NA Digest Sunday, February 13, 1994 Volume 94 : Issue 7

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

Submissions for NA Digest:

Mail to

Information about NA-NET:

Mail to


From: Arne Raeithel <>
Date: Sun, 6 Feb 1994 19:35:16 +0100 (MET)
Subject: A Meta-question

What kind of "digest" is this, I keep asking myself the latter half
of my exposition to >NA Digest< (now over six months). I used to read
"Reader's Digest" back in the fifties; my aunt Lore kept admonishing
me that there is no better way to learn English.

But that publication had questions AND answers, not very convincing ones,
to be sure, but well-written, at least.

In contrast, the na.digest is nothing but a medium to post problems, with
solution attempts going on behind the scenes. Maybe my discipline (psycho-
logy) gravely misleads me in the expectation that it could be otherwise.

Or, I missed a statement of policy by the list-owners saying that
"digest" does not mean to take in nourishing news, rather to ask for them
to be delivered through private two-way channels.

Arne Raeithel
Dept of Psychology
University of Hamburg, Germany.

[Note from Editor: It is certainly true that NA Digest is not the same
kind of publication as Reader's Digest. In the e-mail/network world,
a "digest" is merely an edited collection of contributions obtained from
other sources. In the case of NA Digest, we just collect the items
mailed to the input queue, We certainly
have no policy against "nourishing" news. -- Cleve]


From: Bill Briggs <>
Date: Mon, 7 Feb 94 07:53:46 -0700
Subject: Diffusion Tomography

Is anyone familiar with recent work (<2 years) on the diffusion tomography
or infrared imaging problem? Specifically, Grunbaum have written
on the isotropic problem. Is there any work or efficient algorithms for
the non-isotropic (2-d or 3-d) problem?

Thanks, Bill Briggs


From: Daniel Zwillinger <>
Date: Tue, 8 Feb 1994 11:24:00 -0500
Subject: Three-dimensional Packing Problem

In our work for the Department of Transportation we have been
simulating packed luggage (for the purpose of testing programs that
look at X-ray images for weapons). One step in each simulation is
packing a collection of rectangular parallelepipeds (the objects to be
packed) into another rectangular parallelepiped (the suitcase).

Note that this is not a bin-packing problem, and that each rectangular
parallelepiped can be rotated into any orientation. We believe that
we are aware of the relevant literature, and have found no articles
that directly address this type of three-dimensional packing problem.

Has anyone else considered this problem? We have a working C code
that we would be eager to compare against other codes performing the
same task.

Daniel Zwillinger email:
63 Greylock Road phone: 617/244-5267
Newton, MA 02160 fax: 617/244-5267


From: Jalel Rejeb <>
Date: Wed, 9 Feb 94 16:22:07 EST
Subject: A Diagonal Matrix Updating Problem

Consider the following matrix inverse problem:

(I + ACD) X = B


A, and C are nonsingular NxN matrices.
D is a diagonal matrix.
B, and X are Nx1 matrices.
And I is the NxN identity matrix.

We need to solve for the unknown vector X.

However, the above matrix equation appears in iterative scheme where at
each new iterative step, only the elements of matrix D are
changed (updated).
Is there a more economical procedure other than taking the inverse
of (I +ACD) at each step of the iteration since only D is changed?
ALL suggestions are appreciated!


Jalel Rejeb
Northeast Parallel Architectures Center


From: Richard Petti <petti@NEWTON.Macsyma.COM>
Date: Wed, 9 Feb 1994 13:36-0500
Subject: PDEase Finite Element Software


Macsyma Inc. is now shipping the first of a new class of software products
for finite element analysis. The new product, PDEase/2 for PC DOS, offers
flexibility, simplicity of input, a lot of automated numerical expertise,
graphics, and a low price.

User-Defined Equations

The user can specify the equations which PDEase solves. The package can solve
a wide range of static and time-dependent systems which combine models of
solid mechanics, heat transfer, reaction/diffusion, fluid mechanics, and
electro/magnetostatics. The program comes with over 50 demonstration problems
covering these disciplines. The current version solves mixed elliptic /
parabolic systems and some hyperbolic systems in 2 1/2 dimensions (meaning
that a third dimension can be present, but the field variables cannot depend
on the third dimension).

Automated Expertise

Users need not specify grid information. The software automatically generates
the finite element grid, refines the model to fit error tolerances (default
tolerances or those set by the user), chooses the appropriate nonlinear
equation solvers to balance solution speed with stability, and adaptively
selects time steps in time-dependent problems. While many options can be
selected by users, the package goes to great lengths to adapt to problems
without human guidance.


PDEase creates 2D and 3D graphics under DOS. You can optionally create 3D
graphics under Windows using Macsyma's plot file format and the Macsyma
Graphics Viewer.

Works with Macsyma for Symbolic-Numerical Analysis

PDEase can be used in conjunction with Macsyma, the general-purpose
symbolic-numerical mathematics package. Macsyma can generate and simplify
complicated systems of symbolic equations in any smooth coordinate system.
PDEase can read equations generated by Macsyma. Macsyma Inc. is building into
Macsyma the ability to generate entire batch input files for PDEase.


PDEase is priced under $500 in the U.S. and Canada. Academic and quantity
discounts are available.

System Requirements

PDEase runs on personal computers with 386/387 or greater processors running
DOS, with 4 megabytes of RAM.

A version of PDEase for SPARC workstations running SunOS will ship in March.

For more information contact:

Macsyma Inc.
20 Academy Street
Arlington, MA 02174
tel: 617-646-4550
fax: 627-646-3161

PDEase is a trademark of SPDE, Inc.
Macsyma is a registered trademark of Macsyma Inc.


From: Grigorii V. Gadiyak <>
Date: Mon, 7 Feb 1994 10:33:41 +0700
Subject: Simulation of Devices and Technologies

JULY 5-7, 1994
- Simulation of semiconductor devices
- Simulation of processes in thin insulator films in strong electric
- Simulation of technologies
- Simulation of Carbon Clusters
- Numerical methods for simulation in micro electronics
We suppose that 60-80 scientists will take part in the Seminar
and 5-10 invited lectures will given.
Institute of Computational Technologies of Siberian Branch of
the Russian Academy of Sciences
Institute of Physics and Power Engineering
Institute of Mathematical Modeling of Russian Academy of
Institute of Nuclear Power Engineering

The second Announcement will provide you with a more detailed
information. Please let us know also your telex and the phone number
and e-mail in your letter. We would be very grateful if you inform us
about your decision concerning the participation in the Seminar at
your earliest convenience.
All the correspondence should be sent to the V.P.Ginkin.
Addresses for correspondence:
Bondarenko Sq.1.,Obninsk, Kaluga Region,249020, Russia
Telex: 412509 URAN SU, e-mail:
or to Prof.G.Gadiyak Ac.Lavrentjev Ave. 6, Novosibirsk, 630090,
telex 133122 SOVET SU, FAX: 383-2-35-12-42, E-mail


From: Ivan Dimov <>
Date: Mon, 07 Feb 94 17:50:02 BG
Subject: Conference on Numerical Methods in Bulgaria

Second Announcement and Call for Papers
3rd International Conference on
Numerical Methods and Applications
NM&A - O(h^3)
August 21-26, 1994, Sofia, Bulgaria

Organizer: Bulgarian Academy of Sciences
Co-organizers: University of Sofia, Technical University - Sofia

The conference aims at providing a forum for presentations and
discussions of recent research in numerical methods and their applications.
Participants from all over the world are expected. The subject of the
conference will range from basic research to applications in physics,
mechanics, engeneering, environmental sciences and other areas, including but
not limited within the following topics:

- Finite Difference and Finite Element Methods
- Boundary Element Method and other Dimension Reduction Methods
- Numerical Methods of Approximation Theory
- Monte Carlo Methods
- Preconditioning Methods
- Parallel Algorithms
- Applications of Numerical Methods

Chairman: Bl. Sendov
A.A. Abramov, O. Axelsson, E. Bontchev, B. Boyanov,
T.F. Chan, I. Dimov, D.J. Evans, M. Kaschiev, R.D. Lazarov, S. Margenov,
G.A. Mikhailov, St. Radev, K.K. Sabelfeld, V. Thomee, M. Vajtersic,
P.S. Vassilevski, L. Xanthis

KEY LECTURERS (having confirmed participation):
A.A. Abramov, O. Axelsson, N.S. Bakhvalov, T.F. Chan, A. Donchev,
D.J. Evans, R. Ewing, R.D. Lazarov, S. McKee, C.A. Micchelli,
G.A. Mikhailov, G.V. Milovanovic, W. Proskurowski, K.K. Sabelfeld,
A.A. Samarskij, M. Sapagovas, O. Spaniol, V. Thomee, R. Vaccaro,
M. Vajtersic, Z. Zlatev, L. Xanthis

The authors should present 3 copies of a camera ready full
length paper (8 pages) by surface mail and the corresponding LaTeX file
by e-mail (or on a diskette) by April 20, 1994.


Submission of registration forms February 20, 1994
Submission of papers April 20, 1994
Preregistration May 31, 1994

The third announcement will provide further information about
accommodation, social events, etc.

Chairman: I. Dimov
Secretary: A. Karaivanova
T. Dimov, St. Dimova, St. Fidanova, K. Georgiev, R. Georgiev,
T. Gjurov, O. Iliev, N. Kolkovska, S. Margenov, P. Marinov, O. Tonev,
E. Varbanova, L. Zikatanov

The mailing address of the Organizing Committee:
Acad. G. Bonchev str., bl. 25A
BG - 1113 Sofia, Bulgagia
fax: (+359 2) 70 72 73


From: Heinz W. Engl <>
Date: Tue, 08 Feb 1994 09:57:49 EST
Subject: European Consortium for Mathematics in Industry


The European Consortium for Mathematics in Industry (ECMI)
holds its 8th International Conference at the University of
Kaiserslautern, Germany, between September 6 and 10, 1994
(in cooperation with the German Mathematical Union, DMV).

The objective of the conference is to provide a forum for
the presentation of work on the applications of mathematics
to industrial problems. It will provide a forum for
academics and industrialists to meet and discuss
mathematical problems of mutual interest. It should also
give advanced students a first-hand impression of the
challenges and opportunities for mathematicians in European

The scientific program will focus mainly on seven topics,
namely mathematical methods in
- aerospace industry
- automotive industry
- chemical industry
- chip production
- construction industry
- finance
- metallurgic processes.

The core of each topic is a special session consisting of
one invited lecture and four selected lectures; contributed
talks related to these topics will be placed on the
afternoon following the respective special session.
Contributed papers and minisymposia in other areas of
industrial mathematics are also welcome.

If interested, please request registration material from the
chairman of the organizing committee

Prof.Helmut Neunzert
Fachbereich Mathematik
POBox 3049
D-67653 Kaiserslautern, Germany

The deadline for the submission of abstracts is April 10.


From: Ilse Ipsen <>
Date: Tue, 8 Feb 94 14:38:42 EST
Subject: Workshop on Markov Chains


1995 International Workshop on the

North Carolina State University
Raleigh, North Carolina, USA.
January 16-18, 1995

The aims of the workshop are twofold. To foster cooperation among
researchers and practitioners working on diverse aspects of the
numerical solution of Markov chains; and to provide an opportunity
for researchers to present their latest results. The collection of
presentations intends to be an authoritative overview of the field,
including its developments, current status and projections for future
directions. With this in mind, the program will consist of both
invited and contributed papers. The workshop proceedings will be

Matrix Generation Techniques
Computation of Stationary Probability Vectors:
Direct Solution Methods
Iterative Solution Methods
Recursive Solution Methods (Incl. those of Neuts)
Domain Decomposition Methods
Incomplete Factorizations
Computation of Transient Solutions:
ODE Solvers
Krylov Subspace Methods
Very Large State Spaces
Sensitivity Analysis
Stochastic Petri Nets
Stochastic Automata Networks
Markov Reward Models
Infinite Markov Chains
Parallel and Distributed Implementations
Software Demonstrations

Chairman: William J. Stewart, NCSU
Local Program Ilse Ipsen, NCSU
Committee: Carl Meyer, NCSU
Erol Gelenbe, Duke
Kishor Trivedi, Duke
Sandy Stidham, UNC
Bob Plemmons, WFU

G. Fishman U. of North Carolina, USA
G. Golub Stanford University, USA
W.K. Grassmann U. of Saskatchewan, Canada
U. Krieger Deutsche Bundespost Telekom, Germany
G. Latouche U. Libre de Bruxelles, Belgium
D. Mitra AT&T, Bell Labs, USA
R. Muntz UCLA, USA
M.F. Neuts U. of Arizona, USA
B. Plateau IMAG, France
Y. Saad U. of Minnesota, USA
P. Schweitzer U. of Rochester, USA
P. Semal Philips Research Labs, Germany
G.W. Stewart U. of Maryland, USA
W. Whitt AT&T Bell Labs, USA

June 1, 1994 Deadline for Paper Submission
October 1, 1994 Notification of Acceptance
December 1, 1994 Final Version Due
January 16-18, 1995 Workshop

Submit five copies of a full paper to W.J. Stewart by
June 1, 1994. Papers should be no longer than 20 double
spaced typewritten pages, including tables and figures.
All correspondance should be addressed to

William J. Stewart
Department of Computer Science
Box 8206
N. Carolina State University
Raleigh, N.C. 27695-8206, USA

Phone: (919) 515-7824
FAX: (919) 515-7896


From: Michael Olesen <>
Date: Fri, 11 Feb 1994 14:20:02 -0600 (CST)
Subject: Workshop on Domain-Based Parallelism and Problem Decomposition

Workshop on Domain-Based Parallelism and Problem Decomposition
Methods in Computational Science and Engineering

Sponsored by University of Minnesota Supercomputer Institute and
Army High Performance Computing Research Center

In cooperation with the SIAM Activity Group on Supercomputing

April 25-26, 1994

On April 25 and 26, 1994, the Supercomputer Institute will host a
Workshop on Domain-Based Parallelism and Problem Decomposition Methods in Computational Science and Engineering, organized in cooperation with the SIAM Activity Group on Supercomputing and the
Army High Performance Computing Research Center. The symposium,
which will be held at the Supercomputer Institute in Minneapolis,
will focus on problem decomposition algorithms in engineering and
physical sciences, and lecturers will take a broad perspective
suitable for an interdisciplinary audience of engineers, physical
scientists, and applied mathematicians.

There will be a poster session for contributed papers on Monday,
April 25. The abstract deadline in March 18.

The registration fee is $150 fro registrations received prior to
March and $175 after March 1.

For more information or to obtain a registration form, contact one
of the organizers, David Keyes of Yale University, Youcef Saad or
Donald Truhlar of the University of Minnesota, or contact the
Symposium Administrator, Michael Olesen, Supercomputer Institute,
1200 Washington Avenue South, Minneapolis, MN 55415, phone: (612)
624-1356, fax: (612) 624-8861, e-mail:

Michael J. Olesen
Research Programs Administrator
University of Minnesota
Supercomputer Institute
1200 Washington Avenue South
Minneapolis, MN 55415

office: (612) 624-1356
fax: (612) 624-8861


From: Omar Ghattas <>
Date: Tue, 08 Feb 94 20:08:04 EST
Subject: Postdoctoral Position at U. of Pittsburgh Medical Center

The Artificial Heart Program, Department of Surgery, University of
Pittsburgh Medical Center seeks an outstanding individual for a
postdoctoral research position in computational fluid dynamics. You
will join an interdisciplinary team of bioengineers and surgeons from
the Artificial Heart Program, and computational scientists from
Carnegie Mellon University, working on computer-based design of a
new-generation artificial heart.

You should have a PhD in engineering or applied mechanics, a strong
background in fluid mechanics, and significant research experience in
CFD. Experience in turbomachinery, viscous incompressible flows,
non-Newtonian fluids, computational optimization, and high performance
computing is desirable, as is familiarity with commercial CFD codes.
Because of the team nature of the project, excellent communication and
interpersonal skills are essential.

If you wish to be considered for this position, please send your
resume and a letter of interest (postscript and latex accepted, ascii
preferred) to as soon as possible.

Omar Ghattas
Computational Mechanics Laboratory
Carnegie Mellon University


From: Lutz H. Tack <>
Date: Wed, 09 Feb 94 14:51:50 MEZ
Subject: Postdoctoral Position in Aachen, Germany


Fellowship to be performed at the RWTH AACHEN within the
European Union (EU) Program "Human Capital and Mobility"
"Numerical and Physical Study of Material Forming Processes"
EU 1991-1994

Close linking of metal-forming technology with fundamental data in the
fields of physical metallurgy and materials science is of crucial
importance in planning and designing processes. The planned project
essentially establishes links of this kind by combining deformation models
(based on the finite element method, FEM) with material models
(mathematical descriptions of developments in the microstructure) to allow
prediction of the local structure of the product during the forming process,
with subsequent design of a material-optimized process by means of
FEM computations with controlled modification of the decisive parameters.

Candidate Profile:
The applicant should have a background in metallurgy and interest in
constitutive modelling. Good experience in Finite Element Methods for
simulation of bulge-forming operations and good knowledge of FORTRAN
are required.
Funding for this position is provided by a grant from the EU, therefore
the candidate must have EU nationality and must not be German.

Research Period:
Six months

Kai Karhausen
Phone: +49/241/805916


From: Kent Budge <>
Date: Mon, 7 Feb 94 10:16:34 -0700
Subject: Position at Sandia national Laboratories

I am looking for prospective candidates to fill a postdoctoral position
at Sandia National Laboratories, Department 1431 (Computational Physics
Research and Development.) Please let me know of any student of whom
you are aware who might be interested and qualified. Feel free to
forward this message to your colleagues.

The successful candidate will share responsibility for the development
of a library of C++ classes useful for a wide variety of numerical
applications, including the solution of partial differential equations,
linear algebra, and visualization. The candidate must have a Ph.D. in
computer science, applied mathematics, or the physical sciences. A
strong background in C++ programming is a must. The candidate must be
comfortable working on parallel computers. The candidate should be
able to work well in a team environment and should be comfortable
interacting with industry partners. Good presentation and technical
writing skills are also important. The candidate must be a U.S.

The appointment is for two years.

We wish to fill this position as soon as possible. For more
information, please contact Kent Budge at or by
phone at (505) 844-8244.


From: Kenneth Appel <>
Date: Mon, 7 Feb 94 14:51:16 -0500
Subject: Position at University of New Hampshire

The Department of Mathematics of the University of New Hampshire invites
applications for a tenure track (possibly senior) position in applied
mathematics. The Institute for the Study of Earth, Oceans, and Space at
UNH is a source of many interesting problems in applied mathematics and
(funded) collaborative work is possible. It is possible that there will be
a position in pure mathematics if funding becomes available. Strong
commitment to teaching in both undergraduate and graduate courses is
expected. the position(s) will begin in the fall semester of 1994. Review
of applications will begin on April 1, 1994 but applications will be
considered until the positions are filled.

Resume, e-mail address if possible, and three letters of recommendation should
be sent to Kenneth Appel, Chair, Department of Mathematics, University of New
Hampshire, Durham, N.H., 03824 ( The University of New
Hampshire is an EO/AA employer and encourages applications from women and
minority candidates.


From: Pardalos <>
Date: Wed, 9 Feb 94 14:53:23 EST
Subject: Contents, Global Optimization

Table of Contents
Vol. 4 No. 1 (1994)

I.M. Bomze and G. Danninger: A finite algorithm for solving general quadratic

F. Schoen: On a new stochastic global optimization algorithm based on censored

W. Baritompa: Accelerations for a variety of global optimization methods

H. Konno, T. Kuno and Y. Yajima: Global minimization of a generalized convex
multiplicative function

J.E. Falk and S.W. Palocsay: Image space analysis of generalized fractional

M.Sniedovich, E. Macalalag and S. Findlay: The simplex method as a global
optimizer: A C-programming perspective

Software Section

Book Review
R. Horst and H. Tuy: Global Optimization - Deterministic Methods (H. Benson)


From: Eduardo Sontag <>
Date: Mon, 7 Feb 94 09:24:22 EST
Subject: Contents, Control, Signals, and Systems


Arkadii S. Nemirovskii,
Several NP-hard problems arising in robust stability analysis, pp. 99-105.

J-B. Pomet, R.M. Hirschorn, and W.A. Cebuhar,
Dynamic ouput feedback regulation for a class of nonlinear systems,
pp. 106-124.

Hirosi Sugiyama,
A Sampling Theorem with Equally-Spaced Sampling Points on the Negative Time
Axis, pp. 125-134.

Vincent Blondel and M. Gevers,
The simultaneous stabilizability question of three linear systems is
undecidable, pp. 135-145.

Andrea Baccioti and Paolo Boieri,
Linear stabilizability of planar nonlinear systems, pp. 146-165.

Y. Kannai and G. Weiss,
Approximating signals by fast impulse sampling, pp. 166-179.

Richard Vinter and J. Kotsiopoulus,
Dynamic programming for free time problems with endpoint constraints,
pp. 180-193.


End of NA Digest