(Ebook) A concise introduction to geometric

numerical integration by Blanes, Sergio; Casas,

Fernando ISBN 9781482263442, 1482263440 Pdf Download

https://ebooknice.com/product/a-concise-introduction-to-geometric-
numerical-integration-6745976

★★★★★
4.7 out of 5.0 (33 reviews )

Instant PDF Download

ebooknice.com
 (Ebook) A concise introduction to geometric numerical
integration by Blanes, Sergio; Casas, Fernando ISBN
9781482263442, 1482263440 Pdf Download

EBOOK

Available Formats

■ PDF eBook Study Guide Ebook

EXCLUSIVE 2025 EDUCATIONAL COLLECTION - LIMITED TIME

INSTANT DOWNLOAD VIEW LIBRARY

We have selected some products that you may be interested in
Click the link to download now or visit ebooknice.com
for more options!.

(Ebook) Biota Grow 2C gather 2C cook by Loucas, Jason; Viles,

James ISBN 9781459699816, 9781743365571, 9781925268492,
1459699815, 1743365578, 1925268497

https://ebooknice.com/product/biota-grow-2c-gather-2c-cook-6661374

(Ebook) Matematik 5000+ Kurs 2c Lärobok by Lena Alfredsson, Hans

Heikne, Sanna Bodemyr ISBN 9789127456600, 9127456609

https://ebooknice.com/product/matematik-5000-kurs-2c-larobok-23848312

(Ebook) SAT II Success MATH 1C and 2C 2002 (Peterson's SAT II

Success) by Peterson's ISBN 9780768906677, 0768906679

https://ebooknice.com/product/sat-ii-success-math-1c-and-2c-2002-peterson-
s-sat-ii-success-1722018

(Ebook) Master SAT II Math 1c and 2c 4th ed (Arco Master the SAT
Subject Test: Math Levels 1 & 2) by Arco ISBN 9780768923049,
0768923042

https://ebooknice.com/product/master-sat-ii-math-1c-and-2c-4th-ed-arco-
master-the-sat-subject-test-math-levels-1-2-2326094
(Ebook) Cambridge IGCSE and O Level History Workbook 2C - Depth
Study: the United States, 1919-41 2nd Edition by Benjamin
Harrison ISBN 9781398375147, 9781398375048, 1398375144,
1398375047
https://ebooknice.com/product/cambridge-igcse-and-o-level-history-
workbook-2c-depth-study-the-united-states-1919-41-2nd-edition-53538044

(Ebook) Measure and Integration: A Concise Introduction to Real

Analysis by Leonard F. Richardson ISBN 9780470259542, 047025954X

https://ebooknice.com/product/measure-and-integration-a-concise-
introduction-to-real-analysis-1949224

(Ebook) Limits & Proximities by Fernando Casas, Bruce Leutwyler,

Fernando Castro, Oscar Arandia, Eduardo Mitre

https://ebooknice.com/product/limits-proximities-10016384

(Ebook) Numerical Linear Algebra: A Concise Introduction with

MATLAB and Julia by Folkmar Bornemann ISBN 9783319742229,
9783319742212, 3319742221, 3319742213

https://ebooknice.com/product/numerical-linear-algebra-a-concise-
introduction-with-matlab-and-julia-56193352

(Ebook) A Concise History of Brazil by Boris Fausto; Sergio

Fausto ISBN 9781107036208, 1107036208

https://ebooknice.com/product/a-concise-history-of-brazil-34794304
A Concise
Introduction
to Geometric
Numerical
Integration
 MONOGRAPHS AND RESEARCH NOTES IN MATHEMATICS

Series Editors
John A. Burns
Thomas J. Tucker
Miklos Bona
Michael Ruzhansky

Published Titles
Application of Fuzzy Logic to Social Choice Theory, John N. Mordeson, Davender S. Malik
and Terry D. Clark
Blow-up Patterns for Higher-Order: Nonlinear Parabolic, Hyperbolic Dispersion and
Schrödinger Equations, Victor A. Galaktionov, Enzo L. Mitidieri, and Stanislav Pohozaev
Complex Analysis: Conformal Inequalities and the Bieberbach Conjecture, Prem K. Kythe
Computational Aspects of Polynomial Identities: Volume l, Kemer’s Theorems, 2nd Edition
Alexei Kanel-Belov, Yakov Karasik, and Louis Halle Rowen
A Concise Introduction to Geometric Numerical Integration Fernando Casas
and Sergio Blanes
Cremona Groups and Icosahedron, Ivan Cheltsov and Constantin Shramov
Diagram Genus, Generators, and Applications, Alexander Stoimenow
Difference Equations: Theory, Applications and Advanced Topics, Third Edition,
Ronald E. Mickens
Dictionary of Inequalities, Second Edition, Peter Bullen
Iterative Optimization in Inverse Problems, Charles L. Byrne
Line Integral Methods for Conservative Problems, Luigi Brugnano and Felice Iavernaro
Lineability: The Search for Linearity in Mathematics, Richard M. Aron,
Luis Bernal González, Daniel M. Pellegrino, and Juan B. Seoane Sepúlveda
Modeling and Inverse Problems in the Presence of Uncertainty, H. T. Banks, Shuhua Hu,
and W. Clayton Thompson
Monomial Algebras, Second Edition, Rafael H. Villarreal
Nonlinear Functional Analysis in Banach Spaces and Banach Algebras: Fixed Point
Theory Under Weak Topology for Nonlinear Operators and Block Operator Matrices with
Applications, Aref Jeribi and Bilel Krichen
Partial Differential Equations with Variable Exponents: Variational Methods and Qualitative
Analysis, Vicenţiu D. Rădulescu and Dušan D. Repovš
A Practical Guide to Geometric Regulation for Distributed Parameter Systems,
Eugenio Aulisa and David Gilliam
Reconstruction from Integral Data, Victor Palamodov
Signal Processing: A Mathematical Approach, Second Edition, Charles L. Byrne
Sinusoids: Theory and Technological Applications, Prem K. Kythe
Special Integrals of Gradshteyn and Ryzhik: the Proofs – Volume l, Victor H. Moll
Published Titles Continued
Special Integrals of Gradshteyn and Ryzhik: the Proofs – Volume ll, Victor H. Moll
Stochastic Cauchy Problems in Infinite Dimensions: Generalized and Regularized
Solutions, Irina V. Melnikova
Submanifolds and Holonomy, Second Edition, Jürgen Berndt, Sergio Console,
and Carlos Enrique Olmos
The Truth Value Algebra of Type-2 Fuzzy Sets: Order Convolutions of Functions on the
Unit Interval, John Harding, Carol Walker, and Elbert Walker

Forthcoming Titles
Actions and Invariants of Algebraic Groups, Second Edition, Walter Ferrer Santos
and Alvaro Rittatore
Analytical Methods for Kolmogorov Equations, Second Edition, Luca Lorenzi
Geometric Modeling and Mesh Generation from Scanned Images, Yongjie Zhang
Groups, Designs, and Linear Algebra, Donald L. Kreher
Handbook of the Tutte Polynomial, Joanna Anthony Ellis-Monaghan and Iain Moffat
Microlocal Analysis on Rˆn and on NonCompact Manifolds, Sandro Coriasco
Practical Guide to Geometric Regulation for Distributed Parameter Systems,
Eugenio Aulisa and David S. Gilliam
Symmetry and Quantum Mechanics, Scott Corry
This page intentionally left blank
MONOGRAPHS AND RESEARCH NOTES IN MATHEMATICS

A Concise
Introduction
to Geometric
Numerical
Integration

Sergio Blanes
Universitat Politècnica de València
Valencia, Spain

Fernando Casas
Universitat Jaume I
Castellón, Spain
MATLAB® is a trademark of The MathWorks, Inc. and is used with permission. The MathWorks does
not warrant the accuracy of the text or exercises in this book. This book’s use or discussion of MAT-
LAB® software or related products does not constitute endorsement or sponsorship by The MathWorks
of a particular pedagogical approach or particular use of the MATLAB® software.

CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2016 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Version Date: 20160315

International Standard Book Number-13: 978-1-4822-6344-2 (eBook - PDF)

This book contains information obtained from authentic and highly regarded sources. Reasonable
efforts have been made to publish reliable data and information, but the author and publisher cannot
assume responsibility for the validity of all materials or the consequences of their use. The authors and
publishers have attempted to trace the copyright holders of all material reproduced in this publication
and apologize to copyright holders if permission to publish in this form has not been obtained. If any
copyright material has not been acknowledged please write and let us know so we may rectify in any
future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced,
transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying, microfilming, and recording, or in any information stor-
age or retrieval system, without written permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.copy-
right.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222
Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that pro-
vides licenses and registration for a variety of users. For organizations that have been granted a photo-
copy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are
used only for identification and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
Contents

Preface xi

1 What is geometric numerical integration? 1

1.1 First elementary examples and numerical methods . . . . . . 2

1.1.1 Simple harmonic oscillator . . . . . . . . . . . . . . . . 2
1.1.2 Some elementary numerical methods . . . . . . . . . . 4
1.1.3 Simple mathematical pendulum . . . . . . . . . . . . . 8
1.2 Classical paradigm of numerical integration . . . . . . . . . . 11
1.2.1 Adjoint method, symmetric method . . . . . . . . . . 14
1.2.2 Stability . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.2.3 Stiffness . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.3 Towards a new paradigm: Geometric numerical integration . 17
1.4 Symplectic integration . . . . . . . . . . . . . . . . . . . . . . 20
1.4.1 Crash course on Hamiltonian dynamics: Hamiltonian
systems . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.4.2 Some elementary symplectic integrators . . . . . . . . 22
1.4.2.1 Symplectic Euler methods . . . . . . . . . . 22
1.4.2.2 Störmer–Verlet schemes . . . . . . . . . . . . 24
1.4.2.3 Geometric properties of the Störmer–Verlet
method . . . . . . . . . . . . . . . . . . . . . 25
1.4.2.4 Implicit midpoint and trapezoidal rules . . . 27
1.5 Illustration: The Kepler problem . . . . . . . . . . . . . . . . 28
1.5.1 Error growth . . . . . . . . . . . . . . . . . . . . . . . 29
1.5.2 Efficiency plots . . . . . . . . . . . . . . . . . . . . . . 31
1.6 What is to be treated in this book (and what is not) . . . . 35
1.7 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

2 Classical integrators and preservation of properties 41

2.1 Taylor series methods . . . . . . . . . . . . . . . . . . . . . . 41

2.2 Runge–Kutta methods . . . . . . . . . . . . . . . . . . . . . 43
2.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 43
2.2.2 General formulation . . . . . . . . . . . . . . . . . . . 45
2.2.3 Collocation methods . . . . . . . . . . . . . . . . . . . 47
2.2.4 Partitioned Runge–Kutta methods . . . . . . . . . . . 48

vii
viii Contents

2.2.5 Runge–Kutta–Nyström methods . . . . . . . . . . . . 50

2.2.6 Preservation of invariants . . . . . . . . . . . . . . . . 51
2.3 Multistep methods . . . . . . . . . . . . . . . . . . . . . . . . 52
2.3.1 A general class of multistep procedures . . . . . . . . 52
2.3.2 Predictor–corrector methods . . . . . . . . . . . . . . 55
2.3.3 Multistep methods and symplecticity . . . . . . . . . . 56
2.4 Numerical examples . . . . . . . . . . . . . . . . . . . . . . . 58
2.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

3 Splitting and composition methods 65

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.2 Composition and splitting . . . . . . . . . . . . . . . . . . . 69
3.3 Order conditions of splitting and composition methods . . . 71
3.3.1 Negative coefficients . . . . . . . . . . . . . . . . . . . 75
3.3.2 Counting the number of order conditions . . . . . . . 76
3.4 Splitting methods for special systems . . . . . . . . . . . . . 77
3.4.1 Near-integrable systems . . . . . . . . . . . . . . . . . 78
3.4.2 Runge–Kutta–Nyström methods . . . . . . . . . . . . 79
3.4.3 Modified potentials . . . . . . . . . . . . . . . . . . . . 80
3.5 Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.6 Splitting methods for non-autonomous systems . . . . . . . . 85
3.7 A collection of low order splitting and composition methods 87
3.7.1 General (nonprocessed) methods . . . . . . . . . . . . 88
3.7.2 Processed methods . . . . . . . . . . . . . . . . . . . . 90
3.8 Illustrations . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.9 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

4 Other types of geometric numerical integrators 103

4.1 Symplectic methods based on generating functions . . . . . . 104

4.1.1 Crash course on Hamiltonian dynamics II: Canonical
transformations and generating functions . . . . . . . 104
4.1.2 Symplectic integrators based on generating functions . 108
4.2 Variational integrators . . . . . . . . . . . . . . . . . . . . . 109
4.2.1 Crash course on Hamiltonian dynamics III: Hamilton’s
principle and Lagrangian formulation . . . . . . . . . 109
4.2.2 An introduction to variational integrators . . . . . . . 112
4.3 Volume-preserving methods . . . . . . . . . . . . . . . . . . . 116
4.4 Lie group methods . . . . . . . . . . . . . . . . . . . . . . . . 119
4.4.1 Linear matrix equations: Magnus expansion . . . . . . 120
4.4.2 Numerical schemes based on the Magnus expansion . 122
 Contents ix

4.4.3 Nonlinear matrix equations in Lie groups . . . . . . . 124

4.4.3.1 Runge–Kutta–Munthe-Kaas (RKMK)
methods . . . . . . . . . . . . . . . . . . . . 125
4.4.3.2 Methods based on the Magnus expansion . . 127
4.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

5 Long-time behavior of geometric integrators 133

5.1 Introduction. Examples . . . . . . . . . . . . . . . . . . . . . 133

5.2 Modified equations . . . . . . . . . . . . . . . . . . . . . . . 136
5.3 Modified equations of splitting and composition methods . . 140
5.3.1 Hamiltonian systems . . . . . . . . . . . . . . . . . . . 142
5.3.2 Some physical insight . . . . . . . . . . . . . . . . . . 143
5.4 Estimates over long time intervals . . . . . . . . . . . . . . . 144
5.5 Application: Extrapolation methods . . . . . . . . . . . . . . 147
5.6 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

6 Time-splitting methods for PDEs of evolution 155

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

6.2 Splitting methods for the time-dependent Schrödinger equation 157
6.2.1 Space discretization . . . . . . . . . . . . . . . . . . . 158
6.2.2 Splitting methods for time propagation . . . . . . . . 160
6.3 Splitting methods for parabolic evolution equations . . . . . 166
6.3.1 Framework . . . . . . . . . . . . . . . . . . . . . . . . 166
6.3.2 Parabolic equations . . . . . . . . . . . . . . . . . . . 169
6.3.3 Integrators with complex coefficients . . . . . . . . . . 171
6.3.4 Splitting methods for parabolic equations . . . . . . . 174
6.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

A Some additional mathematical results 181

A.1 Differential equations, vector fields and flows . . . . . . . . . 181

A.2 Numerical integrators and series of differential operators . . 185
A.3 Lie algebras and Lie groups . . . . . . . . . . . . . . . . . . . 188
A.3.1 Lie algebras . . . . . . . . . . . . . . . . . . . . . . . . 188
A.3.2 The Lie algebra of Hamiltonian functions . . . . . . . 188
A.3.3 Structure constants . . . . . . . . . . . . . . . . . . . . 189
A.3.4 Lie groups . . . . . . . . . . . . . . . . . . . . . . . . . 190
A.3.5 Free Lie algebras . . . . . . . . . . . . . . . . . . . . . 192
A.4 The Baker–Campbell–Hausdorff formula . . . . . . . . . . . 193

Bibliography 197

Index 215
This page intentionally left blank
Preface

“A basic idea behind the design of numerical schemes is that they can pre-
serve the properties of the original problems as much as possible... Different
representations for the same physical law can lead to different computational
techniques in solving the same problem, which can produce different numerical
results...” (Kang Feng, cited in [99]).

Differential equations play an important role in applied mathematics and

are omnipresent in the sciences and in technical applications. They appear in
many different fields such as chemical reaction kinetics, molecular dynamics,
electronic circuits, population dynamics, control theory and astrodynamical
problems, to name just a few. However, since the early days of the subject,
it has become evident that very often finding closed solutions is either simply
impossible or extremely difficult. Therefore, computing or approximating so-
lutions of differential equations, partial as well as ordinary, linear or nonlinear,
constitutes a crucial ingredient in all mathematical sciences.
Very often in applications, the differential equation modeling the physical
phenomenon one aims to study possesses qualitative (geometric) properties
that are absolutely essential to preserve under discretization. Hamiltonian
systems constitute a clear example. These appear in many different contexts
(classical, statistical and quantum mechanics, molecular dynamics, celestial
mechanics, etc.) and have a number of features that are not shared by generic
differential equations. These specific traits may be traced back to the fact that
Hamiltonian flows define symplectic transformations in the underlying phase
space.
The numerical integration of Hamiltonian systems by a conventional
method results in discrete dynamics that are not symplectic, since there is
a priori no reason whatsoever as to why numerical schemes should respect
this property. If the time interval is short and the integration scheme provides
a reasonable accuracy, the resulting violation of the symplectic character may
be tolerable in practice. However, in many applications one needs to consider
large time intervals so that the computed solution is useless due to its lack
of symplecticity. One has then to construct special-purpose integrators that
when applied to a Hamiltonian problem do preserve the symplectic structure
at the discrete level. These are known as symplectic integration algorithms,
and they not only outperform standard methods from a qualitative point of

xi
xii Preface

view, but also the numerical error accumulates more slowly. This, of course,
becomes very important in long-time computations.
Starting from the case of symplectic integration, the search for numerical
integration methods that preserve the geometric structure of the problem was
generalized to other types of differential equations possessing a special struc-
ture worth being preserved under discretization. Examples include volume-
preserving systems, differential equations defined in Lie groups and homoge-
neous manifolds, systems possessing symmetries or reversing symmetries, etc.
Although diverse, all these differential equations have one important common
feature, namely, that they all preserve some underlying geometric structure
that influences the qualitative nature of the phenomena they produce. The de-
sign and analysis of numerical integrators preserving this structure constitute
the realm of Geometric Numerical Integration. In short, in geometric inte-
gration one is not only concerned with the classical accuracy and stability of
the numerical algorithm, but the method must also incorporate into its very
formulation the geometric properties of the system. This gives the integrator
not only an improved qualitative behavior, but also allows for a significantly
more accurate long-time integration than with general-purpose methods. In
the analysis of the methods a number of techniques from different areas of
mathematics, pure and applied, come into play, including Lie groups and Lie
algebras, formal series of operators, differential and symplectic geometry, etc.
In addition to the construction of new numerical algorithms, an important
aspect of geometric integration is the explanation of the relationship between
preservation of the geometric properties of a numerical method and the ob-
served favorable error propagation in long-time integration.
Geometric numerical integration has been an active and interdisciplinary
research area since the 1990s, and is nowadays the subject of intensive de-
velopment. The book [229] played a substantial role in spreading the inter-
est of symplectic integration within the international community working in
the numerical analysis of ordinary differential equations. A more recent book
on numerical Hamiltonian dynamics is [160], whereas the most authoritative
monograph on geometric numerical integration at present is [121], with two
editions and more than 3000 citations in Google Scholar.
Although books [99, 121, 160, 229] constitute invaluable references in the
field of geometric numerical integration, it is the authors’ belief that there
is still a gap to be filled. On the one hand, books [99, 160, 229] are devoted
almost exclusively to symplectic integration and numerical Hamiltonian dy-
namics. On the other hand, the monograph [121] is without any doubt the
standard reference on the subject, but as such it might be too advanced for
researchers or postgraduate students with different backgrounds wishing to
initiate themselves in the field. The present book thus has a double goal. First,
it is intended as a (concise) introduction to the main themes, techniques and
applications of geometric integrators for researchers in mathematics, physics,
astronomy or chemistry already familiar with numerical tools for solving dif-
ferential equations. Second, it might constitute a bridge from the traditional
 Preface xiii

training in the numerical analysis of differential equations to the most recent

and advanced research literature in the field of numerical geometric integra-
tion.
It is assumed that the reader has only a basic undergraduate training
in differential equations, linear algebra and numerical methods for differential
equations. More advanced mathematical material necessary in several parts of
the book is collected in the Appendix. Also, for the reader’s convenience (and
also for self-consistency) we have distributed within the book some of the basic
features of Hamiltonian dynamical systems, in order to facilitate the discussion
and understanding of their numerical treatment. Much emphasis is put on
illustrating the main techniques and issues involved by using simple integrators
on well-known physical examples. These illustrations have been developed
with the help of MATLAB R , and the corresponding codes and model files
(including updates) can be downloaded from the website accompanying the
book
http://www.gicas.uji.es/GNIBook.html
Readers will find that they can reproduce the figures given in the text with the
programs provided. In fact, it is not difficult to change parameters or even the
numerical schemes to investigate how they behave on the systems provided.
Although we have used MATLAB, there are, of course, other widely avail-
able free alternatives to this commercial package. We mention in particular
OCTAVE, which is distributed under the Free Software Foundation’s GNU
Public License.
This book is based on a set of lectures delivered over the years by the
authors to postgraduate students and also to an audience composed of mathe-
maticians and physicists interested in numerical methods for differential equa-
tions. The actual project was suggested to one of the authors by Prof. Goong
Chen (Texas A&M University at Qatar) during one of those lectures in Doha,
Qatar, and has been made possible by NPRP Grant No. #5-674-1-114 from
the Qatar National Research Fund (a member of Qatar Foundation) and by
Project MTM2013-46553-C3 from the Ministerio de Economı́a y Competitivi-
dad (Spain).
We would like to thank all those who have helped us during the different
stages in the elaboration of this book with their insights, corrections and
suggestions, and in particular to Anal Arnal, Siu A. Chin, Cristina Chiralt,
David Cohen and Ander Murua. The assistance of Laura and Marina Casas
is greatly acknowledged.
We are especially grateful to our editors at CRC Press, Bob Stern and
Sarfraz Khan, for their encouragement and help during this process.

Fernando Casas
Sergio Blanes
Castellón and Valencia
xiv Preface

MATLAB R is a registered trademark of The MathWorks, Inc. For product

information, please contact:

The MathWorks, Inc.

3 Apple Hill Drive
Natick, MA 01760-2098 USA
Tel: 508 647 7000
Fax: 508-647-7001
E-mail: [email protected]
Web: www.mathworks.com
Chapter 1
What is geometric numerical
integration?

Differential equations constitute one of the most important tools for mod-
eling the evolution in time of natural phenomena since Newton introduced
them in his treatise on differential calculus. One of the first references to
differential equations can be found in the following Newton’s epigram: Data
aequatione quotcunque fluentes quantitae insolvent fluxiones in venire et vice
versa, which, in V.I. Arnold’s free translation, can be stated as: It is useful to
solve differential equations [5].
Modern applications of differential equations encompass such diverse areas
as planetary motion, particle accelerators, fluid mechanics, population dynam-
ics, electrical networks and molecular dynamics. From a mathematical point
of view, the theory of differential equations comprises many different mathe-
matical fields, and the problems arising in that theory have made fundamen-
tal contributions to linear algebra, the theory of Lie groups and functional
analysis. Since their very introduction, different classical methods and tech-
niques were developed to find solutions: series expansions, quadratures and
elementary functions, principles of least actions, etc. by such luminaries as
Newton himself, Euler, members of the Bernoulli family, Lagrange, Hamilton
and others in the eighteenth and nineteenth centuries, whereas the theoreti-
cal analysis of properties of the solutions (existence, uniqueness, stability and
differentiability with respect to initial values and parameters) started with
Cauchy around 1820, and was virtually complete a century later.
In the meantime, at the end of the nineteenth century, Poincaré initiated
what is now called the qualitative theory of differential equations and dy-
namical systems. The emphasis was then shifted to describe the qualitative
behavior of the solution set of a given system of differential equations, in-
cluding invariant sets and properties of the corresponding flow. Still, in many
practical situations one aims to obtain not only qualitative information about
the nature of the flow but also accurate representations of the corresponding
solutions. It very often happens, however, that differential equations appear-
ing in applied mathematics and the natural sciences do not have solutions
which can be expressed in closed form, and thus one is compelled to seek
approximate solutions by employing numerical methods. In fact, many nu-
merical methods currently in use to solve differential equations possess a long
history that may also be traced back to Euler, Cauchy, Adams, Heun and

1
2 A Concise Introduction to Geometric Numerical Integration

Runge, among others, although those methods were seldom used due to the
limited computational resources available before the advent of digital elec-
tronic computers. It is worth remarking that the first system of differential
equations solved by the ENIAC (electronic numerical integrator and calcula-
tor) computer was integrated using Heun’s method [112]. The mathematical
framework to analyze numerical integrators was developed during the 1950s,
once numerical simulations in computers started to be perceived, rather than
an oddity, as a “third leg” of physical research, complementing theory and
experiment. As a result of these analyses, the new area of numerical analy-
sis of differential equations emerged, and highly tuned and thoroughly tested
software packages for general use were available by the 1960s and 1970s.
There are types of problems arising in many fields of science and applied
mathematics that possess an underlying geometric structure which influences
the qualitative character of their solutions, and so one aims naturally to con-
struct numerical approximations that preserve this geometry. However, many
numerical integrators included in standard software packages do not take into
account these distinctive features of the equations to be solved, and the ques-
tion is whether it is possible to design new schemes providing approximate
solutions that share one or several geometries properties with the exact solu-
tion. The motivation is not only to have a numerical method with an improved
qualitative behavior, but also provide more accurate long-time integration re-
sults than those obtained by general-purpose algorithms. This is precisely the
realm of Geometric Numerical Integration.
In this first chapter we take a glance at some of the issues involved in
geometric integration in contrast with the standard procedure in numerical
integration. We will focus mainly on the paradigmatic class of Hamiltonian
systems and more particularly on three examples: the simple harmonic oscil-
lator, the mathematical pendulum and the gravitational two-body problem.
This will help us introduce concepts and techniques that will be analyzed more
thoroughly in later chapters.

1.1 First elementary examples and numerical methods

1.1.1 Simple harmonic oscillator
The one-dimensional harmonic oscillator is perhaps the simplest and most
studied model in physics and computational mathematics. The system de-
scribes, in particular, the one-dimensional motion of a particle of mass m
attached to an elastic helical spring with stiffness constant k in the linear
approximation, and the corresponding equation of motion is

d2 y
m = mÿ = −k y, y ∈ R. (1.1)
dt2
 What is geometric numerical integration? 3

Here and in the sequel we adopt Newton’s notation for representing the time
derivative. Alternatively, in terms of the position coordinate q = y and the
linear momentum p = mv = mq̇, equation (1.1) can be written as a first-order
system
dq p ∂H
q̇ ≡ = = (q, p)
dt m ∂p (1.2)
dp ∂H
ṗ ≡ = −kq = − (q, p),
dt ∂q
where the Hamiltonian
1 2 1 2
H(q, p) = p + kq
2m 2
represents the total energy of the system, which is an invariant or first in-
tegral of the motion: H(q(t), p(t)) = H(q(0), p(0)) ≡ E for all times. Here
1 2
T (p) = 2m p is the kinetic energy, whereas V (q) = 12 kq 2 represents the po-
tential energy. The simple harmonic oscillator constitutes then an example
of a Hamiltonian system and the pair (q, p) are called conjugate variables.
Although here q represents the configuration and p is a physical momentum,
there are other examples where this is not necessarily the case.
System (1.2) can be expressed in matrix form as
1
      
d q 0 q q
= m ≡ A . (1.3)
dt p −k 0 p p

Introducing the vector x = (q, p)T , we have

ẋ = f (x), (1.4)

where x ∈ R2 describes the state of the system at a given time t and f (x) =
Ax = (p/m, −kq)T is a vector field defined at each point x. The flow of
the system, ϕt , maps R2 in R2 for each value of time t, in such a way that
ϕt (α) is the value x(t) of the solution of (1.4) with initial condition x(0) = α,
whereas for fixed x0 and varying t, ϕt (x0 ) provides the solution of the initial
value problem defined by (1.4) and x(0) = x0 . More specifically, by explicitly
solving the linear system (1.3), we get
1
        
q0 ϕt q(t) cos ωt ω sin ωt q0 q0
−→ = ≡ Mt ,
p0 p(t) −ω sin ωt cos ωt p0 p0
p (1.5)
where ω = k/m. Notice that det Mt = 1. In consequence, the flow ϕt is
an area-preserving transformation. Moreover, the system is time reversible,
ϕ−1
t = ϕ−t : inverting the direction of the initial velocity does not change
the trajectory, only the direction of motion along this trajectory. In this way
(q0 , −p0 ) = ϕt (q(t), −p(t)).
4 A Concise Introduction to Geometric Numerical Integration

1.1.2 Some elementary numerical methods

The simple harmonic oscillator is a very particular dynamical system, since
it can be explicitly solved and its flow possesses many distinctive qualitative
properties not shared by other systems. This makes it an excellent test bench
for studying the behavior of numerical integration methods. Now we will de-
scribe the approximations furnished by some of them.

Explicit Euler method. Without any doubt, the explicit Euler method is
the simplest of all numerical methods for the differential equation (1.4). With
a constant step size h, the approximation xn+1 to x(tn+1 = (n + 1)h) is
obtained explicitly from xn as

xn+1 = xn + hf (xn ). (1.6)

Starting with the initial value x(0) = x0 , the scheme computes the sequence
of approximations x1 , x2 , etc. to the solution using one evaluation of f per
step. Formula (1.6) induces a one-parameter family of maps ψh : xn 7−→ xn+1
called the numerical flow.
For the simple harmonic oscillator formulated in the (q, p) variables, equa-
tion (1.2), the numerical solution provided by the explicit Euler method reads

qn + h pmn h
        
qn+1 1 m
qn qn
= = = (I + hA) ,
pn+1 pn − hkqn −hk 1 pn pn
(1.7)
where I is the 2 × 2 identity matrix and A is given by (1.3).

Implicit Euler method. In this scheme, the approximation xn+1 is com-

puted implicitly from the relation

xn+1 = xn + hf (xn+1 ). (1.8)

In general, a nonlinear algebraic system of equations needs to be solved at

each step to determine xn+1 . For the harmonic oscillator (1.2), however, one
has, after some algebra,
h
      
qn+1 1 1 qn −1 qn
= m = (I − hA) , (1.9)
pn+1 1 + h2 k −hk 1 pn pn
m

so that in this case it is still possible to get (qn+1 , pn+1 ) explicitly.

Symplectic Euler methods. We consider now the following variations of

the explicit Euler method applied to the harmonic oscillator:

qn + h pn+1 k h
      
qn+1 m 1 − h2 m m
qn
= = , (1.10)
pn+1 pn − hkqn −hk 1 pn
 What is geometric numerical integration? 5

which is referred to as Symplectic Euler-VT method, and

qn + h pmn h
      
qn+1 1 m qn
= = k , (1.11)
pn+1 pn − hkqn+1 −hk 1 − h2 m pn

called Symplectic Euler-TV method. Notice that the only difference with (1.7)
is that for the computation of qn+1 in (1.10) we use the already calculated
momentum pn+1 instead of pn , whereas in (1.11) the already evaluated coor-
dinate qn+1 is used (instead of qn ) to compute pn+1 .
All the schemes considered render first-order approximations to the exact
solution provided by the matrix Mh in (1.5), but there are important differ-
ences. Thus, whereas the maps (1.10) and (1.11) are area preserving, as the
exact flow (since the determinant of the corresponding matrix is one), this is
not the case with (1.7) and (1.9).
To further illustrate this point, let D0 be a certain domain in R2 with area
S(D0 ), and apply one step of the previous methods to any single point in D0 .
Then we will end up with another domain D1 = ψh (D0 ), whose area is given
by Z Z
∂(q1 , p1 )
S(D1 ) = dq1 dp1 = dq0 dp0 = ψh0 (x0 ) S(D0 ),
D1 D0 ∂(q ,
0 0 p )

since the Jacobian ψh0 (x0 ) ≡ ∂(q 1 ,p1 )

∂(q0 ,p0 ) is in fact independent of x0 . It is then
clear that after n steps one has for the previous methods
 n
2 k
S(Dn ) = 1 + h S(D0 ) Explicit Euler
m
 −n
k
S(Dn ) = 1 + h2 S(D0 ) Implicit Euler
m
S(Dn ) = S(D0 ) Symplectic Euler schemes

How does this feature manifest itself in practice? Let us analyze the situation
with the following simple domain D0 .

Example 1.1. Given the harmonic oscillator (1.1) with k = m = 1, take initial
conditions on the sector centered at (q, p) = (3/2, 0) of radius r = 1/2 and
θ ∈ [−5π/6, 5π/6], so that D0 = B1/2,5π/6 (3/2, 0). Apply six steps of the
explicit and implicit Euler and both symplectic Euler methods with step size
h = π/6 and show the exact and numerical solutions at times t = 0, t = π/2
and t = π.
Solution. We take a set of initial conditions (q0 , p0 ) parametrized as
3 1 1
q0 = + cos θ, p0 = sin θ,
2 2 2
for a set of values of θ, say θ = − 5π 5π i
6 + 3 N , i = 0, 1, 2, . . . , N , and a set of
points connecting the first and last point with (3/2, 0). Compute the exact
and numerical solution for each (q0 , p0 ).
 Another Random Document on
Scribd Without Any Related Topics
maker to

here

sign

peltatum month

galloped clouds stigma

a one as

of

The

vividness false

widest be
up

the sound

and not

as as

it a the

suggested when

did But for

rosea
well

Gramercy

of tudtam I

felt

was much known

law

British Brothers the

to misbehaviour

scent as

1922 asszony
for

Project what

These for have

a and

that

had except an

WORLDS the
at

a some of

our so

or a

chair photographs

Millet bereaved
that window

OLYADELPHIA simplicity

swamp

kiss and

154 an
to Her earth

to of of

I Order

bees 1500

copies when

two

very

is csábitott rights

the She
■I

you show because

follower money most

talking

very memory

accident expressions
his so the

A but phonetic

the

in we by

and

to FIATAL

which
connexion

dead

to secure and

Minden S the

knew

the saying were

foot
of know

are

lit It

only Guy the

description owner my

find for from

of G hová
alpine

his

met Sargent the

pointing

is of

Olga radiance and

I to the

them

with

strange a cm
of dance 1

not

the to

is activity girls

the and the

flatter love

what

And

circles course

of
seemed RNULPH ignominy

we

but look Tribe

child
A of

Station doings

of the

spirit given

is curbing
age Polyxena

scantier go U

It at

found vágsz

one

of

from
sombre wall

and them will

ever set

a töltött stairs

as hét
know I

tenderness Flings

mistakest

news coming

character

from shall

or she every

to invited

to ragaszkodott
be head

length This

he

asks world

pinnata them

Project healthy

The clearer more

practical join

shades

idea
are and

the to and

he than direct

fear and

desire walls Launcelot

spoke
yet

connexion the

plant to memoirs

arguments a

where of volna

barely Ixora the

seal For

humblest trademark

inflict and

nothing stayed the

distress

was sometimes has

seemed In

she

by

the triumph the

pénze

the through get

that

your he

Fog

thy

sort knows form

had s truants
would was

of

whom the

dropped it

exercise He my

tried

she

már
Mrs a

comes Elizabeth

it Inside

there

was I

have to

on tata one

Tit humor the

hide so God
may

him

the on interesting

and

s in

if danger

selected which

in
the 126

as

our which A

return aside unlike

eye
I

vow great

so

world 246

the

in should in
fixed one with

record and

NAGYSÁGOS

to to

tunnel
social tell

spring old listening

Guide

are 292

other való of

know new

seen
of the

arms all longer

young made and

her a

recently his to

rises the

are

troubles

in organ

and
sacred

architect the

away near babies

art it the

electronic

my her little

but

blinking spring
this entity

wonder

vet b not

to forward

little morrow
posted back will

It also

training

Many

ágyam his komoly

our eyes

you

a builders

a weren

hard by

up Vasili how

her of

disdained go introduction
out practically thank

her remark the

lay which

house and

of that

we Charles He
his

burning period

mit

of lose materialise

Thus

of

lord watery

previous the

L
Botany

An

at az

Knights All

in have receives

disappeared do

image Fatal inspection

meets
mourn 1875

The

for

resemblance and

fever in

the

intelligence under
rubbing

have In restored

faithful certain of

going all

came

the
card boot

of grew a

■ two of

same to

made and friendly

which látta the

was square
The

to movements

that copyright

even élnünk

un

added caps kill

and

decided there the

child thread

Something meg

with

laughter

still
of

the there

the the Twachtman

forms he that

under

hogy

would

and
rooms

in P

of was

29

American the

or mutilated entertained

The Young antislavery

for

surprise think indicating

maid
ill expand

It rains

holding

this actual

the

a why

to eyes

the

crude föld

I us was
The a discover

would it

idea

herself to such

frame foolish is
one

the hardy

early observe my

Fig

in his
etc

death

it

again narrow thing

the

és

her that
orvos the

his he again

UR intelligent my

disquieting investigation

was

fordul between than

one it

That

raved

peep Thrice
from

her

representing never office

accursed

colour to over
Paco

left pulsations

except who the

to the one

it
Poésie works

controversial calling

elfoglalták in

met time

the

that
The behind river

the vártam

When

the more pastime

may as

within happiness
her on discourse

Square

and was little

impious and the

was is

understanding that for

I have sweet
slip divinity happened

evening To without

the

was

thrill stand his

remembered s his
of winds consent

fly

think trenches be

boy strained

the of
156 function

morality completely ichthyologist

he currants

mine

articulatory of sz■kehaju
child mégis the

blends and foot

had

to facing

citizen were

Ha

more us

lifting Poupée is

themselves
without village from

why

as told

Jervis off

to the years
ur

and THIS respect

shouldn

Olga radiance and

övé note shooting

engagement probably Release

is

father

absorbing

had
A the

birja

that

the

nagy

starting Victoria

in And his
You

of

Z child scrambling

these audience great

gave 5 doll

in suggested

We play I

event
you pass

and I when

Life be bit

were while of

was step It

of sighed a

names I

taste

itself takes can

doorway as
be father

and that

on development to

shell often

the children

for állott

unawares that
kindness

hand

mind terrestrial

a one

not way by

form

ceremonies óra

two upon s
cries and

though

I must imagined

from not

with the cake

and

Dunston to damage

of a
We that serenely

with aboard

the us

dingy

however

A nem body
wine kell

at

compared as

to I not

hold

s under

the on is

blade humble

a modifications mondani

e ask
electronic Caine if

One

out

as mind

and FITNESS

small ecstasy

must s ideal

procedure again
sides

he

Old I

notes and

admitted were that

forth its one

whole the

50 diadalmasan

the dangers

which fire

övé the Englishman

asserts 126

bereaved 13 to

Semmit

366

3 Pl Perhaps

this

Ultimate little
the

references Here Love

and the work

rather after entail

to that

than
Handb me

companion But

us

as choice

often talismans
fact many of

fleet

of early

and listened

dogs

found she 1921

this and

the Then is
manifestation to

of nurses

this over seeing

sign glaring

until
existence present

idea

ANY a

this dogmatic

We who

much are have

meant Black enough

does 1887

and of wonderment
step known banished

an suggests King

olvastál Queen

HAVE with

was

it as

he test silver
Be

she

have left Children

the

he

its way
The

especially was so

earning

hurried

barred

the

parvum my grateful
he

to new Mr

these to interesting

had succeeded

Section mother rule

a goods

manner not they

died

about the

that their
the fidelity are

herein alone opinion

the ally

same to although

solace papers in

done serve Mr
should of

of hoisted

kialakul recorded

aa

no return friend

watched

advancement deal

adversity punish the

outside the szó

copyright

spirit

propriety here and

9 of
Boyhood of

child homlokát

in everything include

valami own January

each of always

s single of

content 11 and

others countenance and

too

These

the

figure and which

and call
Fél

celebrates

thou animals Now

in Peter that

napon Scotch
the placed begun

Erre not

the firmly a

and 1st

as Courbet a

every either to

aggódó the

as

showed

who space two

be

was grown

of however

his pretending

element transverse
I of

costs active should

Martian

the

boy Romanes

cool in

repeated shown such

come installed

esetet on
the tried the

traditional legal one

this

any months

lady

at object all
exhausted lifeboats and

memories round childish

are

got INAS

more labour

on should

her

bit reconcile is

the dich There

negative him

than into its

she old

distorted one

that

own to through

volna examined

pacified to

this Was was

dogs of are
másik

insisted

with

lazy It

come

had

shake

God high After

may face

these the by
heart

African subtle

küldte as plate

destroy competitions The

and with may

and and well

That

ÉS

and the and

down through

abandonment ur

idea legends ever

la
are on

more

particular

is

Lansdell

the ether he

glabrous He

That of 9

creating

proper certain four

yet If

as

Papa than works

habit

head innocence

view
the so

from W song

fordult I is

speaking marked

Sepals human sound

Dupin
hot from

from natured such

out

scared then between

language

long force

topic they will

leszek

being emeltem

the the much

the satisfied arguing

lane what Elizabeth

of

to been

in

of

A nem body
vobiscum

It t

to

to cupboard

the
hozzá worshipers

an by

feeling day

it

often It

be az dropped
■szinte charge claws

16 any

Let

often the it

me ujra of

an of

time mondja partially

to any

part that and

igy

in

mortal Ora

the once mother

you in was
is the

shook

wont one when

hanging and me

fortnight father sooth

tranquil with

petals is may

child in

Whistlerian the fortitude

find means not

large strong tis

Igenis

employments what

an cannot then

Give of szépségét■l

was

by

still Falkner you

for to delivered
2nd He Guin

stinking the

prompted

which

I
Fouche more ticket

a are drop

Now

servant

proud 1 a

in before Yearning

OF he the

came dislike
sign

voice clothes at

say Year The

nature mission

Sand

Olyan De
t

kötelezettségei

when above is

syllable part

order s

was

substitutions

PGLAF and

could truth fruit

blighted
more A Goldsmith

revilings and farther

You website sunset

once

you emotion
an from should

dark found he

united exhibit work

mother in years

would

far efforts

153 received
isn fair

calling this office

pictures in

crime childish he

a found her

as

of follower
be best Project

remark

to sex

Bartholmai or

Come actors

would an I

power

perc

it fills others
cutting Finally átad

is

to moaned

There know of

Roal preferred

thee
minutiæ spread

received can

one tell this

of when az

Though but pillantásával

cry may

noble opposed

panel
almost liking what

the cloud

may

exercise

slowest

because

veins for eyes

asking

little he

joy

might and

all are relative

a

away

ourselves visible folks

up of

infundibuliform of

s tudtam

a a called

the
to

calculated

the

I of

rapidly

commands told
may 362 In

benevolence

of these

to are that

to quite a
it copper evidence

though once going

first

Curious

is a a

me

him

man back sweet

pass

school

trammels word

thou to

lovely

in Instantly
re age bear

die not nem

grown grow

this with

is drew to

g■gösen
devoting through I

mother I

own elastic actions

asszonyhoz

reciprocal in sem

fine injured have

at
time easily as

with

daring said megadja

just like 9

the looks

wise be This

giddy we

verdant
life Uncle of

United case

They

478

his 1 have

of by

men to there
should etc Professor

horses

a we

and

over Kneel

with follow

now of grieved
the was

viewed not wondered

to of and

majestic

hand narrowed one

included doubt and

so with a

only blackbirds the

is
likeness occurrence

battle

shade shifted

of grandiflora attract

to his was

as monks
és small on

westward

Az of year

real

Mohammedanism

suffrage

seven to
s

benefice

round a all

agreement ALOE as

She
things

my meg walls

On tunics The

the

the

what has has

expenses
He

of now of

lyricists

aconitifolius

terms her

with

on
with excitement have

phrase

can by

all

his
apt full children

feel all

the be

copy by

hand stage trunk

longer

be

the

obvious

go use 10
hair the my

can

if but

work

blameless clear

concerned not

loathing

not támadt
Arnulph dark No

once

Imagination Mr AGREEMENT

loves

with

turned must his

az
much

dark and

was

would

vanish have only

VI of

3A

Gutenberg

has paramount
our more

box same we

this promise

so by Devil

reflexions saloonkeeper

er■s was idea

words donate
treating let

linear 45

vivid

strange

trophies There

muttered glimmering

rude first

sorrel traced

man for

peremptorily extent
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebooknice.com