From: Ken Driessel <email@example.com>
Date: Tue, 20 Apr 93 15:04:31 -0600
Subject: Looking for Book 'Approximate Methods of Higher Analysis'
To: readers of na.digest --
I would like to buy a copy of the book 'Approximate Methods of Higher
Analysis' by L.V. Kantorovich and V.I. Krylov (translated by C.D.
Benster) published by Interscience in 1958. It is now out of print.
Please contact me if you have a copy to sell.
-- Ken Driessel
From: firstname.lastname@example.org (Marcus Speh)
Date: Mon, 19 Apr 93 10:46:52 +0200
Subject: Re: C++ for Numerical Analysis
>>>>> On Fri, 16 Apr 93 16:20:03 BSC, Walter Figueiredo Mascarenhas
|> If you think that FORTRAN is THE language for numerical analysis and
|> would not waste your time with C++, please let me know why.
FYI: If you have access to the World Wide Web [WWW], you may want to
have a look at the DESY Entry Point to WWW (accessible e.g. from
the HEP Overview Page). Under "Documents provided by Users", you'll
find a link to "C++ for HEP applications". There I have collected
a couple of documents: the summary of the 1992 SIAM workshop
"Scientific Computing with C++" may find your particular interest.
The WWW address for the C++ Page is:
Accessible e.g. from the DESY Home Page at
If you dont know what WWW is, but you have got a telnet server, try
this brings you to the WWW Overview Page at CERN (simple line mode [*])
from where you reach the links "HEP" -> "DESY" -> "C++".
There, you will also find preliminary information on work in
progress on a C++ Class Library for MultiGrid Programming which I hope
to finish by this summer. In my own field, large-scale computations in
lattice field theory, there have been similar approaches based on
C (the CANOPY tool set) and APESE (language on the dedicated APE-100)
The only only example where I have directly compared C++, C and FORTRAN
are Multigrid applications. Here, C++ is conceptually clearly superior to
FORTRAN [see the summary report mentioned above, and related papers by U.
Ruede, available via anonymous FTP from casper.cs.yale.edu as
As for speed, you have to pay a price for some additional
structures, both for storage and CPU (e.g. virtual functions);
depending on this choice of structures (but also on your environment:
C++ compiler or C++ Cfront, scalar/vector/parallel machine...) I have
heard estimates of 5-20% overhead (which is quite a gap). Since
optimization is usually the last step or program development, I
cannot give reliable numbers yet.
As for stability, though there is no ANSI standard yet, C++ has
become a quasi-standard OOP tool (which is often criticised) which I
have found to be sufficiently stable in practical work (same is true
for the differences between compilers -- on fast machines, I have
used a CRAY Cfront and GNU g++ so far). Based on work in the
standardization committee, many a set of "coding rules" has evolved, and
several libraries are already available (though I dont have
experience with any of them). All this information is on the net
(comp.std.c++,comp.lang.c++) or on the World Wide Web [see above].
I'd like to know if you have problems accessing the documents on the
Hope this helps--
[*] A beautiful WWW browser, "XMosaic", is available precompiled for many
machines via anon FTP from ftp.ncsa.uiuc.edu in directory Web/xmosaic.
From: George A. Anastassiou <ANASTASG@hermes.msci.memst.edu>
Date: 20 Apr 93 16:39:14 CDT
Subject: New Book on Probability and Approximation Theory
Pitman Research Notes In Mathematical Series
"Moments in probability and approximation theory"
George A. Anastassiou
The use of probabilistic methods in other mathematical disciplines
has become a trend in recent years, since they produce simple and
elegant proofs usually leading to optimal results. This research
monograph in approximation theory and probability theory falls into
this category. Using methods from geometric moment theory, the
author first solves some very important basic moment problems, and
then develops in parallel the theories of convergence of positive
linear operators to the unit/weak convergence of finite measures to
the Dirac measure, both with rates. The results produced are
quantitative inequalities and most of them are either sharp or
nearly sharp. Many examples connecting the material to other
topics are given.
Readership: Researchers in approximation theory, probability
theory, numerical analysis, statistics, applied analysis, classical
analysis, measure theory, functional analysis, and related fields.
From: John Scales <jscales@dix.Mines.Colorado.EDU>
Date: Thu, 22 Apr 93 09:29:37 -0600
Subject: Free Books/Software
We have decided to give away some lecture notes and software that
we've written over the years. More is in the works and we hope
that others will join us in making their works available for all
to share. None of the authors, the Colorado School of Mines, New
England Research, or anyone else you can imagine makes any prom-
ises or guarantees about anything in these documents/codes.
If you want any of this material, here's what you do. Ftp to the
Internet site hilbert.mines.colorado.edu (184.108.40.206) and log
in as anonymous. Any password will do. Dir doesn't work here so
you need to use ls (or mls if you're coming in from abroad). CD
to pub and you will see, among other things, the following direc-
Directory uga contains a preliminary release of UGA (uni-
processor genetic algorithm) and associated class libraries. The
code was developed on Sun Sparcstations using AT&T's C++ com-
piler. Versions which will compile using the GNU C++ compiler on
various architectures are in preparation.
The lecture notes are in directory papers. Subdirectories in-
What you will see in these directories are compressed postscript
files. Migration contains lecture notes for a graduate course in
seismic imaging. It's about 200 pages long and has roughly 55
Conjugate_gradient contains notes for a short course on the use
of sparse matrix methods for inverse calculations. About 40
pages with no figures--yet.
Theoretical_seismology contains lecture notes for a graduate
course in theoretical seismology. About 120 pages with a few
figures. We expect that many figures will be added in the next
few months or so, at which time we'll put out a revised version.
So, there you have it. If you have any books, lecture notes or
software that you would be willing to distribute freely, drop us
a line and we'll be happy to include your contribution. And
remember, you get what you pay for: please don't send any irate
email asking how you get from equation 8.4.6 to 8.4.7 on page so
and so of some document.
John A. Scales Martin L. Smith
Center for Wave Phenomena New England Research
Deptartment of Geophysics White River Junction
Colorado School of Mines Vermont, 05001
Golden, Colorado 80401 email@example.com
From: Federico Malucelli <firstname.lastname@example.org>
Date: Tue, 20 Apr 93 17:25:36 METDST
Subject: Netflow93: Deadline update
Theory and Practice
October 3-7, 1993
Centro Studi "Cappuccini", Cassa di Risparmio di S. Miniato,
San Miniato (PI) Italy.
Mathematical Programming Society
Hewlett-Packard Italy Ltd.
Italian Operations Research Society
CNR - Progetto Finalizzato Trasporti
Universita' di Pisa
This workshop will emphasize recent theoretical and computational advances in
the field of network optimization, related continuous and combinatorial
optimization problems, implementations and applications. Netflow93 will be
organized in sequential sessions of invited and contributed papers with
sufficient time for informal discussions among participants.
A variety of topics will be accommodated, such as single and multicommodity
network flows, matching, network design, scheduling theory and algorithms,
computational complexity, graph partitioning algorithms, nonlinear network
optimization, combinatorial optimization, large-scale network optimization,
fast approximation algorithms, implementations and computational
experimentation, as well as significant applications.
INTERNATIONAL PROGRAM COMMITTEE
G. Gallo, D. Goldfarb, M.D. Grigoriadis, M. Grotschel, L.G. Khachiyan,
F.T. Leighton, J.K. Lenstra, T.L. Magnanti, J.B. Orlin, W.R. Pulleyblank,
B. Simeone, E. Tardos, R.E. Tarjan, P. Toint.
Giorgio Gallo and Mike Grigoriadis.
The workshop will take place in the Centro Studi "Cappuccini", a former
monastery which has been renovated and converted into a conference center.
It has been made available to Netflow93 by "Cassa di Risparmio di San Miniato".
San Miniato, almost equidistant from Pisa, Firenze and Siena, is a charming
Tuscan medieval town whose origins date back to the tenth century when it was
a fortress. The Center is accessible either from Pisa or from Firenze by train.
Alternatives and detailed directions on how to get there will be provided to
INFORMATION and CORRESPONDENCE
Please send all inquiries, applications for registration and Extended Abstracts to:
Federico Malucelli, Chairman
Local Organizing Committee
c/o Dipartimento di Informatica
Universita' di Pisa
Corso Italia 40, 56125 Pisa, Italy.
Phone: +39 -50 510216
Fax: +39 -50 510226
From: Jinchao Xu <email@example.com>
Date: Thu, 22 Apr 93 16:54:11 -0400
Subject: Conference on Domain Decomposition
2nd Call for Papers and Registration Information
SEVENTH INTERNATIONAL CONFERENCE ON
DOMAIN DECOMPOSITION METHODS IN SCIENTIFIC AND ENGINEERING COMPUTING
Keller Conference Center
Pennsylvania State University
University Park, Pennsylvania
October 27--30, 1993
ABOUT THE CONFERENCE:
Domain decomposition refers to a class numerical methods for obtaining
solutions of scientific and engineering problems by combining solutions
to problems posed on physical subdomains, or, more generally, by
combining solutions to appropriately constructed subproblems. It has
become a subject of intense interest recently because of its
suitability for implementation on high-performance computer
architectures. The aim of this conference, and of similar gatherings
in the series, is to bring together researchers with different
backgrounds working in this active area to discuss recent and
prospective advances and to promote interaction among applied
scientists, computer scientists, and numerical analysts.
FORMAT AND THEMES:
The conference will feature invited lectures, selected contributed
papers and poster presentations. Themes of the conference will range
from basic theoretical research to industrial applications related to
domain decomposition (DD) method, including:
numerical analysis of DD methods
block and substructuring methods
multigrid and multilevel methods
fictitious domain methods
DD methods for high order and spectral methods
DD methods for nonlinear and time dependent problems
DD methods in computational fluid dynamics and structural mechanics
general iterative and preconditioning methods
The conference proceedings will be published by the American
Mathematical Society in the series Contemporary Mathematics. Refereed
papers from both invited and contributed talks may appear in the
Potential contributors should submit an abstract of no more than one
page by May 15, 1993 to the conference secretary at the address below.
The submission of the abstract through e-mail is preferred. Plain TeX,
AMSTeX and LaTeX are all welcome and sample tex macros are available
via anonymous ftp to ftp.math.psu.edu (220.127.116.11) in the
directory pub/ddm7. Plain text (ASCII) files are also acceptable.
Submitted abstracts will be reviewed by the Scientific Committee of
the conference and authors will be notified by July 15, 1993.
Abstracts May 15, 1993
Lodging Reservations Sept. 27, 1993
Early Registration Sept. 27, 1993
The organizing committee, chaired by J. Xu, consists of members of Penn
State and Indiana University-Purdue University at Indianapolis
(IUPUI): D. Arnold, J. Bona, M. Chen, A. Haghighat, J. Shen, S. Tavener,
and J. Xu from Penn State, and R. Chin and A. Ecer from IUPUI.
J. Bramble (Cornell), T. Chan (U.C.L.A.), R. Chin (IUPUI), P. Deuflhard
(Konrad-Zuse-Zentrum, Berlin), R. Glowinski (Univ. of Houston), G. Golub
(Stanford), D. Keyes (Yale), Y. Kuznetsov (Russia Academy of
Science), J. Periaux (GAMNI/SMAI, Paris), A. Quarteroni (Politecnico di
Milano, Italy), O. Widlund (Courant Institute), and J. Xu (Penn State).
R. Bank (UCSD), J. Bramble (Cornell), F. Brezzi (Pavia, Italy),
R. Brown (MIT), T. Chan (UCLA), W. Coughran (Bell Lab), P. Deuflhard
(Konrad-Zuse-Zentrum, Germany), R. Ewing (TAMU), H. Kawarada (Chiba,
Japan), Y. Kuznetsov (Russia Academy of Science), P. Le Tallec (INRIA,
France), J. Mandel (Univ. of Colorado at Denver), Y. Maday (Univ. de
Pierre et Marie Curie, France), T. Mathew (univ. of Wyoming), M. Mu
(Purdue), S. Nepomnyaschikh (Novosibirsk, Russia), P. Oswald (Jena,
Germany), J. Periaux (GAMNI/SMAI, France), R. Scott (Univ. of
Houston), J. Scroggs (NC State), Z. Shi (Academia Sinica, China),
H. Simon (Ames), B. Smith (UCLA), M. Wheeler (Rice), D. P. Young
(Boeing), H. Yserentant (Tuebingen, Germany).
FEES AND REGISTRATION:
The early registration fee (before Sep. 27) is $175 ($50 for students).
The registration fee after Sep. 27 is $200 ($60 for students). The fee
covers the opening reception, instruction, materials, conference
proceedings, the conference banquet held Friday evening at 7:00pm and
break refreshments. It may be paid by check, money order,
VISA/MasterCard, or request to bill employer. To register, contact:
Penn State University
409 Keller Conference Center Tel. 814-863-1744
University Park, PA 16802-1304 Fax. 814-865-3749
Registration will be acknowledged by mail. Refunds will be made for
cancellations received by Sep. 27. After that, the participant or
organization will be held responsible for the fee. Anyone who is
registered but cannot attend may send a substitute. To receive a permit
for on-campus parking, add $12 (nonrefundable) to your fee payment.
TIME AND LOCATION:
Those who arrive early will be able to register at a reception on
Tuesday, October 26 at 6:30pm. Final registration will start at
8:00am on Wednesday, October 27. The conference will begin at 9:00am
Wednesday and will end on Saturday, October 30. The conference will be
held at the Keller Conference Center on Penn State's University Park
Campus in State College, Pennsylvania.
A modest amount of funding is anticipated from National Science
Fundation to support travel for graduate students and postdoctoral
researchers who lack other applicable funding (applications are
especially encouraged from women, minorities, and persons with
disablities). To apply, please have a letter of recommendation by a
faculty adviser sent to Jinchao Xu or the conference secretary
at the address below.
FOR MORE INFORMATION:
About program content, contact: To contribute a paper, contact:
Dr. Jinchao Xu Ms. R. Manning, Conference secretary
Department of Mathematics Department of Mathematics
Penn State University Penn State University
University Park, PA 16802-6403 University Park, PA 16802-6403
Tel. 814-865-1110 Tel. 814-865-7527
Fax. 814-865-3735 Fax. 814-865-3735
e-mail: firstname.lastname@example.org e-mail: email@example.com
More detailed and updated information relevant to the conference (such
as registration form, hotel information) may be obtained via anonymous
ftp to ftp.math.psu.edu (18.104.22.168) in the directory pub/ddm7.
More than 50 abstracts alrealy submitted via email (in TeX or Ascii
files) may also be found in the subdirectary abstracts/.
From: Zhiqiang Cai <firstname.lastname@example.org>
Date: Mon, 19 Apr 93 11:50:00 PDT
Subject: Visiting Position at the University of Southern California
University of Southern California
Department of Mathematics
Applications are invited for a visiting position in Numerical Analysis
and Parallel Computation for the academic year 1993/94 at an Assistant
Professor level, which might lead to a Tenure-Track position in 1994.
Preference will be given to candidates who has experience of parallel
computing and/or whose area is on numerical solution of elliptic or
parabolic partial differential equations.
Applicants should send a current curriculum vita and the names of three
Center for Applied Mathematical Sciences
University of Southern California
1042 W. 36th Place, DRB 155
Los Angels, CA 90089-1113
From: SIAM <email@example.com>
Date: Fri, 23 Apr 93 08:06:30 EST
Subject: Contents, SIAM Computing
SIAM Journal on Computing
AUGUST 1993 VOLUME 22, NUMBER 4
Regular Resolution Versus Unrestricted Resolution
Local Randomness in Polynomial Random Number and Random Function
H. Neiderreiter and C. P. Schnorr
Applying Coding Theory to Sparse Interpolation
A. Dur and J. Grabmeier
Exact Identification of Read-Once Formulas Using Fixed Points of
Sally A. Goldman, Michael J. Kearns, and Robert E. Schapire
Asynchronous Fault-Tolerant Total Ordering Algorithms
Louise E. Moser, P. M. Melliar-Smith, and Vivek Agrawala
A Fibonacci Version of Kraft's Inequality Applied to Discrete
Arthur S. Goldstein and Edward M. Reingold
Counting Circular Arc Intersections
Pankaj K. Agarwal, Marco Pellegrini, and Micha Sharir
Ray Shooting and Parametric Search
Pankaj K. Agarwal and Jiri Matousek
Learning in the Presence of Malicious Errors
Michael Kearns and Ming Li
Small-Bias Probability Spaces: Efficient Constructions and
Joseph Naor and Moni Naor
Deciding Properties of Nonregular Programs
David Harel and Danny Raz
Multiple Communication in Multihop Radio Networks
Reuven Bar-Yehuda, Amos Israeli, and Alon Itai
From: SIAM <firstname.lastname@example.org>
Date: Thu, 22 Apr 93 11:39:50 EST
Subject: Contents, SIAM Numerical Analysis
Table of Contents
SIAM Journal on Numerical Analysis
vol 30-4, Aug. 1993
A Posteriori Error Estimates Based on Hierarchical Bases
Randolph E. Bank and R. Kent Smith
Multiplicative Schwarz Algorithms for Some Nonsymmetric and Indefinite Problems
Xiao-Chuan Cai and Olof B. Widlund
Convergence Analysis of the Schwarz Algorithm and Multilevel Decomposition
Iterative Methods II: Nonselfadjoint and Indefinite Elliptic Problems
A New Mixed Finite Element for the Stokes and the Elasticity Problem
M. Farhloul and M. Fortin
A Fully Discrete Adaptive Nonlinear Chernoff Formula
R. H. Nochetto, M. Paolini, and C. Verdi
Bending-Dominated Deformation of Thin Spherical Shells:
Analysis and Finite-Element Approximation
The Computation of Water Waves Modelled by Nekrasov's Equation
G.A. Chandler and I.G. Graham
Combined Finite Element and Pseudospectral Method for the Two-Dimensional
Evolutionary Navier-Stokes Equations
Guo Ben-yu and Ma He-ping
Rosenbrock Methods for Partial Differential Equations and
Fractional Orders of Convergence
A. Ostermann and M. Roche
Stabilization of Unstable Procedures: The Recursive Projection Method
Gautam M. Shroff and Herbert B. Keller
Computation of Takens-Bogdanov Type Bifurcations with Arbitrary Codimension
On the Computation of Multiple Bifurcations with Multiple
Parameters and Symmetry
Numerical Computation of Saddle-Node Homoclinic Bifurcation Points
Stability of Parallel Explicit ODE Methods
Hon-Wah Tam and Robert D. Skeel
Circulant Preconditioners for Complex Toeplitz Matrices
Raymond H. Chan and Man-Chung Yeung
An $\epsilon$-Free A Posteriori Stopping Rule for Certain
Iterative Regularization Methods
End of NA Digest