Textbook Computer Algebra in Scientific Computing 16Th International Workshop Casc 2014 Warsaw Poland September 8 12 2014 Proceedings 1St Edition Vladimir P Gerdt Ebook All Chapter PDF

Computer Algebra in Scientific
Computing 16th International

Workshop CASC 2014 Warsaw Poland
September 8 12 2014 Proceedings 1st
Edition Vladimir P. Gerdt
Visit to download the full and correct content document:
https://textbookfull.com/product/computer-algebra-in-scientific-computing-16th-interna
tional-workshop-casc-2014-warsaw-poland-september-8-12-2014-proceedings-1st-ed
ition-vladimir-p-gerdt/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Computer Algebra in Scientific Computing 20th

International Workshop CASC 2018 Lille France September
17 21 2018 Proceedings Vladimir P. Gerdt
https://textbookfull.com/product/computer-algebra-in-scientific-
computing-20th-international-workshop-casc-2018-lille-france-
september-17-21-2018-proceedings-vladimir-p-gerdt/
Computer Algebra in Scientific Computing 18th

International Workshop CASC 2016 Bucharest Romania
September 19 23 2016 Proceedings 1st Edition Vladimir
P. Gerdt
computing-18th-international-workshop-casc-2016-bucharest-
romania-september-19-23-2016-proceedings-1st-edition-vladimir-p-
gerdt/
Computer Algebra in Scientific Computing 22nd

International Workshop CASC 2020 Linz Austria September
14 18 2020 Proceedings François Boulier
computing-22nd-international-workshop-casc-2020-linz-austria-
september-14-18-2020-proceedings-francois-boulier/
Intelligent Systems 2014 Proceedings of the 7th IEEE

International Conference Intelligent Systems IS 2014
September 24 26 2014 Warsaw Poland Volume 1
Mathematical Foundations Theory Analyses 1st Edition P.
Angelov
https://textbookfull.com/product/intelligent-
systems-2014-proceedings-of-the-7th-ieee-international-
conference-intelligent-systems-
Computer Safety Reliability and Security 33rd
International Conference SAFECOMP 2014 Florence Italy
September 10 12 2014 Proceedings 1st Edition Andrea
Bondavalli
https://textbookfull.com/product/computer-safety-reliability-and-
security-33rd-international-conference-safecomp-2014-florence-
italy-september-10-12-2014-proceedings-1st-edition-andrea-
bondavalli/
Discrete Geometry for Computer Imagery 18th IAPR

International Conference DGCI 2014 Siena Italy
September 10 12 2014 Proceedings 1st Edition Elena
Barcucci
https://textbookfull.com/product/discrete-geometry-for-computer-
imagery-18th-iapr-international-conference-dgci-2014-siena-italy-
september-10-12-2014-proceedings-1st-edition-elena-barcucci/
Principles and Practice of Constraint Programming 20th

International Conference CP 2014 Lyon France September
8 12 2014 Proceedings 1st Edition Barry O’Sullivan
(Eds.)
https://textbookfull.com/product/principles-and-practice-of-
constraint-programming-20th-international-conference-
cp-2014-lyon-france-september-8-12-2014-proceedings-1st-edition-
barry-osullivan-eds/
Parameterized and Exact Computation 9th International

Symposium IPEC 2014 Wroclaw Poland September 10 12 2014
Revised Selected Papers 1st Edition Marek Cygan
https://textbookfull.com/product/parameterized-and-exact-
computation-9th-international-symposium-ipec-2014-wroclaw-poland-
september-10-12-2014-revised-selected-papers-1st-edition-marek-
cygan/
Augmented Environments for Computer Assisted

Interventions 9th International Workshop AE CAI 2014
Held in Conjunction with MICCAI 2014 Boston MA USA
September 14 2014 Proceedings 1st Edition Cristian A.
Linte
https://textbookfull.com/product/augmented-environments-for-
computer-assisted-interventions-9th-international-workshop-ae-
cai-2014-held-in-conjunction-with-miccai-2014-boston-ma-usa-
Vladimir P. Gerdt Wolfram Koepf
Werner M. Seiler Evgenii V. Vorozhtsov (Eds.)
LNCS 8660
Computer Algebra
in Scientific Computing
16th International Workshop, CASC 2014
Warsaw, Poland, September 8–12, 2014
Proceedings
123
Lecture Notes in Computer Science 8660
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen
Editorial Board
David Hutchison
Lancaster University, UK
Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA
Josef Kittler
University of Surrey, Guildford, UK
Jon M. Kleinberg
Cornell University, Ithaca, NY, USA
Alfred Kobsa
University of California, Irvine, CA, USA
Friedemann Mattern
ETH Zurich, Switzerland
John C. Mitchell
Stanford University, CA, USA
Moni Naor
Weizmann Institute of Science, Rehovot, Israel
Oscar Nierstrasz
University of Bern, Switzerland
C. Pandu Rangan
Indian Institute of Technology, Madras, India
Bernhard Steffen
TU Dortmund University, Germany
Demetri Terzopoulos
University of California, Los Angeles, CA, USA
Doug Tygar
University of California, Berkeley, CA, USA
Gerhard Weikum
Max Planck Institute for Informatics, Saarbruecken, Germany
Vladimir P. Gerdt Wolfram Koepf
Werner M. Seiler Evgenii V. Vorozhtsov (Eds.)
Computer Algebra
in Scientific Computing
16th International Workshop, CASC 2014
Warsaw, Poland, September 8-12, 2014
Proceedings
13
Volume Editors
Vladimir P. Gerdt
Joint Institute of Nuclear Research
Laboratory of Information Technologies (LIT)
Dubna, Russia
E-mail: gerdt@jinr.ru
Wolfram Koepf
Universität Kassel, Institut für Mathematik
Kassel, Germany
E-mail: koepf@mathematik.uni-kassel.de
Werner M. Seiler
Universität Kassel, Institut für Mathematik
Kassel, Germany
E-mail: seiler@mathematik.uni-kassel.de
Evgenii V. Vorozhtsov
Russian Academy of Sciences
Institute of Theoretical and Applied Mechanics
Novosibirsk, Russia
E-mail: vorozh@itam.nsc.ru
ISSN 0302-9743 e-ISSN 1611-3349

ISBN 978-3-319-10514-7 e-ISBN 978-3-319-10515-4
DOI 10.1007/978-3-319-10515-4
Springer Cham Heidelberg New York Dordrecht London
Library of Congress Control Number: 2014946203
LNCS Sublibrary: SL 1 – Theoretical Computer Science and General Issues

© Springer International Publishing Switzerland 2014
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed. Exempted from this legal reservation are brief excerpts in connection
with reviews or scholarly analysis or material supplied specifically for the purpose of being entered and
executed on a computer system, for exclusive use by the purchaser of the work. Duplication of this publication
or parts thereof is permitted only under the provisions of the Copyright Law of the Publisher’s location,
in its current version, and permission for use must always be obtained from Springer. Permissions for use
may be obtained through RightsLink at the Copyright Clearance Center. Violations are liable to prosecution
under the respective Copyright Law.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
While the advice and information in this book are believed to be true and accurate at the date of publication,
neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or
omissions that may be made. The publisher makes no warranty, express or implied, with respect to the
material contained herein.
Typesetting: Camera-ready by author, data conversion by Scientific Publishing Services, Chennai, India
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)
Preface
Since their start in 1998, the International Workshops on Computer Algebra in

Scientific Computing have provided excellent international forums for sharing
knowledge and results in computer algebra (CA) methods, systems, and CA
applications in scientific computing. The aim of the 16th CASC Workshop was
to provide a platform to the researchers and practitioners from both academia
as well as industry to meet and share cutting-edge development in the field.
Last autumn, Ernst Mayr, a co-founder of CASC and co-chair of the CASC
conference series since its foundation in 1997 and holding the first conference in
April 1998 retired from his positions as General Chair and Proceedings Editor
of CASC. He decided to step down to spend more time on his other professional
activities and with his family. The importance of his contributions to establish-
ing CASC as a renowned conference series and to its evident progression from
1998 until today cannot be overestimated. When Ernst informed his co-chairs
about his retirement plans, he suggested Werner M. Seiler as his successor and
expressed his readiness to assist in future CASC activities. All chairs of CASC
wish Ernst good health and continuing success in his professional work and hope
to see him again at many further CASC conferences. Werner gladly accepted the
honorable invitation to take over as General Chair and Co-editor of the proceed-
ings. He will be supported in his work by Andreas Weber, who has agreed to
assume the new position of Publicity Chair of CASC. They will try to keep
CASC in the great spirit created by Ernst.
This year the CASC conference was held in Poland, where research in the
field of CA becomes more and more popular. At present, research on the develop-
ment and application of methods, algorithms, and programs of CA is performed
at universities of Bialystok, Kraków, Lublin, Siedlce, Toruń, Warsaw, Zielona
Góra and others. In many universities, for example, AGH University of Science
and Technology in Kraków, Nicolaus Copernicus University in Toruń, Warsaw
University and Warsaw University of Life Sciences, regular courses have been
introduced for students on symbolic computation and application of computer
algebra to the theory of differential equations, dynamical systems, coding theory,
and cryptography. In connection with the above, it was decided to hold the 16th
CASC Workshop in Warsaw to draw the attention of young researchers to this
interesting and important field of applied mathematics and computer science.
This volume contains 33 full papers submitted to the workshop by the par-
ticipants and accepted by the Program Committee after a thorough reviewing
process. Additionally, the volume includes two invited talks.
Studies in polynomial algebra are represented by contributions devoted to
factoring sparse bivariate polynomials using the priority queue, the construc-
tion of irreducible polynomials by using the Newton index, real polynomial
root finding by means of matrix and polynomial iterations, application of the
VI Preface
eigenvalue method with symmetry for solving polynomial systems arising in the
vibration analysis of mechanical structures with symmetry properties, applica-
tion of Gröbner systems for computing the (absolute) reduction number of poly-
nomial ideals, the application of cylindrical algebraic decomposition for solving
the quantifier elimination problems, certification of approximate roots of overde-
termined and singular polynomial systems via the recovery of an exact rational
univariate representation from approximate numerical data, and new parallel
algorithms for operations on univariate polynomials (multi-point evaluation, in-
terpolation) based on subproduct tree techniques.
Several papers are devoted to using CA for the investigation of various math-
ematical and applied topics related to ordinary differential equations (ODEs):
application of CAS Mathematica for the investigation of movable singularities
of the complex valued solutions of ODEs, the decidability problem for linear
ODE systems with variable coefficients, conversion of nonlinear ODEs to inte-
gral equations for the purpose of parameter estimation.
The invited talk by L. Plaskota deals with the information-based complexity
(IBC) as a branch of computational complexity that studies continuous prob-
lems, for which available information is partial, noisy, and priced. The basic ideas
of IBC are presented, and some sample results on optimal algorithms, complexity,
and tractability of such problems are given. The focus is on numerical integration
of univariate and multivariate functions.
A number of papers deal with applications of symbolic and symbolic-numeric
computations for investigating and solving partial differential equations (PDEs)
in mathematical physics: symbolic solution of algebraic PDEs, testing unique-
ness of analytic solutions of PDE with boundary conditions, construction of
high-order difference schemes for solving the Poisson equation, symbolic-numeric
solution of the Burgers and Korteweg–de Vries–Burgers equations at very high
Reynolds numbers, and derivation of new analytic solutions of the PDEs gov-
erning the motion of a special Cosserat rod-like fiber.
Application of symbolic and symbolic-numeric algorithms in mechanics and
physics is represented by the following themes: analytic calculations in Maple
for modeling smoothly irregular integrated optical waveguide structures, the
integrability of evolutionary equations in the restricted three-body problem with
variable masses, quantum tunneling problem of diatomic molecule through
repulsive barriers, symbolic-numeric solution of the parametric self-adjoint Sturm–
Liouville problem, obtaining new invariant manifolds in the classic and general-
ized Goryachev–Chaplygin problem of rigid body dynamics.
The invited talk by G. Regensburger, article written jointly with S. Müller,
addresses a recent extension of chemical reaction network theory (CRNT), called
generalized mass-action systems, where reaction rates are allowed to be power-
laws in the concentrations. In particular, the kinetic orders (the real exponents)
can differ from the corresponding stoichiometric coefficients. As with mass action
kinetics, complex balancing equilibria (determined by the graph Laplacian of the
underlying network) can be characterized by binomial equations and parameter-
ized by monomials. However, uniqueness and existence for all rate constants and
Preface VII
initial conditions additionally depend on sign vectors of the stoichiometric and

kinetic-order subspaces. This leads to a generalization of Birch’s theorem, which
is robust with respect to certain perturbations in the exponents. Finally, the
occurrence of multiple complex balancing equilibria is discussed. The presenta-
tion is focused on a constructive characterization of positive real solutions to
generalized polynomial equations with real and symbolic exponents.
The other topics include the application of the CAS GAP (Groups, Al-
gorithms, Programming) for construction of directed strongly regular graphs,
the application of CAS Mathematica for derivation of a four-points piecewise-
quadratic interpolant, the efficient calculation of the determinant of a generalized
Vandermonde matrix with CAS Mathematica, the solution of systems of linear
inequalities by combining virtual substitution with learning strategies with the
use of the REDUCE package Redlog, finding a generic position for an al-
gebraic space curve, the application of CAS Relview for solving problems of
voting systems, computation of the truncated annihilating ideals for algebraic
local cohomology class attached to isolated hypersurface singularities, optimal
estimations of Seiffert-type means by some special Gini means, solving para-
metric sparse linear systems by local block triangularization, computation of
the topology of an arrangement of algebraic plane curve defined by implicit and
parametric equations, the use of CASs Maxima and Mathematica for studying
the coherence and large-scale pattern formation in coupled logistic-map lattices,
and enumeration of all Schur rings over the groups of orders up to 62.
Our particular thanks are due to the dean of the Faculty of Applied Infor-
matics and Mathematics, Arkadiusz Orlowski, and the members of the CASC
2014 local Organizing Committee in Warsaw (Warsaw University of Life Sci-
ences), i.e., Alexander Prokopenya (chair of the local Organizing Committee),
and Ryszard Kozera, Luiza Ochnio, and Artur Wilinski, who have ably handled
all the local arrangements in Warsaw. Furthermore, we want to thank all the
members of the Program Committee for their thorough work. Finally we are
grateful to W. Meixner for his technical help in the preparation of the camera-
ready manuscript for this volume and the design of the conference poster.
July 2014 V.P. Gerdt

W. Koepf
W.M. Seiler
E.V. Vorozhtsov
Organization
CASC 2014 was organized jointly by the Institute of Mathematics at the Uni-
versity of Kassel, Kassel, Germany, and the Faculty of Applied Informatics and
Mathematics, Warsaw University of Life Sciences (SGGW), Warsaw, Poland.
Workshop General Chairs

Vladimir P. Gerdt (Dubna)
Werner M. Seiler (Kassel)
Program Committee Chairs

Wolfram Koepf (Kassel)
Evgenii V. Vorozhtsov (Novosibirsk)
Program Committee
François Boulier (Lille) Georg Regensburger (Linz)
Hans-Joachim Bungartz (Munich) Eugenio Roanes-Lozano (Madrid)
Jin-San Cheng (Beijing) Valery Romanovski (Maribor)
Victor F. Edneral (Moscow) Markus Rosenkranz (Canterbury)
Dima Grigoriev (Lille) Doru Stefanescu (Bucharest)
Jaime Gutierrez (Santander) Thomas Sturm (Saarbrücken)
Sergey A. Gutnik (Moscow) Jan Verschelde (Chicago)
Jeremy Johnson (Philadelphia) Stephen M. Watt
Victor Levandovskyy (Aachen) (W. Ontario, Canada)
Marc Moreno Maza (London, Canada) Andreas Weber (Bonn)
Alexander Prokopenya (Warsaw) Kazuhiro Yokoyama (Tokyo)
Additional Reviewers
Parisa Alvandi Jorge Caravantes
Hirokazu Anai Francisco-Jesus Castro-Jimenez
Alexander Batkhin Changbo Chen
Johannes Blömer Colin Denniston
Charles Bouillaguet Gema M. Diaz-Toca
Jürgen Bräckle Matthew England
X Organization
Ruyong Feng Bernard Mourrain

Mario Fioravanti Hirokazu Murao
Mark Giesbrecht Ioana Necula
Joris van der Hoeven Masayuki Noro
Martin Horvat Alina Ostafe
Silvana Ilie Alfredo Parra
Fredrik Johansson Roman Pearce
Wolfram Kahl Marko Petkovšek
Manuel Kauers Daniel Robertz
Irina Kogan Vikram Sharma
Marek Kosta Takeshi Shimoyama
Ryszard Kozera Kristof Unterweger
Dmitry Kulyabov Luis Verde-Star
François Lemaire Konrad Waldherr
Wei Li Uwe Waldmann
Gennadi Malaschonok Uli Walther
Thanos Manos Hitoshi Yanami
Michael Monagan Hangzhou Zhejiang
Local Organization
Alexander Prokopenya (Chair) Luiza Ochnio
Ryszard Kozera Artur Wilinski
Publicity Chair
Andreas Weber (Bonn)
Website
http://wwwmayr.in.tum.de/CASC2014/
Table of Contents
Computable Infinite Power Series in the Role of Coefficients of Linear

Differential Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Sergei A. Abramov and Moulay A. Barkatou
Relation Algebra, RelView, and Plurality Voting . . . . . . . . . . . . . . . . . . . 13
Rudolf Berghammer
An Algorithm for Converting Nonlinear Differential Equations to
Integral Equations with an Application to Parameter Estimation from
Noisy Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
François Boulier, Anja Korporal, François Lemaire,
Wilfrid Perruquetti, Adrien Poteaux, and Rosane Ushirobira
Truth Table Invariant Cylindrical Algebraic Decomposition by Regular
Chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Russell Bradford, Changbo Chen, James H. Davenport,
Matthew England, Marc Moreno Maza, and David Wilson
Computing the Topology of an Arrangement of Implicit and Parametric
Curves Given by Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Jorge Caravantes, Mario Fioravanti, Laureano Gonzalez–Vega, and
Ioana Necula
Finding a Deterministic Generic Position for an Algebraic Space
Curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Jin-San Cheng and Kai Jin
Optimal Estimations of Seiffert-Type Means by Some Special Gini
Means . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Iulia Costin and Gheorghe Toader
CAS Application to the Construction of High-Order Difference Schemes
for Solving Poisson Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Grigoriy M. Drozdov and Vasily P. Shapeev
On Symbolic Solutions of Algebraic Partial Differential Equations . . . . . . 111
Georg Grasegger, Alberto Lastra, J. Rafael Sendra, and
Franz Winkler
Eigenvalue Method with Symmetry and Vibration Analysis of Cyclic
Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Aurelien Grolet, Philippe Malbos, and Fabrice Thouverez
XII Table of Contents
Symbolic-Numerical Solution of Boundary-Value Problems with

Self-adjoint Second-Order Differential Equation Using the Finite
Element Method with Interpolation Hermite Polynomials . . . . . . . . . . . . . 138
Alexander A. Gusev, Ochbadrakh Chuluunbaatar,
Sergue I. Vinitsky, Vladimir L. Derbov,
Andrzej Góźdź, Luong Le Hai, and Vitaly A. Rostovtsev
Sporadic Examples of Directed Strongly Regular Graphs Obtained By

Computer Algebra Experimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Štefan Gyürki and Mikhail Klin
On the Parallelization of Subproduct Tree Techniques Targeting

Many-Core Architectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Sardar Anisul Haque, Farnam Mansouri, and Marc Moreno Maza
Deterministically Computing Reduction Numbers of Polynomial

Ideals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
Amir Hashemi, Michael Schweinfurter, and Werner M. Seiler
A Note on Global Newton Iteration Over Archimedean and

Non-Archimedean Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
Jonathan D. Hauenstein, Victor Y. Pan, and Agnes Szanto
Invariant Manifolds in the Classic and Generalized Goryachev–

Chaplygin Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
Valentin Irtegov and Tatyana Titorenko
Coherence and Large-Scale Pattern Formation in Coupled Logistic-Map

Lattices via Computer Algebra Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
Maciej Janowicz and Arkadiusz Orlowski
On the Computation of the Determinant of a Generalized Vandermonde

Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
Takuya Kitamoto
Towards Conflict-Driven Learning for Virtual Substitution . . . . . . . . . . . . 256

Konstantin Korovin, Marek Košta, and Thomas Sturm
Sharpness in Trajectory Estimation for Planar Four-points

Piecewise-Quadratic Interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
Ryszard Kozera, Lyle Noakes, and Piotr Szmielew
Scheme for Numerical Investigation of Movable Singularities of the

Complex Valued Solutions of Ordinary Differential Equations . . . . . . . . . . 286
Radoslaw Antoni Kycia
Generalized Mass-Action Systems and Positive Solutions of Polynomial

Equations with Real and Symbolic Exponents (Invited Talk) . . . . . . . . . . 302
Stefan Müller and Georg Regensburger
Table of Contents XIII
Lie Symmetry Analysis for Cosserat Rods . . . . . . . . . . . . . . . . . . . . . . . . . . . 324

Dominik L. Michels, Dmitry A. Lyakhov, Vladimir P. Gerdt,
Gerrit A. Sobottka, and Andreas G. Weber
Real Polynomial Root-Finding by Means of Matrix and Polynomial

Iterations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
Victor Y. Pan
On Testing Uniqueness of Analytic Solutions of PDE with Boundary

Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
Sergey V. Paramonov
Continuous Problems: Optimality, Complexity, Tractability

(Invited Talk) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
Leszek Plaskota
On Integrability of Evolutionary Equations in the Restricted

Three-Body Problem with Variable Masses . . . . . . . . . . . . . . . . . . . . . . . . . . 373
Alexander N. Prokopenya, Mukhtar Zh. Minglibayev, and
Baglan A. Beketauov
Factoring Sparse Bivariate Polynomials Using the Priority Queue . . . . . . 388

Fatima K. Abu Salem, Khalil El-Harake, and Karl Gemayel
Solving Parametric Sparse Linear Systems by Local Blocking . . . . . . . . . . 403

Tateaki Sasaki, Daiju Inaba, and Fujio Kako
Analytical Calculations in Maple to Implement the Method of

Adiabatic Modes for Modelling Smoothly Irregular Integrated Optical
Waveguide Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
Leonid A. Sevastyanov, Anton L. Sevastyanov, and
Anastasiya A. Tyutyunnik
CAS Application to the Construction of the Collocations and Least

Residuals Method for the Solution of the Burgers and Korteweg–de
Vries–Burgers Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432
Vasily P. Shapeev and Evgenii V. Vorozhtsov
An Algorithm for Computing the Truncated Annihilating Ideals for an

Algebraic Local Cohomology Class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
Takafumi Shibuta and Shinichi Tajima
Applications of the Newton Index to the Construction of Irreducible

Polynomials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 460
Doru Ştefănescu
XIV Table of Contents
Symbolic-Numeric Algorithm for Solving the Problem of Quantum

Tunneling of a Diatomic Molecule through Repulsive Barriers . . . . . . . . . . 472
Sergue I. Vinitsky, Alexander A. Gusev, Ochbadrakh Chuluunbaatar,
Luong Le Hai, Andrzej Góźdź, Vladimir L. Derbov, and
Pavel Krassovitskiy
Enumeration of Schur Rings Over Small Groups . . . . . . . . . . . . . . . . . . . . . 491

Matan Ziv-Av
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501

Computable Infinite Power Series in the Role
of Coefficients of Linear Differential Systems
Sergei A. Abramov1, and Moulay A. Barkatou2

1
Computing Centre of the Russian Academy of Sciences, Vavilova, 40,
Moscow 119333, Russia
sergeyabramov@mail.ru
2
Institut XLIM, Département Mathématiques et Informatique,
Université de Limoges; CNRS, 123, Av. A. Thomas,
87060 Limoges cedex, France
moulay.barkatou@unilim.fr
Abstract. We consider linear ordinary differential systems over a dif-

ferential field of characteristic 0. We prove that testing unimodularity
and computing the dimension of the solution space of an arbitrary sys-
tem can be done algorithmically if and only if the zero testing problem
in the ground differential field is algorithmically decidable. Moreover, we
consider full-rank systems whose coefficients are computable power series
and we show that, despite the fact that such a system has a basis of for-
mal exponential-logarithmic solutions involving only computable series,
there is no algorithm to construct such a basis.
1 Introduction
Linear ordinary differential systems with variable coefficients appear in various
areas of mathematics. Power series are very important objects in the represen-
tation of the solutions of such systems as well as of the systems themselves. The
representation of infinite series lies at the core of computer algebra. A general
formula that expresses the coefficients of a series is not always available and may
even not exist. One natural way to represent the series is the algorithmic one,
i.e., providing an algorithm which computes its coefficients. Such algorithmic
representation of a concrete series is not, of course, unique. This non-uniqueness
is one of the reasons for undecidability of the zero testing problem for such
computable series.
At first glance, it may seem that if we cannot decide algorithmically whether
a concrete coefficient of a system is zero or not, then we will not be able to solve
any more or less interesting problem related to the search of solutions. However,
this is not completely right: at least, if we know in advance that the system is of
full rank then some of the problems can still be solved. For example, we can find

Supported in part by the Russian Foundation for Basic Research, project no. 13-01-
00182-a. The first author thanks also Department of Mathematics and Informatics
of XLIM Institute of Limoges University for the hospitality during his visits.
V.P. Gerdt et al. (Eds.): CASC Workshop 2014, LNCS 8660, pp. 1–12, 2014.

c Springer International Publishing Switzerland 2014
2 S.A. Abramov and M.A. Barkatou
Laurent series [3] and regular [6] solutions. Some non-trivial characteristics can be
computed as well, e.g., the so called “width” of the system [3]. Nevertheless, many
of the problems are undecidable. For example, we cannot answer algorithmically
the following question: does a given full-rank system with power series coefficients
have a formal exponential-logarithmic solution which is not regular? We prove this
undecidability in the present paper. It is also shown that if exponential-logarithmic
solutions of a given full-rank system exist then there exists a basis of the space of
those solutions such that all the series which appear in the elements of the basis are
computable; the exact formulation is given in Proposition 7 of this paper.
So, we know that there exists a basis of the solution space which consists
of computable objects, but we are not able to find this basis algorithmically.
This is analogous to some facts of constructive mathematical analysis. In fact,
the notion of a constructive real number (computable point) is fundamental
in that discipline: “... an algorithm which finds the zeros of any alternating,
continuous, computable function is impossible. At the same time, there cannot
be a computable function that assumes values of different signs at the ends of
a given interval and does not vanish at any computable point of this interval
(a priori, it is impossible to rule out the existence of computable alternating
functions whose zeros are all ‘noncomputable’). These results are due to Tseitin
[21] ...” ([14, p. 5], see also [16, §24]).
We prove in the same direction that testing unimodularity, i.e., the invertibil-
ity of the corresponding operator and computing the dimension of the solution
space of an arbitrary system can be done algorithmically if and only if the zero
testing problem in the ground differential field is algorithmically decidable. As
a consequence, these problems are undecidable when the coefficients are power
series or Laurent series which are represented by arbitrary algorithms.
If the algorithmic way of series representation is used then some of the prob-
lems related to linear ordinary systems are decidable while others are not. Note
that the above mentioned algorithms for finding Laurent series solutions and reg-
ular solutions are implemented in Maple [23]. The implementation is described
in [3,6] and, is available at http://www.ccas.ru/ca/doku.php/eg.
The rest of the paper is organized as follows: After stating some prelimi-
naries in Section 2, we give in Section 3 a review of some results related to
systems whose coefficients belong to a field K of characteristic zero. The field K
is supposed to be a constructive differential field, i.e., there exist algorithms for
the field operations, differentiation, and for zero testing. The problems that are
listed in Section 3 can be solved algorithmically. On the other hand, we show
in Section 4 that the same problems are algorithmically undecidable, if the field
K is semi-constructive, i.e., there exist algorithms for the field operations and
differentiation but there is no algorithm for zero testing. Finally, we consider in
Section 5 semi-constructive fields of computable formal Laurent series in the role
of coefficient field of systems of linear ordinary differential systems.
The results of this paper supplement known results on the zero testing problem
and some algorithmically undecidable problems related to differential equations
(see, e.g., [10], [13]).
Computable Infinite Power Series 3
2 Preliminaries
The ring of m × m matrices with entries belonging to a ring R is denoted by
Matm (R). We use the notation [M ]i,∗ , 1 i m, for the 1 × m-matrix which is
the ith row of an m × m-matrix M . The notation M T is used for the transpose
of a matrix (vector) M .
If F is a differential field with derivation ∂ then Const (F ) = {c ∈ F | ∂c = 0}
is the constant field of F .
2.1 Differential Universal and Adequate Field Extensions

Let K be a differential field of characteristic 0 with derivation ∂ = .
Definition 1. An adequate differential extension Λ of K is a differential field
extension Λ of K such that any differential system
∂y = Ay, (1)
with A ∈ Matm (K) has a solution space of dimension m in Λm over Const (Λ).
If Const (K) is algebraically closed then there exists a unique (up to a dif-
ferential isomorphism) adequate differential extension Λ such that Const (Λ) =
Const (K) which is called the universal differential field extension of K [18,
Sect. 3.2]. For any differential field K of characteristic 0 there exists a differ-
ential extension whose constant field is algebraically closed. Indeed, this is the
algebraic closure K̄ with the derivation obtained by extending the derivation of
K in the natural way. In this case, Const (K̄) = Const (K) (see [18, Exercises
1.5, 2:(c),(d)]). Existence of the universal differential extension for K̄ implies
that there exists an adequate differential extension for K, i.e., for an arbitrary
differential field of characteristic zero.
In the sequel, we denote by Λ a fixed adequate differential extension of K,
and we suppose that the vector solutions of systems in the form (2) lie in Λm .
In addition to the first-order systems of the form (1), we also consider the
differential systems of arbitrary order r 1. Each of these systems can be
represented, e.g., in the form
Ar y (r) + Ar−1 y (r−1) + · · · + A0 y = 0, (2)
where the matrices A0 , A1 , . . . , Ar belong to Matm (K), m 1, and Ar (the

leading matrix of the system) is non-zero. The system (2) can be written as
L(y) = 0 where
L = Ar ∂ r + Ar−1 ∂ r−1 + · · · + A0 . (3)
The number r is the order of L (we write r = ord L). The operator (3) can be
alternatively represented as a matrix in Matm (K[∂]):
⎛ ⎞
L11 . . . L1m
⎝... ... ... ⎠, (4)
Lm1 . . . Lmm
Lij ∈ K [∂], i, j = 1, . . . , m, with maxi,j ord Lij = r. We say that the operator
L ∈ Matm (K [∂]) (as well as the system L(y) = 0) is of full rank, if the rows
(Li1 , . . . , Lim ), i = 1, . . . , m, of matrix (4) are linearly independent over K [∂].
The matrix Ar is the leading matrix of both the system L(y) = 0 and operator
L, regardless of representation form.
2.2 Universal Differential Extension of Formal Laurent Series Field
Let K0 be a subfield of the complex number field C and K be the field K0 ((x))
of formal Laurent series with coefficients in K0 , equipped with the derivation
d
∂ = dx . As it is well known [20, Sect. 110], if K0 is algebraically closed then the
universal differential field extension Λ is the quotient field of the ring generated
by expressions of the form
eP (x) xγ (ψ0 + ψ1 ln x + · · · + ψs (ln x)s ), (5)
where in any such expression
– P (x) ∈ K0 [x−1/q ], q is a positive integer,

– γ ∈ K0 ,
– s is a non-negative integer and
ψj ∈ K0 [[x1/q ]], (6)
j = 0, 1, . . . , s.
In fact, system (1) has m linearly independent solutions b1 (x), . . . , bm (x) such
that
bi (x) = ePi (x) xγi Ψi (x), (7)
where the factor ePi (x) xγi is common for all components of bi , and
γi ∈ K0 , qi is a positive integer, Pi (x) ∈ K0 [x−1/qi ], Ψi (x) ∈ K0m [[x1/qi ]][ln x],
i = 1, . . . , m.
Definition 2. Solutions of the form (7) will be called (formal) exponential-log-

arithmic solutions. If q = 1 and P (x) = 0 then the solutions (7) are called
regular.
Remark 1. If K0 is not algebraically closed then there exists a simple algebraic

extension K1 of K0 (specific for each system) such that system (1) has m linearly
independent solutions of the form (7) with γi ∈ K1 , Pi (x) ∈ K1 [x−1/qi ], Ψi (x) ∈
K1m [[x1/qi ]][ln x], i = 1, . . . , m.
2.3 Row Frontal Matrix and Row Order

Let a full-rank operator L ∈ Matm (K[∂]) be of the form (3). If 1 i m then
define αi (L) as the biggest integer k, 0 k r, such that [Ak ]i,∗ is a nonzero
row. The matrix M ∈ Matm (K) such that [M ]i,∗ = [Aαi (L) ]i,∗ , i = 1, . . . , m, is
the row frontal matrix of L. The vector (α1 (L), . . . , αm (L)) is the row order of L.
We will write simply (α1 , . . . , αm ), when it is clear which operator is considered.
Definition 3. An operator U ∈ Matm (K[∂]) is unimodular (or invertible) if
there exists Ū ∈ Matm (K[∂]) such that Ū U = U Ū = Im . An operator in
Matm (K[∂]) is row reduced if its row frontal matrix is invertible.
The following proposition is a consequence of [9, Thm. 2.2]:
Proposition 1. Let L ∈ Matm (K [∂]) then there exist U, L̆ ∈ Matm (K [∂]) such
that U is unimodular and L̆ defined by
L̆ = U L (8)
and represented in the form (4), has k zero rows, where 0 k m, and the row
frontal matrix of L̆ is of rank m − k over K. The operator L is of full rank if
and only if k = 0, and in this case the operator L̆ in (8) is row reduced.
We will say that the system (2) is unimodular whenever the corresponding
matrix (4) is.
3 When K Is a Constructive Field

Definition 4. A ring (field) K is said to be constructive if there exist algorithms
for performing the ring (field) operations and an algorithm for zero testing in K
This definition is close to the definition of an explicit field given in [11].
Suppose that K is a constructive field. Then the proof of the already men-
tioned theorem [9, Thm. 2.2] gives an algorithm for constructing U, L̆. We will
refer to this algorithm as RR (Row-Reduction).
3.1 The Dimension of the Solution Space of a Given Full Rank

System
Proposition 2. ([1]) Let L ∈ Matm (K [∂]) be row reduced, and denote by α =
(α1 , . . . , αm ) its
row order. Then the dimension of its solution space VL is given
by: dim VL = m i=1 αi .
Hence, when the field K is constructive we can apply algorithm RR, and
compute, by Proposition 2, the dimension of the solution space of a given full-
rank system.
Note that in the case when K is the field of rational functions of x over a field
d
of characteristic zero with ∂ = dx , some inequalities close to the formula given
in Proposition 2 can be derived from the results of [12].
3.2 Recognizing the Unimodularity of an Operator and Computing

the Inverse Operator
The following property of unimodular operators is a direct result of Proposition
2.
Proposition 3. [2] Let L ∈ Matm (K [∂]) be of full rank. Then L is unimodular
if and only if dim VL = 0. Moreover, in the case when the row frontal matrix of
L is invertible, L is unimodular if and only if ord L = 0.
Algorithm RR allows one to compute a unimodular U ∈ Matm (K [∂]) such
that the operator L̆ = U L has an invertible row frontal matrix. Proposition 3
implies that L is unimodular if and only if L̆ is an invertible matrix in Matm (K).
In this case (L̆)−1 U L = Im , i.e., (L̆)−1 U is the inverse of L. Hence the following
proposition holds (taking into account Proposition 1, we need not assume that
L is of full rank):
Proposition 4. Let K be constructive and L ∈ Matm (K [∂]). One can recognize
algorithmically whether L is unimodular or not, and compute the inverse operator
if it is.
4 When the Zero Testing Problem in K Is Undecidable

It is easy to see that if the zero testing problem in K is undecidable then the
problem of recognizing whether a given L ∈ Matm (K [∂]) is of full rank is unde-
cidable. Indeed, let u ∈ K, then the operator

u∂ ∂ u 1 0 0
L= = ∂+
0 1 0 0 0 1
is of full rank if and only if u = 0, and any algorithm to recognize whether a
given L ∈ Matm (K [∂]) is of full rank can be used for zero testing in K.
Furthermore, it turns out that if the zero testing problem in K is undecidable
then even with a prior knowledge that operators under consideration are of full
rank, many questions about those operators remain undecidable.
Proposition 5. Let the zero testing problem in K be undecidable. Then for
m 2 the following problems about a full-rank operator L ∈ Matm (K [∂]) are
undecidable:
(a) computing dim VL ,
(b) testing unimodularity of L.
Proof. (a) Let u ∈ K and

u∂ + 1 ∂ u 1 1 0
L= = ∂+ . (9)
0 1 0 0 0 1
If u = 0 then L is unimodular:
−1
1 ∂ 1 −∂
=
0 1 0 1
and, therefore, dim VL = 0. If u = 0 then dim VL = 1 by Proposition 2. We have
0 if u = 0,
dim VL =
1 if u = 0.
This implies that if we have an algorithm for computing the dimension then we
have an algorithm for the zero testing problem.
(b) As we have seen the operator L of the form (9) is unimodular if and only
if u = 0.
As a consequence of Propositions 4, 5 we have the following:

Testing unimodularity and determining the dimension of the solution space of
an arbitrary full-rank system can be done algorithmically if and only if the zero
testing problem in K can be solved algorithmically.
One of the general causes of difficulties in the zero testing problem in K may
be associated with non-uniqueness of representation of the elements of K [11,
Sect. 2]. This is illustrated in Section 5.1.
5 Computable Power Series

5.1 Semi-constructive Fields
Let K be the field K0 ((x)) where K0 is a constructive field of characteristic
0. The field K contains the set K|c of computable series, whose sequences of
coefficients can be represented algorithmically. That is to say that for each series
a(x) ∈ K|c there exists an algorithm Ξa to compute the coefficient ai ∈ K0
for a given i; arbitrary algorithms which are applicable to integer numbers and
return elements of K0 are allowed. For this set to be considered as a constructive
differential subfield of K, it would be necessary to define algorithmically on K|c
d
the field operations of the field K, the unary operation dx , and a zero testing
algorithm as well. However, in accordance with the classical results of Turing
[22], we are not able to solve algorithmically the zero testing problem in K|c .
As mentioned in Section 4, the undecidability of the zero testing problem is
quite often associated with the fact that the elements of the field (or ring) under
consideration can be represented in various ways, and for some of which the test
is evident while for the others is not. This holds for K|c as well.
Remark 2. The field K|c is smaller than the field K because not every sequence
of coefficients can be represented algorithmically. Indeed, the set of elements of
K|c is countable (each of the algorithms is a finite word in some fixed alphabet)
while the cardinality of the set of elements of K is uncountable.
If the only information we possess about the elements of K|c is an algorithm

to compute their coefficients then the problem of finding the valuation of a
given a(x) ∈ K|c , val a(x), is undecidable even in the case when it is known
in advance that a(x) is not the zero series. This implies that when we work
with elements of K|c , i.e., with computable Laurent series, we cannot compute
a−1 (x) for a given non-zero a(x) ∈ K|c , since the coefficient of x−1 of the
series a (x)a−1 (x) ∈ K|c is equal to val a(x), i.e., is equal to the value that
we are not able to find algorithmically knowing only Ξa . This means that a
suitable representation has to contain some additional information besides a
corresponding algorithm. The value val a(x) cannot close the gap, since we have
no algorithm to compute the valuation of the sum of two series. However, we
can use a lower bound of the valuation instead: observe that if we know that a
series a(x) is non-zero then using a valuation lower bound we can compute the
exact value of val a(x). Thus, we can use as the representation of a(x) ∈ K|c a
pair of the form
(Ξa , μa ), (10)
where Ξa is an algorithm for computing the coefficient ai ∈ K0 for a given i, and

the integer μa is a lower bound for the valuation of a(x). A computable
Laurent
series a(x), represented by a pair of the form (10) is equal to ∞ i=μa Ξ a (i)xi
.
Of course, there exist other ways to represent computable Laurent series. For
example, one can use a pair (Ξa , pa (x)), where the algorithm Ξa represents a
power series that is the regular part of a(x) while pa (x) ∈ K0 [x−1 ] represents
explicitly its singular part. We can also represent each Laurent series as a fraction
of two power series (the latter are represented algorithmically, this is possible as
the field of Laurent series is the quotient field of the ring of power series). So a
Laurent series can be represented as a couple (a(x), b(x)) of power series with
b(x) nonzero.
We can define the field structure on K|c : all field operations can be performed
algorithmically. Since we do not have an algorithm for solving the zero testing
problem in K|c , we use for K|c the term “semi-constructive field” instead.
Definition 5. A ring (field) is semi-constructive if there are algorithms to per-

form the ring (field) operations, but there exists no algorithm to solve the zero
testing problem.
Observe that if the standard representation form is used for rational functions,
i.e., for elements in K0 (x), then the field K0 (x) is constructive.
Remark 3. Consider for the ring R = K0 [[x]] its semi-constructive sub-ring

R|c of computable power series. In this case we do not need to include a lower
bound for the valuation into a representation of a series a(x) ∈ R|c , since 0 is
such a bound.
5.2 Systems with Computable Power Series Coefficients
Below we suppose that K0 is a constructive field of characteristic 0, K = K0 ((x)),

R = K0 [[x]], and
K|c , R|c
are a semi-constructive field and, resp., a semi-constructive ring as in Section

5.1. We will consider systems of the form

d
L(y) = 0, L ∈ Matm R|c . (11)
dx
It follows from Proposition 5 that the problems (a) and (b) listed in that
proposition are undecidable if L is as in (11). At first glance, it seems that such
undecidability is mostly caused by the inability to distinguish zero and nonzero
coefficients of operators and systems. However, even if we know in advance which
of the coefficients of an operator L are null, we, nevertheless, cannot solve prob-
lems (a) and (b) of Proposition 5 algorithmically. Let u(x) ∈ R|c and
d d

(u(x)x + 1) dx + 1 dx u(x)x + 1 1 d 1 0
L= = + .
1 1 0 0 dx 1 1
For such an operator, we know in advance which of its coefficients are equal to
zero, but we do not know whether the power series u(x) is equal to zero. It is
easy to see that
0 if u(x) = 0,
dim VL =
1 if u(x) = 0.
5.3 On Formal Exponential-Logarithmic Solutions

In [3,6], it was proven that the problems of existence of Laurent series solutions
and regular solutions (see Definition 2) for a given system (11) are decidable. A
regular solution has the form xγ w(x), where γ ∈ K̄0 , and w(x) ∈ K̄0 ((x))m [ln x];
m
in the context of [6], w(x) ∈ K̄0 ((x))|c [ln x]. In those papers, it was proven
also that if non-zero Laurent series or regular solutions exist then we can con-
struct them, i.e., find a lower bound for valuations of all involved Laurent series
as well as any number of terms of the series; for regular solutions we also find the
corresponding values of γ, the degrees of polynomials in ln x etc. It was shown
also that instead of K̄0 which is the algebraic closure of K0 some simple algebraic
extension K1 of K0 may be used.
Remark 4. The power series which appear in [3,6] as coefficients of a given
system can be represented not only by algorithms as described above but also as
“black boxes”, i.e., by procedures of unknown internal form.
Proposition 6. Let m be an integer, m 2, and K0 be a constructive subfield
of C. Then for a given full-rank system of the form (11),
(i) the question whether nonzero Laurent series solutions exist as well as the
question whether nonzero regular solutions exist are algorithmically decidable;
(ii) the question whether nonzero formal exponential-logarithmic solutions ex-
ist is algorithmically undecidable;
(iii) the question whether nonzero formal exponential-logarithmic solutions
which are not regular solutions exist is algorithmically undecidable.
Proof. (i) This follows from [3,6], as it was explained in the beginning of this
section.
(ii) A given L is unimodular if and only if the system (11) has no non-zero
formal exponential-logarithmic solution, and the claim follows from Proposition
5 (problem (b)).
(iii) A full-rank operator L is evidently unimodular if and only if it has no
regular solution and no exponential-logarithmic solution which is not regular. By
(i), we can test whether the system L(y) = 0 has no regular solution. Thus, if
we are able to test whether this system has no exponential-logarithmic solution
which is not regular then we can test whether L is unimodular or not. However,
this is an undecidable problem by Proposition 5 (problem (b)).
Proposition 7. Let m be an integer number, m 2, K0 be a constructive subset
of C. Let L(y) = 0 be a full-rank system of the form (11), and d = dim VL . Then
VL has a basis b1 (x), . . . , bd (x) consisting of exponential-logarithmic solutions
such that any Ψi (x) from (7) is of the form Ψi (x) = Φi (x1/qi ) where qi is a
non-negative integer,
Φi (x) ∈ ((K1 [[x]]) |c )m [ln x], (12)
and K1 is a simple algebraic extension of K0 . In addition to (12), γi ∈ K1 ,
Pi (x) ∈ K1 [x−1/qi ], i = 1, . . . , d.
Proof. It follows from, e.g., [4,5,8], that for any operator L of full rank there
exists an operator F such that the leading matrix of F L is invertible. The system
F L(y) = 0 is equivalent to a first order system of the form y = Ay, A ∈
Matms (K((x))), s = ord F L. It is known ([7]) that for a first-order system
there exists a simple algebraic extension K1 of K0 such that those γi and the
coefficients of Pi (x) which appear in its solutions of the form (7), belong to K1 .
The field K1 is constructive since K0 is. Obviously, qi ∈ N.
The substitution qi
x = tqi , y(tqi ) = z(t)ePi (t ) ,
Pi (tqi ) ∈ K1 [1/t], into the original system L(y) = 0 transforms it into a full-rank
system which can be represented as

d
L̃(z) = 0, L̃ ∈ Matm (K1 [[t]])|c .
dt
The Laurent series that appear in the regular solutions of this new system can
be taken to be computable, as it follows from [3,6] (see the beginning of this
section).
Thus, the series that appear in the representation of solutions are computable
(Proposition 7), but we cannot find them algorithmically (Proposition 6). In fact,
Proposition 7 guarantees existence. However, the operator F mentioned therein
cannot be constructed algorithmically.
Remark 5. In the case of first-order systems of the form (1), the questions
formulated in Proposition 6(ii, iii) are decidable. This follows from the fact that
for constructing exponential-logarithmic solutions of a system of this form one

needs only a finite number of terms of the entries (which are Laurent series) of
A, and the number of those terms can be computed in advance ([15,7,17]). This
holds also for higher-order systems whose leading matrices are invertible.
It is proven ([19]) that if the dimension d of the space of exponential-logarith-
mic solutions is known in advance then the basis b1 , . . . , bd which is mentioned in
Proposition 7 can be constructed algorithmically. The corresponding algorithm
is implemented in Maple.
As we see, if the algorithmic representation of series is used and if arbitrary
algorithms representing series are admitted then some of the problems related to
linear ordinary differential systems are decidable, while others are not. There is
a subtle border between them, and a careful formulation of each of the problems
under consideration is absolutely necessary. A small change in the formulation
of a decidable problem can transform it into an undecidable one, and vice versa.
Acknowledgments. The authors are thankful to S. Maddah, M. Petkovšek, A.

Ryabenko, M. Rybowicz, S. Watt for interesting discussions, and to anonymous
referees for their useful comments.
References
1. Abramov, S.A., Barkatou, M.A.: On the dimension of solution spaces of full rank
linear differential systems. In: Gerdt, V.P., Koepf, W., Mayr, E.W., Vorozhtsov,
E.V. (eds.) CASC 2013. LNCS, vol. 8136, pp. 1–9. Springer, Heidelberg (2013)
2. Abramov, S.A., Barkatou, M.A.: On solution spaces of products of linear differential
or difference operators. ACM Communications in Computer Algebra (accepted)
3. Abramov, S.A., Barkatou, M.A., Khmelnov, D.E.: On full-rank differential systems
with power series coefficients. J. Symbolic Comput. (accepted)
4. Abramov, S.A., Khmelnov, D.E.: Desingularization of leading matrices of systems
of linear ordinary differential equations with polynomial coefficients. In: Inter-
national Conference “Differential Equations and Related Topics” Dedicated to
I.G.Petrovskii, Moscow, MSU, May 30-June 4, p. 5. Book of Abstracts (2011)
5. Abramov, S.A., Khmelnov, D.E.: On singular points of solutions of linear differen-
tial systems with polynomial coefficients. J. Math. Sciences 185(3), 347–359 (2012)
6. Abramov, S.A., Khmelnov, D.E.: Regular solutions of linear differential systems
with power series coefficients. Programming and Computer Software 40(2), 98–106
(2014)
7. Barkatou, M.A.: An algorithm to compute the exponential part of a formal fun-
damental matrix solution of a linear differential system. Applicable Algebra in
Engineering, Communication and Computing 8, 1–23 (1997)
8. Barkatou, M.A., El Bacha, C., Labahn, G., Pflügel, E.: On simultaneous row and
column reduction of higher-order linear differential systems. J. Symbolic Com-
put. 49(1), 45–64 (2013)
9. Beckermann, B., Cheng, H., Labahn, G.: Fraction-free row reduction of matrices
of Ore polynomials. J. Symbolic Comput. 41(5), 513–543 (2006)
10. Denef, J., Lipshitz, L.: Power series solutions of algebraic differential equations.
Math. Ann. 267, 213–238 (1984)
11. Frölich, A., Shepherdson, J.C.: Effective procedures in field theory. Phil. Trans. R.
Soc. Lond. 248(950), 407–432 (1956)
12. Grigoriev, D.: NC solving of a system of linear differential equations in several
unknowns. Theor. Comput. Sci. 157(1), 79–90 (1996)
13. van der Hoeven, J., Shackell, J.R.: Complexity bounds for zero-test algorithms. J.
Symbolic Comput. 41(4), 1004–1020 (2006)
14. Kushner, B.A.: Lectures on Constructive Mathematical Analysis (Translations of
Mathematical Monographs) Amer. Math. Soc. (1984)
15. Lutz, D.A., Schäfke, R.: On the identification and stability of formal invariants for
singular differential equations. Linear Algebra and Its Applications 72, 1–46 (1985)
16. Martin-Löf, P.: Notes on Constructive Mathematics. Almquist & Wiskell, Stokholm
(1970)
17. Pflügel, E.: Effective formal reduction of linear differential systems. Applicable
Algebra in Engineering, Communication and Computation 10(2), 153–187 (2000)
18. van der Put, M., Singer, M.F.: Galois Theory of Linear Differential Equations.
Grundlehren der mathematischen Wissenschaften, vol. 328. Springer, Heidelberg
(2003)
19. Ryabenko, A.: On exponential-logarithmic solutions of linear differential systems
with power series coefficients (In preparation)
20. Schlesinger, L.: Handbuch der Theorie der linearen Differentialgleichungen, vol. 1.
Teubner, Leipzig (1895)
21. Tseitin, G.S.: Mean-value Theorems in Constructive Analysis. Problems of the
Constructive Direction in Mathematics. Part 2. Constructive Mathematical Anal-
ysis. Collection of Articles: Trudy Mat. Inst. Steklov, Acad. Sci. USSR 67, 362–384
(1962)
22. Turing, A.: On computable numbers, with an application to the Entscheidungs-
problem. Proc. London Math. Soc., Series 2 42, 230–265 (1936)
23. Maple online help, http://www.maplesoft.com/support/help/
Relation Algebra, RelView, and Plurality Voting
Rudolf Berghammer
Institut für Informatik, Universität Kiel, Olshausenstraße 40, 24098 Kiel, Germany
rub@informatik.uni-kiel.de
Abstract. We demonstrate how relation algebra and a supporting tool

can be combined to solve problems of voting systems. We model plurality
voting within relation algebra and present relation-algebraic specifications
for some computational tasks. They can be transformed immediately into
the programming language of the BDD-based Computer Algebra system
RelView, such that this tool can be used to solve the problems in question
and to visualize the computed results. The approach is extremely formal,
very flexible and especially appropriate for prototyping, experimentation,
scientific research, and education.
1 Introduction
Since centuries systematic experiments are an accepted means for doing science.
In the meantime they have also become important in formal fields, such as math-
ematics, scientific computing, and theoretical computer science. Here computer
support proved to be very useful, e.g., for computing results and for discover-
ing mathematical relationships by means of visualization and animation. Used
are general Computer Algebra systems, like Maple and Mathematica, but
frequently also tools that focus on specific mathematical objects and (in many
cases: algebraic) structures and particular applications. RelView (cf. [2,12,15])
is a tool of the latter kind. It can be regarded as specific purpose Computer Alge-
bra system and its main purpose is the mechanization of heterogeneous relation
algebra in the sense of [13,14] and the visualization of relations. Computational
tasks can be expressed by relational functions and programs. These frequently
consist of only a few lines that present the relation-algebraic specification of the
notions in question. In view of efficiency the implementation of relations via bi-
nary decision diagrams (BDDs) proved to be superior to many other well-known
implementations, like Boolean matrices or successor lists. Their use in RelView
led to an amazing computational power, in particular if hard problems are to
solve and this is done by the search of a huge set of objects.
In [4,6] it is demonstrated how relation algebra and RelView can be com-
bined to solve computational problems of voting systems. The papers consider
two specific systems, known as approval voting and Condorcet voting. Our paper
constitutes a continuation of this work. We concentrate on plurality voting, a
further popular voting system that simply selects the alternatives with the high-
est number of first place votes and is widely used through the world. After the
presentation of the relation-algebraic preliminaries we model two versions of this
V.P. Gerdt et al. (Eds.): CASC Workshop 2014, LNCS 8660, pp. 13–27, 2014.

c Springer International Publishing Switzerland 2014
14 R. Berghammer
kind of voting within relation algebra. Based on these models, we then present
relation-algebraic specifications for computing dominance relations and (sets of)
winners. Next, for the second model we show how relation-algebraically to solve
hard so-called control problems. For a control of an election the assumption is
that the authority conducting the election – usually called the chair – knows all
individual preferences of the voters and tries to achieve (in case of constructive
control ) or to avoid (in case of destructive control ) win for a specific alternative
by a strategic manipulation of the set of voters and alternatives, respectively.
The chair’s knowledge of all individual preferences and the ability to manipulate
by ‘dirty tricks’ (mistimed meetings, excuses like ‘to expensive’ or ‘legally not
allowed’, etc.) are worst-case assumptions that are not entirely unreasonable in
some settings, for instance, in case of commissions of political institutions. All
relation-algebraic specifications we will present can be transformed immediately
into the programming language of the Computer Algebra system RelView such
that this tool can be used to compute results and to visualize them. Finally, we
evaluate our approach in view of its advantages and its drawbacks.
2 Relation-Algebraic Preliminaries
In this section we recall some preliminaries of (typed, heterogeneous) relation
algebra. For more details, see [13,14], for example.
Given two sets X and Y we write [X ↔ Y ] for the set of all (binary) relations
with source X and target Y , i.e., for the powerset 2X×Y . Furthermore, we write
R : X ↔ Y instead of R ∈ [X ↔ Y ] and call then X ↔ Y the type of R. If the two
sets X and Y of R’s type are finite, then we may consider R as a Boolean matrix.
Since a Boolean matrix interpretation is well suited for many purposes and also
used by the RelView tool as the main possibility to visualize relations, in the
present paper we frequently use matrix terminology and notation. In particular,
we speak about the entries, rows and columns of a relation/matrix and write
Rx,y instead of (x, y) ∈ R or x R y. We assume the reader to be familiar with the
basic operations on relations, viz. RT (transposition), R (complement), R ∪ S
(union), R ∩ S (intersection), and R; S (composition), the predicates R ⊆ S
(inclusion) and R = S (equality), and the special relations O (empty relation), L
(universal relation), and I (identity relation). In case of O, L, and I we overload
the symbols, i.e., avoid the binding of types to them.
T
By syq(R, S) := RT ; S ∩ R ; S we define the symmetric quotient of two re-
lations R : X ↔ Y and S : X ↔ Z. From this definition we get the typing
syq(R, S) : Y ↔ Z and that for all y ∈ Y and z ∈ Z it holds
syq(R, S)y,z ⇔ ∀ x ∈ X : Rx,y ↔ Sx,z . (1)
Descriptions of the form (1) with relationships expressed by logical formulae and
indices are called point-wise. Such descriptions of symmetric quotients and the
constructs we introduce now for modeling sets and direct products will play a
fundamental role in the remainder of the paper. We will use them to get relation-
algebraic (or point-free) specifications, which then can be treated by RelView.
Relation Algebra, RelView, and Plurality Voting 15
Vectors are a first well-known relational means to model sets. For the purpose
of this paper it suffices to define them as relations r : X ↔ 1 (we prefer lower
case letters in this context) with a specific singleton set 1 = {⊥} as target.
Then relational vectors can be considered as Boolean column vectors. To be
consonant with the usual notation, we always omit the second subscript, i.e.,
write rx instead of rx,⊥ . Given r : X ↔ 1 and Y ∈ 2X we then define that
r models the subset Y of X :⇔ ∀ x ∈ X : x ∈ Y ↔ rx . (2)
A point is a specific vector with precisely one 1-entry if considered as a Boolean
column vector. Consequently, it models a singleton subset {x} of X and we then
say that it models the element x of X. In conjunction with a powerset 2X we
will also use the membership relation M : X ↔ 2X , point-wisely defined by
Mx,Y :⇔ x ∈ Y, (3)
for all x ∈ X and Y ∈ 2 . The following lemma shows how symmetric quotients
X
and membership relations allow to describe a specific relationship between sets

relation-algebraically. The construction set(R) of the lemma is an instance of
the power transpose construction of [14] and will play a prominent role in the
remainder of the paper, too.
Lemma 2.1. For a given relation R : X ↔ Y and M : X ↔ 2X as a membership
relation we define
set(R) := syq(R, M) : Y ↔ 2X .
Then we have set(R)y,Z iff Z = {x ∈ X | Rx,y }, for all y ∈ Y and Z ∈ 2X .
Proof. Given arbitrary y ∈ Y and Z ∈ 2X , we get the result as follows:
set(R)y,Z ⇔ syq(R, M)y,Z
⇔ ∀ x ∈ X : Rx,y ↔ Mx,Z by (1)
⇔ ∀ x ∈ X : Rx,y ↔ x ∈ Z by (3)
⇔ Z = {x ∈ X | Rx,y }
In each of our later applications we will assume the sets in question to be finite.
On powersets 2X of finite sets X we then will use the size comparison relation
S : 2X ↔ 2X , which is point-wisely defined by
SY,Z :⇔ |Y | ≤ |Z|, (4)
for all Y, Z ∈ 2 . In the next result it is shown how to compare for a finite set
X
X the specific sets {x ∈ X | Rx,y } of Lemma 2.1 w.r.t. their sizes within the
language of relation algebra.
Lemma 2.2. Let relations Q : X ↔ Y and R : X ↔ Z be given, where X is
finite, and S : 2X ↔ 2X be a size-comparison relation. Furthermore, let
scomp(Q, R) := set(Q); ST ; set(R)T : Y ↔ Z.
Then we have scomp(Q, R)y,z iff |{x ∈ X | Qx,y }| ≥ |{x ∈ X | Rx,z }|, for all
y ∈ Y and z ∈ Z.
16 R. Berghammer
Proof. We get for arbitrary y ∈ Y and z ∈ Z the claim as follows:
scomp(Q, R)y,z
⇔ (set(Q); ST ; set(R)T )y,z
⇔ ∃ U ∈ 2X : set(Q)y,U ∧ ∃ V ∈ 2X : ST T
U,V ∧ set(R)V,z
⇔ ∃ U ∈ 2X : U = {x ∈ X | Qx,y } ∧
∃ V ∈ 2X : |U | ≥ |V | ∧ V = {x ∈ X | Rx,z } Lemma 2.1, (4)
⇔ |{x ∈ X | Qx,y }| ≥ |{x ∈ X | Rx,z }|
As a general assumption, in the remainder of this paper we always assume pairs

u ∈ X×Y to be of the form (u1 , u2 ). To model direct products X×Y relation-
algebraically, the projection relations π : X×Y ↔ X and ρ : X×Y ↔ Y have
been proven as the convenient means. They are the relational variants of the
well-known projection functions and, in point-wise definitions, given by
πu,x :⇔ u1 = x ρu,y :⇔ u2 = y, (5)
for all u ∈ X×Y , x ∈ X and y ∈ Y . Projection relations enable us to specify

the pairing operation of functional programming with relation-algebraic means.
Assuming π and ρ as above, the (right) pairing (also called fork ) of R : Z ↔ X
and S : Z ↔ Y is defined as [R, S]] := R; π T ∩ S; ρT . This leads to Z ↔ X×Y as
type of [R, S]] and to a point-wise description saying that
[R, S]]z,u ⇔ Rz,u1 ∧ Sz,u2 , (6)
for all z ∈ Z and u ∈ X×Y . Using the projection relations π : X×Y ↔ X and
ρ : X×Y ↔ Y , we also are able to define functions which establish an isomor-
phism between the Boolean lattices [X ↔ Y ] and [X×Y ↔ 1 ]. The direction from
[X ↔ Y ] to [X×Y ↔ 1 ] is given by the function vec, with vec(R) := (π; R ∩ρ); L,
and that from [X×Y ↔ 1 ] back to [X ↔ Y ] by the inverse function rel, with
rel(v) := π T ; (ρ∩v; LT ). In point-wise descriptions the definitions say that Ru1 ,u2
iff vec(R)u and vu iff rel(v)u1 ,u2 , for all u ∈ X×Y .
As from now we assume all constructions of Section 2 to be at hand. Except
the functions set, scomp, vec and rel they are available in the programming
language of RelView as or via pre-defined operations and the implementing
BDDs of relations are comparatively small. For example, a size-comparison re-
lation S : 2X ↔ 2X can be implemented by a BDD with O(|X|2 ) nodes and for
a membership relation M : X ↔ 2X even O(|X|) nodes suffice (for details, see
[11,12]). An implementation of set, scomp, vec, and rel in the tool is nothing else
than the translation of their relation-algebraic definitions into RelView-code;
the RelView-programs for set and scomp can be found in Section 5.
3 Relation-Algebraic Models of Plurality Voting

In Social Choice Theory (see e.g., [7] for an overview) an election consists of
a non-empty and finite set N of voters (agents, individuals), normally defined
Another random document with
no related content on Scribd:
sacraments and scriptures of the Persian, Jewish, Christian, “post-
Classical” and Manichæan religions.
“Space”—speaking now in the Faustian idiom—is a spiritual
something, rigidly distinct from the momentary sense-present, which
could not be represented in an Apollinian language, whether Greek
or Latin. But the created expression-space of the Apollinian arts is
equally alien to ours. The tiny cella of the early-Classical temple was
a dumb dark nothingness, a structure (originally) of perishable
material, an envelope of the moment in contrast to the eternal vaults
of Magian cupolas and Gothic naves, and the closed ranks of
columns were expressly meant to convey that for the eye at any rate
this body possessed no Inward. In no other Culture is the firm
footing, the socket, so emphasized. The Doric column bores into the
ground, the vessels are always thought of from below upward
(whereas those of the Renaissance float above their footing), and
the sculpture-schools feel the stabilizing of their figures as their main
problem. Hence in archaic works the legs are disproportionately
emphasized, the foot is planted on the full sole, and if the drapery
falls straight down, a part of the hem is removed to show that the
foot is standing. The Classical relief is strictly stereometrically set on
a plane, and there is an interspace between the figures but no depth.
A landscape of Claude Lorrain, on the contrary, is nothing but space,
every detail being made to subserve its illustration. All bodies in it
possess an atmospheric and perspective meaning purely as carriers
of light and shade. The extreme of this disembodiment of the world
in the service of space is Impressionism. Given this world-feeling,
the Faustian soul in the springtime necessarily arrived at an
architectural problem which had its centre of gravity in the spatial
vaulting-over of vast, and from porch to choir dynamically deep,
cathedrals. This last expressed its depth-experience. But with it was
associated, in opposition to the cavernous Magian expression-
space,[190] the element of a soaring into the broad universe. Magian
roofing, whether it be cupola or barrel-vault or even the horizontal
baulk of a basilica, covers in. Strzygowski[191] has very aptly
described the architectural idea of Hagia Sophia as an introverted
Gothic striving under a closed outer casing. On the other hand, in the
cathedral of Florence the cupola crowns the long Gothic body of
1367, and the same tendency rose in Bramante’s scheme for St.
Peter’s to a veritable towering-up, a magnificent “Excelsior,” that
Michelangelo carried to completion with the dome that floats high
and bright over the vast vaulting. To this sense of space the
Classical opposes the symbol of the Doric peripteros, wholly
corporeal and comprehensible in one glance.
The Classical Culture begins, then, with a great renunciation. A
rich, pictorial, almost over-ripe art lay ready to its hand. But this
could not become the expression of the young soul, and so from
about 1100 B.C. the harsh, narrow, and to our eyes scanty and
barbaric, early-Doric geometrical style appears in opposition to the
Minoan.[192] For the three centuries which correspond to the flowering
of our Gothic, there is no hint of an architecture, and it is only at
about 650 B.C., “contemporarily” with Michelangelo’s transition into
the Baroque, that the Doric and Etruscan temple-type arises. All
“Early” art is religious, and this symbolic Negation is not less so than
the Egyptian and the Gothic Affirmation. The idea of burning the
dead accords with the cult-site but not with the cult-building; and the
Early Classical religion which conceals itself from us behind the
solemn names of Calchas, Tiresias, Orpheus and (probably)
Numa[193] possessed for its rites simply that which is left of an
architectural idea when one has subtracted the architecture, viz., the
sacred precinct. The original cult-plan is thus the Etruscan templum,
a sacred area merely staked off on the ground by the augurs with an
impassable boundary and a propitious entrance on the East side.[194]
A “templum” was created where a rite was to be performed or where
the representative of the state authority, senate or army, happened
to be. It existed only for the duration of its use, and the spell was
then removed. It was probably only about 700 B.C. that the Classical
soul so far mastered itself as to represent this architectural Nothing
in the sensible form of a built body. In the long run the Euclidean
feeling proved stronger than the mere antipathy to duration.
Faustian architecture, on the contrary, begins on the grand scale
simultaneously with the first stirrings of a new piety (the Cluniac
reform, c. 1000) and a new thought (the Eucharistic controversy
between Berengar of Tours and Lanfranc 1050),[195] and proceeds at
once to plans of gigantic intention; often enough, as in the case of
Speyer, the whole community did not suffice to fill the cathedral,[196]
and often again it proved impossible to complete the projected
scheme. The passionate language of this architecture is that of the
poems too.[197] Far apart as may seem the Christian hymnology of
the south and the Eddas of the still heathen north, they are alike in
the implicit space-endlessness of prosody, rhythmic syntax and
imagery. Read the Dies Iræ together with the Völuspá,[198] which is
little earlier; there is the same adamantine will to overcome and
break all resistances of the visible. No rhythm ever imagined radiates
immensities of space and distance as the old Northern does:
Zum Unheil werden—noch allzulange

Männer und Weiber—zur Welt geboren
Aber wir beide —bleiben zusammen
Ich und Sigurd.
The accents of the Homeric hexameter are the soft rustle of a leaf
in the midday sun, the rhythm of matter; but the “Stabreim” likes
“potential energy” in the world-pictures of modern physics, creates a
tense restraint in the void without limits, distant night-storms above
the highest peaks. In its swaying indefiniteness all words and things
dissolve themselves—it is the dynamics, not the statics, of language.
The same applies to the grave rhythm of Media vita in morte sumus.
Here is heralded the colour of Rembrandt and the instrumentation of
Beethoven—here infinite solitude is felt as the home of the Faustian
soul. What is Valhalla? Unknown to the Germans of the Migrations
and even to the Merovingian Age, it was conceived by the nascent
Faustian soul. It was conceived, no doubt, under Classic-pagan and
Arabian-Christian impressions, for the antique and the sacred
writings, the ruins and mosaics and miniatures, the cults and rites
and dogmas of these past Cultures reached into the new life at all
points. And yet, this Valhalla is something beyond all sensible
actualities floating in remote, dim, Faustian regions. Olympus rests
on the homely Greek soil, the Paradise of the Fathers is a magic
garden somewhere in the Universe, but Valhalla is nowhere. Lost in
the limitless, it appears with its inharmonious gods and heroes the
supreme symbol of solitude. Siegfried, Parzeval, Tristan, Hamlet,
Faust are the loneliest heroes in all the Cultures. Read the wondrous
awakening of the inner life in Wolfram’s Parzeval. The longing for the
woods, the mysterious compassion, the ineffable sense of
forsakenness—it is all Faustian and only Faustian. Every one of us
knows it. The motive returns with all its profundity in the Easter
scene of Faust I.
“A longing pure and not to be described

drove me to wander over woods and fields,
and in a mist of hot abundant tears
I felt a world arise and live for me.”
Of this world-experience neither Apollinian nor Magian man,

neither Homer nor the Gospels, knows anything whatever. The
climax of the poem of Wolfram, that wondrous Good Friday morning
scene when the hero, at odds with God and with himself, meets the
noble Gawan and resolves to go on pilgrimage to Tevrezent, takes
us to the heart of the Faustian religion. Here one can feel the
mystery of the Eucharist which binds the communicant to a mystic
company, to a Church that alone can give bliss. In the myth of the
Holy Grail and its Knights one can feel the inward necessity of the
German-Northern Catholicism. In opposition to the Classical
sacrifices offered to individual gods in separate temples, there is
here the one never-ending sacrifice repeated everywhere and every
day. This is the Faustian idea of the 9th-11th Centuries, the Edda
time, foreshadowed by Anglo-Saxon missionaries like Winfried but
only then ripened. The Cathedral, with its High Altar enclosing the
accomplished miracle, is its expression in stone.[199]
The plurality of separate bodies which represents Cosmos for the
Classical soul, requires a similar pantheon—hence the antique
polytheism. The single world-volume, be it conceived as cavern or as
space, demands the single god of Magian or Western Christianity.
Athene or Apollo might be represented by a statue, but it is and has
long been evident to our feeling that the Deity of the Reformation
and the Counter-Reformation can only be “manifested” in the storm
of an organ fugue or the solemn progress of cantata and mass. From
the rich manifold of figures in the Edda and contemporary legends of
saints to Goethe our myth develops itself in steady opposition to the
Classical—in the one case a continuous disintegration of the divine
that culminated in the early Empire in an impossible multitude of
deities, in the other a process of simplification that led to the Deism
of the 18th Century.
The Magian hierarchy of heaven—angels, saints, persons of the
Trinity—has grown paler and paler, more and more disembodied, in
the sphere of the Western pseudomorphosis,[200] supported though it
was by the whole weight of Church authority, and even the Devil—
the great adversary in the Gothic world-drama[201]—has disappeared
unnoticed from among the possibilities of the Faustian world-feeling.
Luther could still throw the inkpot at him, but he has been passed
over in silence by perplexed Protestant theologians long ago. For the
solitude of the Faustian soul agrees not at all with a duality of world
powers. God himself is the All. About the end of the 17th Century
this religiousness could no longer be limited to pictorial expression,
and instrumental music came as its last and only form-language: we
may say that the Catholic faith is to the Protestant as an altar-piece
is to an oratorio. But even the Germanic gods and heroes are
surrounded by this rebuffing immensity and enigmatic gloom. They
are steeped in music and in night, for daylight gives visual bounds
and therefore shapes bodily things. Night eliminates body, day soul.
Apollo and Athene have no souls. On Olympus rests the eternal light
of the transparent southern day, and Apollo’s hour is high noon,
when great Pan sleeps. But Valhalla is light-less, and even in the
Eddas we can trace that deep midnight of Faust’s study-broodings,
the midnight that is caught by Rembrandt’s etchings and absorbs
Beethoven’s tone colours. No Wotan or Baldur or Freya has
“Euclidean” form. Of them, as of the Vedic gods of India, it can be
said that they suffer not “any graven image or any likeness
whatsoever”; and this impossibility carries an implicit recognition that
eternal space, and not the corporeal copy—which levels them down,
desecrates them, denies them—is the supreme symbol. This is the
deep-felt motive that underlies the iconoclastic storms in Islam and
Byzantium (both, be it noted, of the 7th century), and the closely
similar movement in our Protestant North. Was not Descartes’s
creation of the anti-Euclidean analysis of space an iconoclasm? The
Classical geometry handles a number-world of day, the function-
theory is the genuine mathematic of night.
II
That which is expressed by the soul of the West in its
extraordinary wealth of media—words, tones, colours, pictorial
perspectives, philosophical systems, legends, the spaciousness of
Gothic cathedrals and the formulæ of functions—namely its world-
feeling, is expressed by the soul of Old Egypt (which was remote
from all ambitions towards theory and literariness) almost exclusively
by the immediate language of Stone. Instead of spinning word-
subtleties around its form of extension, its “space” and its “time,”
instead of forming hypotheses and number-systems and dogmas, it
set up its huge symbols in the landscape of the Nile in all silence.
Stone is the great emblem of the Timeless-Become; space and
death seem bound up in it. “Men have built for the dead,” says
Bachofen in his autobiography, “before they have built for the living,
and even as a perishable wooden structure suffices for the span of
time that is given to the living, so the housing of the dead for ever
demands the solid stone of the earth. The oldest cult is associated
with the stone that marks the place of burial, the oldest temple-
building with the tomb-structure, the origins of art and decoration
with the grave-ornament. Symbol has created itself in the graves.
That which is thought and felt and silently prayed at the grave-side
can be expressed by no word, but only hinted by the boding symbol
that stands in unchanging grave repose.” The dead strive no more.
They are no more Time, but only Space—something that stays (if
indeed it stays at all) but does not ripen towards a Future; and hence
it is stone, the abiding stone, that expresses how the dead is
mirrored in the waking consciousness of the living. The Faustian soul
looks for an immortality to follow the bodily end, a sort of marriage
with endless space, and it disembodies the stone in its Gothic thrust-
system (contemporary, we may note, with the “consecutives” in
Church music[202]) till at last nothing remained visible but the
indwelling depth- and height-energy of this self-extension. The
Apollinian soul would have its dead burned, would see them
annihilated, and so it remained averse from stone building
throughout the early period of its Culture. The Egyptian soul saw
itself as moving down a narrow and inexorably-prescribed life-path to
come at the end before the judges of the dead (“Book of the Dead,”
cap. 125). That was its Destiny-idea. The Egyptian’s existence is that
of the traveller who follows one unchanging direction, and the whole
form-language of his Culture is a translation into the sensible of this
one theme. And as we have taken endless space as the prime
symbol of the North and body as that of the Classical, so we may
take the word way as most intelligibly expressing that of the
Egyptians. Strangely, and for Western thought almost
incomprehensibly, the one element in extension that they emphasize
is that of direction in depth. The tomb-temples of the Old Kingdom
and especially the mighty pyramid-temples of the Fourth Dynasty
represent, not a purposed organization of space such as we find in
the mosque and the cathedral, but a rhythmically ordered sequence
of spaces. The sacred way leads from the gate-building on the Nile
through passages, halls, arcaded courts and pillared rooms that
grow ever narrower and narrower, to the chamber of the dead,[203]
and similarly the Sun-temples of the Fifth Dynasty are not “buildings”
but a path enclosed by mighty masonry.[204] The reliefs and the
paintings appear always as rows which with an impressive
compulsion lead the beholder in a definite direction. The ram and
sphinx avenues of the New Empire have the same object. For the
Egyptian, the depth-experience which governed his world-form was
so emphatically directional that he comprehended space more or
less as a continuous process of actualization. There is nothing rigid
about distance as expressed here. The man must move, and so
become himself a symbol of life, in order to enter into relation with
the stone part of the symbolism. “Way” signifies both Destiny and
third dimension. The grand wall-surfaces, reliefs, colonnades past
which he moves are “length and breadth”; that is, mere perceptions
of the senses, and it is the forward-driving life that extends them into
“world.” Thus the Egyptian experienced space, we may say, in and
by the processional march along its distinct elements, whereas the
Greek who sacrificed outside the temple did not feel it and the man
of our Gothic centuries praying in the cathedral let himself be
immersed in the quiet infinity of it. And consequently the art of these
Egyptians must aim at plane effects and nothing else, even when it
is making use of solid means. For the Egyptian, the pyramid over the
king’s tomb is a triangle, a huge, powerfully expressive plane that,
whatever be the direction from which one approaches, closes off the
“way” and commands the landscape. For him, the columns of the
inner passages and courts, with their dark backgrounds, their dense
array and their profusion of adornments, appear entirely as vertical
strips which rhythmically accompany the march of the priests. Relief-
work is—in utter contrast to the Classical—carefully restricted in one
plane; in the course of development dated by the Third to the Fifth
dynasties it diminishes from the thickness of a finger to that of a
sheet of paper, and finally it is sunk in the plane.[205] The dominance
of the horizontal, the vertical and the right angle, and the avoidance
of all foreshortening support the two-dimensional principle and serve
to insulate this directional depth-experience which coincides with the
way and the grave at its end. It is an art that admits of no deviation
for the relief of the tense soul.
Is not this an expression in the noblest language that it is possible
to conceive of what all our space-theories would like to put into
words? Is it not a metaphysic in stone by the side of which the
written metaphysics of Kant seems but a helpless stammering?
There is, however, another Culture that, different as it most
fundamentally is from the Egyptian, yet found a closely-related prime
symbol. This is the Chinese, with its intensely directional principle of
the Tao.[206] But whereas the Egyptian treads to the end a way that is
prescribed for him with an inexorable necessity, the Chinaman
wanders through his world; consequently, he is conducted to his god
or his ancestral tomb not by ravines of stone, between faultless
smooth walls, but by friendly Nature herself. Nowhere else has the
landscape become so genuinely the material of the architecture.
“Here, on religious foundations, there has been developed a grand
lawfulness and unity common to all building, which, combined with
the strict maintenance of a north-south general axis, always holds
together gate-buildings, side-buildings, courts and halls in the same
homogeneous plan, and has led finally to so grandiose a planning
and such a command over ground and space that one is quite
justified in saying that the artist builds and reckons with the
landscape itself.”[207] The temple is not a self-contained building but a
lay-out, in which hills, water, trees, flowers, and stones in definite
forms and dispositions are just as important as gates, walls, bridges
and houses. This Culture is the only one in which the art of
gardening is a grand religious art. There are gardens that are
reflections of particular Buddhist sects.[208] It is the architecture of the
landscape, and only that, which explains the architecture of the
buildings, with their flat extension and the emphasis laid on the roof
as the really expressive element. And just as the devious ways
through doors, over bridges, round hills and walls lead at last to the
end, so the paintings take the beholder from detail to detail whereas
Egyptian relief masterfully points him in the one set direction. “The
whole picture is not to be taken at once. Sequence in time
presupposes a sequence of space-elements through which the eye
is to wander from one to the next.”[209] Whereas the Egyptian
architecture dominates the landscape, the Chinese espouses it. But
in both cases it is direction in depth that maintains the becoming of
space as a continuously-present experience.
III
All art is expression-language.[210] Moreover, in its very earliest
essays—which extend far back into the animal world—it is that of
one active existence speaking for itself only, and it is unconscious of
witnesses even though in the absence of such the impulse to
expression would not come to utterance. Even in quite “late”
conditions we often see, instead of the combination of artist and
spectator, a crowd of art-makers who all dance or mime or sing. The
idea of the “Chorus” as sum total of persons present has never
entirely vanished from art-history. It is only the higher art that
becomes decisively an art “before witnesses” and especially (as
Nietzsche somewhere remarks) before God as the supreme witness.
[211]
This expression is either ornament or imitation. Both are higher

possibilities and their polarity to one another is hardly perceptible in
the beginnings. Of the two, imitation is definitely the earlier and the
closer to the producing race. Imitation is the outcome of a
physiognomic idea of a second person with whom (or which) the first
is involuntarily induced into resonance of vital rhythm (mitschwingen
im); whereas ornament evidences an ego conscious of its own
specific character. The former is widely spread in the animal world,
the latter almost peculiar to man.
Imitation is born of the secret rhythm of all things cosmic. For the
waking being the One appears as discrete and extended; there is a
Here and a There, a Proper and an Alien something, a Microcosm
and a Macrocosm that are polar to one another in the sense-life, and
what the rhythm of imitation does is to bridge this dichotomy. Every
religion is an effort of the waking soul to reach the powers of the
world-around. And so too is Imitation, which in its most devoted
moments is wholly religious, for it consists in an identity of inner
activity between the soul and body “here” and the world-around
“there” which, vibrating as one, become one. As a bird poises itself in
the storm or a float gives to the swaying waves, so our limbs take up
an irresistible beat at the sound of march-music. Not less contagious
is the imitation of another’s bearing and movements, wherein
children in particular excel. It reaches the superlative when we “let
ourselves go” in the common song or parade-march or dance that
creates out of many units one unit of feeling and expression, a “we.”
But a “successful” picture of a man or a landscape is also the
outcome of a felt harmony of the pictorial motion with the secret
swing and sway of the living opposite; and it is this actualizing of
physiognomic rhythm that requires the executant to be an adept who
can reveal the idea, the soul, of the alien in the play of its surface. In
certain unreserved moments we are all adepts of this sort, and in
such moments, as we follow in an imperceptible rhythm the music
and the play of facial expression, we suddenly look over the
precipice and see great secrets. The aim of all imitation is effective
simulation; this means effective assimilation of ourselves into an
alien something—such a transposition and transubstantiation that
the One lives henceforth in the Other that it describes or depicts—
and it is able to awaken an intense feeling of unison over all the
range from silent absorption and acquiescence to the most
abandoned laughter and down into the last depths of the erotic, a
unison which is inseparable from creative activity. In this wise arose
the popular circling-dances (for instance, the Bavarian Schuhplattler
was originally imitated from the courtship of the woodcocks) but this
too is what Vasari means when he praises Cimabue and Giotto as
the first who returned to the imitation of “Nature”—the Nature, that is,
of springtime men, of which Meister Eckart said: “God flows out in all
creatures, and therefore all created is God.” That which in this world-
around presents itself to our contemplation—and therefore contains
meaning for our feelings—as movement, we render by movement.
Hence all imitation is in the broadest sense dramatic; drama is
presented in the movement of the brush-stroke or the chisel, the
melodic curve of the song, the tone of the recitation, the line of
poetry, the description, the dance. But everything that we experience
with and in seeings and hearings is always an alien soul to which we
are uniting ourselves. It is only at the stage of the Megalopolis that
art, reasoned to pieces and de-spiritualized, goes over to naturalism
as that term is understood nowadays; viz., imitation of the charm of
visible appearances, of the stock of sensible characters that are
capable of being scientifically fixed.
Ornament detaches itself now from Imitation as something which
does not follow the stream of life but rigidly faces it. Instead of
physiognomic traits overheard in the alien being, we have
established motives, symbols, which are impressed upon it. The
intention is no longer to pretend but to conjure. The “I” overwhelms
the “Thou.” Imitation is only a speaking with means that are born of
the moment and unreproduceable—but Ornament employs a
language emancipated from the speaking, a stock of forms that
possesses duration and is not at the mercy of the individual.[212]
Only the living can be imitated, and it can be imitated only in
movements, for it is through these that it reveals itself to the senses
of artists and spectators. To that extent, imitation belongs to Time
and Direction. All the dancing and drawing and describing and
portraying for eye and ear is irrevocably “directional,” and hence the
highest possibilities of Imitation lie in the copying of a destiny, be it in
tones, verses, picture or stage-scene.[213] Ornament, on the contrary,
is something taken away from Time: it is pure extension, settled and
stable. Whereas an imitation expresses something by accomplishing
itself, ornament can only do so by presenting itself to the senses as
a finished thing. It is Being as such, wholly independent of origin.
Every imitation possesses beginning and end, while an ornament
possesses only duration, and therefore we can only imitate the
destiny of an individual (for instance, Antigone or Desdemona), while
by an ornament or symbol only the generalized destiny-idea itself
can be represented (as, for example, that of the Classical world by
the Doric column). And the former presupposes a talent, while the
latter calls for an acquirable knowledge as well.
All strict arts have their grammar and syntax of form-language,
with rules and laws, inward logic and tradition. This is true not merely
for the Doric cabin-temple and Gothic cottage-cathedral, for the
carving-schools of Egypt[214] and Athens and the cathedral plastic of
northern France, for the painting-schools of the Classical world and
those of Holland and the Rhine and Florence, but also for the fixed
rules of the Skalds and Minnesänger which were learned and
practised as a craft (and dealt not merely with sentence and metre
but also with gesture and the choice of imagery[215]), for the
narration-technique of the Vedic, Homeric and Celto-Germanic Epos,
for the composition and delivery of the Gothic sermon (both
vernacular and Latin), and for the orators’ prose[216] in the Classical,
and for the rules of French drama. In the ornamentation of an art-
work is reflected the inviolable causality of the macrocosm as the
man of the particular kind sees and comprehends it. Both have
system. Each is penetrated with the religious side of life—fear and
love.[217] A genuine symbol can instil fear or can set free from fear;
the “right” emancipates and the “wrong” hurts and depresses. The
imitative side of the arts, on the contrary, stands closer to the real
race-feelings of hate and love, out of which arises the opposition of
ugly and beautiful. This is in relation only with the living, of which the
inner rhythm repels us or draws us into phase with it, whether it be
that of the sunset-cloud or that of the tense breath of the machine.
An imitation is beautiful, an ornament significant, and therein lies the
difference between direction and extension, organic and inorganic
logic, life and death. That which we think beautiful is “worth copying.”
Easily it swings with us and draws us on to imitate, to join in the
singing, to repeat. Our hearts beat higher, our limbs twitch, and we
are stirred till our spirits overflow. But as it belongs to Time, it “has its
time.” A symbol endures, but everything beautiful vanishes with the
life-pulsation of the man, the class, the people or the race that feels it
as a specific beauty in the general cosmic rhythm.[218] The “beauty”
that Classical sculpture and poetry contained for Classical eyes is
something different from the beauty that they contain for ours—
something extinguished irrecoverably with the Classical soul—while
what we regard as beautiful in it is something that only exists for us.
Not only is that which is beautiful for one kind of man neutral or ugly
for another—e.g., the whole of our music for the Chinese, or
Mexican sculpture for us. For one and the same life the accustomed,
the habitual, owing to the very fact of its possessing duration, cannot
possess beauty.
And now for the first time we can see the opposition between
these two sides of every art in all its depth. Imitation spiritualizes and
quickens, ornament enchants and kills. The one becomes, the other
is. And therefore the one is allied to love and, above all—in songs
and riot and dance—to the sexual love, which turns existence to face
the future; and the other to care of the past, to recollection[219] and to
the funerary. The beautiful is longingly pursued, the significant instils
dread, and there is no deeper contrast than that between the house
of the living and the house of the dead.[220] The peasant’s cottage[221]
and its derivative the country noble’s hall, the fenced town and the
castle are mansions of life, unconscious expressions of circling
blood, that no art produced and no art can alter. The idea of the
family appears in the plan of the proto-house, the inner form of the
stock in the plan of its villages—which after many a century and
many a change of occupation still show what race it was that
founded them[222]—the life of a nation and its social ordering in the
plan (not the elevation or silhouette) of the city.[223] On the other
hand, Ornamentation of the high order develops itself on the stiff
symbols of death, the urn, the sarcophagus, the stele and the temple
of the dead,[224] and beyond these in gods’ temples and cathedrals
which are Ornament through and through, not the expressions of a
race but the language of a world-view. They are pure art through and
through—just what the castle and the cottage are not.[225]
For cottage and castle are buildings in which art, and, specifically,
imitative art, is made and done, the home of Vedic, Homeric and
Germanic epos, of the songs of heroes, the dance of boors and that
of lords and ladies, of the minstrel’s lay. The cathedral, on the other
hand, is art, and, moreover, the only art by which nothing is imitated;
it alone is pure tension of persistent forms, pure three-dimensional
logic that expresses itself in edges and surfaces and volumes. But
the art of villages and castles is derived from the inclinations of the
moment, from the laughter and high spirit of feasts and games, and
to such a degree is it dependent on Time, so much is it a thing of
occasion, that the troubadour obtains his very name from finding,
while Improvisation—as we see in the Tzigane music to-day—is
nothing but race manifesting itself to alien senses under the
influence of the hour. To this free creative power all spiritual art
opposes the strict school in which the individual—in the hymn as in
the work of building and carving—is the servant of a logic of timeless
forms, and so in all Cultures the seat of its style-history is in its early
cult architecture. In the castle it is the life and not the structure that
possesses style. In the town the plan is an image of the destinies of
a people, whereas the silhouette of emergent spires and cupolas
tells of the logic in the builders’ world-picture, of the “first and last
things” of their universe.
In the architecture of the living, stone serves a worldly purpose,
but in the architecture of the cult it is a symbol.[226] Nothing has
injured the history of the great architectures so much as the fact that
it has been regarded as the history of architectural techniques
instead of as that of architectural ideas which took their technical
expression-means as and where they found them. It has been just
the same with the history of musical instruments,[227] which also were
developed on a foundation of tone-language. Whether the groin and
the flying buttress and the squinch-cupola were imagined specially
for the great architectures or were expedients that lay more or less
ready to hand and were taken into use, is for art-history a matter of
as little importance as the question of whether, technically, stringed
instruments originated in Arabia or in Celtic Britain. It may be that the
Doric column was, as a matter of workmanship, borrowed from the
Egyptian temples of the New Empire, or the late-Roman domical
construction from the Etruscans, or the Florentine court from the
North-African Moors. Nevertheless the Doric peripteros, the
Pantheon, and the Palazzo Farnese belong to wholly different worlds
—they subserve the artistic expression of the prime-symbol in three
different Cultures.
IV
In every springtime, consequently, there are two definitely
ornamental and non-imitative arts, that of building and that of
decoration. In the longing and pregnant centuries before it, elemental
expression belongs exclusively to Ornamentation in the narrow
sense. The Carolingian period is represented only by its ornament,
as its architecture, for want of the Idea, stands between the styles.
And similarly, as a matter of art-history, it is immaterial that no
buildings of the Mycenæan age have survived.[228] But with the dawn
of the great Culture, architecture as ornament comes into being
suddenly and with such a force of expression that for a century mere
decoration-as-such shrinks away from it in awe. The spaces,
surfaces and edges of stone speak alone. The tomb of Chephren is
the culmination of mathematical simplicity—everywhere right angles,
squares and rectangular pillars, nowhere adornment, inscription or
desinence—and it is only after some generations have passed that
Relief ventures to infringe the solemn magic of those spaces and the
strain begins to be eased. And the noble Romanesque of
Westphalia-Saxony (Hildesheim, Gernrode, Paulinzella, Paderborn),
of Southern France and of the Normans (Norwich and Peterborough)
managed to render the whole sense of the world with indescribable
power and dignity in one line, one capital, one arch.
When the form-world of the springtime is at its highest, and not
before, the ordained relation is that architecture is lord and ornament
is vassal. And the word “ornament” is to be taken here in the widest
possible sense. Even conventionally, it covers the Classical unit-
motive with its quiet poised symmetry or meander supplement, the
spun surface of arabesque and the not dissimilar surface-patterning
of Mayan art, and the “Thunder-pattern”[229] and others of the early
Chóu period which prove once again the landscape basis of the old
Chinese architecture without a doubt. But the warrior figures of
Dipylon vases are also conceived in the spirit of ornament, and so, in
a far higher degree still, are the statuary groups of Gothic cathedrals.
“The figures were composed pillarwise from the spectator, the
figures of the pillar being, with reference to the spectator, ranked
upon one another like rhythmic figures in a symphony that soars
heavenward and expands its sounds in every direction.”[230] And
besides draperies, gestures, and figure-types, even the structure of
the hymn-strophe and the parallel motion of the parts in church
music are ornament in the service of the all-ruling architectural idea.
[231]
The spell of the great Ornamentation remains unbroken till in the
beginning of a “late” period architecture falls into a group of civic and
worldly special arts that unceasingly devote themselves to pleasing
and clever imitation and become ipso facto personal. To Imitation
and Ornament the same applies that has been said already of time
and space. Time gives birth to space, but space gives death to time.
[232]
In the beginning, rigid symbolism had petrified everything alive;
the Gothic statue was not permitted to be a living body, but was
simply a set of lines disposed in human form. But now Ornament
loses all its sacred rigour and becomes more and more decoration
for the architectural setting of a polite and mannered life. It was
purely as this, namely as a beautifying element, that Renaissance
taste was adopted by the courtly and patrician world of the North
(and by it alone!). Ornament meant something quite different in the
Egyptian Old Kingdom from what it meant in the Middle; in the
geometric period from what it meant in the Hellenistic; at the end of
the 12th Century from what it meant at the end of Louis XIV’s reign.
And architecture too becomes pictorial and makes music, and its
forms seem always to be trying to imitate something in the picture of
the world-around. From the Ionic capital we proceed to the
Corinthian, and from Vignola through Bernini to the Rococo.
At the last, when Civilization sets in, true ornament and, with it,
great art as a whole are extinguished. The transition consists—in
every Culture—in Classicism and Romanticism of one sort or
another, the former being a sentimental regard for an Ornamentation
(rules, laws, types) that has long been archaic and soulless, and the
latter a sentimental Imitation, not of life, but of an older Imitation. In
the place of architectural style we find architectural taste. Methods of
painting and mannerisms of writing, old forms and new, home and
foreign, come and go with the fashion. The inward necessity is no
longer there, there are no longer “schools,” for everyone selects
what and where it pleases him to select. Art becomes craft-art
(Kunstgewerbe) in all its branches—architecture and music, poetry
and drama—and in the end we have a pictorial and literary stock-in-
trade which is destitute of any deeper significance and is employed
according to taste. This final or industrial form of Ornament—no
longer historical, no longer in the condition of “becoming”—we have
before us not only in the patterns of oriental carpets, Persian and
Indian metal work,` Chinese porcelain, but also in Egyptian (and
Babylonian) art as the Greeks and Romans met it. The Minoan art of
Crete is pure craft-art, a northern outlier of Egyptian post-Hyksos
taste; and its “contemporary,” Hellenistic-Roman art from about the
time of Scipio and Hannibal, similarly subserves the habit of comfort
and the play of intellect. From the richly-decorated entablature of the
Forum of Nerva in Rome to the later provincial ceramics in the West,
we can trace the same steady formation of an unalterable craft-art
that we find in the Egyptian and the Islamic worlds, and that we have
to presume in India after Buddha and in China after Confucius.
Now, Cathedral and Pyramid-temple are different in spite of their

deep inward kinship, and it is precisely in these differences that we
seize the mighty phenomenon of the Faustian soul, whose depth-
impulse refuses to be bound in the prime symbol of a way, and from
its earliest beginnings strives to transcend every optical limitation.
Can anything be more alien to the Egyptian conception of the State
—whose tendency we may describe as a noble sobriety—than the
political ambitions of the great Saxon, Franconian and Hohenstaufen
Emperors, who came to grief because they overleapt all political
actualities and for whom the recognition of any bounds would have
been a betrayal of the idea of their rulership? Here the prime symbol
of infinite space, with all its indescribable power, entered the field of
active political existence. Beside the figures of the Ottos, Conrad II,
Henry VI and Frederick II stand the Viking-Normans, conquerors of
Russia, Greenland, England, Sicily and almost of Constantinople;
and the great popes, Gregory VII and Innocent III—all of whom alike
aimed at making their visible spheres of influence coincident with the
whole known world. This is what distinguishes the heroes of the Grail
and Arthurian and Siegfried sagas, ever roaming in the infinite, from
the heroes of Homer with their geographically modest horizon; and
the Crusades, that took men from the Elbe and the Loire to the limits
of the known world, from the historical events upon which the
Classical soul built the “Iliad” and which from the style of that soul we
may safely assume to have been local, bounded, and completely
appreciable.
The Doric soul actualized the symbol of the corporally-present
individual thing, while deliberately rejecting all big and far-reaching
creations, and it is for this very good reason that the first post-
Mycenæan period has bequeathed nothing to our archæologists.
The expression to which this soul finally attained was the Doric
temple with its purely outward effectiveness, set upon the landscape
as a massive image but denying and artistically disregarding the
space within as the μὴ ὄv, that which was held to be incapable of
existence. The ranked columns of the Egyptians carried the roof of a
hall. The Greek in borrowing the motive invested it with a meaning
proper to himself—he turned the architectural type inside out like a
glove. The outer column-sets are, in a sense, relics of a denied
interior.[233]
The Magian and the Faustian souls, on the contrary, built high.
Their dream-images became concrete as vaultings above significant
inner-spaces, structural anticipations respectively of the mathematic
of algebra and that of analysis. In the style that radiated from
Burgundy and Flanders rib-vaulting with its lunettes and flying
buttresses emancipated the contained space from the sense-
appreciable surface[234] bounding it. In the Magian interior "the
window is merely a negative component, a utility-form in no wise yet
developed into an art-form—to put it crudely, nothing but a hole in
the wall."[235] When windows were in practice indispensable, they
were for the sake of artistic impression concealed by galleries as in
the Eastern basilica.[236] The window as architecture, on the other
hand, is peculiar to the Faustian soul and the most significant symbol
of its depth-experience. In it can be felt the will to emerge from the
interior into the boundless. The same will that is immanent in

Textbook Computer Algebra in Scientific Computing 16Th International Workshop Casc 2014 Warsaw Poland September 8 12 2014 Proceedings 1St Edition Vladimir P Gerdt Ebook All Chapter PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Textbook Computer Algebra in Scientific Computing 16Th International Workshop Casc 2014 Warsaw Poland September 8 12 2014 Proceedings 1St Edition Vladimir P Gerdt Ebook All Chapter PDF

Uploaded by

Copyright:

Available Formats

Computer Algebra in Scientific

Computing 16th International

Computer Algebra in Scientific Computing 20th

Computer Algebra in Scientific Computing 18th

Computer Algebra in Scientific Computing 22nd

Intelligent Systems 2014 Proceedings of the 7th IEEE

Discrete Geometry for Computer Imagery 18th IAPR

Principles and Practice of Constraint Programming 20th

Parameterized and Exact Computation 9th International

Augmented Environments for Computer Assisted

ISSN 0302-9743 e-ISSN 1611-3349

Library of Congress Control Number: 2014946203

LNCS Sublibrary: SL 1 – Theoretical Computer Science and General Issues

Since their start in 1998, the International Workshops on Computer Algebra in

initial conditions additionally depend on sign vectors of the stoichiometric and

July 2014 V.P. Gerdt

Workshop General Chairs

Program Committee Chairs

Ruyong Feng Bernard Mourrain

Computable Inﬁnite Power Series in the Role of Coeﬃcients of Linear

Symbolic-Numerical Solution of Boundary-Value Problems with

Sporadic Examples of Directed Strongly Regular Graphs Obtained By

On the Parallelization of Subproduct Tree Techniques Targeting

Deterministically Computing Reduction Numbers of Polynomial

A Note on Global Newton Iteration Over Archimedean and

Invariant Manifolds in the Classic and Generalized Goryachev–

Coherence and Large-Scale Pattern Formation in Coupled Logistic-Map

On the Computation of the Determinant of a Generalized Vandermonde

Towards Conﬂict-Driven Learning for Virtual Substitution . . . . . . . . . . . . 256

Sharpness in Trajectory Estimation for Planar Four-points

Scheme for Numerical Investigation of Movable Singularities of the

Generalized Mass-Action Systems and Positive Solutions of Polynomial

Lie Symmetry Analysis for Cosserat Rods . . . . . . . . . . . . . . . . . . . . . . . . . . . 324

Real Polynomial Root-Finding by Means of Matrix and Polynomial

On Testing Uniqueness of Analytic Solutions of PDE with Boundary

Continuous Problems: Optimality, Complexity, Tractability

On Integrability of Evolutionary Equations in the Restricted

Factoring Sparse Bivariate Polynomials Using the Priority Queue . . . . . . 388

Solving Parametric Sparse Linear Systems by Local Blocking . . . . . . . . . . 403

Analytical Calculations in Maple to Implement the Method of

CAS Application to the Construction of the Collocations and Least

An Algorithm for Computing the Truncated Annihilating Ideals for an

Applications of the Newton Index to the Construction of Irreducible

Symbolic-Numeric Algorithm for Solving the Problem of Quantum

Enumeration of Schur Rings Over Small Groups . . . . . . . . . . . . . . . . . . . . . 491

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501

Sergei A. Abramov1, and Moulay A. Barkatou2

Abstract. We consider linear ordinary diﬀerential systems over a dif-

2.1 Diﬀerential Universal and Adequate Field Extensions

Ar y (r) + Ar−1 y (r−1) + · · · + A0 y = 0, (2)

where the matrices A0 , A1 , . . . , Ar belong to Matm (K), m 1, and Ar (the

2.2 Universal Diﬀerential Extension of Formal Laurent Series Field

eP (x) xγ (ψ0 + ψ1 ln x + · · · + ψs (ln x)s ), (5)

where in any such expression

– P (x) ∈ K0 [x−1/q ], q is a positive integer,

ψj ∈ K0 [[x1/q ]], (6)

γi ∈ K0 , qi is a positive integer, Pi (x) ∈ K0 [x−1/qi ], Ψi (x) ∈ K0m [[x1/qi ]][ln x],

Deﬁnition 2. Solutions of the form (7) will be called (formal) exponential-log-

Remark 1. If K0 is not algebraically closed then there exists a simple algebraic

2.3 Row Frontal Matrix and Row Order

3 When K Is a Constructive Field

3.1 The Dimension of the Solution Space of a Given Full Rank

3.2 Recognizing the Unimodularity of an Operator and Computing