidates coming from one of the member
states of the European Union are particularly asked to apply.

Herman te Riele at CWI Amsterdam
(tel. +31 20 5924106, email herman@cwi.nl) or
Gerard Sleijpen at Utrecht University
(tel. +31 30 2531732, email sleijpen@math.ruu.nl).

Applicants are requested to send their application letter with curriculum
vitae, and two letters of recommendation (arranged to be sent) within
three weeks after this announcement to the Personnel Department of
CWI, Kruislaan 413, 1098 SJ Amsterdam, The Netherlands.

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

From: Ken Jackson <krj@cs.toronto.edu>
Date: Tue, 23 Jan 1996 09:40:57 -0500
Subject: Postdoctoral Position at the University of Toronto

The Numerical Computing/Analysis Group in the Computer Science
Department at the University of Toronto has an opening for a
Postdoctoral Fellow. We are quite flexible about the starting date.
We would prefer the candidate to begin in the summer or early fall of
1996, but we would consider other times as well. Our main area of
interest is differential equations, particularly initial and boundary
value problems for ODEs and related problems such as DAEs and DDEs.

The "official" ad follows. Although it states a deadline of 31 January
1996 for applications, we will consider applications until at least
29 February 1996 for this position.

The Department of Computer Science, University of Toronto, has received
funding from various granting agencies. Funding permitting,
Postdoctoral Fellowships are available in all areas of Computer
Science.

Applications, including a curriculum vitae and a list of publications,
should be sent to: Professor Wayne H. Enright, Chairman, Department of
Computer Science, University of Toronto, Toronto, Ontario, M5S 3G4,
is January 31st, 1996.

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

From: Barry Smith <bsmith@mcs.anl.gov>
Date: Thu, 25 Jan 1996 13:52:35 -0600
Subject: Summer Graduate Student Positions at Argonne

Givens Research Associates
Mathematics and Computer Science Division
Argonne National Laboratory

Argonne National Laboratory invites graduate students to apply
for positions as 1996 Givens Research Associates in the Mathematics
and Computer Science (MCS) Division. Four to six appointments will be
made, for 10 to 12 summer weeks, and may be renewed for a second
summer.

These positions are intended to encourage graduate students who are
beginning careers in numerical analysis or computational mathematics.
Candidates must be in an accredited Ph.D. program in mathematics,
applied mathematics, computer science, or a related field. The
Associates will participate in the research activities of the MCS
Division, which has strong programs in software tools and in numerical
methods for linear algebra, automatic differentiation, optimization,
and partial differential equations.

The Mathematics and Computer Science Division supports an advanced
scientific workstations, a scientific visualization laboratory, and
state-of-the-art parallel computers (including a 128-node IBM SP).
http://www.mcs.anl.gov/index.html or contact Jan Griffin at
griffin@mcs.anl.gov. Applicants should send or e-mail their current
resume, and the names and addresses of three persons willing to write
letters of reference to Dr. Deon Ettinger, Division of Educational
Programs, Argonne National Laboratory, 9700 S. Cass Avenue, Argonne,
IL 60439, (708-252-4272; ettinger@dep.anl.gov). This material should be
received at Argonne by February 20, 1996.

During the appointment period, participants receive a stipend of
$540.00 to$580.00 per week (depending on experience level) and a
housing allowance*. Transportation expenses are reimbursed for one
round-trip between the Laboratory and the participant's home or
university for round-trip distances greater than 100 miles.

* Stipends and housing allowance are considered taxable income by the
IRS. The Laboratory is required by law to withhold tax on this
income.

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

From: AOH Axelsson <axelsson@sci.kun.nl>
Date: Mon, 22 Jan 1996 17:25:22 +0100
Subject: Contents, Numerical Linear Algebra with Applications

CONTENTS
Numerical Linear Algebra with Applications
Volume 2, Issue 5, 1995
Special issue dedicated to David M. Young, Jr.

The RSCG Algorithm on Distributed Memory Architectures
L.A. Freitag and J.M. Ortega

A Modified Direct Preconditioner for Indefinite Symmetric
Toeplitz Systems:
P. Saylor and P. Concus

Some Remarks on the Barrier Lemma and K-monotonicity
I. Marek

Completely Parallelizable Preconditioning Methods
I. Gustafsson and G. Lindskog

On a Generalized Conjugate Gradient Orthogonal Residual Method
O. Axelsson and M. Makarov

CONTENTS
Numerical Linear Algebra with Applications
Volume 2, Issue 6, 1995

Multilevel Preconditioners for Discretizations of the
Biharmonic Equation by Rectangular Finite Elements
P.Oswald

Implicit Cholesky Algorithms for Singular Values and Vectors of
Triangular Matrices
K.V. Fernando and B.N. Parlett

A Saxpy Formulation for Plane Rotations
T.J. Ypma

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

From: Ake Bjorck <akbjo@math.liu.se>
Date: Wed, 24 Jan 1996 11:09:01 +0100
Subject: Contents, BIT

CONTENTS BIT Volume 36, No. 1 (1996)
ISSN 0006-3835

Comparing stability properties of three methods in DAEs or
ODEs with invariants
Anders Barrlund, pp. 1--18

Stabilized multistep methods for index 2 Euler-Lagrange DAEs
C. Ar\'evalo, C. F\"uhrer and G. S\"oderlind, pp. 1--13

An algorithm and a stability theory for downdating the ULV decomposition
J. L. Barlow, P. A. Yoon, and H. Zha, pp. 14--40

A graph isomorphism algorithm using pseudoinverses
J. M. Bennet and J. J. Edwards, pp. 41--53

Preconditioning waveform relaxation iterations for differential systems
K. Burrage, Z. Jackiewicz, S. P. N\o rsett, and R. A. Renaut, pp. 54--76

Some linear stability results for iterative schemes for implicit
Runge-Kutta methods
G. J. Cooper, pp. 77--85

An incomplete factorization preconditioning method based on modification of
element matrices
I. Gustafsson, pp. 86--100
A preconditioner for constrained and weighted least squares problems
with Toeplitz structure
X.-Q. Jin, pp. 101--109

Higher-order quadratures for circulant preconditioned Wiener-Hopf equations
F.-R. Lin and M. K. Ng, 110--121
Enhancing energy conserving methods
S. Reich, pp. 122--134

On the extreme spectral properties of Toeplitz matrices generated by $L^1$
functions with several minima/maxima
S. Serra, pp. 135--142

Further results on convergence and stability of a generalization of the
Richardson extrapolation process
A. Sidi, pp. 143--157

On the construction of arbitrary order schemes for the many dimensional wave
equation
J. Tuomela, pp. 158--165

Accurate downdating of a modified Gram-Schmidt QR decomposition
K. Yoo and H. Park, pp. 166--181

SCIENTIFIC NOTES

A note on the Le Verrier-Fadeev's method
N. Kjurkchiev, 182--186

New Books p. 187

Contributions in LaTeX are preferred. Information for subscription and style
files are available from the Editor or from the WWW server for BIT at:
http://math.liu.se/BIT/

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

Date: Wed, 24 Jan 1996 09:57:05 -0500 (EST)
Subject: Contents, Approximation Theory

In Memoriam

In Memoriam Arun Kumar Varma (1934--1994)
1--11

Regular Articles

Satya Deo
On projective dimension of spline modules
12--30

V. E. Maiorov and G. W. Wasilkowski
Probabilistic and average linear widths in $L_\infty$-norm with respect
to $r$-fold Wiener measure
31--40

T. N. T. Goodman, K. G. Ivanov, and A. Sharma
Hermite interpolation in the roots of unity
41--60

Jos\'e A. Adell, F. Germ\'an Bad\'ia, Jes\'us de la Cal and Fernando Plo
On the property of monotonic convergence for beta operators
61--73

Hong-Tae Shim and Hans Volkmer
On the Gibbs phenomenon for wavelet expansions
74--95

Antonio J. Duran and Pedro Lopez-Rodriguez
Orthogonal matrix polynomials: zeros and Blumenthal's theorem
96--118

Ward Cheney and Alphonse Magnus
JATBits
119--122

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

From: Baltzer Science Publishers <publish@baltzer.nl>
Date: Fri, 26 Jan 1996 10:41:18 +0100
Subject: Contents, Advances in Computational Mathematics

CONTENTS:
Advances in Computational Mathematics, Volume 4, 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 4, No. IV, 1995

pp. 293-330; F. Plantevin, Wavelets on irregular meshes

pp. 331-356; B.N. Khoromskij and S. Proessdorf, Multilevel preconditioning
on the refined interface and optimal boundary solvers for the Laplace
equation

pp. 357-388; G. Plonka, K. Selig and M. Tasche, On the construction of
wavelets on a bounded interval

pp. 389-396; H. Wendland, Piecewise polynomial, positive definite and
compactly supported radial functions of minimal degree

Submissions of articles and proposals for special issues are to be

John C. Mason
School of Computing and Mathematics, University of Huddersfield,
Queensgate, Huddersfield, HD1 3DH, United Kingdom
E-mail: j.c.mason@hud.ac.uk

or

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