Handbook of Splines

Mathematics and Its Applications

Managing Editor:

M . HAZEWINKEL
Centre for Mathematics and Computer Science, Amsterdam, The Netherlands

Volume 462
Handbook of
Splines
by

Gheorghe Micula
University of Cluj-Napoca,
Cluj-Napoca, Romania

and

Sanda Micula
Western Oregon University,
Monmouth, Oregon

SPRINGER SCIENCE+BUSINESS MEDIA, B.V.

A CLP. Catalogue record for this book is available from the Library of Congress.

ISBN 978-94-010-6244-2 ISBN 978-94-011-5338-6 (eBook)

DOI 10.1007/978-94-011-5338-6

Printed on acid-free paper

All Rights Reserved

©1999 Springer Science+Business Media Dordrecht
Originally published by Kluwer Academic Publishers in 1999
Softcover reprint of the hardcover 1st edition 1999
No part of the material protected by this copyright notice may be reproduced or
utilized in any form or by any means, electronic or mechanical,
including photocopying, recording or by any information storage and
retrieval system, without written permission from the copyright owner
 CONTENTS

PREFACE Xl

1 SPLINE FUNCTIONS AND THE REPRESENTATION

OF LINEAR FUNCTIONALS 1
1.1 Some Problems Leading to Spline Functions . . . . . .. 1
1.2 Polynomial Spline Functions. Definitions and Properties 3
1.3 Local Bases for Spline Spaces . . . . 7
1.4 Periodic Spline Functions. . . . . . . 23
1.5 Natural Polynomial Spline Functions 26
1.6 Discrete Spline Functions. . . . . . . 31
1. 7 Existence and Uniqueness of Interpolating Spline Func-
tions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
1.8 Practical Methods for Calculating Cubic Spline Functions 53
1.9 Another Kind of End Conditions for Cubic Spline Func-
tions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
1.10 Representation of Linear Functionals. Applications in
Numerical Integration and Numerical Differentiation of
Functions . . . . . . . . . . . . . . . . . . . . . . . . . . 62
1.11 Optimal Approximation of Linear Functionals. Optimal
Approximation Formulas. . . . . . . . . . . . . . . . . . 69

2 MULTIVARIATE SPLINE FUNCTIONS 79

2.1 Definitions and Preliminaries. . . . . . . . . . . . . . . . 79
2.2 Bicubic Spline Interpolation Functions of Hermite - Type 82
2.3 Dimension of Bivariate Spline Spaces . . . . . . 86
2.4 Bezier Nets and Local Bases . . . . . . . . . . . . . .. 89
2.5 Estimate of the Approximation Power of Splines. . .. 93
2.6 Interpolation and Fitting. Extremal Property of Spline 95
2.7 On the Space of Trivariate Splines. . . . . . . . . 98

3 NONLINEAR SETS OF SPLINE FUNCTIONS 103

3.1 Introduction..................... . 103
3.2 Rational Spline Functions . . . . . . . . . . . . . . 104
3.3 Existence and Uniqueness of Interpolating Rational Spline
Functions . . . . . . . . . . . . . . . . . . . . . . . . . . 105
vi HANDBOOK OF SPliNES

3.4 Regular Spline Functions . . . . . . . . . . . . . . . . . . 111

3.5 Nonlinear Spline Functions to the Numerical Solution of
Differential Equations . . . . . . . . . . . . . . . . . . . . 119

4 NUMERICAL TREATMENT OF THE INTEGRAL E-

QUATIONS 129
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 129
4.2 Spline - Galerkin Method for Solving Linear Fredholm
Integral Equation of Second Kind . . . . . . . . . . . . . 133
4.3 The Discrete Spline - Galerkin Method for Nonlinear
Fredholm Integral Equations . . . . . . . . . . . . . . . . 137
4.4 Spline Collocation Methods for Volterra Integral Equations 151
4.4.1 Description of the Spline Collocation Method .. 151
4.4.2 Iterated Collocation Approximations . . . . . . . 159
4.5 Polynomial Spline Collocation for Volterra Integro-dif-
ferential Equations . . . . . . . . . . . . . . . . . . . . . 164
4.5.1 Description of Repeated Spline Collocation Method166
4.5.2 The Discretized Repeated Collocation Method . . 173

5 NUMERICAL SOLUTION OF ORDINARY DIFFER-

ENTIAL EQUATIONS 179
5.1 Polynomial Spline Functions of Even Degree. Basic Def-
initions and Properties . . . . . . . . . . . . . . . . . . . 179
5.2 The Numerical Solutions of Differential Equations by
Spline Functions of Even Degree . . . . . . . . . . . . . . 186
5.3 The Numerical Solutions of Retarded Differential Equa-
tion by Natural Spline Functions of Even Degree . . . . . 195
5.4 The Spline Collocation Method . . . . . . . . . . . . . . 202
5.4.1 Preliminary Notions. Setting the Problem of Col-
location . . . . . . . . . . . . . . . . . . . . . . . 202
5.4.2 Collocation with Splines from S~ (L\N) . . . . . . 204
5.4.3 Collocation with Splines from S::;:(L\N)' Runge-
Kutta Collocation . . . . . . . . . . . . . . . . . . 205
5.4.4 Collocation with Splines from S~ (L~N), 1:::; d:::; m 208
5.4.5 A Class of Deficient-Spline Methods of Nordsieck
Type (DSIN) . . . . . . . . . . 211
5.4.6 Stability of Spline Collocation . . . . . . . . . . . 215
 vii

5.5 Numerical Solution of Boundary Value Problems by Spline

Functions . . . . . . . . . . . . . . . . . . . . . . . . . . 222
5.5.1 Cubic Spline Method for Linear Second - Order
Two - Point Boundary Value Problems . . . . . . 222
5.5.2 A Fourth-Order Cubic Spline Method for Second
- Order Nonlinear Two-Point Boundary Value Prob-
lems . . . . . . . . . . . . . . . . . . . . . . . . . 228

6 SPLINES AND FINITE ELEMENTS 235

6.1 Introduction . . . . . . . . .235
6.2 Lagrange Finite Elements . . . . . . .236
6.3 Hermite Finite Elements . . . . . . . .240
6.4 Triangulations and Spline Functions . . 243
6.5 Family of Lagrange Finite Elements . . 245
6.6 Interpolation by Continuous Spline Functions . 248
6.7 Hermite Interpolation on a Triangulation and Spline Func-
tions of the Class G1 (0) and of the Class GO(O) . . . . . 251

7 FINITE ELEMENT METHOD FOR SOLUTION OF

BOUNDARY PROBLEMS FOR PARTIAL DIFFER-
ENTIAL EQUATIONS 257
7.1 Variational Problems . . . . . . . . . . . . . . . . . . . . 257
7.1.1 The space HJ(O) . . . . . . . . . . . . . . . . . . 257
7.1.2 Variational Formulations ofthe Elliptic Dirichlet's
and Neumann's Problems . . . . . . . . . . . . . 260
7.1.3 General Variational problems . . . . . . . . . . . 261
7.1.4 Other Variational Formulations for Boundary Value
Elliptic Problems . . . . . . . . . . . . . . . . . . 265
7.2 Approximations of the Variational Problems and the Fi-
nite Element Methods . . . . . . . . . . . . . . . . . . . 270
7.2.1 Galerkin's Method for Approximation of the Vari-
ational Problems . . . . . . . . . . . . . . . . . . 270
7.2.2 Finite Element Method . . . . . . . . . . . . . . . 272
7.2.3 The Influence of the Cubature Formulae on the
Approximation by Finite Element Method . . . . 276
7.3 Spline Collocation Methods for Parabolic and Hyper-
bolic Problems in two Space Variables . . . . . . . . . . 281
viii HANDBOOK OF SPliNES

7.4 A Finite Element Solution for the Korteweg-de Vries-

Burgers Equation Using Splines . . . . . . . . . . . . . . 288
7.4.1 Introduction . . . . . . . . . . . . . . . . . . . . . 288
7.4.2 Finite Element Solution for the Korteweg-de Vries-
Burgers Equation using Quadratic B-splines . . . 289

8 SPLINE FUNCTIONS IN COMPUTER AIDED GEO-

METRIC DESIGN 295
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 295
8.2 Spline Curves, Spline Surfaces and their B-Spline Rep-
resentations . . . . . . . . . . . . . . . . . . . . . . . . . 302
8.3 Spline Curve Interpolants and Spline Curve Approximants306
8.4 Spline Surface Interpolants and Spline Surface Approxi-
mants . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
8.5 On Non-Uniform Rational B-Splines (NURBS) in CAGD 314
8.6 Surface Interpolation and Smoothing by "Thin Plate"
Splines . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321

9 FROM SPLINE TO FRACTALS 325

9.1 Introduction . . . . . . . . . . . . .325
9.2 The Constrained Fractal Approach .326
9.3 Variational Splines .. . . . . . . . .328
9.4 Converting Energies into Probabilities .332
9.5 Multiresolution Stochastic Relaxation . .334

10 BOX SPLINES 337

10.1 Introduction . . . . . . . . . . . . . . . . . . .337
10.2 Box Splines. Definition and Basic Properties .339
10.3 Multivariate Truncated Powers . . . . . . . .344
10.4 Box Spline Series . . . . . . . . . . . . . . . . 347
10.5 Box Spline-Based Approach to the Numerical Method
for Some Partial Differential Equations . . . . . . . . . . 349

11 SPLINE WAVELETS 357

11.1 Introduction . . . . .357
11.2 Multiresolution Analysis, Splines and Wavelets. . 359
11.3 Wavelet Decomposition and Reconstruction .. . 363
 ix

11.4 The Two Scale Relation . . . . . . . . . . . . . .365

11.5 An Interpolatory Graphical Display Algorithm. .367
11.6 Compactly Supported Wavelets . . . . . . . . . .369
11.7 Computation of Cardinal Spline Wavelets. . . . .372
11.8 Spline-wavelets in Numerical Solution of Partial Differ-
ential Equation Problems. . . . . . . . . . . . . . . . . . 378

12 REF ERE NeE S 383

12.1 Books, Monographs and Proceedings .383
12.2 Original Papers . . . . .397
12.3 Doctoral Dissertations .. . . . . . . .589

INDEX 601
 PREFACE

The purpose of this book is to give a comprehensive introduction

to the theory of spline functions, together with some applications to
various fields, emphasizing the significance of the relationship between
the general theory and its applications.
At the same time, the goal of the book is also to provide new ma-
terial on spline function theory, as well as a fresh look at old results,
being written for people interested in research, as well as for those who
are interested in applications.
The theory of spline functions and their applications is a relatively
recent field of applied mathematics.
In the last 50 years, spline function theory has undergone a won-
derful development with many new directions appearing during this
time. This book has its origins in the wish to adequately describe this
development from the notion of 'spline' introduced by 1. J. Schoenberg
(1901-1990) in 1946, to the newest recent theories of 'spline wavelets' or
'spline fractals'. Isolated facts about the functions now called 'splines'
can be found in the papers of L. Euler, A. Lebesgue, G. Birkhoff, J.
Favard, L. Tschakaloff, L. Collatz, T. Popoviciu, D. V. Ionescu. How-
ever, the theory of spline functions has been developed in the last 40
years through the effort of many mathematicians. As late as 1960, there
were no more than a handful of papers mentioning spline functions by
name. Today, less than 40 years later, there are, to our knowledge, more
than 357 books, monographies and conference reports, many thousands
original papers, and more than 311 dissertations for a doctoral degree
or habilitation, on various aspects of spline functions and their appli-
cations, and it is still an active research area.
The rapid development of spline functions is due primarily to their
great usefulness in applications. Classes of spline functions posses many
nice structural properties as well as excellent approximation powers.
Since they are easy to evaluate and manipulate on a computer, a myr-
iad of applications to the numerical solution of a variety of problems
in applied mathematics has been found. The enormous literature pub-
lished during the last decades, shows that the actual development of
spline theory has an essential influence on large areas of modern nu-
xi
xii HANDBOOK OF SPUNES

merical mathematics, such as:

• data fitting, interpolation and approximation

• numerical integration and differentiation

• numerical treatment of integral, differential and partial differen-

tial equations

• optimal approximation

• calculation of eigenvalues and eigenfunctions of operators

• control theory

• computer aided geometric design and computational geometry

• numerical methods of probability theory and statistics

• wavelets and fractals

While this book attempts to give a comprehensive treatment of the

basic methods in spline functions and their applications, it is not meant
to provide solutions to all problems that have arisen in this field in the
last years. In particular, no results are included concerning spline func-
tions of a complex variable, nor those concerning the stimulation and
approximation of stochastic processes by real spline functions. Most
of the sections contain remarks directed to the literature involving re-
lated results and applications or further developments of the material
covered in these sections.
There is no claim for completeness and there is definitely a num-
ber of readers that will find some of their favorite areas missing. The
omissions in some instances are deliberate and in others, no doubt,
oversights. However, it is our hope that enough information has been
included which may be of interest to the reader.
A detailed analysis of many of the most important applications in-
volves a great deal of material in the specific area from which they are
taken. Therefore, an attempt has been made by the authors to include
comprehensive (if not quite exhaustive) references to the literature, to
 xiii

enable the interested readers to find as much of the omitted material

as they wish.
The book is organised as follows:
Chapter 1 introduces the polynomial spline functions and their
fundamental properties. The existence and uniqueness of interpolating
spline functions (natural, discrete and periodic) are investigated, and
local bases for spline spaces are constructed. Practical methods for
calculating cubic spline functions with various end conditions are given.
The optimal approximation of linear functionals concludes this chapter.
Chapter 2 is devoted to multivariate spline functions (especially
the bivariate case). The dimension, Bezier nets, and local bases of
bivariate spline spaces are discussed. The estimate of the approxima-
tion power, as well as the extremal property of interpolating bivariate
splines are underlined.
Chapter 3 treats the nonlinear sets of spline functions, especially
rational and regular spline functions and their application to the nu-
merical solution of ordinary differential equations.
Chapter 4 presents the most important methods for numerical
treatment of integral equations problems by means of spline functions.
The spline - Galerkin method and the spline collocation method are
derived for Fredholm integral equations, as well as for Volterra integral
and integro - differential equations.
Chapter 5 contains a detailed presentation of the application of
spline functions to the numerical solution of ordinary differential equa-
tions with given initial boundary conditions, as well as for the delay
differential equation problems. The connection between spline colloca-
tion methods and special Runge - Kutta - Nordsieck methods, and also
the stability of spline collocation methods are investigated. The cubic
spline method for two point boundary value problems for both linear
and nonlinear equations is presented.
Chapter 6 shows that the most natural framework for using spline
functions is the finite element methods. The Lagrange and Hermite
finite elements are introduced by means of spline interpolating functions
on a suitable triangulation.
Chapter 7 is a thorough presentation of the finite element method
for solution of boundary value problems for partial differential equa-
tions.
xiv HANDBOOK OF SPliNES

Using suitable spaces of spline functions the general finite element

methods for elliptic Dirichlet and Neumann problems are developed.
The spline collocation methods for parabolic and hyperbolic problems
in two space variables are also discussed.
A finite element solution for the Korteweg - de - Vries - Burgers
Equations using splines proves that important cases of nonlinear equa-
tion problems can be numerically solved by spline functions.
Chapter 8 is devoted to spline curves, spline surfaces and their B -
spline representation for Computer Aided Geometric Design (CAGD).
The rational point of view is also studied, especially the problem and
the usefulness of Non - Uniform Rational B - Splines (NURBS) in
CAGD. The chapter ends with surface interpolation and smoothing by
"Thin Plate}' splines.
Chapter 9 describes briefly a model of shape which combines deter-
ministic splines and stochastic fractals inheriting their complementary
features. The main text develops the variational splines in a single
variable and it describes how to discretize the spline energy expressions
using finite elements. It is shown how to associate a probability dis-
tribution with variational splines through the Boltzmann distribution
and that samples from this distribution have fractal statistics. The
multiresolution stochastic relaxation algorithm for sampling the Boltz-
mann distribution to obtain constrained fractals is investigated.
Chapter 10 introduces the notions of box splines and multivariate
truncated powers, following the convolution procedure. Some of the
basic properties, including a recurrence relation, relationships between
box splines and truncated powers, and approximation order are pre-
sented. A multivariate box splines framework for the formulation of
numerical methods for some partial differential equations is also inves-
tigated.
Chapter 11 is a brief introduction to one of the most exciting
recent developments in modern mathematics, which is called "wavelet
analysis" or as it is known ussually ,"wavelets". The Multiresolution
Analysis, Splines and Wavelets form a very important mathematical
triplet. The main results concerning Compactly Supported Wavelets
and Computation of Cardinal Spline Wavelets are presented. Some
new aspects of the numerical methods using spline wavelets for the
solution of evolution partial differential equations are discussed.
 xv

The references to the literature contained at the end of this book

try to be the most exhaustive possible. To make the list as complete as
possible, the first author has exhausted all sources available over many
years. All publications that have become known to us until August
1998 are listed, subdivided in three sections:

1. books, monographies and conference reports

2. original papers

3. dissertations for doctoral degree or habilitation

Papers published in conference proceedings and reports are not men-

tioned separately. For each cited reference, we mentioned only the year
of publication after the name of the author, which may force the reader
to find the paper himself, in the case where more than one paper was
published that year.
It has taken us much longer than we expected to produce this book.
Nevertheless, we think that the ideas and the content of this book are
still timely, and we trust that these topics are of ongoing interest to a
large segment of the community devoted to research and development
of splines for use in numerical treatment of differential and integral
equations and many other fields of applied mathematics.
We can only hope that the readers learn as much from reading this
book as the authors did from writing it.
The authors would like to thank many people who have helped us
with information, hints, advice, fruitful discussions, or who have sent us
their latest publications and results, or helped us in many other ways.
The first author wishes to express his thanks to Alexander von Hum-
boldt Foundation and to DAAD (German Scientific Exchange Service)
for granting him several research stays in Germany and also to Fulbright
Foundation for granting him one academic year (1990/1991) at the Uni-
versity of Kentucky, Lexington, U.S.A. Thus, he has had many oppor-
tunities for stimulating exchange of ideas with many mathematicians
and also excellent mathematical sources of information. He would also
like to take this opportunity to aknowledge some of his colleagues for
their cooperation, help and assistance in the preparation of this book,
during the time he has been a visiting professor or researcher for longer
xvi HANDBOOK OF SPliNES

or shorter time at their universities abroad. In particular, he would

especially like to mention Professors: A. Bellen (Univ. of Trieste), K.
Bohmer (Univ. of Marburg), R. Gorenflo (Free Univ. of Berlin), W.
Jager (Univ. of Heidelberg), H. N. Miilthei (Univ. of Mainz), J. Stoer
(Univ. of Wiirzburg), W. Tornig (Univ. of Darmstadt), W. Wendland
(Univ. of Stuttgart), G. Fairweather (Univ. of Kentucky at Lexington),
L. L. Schumaker (Vanderbilt Univ. at Nashville, Tennessee), J. Butcher
(Univ. of Auckland), H. Brunner (Memorial Univ. of Newfoundland,
St. John's), Reng Hong Wang (Univ. of Dalian).
The second author wishes to aknowledge her great debt and grati-
tude to Professor K. Atkinson for introducing her to applied mathemat-
ics research and for guiding her as a Ph. D. student at the University
of Iowa, Iowa City, U. S. A.
Both authors are deeply indebted to their colleagues in the De-
partment of Applied Mathematics of the University of Cluj-Napoca,
Romania, for their consideration and friendliness, as well as for sharing
their valuable insight.
Finally, we would like to thank the staff and members of Kluwer
Academic Publishers for thier cooperation and understanding.

G. Micula and Sanda Micula,

Cluj-Napoca, Romania and Monmouth, Oregon, U. S. A.,
August 1998