## NA Digest Monday, August 6, 2007 Volume 07 : Issue 32

Today's Editor:
Tamara G. Kolda
Sandia National Labs
tgkolda@sandia.gov

### Submissions for NA Digest:

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

### Information via email about NA-NET:

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

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

From: Joseph Traub <traub@cs.columbia.edu>
Date: Sun, 29 Jul 2007 15:38:53 -0400 (EDT)
Subject: Philip Rabinowitz

In addition to "Methods of Numerical Integration" with
Philip Davis, Rabinowitz was the co-author of
"A First Course in Numerical Analysis" with Anthony
Ralston. I think that was one of the finest general
numerical analysis texts of its day. Its some 550 pages
and contains an enormous amount of excellent material.

Joseph Traub

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

From: Gene H Golub <golub@stanford.edu>
Date: Fri, 3 Aug 2007 05:04:28 -0700 (PDT)
Subject: Matrix Problem

My colleague, Rajendra Bhatia, has the following problem.

Let T(a) be an nxn upper triangular matrix with the diagonal elements =
a, and the off-diagonal elements for j > i all equal to - 1. Can you
determine the exact singular values? (Bhatia knows the solu'n for a few
values of a.) We know the asymptotic values for some values of a. Let
Bhatia (rbh@isid.ac.in) and me know of your solution.

Gene Golub

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

From: Evgenii Rudnyi <evgenii@rudnyi.ru>
Date: Sat, 04 Aug 2007 19:12:40 +0200
Subject: Re: Packages vs. languages

>I would be interested in contributions to a discussion of the following
>related topic: the relative merits of commercial CFD/CHT packages (CFX,
>FIDAP, Fluent, Phoenics, etc. etc.) and programming languages (APL, BASIC,
>C++, Fortran, etc. etc.) for teaching computational fluid dynamics and
>heat transfer.

I believe the answer to this question depends on the goal, that is, whom
you are going to teach. However, to get acquainted with modern
commercial software makes sense for everyone who is going to learn
computational fluid dynamics and heat transfer.

Let us start with the biggest audience: people who plan just use CFD in
practice. Here one of commercial packages is the best choice. I am
working for CADFEM and it has great experience of teaching engineers.
See for example ESoCAET - European School of Computer Aided Engineering

http://www.esocaet.com/

Then we can consider people who plan to contribute to the future
development. Here programming should be already an important part of
education. Yet,

1) It makes sense to learn commercial packages anyway, as they should
learn the state of the art in the field. Otherwise, there is a danger
that they reinvent the wheel in the future.

2) They should learn programming first. In my view, learning programming
with CFD simultaneously would be really a bad choice.

3) Here one should probably choose not between programming languages but
rather between available libraries.

Best wishes,

Evgenii Rudnyi

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

From: Mike Sussman <sussmanm@math.pitt.edu>
Date: Mon, 30 Jul 2007 22:44:20 -0400
Subject: Re: Packages vs languages

Graham de Vahl Davis posted a question about the relative merits of
commercial packages (CFX, FIDAP, etc.) and programming languages (C++,
Fortran, etc.) for teaching computational fluid dynamcis and heat
transfer. It seems to me that the response depends heavily on to whom
the course would be directed. Are you teaching theory, are you teaching
practice, or are you teaching applications?

A theoretician ultimately is going to require a programming language in
order to try out new algorithms.

A person interested in the practice of CFD will be interested in
commercial packages and the most successful practical methods, but will
also be interested in writing code, at least to demonstrate why some
methods are not successful. This person may also be interested in
writing code that interfaces with a commercial package.

A person interested in applications will want a good commercial package
with sophisticated front ends (mesh generation, etc.) and back ends
(motion pictures). It is a waste of this person's time to learn a
general-purpose programming language.

And remember, too, that there is a third option: general-purpose
problem-solving environments such as Free-FEM, Deal.II, and Overture
come to mind. These provide for problem descriptions at a much higher
level than programming languages yet are applicable to a wide range of
problems using a wide range of techniques.

Mike Sussman
sussmanm@math.pitt.edu

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

From: trhoffendjr@mmm.com
Date: Mon, 30 Jul 2007 11:24:31 -0500
Subject: Re: Packages vs languages

merits of teaching commercial CFD packages vs teaching how to
program the basic algorithms. I think that it is fair to extend
this question to other types of commercial modeling packages
(e.g. finite elements, ray tracing, etc.).

For engineering and science students, it is my opinion that a
solid core in a programming language, numerical methods, and
scientific computing are preferable over learning to use a specific
package. I do not prefer teaching one language over another. My
opinion is that the important aspect of learning any particular
programming language is "learning to learn" a programming language so
that others may be learned and used confidently in the future. Base
numerical methods are required to understand the elementary building
blocks of larger scale codes and algorithms, for example what it
means for a quantity to be large and small numerically, interpolation,
solving linear and nonlinear equations, and computing eigenvalues. By
scientific computing I mean learning how to combine numerical methods
to build larger scale computer codes to solve a particular problem and
developing a basic understanding of how the program runs on different
computers, memory architectures, in/out of cache, etc. The base
scientific computing step connects to the theory and experiments taught
in the rest of the curriculum (fluid mechanics, mechanics,
electromagnetics, optics, networks, etc.).

My expectation that a new engineer or science graduate should have the
ability to write a program to solve a simple problem and/or to process
and analyze data has turned lately into wishful thinking. I see code
runners that are lost without the base code that they learned to use in
college and/or people that are unable and/or afraid and/or unwilling to
spend the hours necessary to write a code/macro/script to accomplish
something slightly different than what is provided in a commercial code
or to streamline their design/development process.

Ability to learn and use a commercial code is important for design and
development. Knowing how to use a specific code is much less important
to me, especially for a new hiree, than having a good understanding the
the core science/engineering, having the ability to learn something new
and being agressive about writing and using software tools as needed.

Finally, I would like to comment that in my experience, even though many
commercial codes are "closed", communicating with the developer's
technical staff is often an email, online forum post, or phone call away.
The best developers I have worked with have people with a wealth of knowledge
that have helped me understand how their code works and how to accomplish
an API or plugin scheme that goes with their code. They may also have
documentation with enough background information and literature references
to obviate the need for a textbook. So it is often possible to learn
something new and/or review something old by taking advantage of their
support.

Please note that these opinions are my own opinions and not those of 3M
Company.

Thank you and best wishes,
TRH

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

From: "Schoenauer, Willi" <Willi.Schoenauer@iwr.fzk.de>
Date: Fri, 3 Aug 2007 18:10:40 +0200
Subject: Comment to packages versus languages

In NA Digest Vol. 07, Issue 31, Graham De Vahl Davis asked for a
discussion "packages versus languages". Here are my comments:

The situation: For numerical simulations you must bring together TWO
experts: The engineer that has the technical problem, defines the PDEs,
measures the empirical coefficients and then wants to solve the PDEs.
Now a quite different expert is needed, a computer scientist that knows
the numerical methods, is a programming specialist (HE must know the
programming languages) and he is a specialist for parallelization for
supercomputers. This computer science specialist can be replaced to a
certain extent by standard software packages for everyday problems.

We have developed a black-box solver (FDEM) for arbitrary nonlinear
systems of PDEs on unstructured grids that computes with the solution an
error estimate so that we know the quality of the solution. The code is
efficiently parallelized. We offer a service where in a close
cooperation the engineer gives us his PDEs, we solve them and return the
result in the requested form. I am presently addressing university
institutes and research divisions of high-tech manufacturers for
cooperations in the above mentioned sense. The manufacturers enforce the
use of standard software and university institutes that want to
cooperate with them MUST use the same software. So their research can
never go beyond the possibilities of that software. University
institutes usually cannot afford a computer scientist, so they use also
as far as possible standard software. In most cases these people do no
longer know the PDEs they are solving. They enter some coefficients in
some input tables of the software, that's all. I have now addressed 800
institutions and we are now creating cooperations.

My conclusion: If you teach thermodynamics or fluid dynamics the
students should be able to define the PDEs and boundary conditions for
their phenomena. Then they must look for software that gives wit the
result an error estimate, i.e. a measure for the quality of the
numerical result. As up to now standard software does not deliver an
error estimate, they must do mesh refinement tests until the numerical
error is small. Then they compare the numerical result with the
measurements. If the two results do not coincide the error must be in
their model because the numerical error is small. So you must teach how
to use standard software CRITICALLY. As an example: a cooperating
institute calculated the chemically reacting flow in a micro-reactor
with FLUENT on a very coarse grid. Our error estimator gave 200% error
for the reaction product. We needed with FDEM one hour computation time
on 250 processors of a parallel supercomputer to compute the reaction
product with an error estimate in the 1% range!!!

Willi Schoenauer
Forschungszentrum Karlsruhe (FZK)
Institut fuer Wissenschaftliches Rechnen (IWR) Tel. +49(0)7247/82-8625
E-mail: Willi.Schoenauer@iwr.fzk.de

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

From: Walter Gander <gander@inf.ethz.ch>
Date: Wed, 1 Aug 2007 07:42:41 +0200 (CEST)
Subject: Conference Report on 2nd Intl Conf on Matrix Methods & Operator Eqns

Report on
II INTERNATIONAL CONFERENCE ON MATRIX METHODS AND OPERATOR EQUATIONS
http://bach.inm.ras.ru
Moscow July 23-27, 2007
Walter Gander and Gene Golub

Following the ICIAM07 meeting in Zurich, a satellite conference was organized
in Moscow; many participants took the opportunity to visit both places. Thus
there was a good mix of speakers from 16 countries and delegates from
Russian-, Ukrainian-, European- and US; this also included the distinguished
Indian scientist Rajendra Bhatia. It was great that many young researchers
from East and West participated. Remarkably, most young Eastern Europeans
could communicate in excellent English, the official conference language.

Eugene Tyrtyshnikov did an excellent job as local organizer. Most foreign
visitors were located in Hotel Soyuz, a former military institution. The
conference site was in the new building of the famous Steklov Institute of
Mathematics of the Russian Academy of Science. Guri Marchuk, the grand old
man and former director of the academy opened the conference with some warm
welcome words. The social program included a visit to the Kremlin and a
banquet in a magnificent hall of the impressive main building of the Russian

The Scientific Program was organized in 21 plenary talks and 81 talks parallel
in three sessions divided by the topics: "Matrices and Algorithms", "Algebra
and Matrices" and "Matrices and Applications". The spectrum of the talks was
broad it reached from classical topics from orthogonal polynomials and best
approximation to modern development such as quasi-separable matrices, tensor
factorization and approximation by matrices of low tensor rank. Almost all
talks were in English. A presentation was made by Vladimir Khazanov of joint
work with Vera Kublanovskaya, the great Russian numerical analyst and
co-inventor of the QR method for computing eigenvalues.

Although no best paper award was given, the company Neurok Techsoft sponsored
for the first time "Matrix Prizes" for best young researchers. The first prize
went to Ivan Oseledets from the Institute of Numerical Mathematics, RAS,
Moscow for his excellent contribution: "A new class of low-tensor rank matrices
closed under inversion". Three second prizes went to Olga Markova (Lomonosov
Moscow State University) for the paper "Matrix algebras and their length" ,
Federico Poloni (Scuola Normale Superiore, Pisa, Italy) for the paper
"Exploiting rank structure in the solution of a class of nonsymmetric
algebraic Riccati equations", and Dmitry Savostyanov (Institute of Numerical
Mathematics, RAS, Moscow) for the paper "Matrix tricks in tensor
approximation: 3D-cross method and maximum-volume principle".

Hans Schneider closed the conference with a gracious comment thanking Eugene
Tyrtyshnikov with the words that he was very pleased to see how the field of
Linear Algebra continues to flourish and noting that many new applications and
developments arise from Numerical Linear Algebra.

This is the second meeting of this nature. Another meeting is planned two
years hence. We hope we will see many of our colleagues from abroad at this
exciting event.

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

From: Daniel B Szyld <szyld@temple.edu>
Date: Tue, 31 Jul 2007 16:12:23 -0400
Subject: First Annoncement: meeting in NY Sept. 08, in honor of Olof Widlund

First Announcement:
"Fast Algorithms for Scientific Computing",
A Symposium in Honor of Olof B. Widlund on the Occasion of His 70th Birthday,
19-20 September 2008, Courant Institute, New York University, New York.

There will be about two dozen lectures covering a wide area of
numerical analysis, scientific computing, and relevant applications.
There will also be contributed poster presentations.

For a list of invited speakers, hotel information, and (eventually)
details about abstract submissions, see the conference

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

From: Peter Moore <pmoore@smu.edu>
Date: Thu, 02 Aug 2007 14:47:33 -0500
Subject: Ad for position at SMU

Applications are invited for one senior level (tenured) position,
to begin in the fall semester of 2008.
Applicants must have
a Ph.D., provide evidence of outstanding research, and have a strong
commitment to teaching at all levels. Applicants in all areas of
computational mathematics are encouraged. The Department of
Mathematics offers an active doctoral program in computational and applied
mathematics.

To apply, send a letter of application with a curriculum vitae, a list of
publications, research and teaching statements and the names of
three references to: The Faculty Search
Committee, Department of Mathematics, Southern Methodist University,
P.O. Box 750156, Dallas, Texas, 75275-0156.
The Search Committee can be contacted by sending e-mail to
mathsearch@mail.smu.edu. (Tel: (214)768-2452; Fax: (214)768-2355).

To ensure full consideration for the position, the application must be
received by October 1, 2007, but the committee will continue to accept
applications until the position is filled. The committee will notify
applicants of its employment decision after the position is filled.

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

From: Georg Propst <georg.propst@uni-graz.at>
Date: Wed, 1 Aug 2007 04:50:12 -0400
Subject: Faculty Position Applied Mathematics, University of Graz, Austria

Die Karl-Franzens-Universitaet Graz besetzt am Institut fuer Mathematik und
Wissenschaftliches Rechnen der Naturwissenschaftlichen Fakultaet eine

Professur fuer Angewandte Mathematik 2

(unbefristetes Arbeitsverhaeltnis nach dem Angestelltengesetz;
voraussichtlich zu besetzen ab 01.10.2008)

Die/Der Bewerber/in soll das Gebiet Partielle Differentialgleichungen,
Dynamische Systeme und Modellierung, bevorzugt im Bereich der life-sciences,
vertreten. Die/Der erfolgreiche Kandidat/in ist eine international anerkannte
Forscherpersoenlichkeit und kann auf wissenschaftliche Beitraege sowohl im
Bereich der Analysis als auch auf Anwendungen und deren numerische Realisierung
verweisen. Voraussetzung fuer die Bewerbung ist eine fachspezifische Habilita-
tion oder gleich zu wertende wissenschaftliche Leistung. Weiters wird die
Bereitschaft und Befaehigung zur Mitarbeit bei einschlaegigen interdiszipli-
naeren Forschungsprojekten, insbesondere am Spezialforschungsbereich
"Mathematical Optimization and Applications in Biomedical Sciences", sowie zur
selbstaendigen Einwerbung von Drittmitteln vorausgesetzt. Weiters wird
Erfahrung in der fachspezifischen Lehre und Mitwirkung bei der inneruniversi-
taeren Selbstverwaltung erwartet. Zudem erwarten wir Kompetenz im Bereich des
Gender Mainstreaming.
Die Universitaet Graz strebt eine Erhoehung des Frauenanteils insbesondere in
Leitungsfunktionen an und fordert daher qualifizierte Frauen ausdruecklich zur
Bewerbung auf. Bei gleicher Qualifikation werden Frauen vorrangig aufgenommen.

Bewerbungen sind unter Angabe der Kennzahl 22/25/99 ex 2006/07 bis 10.09.2007
einzureichen. Informationen zu den Bewerbungsmodalitaeten und weitere Voraus-
setzungen finden Sie im Mitteilungsblatt (21. Stueck) vom 01.08.2007 unter
http://www.uni-graz.at/zvwww/miblatt.html

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

Date: Thu, 2 Aug 2007 09:39:14 +0100
Subject: Post-doctoral Research Officer, Cranfield University, Shrivenham, UK

There is a vacancy for a post-doctoral research officer in Applied
mathematics and Scientific Computation Group, Engineering Systems
Department. The post is a fixed term contract for one year. The salary
is in the range of GBP 27,665 - GBP30,835.

You will be required to teach mathematics and statistics from very
elementary level to advanced computational mathematics on our
post-graduate courses. You will also be required to contribute to
current short course programme. As a research officer you will be
responsible to enhance the capabilities of Air3D, a compressible flow
solver.

Some of the enhancements include algorithmic performance, development of
graphical user interface and parallelization of the program. You will
also be required to contribute and deliver lectures/tutorials on a range
of topics including computational partial differential equations,
programming in C, high performance and parallel computing.

Please apply online at www.cranfield.ac.uk/hr or alternatively,
application forms and further details may be obtained from the Human
Resources Dept, Cranfield University, DCMT, Shrivenham, Swindon, Wilts,
SN6 8LA Telephone 01793 785758, quoting Ref R/ESD74
Closing date for receipt of applications is 25th August 2007.

Venkat V S S Sastry

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

From: Jiguang Sun <sunjiguang@yahoo.com>
Date: Tue, 31 Jul 2007 08:26:16 -0700 (PDT)
Subject: PostDoc Position in Inverse Scattering Theory / Computation EM at DSU

A postdoc/research associate/visiting professor
position is open in the areas of computational EM
and/or inverse scattering theory at the Applied
Mathematical Research Center, Delaware State
University. Excellent programming skills is a plus.

The position is for a term of one year and can be
renewed. Salary is competitive.
recomendation letters to Prof. Jiguang Sun at
jsun@desu.edu.

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

From: "Ismael Vaz" <aivaz@dps.uminho.pt>
Date: Thu, 2 Aug 2007 15:37:24 +0100
Subject: Pos-Doc position available - Portugal

Please disseminate the following announcement to potential candidates.

A post-doctoral position in the University of Minho, Portugal, is available.
Candidates must have at least 3 years experience after PhD.

Job/Fellowship Reference: C2007-UMINHO-ALGORITMI-04 - Post doctoral position
in Global and Mixed Computational Optimization

Main research field: Mathematics

Sub research field: Computational mathematics

See

for further information or contact emgpf@dps.uminho.pt.

The deadline for application is 30 August 2007.

Yours sincerely,
Edite Fernandes

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

From: Massimo Fornasier <mfornasi@Math.Princeton.EDU>
Date: Mon, 30 Jul 2007 15:04:46 -0400 (EDT)
Subject: Doctoral position at RICAM (Linz, Austria)

Linz, July 27, 2007

Doctoral position at RICAM in mathematical image processing

RICAM, the Johann Radon Institute for Computational and Applied
Mathematics (Austrian Academy of Sciences), is looking for 1 doctoral
student in applied and computational mathematics or computer science
interested in pursuing research in computational harmonic analysis and
partial differential equations for advanced image processing. We emphasize
the interdisciplinary approach which will involve both a variety of
mathematical tools and applications in art restoration. The student will
be a member of the WWTF project "Five Senses - Call 2006", Mathematical
Methods for Image Analysis and Processing in Visual Arts research project
at RICAM, Linz, Austria.

Deadline for applications: September 15, 2007

Starting date of the PhD: November 1

The salary is determined according to the rules of the Austrian Research
Fund FWF
(http://www.fwf.ac.at/de/projects/personalkostensaetze_2007.html), i.e.,
30.860,00 Euro/year brutto.

Work location: RICAM, Linz, Austria.

Applications should be addressed to Dr. Massimo Fornasier
(massimo.fornasier@oeaw.ac.at) and they should consist of:

1. a presentation letter
2. curriculum vitae et studiorum (including grades for each exam)
3. diploma thesis, if already finished, title and abstract otherwise
4. two letters of recommendation (to be addressed to the email above)

The ideal candidate has a Bachelor in Mathematics, Computer Science, or
Information Engineering with courses successfully achieved in numerical
analysis and experience in C/C++ and Matlab.

The time requirements for a PhD thesis is typically in the range of 3
years. The doctoral stud-ies will be focussed (although not limited) on
variational calculus and inverse problems for image reconstruction and
applications in art restoration.
In particular, we are interested in investigating efficient algorithms for
the minimization of functionals promoting sparse recovery. The algorithms
may include randomized components. Studies on random matrices and subspace
correction (domain decomposition) methods will be a crucial issue.
Depending on the particular interests of the doctoral student, related
investigations in inverse problems for parameters detection in PDEs,
compressed sensing, and learning theory can also be pursued.

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

From: Paul Dellar <dellar@maths.ox.ac.uk>
Date: Fri, 3 Aug 2007 08:02:10 -0400
Subject: DPhil studentship, Mathematical Institute, University of Oxford

Applications are invited for an EPSRC-funded project studentship to
work on "Lattice Boltzmann methods for fluids, plasmas, and quantum
systems" with Dr Paul Dellar at the Oxford Centre for Industrial and
Applied Mathematics (OCIAM), part of the Mathematical Institute at the
University of Oxford. A DPhil is Oxford's equivalent of a PhD.

The project will explore applications of the lattice Boltzmann method,
currently used primarily to simulate incompressible fluid flow, to
other physical systems such as Bose-Einstein condensates and
semiconductors.

The studentship is is available for 3.5 years from 1st October 2007. It
includes a stipend, currently GBP 12,600 per annum (tax free), and
payment of University and College fees at the rate applicable to EU
citizens. In principle the studentship is available to applicants of
any nationality, but it does not include payment of the higher fees
charged to non-EU citizens.

The studentship is attached to St Anne's College.

Further details and the application procedure may be found via
http://www.maths.ox.ac.uk/notices/vacancies/

Applications must arrive by lunchtime on 17th August 2007.

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

From: "Melvin Brown" <melvin.brown@smithinst.co.uk>
Date: Mon, 30 Jul 2007 11:14:14 +0100
Subject: Glass Corrosion: Industrial Case Studentship, Imperial College & Pilkington

Glass Corrosion: Parameter estimation in reaction diffusion problems
involving ionic species with limited data

An EPSRC Industrial CASE studentship is available in collaboration with
Pilkington Group Limited and the Mathematics Knowledge Transfer Network in
the UK. The studentship will be based at Imperial College, London in the
Department of Materials and will be supervised by the Head of Department,
Professor Bill Lee and Dr Paul Tangney, who has recently transferred from
Lawrence Berkeley Laboratory. In addition, mathematical support will be
provided by Dr Daniel Lesnic from the Department of Applied Mathematics at
the University of Leeds, who is an acknowledged expert in mathematical
inverse problems.

Many problems in the manufacture and use of glass are associated with the
thermally activated diffusion of ionic species and resultant redox reactions
at the diffusion interface between species with different oxidation states.
This applies to all types of glass, including ancient museum glass of
historical significance, modern float glass and novel glasses used for the
encapsulation and long term storage of nuclear waste. Recent work at
Imperial College has shown that dynamic SIMS (Secondary Ion Mass
Spectrometry) can produce accurate and well defined depth profiles of the
relevant ionic species in glass. These profiles are rich in information and
represent a classic example of a mathematical inverse problem. The aim of
the project is to work back from the depth profiles to elucidate the detail
of the ionic diffusion processes and redox reactions that produced them.

The primary focus of the experimental work will be on the SIMS technique
using state-of-the-art instruments at Imperial College and other
Institutions within Europe, including the University of Warwick and
instrument manufacturers in Germany and France. However, other surface
analysis techniques, such as XPS, SEM, etc. will be used to cross-check and
calibrate the SIMS data. As part of the research the student will have the
opportunity to spend a minimum of 3 months with the industrial sponsor
during the course of their studies, and, as a Mathematics KTN student, there
will be opportunities to attend courses, events and workshops at the
University of Leeds and the Smith Institute.

An individual with a first degree in any of the scientific disciplines with
significant mathematical content and strong interest in mathematical
modelling of physical process will be considered for this opportunity but an
interest in instrumental surface analysis methods is also desirable.

EPSRC CASE studentships are restricted to UK citizens or individuals
normally resident in the UK for the past 3 years. This studentship will be
provided at the standard EPSRC rate of a bursary of £14,600 and home level
fees.

Although the studentship will be available to students who wish to start
later, we are particularly interested in anyone who wishes to commence in or
before October 2007.

Applications will be processed when received, with the expectation that the
post will be filled by 17th August 2007. To apply, please email a copy of
your CV, including the names of 3 referees, to Professor Bill Lee, email
w.e.lee@imperial.ac.uk. Applicants should also complete an on-line
application for a PhD which is available under how to apply at:
www.imperial.ac.uk/pgprospectus. For assistance with the on-line application

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

From: "Do Y. Kwak" <kdy@kaist.ac.kr>
Date: Sun, 29 Jul 2007 19:33:54 -0400
Subject: Contents, Journal of KSIAM

Call for paper and Contents of Journal of KSIAM
http://www.ksiam.org/

J. KSIAM Vol 11, no 2, 2007
1. Comment on: ¡°Hall and Ion-Slip effects on magneto-micropolar
fluid with combined forced and free convection in boundary
layer flow over a horizontal plate¡±
by Mostafa A. A. Mahmoud
2. Coding theorems on a geeneralized information measures
by M.A.K Baig and Rayees Ahmad Dar

3. Signed degree sequences in signed 3-partite graphs
by S. Pirzada and F. A. Dar
4. On some models leading to quasinegative-binomial distribution
by Sheikh Bilal

5. Teaching applied mathematics for engineers- A new teaching paradigm
based on industrial mathematics

Veli-Matti Taavitsainen

6. Fine segmentation using geometric attraction-driven flow and edge-regions
Jooyoung Hahn and Chang-Ock Lee
7. Direction filter bank-based fingerprint image enhancement using ridge
curvature classification
by Joon Jae Lee, Byung Gook Lee, Chul Hyun Park

8. The process of the development of hypoxia in an abnormal blood flow II

by Minkyu Kwak and Jaegwi Go

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

From: Martin Berzins <mb@sci.utah.edu>
Date: Wed, 1 Aug 2007 08:46:36 -0600 (MDT)
Subject: Contents, Applied Numerical Mathematics, Vol. 57, Iss. 10, 2007 (fwd)

Applied Numerical Mathematics
Volume 57, Issue 10, Pages 1081-1162 (October 2007)

1. Editorial Board Page
2. Monotone iterative technique for numerical solutions of fourth-order nonlinear elliptic boundary value problems
Pages 1081-1096
Yuan-Ming Wang
3. Continuous parallel-iterated RKN-type PC methods for nonstiff IVPs
Pages 1097-1107
Nguyen Huu Cong and Nguyen Van Minh
4. Convergence of an adaptive hp finite element strategy in one space dimension
Pages 1108-1124
W. Dörfler and V. Heuveline
5. Mathematical modelling and numerical simulation of a glow-plug
Pages 1125-1144
L. Formaggia, S. Micheletti, R. Sacco and A. Veneziani
6. Validated solutions of initial value problems for parametric ODEs
Pages 1145-1162
Youdong Lin and Mark A. Stadtherr

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

From: Martin Berzins <mb@sci.utah.edu>
Date: Tue, 31 Jul 2007 16:59:09 -0600 (MDT)
Subject: Contents, Applied Numerical Mathematics, Vol. 57, Iss. 9, 2007

Applied Numerical Mathematics
Volume 57, Issue 9, Pages 975-1080 (September 2007)

2. Quasistatic evolution of damage in an elastic body: numerical analysis and computational experiments
Pages 975-988
M. Campo, J.R. Fernández, K.L. Kuttler and M. Shillor
3. Optimal convergence recovery for the Fourier-finite-element approximation of Maxwell's equations in nonsmooth axisymmetric
domains
Pages 989-1007
Boniface Nkemzi
4. Solving aerosol coagulation with size-binning methods
Pages 1008-1020
Edouard Debry and Bruno Sportisse
5. A spectral finite volume transport scheme on the cubed-sphere
Pages 1021-1032
Vani Cheruvu, Ramachandran D. Nair and Henry M. Tufo
6. Enlarging neighborhoods of interior-point algorithms for linear programming via least values of proximity measure functions
Pages 1033-1049
Y.B. Zhao
7. Finite volume evolution Galerkin (FVEG) methods for three-dimensional wave equation system
Pages 1050-1064
M. Lukáÿÿová-Medvid'ová, G. Warnecke and Y. Zahaykah
8. A posteriori error estimators for locally conservative methods of nonlinear elliptic problems
Pages 1065-1080
Kwang Y. Kim

------------------------------
End of NA Digest

**************************
-------