NA Digest Monday, June 12, 1995 Volume 95 : Issue 24

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: Cleve Moler <>
Date: Mon Jun 12 09:25:57 EDT 1995
Subject: Accidental Posting to the NA Digest Mailing List

A few days ago, a posting about an integral equation with a singular
kernel was accidentally made to the NA Digest mailing list. We have
now plugged the hole in the NA Net software that allowed such
unmoderated postings.

-- Cleve


From: Angela Kunoth <>
Date: Mon, 5 Jun 1995 12:35:12 -0500
Subject: Computing Refinable Integrals

Software: Computing Refinable Integrals

Applications: * Evaluating refinable functions in 1D-3D
* Computing inner products in 1D-3D

Furthermore, the code can be used to

* Compute entries of stiffness matrices
in Wavelet-Petrov-Galerkin schemes in 1D-3D

Based on the theory developed in [W. Dahmen, C.A. Micchelli, Using
the refinement equation for evaluating integrals of wavelets, Siam
J. Numer. Anal. 30, 1993, 507-537], I have written a C++-program
that computes function values and derivatives of (possibly different)
refinable functions (i.e., generators of multiresolution analyses)
and integrals of refinable functions for up to four factors in the
integral in one and two dimensions. In three dimensions, integrals
with at most three factors can be computed. The routines calculate
the desired values exactly up to round-off while avoiding any
quadrature rules by using the refinement equations for the computations.
As input data, only the mask coefficients of the refinable functions
and some parameters like the dimension of the underlying domain and
the number of refinable functions are needed.

A comprehensive documentation of the program in latex-format called

Computing Refinable Integrals --- Documentation of the Program
--- Version 1.1 ---

which summarizes the theory and contains several examples (Daubechies'
generators, B- and Box Splines) can be obtained via anonymous ftp from
or alternatively, inn.dvi.Z, inn.dvi.gz.

The documentation contains directions how to obtain the program.

In a future version, I intend to include a special treatment of the
tensor product case. Also I hope to modify the program soon such that
four factors in integrals in three dimensions can be computed.

Angela Kunoth


From: Steve Zitney <>
Date: Mon, 5 Jun 95 17:47:57 CDT
Subject: Looking for Bordered Block Diagonal Ordering Software

We are looking for computer software to order unsymmetric sparse
matrix systems into Bordered Block Diagonal (BoBD) form. We are
interested in general equation ordering rather than element
ordering for finite-element problems.

Thanks for your help,

Stephen E. Zitney
Cray Research, Inc.
655-E Lone Oak Drive
Eagan, MN 55121


From: Simon Chamlian <>
Date: Fri, 9 Jun 95 09:07:44 EDT
Subject: Generalized Eigenvalue Problems


Is there a "good" reference (book) that treats in great care and
detail the generalized eigenvalues problem (A.x=lambda B.x) and the
difficulties involved with it. Namely,

1) What happens if the matrices A and B are symmetric but not
positive definite (the physics of the problem does not allow
you to get positive definite). What is the numerical stability
of the LR decomposition ( inv(L)*A*inv(U) U.x = lambda U.x,
where inv(L) denotes the inverse of L).

2) What is the numerical stability of the Cholesky decomposition
(inv(L)*A*inv(Lt) Lt.x = lambda Lt.x, where Lt denotes L
transpose) when A and B are positive definite. Is it possible
to get complex eigenvalues, du to numerical instability, with
symmetric positive definite matrices?

3) How to shrink the problem of 1 dimension in order to remove
the null space of dimension 1. In other words, if it exists a
vector y of dimension n such that A.y=0 and B.y=0, how to find
a matrix V (n by m with m = n-1) so that we can express the
problem as: Vt.A.V.z = lambda Vt.B.V.z , with Vt.A.V full
rank (z is the new eigenvector with dimension m).

4) etc.

Thanks in advance.



From: Peter Pacheco <>
Date: Tue, 6 Jun 1995 02:36:43 -0700
Subject: Broadcast of Parallel Linear Algebra Conference

NSF-CBMS Regional Conference
Numerical Linear Algebra on Parallel Processors
University of San Francisco
June 12-16, 1995

The NSF-CBMS Conference, Numerical Linear Algebra on Parallel Processors
will be "broadcast live" on the internet by the San Diego Supercomputer
Center. Information on accessing the broadcasts can be obtained from
SDSC-TV at the URL

Information on the conference schedule can be obtained from the University
of San Francisco at the URL

If you have questions about the conference, please contact Peter Pacheco,
Department of Mathematics, University of San Francisco, San Francisco,
CA 94117, 415-666-6630,


From: Daniel Hershkowitz <>
Date: Mon, 5 Jun 1995 17:03:04 +0300 (EET DST)
Subject: News from ILAS Information Center


We are happy to announce the establishment of a mirror site for IIC
in Chemnitz, Germany.

The URL address of this new mirror site is

The database in Chemnitz gets updated on a daily basis.

Daniel Hershkowitz e-mail:
Mathematics Department URL:
Technion Tel: 972-4-294282 (office)
Haifa 32000 972-4-244626 (home)
Israel 972-4-324654 (fax)


From: Seppo Seikkala <>
Date: Wed, 07 Jun 1995 10:00:04 -0400
Subject: Integral Methods in Science and Engineering

4th International Conference
Integral Methods in Science and Engineering
Oulu, Finland
June 17-20, 1996

Invited Speakers:
D.L.Colton (University of Delaware, USA)
L.Gaul (University of Stuttgart, Germany)
R.Kannan (University of Texas, USA)
A.Klarbring (Linkoping University, Sweden)
N.Morozov (University of St.Petersburg, Russia)
O.A.Ladyzhenskaya (Steklov Mathematical Institute, Russia)
G.Vainikko (Technical University of Helsinki, Finland),
O.C.Zienkiewicz, (University College of Swansea, UK)

Integral equations
Ordinary and partial differential equations
Finite element methods
Conservation laws
Hybrid approaches
Vortex methods
Other integral methods in science and engineering.

Call for Papers: Authors of contributed papers are requested to submit,
before February 15, 1996, an abstract containing the topic of the talk,
a summary (not exceeding 300 words), and the full institutional address
including the telephone and fax numbers and e-mail address.

Division of Mathematics
Faculty of Technology
University of Oulu
90570 Oulu, Finland
tel.:358 81 553 2656
fax: 358 81 553 2664;

The Second Announcement will be sent to all prospective participants
who notify their interest before November 1, 1995.


From: Rob Corless <>
Date: Thu, 8 Jun 1995 08:55:30 -0800 (PDT)
Subject: Potential ISSAC in Hawaii, 1997

What follows is a short questionnaire trying to assess the level
of interest in having ISSAC '97 (ISSAC is the main meeting for
symbolic computation) in Hawaii. We wish to include numerical
analysts in the consideration of this meeting, and hence I am
posting the questionnaire here also. Please take the time to

Thank you,

Rob Corless


We (Bruce Char and Rob Corless) are considering putting in a bid in to hold
ISSAC 97 in Hawaii. There is little doubt in our minds that most
potential attendees would find Hawaii a pleasant venue in
which to hold a meeting. Our preliminary investigations into local
arrangements and travel makes us believe it will be a satisfactory
venue from those points of view as well.

Our current thinking has centered on Kihei Maui, on the shore of
southeastern Maui, a few miles from the Maui Supercomputing Center.
(See the web page for more information
on the Supercomputing Center, including pictures of Maui.) We have a
quote from the Maui Intercontinental Hotel of approximately $100US
per night per room, single or double. There are accommodation alternatives
(e.g. rental of condominiums that house four or six) in the area of
Kihei as well. Living expenses are otherwise comparable to the
continental USA. There are regularly scheduled commercial flights
direct from the US mainland to Maui; ground transfer form the Maui
airport to Kihei would be approximately 30 minutes via van.

Air fare information at this early date is sketchy and vague; the
current "21 day advance reservation" price is approximately $1100 US
return from Philadelphia to Maui; from Vancouver, with a block booking
of 100 seats, the price is $499 (Canadian) return to Honolulu. Return
airfare between Honolulu and Maui is approximately $150US. Qantas
flies from Sydney to Honolulu for $1295 (Australian). The fact that
Canadian prices appear to be the cheapest above may reflect the fact
that a Canadian travel agent did most of the price search; cheaper
flights may be available through your local agent. (For example last
year during a sale, the return airfare between Philadelphia and
Honolulu was approximately $700). The agent was unable to give
quotes for flights from Tokyo or from Europe, which would clearly be
useful to know.

Expressions of support have been received for our bid both by the Maui
Supercomputing Center and the Soft Warehouse (creators of Derive).
They would assist us in making local arrangements. The Supercomputing
Center has extensive prior experience in setting up computing meetings
at the Intercontinental Hotel.

There is also the possibility of running another meeting (Stanly
Steinberg's Applied Computer Algebra meeting) `back to back' with
ISSAC, to optimize the use of air transportation and clout with hotel
bookings. We would appreciate suggestions for other meetings
that we try to could co-schedule with ISSAC '97.

We are also strongly interested in encouraging people who think
of themselves as numerical analysts to come to this meeting. Some
ISSAC meetings in the past have had a numerical flavour; with the
Maui High Performance Computer Center supporting the meeting, it
seems quite reasonable to expect that this meeting will have a
similar sub-theme of numerical/symbolic interaction.

Please take the time to respond to the following questionnaire.

1. If ISSAC '97 was held on one of the Hawaiian islands, would you
be interested in attending?

2. Would you be able to come?

3. Block-bookings or charter flights may significantly lower the
cost of airfare. This requires planning ahead. Would you be
able to do so?

4. The usual date for ISSAC is in the (northern hemisphere) summer.
The climate of Hawaii is uniformly pleasant, and a summer conference
there would provide a welcome change for those living in hot, muggy
areas in (for example) the U.S. and Canada. However, several people
have indicated informally that they would like to have an ISSAC meeting
in Hawaii during the (northern hemisphere) winter, say January or
February. How do you feel about that?

5. Would you classify yourself as a numerical analyst or a computer
algebraist (for want of a better name for people interested in
symbolic computation) or both or neither?

6. Please give us any other comments you have.

Send all responses to or

Thank you for your attention.

Bruce Char and Rob Corless


From: Bruce Char <>
Date: Sun, 04 Jun 95 15:12:16 EDT
Subject: Position at Drexel University


Applications are invited for a tenure-track assistant professor or visiting
position beginning September 1, 1995 or later. Applicants must possess a
doctoral degree in computer science or equivalent and show a strong record
and commitment to teaching and research. We are especially interested in
applicants working in parallel processing, software engineering, computer
networking, or scientific computation.

Drexel University is a private, nonsectarian urban university, founded in
1891. Total enrollment is approximately 10,000 students, including 5200
full-time day undergraduates. A pioneer in cooperative education since
1919, Drexel operates one of the largest co-operative education programs in
the nation. It is in downtown Philadelphia (University City), conveniently
located in the midst of Philadelphia's academic, cultural, and historical
resources. Its proximity to Philadelphia's main rail and highway links
makes the US East Coast from Washington to Boston easily accessible.

The MCS Department is a broad-based academic unit in the College of Arts
and Sciences, offering undergraduate and graduate degree programs in both
Mathematics and Computer Science. It is among the largest departments at
Drexel, with 30 full-time faculty, 270 undergraduates, and 140 graduate
students. Student societies include ACM, Upsilon Pi Epsilon, Pi Mu Epsilon,
and SIAM student chapters. The department also offers minors in mathematics
and computer science. The undergraduate computer science program is
accredited by the Computing Science Accreditation Board. The departmental
computing environment includes Sun4 servers, Unix workstations, X
terminals, and Macintoshes connected to the Drexel campus network backbone
and the Internet.

Send a resume to Search Committee Chair, Department of Mathematics and
Computer Science, Drexel University, Philadelphia, PA 19104 (phone:
215-895-2668; fax: 215-895-1582; e-mail:
Evaluations will begin July 1st and continue until the position is filled.
Drexel University is an Affirmative Action/Equal Opportunity Employer.


From: Mike Osborne <>
Date: Fri, 9 Jun 1995 10:43:22 +1000
Subject: Position at the Australian National University


The Australian National University and Fujitsu Limited, the world's second
largest computer manufacturer, are collaborating in developing numerical
algorithms and software in parallel vector environment for new generation
parallel supercomputers such as the 350 Gflops VPP500 and the VPP300.

A mathematician with experience in algorithm and software development is
sought to join this project. This is an exciting opportunity to contribute
to the development of software for some of the most powerful supercomputers
of the decade. The appointee will be under the direction of Professor M.R.
Osborne from the School of Mathematical Sciences and Professor R.P. Brent
from the Computer Sciences Laboratory, RSISE, and will be required to work
with other team members in these groups and in the Supercomputer Facility
headed by Dr. R. Gingold. Appointees will be expected to interact closely
with these groups which are also engaged in the development of algorithms
and software for the experimental multi-processor AP1000 supercomputer.

We are seeking candidates with a strong background in computational
methods, experience in parallel and/or vector processing, and demonstrated
skills in developing numerical algorithms and implementing them in
practice. The successful candidate will be required to develop software in
some of the following areas: numerical linear algebra, numerical solution
of ordinary and partial differential equations, mathematical programming,
numerical approximations, signal processing. Experience in software
development is desirable. The ability to complete projects on schedule is
essential. It is important that candidates be able to take up the post at
an early date.

An initial appointment of between one and three years will be negotiated.
Extension of the appointment may be possible, but will depend on continuing
collaborative arrangements with Fujitsu. Persons wishing to take the post
on secondment from other positions are encouraged to apply.

The ANU has assembled one of the most powerful and broadly-based advanced
computing environments to be found in a university. Advanced computing
resources at the ANU include a Fujitsu VP2200 vector processor, a CM5 from
Thinking Machines Corporation and a 128 processor Fujitsu AP1000.

The level of appointment, depending on the candidate's experience and
qualifications, will lie within the following salary ranges:

Research Fellow [Level B] $A42,198 - $A50,111 p.a.
Fellow [Level C} $A51,692 - $A59,605 p.a.

Further information including selection criteria is available from Dr R
Gingold, Supercomputer Facility, phone: (06) 249 3437, Fax (06) 279 8199,
and email, Professor R Brent, email,
or Professor M Osborne, email

Closing Date: 30 June 1995 Ref: SMS 31.5.1

Email addresses and/or fax numbers should be provided for referees so that
appointment procedures can be expedited.

APPLICATIONS addressing the selection criteria should be submitted in
duplicate to The Secretary, The Australian National University, Canberra,
ACT 0200 quoting reference number and including curriculum vitae, list of
publications and names and addresses of at least three referees. Further
information is available from either of the contacts listed above. The
University has a "no smoking" policy effective in all University buildings
and vehicles.


From: SIAM <>
Date: Wed, 07 Jun 95 10:15:37 EST
Subject: Contents, SIAM Computing

SIAM Journal on Computing
Volume 24, Number 4
August 1995

SPARSE Reduces Conjunctively to TALLY
Harry Buhrman, Edith Hemaspaandra, and Luc Longpre

Size--Depth Tradeoffs for Algebraic Formulas
Nader H. Bshouty, Richard Cleve, and Wayne Eberly

Learning Arithmetic Read-Once Formulas
Nader H. Bshouty, Thomas R. Hancock, and Lisa Hellerstein

Amortized Communication Complexity
Tomas Feder, Eyal Kushilevitz, Moni Naor, and Noam Nisan

An Optimal Execution Time Estimate of Static Versus Dynamic Allocation in
Multiprocessor Systems
Hakan Lennerstad and Lars Lundberg

Computing the Degree of Determinants Via Combinatorial Relaxation
Kazuo Murota

Scheduling Tasks with AND/OR Precedence Constraints
Donald W. Gillies and Jane W.-S. Liu

Work-Preserving Speed-Up of Parallel Matrix Computations
Victor Y. Pan and Franco P. Preparata

Integer Linear Programs and Local Search for Max-Cut
Svatopluk Poljak

Easily Checked Generalized Self-Reducibility
Lane A. Hemaspaandra and Riccardo Silvestri

Approximating the Minimum Equivalent Digraph
Samir Khuller, Balaji Raghavachari, and Neal Young

Fixed-Parameter Tractability and Completeness I: Basic Results
Rod G. Downey and Michael R. Fellows


From: SIAM <>
Date: Wed, 07 Jun 95 16:30:18 EST
Subject: Contents, SIAM Applied Mathematics

SIAM Journal on Applied Mathematics
AUGUST 1995 Volume 55, Number 4

A Study of Singularity Formation in the Kelvin-Helmholtz Instability with
Surface Tension
Michael Siegel

Computations of Steep Gravity Waves by a Refinement of Davies-Tulin's
Jean-Marc Vanden-Broeck and Touvia Miloh

Fluid-Solid Interaction: Acoustic Scattering by a Smooth Elastic Obstacle
C. J. Luke and P. A. Martin

Nonlinear Oscillations in a Resonant Gas Column: An Initial-Boundary-Value
Meng Wang and David R. Kassoy

Low or High Peclet Number Flow Past a Prolate Spheroid in a Saturated Porous
Louis A. Romero

Forced Convection Past a Slender Body in a Saturated Porous Medium
Louis A. Romero

Models of Vortices in Anisotropic Superconductors
Richard A. Klemm

The Accurate Dynamic Modelling of Contaminant Dispersion in Channels
Simon D. Watt and Anthony J. Roberts

Constant Front Speed in Weakly Diffusive Non-Fickian Systems
David A. Edwards

An Exact Solution of Stikker's Nonlinear Heat Equation
Allan R. Willms

The Effect of a Refractory Period on the Power Spectrum of Neuronal Discharge
Joel Franklin and Wyeth Bair

Relationships Between a Potential and Its Scattering Frequencies
George Majda and Musheng Wei

Dynamic Bifurcation in Hamiltonian Systems with One Degree of Freedom
Norman R. Lebovitz and Adriana I. Pesci

Recognizing Convergent Orbits of Discrete Dynamical Systems
Stanley Ocken


From: Panos Pardalos <>
Date: Sat, 10 Jun 1995 15:22:41 -0400
Subject: Contents, Global Optimization

Table of Contents
Vol.6, No.4 (June 1995)



Editors: P.M. Pardalos and P.D. Panagiotopoulos


Quasidifferentiability in Nonsmooth, Nonconvex Mechanics 327-345

A. ROHDE and G.E. STAVROULAKIS/ Path-Following Energy Optimiza-
tion in Unilateral Contact problems 347-365

DANIEL GOELEVEN and MICHEL THERA/ Semicoercive Variational
Hemivariational Inequalities 367-381

Z. NANIEWICS/ On Variational Aspects of Some Nonconvex Nonsmooth
Global Optimization Problem 383-400

Solution of Hemivariational Inequalities by Nonsmooth Optimiza-
tion Methods 401-425

On the Numerical Treatment of Nonconvex Energy Problems of
Mechanics 427-448


End of NA Digest