NA Digest Monday, August 6, 2012 Volume 12 : Issue 32

Today's Editor:
Daniel M. Dunlavy
Sandia National Labs

Submissions for NA Digest:

Mail to

Information via email about NA-NET:

Mail to


From: Art Werschulz <>
Date: Thu, 2 Aug 2012 17:14:29 -0400
Subject: In memoriam, Chris Sikorski

After a yearlong battle with cancer, School of Computing Professor
Krzysztof (Kris) Sikorski passed away on July 21, 2012.

A native of Poland, Sikorski received his Ph.D. in computer science at
the University of Utah in 1982. He then joined Columbia University as
a computer science faculty member, where he taught for 3 years before
returning to the University of Utah.

“Kris was a quiet, studious man who was a great colleague to many and
who brought a significant mathematical rigor to our curriculum and
research programs. He was, by nature, a man who held high standards
for both himself and his colleagues,” said Al Davis, department chair
and professor of computer science at the University of Utah. “His
wisdom has affected many, including both faculty and his students. He
will be missed and he will never be replaced. Kris’ legacy will endure
in the students that he influenced.”

Sikorski’s research focused on parallel scientific computation and
computational complexity, with special emphasis on optimal methods for
solving fixed-point problems. His contributions include the
development of tight complexity bounds and methods that are close to
or achieve these bounds, Davis adds.

In addition to his research and teaching at the University of Utah,
Sikorski served as Chair of the International Federation for
Information Processing group on continuous algorithms and complexity.
“In trying to think of one word that captures Kris as a researcher, as
an educator, as a fellow colleague, and as a friend—the word that
comes to mind is sincere,” said computer science professor Mike
Kirby. “Kris will indeed be truly missed.”

Sikorski is survived by his wife Elizabeth, son John (Monica),
grandchildren John Joseph and Eirene, sister Elizabeth, and extended
family members.


From: Timo Berthold <>
Date: Wed, 1 Aug 2012 10:57:48 -0400
Subject: SCIP Optimization Suite 3.0 released

Dear colleagues,

We are happy to inform you that version 3.0 of the SCIP Optimization
Suite has been released. It is comprised of SCIP 3.0, SoPlex 1.7,
ZIMPL 3.3 and initial releases of GCG 1.0, a generic
branch-cut-and-price solver, and UG 0.7, a parallel framework for
mixed integer (non-)linear programming.

The SCIP Optimization Suite 3.0 provides many new features,
performance improvements, and bug fixes. Some highlights are: six new
presolving and propagation plugins, an iterative refinement procedure
for computing high-precision LP solutions, further improved support of
MINLP, and beta-versions of an AMPL and a MATLAB interface. The SCIP
Optimization Suite can be obtained via

Best regards,
the SCIP team and friends

Release notes of the individual software packages:

SCIP 3.0:
SoPlex 1.7:
ZIMPL 3.3:
GCG 1.0:


From: Joseph Grcar <>
Date: Tue, 31 Jul 2012 14:29:20 -0700
Subject: Bay Area Scientific Computing Day

Hello NA Digest,

I suppose I shouldn't ask this question: who dropped the ball on the
San Francisco Bay Area Scientific Computing Day? This annual
gathering was started by Gene Golub and Beresford Parlett for the
benefit of students. Originally called Numerical Analysis Day, it
rotated between Stanford and Berkeley occasionally with sponsorship
from some of the energy labs. Despite the many corporations here that
in one way or another rely on the intersection of mathematics and
computing, there was no industrial support.

Joseph Grcar


From: Pavel Solin <>
Date: Wed, 1 Aug 2012 17:41:58 -0400
Subject: NCLab High-School CAD Challenge 2012

NCLab High-School CAD Challenge 2012
CAD is fun: Learn, create, win an iPad!
Submission open until November 30, 2012

Dear NA-Digesters, if you have a teenager in
high school, please pass to them the following

Pavel Solin


From: Martin Eigel <>
Date: Sun, 5 Aug 2012 09:28:05 -0400
Subject: Berlin PUM Workshop, Germany, Aug 2012

The "Berlin PUM Workshop 2012" will take place August 22-24 at the
Humboldt- University in Berlin, Germany. The focus lies on the
analysis and application of numerical approaches based on the
Partition of Unity Method such as GFEM, XFEM and Meshfree
Methods. Practitioners as well as numerical analysts are invited to
join in the discussion of recent results and future developments in
this field.

The invited speakers are:
Stéphane Bordas (Cardiff School of Engineering, UK)
Armando Duarte (University of Illinois, USA)
Markus Melenk (TU Wien, Austria)
Yves Renard (INSA de Lyon, France)
Marc Alexander Schweitzer (Universität Stuttgart, Germany)

For more details visit

You may contact for
questions regarding the workshop.

With kind regards from the organizers,
Andreas Byfut, Martin Eigel, Andreas Schröder


From: Andrea Cangiani <>
Date: Sun, 5 Aug 2012 18:25:10 -0400
Subject: Disc Meth for Polygonal & Polyhedral Meshes, Italy, Sep 2012

Workshop on Discretization Methods for Polygonal and Polyhedral Meshes
September 17-19, 2012
University of Milan-Bicocca, Milan, Italy

Main topics of the Workshop include (but are not limited to):
Mimetic Finite Difference methods
Virtual Element Methods
Polygonal/Polyhedral Finite Element Methods
Discontinuous Galerkin Methods
Extended Finite Element Methods
Isogeometric analysis

Invited Speakers: F. Bassi (Univ. of Bergamo), F. Brezzi (IMATI
Pavia), A. Buffa (IMATI Pavia), K. Hormann (Univ. of Lugano),
P. Houston (Univ. of Nottingham), K. Lipnikov (LANL), D.L. Marini
(Univ. of Pavia), G. Paulino (Univ. of Illinois), G. Sangalli
(Univ. of Pavia), N. Sukumar (UC Davis)

There is no registration fee, please register by September 12. For
more details and registration go to

or contact directly the organizers
Lourenço Beirão da Veiga (University of Milan, Italy)
Andrea Cangiani (University of Leicester, UK)
Gianmarco Manzini (Los Alamos National Laboratory, US)
Alessandro Russo (University of Milano-Bicocca, Milan, Italy)


From: Marc Van Barel <>
Date: Fri, 3 Aug 2012 10:32:59 +0200
Subject: Numerical Linear & Multilinear Algebra, Belgium, Sep 2012

The conference on "Structured Numerical Linear and Multilinear Algebra
Problems: Analysis, Algorithms and Applications" will be held at the
Department of Computer Science of the Catholic University of Leuven,
Belgium, 10-14 September 2012.

This is the fifth edition after the conferences held in Cortona
in 1996, 2000, 2004 and 2008.

The conference is intended to be very informal with a strong
interaction between the participants. Therefore parallel sessions are
avoided and the number of talks is limited to 50 where each lecture is
30 minutes. The preliminary programme and corresponding abstracts,
can be found on the website. If you intend to participate,
registration is open till the end of August.


From: Amik St-Cyr <>
Date: Mon, 6 Aug 2012 10:53:05 -0400
Subject: Computational Science Researcher Position, Royal Dutch Shell

I would like to post this new opening in our group (Computation &
Modeling) at Shell (2 such positions are available).

Creative Minds Wanted!

As a Senior Computational Science Researcher, you’ll play a pivotal
role in Shell's R&D organization with discovering and inventing
cutting-edge emerging technologies on the interface of computer
science, engineering and the physical sciences.

Computational research is growing rapidly within Shell’s global R&D
program in our quest to be world's most innovative energy company. We
are seeking innovative and forward-thinking computer science Ph.D.s
globally to be based in Houston, Texas and Bangalore, India.

Description and Responsibilities:

The Senior Computational Science Researchers in Houston, TX, will be
renowned technical leaders working with other top talent in this field
and fostering a national and international technical network with
universities, private and government research institutes, industrial
R&D companies and peers in Shell. The goal for this team of 10-15
scientists is to help solve the challenges associated with Big Data
Analytics and multi-scale modeling in geoscience and petroleum
engineering. We provide state-of-the-art computational resources
including large-scale HPC platforms. Areas currently being explored
related to this position include: Artificial Intelligence/Machine
Learning, algorithms, operating systems and programming models,
application performance analysis and tuning, statistical and
probability theory, application programming and very large scale
parallel computing.

This position offers a competitive compensation package and best in
class benefits. The successful candidate will have the advantage of
getting involved in leading research, attending conferences and
workshops and publishing. Shell fosters a safe, dynamic and
collaborative work environment.

-Candidates must have an PhD in Computer Science or Applied
-Must have 5+ years of hands-on post-PhD experience. Industry
experience preferred.
-Must have advanced reporting and presentation skills for frequent
internal and external publications and presentations
-Must show recognition for scientific and professional accomplishments
via papers, books, teaching or professional service.
-Must have advanced teamwork and leadership skills, must be willing
and able to coach and inspire others

With technically innovative people—and state-of-the-art equipment and
facilities—Shell is a worldwide leader in the development and
application of technology in the energy industry. Globally, Shell
operates in more than 80 countries and territories, is a recognized
innovator in oil and gas exploration and production technology and is
one of the world’s leading oil and gas producers, gasoline and natural
gas marketers and petrochemical manufacturers. At Shell, our
commitment is to satisfy the world’s growing need for energy with
economically, socially and environmentally responsible solutions.

If you want to work with a global leader and some of the finest minds
in the business, discover how you can propel your career while helping
us solve some of the world’s biggest energy challenges. Let’s deliver
better energy solutions together!


From: Steven Lee <>
Date: Sat, 4 Aug 2012 19:07:03 -0400
Subject: Program Mgr Position, DOE Adv Sci Computing Research

The Office of Advanced Scientific Computing Research (ASCR), in the
Department of Energy’s Office of Science, has posted 2 job summary
descriptions for upcoming employment opportunities as a Computer
Science Program Manager and an Applied Mathematics Program Manager in
ASCR – see the following for preliminary information:

These positions will appear in USAJOBS soon. It is essential that
applicants are prepared to provide transcripts for graduate and
undergraduate education when submitting their application.


From: Michaela Seiwald <>
Date: Mon, 6 Aug 2012 05:00:48 -0400
Subject: University Assistant Position

University Assistant with doctorate

40 hours a week; fixed-term employment for the period of 6 years with
qualification agreement; Envisaged Job Starting Date 01. October 2012

Job specification:
- Research and teaching in applied mathematics with emphasis on
continuous optimization, numerical analysis and analysis
- Cooperation in the reseach network "Mathematical Optimization and
Applications in the Biomedical Sciences"
- Collaboration in supervision of students
- Cooperation in organisational and administrative matters

Professional qualifications:
- Doctorial degree in a mathematical branch of study
- Solid knowledge in numerical analysis, differential equations and
functional analysis, mathematical optimization (continuous) and its
algorithmic realisation

Personal profile:
- High motivation for scientific excellence
- Capacity for teamwork
- Ability to teach in german language

Classification: Salary scheme of the Universitäten-KV (University
Collective Agreement): A2

Minimum salary:
The minimum salary as stated in the collective agreement and according
to the classification scheme is EUR 4004.70 gross/month. This minimum
salary may be higher due to previous employment periods eligible for
inclusion and other earnings and remunerations.

We offer you a job with a lot of responsibility and variety. You can
expect an enjoyable work climate, flexible work hours and numerous
possibilities for further education and personal development. Take
advantage of the chance to enter into a challenging work environment
full of team spirit and enthusiasm for your job.

Application Deadline: 29. August 2012
Reference Number: MB/109/99 ex 2011/12

If you are interested, please submit your application documents within
the stated deadline. Make sure to indicate the reference number on
your application and please send your CV, photo and relevant letters
of recommendation to:

Karl-Franzens-Universität Graz
Universitätsplatz 3
8010 Graz
or by e-mail


From: Gerhard Starke <>
Date: Sat, 4 Aug 2012 06:15:00 -0400
Subject: Professorship Position, Numerical Analysis, Univ of Hannover

The Faculty of Mathematics and Physics of the Leibniz Universität
Hannover invites applications for a University Professorship in
Numerical Analysis (Salary Scale W 3 BBes0) starting at the earliest
possible date.

For more information please go to

Deadline: September 30, 2012


From: Emil Constantinescu <>
Date: Wed, 1 Aug 2012 17:01:47 -0400
Subject: Postdoc Position, Time Integration, Argonne

Postdoctoral Appointee, Laboratory for Advanced Numerical Simulations

The Mathematics and Computer Science (MCS) Division at Argonne
National Laboratory invites outstanding candidates to apply for a
postdoctoral position focusing on the development of scalable time
integration algorithms and software for time-dependent multimodel
partial differential equations. The successful candidate will design
such new algorithms as implicit-explicit, multirate, and error
control; extend the time stepping component (TS) of the Portable
Extensible Toolkit for Scientific computing (PETSc), and collaborate
with applications scientists to apply these techniques to large-scale
scientific problems.

The successful candidate will work as part of a multidisciplinary
research team involving mathematicians, computer scientists, and
domain scientists. Candidates should have a background in numerical
techniques for time integration and the solution of ordinary and
partial differential equations and linear and nonlinear solvers. Also,
expertise in one or more of Fortran/C/C++ is required, as well as
experience with numerical algorithms and software for large-scale
applications based on partial differential equations and parallel
programming algorithms and software with MPI.

Candidates must have a PhD in applied mathematics, computer science,
computational science, or a related field. Candidates should submit an
application via the Argonne website under division postdoctoral job
openings for
job requisition 319443. The application must include a curriculum
vitae; list of publications, abstracts, and significant presentations;
and the names of three references, other than Argonne staff, who can
attest to the candidate's ability and potential.

For further information about division postdoctoral appointments at
Argonne, see the website. Questions can be addressed by email to Emil
Constantinescu ( MCS Division.


From: "Craster, Richard V" <>
Date: Thu, 2 Aug 2012 12:00:18 +0000
Subject: Postdoc Position, Imperial College

A postdoc position on the Multiscale modelling of surfactant laden
flows at Imperial College is now available and applications are
invited. Details are online

or below.

Applications are invited for a researcher to work in the Department of
Chemical Engineering at Imperial College London based at the South
Kensington Campus. The research project is on multi-scale modeling,
and analysis, for the engineering and control of surfactant-laden
flows. The successful candidate will integrate coarse-grained
molecular simulations with continuum-level modeling thereby creating a
macroscale model that none the less encapsulates the dominant
micro-scale behaviour. This will describe the wetting and spreading of
drops doped with super-spreading surfactants and unveil the underlying
physics behind this phenomenon. You will also be expected to produce
independent and original research in the area of computer simulations
of related coarse grained and continuum-scale models.

The successful applicant will be fully funded by the EPSRC and will
interact with a truly multidisciplinary group spanning several
departments at Imperial College London, with experimental partners at
University of Loughborough.

You should hold, or be near completion of a PhD (or equivalent) in
Chemical or Mechanical Engineering, Physics, Chemistry or a closely
related discipline. You should have a strong background on the
development and use of computer simulation methods, as well as
statistical mechanics, thermodynamics and fluid mechanics. Excellent
knowledge of FORTRAN and C++, Unix, numerical analysis and a proven
publication track record are essential.

Informal enquiries may be made to Professor Omar Matar


From: "Hochstenbach, M.E." <>
Date: Sat, 4 Aug 2012 13:06:30 +0200
Subject: 2 PhD Positions, TU Eindhoven, Netherlands

Two 4-year PhD positions are available at TU Eindhoven, starting
date negotiable. Research topics include generalized matrix
eigenvalue problems and Krylov methods, ill-posed problems,
and probabilistic methods.

Required: MSc in applied mathematics (numerical analysis or closely
related field), good analytical and communication skills.

For informal inquiries and applications, please contact
Dr. Michiel Hochstenbach, TU Eindhoven, by email.
Please include: presentation letter, CV, list of MSc. grades,
list of (at least 3) people for recommendation.

See also


From: "Lucking, Laura - Chichester" <>
Date: Wed, 1 Aug 2012 14:53:31 +0100
Subject: Contents, Statistical Analysis and Data Mining, 5(4)

Statistical Analysis and Data Mining - Volume 5, Issue 4


Review of statistical network analysis: models, algorithms, and
software (pages 243-264) M. Salter-Townshend, A. White, I. Gollini and
T. B. Murphy

Research Articles

Effective graph classification based on topological and label
attributes (pages 265-283) Geng Li, Murat Semerci, Bülent Yener and
Mohammed J. Zaki

From black and white to full color: extending redescription mining
outside the Boolean world (pages 284-303) Esther Galbrun and Pauli

Multi-view predictive partitioning in high dimensions (pages 304-321)
Brian McWilliams and Giovanni Montana

Interval Archetypes: A New Tool for Interval Data Analysis (pages
322-335) Maria R. D'Esposito, Francesco Palumbo and Giancarlo Ragozini

A resampling approach for interval-valued data regression (pages
336-348) Jeongyoun Ahn, Muliang Peng, Cheolwoo Park and Yongho Jeon

Nearest-neighbors medians clustering (pages 349-362) Daniel Peña,
Júlia Viladomat and Ruben Zamar

End of NA Digest