## NA Digest Sunday, July 2, 2006 Volume 06 : Issue 27

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: Jim Demmel <demmel@cs.berkeley.edu>
Date: Mon, 26 Jun 2006 14:44:11 -0700
Subject: Query: Iterative refinement and error estimates for least-squares

An upcoming LAPACK release will include our extended iterative
refinement algorithm for linear systems, and we would like to
extend that algorithm to overdetermined least squares problems.
LAPACK currently does not provide refinement for least squares
problems. We are free to invent an interface. There are many
possible data to return, so we are looking for community input.
What error estimates and returned data are useful to least
squares applications?

We apply our refinement algorithm [1] to Bjorck's formulation as
an augmented linear system [I A; A' 0] [r; x] = [b; 0]. Our
algorithm reduces the forward normwise and componentwise errors
of a linear system to sqrt(n)*eps unless the system is too
ill-conditioned. We either return an accurate answer or warn
that the condition number is too large for this guarantee. In
contrast to linear systems, solving least squares problems
produces *two* vectors of interest, the solution x and the
residual r, doubling the number of potentially useful error
and condition number estimates.

We can provide estimates for any subset of the following forward
errors for the refined x and r:

normwise error: componentwise error:
solution x: norm(x-xtrue)/norm(xtrue) norm((x-xtrue)./xtrue)
residual r: norm(r-rtrue)/norm(rtrue) norm((r-rtrue)./rtrue)
r rel. to b: norm(r-rtrue)/norm(b) norm((r-rtrue)./b)

where the norm used could be the infinity-norm or 2-norm. Each
of these bounds has a corresponding condition number. Which of
these possibilities are most useful to applications and
environments?

Linear systems already require a complicated interface to return
normwise and componentwise condition numbers and error estimates
for x. We would prefer not to flood users with unneeded data or
take time computing unneeded bounds. What do users actually
need?

[1] http://www.netlib.org/lapack/lawnspdf/lawn165.pdf

From:
Jim Demmel <demmel@cs.berkeley.edu>,
Jason Riedy <ejr@cs.berkeley.edu>,
Yozo Hida <yozo@cs.berkeley.edu>,
Xiaoye Li <xsli@lbl.gov>,
David Vu <dtv@cs.berkeley.edu>,
Meganath Vishvanath <vishvana@cs.berkeley.edu>

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

From: Des Higham <aas96106@maths.strath.ac.uk>
Date: Wed, 28 Jun 2006 15:10:38 +0100 (BST)
Subject: SIAM Prize for Peter Kloeden

Congratulations to Peter Kloeden on being announced by SIAM as
the 2006 recipient of the
W. T. and Idalia Reid Prize in Mathematics
http://www.siam.org/meetings/an06/prizes.php

Peter's work covers a wide range of topics in applied analysis,
but NA Digest readers may be most familiar with his
pioneering work on numerical dynamics
P. E. Kloeden and J. Lorenz,
Stable attracting sets in dynamical systems and in their
one step discretizations,
SIAM J. Numer. Anal., 23, 1986, 986--995,
and on numerical methods for stochastic differential equations
P. E. Kloeden and E. Platen,
Numerical Solution of Stochastic Differential Equations,
Springer-Verlag, 1992.

Details of the The W.T. and Idalia Reid Prize in Mathematics Lecture:
Random Attractors and the Preservation of Synchronization
in the Presence of Noise,
taking place at the SIAM Annual Conference, can be found at
http://www.siam.org/meetings/an06/reidprize.php

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

From: "Brenner, Alfred" <abrenner@ida.org>
Date: Thu, 29 Jun 2006 11:09:45 -0400
Subject: Nominations Solicited for the Sidney Fernbach Memorial Award

Nominations Solicited for the Sidney Fernbach Memorial Award

We are seeking nominations for the prestigious Sidney Fernbach Memorial Award,
established in 1992 in memory of Sidney Fernbach, one of the pioneers in the
development and application of high performance computers for the solution of
large computational problems. A certificate and \$2000 are awarded for
outstanding contributions in the application of high performance computers
using innovative approaches. The winner will be announced at the SC06
Conference in Tampa FL. The nomination poster may be viewed at
http://www.interactions.org/sgtw/2006/0621/ieeAwards-2006-1MB.pdf

nomination forms and submission instructions can be found at
http://www.computer.org/awards/ .

Please direct any questions to Sangtae Kim <kim55@purdue.edu>, chair of the
2006 Fernbach Award Committee.

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

From: Steven Leon <sleon@umassd.edu>
Date: Thu, 29 Jun 2006 15:56:57 -0400
Subject: Workshop Report: Algorithms for Modern Massive Data Sets (MMDS)

Workshop Report: Algorithms for Modern Massive Data Sets (MMDS)

The workshop was held at Stanford University, June 21-24, 2006. It
was sponsored by the National Science Foundation, the Stanford
Computer Forum, Yahoo! Research and Ask.com.

The objectives of the workshop were to explore novel techniques for
modeling and analyzing massive, high-dimensional, and
nonlinear-structured data and to bring together computer scientists,
computational and applied mathematicians, statisticians, and
practitioners to promote cross-fertilization of ideas.

The workshop is evidence that MMDS is a popular and growing area of
research activity. The number of applicants for the workshop was far
more than was originally anticipated by the organizing committee and
consequently it was necessary limit enrollments. In total there were
232 participants representing a wide spectrum of the general
scientific community. Altogether there were 45 talks and 24 poster
presentations. It was indeed a massive amount of information to assimilate.

The focus of the first day of the workshop was on linear algebra
basics with an emphasis on sampling methods and matrix
factorizations. During the second day the spotlight was on industrial
applications and sampling methods. Day 3 featured talks on kernel
learning and applications. The final day of the workshop focused on
tensor-based applications. The final evening featured a poster
session and reception/banquet.

There was an excellent mix of both theory and applications. The
theory included talks focusing on large scale numerical linear
algebra, kernel-based nonlinear structure extraction, tensor-based
multilinear structure extraction, missing value estimation, and
sampling-based algorithms. The applications included: analyzing
microarray data and high-throughput chemical data in pharmaceutical
applications; identifying gene products, elucidating protein folding
pathways; detecting and classifying cancer; modeling combinational
structure of large social, computer, and communication networks;
identifying potential terrorist cells in communication networks;
identifying noisy targets and faces in realistic settings; improving
internet search engines; analyzing remote sensing data for
environmental planning, weather forecasting, and public health contamination.
Further details including abstracts and slides are available at

http://www.stanford.edu/group/mmds/

The workshop was extremely well coordinated. The organizers Gene
Golub, Michael Mahoney, Lek-Heng Lim, and Petros Drineas are to be
congratulated for a job done.

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

From: Tim Davis <davis@cise.ufl.edu>
Date: Tue, 27 Jun 2006 16:23:27 -0400
Subject: Horror Matrices

Postal regulations in the United Kingdom prohibit the sending of "Horror
comics and matrices" in regular mail:

http://pe.usps.gov/text/imm/immicl/immicleg_021.html

As a result, this is to announce that singular and extremely
ill-conditioned test matrices in the UF sparse matrix collection will no
longer be available for download to sites in the United Kingdom. Also
prohibited are sparse matrices with random nonzero pattern, unless you
do research in iterative methods. This is because random sparse
matrices experience catastrophic fill-in when factorized with direct
methods, as shown by one well known United Kingdom researcher in sparse
direct methods [1]. Random sparse matrices are truly a horror to direct
methods. One can only surmise that [1] was inspired by Her Majesty's
desire to protect Her Citizens against the Ravages of Horrible Matrices.

[1] Duff, I. S., "On the number of nonzeros added when Gaussian
elimination is performed on sparse random matrices", Mathematics of
Computation, vol 28, 1974, pp. 219-230.

Horror matrices have a long tradition in UK literature, as seen
below in a little-known variant of Lewis Carroll's "Jabberwocky".

% ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Twas fill-in, and the slow transpose
Did dir'ect factor in archive
All tipsy were the banded 'ones
And memory paths outgave.

"Beware the 'OrrorMat, my son!
The cols that peak, the try's that catch!
Beware the Netlib blurb, and shun
the fractal banded mats!

He took is direct slash in hand:
Long time the max cond foe he sought ---
So rested he by the elim' tree
And stood while(1) in thought.

Add, as in out-core work he stood,
The 'OrrorMat, with speyes of NaN,
And core dumped as it ran!

For one! to n! and 'til the end
The direct slash went snicker-snack!
He left it tril, and with diag
He went condesting back.

"And, has thou slain the 'OrrorMat?
Gzip to tars, my direct code!
O factorize day! tril(U)! tril(A)!'
He posted in his ode.

Twas fill-in, and the slow transpose
Did dir'ect factor in archive
All tipsy were the banded 'ones
And memory paths outgave.

% ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

perm = [9 8 1 4 3 2 5 6 7]
author (perm)

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

From: Dugald Duncan <D.B.Duncan@ma.hw.ac.uk>
Date: Fri, 30 Jun 2006 14:37:26 +0100
Subject: Scottish Computational Mathematics Symposium, Sep 14, 2006

15th Scottish Computational Mathematics Symposium (First Announcement)

10:30-17:30 Thursday September 14th, 2006
Heriot-Watt University, Riccarton Campus, Edinburgh, UK

SPEAKERS:

Assyr Abdulle (Basel, moving to Edinburgh)
Lyonnel Boulton (Heriot-Watt)
Chris Budd (Bath)
Mike Giles (Oxford)
Tony Shardlow (Manchester)

DETAILS:

Anyone interested is welcome. The meeting is supported by the London
Mathematical Society, and they have
allocated some funding for travel for UK-based PhD students (apply as

Registration, student funding and other details are at:
http://www.ma.hw.ac.uk/scms

Dugald Duncan

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

From: Charles Augarde <charles.augarde@durham.ac.uk>
Date: Thu, 29 Jun 2006 11:37:11 +0100
Subject: Advances in Continuum Mechanics, Summer School, Sep 10-16, 2006

The nonlinear deformation of solids
EPSRC Mathematics for Engineers Summer School
Durham University, UK
September 10-16 2006

The theme of this school is the nonlinear deformation of solids within
the framework of Continuum Mechanics. The topic is of central importance
to many areas of engineering research ranging from the constitutive
modelling of cell tissue to wave propagation in porous media and the
structural integrity of ageing nuclear reactor pressure vessels.
The motivation for this Summer School comes from the desire to remove
some of the research barriers experienced by engineers by providing high
level core mathematics instruction in advanced continuum mechanics
(specifically, the nonlinear deformation of solids) from leading
academics and back-up support in practical tutorial sessions.
Speakers
Roger Crouch, Jon Trevelyan & Charles Augarde - Engineering, Durham
Brian Straughan & James Blowey - Mathematics, Durham
Tony Spencer - Mathematical Sciences , Nottingham
Alison Ramage - Mathematics, Strathclyde
Mauro Fabrizio - Engineering, Bologna.
Paul Steinmann, - TU Kaiserslautern.
Davide Bigoni - Trento.

The School is primarily aimed at first or second year engineering PhDs
for whom the fees are heavily subsidised. Other attendees are welcome
provided space is available.

Full details are given at this URL
http://www.dur.ac.uk/maths.summerschool06

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

From: Jamila Sam <jamila.sam@epfl.ch>
Date: Thu, 29 Jun 2006 16:56:09 +0200
Subject: Extended deadline, IntCP 2006, Sep 25, 2006

CALL FOR PAPERS
IntCP 2006 workshop
Interval Analysis and Constraint Propagation for Applications
http://liawww.epfl.ch/Events/IntCP2006
Cité des Congrčs, Nantes, France
September 25th, 2006

Held in conjunction with the
Eleventh International Conference on Principles
and Practice of Constraint Programming (CP 2006)
(http://www.sciences.univ-nantes.fr/cp06)

* Important Dates:
10 July 2006 - Submission deadline
27 July 2006 - Notification of acceptance
06 Aug 2006 - Final camera-ready copies
15 Aug 2006 - Early registration deadline<br>
25 Sept 2006 - Workshop day

* Description and goals:
The goal of this workshop is to bring together researchers and
practitioners working on constraint propagation and interval analysis,
providing a forum to present ongoing research and practices and to
discuss techniques for real-world requirements.

* Submissions:
People wishing to give a talk should submit an extended abstract of
at least 2 pages. Submissions must be formatted using LNCS packages
(see CP formatting instructions). The title page should include the

Please, email all submissions in postscript or pdf format to
** intcp-sub@essi.fr ** by July 2nd 2006, specifying the name of
the contact author in the message.

* Contacts:
Send questions about the workshop to : intcp-sub@essi.fr

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

From: RenataGimnez <renata.gimenez@bsc.es>
Date: Wed, 28 Jun 2006 11:29:55 +0200
Subject: Parallel Computing for Large-scale CFD, Barcelona, Oct 24-26, 2006

Parallel Computing for Large-scale CFD
October 24th - 27th, 2006 - Barcelona, Spain

THE COURSE
In this course, the CFD community will be taught the essence of what is
needed for researchers to take advantage of the machines' power, and an
extra effort is spent to hold the researchers' hand in parallelizing,
debugging and optimizing their respective codes.

The scaling of CFD algorithms has not yet matched peak speed, and the
programming burden for parallel machines remains heavy. Hence, the
applications must be programmed to exploit parallelism in the most efficient
way possible. Today, the responsibility for achieving the vision of scalable
parallelism remains in the hands of the application developers.

This course illustrates the state-of-the-art parallel computing and links
theory to applications for large-scale CFD, through demonstrations and
training. This course should be of value to all involved in parallel
computing and is structured to be of equal interest to CFD programmers, and
code developers in general.

THE INSTRUCTORS
Prof. Hasan AKAY (Purdue University at Indianapolis, USA)
Dr. José María CELA (BSC, Spain)
Dr. Guillaume HOUZEAUX (BSC, Spain)
Dr. Jesús LABARTA (BSC, Spain)
Dr. Mariano VÁZQUEZ (BSC, Spain)

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

From: Gaetano Zanghirati <g.zanghirati@unife.it>
Date: Thu, 29 Jun 2006 23:57:49 +0200
Subject: School in Computational Cell Biology, Emilia, Italy, Sep 4-6, 2006

Second International School in Computational Cell Biology - SCCB2006
COMPUTATIONAL METHODS IN MULTISCALE PROCESSES FOR
PROTEIN INTERACTIONS
Department of Pure and Applied Mathematics,
University of Modena and Reggio Emilia, Italy
September 4-6, 2006
All information available at: http://dm.unife.it/SCCB2006

Aims and scope: The school is addressed to students and researchers either
with an applied mathematics, computer science or engineering background who
are interested in cell biology applications or to cellular biologists and
biochemists who wish to gain new quantitative insights. In particular, it will
focus on multiscale processes for protein interactions.

Lecturers
* Kevin Burrage, The University of Queensland, Brisbane, Australia
* Johan Suykens, Katholieke Universiteit, Leuven, Belgium
* Andrew Torda, University of Hamburg, Hamburg, Germany

Topics:
- Systems Biology will be a significant scientific discipline in the twenty
first century. Modelling and simulation will become key components: we will
need to understand the interplay between the deterministic and stochastic,
between the continuous and the discrete, and between small scale and large
scale. This set of lectures attempts to explore these interplays. (Kevin
Burrage)
- The lectures aim at providing a systematic overview of support vector
machines and kernel based methods in supervised, unsupervised and
semi-supervised learning. Applications towards the specific needs for
microarray data analysis will be explained. Different case studies will be
considered. (Johan Suykens)
- Protein force fields range from models based on physics to ones based on
naive statistics and entertaining optimisation procedures. Here we will
discuss how to build, break and apply score functions, when they are fun and
when they are tragic. (Andrew Torda)

Organizing Committee:
Marco Prato, University of Modena and Reggio Emilia, Italy.
Margherita Carletti, University of Urbino, Italy.
Gaetano Zanghirati, University of Ferrara, Italy.

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

From: "Martin H. Gutknecht" <mhg@math.ethz.ch>
Date: Sat, 01 Jul 2006 10:55:06 +0200
Subject: ICIAM 2007: minisymposium proposals solicited by August 2006

The Sixth International Congress on Industrial and Applied Mathematics,
ICIAM 2007, will be held in Zurich, Switzerland from 16-20 July 2007.
It will take place at the two adjacent central campuses of ETH Zurich
and the University of Zurich. The old town is just a few minutes away.

Currently, minisymposia proposals are being solicited:

Deadline for proposals: 31 Aug 2006
Notification of acceptance: 30 Sep 2006

An ICIAM minisymposium proposal requires a title and a short abstract or
description (to be printed in the program) plus a short justification, which
includes a tentative list of speakers. An ICIAM minisymposium consists of
either one session (4 talks of 30 min) or two such sessions. A participant
may organize several minisymposia, but should speak only in one of them. He or
she may also speak in other minisymposia that are organized by other persons.
The proposal has to be submitted electronically via the conference
registration system, which is accessible from

http://www.iciam07.ch/registration

ICIAM discourages minisymposia in which most of the speakers or all the
coauthors come from the same organization. For additional rules and
recommendations on ICIAM minisymposia see:

http://www.sam.math.ethz.ch/~iciam07/RegGuide/node8.html

The Congress will feature
* 27 invited ICIAM speakers
* the AWM/EWM Olga-Taussky Todd Lecture
* the GAMM Prandtl Lecture
* the GAMM von Mises Lecture
* the SIAM von Neumann Lecture
* an Euler Lecture
* a Public Lecture
* hundreds of minisymposia
* contributed papers, poster sessions
* ceremonies for five ICIAM prizes, the ETH/SIAM Peter Henrici Prize,
the GAMM Richard von Mises Prize, and the ANL/NAG Wilkinson Prize
* five Industry Days
* exhibitions
* five Embedded Meetings:
- the Annual GAMM Meeting 2007 (with additional invited speakers)
- a meeting of the African Mathematical Union
- a meeting of the China Society for Industrial and Applied
Mathematics
- a meeting of the Union Matematica de America Latina y el Caribe
- a meeting of the European Society for Mathematical and
Theoretical Biology

* 31 Aug 2006 deadline for minisymposia proposals
* 12 Nov 2006 deadline for abstract submission
* 15 Jan 2007 deadline for early-bird registration

For further information keep visiting http://www.iciam07.ch.

A list of satellite conferences (held shortly before or after ICIAM07)
is compiled under http://www.iciam07.ch/satellite-conferences.

ICIAM 2007 is being organized on behalf of the ICIAM organization
by a joint committee of ETH Zurich and the University of Zurich:

Rolf Jeltsch (Congress Director, ETH Zurich)
Walter Gander / Martin Gutknecht / Petros Koumoutsakos (ETH Zurich)
Erwin Bolthausen / Michel Chipot / Stefan Sauter (University Zurich)

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

From: "T.Terlaky" <terlaky@mcmaster.ca>
Date: Tue, 27 Jun 2006 09:41:23 -0400
Subject: Tenure-track/Tenured Faculty Position at McMaster University

McMaster University
School of Computational Engineering and Science
Department of Computing and Software
Tenure-track/Tenured Faculty Position

The Department of Computing and Software jointly with the School of
Computational Engineering and Science invites applications for a
tenure-track/tenured position in computational engineering and
science. Preference will be given to candidates with experience in
computational design, optimization, or signal processing. Industry/practical
experience and the ability to be, or to become a Professional Engineer in the
Province of Ontario is a definite asset.

The position/rank is open with the expectation of the appointment of an
established high profile researcher at an Associate Professor
level. Exceptionally strong candidates may be considered for being appointed
with tenure and/or at the Full Professor level. It is expected that the
successful candidate will take the position by September 2006.

The McMaster School of Computational Engineering and Science is the first
education and research programs at the Master's and Ph.D. levels. It is an
equal partnership between the Faculties of Engineering and Science, with
outreach to the School of Business and the Faculty of Health Science. The
three major thrusts of the School are: Computational Physical Sciences;
Computational Optimization, Design and Control; and Computational Biosciences.

The Department of Computing and Software has a component of 26 faculty
Software Engineering. The appointment is within the Faculty of Engineering,
one of the most research intensive Faculties of Engineering in Canada.

Applications and enquiries should be forwarded to:
The Chair of the Department of Computing and Software
McMaster University
Information Technology Building
1280 Main Street West

http://www.cas.mcmaster.ca and http://computational.mcmaster.ca
McMaster University is strongly committed to diversity within its community
and especially welcomes applications from visible minority group members,
women, Aboriginal persons, persons with disabilities and others who may
contribute to the further diversification of ideas. All qualified candidates
are encouraged to apply; however, Canadians and permanent residents will be
given priority.

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

From: "Chu Delin" <matchudl@nus.edu.sg>
Date: Fri, 30 Jun 2006 10:31:40 +0800
Subject: Research Fellow Position at National University of Singapore

The Scientific Computing group in Department of Mathematics at National
University of Singapore is seeking to hire a research fellow to conduct
research in the area of (Numerical) Linear algebra. Salary will be
determined based on University's scales, depending on qualifications.

Interested candidates should send a CV and publication list to

Dr Delin Chu
Department of Mathematics
National University of Singapore
2 Science Drive 2
Singapore 117543
Email: matchudl@nus.edu.sg

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

From: Frank Knoben <frank@igpm.rwth-aachen.de>
Date: Tue, 27 Jun 2006 13:36:20 +0200
Subject: PhD / Postdoc Position in Numerical Analysis at RWTH Aachen

The development and analysis of numerical methods along with the
relevant theoretical foundations for a wide range of problem classes
is a central research area at the IGPM (Institut fuer Geometrie und
Praktische Mathematik).

Primary examples are adaptive solution strategies or nonlinear data
compression and analysis techniques. In particular, recent wavelet
concepts have opened new perspectives by closely intertwining the
discretization with the solution process aiming at realizing desired
target accuracies by using as few degrees of freedom as possible while
keeping storage and computational complexity proportional to this
minimal problem size.

These activities are closely intertwined with current
interdisciplinary research projects with partners from engineering
sciences, e.g., Collaborative Research Center SFB401 "Modulation of
Flow and Fluid-Structure Interaction at Airplane Wings".

Currently we are looking for candidates that hold a master/diploma in
mathematics. The candidate should have a strong background in the
numerical analysis of partial differential equations or computational
harmonic analysis.

Of particular interest are conservation laws, finite volume
discretizations, multilevel schemes, wavelet techniques. In this
context, good programming skills in C++ and a basic knowledge in
computational fluid dynamics are desirable but not a must.

We can offer a full PhD/Postdoc position which can be filled
immediately. The salary is according to BAT IIa.

Employment equity is one of our goals. We also welcome applications from
persons with disabilities.

For further details, see our homepage

http://www.igpm.rwth-aachen.de

or contact

Prof. Dr. W. Dahmen (dahmen@igpm.rwth-aachen.de)
Priv.-Doz. Dr. S. Mueller (mueller@igpm.rwth-aachen.de)

Applications should be sent in pdf format to

mueller@igpm.rwth-aachen.de

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

From: "Boglaev, Igor" <I.Boglaev@massey.ac.nz>
Date: Thu, 29 Jun 2006 16:24:06 +1200
Subject: Lecturer/Senior Lecturer position at Massey University

Applications are invited for the post of Lecturer/Senior Lecturer in
Mathematics from candidates with proven research
expertise in applied, computational or discrete mathematics. The
successful applicant will be required to teach in a variety of
courses in both internal and extramural modes and would be expected to
play a full part in the life of the mathematics subject area.
View detailed job description and apply online at
http://jobs.massey.ac.nz/positiondetail.asp?p=3987
Enquiries of an academic nature should be addressed to Dr Kee Teo,
Discipline Leader of Mathematics (telephone +64-6-356 9099 extension
3572, email K.L.Teo@massey.ac.nz).

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

From: <Nicolas.Gauger@dlr.de>
Date: Wed, 28 Jun 2006 16:29:32 +0200
Subject: PhD Student Position, DFG (German Research Foundation)

Within the DFG (German Research Foundation) funded SPP 1253 "Optimization with
partial differential equations" the German Aerospace Center (DLR Braunschweig)
has acquired the project "Multilevel parameterizations and fast multigrid
methods for aerodynamic shape optimization" and therefore aims to employ a

Scientific Researcher (PhD Student)
1/2 TVöD EG 13 (1/2 BAT 2A)

for two years with the option of one year extension, starting as soon as
possible.

The position gives the opportunity to work in a highly motivated
interdisciplinary team on a topic on the cutting-edge of applied mathematics
and scientific computing. There will be a strong collaboration with the Group
of Prof. Schulz at Universität Trier.

Project Summary:

Numerical flow simulation is an integral part of the construction process of
commercial aircrafts. Due to the advance of computing power and improvements
in the efficiency of simulation algorithms, it is nowadays conceivable to
design parts of aircrafts or even whole aircrafts on the computer
only. Therefore, numerical shape optimization will play a strategic role for
future aircraft design and for the respective industry. Current limitations of
numerical optimization methods in this field lie mainly in the high complexity
of numerical flow models, which causes long computation times in the order of
weeks and months, if a standard optimization algorithm is coupled with a flow
solver. And the computation time typically even depends on the shape
parameterization. That leads to questions of highly efficient optimization
algorithms. Also, the optimal representation of shapes on the computer is in
principle open, which leads to questions of parameterizations and local design
refinements. Both questions are of fundamental concern for aerodynamic shape
design and, of course, interact.

In this application, we propose to investigate multilevel shape
parameterizations and fast optimization algorithms exploiting the arising
multilevel structures in the shape and in the flow problem. The algorithmic
paradigm favored is that of an overall multigrid optimization method for all
variables involved, which is in contrast to traditional optimization
methods, which require a numerical effort which is equivalent to only a few
simulation runs - regardless of the resolution of the discretizations. The
goal of the project is to explore the exploitation of this potential in the
best way.

Qualifications:
- Diploma or MSc in mathematics, computer science, natural or
engineering science.
- Programming experience in C/C++ and FORTRAN.
- A background in Applied Mathematics, in particular Nonlinear
Optimization, Automatic Differentiation and Theory and Numerics of
Partial Differential Equations is an advantage.
- The candidate must be committed to interdisciplinary research and
project-oriented work.

Applicants should submit a detailed resume.

We explicitly encourage female researchers to apply for the offered
position. Among equally qualified applicants, disabled candidates will be
given preference.

Prof. Dr. Nicolas Gauger
German Aerospace Center (DLR)
Institute of Aerodynamics and Flow Technology
Numerical Methods Branch
Lilienthalplatz 7
38108 Braunschweig, Germany
Email: Nicolas.Gauger@dlr.de <mailto:Nicolas.Gauger@dlr.de>
Phone: +49 531 295-3339

Email applications are welcome.

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

From: "W. Hundsdorfer" <Willem.Hundsdorfer@cwi.nl>
Date: Tue, 27 Jun 2006 17:34:47 +0200
Subject: PhD position CWI Amsterdam

The research group "Modelling, Analysis and Simulation"
at CWI, Amsterdam, has an open position for a

PhD student Numerical Mathematics

on the project "Monotonicity Preservation for Time Stepping Schemes."

The research topic is the analysis and development of time stepping
schemes for the numerical solution of time-dependent partial
differential equations, in particular for hyperbolic equations
with stiff source and relaxation terms. The project description can
be found at <http://www.cwi.nl/~willem/Proj_MONGMM.html>

Applications are to be sent to Willem Hundsdorfer
<willem.hundsdorfer@cwi.nl>.

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

From: "Prof. Ralf Hiptmair" <hiptmair@sam.math.ethz.ch>
Date: Mon, 26 Jun 2006 07:52:30 +0200
Subject: PhD position in Applied Mathematics at ETH Zurich

The Seminar for Applied Mathematics of ETH Zuerich has open positions to be
filled with PhD students pursuing research projects in the field of numerical
methods for the solution of partial differential equations. Available thesis
topics include analysis and implementation of numerical methods for the
simulations of dense plasmas, fast solvers in computational electromagnetics,
and the numerical analysis of finite element methods for multiscale and
stochastic partial differential equations.

The positions are temporary and limited to 3-4 years after a three month period
of probation. Their holders will be regular ETH research staff and receive a
gross salary of roughly 62,000 Swiss Francs (about EUR 40.000, USD 50.000). The
positions will involve teaching of tutorial lectures, supervision of

For required skills and details of the application procedure see

http://www.sam.math.ethz.ch/news/jobs

Contact: Ms. Pfister, pfister@sam.math.ethz.ch

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

From: <sibjnm@oapmg.sscc.ru>
Date: Mon, 26 Jun 2006 15:41:24 +0700
Subject: Contents, SibJNM (Volume 9, No. 2, 2006)

CONTENTS, Siberian Journal of Numerical Mathematics
Volume 9, No. 2, 2006

For information to contributors and about subscriptions
see http://www.sscc.ru/SibJNM/

Amelkin V.A.
Recalculation, numbering and generation of serial
sequences with detached natural series}
(in Russian), pp.109-121

Godunov S.K., Selivanova S.V.
Experiments on using the resonance effect for spectral analysis of
finite-dimensional skew-symmetric operators
(in Russian), pp.123-136

Khludnev A.M., Leontiev A.N.
A problem of flow through semipermeable obstacle
(in English), pp.173-188

Kuznetsov Yu.I.
The orthogonal and the nodal polynomials
(in Russian), pp.137-145

Miloserdov V.V.
Conditional optimization of discrete-stochastic numerical procedures
with cubic splines applied
(in Russian), pp.147-163

Novikov S.I.
Periodic interpolation with a minimum norm of the \it{m}-th derivative
(in Russian), pp.165-172

Shary S.P.
Inner estimation of solution sets to non-negative interval linear systems
(in Russian), pp.189-206

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

From: "Kendall Atkinson" <Kendall-Atkinson@uiowa.edu>
Date: Mon, 26 Jun 2006 16:29:28 -0500
Subject: Contents, Journal of Integral Equations & Applications

Integral Equations and Applications. For abstracts of these articles and for
other information about the journal, go to the URL
http://rmmc.eas.asu.edu/abstracts/jie/jie.htm

*** Volume 17, Number 4.

"A Class of Maximal Operators Related to Rough Singular Integrals on Product
Spaces", by H. Al-Qassem and Y. Pan

"Marcinkiewicz Functions along Flat Surfaces with Hardy Space Kernels", by

"Iteratively Regularized Gradient Method with a Posteriori Stopping Rule for
2D Inverse Gravimetry Problem", by A. Bakushinsky, A. Smirnova and
M.A. Skinner

"Almost Automorphic Mild Solutions to Some Semi-Linear Abstract Di.erential
Equations with Deviated Argument", by Ciprian G. Gal

"Orthogonal Polynomials for the Solution of Semi Linear Two-Point Boundary
Value Problems", by Olaf Hansen

"Nonlinear Integral Equations with Increasing Operators in Measure Spaces", by
Laszlo Horvath

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

From: Arieh Iserles <ai@damtp.cam.ac.uk>
Date: Wed, 28 Jun 2006 16:47:53 +0100
Subject: Contents, Acta Numerica 15 (2006)

ACTA NUMERICA 15 (2006)

Douglas N. Arnold, Richard S. falk & Ragnar Winther,
"Finite element exterior calculus, homological techniques. and
applications"
1--155

J.C. Butcher
"General linear methods"
157--256

Emmanuel Cand\'es
"Modern statistical estimation via oracle inequalities"
257--325

Lars Eld\'en
"Numerical linear algebra in data mining"
327--384

Robert L. Higdon
"Numerical modelling of ocean circulation"
385--470

G\'erard Meurant & Zden\v{e}k Strako\v{s}
"The Lanczos and conjugate gradient algorithms in finite precision
arithmetic"
471--542

Robert Schaback & Holger Wendland
"Kernel techniques: From machine learning to meshless methods"
543--639

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

From: <sibjnm@oapmg.sscc.ru>
Date: Mon, 26 Jun 2006 15:40:44 +0700
Subject: Contents, SibJNM (Volume 9, No. 3, 2006)

CONTENTS, Siberian Journal of Numerical Mathematics
Volume 9, No. 3, 2006

For information to contributors and about subscriptions
see http://www.sscc.ru/SibJNM/

Bazanov P.V., Djosan O.V.
Methods of face feature extraction of the human identification problem
(in Russian), pp.207-214

Borisov Yu.S.
Visualization of city environment by the plenoptic method
(in Russian), pp.215-223

Ermolin E.N., Kazakov A.Yu.
The impulse-based method for the simultaneous resolution of collisions
between rigid bodies
(in Russian), pp.253-262

Ganelina N.D., Frolovsky V.D.
On constructing the shortest circuits on a set of line segments
(in Russian), pp.241-252

Konouchine A.S., Gaganov V.A., Vezhnevets V.P.
Extending RANSAC-based estimators to handle unknown and varying
noise level
(in English), pp.263-277

Korobeynikov A.V., Turlapov V.E.
Modeling and evaluation of the Stewart platforms
(in English), pp.279-286

Landovsky V.V., Frolovsky V.D.
Integration methods in the problem of modelling a fabric based on
the particles method
(in Russian), pp.287-298

Mestetskiy L.M.
Skeletonization of a multiply-connected polygonal domain based on its
(in Russian), pp.299-314

Pozin A.G.
Volumetric algorithm for 3D surface generation
(in Russian), pp.315-323

Vaganova N.A., Matsokin A.M.
The post-processing algorithms for images and video
(in Russian), pp.225-240

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

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