Practical Optimization - Algorithms and Engineering Applications
B-ngn{Yc T@2#6Tffo by
m% -g ~q j*so9M6|c Andreas Antoniou
hYPl&^ Wu-Sheng Lu
[ 9 {*94M Department of Electrical and Computer Engineering
$rv8K j+ University of Victoria, Canada
KL1/^1 !t$'AoVBq 2007 Springer Science+Business Media, LLC
2@'oe7E E ~Sb Preface
3!XjtVhK?I =,spvy'"*C The rapid advancements in the efficiency of digital computers and the evolution
$PTP/^ of reliable software for numerical computation during the past three
:61Tun decades have led to an astonishing growth in the theory, methods, and algorithms
vK|E>nL of numerical optimization. This body of knowledge has, in turn, motivated
^1iSn)& widespread applications of optimization methods in many disciplines,
}vxw*8d? e.g., engineering, business, and science, and led to problem solutions that were
UO0{):w> considered intractable not too long ago.
DFZkh^PFd Although excellent books are available that treat the subject of optimization
=RA / with great mathematical rigor and precision, there appears to be a need for a
DS+}UO book that provides a practical treatment of the subject aimed at a broader audience
qE&v ; ranging from college students to scientists and industry professionals.
]
o*#t This book has been written to address this need. It treats unconstrained and
C'\-
@/ constrained optimization in a unified manner and places special attention on the
t<#mP@Mz=N algorithmic aspects of optimization to enable readers to apply the various algorithms
^Cu\VV and methods to specific problems of interest. To facilitate this process,
JjPKR?[> the book provides many solved examples that illustrate the principles involved,
*X'Y$x>f and includes, in addition, two chapters that deal exclusively with applications of
^t` k0< unconstrained and constrained optimization methods to problems in the areas of
@$S+ Ne[< pattern recognition, control systems, robotics, communication systems, and the
nw -xSS{ design of digital filters. For each application, enough background information
Jj!vh{ is provided to promote the understanding of the optimization algorithms used
i
UCXAWP to obtain the desired solutions.
}\tdcTMgS Chapter 1 gives a brief introduction to optimization and the general structure
+T|JK7 of optimization algorithms. Chapters 2 to 9 are concerned with unconstrained
G18F&c~ optimization methods. The basic principles of interest are introduced in Chapter
]zj9A]i:a 2. These include the first-order and second-order necessary conditions for
RDW8]=uM a point to be a local minimizer, the second-order sufficient conditions, and the
@e(o129 optimization of convex functions. Chapter 3 deals with general properties of
}Lc-7[/ algorithms such as the concepts of descent function, global convergence, and
sGyeb5c XVI
7X0Lq}G@ rate of convergence. Chapter 4 presents several methods for one-dimensional
k;K)xb[w | optimization, which are commonly referred to as line searches. The chapter
i6dHrx]:, also deals with inexact line-search methods that have been found to increase
sy:[T T!w the efficiency in many optimization algorithms. Chapter 5 presents several
6t>.[Y"v basic gradient methods that include the steepest descent, Newton, and Gauss-
|^: cG4e Newton methods. Chapter 6 presents a class of methods based on the concept of
Gw>^[dmt! conjugate directions such as the conjugate-gradient, Fletcher-Reeves, Powell,
d"}k!
0m and Partan methods. An important class of unconstrained optimization methods
EYtL_hNp}I known as quasi-Newton methods is presented in Chapter 7. Representative
Vq2y4D? methods of this class such as the Davidon-Fletcher-Powell and Broydon-
.aO,8M Fletcher-Goldfarb-Shanno methods and their properties are investigated. The
&+Z,hs9% chapter also includes a practical, efficient, and reliable quasi-Newton algorithm
|L%Z,:yO that eliminates some problems associated with the basic quasi-Newton method.
aoMqSwF= Chapter 8 presents minimax methods that are used in many applications including
D^yRaP*|7 the design of digital filters. Chapter 9 presents three case studies in
QS%%^+E2 which several of the unconstrained optimization methods described in Chapters
\Jx04[= 4 to 8 are applied to point pattern matching, inverse kinematics for robotic
)WRLBFi3 manipulators, and the design of digital filters.
?I6fye7 Chapters 10 to 16 are concerned with constrained optimization methods.
m? eiIrMW Chapter 10 introduces the fundamentals of constrained optimization. The concept
C7rNV0.Fq of Lagrange multipliers, the first-order necessary conditions known as
JJP08oP Karush-Kuhn-Tucker conditions, and the duality principle of convex programming
{G U&a are addressed in detail and are illustrated by many examples. Chapters
|jI#"LbF 11 and 12 are concerned with linear programming (LP) problems. The general
10)jsA properties of LP and the simplex method for standard LP problems are
|SoCRjuCPM addressed in Chapter 11. Several interior-point methods including the primal
pWy=W&0~qf affine-scaling, primal Newton-barrier, and primal dual-path following methods
qaY1xPWz" are presented in Chapter 12. Chapter 13 deals with quadratic and general
iq8GrdL" convex programming. The so-called active-set methods and several interiorpoint
ykl
.1( methods for convex quadratic programming are investigated. The chapter
u[@l~gwL also includes the so-called cutting plane and ellipsoid algorithms for general
w`4=_J=GO convex programming problems. Chapter 14 presents two special classes of convex
+]*zlE\N` programming known as semidefinite and second-order cone programming,
VCY\be which have found interesting applications in a variety of disciplines. Chapter
S=5<^o^h3 15 treats general constrained optimization problems that do not belong to the
?u{~> class of convex programming; special emphasis is placed on several sequential
F<n3 quadratic programming methods that are enhanced through the use of efficient
QKyo`g7 line searches and approximations of the Hessian matrix involved. Chapter 16,
iP
=V8g?L which concludes the book, examines several applications of constrained optimization
+MR.>" for the design of digital filters, for the control of dynamic systems, for
vMA]j>> evaluating the force distribution in robotic systems, and in multiuser detection
n!YKz"$ for wireless communication systems.
!TAlBkj PREFACE xvii
8J:}%DaxL The book also includes two appendices, A and B, which provide additional
AP68V support material. Appendix A deals in some detail with the relevant parts of
PPb7%2r linear algebra to consolidate the understanding of the underlying mathematical
&M46&^Jho principles involved whereas Appendix B provides a concise treatment of the
pOGeruu? basics of digital filters to enhance the understanding of the design algorithms
{_7i8c<s= included in Chaps. 8, 9, and 16.
C<NLE- The book can be used as a text for a sequence of two one-semester courses
iX0i2ek on optimization. The first course comprising Chaps. 1 to 7, 9, and part of
,5i` -OI Chap. 10 may be offered to senior undergraduate or first-year graduate students.
W#^2#sjO The prerequisite knowledge is an undergraduate mathematics background of
)Ub_@)X3%l calculus and linear algebra. The material in Chaps. 8 and 10 to 16 may be
_7H7
dV used as a text for an advanced graduate course on minimax and constrained
8K.R= optimization. The prerequisite knowledge for thi^ course is the contents of the
Id_2PkIN$~ first optimization course.
Wr<j!>J6Ki The book is supported by online solutions of the end-of-chapter problems
/ :
L ?~ under password as well as by a collection of MATLAB programs for free access
u?4:H=;> by the readers of the book, which can be used to solve a variety of optimization
n,jKmA problems. These materials can be downloaded from the book's website:
i*|\KM?P http://www.ece.uvic.ca/~optimization/. "Ue.@> We are grateful to many of our past students at the University of Victoria,
Mmxlp.l in particular, Drs. M. L. R. de Campos, S. Netto, S. Nokleby, D. Peters, and
5]NqRI^0 Mr. J. Wong who took our optimization courses and have helped improve the
FG%X~L<d,) manuscript in one way or another; to Chi-Tang Catherine Chang for typesetting
[BWNRC1 the first draft of the manuscript and for producing most of the illustrations; to
W=j[V
Oq R. Nongpiur for checking a large part of the index; and to R Ramachandran
_Tf4WFu2 for proofreading the entire manuscript. We would also like to thank Professors
Yk)."r&