NA Digest Saturday, May 24, 2003 Volume 03 : Issue 21

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

Submissions for NA Digest:

Mail to

Information via e-mail about NA-NET: Mail to


From: Ken Turkowski <>
Date: Mon, 19 May 2003 12:56:34 -0700
Subject: Quaternion Specification from Vector Transformation

Can a [unit] quaternion be specified by its action on two normalized vectors?

I'm thinking, in particular, about the transformation of two particular
vectors: straight ahead and up.

If you represent points on the unit sphere with angles (theta, phi),
and the ahead and up vectors by
(theta=0, phi = 0), (theta=0, phi = pi/2)
and their transformation by quaternion multiplication
v' = q v q*
then it seems like
(theta_ahead, phi_ahead, theta_up, phi_up)
are 4 parameters that should be able to uniquely specify a quaternion
q = (qw, qx, qy, qz)
Right? The restrictions on the normalized quaternion should be balanced
out by the conformal (angle-preserving) nature of the quaternion
transformation. I realize that this is hand-waving. Is there some
justification for this?

There are many practical formulas to convert between quaternions and
rotation matrices, but the embedding into a higher-dimensional space
and projection seems like a waste of effort.

Ken Turkowski
Immersive Imaging Technologist


From: Leslie Greengard <>
Date: Fri, 23 May 2003 21:51:30 -0400 (EDT)
Subject: Fast Multipole Methods Toolbox

The FMM 2D Toolbox v1.0 for MATLAB is now available from MadMax Optics
for free download. This Toolbox provides three kinds of solvers based on
optimized Fast Multipole Methods (FMMs).

- Particle codes for 2D Coulombic and screened Coulombic interactions

- Solvers for homogeneous boundary value problems in interior or exterior
multiply connected domains with Dirichlet or Neumann conditions:
(Laplace and modified Helmholtz equations)

- Solvers for inhomogeneous boundary value problems in interior or exterior
multiply connected domains with Dirichlet or Neumann conditions:
(Poisson and modified Helmholtz equations)

The PDE solvers are true black-box tools, based on FMM-accelerated integral
equations using high-order quadrature and interpolation schemes.
Solutions are obtained in O(N) time, where N is the number of points in the
discretization. Boundaries are easily described as a sequence
of points. Volume source data (for the Poisson and inhomogeneous modified
Helmholtz equation) can be specified at a set of arbitrary locations,
allowing the solver to be coupled with virtually any underlying data
structure. No grid generation is required.

In order to provide solvers with the maximum of flexibility, we have
incorporated a number of novel features, including algorithms that can be
applied to regions with corners and discontinuous boundary data.
Thus, while the amount of work is of the order O(N), the solvers
are optimized for flexibility rather than speed. The latter would require
stricter control of the user's discretization schemes and more detailed
assumptions about the user's data.

The Toolbox and user manual are available for free download at

Selected 3D FMM algorithms will be available on this website
in the near future.


From: Trevor Misfeldt <>
Date: Mon, 19 May 2003 13:34:48 -0700
Subject: NMath Core from CenterSpace Software

CenterSpace Software is pleased to announce the release of NMath Core 1.0:


NMath Core contains vector, matrix, and complex number classes, random
number generators, and other high-performance functions for object-oriented
numerics on the .NET platform. NMath Core is designed to run with
machine-specific, highly optimized versions of the public domain linear
algebra subroutine libraries known as the BLAS (Basic Linear Algebra
Subroutines) and LAPACK (Linear Algebra PACKage). This gives NMath Core
classes performance levels comparable to C, and often results in
performance an order of magnitude faster than non-platform-optimized
implementations. Fully CLS-compliant, all NMath Core routines are callable
from any .NET language, including C# and Visual Basic.NET.

A free evaluation version of NMath Core is available from our website,
along with complete user documentation, performance benchmarks, whitepaper,
and code examples.

Trevor Misfeldt
CEO, CenterSpace Software


From: Joseph Traub <>
Date: Tue, 20 May 2003 11:19:09 -0400 (EDT)
Subject: Journal of Complexity Best Paper Award


The Award Committee consisting of Fred Hickernell and Peter Mathe
has selected the paper "Topological Complexity of Zero Finding with
Algebraic Operations" by Peter Hertling which appeared in December 2002,
Vol. 18, No. 4, Pages 912-942.

The author will be presented with a plaque and the $3,000 prize at
a suitable location.

Harald Niederreiter
Joseph F. Traub
Henryk Wozniakowski


From: Graham Robertson <>
Date: Wed, 21 May 2003 10:29:40 +0100
Subject: New Book on Differential Equations

New Textbook From Cambridge:
"Differential Equations: Linear, Nonlinear, Ordinary, Partial"
by A.C. King, J. Billingham and S. Otto

This book enables the reader to develop the required skills to thoroughly
understand Differential Equations. The authors focus on the business of
constructing solutions analytically, and interpreting their meaning; MATLAB
is used extensively to illustrate the material.

Preface; Part I. Linear Equations: 1. Variable coefficient, second-order,
linear ordinary differential equations; 2. Legendre functions; 3. Bessel
functions; 4. Boundary value problems, Green's functions and
Sturm-Liouville theory; 5. Fourier series and the Fourier transform; 6.
Laplace transforms; 7. Classification Properties and Complex Variable
Methods for Second Order Partial Differential equations; Part II. Nonlinear
Equations and Advanced Techniques: 8. Existence, uniqueness, continuity and
comparison of solutions of ordinary differential equations; 9. Nonlinear
ordinary differential equations; 10. Group theoretical methods; 11.
Asymptotic methods: basic ideas; 12. Asymptotic methods: differential
equations; 13. Stability, instability and bifurcations; 14. Time-optimal
control in the phase plane; 15. An introduction to chaotic systems;
Appendix 1. Linear algebra; Appendix 2. Continuity and differentiability;
Appendix 3. Power series; Appendix 4. Sequences of functions; Appendix 5.
Ordinary differential equations; Appendix 6. Complex variables; Appendix 7.
A short introduction to MATLAB; Bibliography; Index.

For more information please visit:


From: Vadim Olshevsky <>
Date: Mon, 19 May 2003 19:11:24 -0400
Subject: New Book on Structured Matrices

Just published by AMS and SIAM:

Fast Algorithms for Structured Matrices: Theory and Applications
Edited by: Vadim Olshevsky, University of Connecticut, Storrs, CT
A joint publication of the
American Mathematical Society,
Series: Contemporary Mathematics, Volume: 323
Society for Industrial and Applied Mathematics
Publication Year: 2003
ISBN: 0-8218-3177-1
Paging: 433 pp.

The volume can be ordered at the AMS bookstore

One of the best known fast computational algorithms is the fast Fourier
transform method. Its efficiency is based mainly on the special structure
of the discrete Fourier transform matrix. Recently, many other algorithms
of this type were discovered, and the theory of structured matrices emerged.

This volume contains 22 survey and research papers devoted to a variety of
theoretical and practical aspects of the design of fast algorithms for
structured matrices and related issues. Included are several papers containing
various affirmative and negative results in this direction. The theory of
rational interpolation is one of the excellent sources providing intuition and
methods to design fast algorithms. The volume contains several computational
and theoretical papers on the topic. There are several papers on new
applications of structured matrices, e.g., to the design of fast decoding
algorithms, computing state-space realizations, relations to Lie algebras,
unconstrained optimization, solving matrix equations, etc.

The book is suitable for mathematicians, engineers, and numerical analysts
who design, study, and use fast computational algorithms based on the theory
of structured matrices.

This volume is a joint publication of the American Mathematical Society and
the Society for Industrial and Applied Mathematics.

V. Olshevsky -- Pivoting for structured matrices and rational tangential

G. Heinig -- Inversion of Toeplitz-plus-Hankel matrices with arbitrary rank

D. Fasino and L. Gemignani -- A Lanczos-type algorithm for the QR factorization
of Cauchy-like matrices

D. Fasino, N. Mastronardi, and M. Van Barel -- Fast and stable algorithms for
reducing diagonal plus semiseparable matrices to tridiagonal and bidiagonal form

A. Olshevsky, V. Olshevsky, and J. Wang -- A comrade-matrix-based derivation of
the eight versions of fast cosine and sine transforms

D. A. Bini, L. Gemignani, and B. Meini -- Solving certain matrix equations
by means of Toeplitz computations: algorithms and applications

F. T. Luk and S. Qiao -- A fast singular value algorithm for Hankel matrices

D. Calvetti, L. Reichel, and F. Sgallari -- A modified companion matrix
method based on Newton polynomials

J. Hendrickx, R. Vandebril, and M. Van Barel -- A fast direct method for
solving the two-dimensional Helmholtz equation, with Robbins boundary conditions

C. Di Fiore -- Structured matrices in unconstrained minimization methods

N. Ito, W. Schmale, and H. K. Wimmer -- Computation of minimal state space
realizations in Jacobson normal form

A. Mayo -- High order accurate particular solutions of the biharmonic
equation on general regions

T. Wen, A. Edelman, and D. Gorsich -- A fast projected conjugate gradient
algorithm for training support vector machines

V. Olshevsky and M. A. Shokrollahi -- A displacement approach to decoding
algebraic codes

M. Bollh=F6fer and V. Mehrmann -- Some convergence estimates for algebraic
multilevel preconditioners

D. Noutsos, S. S. Capizzano, and P. Vassalos -- Spectral equivalence and matrix
algebra preconditioners for multilevel Toeplitz systems: a negative result

W. F. Trench -- Spectral distribution of Hermitian Toeplitz matrices
formally generated by rational functions

D. Fasino and S. S. Capizzano -- From Toeplitz matrix sequences to zero
distribution of orthogonal polynomials

K. R. Driessel -- On Lie algebras, submanifolds and structured matrices

H. Dym -- Riccati equations and bitangential interpolation problems with
singular Pick matrices

V. Bolotnikov, A. Kheifets, and L. Rodman -- Functions with Pick matrices
having bounded number of negative eigenvalues

Yu. M. Arlinskii, S. Hassi, H. S. V. de Snoo, and E. R. Tsekanovskii --
One-dimensional perturbations of selfadjoint operators with finite or discrete


From: Axel Voigt <>
Date: Tue, 20 May 2003 09:33:59 +0200
Subject: Workshop in Bonn on Continuum Models for Epitaxial Growth

Last call for papers, new deadline 31 May 2003

Workshop on
Continuum Models for Epitaxial Growth
August, 4th- 6th, 2003
Bonn, Germany

Felix Otto, University of Bonn
Axel Voigt, Research Center Caesar

In order to describe epitaxial thin film growth a large variety of continuum
models at different time and length scales have been derived. The workshop
concentrates on bridging the gap between this different models, with a special
focus on BCF and its variants. We will bring together experts from different
communities to report about the state-of-the-art of modelling epitaxial growth
with the aim to combine microscopic film properties with macroscopic growth
process conditions. The symposium will take place in Bonn, in the newly built
research center caesar. Thin film growth is presently one of the key research
areas of caesar. According to the philosophy of caesar it is carried out in
close cooperation between experimentalists and theorists.

For details, please, visit the workshop Web site at:

We look forward to welcome you in Bonn

Felix Otto and Axel Voigt


From: Jan Hesthaven <>
Date: Wed, 21 May 2003 13:20:30 -0400 (EDT)
Subject: Conference at Brown on Spectral and High-Order Methods

First Announcement
6th International Conference on Spectral
and High-Order Methods (ICOSAHOM'04)
June 21-25, 2004
Brown University, RI, USA

The 6th International Conference on Spectral and High Order
Methods (ICOSAHOM) will continue a tradition established
through the organization of the 1st ICOSAHOM by C. Canuto
and A. Quarteroni in Como, Italy in 1989. This conference
continues to be a main conference venue where significant
advances and new applications of high-order accurate methods
for solving PDE's are presented.

Conference themes include but are not limited to spectral
methods, high-order finite difference and finite element
methods, h-p finite element methods, discontinuous Galerkin
methods, ENO/WENO methods, high-order methods for integral
equations, and wavelet based methods.

Applications include problems appearing in electromagnetics,
fluid and structural dynamics, climate modeling, inverse
problems, efficient solvers and preconditioners for high-
order methods, efficient time-stepping techniques, and
parallel and computational aspects of spectral and high-
order methods.

To highlight recent developments, a number of leading
reseachers have kindly agreed to give invited plenary
presentations. These include

Mark Ainsworth (Strathclyde University, Scotland)
Mejdi Azaiez (ENSCPB, France)
Ivo Babuska (ICES, University of Texas, Austin, USA)
Silvia Bertoluzza (IAN-CNR, Pavia, Italy)
Wei Cai (University of North Carolina, USA)
Monique Dauge (Rennes, France)
Dale Haidvogel (Rutgers University, USA)
Luca Pavarino (University of Milano, Italy)
Tim Phillips (University of Wales, UK)
Eli Turkel (Tel Aviv University, Israel)
Miguel Visbal (WP Air Force Base, USA)

An integral part of ICOSAHOM'04 will be topic-specific
minisymposia, following a tradition that encompasses all
previous ICOSAHOM conferences. The purpose of these
minisymposia are several, e.g., presentation of new and
exciting developments within the scope of the conference,
new areas of applications in the sciences and engineering,
and to emphasize areas of particular recent activity and

To be included on the maillist for future announcements,
view important dates and a prelimenary call for minisymposia
organizers and contributed papers we invite you to visit
the conference webpage

or to contact the conference secretary

Ms. Janice D'Amico
ICOSAHOM'04 Secretary
Division of Applied Mathematics
Brown University
Providence, RI 02912


Best regards

The organizing committee

David Gottlieb
Jan S. Hesthaven
George E. Karniadakis
Chi-Wang Shu


From: Julyan Cartwright <>
Date: Fri, 23 May 2003 20:25:59 +0200
Subject: Conference in Spain, Dynamics Days

Dynamics Days 2003
XXIII Annual Conference
4 decades of chaos 1963-2003

24-27 September
Palma de Mallorca, Spain

Dynamics Days is an annual interdisciplinary conference designed to stimulate
interactions amongst researchers with interests in dynamical systems,
particularly those with nonlinear aspects.

Dynamics Days 2003 will take place from Wednesday 23 September to Saturday
27 September, on the campus of the University of the Balearic Islands,
Palma de Mallorca, Spain.

Organizers: Oreste Piro (Palma), Julyan Cartwright (Granada)



Michael Berry
Oriol Bohigas
Leonid Bunimovich
Yves Couder
Pierre Coullet
Theo Geisel
Ray Goldstein
Celso Grebogi
Mogens Jensen
Rafael de la Llave
Marcelo Magnasco
Giorgio Mantica
Igor Mezic
Heinz Georg Schuster
Jorge Tredicce
Angelo Vulpiani

Apart from the plenary talks by the invited speakers, contributed oral and
poster communications by participants are expected to form a large and
important part of the meeting. Abstract submissions are therefore
enthusiastically encouraged. Most of the accepted oral contributions will
be arranged together with the plenary talks in a single session.

Additionally, a part of the conference will be devoted to special
minisymposium sessions around common selected themes. These sessions will
contain both key note talks and contributions. We anticipate having
minisymposia on the following themes, but additional ones may be
announced later.

Biological systems
Socio- and econodynamics
Quantum chaos
Chaotic communication
Mixing and fluids

Deadlines for submissions: To be considered for an oral presentation,
abstracts must be received by the organizers by 31st July. No limit will be
placed on the number of posters.

The registration fee is 250 euros; there is a reduced fee of 150 euros
for students.

There is a limited amount of funding available to enable students to attend
who could not otherwise afford to do so, especially those from eastern Europe.
Details of the funding and how to apply for it are given on the conference Web
site - see below.

Accommodation for participants may be booked through the conference Web site.

Online registration and further information can be found on the
conference web pages at
These pages are being updated as further information is available, so please
check them regularly.


From: Bettina Felsner <>
Date: Fri, 23 May 2003 12:24:18 +0200
Subject: Faculty Position at Freie Universitaet Berlin


Freie Universitaet Berlin announces an opening in the Department of
Mathematics and Computer Science of the newly established position of

Professor (C4) of "Mathematics in life sciences".

The position is an integral Part of the DFG research Center "Mathematics
for key technologies. Modeling, simulation, and optimization of real-world
processes" ( The rank is Full Professor with tenure (C4).

Qualifications: According to =A7 100 of the Berlin Higher Education Act, a
post-doctoral thesis (Habilitation) or comparable qualifications in
science and education are obligatory. Specific expertise in mathematics
with strong background in its application to the life sciences are required.

Further Qualifications: The successful applicant should be an expert in
applied stochastics and/or statistics, numerical mathematics, discrete
mathematics, or in related aspects of applied mathematics. He or she
should have a rich experience with mathematical methods for the solution
of real life application problems from the life sciences such as medical
or biotechnology, biostatistics, modeling and simulation of complicated
dynamical behaviour in biological, biochemical or neural systems,
biomechanics or tissue engineering. He or she is expected to collaborate
intensively within the DFG research center. Appropriate participation in
the self-organization duties of the Center and in the scientific
education of younger researchers is essential. Association of a junior
research team to the professorship is possible.

Applicants should send a curriculum vitae, a statement of research and
teaching interests, and a list of publications, to the following address:

Freie Universitaet Berlin
Fachbereich Mathematik und Informatik
Arnimallee 2-6
14195 Berlin

Please quote reference number DFG-FZ-C4 in your covering letter. The
deadline for receiving applications is July 3rd, 2003

Freie Universit=E4t Berlin is an equal opportunity employer.

Further information can be obtained from Christof Schuette (scientific), or Bettina Felsner (administrativ),


From: Alexandre Ern <>
Date: Thu, 22 May 2003 15:12:17 +0200
Subject: Postdoctoral Position at Ecole Nationale des Ponts et Chaussees

Applications are invited for a one-year postdoctoral position at INRIA
( in the field of Computational Geosciences.

The position is offered within the Cooperative Research Action DYNAS
funded by INRIA ( DYNAS is a two-year
interdisciplinary project launched in January 2003 involving applied
mathematicians, modelling experts of subsurface flows and geoscience
engineers. The project aims at developping a software tool based on
sound physical modelling and robust numerical methods to simulate the
dynamics of shallow water tables in variably saturated porous media with

a particular emphasis on heavy rainfall episodes.

A preliminary series of test cases neglecting surface runoff has been
established. Numerical methods have been selected with an emphasis on
locally conservative methods such as mixed finite elements, box schemes
and cell-centered finite volume schemes. These methods are currently
being assessed against the test cases and compared with experimental

The postdoctoral research will pursue the following objectives:
- study appropriate boundary conditions coupling the subsurface water
table to the kinematic wave generated by surface runoff,
- investigate the impact of the hydrodynamical properties of the soil
on model predictions,
- based on various numerical tools already available within the DYNAS
project, develop a 3D software package designed to simulate the dynamics
of shallow water tables and the appearance of surface runoff within rain

The candidate should have, or receive shortly, a PhD in the field of
applied mathematics with a strong background in the numerical solution
of partial differential equations and a relevant experience in computer
programming. Some background in the modelling of subsurface flows is an
important asset.

The position will be for twelve months starting in September or October
2003. The postdoctoral fellowship is approximately 1800 euros/month net
and includes basic health insurance. The candidate will be located at
the Cermics laboratory of the Ecole Nationale des Ponts et Chaussees
near Paris ( and will interact
frequently with all the teams participating in the DYNAS project, in
particular the ESTIME project at INRIA-Rocquencourt
( and the Drainage and Barrier Engineering Unit
at Cemagref (, both in the Paris metropolitan area.

Candidates should submit by e-mail (pdf format preferred) or by post
- a cover letter in support of the application,
- a curriculum vitae with a list of publications,
- one to three publications,
- name and address of three references.

The material should be sent by June 30 to:
Prof. Alexandre Ern
6 et 8, avenue Blaise Pascal
F-77455 Marne la Vallee cedex 2, France


From: Andreas Adelmann <>
Date: Mon, 19 May 2003 06:40:43 +0200
Subject: Short Term Position at ETH, Zurich

Optimal two and three dimensional Transpose for Parallel FFT's

Fast Fourier Transform methods are commonly used for numerically
solving partial differential equations. One crucial point is that data
organization in distributed memory computer systems must have each
entire vector corresponding to the current transformation direction
memory resident on one processor. Typically, the organization will
consist of "slabs" or "pencils" containing several entire vectors of the
current transform direction on each independent processor. When
vectors in a new direction are to be Fourier transformed, a reorganization,
or transpose, of the existing layout must be performed. Although log(n)
steps are needed on each vector (of size n) to perform the FFT, it turns out
that a single transpose operation to move the new direction vectors into
contiguous memory locations on processors represents a significant
portion of the total transformation time. Hence the transpose operation
is very important. It is this transposition that we wish to implement

We are looking for a student of computational science, mathematics,
physics, or engineering with solid C or C++ and MPI experience.
Ideally the candidate is close to graduation.

This is a joint project between the Seminar for Applied Mathematics (SAM) of
ETH Zurich and Paul Scherrer Institut (PSI). Duration 3 months, starting
first of July 2003.

Contact Dr. Andreas Adelmann,


From: Frank Tobias <>
Date: Mon, 19 May 2003 12:02:41 -0400
Subject: Staff Position at GlaxoSmithKline

Mathematical Modeler at GlaxoSmithKline

GlaxoSmithKline (GSK) is a world leading research-based pharmaceutical
company with an opportunity available in our Upper Merion, PA facility
with the Scientific Computing and Mathematical Modeling group. This
position will be responsible for conducting research and developing
mathematical models of biological systems for the purposes of advancing
drug discovery and development efforts. This will involve understanding
the complexities of the biology (e.g. the chemical, genetic, cellular,
tissue, spatial, temporal, and geometric aspects), the needs of the drug
development team involved, and formulating a proper computational
approach. That approach will then be translated into a series of
mathematical models which are to be numerically solved and the solutions
delivered to the team. This role will require the ability to work with
biologists and to convey sophisticated mathematical approaches in a
relevant manner. We are looking for that rare individual who can
successfully integrate sophisticated mathematics and computer software
with biological problems. We seek someone who has the motivation and
drive to solve real-world problems.

We are looking for an individual with experience in mathematical biology
or modelling of complex biological systems and who has an excellent
grounding in several of the following: modelling, numerical analysis,
differential equations, function approximation, dynamical systems,
fractals and stochastic processes, and other modern computational
techniques. A good general scientific or engineering background (e.g.
mathematics, physics, chemistry, electrical engineering, etc.) as well as
a biological one is important. While experience with bioinformatics (i.e.
sequence analysis), molecular biology, or genetics is a asset, it is not a
requirement. An advanced degree (Ph.D.) in a scientific field as well as
strong, documented experience in scientific computing are essential.

GlaxoSmithKline offers a competitive benefits and compensation package.
For confidential consideration and efficient processing, please visit our
website: and apply online indicating Job Code Requisition
Number: 10080


From: John Ramsdin <>
Date: Thu, 22 May 2003 12:45:28 +0100
Subject: Studentships at University of Huddersfield

EPSRC Case Studentships in Computational Mathematics

Two PhD three-year awards are available from October 2003 or earlier.

(i) Rational Approximation in Metrology
The study of measurement systems leads to a requirement for fitting data
representing system responses by model curves and surfaces. Most
empirical models available (polynomials, splines, etc) do not model well
the asymptotic behaviour (eg decay, singularities) observed in many
physical systems. The research will take a new look at rational
approximation techniques that are believed to be a powerful method that
can meet this important requirement. The research is sponsored by the
National Physical Laboratory.

(ii) Spectral Methods for Brake Noise Modelling
Brake noise is a long-standing problem for the automotive industry and
there is a long history of research into the causes of what has been
shown to be a complex phenomena. Standard modelling techniques using
finite element packages have met with limited success in predicting when
brake noise can occur. Spectral methods of approximation have proven
successful for problems that are usually expensive on computing time and
should give a valuable new approach in this case. The research is
sponsored by Alcon Components Ltd.

Applicants for either post should hold at least a 2.1 degree, or
equivalent, preferably in a mathematically-related area, such as physics
or engineering. They would join a successful group of practical
researchers. The studentships include registration fees and provides a
salary of approximately =A312k pa.

To apply, send your CV and two referee names by e-mail to:
* for post (i), Professor J. C. Mason, e-mail
* for post (ii), Dr C. J. Talbot, e-mail

University of Huddersfield
School of Computing and Engineering


From: Vladik Kreinovich <>
Date: Wed, 21 May 2003 09:08:18 -0600 (MDT)
Subject: Contents, Reliable Computing

Reliable Computing
Volume 9, issue 4, 2003

Mathematical Research

Worst-Case Simulation of Discrete Linear Time-Invariant Interval Dynamic Systems
Vicenc Puig, Jordi Saludes, Joseba Quevedo

Optimal Bicentered Form
Youkang Fang

Asymptotic Stability of Interval Time-Delay Systems
Svetlana P. Sokolova, Ruslan S. Ivlev


From: ScienceDirect <>
Date: Sat, 24 May 2003 09:51:21 -0400 (EDT)
Subject: Contents, Computational Statistics and Data Analysis

Computational Statistics & Data Analysis Volume 43, Issue 2, Pages 121-282
(28 June 2003)

On the exact distribution of maximally selected rank statistics, Pages 121-137
Torsten Hothorn and Berthold Lausen

Wavelet denoising of Poisson-distributed data and applications, Pages 139-148
C. Charles and J. P. Rasson

Three-way fuzzy clustering models for LR fuzzy time trajectories, Pages 149-177
Renato Coppi and Pierpaolo D'Urso

Discriminating between Weibull and generalized exponential distributions,
Pages 179-196
Rameshwar D. Gupta and Debasis Kundu

Reducing over-dispersion by generalized degree of freedom and propensity score, Pages 197-214
Ie-Bin Lian

On tests of homogeneity based on minimum [phi]-divergence estimator with
constraints, Pages 215-234
M. L. Menendez, J. A. Pardo, L. Pardo and K. Zografos

Design of hybrids for the minimum sum-of-squares clustering problem,
Pages 235-248
Joaquin Pacheco and Olga Valencia

Computing discrete mixtures of continuous distributions: noncentral chisquare,
noncentral t and the distribution of the square of the sample multiple
correlation coefficient, Pages 249-267
Denise Benton and K. Krishnamoorthy


From: ScienceDirect <>
Date: Sat, 24 May 2003 09:51:21 -0400 (EDT)
Subject: Contents, Journal of Approximation Theory

Journal of Approximation Theory Volume 122, Issue 2, Pages 151-334 (June 2003)

Best approximation by linear combinations of characteristic functions of
half-spaces, Pages 151-159
Paul C. Kainen, Vra Krkova and Andrew Vogt

The uniform convergence of subsequences of the last intermediate row of the
Pade table, Pages 160-207
Victor M. Adukov

Sur les fonctions q-Bessel de Jackson, Pages 208-223
Changgui Zhang

Strong asymptotics in Lagrange interpolation with equidistant nodes,
Pages 224-240
Michael I. Ganzburg

Markov inequality for polynomials of degree n with m distinct zeros,
Pages 241-248
David Benko and Tamas Erdelyi

Bounds for symmetric elliptic integrals, Pages 249-259
Edward Neuman

Approximation by neural networks with a bounded number of nodes at each level,
Pages 260-266
G. Gripenberg

Summability of orthogonal expansions of several variables, Pages 267-333
Zhongkai Li and Yuan Xu


End of NA Digest