Introduction to Optimum Design,

Fourth Edition Arora

Visit to download the full and correct content document:
https://textbookfull.com/product/introduction-to-optimum-design-fourth-edition-arora/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Embedded System Design Introduction to SoC System

Architecture 1st Edition Mohit Arora

https://textbookfull.com/product/embedded-system-design-
introduction-to-soc-system-architecture-1st-edition-mohit-arora/

Biota Grow 2C gather 2C cook Loucas

https://textbookfull.com/product/biota-grow-2c-gather-2c-cook-
loucas/

Optimum Design and Manufacture of Wood Products Etele

Csanády

https://textbookfull.com/product/optimum-design-and-manufacture-
of-wood-products-etele-csanady/

Introduction to Polymer Chemistry, Fourth Edition

Carraher Jr.

https://textbookfull.com/product/introduction-to-polymer-
chemistry-fourth-edition-carraher-jr/
Introduction to Electrodynamics Fourth Edition David J.
Griffiths

https://textbookfull.com/product/introduction-to-electrodynamics-
fourth-edition-david-j-griffiths/

Introduction to health care Fourth Edition. Edition

Dakota Mitchell

https://textbookfull.com/product/introduction-to-health-care-
fourth-edition-edition-dakota-mitchell/

Digital and Kalman Filtering An Introduction to

Discrete Time Filtering and Optimum Linear Estimation
2nd Edition S. M. Bozic

https://textbookfull.com/product/digital-and-kalman-filtering-an-
introduction-to-discrete-time-filtering-and-optimum-linear-
estimation-2nd-edition-s-m-bozic/

A Concise Introduction to Pure Mathematics Fourth

Edition Liebeck

https://textbookfull.com/product/a-concise-introduction-to-pure-
mathematics-fourth-edition-liebeck/

Introduction to Human Factors and Ergonomics, Fourth

Edition Robert Bridger

https://textbookfull.com/product/introduction-to-human-factors-
and-ergonomics-fourth-edition-robert-bridger/
INTRODUCTION
TO OPTIMUM
DESIGN
FOURTH EDITION
Jasbir Singh Arora
The University of Iowa,
College of Engineering,
Iowa City, Iowa

AMSTERDAM • BOSTON • HEIDELBERG • LONDON

NEW YORK • OXFORD • PARIS • SAN DIEGO
SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO

Academic Press is an imprint of Elsevier


Academic Press is an imprint of Elsevier
125 London Wall, London EC2Y 5AS, UK
525 B Street, Suite 1800, San Diego, CA 92101-4495, USA
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, USA
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK

Copyright © 2017, 2012, 2004, Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic
or mechanical, including photocopying, recording, or any information storage and retrieval system,
without permission in writing from the publisher. Details on how to seek permission, further infor-
mation about the Publisher’s permissions policies and our arrangements with organizations such as
the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website:
www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the
Publisher (other than as may be noted herein).

Notices
Knowledge and best practice in this field are constantly changing. As new research and experience
broaden our understanding, changes in research methods, professional practices, or medical treatment
may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating
and using any information, methods, compounds, or experiments described herein. In using such infor-
mation or methods they should be mindful of their own safety and the safety of others, including parties
for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume
any liability for any injury and/or damage to persons or property as a matter of products liability,
negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas
contained in the material herein.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

ISBN: 978-0-12-800806-5

For information on all Academic Press publications

visit our website at http://elsevier.com/
 To
Rita
Ruhee
Matthew
and
in memory of my parents
Balwant Kaur
Wazir Singh
 Preface to Fourth Edition
Introduction to Optimum Design, Fourth Edi-
tion presents an organized approach to opti-
mizing the design of engineering systems.
Basic concepts and procedures of optimiza-
tion are illustrated with examples to show
their applicability to engineering design
problems. The crucial step of formulating
a design optimization problem is detailed.
The numerical process of refining an initial
formulation of the optimization problem
is illustrated and how to correct the initial
formulation if it fails to produce a solution
or does not produce an acceptable solution.
This as well as other numerical aspects of
the solution process for optimum design are
consolidated in chapter: Optimum Design:
Numerical Solution Process and Excel Solver.
The pedagogical aspects of the material
are addressed from student as well as teach-
er’s point of view. This is particularly true
for the material of chapters: Introduction
to Design Optimization, Optimum Design
Problem Formulation, Graphical Solution
Method and Basic Optimization Concepts,
Optimum Design Concepts: Optimality Con-
ditions, More on Optimum Design Concepts:
Optimality Conditions, Optimum Design:
Numerical Solution Process and Excel Solver,
Optimum Design with MATLAB®, Linear
Programming Methods for Optimum Design,
More on Linear Programming Methods for
Optimum Design, Numerical Methods for
Unconstrained Optimum Design, More on
Numerical Methods for Unconstrained Opti-
mum Design, and Numerical Methods for
Constrained Optimum Design that forms the
xv
basis for a first course on optimum design.
More detailed calculations are included in
examples to illustrate procedures. Key ideas
of various concepts are highlighted for easy
reference.
Practical applications of optimization
techniques are expanding rapidly, and to
illustrate this, engineering design exam-
ples are presented throughout the text.
Chapters: Optimum Design: Numerical
­
Solution Process and Excel Solver, Opti-
mum Design with MATLAB®, and Practical
Applications of Optimization also address
various aspects of practical applications of
optimization with example problems. Many
projects (identified with a *) based on prac-
tical applications are included as exercises
at the end of c­hapters: Optimum Design
Problem Formulation, Graphical Solution
Method and Basic Optimization Concepts,
Optimum Design: Numerical Solution Pro-
cess and Excel Solver, Optimum Design
with MATLAB®, Practical Applications of
Optimization, Discrete Variable Optimum
Design Concepts and Methods, and Nature-
Inspired Search Methods.
Direct search methods that do not use any
stochastic ideas in their search are presented
in chapter: More on Numerical Methods for
Unconstrained Optimum Design and those
that do are presented in chapter: Nature-
Inspired Search Methods. These methods do
not require derivatives of the problem func-
tions, therefore they are broadly applicable
for practical applications and they are rela-
tively easy to program and use.
xvi Preface to Fourth Edition

The instructor’s manual that accompanies for students. Three courses are suggested,
the text is comprehensive and user-friendly. although several variations are possible.
The original problem statement for each
exercise is included in the solution page for Undergraduate/First-Year Graduate
easy reference. Many MATLAB and Excel
Level Course
files for the problems are also included.
The text is broadly divided into three Topics for an undergraduate and/or first-
parts. Part I presents the basic concepts year graduate course include:
related to optimum design and optimal-
• Formulation of optimization problems
ity conditions in chapters: Introduction to
(see chapters: Introduction to Design
Design Optimization 2 Optimum Design
Optimization and Optimum Design
Problem Formulation, Graphical Solution
Problem Formulation)
Method and Basic Optimization Concepts,
• Optimization concepts using the
Optimum Design Concepts: Optimality Con-
graphical method (see chapter: Graphical
ditions, and More on Optimum Design Con-
Solution Method and Basic Optimization
cepts: Optimality Conditions. Part II treats
Concepts)
numerical methods for continuous variable
• Optimality conditions for unconstrained
smooth optimization problems and their
and constrained problems (see chapter:
practical applications in chapters: Opti-
Optimum Design Concepts: Optimality
mum Design: Numerical Solution Process
Conditions)
and Excel Solver, Optimum Design with
• Use of Excel and MATLAB® illustrating
MATLAB®, Linear Programming Methods
optimum design of practical problems
for Optimum Design, More on Linear Pro-
(see chapters: Optimum Design:
gramming Methods for Optimum Design,
Numerical Solution Process and Excel
Numerical Methods for Unconstrained
Solver and Optimum Design with
Optimum Design, More on Numerical Meth-
MATLAB®)
ods for Unconstrained Optimum Design,
• Linear programming (see chapter: Linear
Numerical Methods for Constrained Opti-
Programming Methods for Optimum
mum Design, More on Numerical Methods
Design)
for Constrained Optimum Design, Practical
• Numerical methods for unconstrained
Applications of Optimization. Part III con-
and constrained problems (see chapters:
tains advanced topics of practical interest on
Numerical Methods for Unconstrained
optimum design, including nature-inspired
Optimum Design and Numerical
metaheuristic methods that do not require
Methods for Constrained Optimum
derivatives of the problem functions in
Design)
chapters: Discrete Variable Optimum Design
Concepts and Methods, Global Optimization The use of Excel and MATLAB should be
Concepts and Methods, Nature-Inspired introduced mid-semester so that students
Search Methods, Multi-objective Optimum have a chance to formulate and solve more
Design Concepts and Methods, and Addi- challenging projects by the semester’s end.
tional Topics on Optimum Design. Note that project exercises and sections with
Introduction to Optimum Design, Fourth advanced material are marked with an aster-
Edition can be used to design several types of isk (*) next to section headings, which means
courses depending on the learning objectives that they may be omitted for this course.


 Preface to Fourth Edition
First Graduate-Level Course
Topics for a first graduate-level course
include
• Theory and numerical methods for
unconstrained optimization (see
chapters: Introduction to Design
Optimization, Optimum Design Problem
Formulation, Graphical Solution Method
and Basic Optimization Concepts,
Optimum Design Concepts: Optimality
Conditions, Numerical Methods for
Unconstrained Optimum Design,
and More on Numerical Methods for
Unconstrained Optimum Design)
• Theory and numerical methods for
constrained optimization (see chapters:
Optimum Design Concepts: Optimality
Conditions, More on Optimum Design
Concepts: Optimality Conditions,
Numerical Methods for Constrained
Optimum Design, and More on
Numerical Methods for Constrained
Optimum Design)
• Linear and quadratic programming (see
chapters: Linear Programming Methods
for Optimum Design and More on Linear
Programming Methods for Optimum
Design)
The pace of material can be a bit faster
for this course. Students should code some
of the algorithms into computer programs
and solve some projects based on practical
­applications.
Second Graduate-Level Course
This course presents advanced topics on
optimum design:
• Duality theory in nonlinear
programming, rate of convergence

xvii
of iterative algorithms, derivation of
numerical methods, and direct search
methods (see chapters: Introduction
to Design Optimization, Optimum
Design Problem Formulation, Graphical
Solution Method and Basic Optimization
Concepts, Optimum Design Concepts:
Optimality Conditions, More on
Optimum Design Concepts: Optimality
Conditions, Optimum Design: Numerical
Solution Process and Excel Solver,
Optimum Design with MATLAB®, Linear
Programming Methods for Optimum
Design, More on Linear Programming
Methods for Optimum Design,
Numerical Methods for Unconstrained
Optimum Design, More on Numerical
Methods for Unconstrained Optimum
Design, and Numerical Methods for
Constrained Optimum Design, More
on Numerical Methods for Constrained
Optimum Design, and Practical
Applications of Optimization)
• Methods for discrete variable problems
(see chapter: Discrete Variable Optimum
Design Concepts and Methods)
• Global optimization (see chapter: Global
Optimization Concepts and Methods)
• Nature-inspired search methods
(see chapter: Nature-Inspired Search
Methods)
• Multi-objective optimization (see chapter:
Multi-objective Optimum Design
Concepts and Methods)
• Response surface methods, robust
design, and reliability-based design
optimization (see chapter: Additional
Topics on Optimum Design)
In this course, students write computer
programs to implement some of the numeri-
cal methods to gain experience with their
coding and to solve practical problems.
 Acknowledgments
I would like to thank my colleague Pro-
fessor Karim Abdel-Malek, Director of the
Center for Computer-Aided Design and Pro-
fessor of Biomedical Engineering at The Uni-
versity of Iowa, for his enthusiastic support
for this book and for the exciting research on
optimization-based modeling, simulation
and biomechanics of human motion.
I would like to thank the following col-
leagues for their contribution to some
chapters: Professor Tae Hee Lee – chapter:
Optimum Design with MATLAB®; Dr Tim
Marler – chapter: Multi-objective Optimum
Design Concepts and Methods; Professor
G. J. Park – chapter: Additional Topics on
Optimum Design; and Dr Marcelo A. da
Silva and Dr. Qian Wang – chapter: Practical
­Applications of Optimization.
Dr Yujiang Xiang, Dr Rajan Bhatt,
Dr Hyun-Joon Chung, John Nicholson and
Robert Lucente provided valuable input for
some chapters.
I appreciate the help given by Jun Choi,
John Nicholson, Palani Permeswaran, Karlin
Stutzman, Dr Hyun-Jung Kwon, Dr Hyun-
Joon Chung, and Dr Mohammad Bataineh
with parts of the instructor’s manual.
I would like to thank instructors at vari-
ous universities who have used the book for
xix
classroom instruction and provided me with
their input.
I acknowledge the input of my current
and former colleagues at The University of
Iowa who have used the book to teach an
undergraduate course on optimum design:
Professors Karim Abdel-Malek, ­ Asghar
Bhatti, Kyung Choi, Vijay Goel, Ray Han,
Harry Kane, George Lance, and Emad
Tanbour.
Thesis works of all of my current and for-
mer graduate students on various topics of
optimization have been invaluable in broad-
ening of my horizons on the subject.
I would like to thank reviewers of various
parts of the book for their suggestions.
I would also like to thank Steve Merken,
Peter Jardim, Kiruthika Govindaraju and
their team at Elsevier for their superb han-
dling of the manuscript and production of
the book.
I value the support of the Department of
Civil and Environmental Engineering, Cen-
ter for Computer-Aided Design, College of
Engineering, and The University of Iowa for
this project.
Finally, I highly appreciate the uncondi-
tional love and support of my family and
friends.
 Key Symbols and Abbreviations

(a · b) Dot product of vectors a and b; aTb ILP Integer linear programming

c(x) Gradient of cost function, ∇f(x) KKT Karush–Kuhn–Tucker
f(x) Cost function to be minimized LP Linear programming
gj(x) jth inequality constraint MV-OPT Mixed variable optimization problem
hi(x) ith equality constraint NLP Nonlinear programming
m Number of inequality constraints PSO Particle swarm optimization
n Number of design variables QP Quadratic programming
p Number of equality constraints RBDO Reliability-based design optimization
x Design variable vector of dimension n SA Simulated annealing
xi ith component of design variable vector x SLP Sequential linear programming
x(k) kth design variable vector SQP Sequential quadratic programming
ACO Ant colony optimization TS Traveling salesman (salesperson)
BBM Branch-and-bound method
CDF Cumulative distribution function
CSD Constrained steepest descent Note: A superscript “*” indicates (1) optimum value for a
DE Differential evolution; Domain elimination variable, (2) advanced material section, and (3) a project-­
GA Genetic algorithm type exercise.

xxi
 C H A P T E R

1
Introduction to Design Optimization

Upon completion of this chapter, you will be able to:

• Describe the overall process of designing • Distinguish between optimum design and
systems optimal control problems
• Distinguish between engineering design and • Understand the notations used for
engineering analysis activities operations with vectors, matrices, and
• Distinguish between the conventional design functions and their derivatives
process and optimum design process

Engineering consists of a number of well-established activities, including analysis, design,

fabrication, sales, research, and development of systems. The subject of this text—the design
of systems—is a major field in the engineering profession. The process of designing and fab-
ricating systems has been developed over centuries. The existence of many complex systems,
such as buildings, bridges, highways, automobiles, airplanes, space vehicles, and others, is
an excellent testimonial to its long history. However, the evolution of such systems has been
slow and the entire process is both time-consuming and costly, requiring substantial human
and material resources. Therefore, the procedure is to design, fabricate, and use a system
regardless of whether it is the best one. Improved systems have been designed only after a
substantial investment has been recovered.
The preceding discussion indicates that several systems can usually accomplish the same
task, and that some systems are better than others. For example, the purpose of a bridge
is to provide continuity in traffic from one side of the river to the other. Several types of
bridges can serve this purpose. However, to analyze and design all possibilities can be time-
consuming and costly. Usually one type is selected based on some preliminary analyses and
is designed in detail.
The design of a system can be formulated as a problem of optimization in which a performance
measure is optimized while all other requirements are satisfied. Many numerical methods of
optimization have been developed and used to design better systems. This text describes the
basic concepts of optimization and numerical methods for the design of engineering systems.
Design process, rather than optimization theory, is emphasized. Various theorems are stated

Introduction to Optimum Design. http://dx.doi.org/10.1016/B978-0-12-800806-5.00001-9

Copyright © 2017 Elsevier Inc. All rights reserved.

3
4 1. Introduction to Design Optimization

as results without rigorous proofs. However, their implications from an engineering point of
view are discussed.
Any problem in which certain parameters need to be determined to satisfy constraints
can be formulated as an optimization problem. Once this has been done, optimization con-
cepts and methods described in this text can be used to solve it. Optimization methods are
quite general, having a wide range of applicability in diverse fields. It is not possible to
discuss every application of optimization concepts and methods in this introductory text.
However, using simple applications, we discuss concepts, fundamental principles, and basic
techniques that are used in most applications. The student should understand them with-
out getting bogged down with notations, terminologies, and details on particular areas of
­application.

1.1 THE DESIGN PROCESS

How Do I Begin to Design a System?

Designing engineering systems can be a complex process. Assumptions must be made to
develop realistic models that can be subjected to mathematical analysis by the available meth-
ods. The models may need to be verified by experiments. Many possibilities and factors must
be considered during the optimization problem formulation phase. Economic considerations
play an important role in designing cost-effective systems. To complete the design of an en-
gineering system, designers from different fields of engineering must usually cooperate. For
example, the design of a high-rise building involves designers from architectural, structural,
mechanical, electrical, and environmental engineering, as well as construction management
experts. Design of a passenger car requires cooperation among structural, mechanical, auto-
motive, electrical, chemical, hydraulics design, and human factor engineers. Thus, in an inter-
disciplinary environment, considerable interaction is needed among design teams to complete
the project. For most applications, the entire design project must be broken down into several
subproblems, which are then treated somewhat independently. Each of the subproblems can
be posed as a problem of optimum design.
The design of a system begins with the analysis of various options. Subsystems and their
components are identified, designed, and tested. This process results in a set of drawings,
calculations, and reports with the help of which the system can be fabricated. We use a sys-
tems engineering model to describe the design process. Although complete discussion of this
subject is beyond the scope of this text, some basic concepts are discussed using a simple
block diagram.
Design is an iterative process. Iterative implies analyzing several trial designs one after an-
other until an acceptable design is obtained. It is important to understand the concept of
a trial design. In the design process, the designer estimates a trial design of the system
based on experience, intuition, or some simple mathematical analyses. The trial design is
then analyzed to determine if it is acceptable. In case it gets accepted, the design process
is terminated. In the optimization process, the trial design is analyzed to determine if it is the
best. Depending on the specifications, “best” can have different connotations for different
systems. In general, it implies that a system is cost-effective, efficient, reliable, and durable.

I. The Basic Concepts

 1.1 The design process 5

FIGURE 1.1 System evolution model.

The basic concepts are described in this text to aid the engineer in designing systems at
minimum cost.
The design process should be well organized. To discuss it, we consider a system evolu-
tion model, shown in Fig. 1.1, where the process begins with the identification of a need that
may be conceived by engineers or nonengineers. The five steps of the model in the figure are
­described in the following paragraphs.
1. The first step in the evolutionary process is to precisely define the specifications for the
system. Considerable interaction between the engineer and the sponsor of the project is
usually necessary to quantify the system specifications.
2. The second step in the process is to develop a preliminary design of the system. Various
system concepts are studied. Since this must be done in a relatively short time, simplified
models are used at this stage. Various subsystems are identified and their preliminary
designs are estimated. Decisions made at this stage generally influence the system’s
final appearance and performance. At the end of the preliminary design phase, a few
promising design concepts that need further analysis are identified.
3. The third step in the process is a detailed design for all subsystems using the iterative
process described earlier. To evaluate various possibilities, this must be done for all
previously identified promising design concepts. The design parameters for the
subsystems must be identified. The system performance requirements must be
identified and formulated. The subsystems must be designed to maximize system
worth or to minimize a measure of the cost. Systematic optimization methods described
in this text aid the designer in accelerating the detailed design process. At the end of
the process, a description of the final design is available in the form of reports and
drawings.
4. The fourth and fifth steps shown in Fig. 1.1 may or may not be necessary for all systems.
They involve fabrication of a prototype system and testing, and are necessary when the
system must be mass-produced or when human lives are involved. These steps may
appear to be the final ones in the design process, but they are not because the system
may not perform according to specifications during the testing phase. Therefore, the
specifications may have to be modified or other concepts may have to be studied. In fact,
this reexamination may be necessary at any point during the design process. It is for this
reason that feedback loops are placed at every stage of the system evolution process, as
shown in Fig. 1.1. This iterative process must be continued until the best system evolves.

I. The Basic Concepts

6 1. Introduction to Design Optimization

Depending on the complexity of the system, this process may take a few days or several
months.
The model described in Fig. 1.1 is a simplified block diagram for system evolution. In
a­ ctual practice, each block may be broken down into several subblocks to carry out the stud-
ies properly and arrive at rational decisions. The important point is that optimization concepts
and methods are helpful at every stage of the process. Such methods, along with the appropriate
software, can be useful in studying various design possibilities rapidly.

1.2 ENGINEERING DESIGN VERSUS ENGINEERING ANALYSIS

Can I Design Without Analysis?

No, You Must Analyze!
It is important to recognize the differences between engineering analysis and design activities.
The analysis problem is concerned with determining the behavior of an existing system or a
trial system being designed for a known task. Determination of the behavior of the system
implies calculation of its response to specified inputs. For this reason, the sizes of various
parts and their configurations are given for the analysis problem; that is, the design of the
system is known. On the other hand, the design process calculates the sizes and shapes of
various parts of the system to meet performance requirements.
The design of a system is an iterative process. We estimate a trial design and analyze it to
see if it performs according to given specifications. If it does, we have an acceptable (feasible)
design, although we may still want to change it to improve its performance. If the trial design
does not work, we need to change it to come up with an acceptable system. In both cases, we
must be able to analyze designs to make further decisions. Thus, analysis capability must be
available in the design process.
This book is intended for use in all branches of engineering. It is assumed throughout that
students understand the analysis methods covered in undergraduate engineering statics and
physics courses. However, we will not let the lack of analysis capability hinder understanding of the
systematic process of optimum design. Equations for analysis of the system are given wherever
feasible.

1.3 CONVENTIONAL VERSUS OPTIMUM DESIGN PROCESS

Why Do I Want to Optimize?

Because You Want to Beat the Competition and Improve Your Bottom Line!
It is a challenge for engineers to design efficient and cost-effective systems without com-
promising their integrity. Fig. 1.2a presents a self-explanatory flowchart for a conventional
design method; Fig. 1.2b presents a similar flowchart for the optimum design method. It is
important to note that both methods are iterative, as indicated by a loop between blocks 6 and
3. Both methods have some blocks that require similar calculations and others that require
different calculations. The key features of the two processes are as follows.

I. The Basic Concepts

 1.3 Conventional versus optimum design process 7

FIGURE 1.2 Comparison of: (a) conventional design method; and (b) optimum design method.

0. The optimum design method has block 0, where the problem is formulated as one of
­optimization (see chapter: Optimum Design Problem Formulation for detailed discus-
sion). An objective function is defined that measures the merits of different designs.
1. Both methods require data to describe the system in block 1.
2. Both methods require an initial design estimate in block 2.
3. Both methods require analysis of the system in block 3.
4. In block 4, the conventional design method checks to ensure that the performance
criteria are met, whereas the optimum design method checks for satisfaction of all of the
constraints for the problem formulated in block 0.
5. In block 5, stopping criteria for the two methods are checked, and the iteration is stopped
if the specified stopping criteria are met.
6. In block 6, the conventional design method updates the design based on the designer’s
experience and intuition and other information gathered from one or more trial designs;
the optimum design method uses optimization concepts and procedures to update the
current design.
The foregoing distinction between the two design approaches indicates that the conven-
tional design process is less formal. An objective function that measures a design’s merit is
not identified. Trend information is usually not calculated; nor is it used in block 6 to make
­design ­decisions for system improvement. In contrast, the optimization process is more for-
mal, ­using trend information to make design changes.

I. The Basic Concepts

8 1. Introduction to Design Optimization

1.4 OPTIMUM DESIGN VERSUS OPTIMAL CONTROL

What Is Optimal Control?

Optimum design and optimal control of systems are separate activities. There are
­ umerous applications in which methods of optimum design are useful in designing sys-
n
tems. There are many other applications where optimal control concepts are needed. In
­addition, there are some applications in which both optimum design and optimal control
concepts must be used. Sample applications of both techniques include robotics and aerospace
structures. In this text, optimal control problems and methods are not described in detail.
However, the fundamental differences between the two activities are briefly explained in
the sequel. It turns out that some optimal control problems can be transformed into opti-
mum design problems and treated by the methods described in this text. Thus, methods
of optimum design are very powerful and should be clearly understood. A simple opti-
mal control problem is described in chapter: Practical Applications of Optimization and is
solved by the methods of optimum design.
The optimal control problem consists of finding feedback controllers for a system to pro-
duce the desired output. The system has active elements that sense output fluctuations. Sys-
tem controls are automatically adjusted to correct the situation and optimize a measure of
performance. Thus, control problems are usually dynamic in nature. In optimum design, on
the other hand, we design the system and its elements to optimize an objective function. The
system then remains fixed for its entire life.
As an example, consider the cruise control mechanism in passenger cars. The idea behind
this feedback system is to control fuel injection to maintain a constant speed. Thus, the sys-
tem’s output (ie, the vehicle’s cruising speed) is known. The job of the control mechanism
is to sense fluctuations in speed depending on road conditions and to adjust fuel injection
accordingly.

1.5 BASIC TERMINOLOGY AND NOTATION

Which Notation Do I Need to Know?

To understand and to be comfortable with the methods of optimum design, a student must
be familiar with linear algebra (vector and matrix operations) and basic calculus. ­Operations of
linear algebra are described in Appendix A. Students who are not comfortable with this ­material
need to review it thoroughly. Calculus of functions of single and multiple variables must also
be understood. Calculus concepts are reviewed wherever they are needed. In this section,
the standard terminology and notations used throughout the text are defined. It is ­important to
­understand and memorize these notations and operations.

1.5.1 Vectors and Points

Since realistic systems generally involve several variables, it is necessary to define and use
some convenient and compact notations to represent them. Set and vector notations serve this
purpose quite well.

I. The Basic Concepts

 1.5 Basic terminology and notation 9

FIGURE 1.3 Vector representation of a point P in 3D space.

A point is an ordered list of numbers. Thus, (x1, x2) is a point consisting of two numbers where-
as (x1, x2, …, xn) is a point consisting of n numbers. Such a point is often called an n-tuple. The
n components x1, x2, …, xn are collected into a column vector as
 x1 
 x2  T
x=  =  x1 x2 . . . xn  (1.1)
  
 xn 

where the superscript T denotes the transpose of a vector or a matrix. This is called an n-vector.
Each number xi is called a component of the (point) vector. Thus, x1 is the first component, x2
is the second, and so on.
We also use the following notation to represent a point or a vector in the n-dimensional
space:
x = ( x1 , x2 , . . . , xn ) (1.2)

In 3-dimensional (3D) space, the vector x = [x1 x2 x3]T represents a point P, as shown in
Fig. 1.3. ­Similarly, when there are n components in a vector, as in Eqs. (1.1) and (1.2), x is
interpreted as a point in the n-dimensional space, denoted as Rn. The space Rn is simply the
collection of all n
­ -dimensional vectors (points) of real numbers. For example, the real line is
R1, the plane is R2, and so on.

The terms vector and point are used interchangeably, and lowercase letters in roman
boldface are used to denote them. Uppercase letters in roman boldface represent matrices.

1.5.2 Sets
Often we deal with sets of points satisfying certain conditions. For example, we may con-
sider a set S of all points having three components, with the last having a fixed value of 3,
which is written as

I. The Basic Concepts

10 1. Introduction to Design Optimization

{
S = x = ( x1 , x2 , x3 ) x3 = 3 } (1.3)

Information about the set is contained in braces ({ }). Eq. (1.3) reads as “S equals the set of
all points (x1, x2, x3) with x3 = 3.” The vertical bar divides information about the set S into two
parts: To the left of the bar is the dimension of points in the set; to the right are the properties
that distinguish those points from others not in the set (eg, properties a point must possess
to be in the set S).
Members of a set are sometimes called elements. If a point x is an element of the set S, then
we write x ∈ S. The expression x ∈ S is read as “x is an element of (belongs to) S.” Conversely,
the expression “y ∉ S” is read as “y is not an element of (does not belong to) S.”
If all the elements of a set S are also elements of another set T, then S is said to be a subset of
T. Symbolically, we write S ⊂ T, which is read as “S is a subset of T” or “S is contained in T.”
Alternatively, we say “T is a superset of S,” which is written as T ⊃ S.
As an example of a set S, consider a domain of the xl – x2 plane enclosed by a circle of radius
3 with the center at the point (4, 4), as shown in Fig. 1.4. Mathematically, all points within and
on the circle can be expressed as

{
S = x ∈R2 }
( x1 − 4)2 + ( x2 − 4)2 ≤ 9 (1.4)

Thus, the center of the circle (4, 4) is in the set S because it satisfies the inequality in Eq. (1.4).
We write this as (4, 4) ∈ S. The origin of coordinates (0, 0) does not belong to the set ­because it
does not satisfy the inequality in Eq. (1.4). We write this as (0, 0) ∉ S. It can be verified that the
following points belong to the set: (3, 3), (2, 2), (3, 2), (6, 6). In fact, set S has an infinite number
of points. Many other points are not in the set. It can be verified that the following points are
not in the set: (1, 1), (8, 8), and (−1, 2).

{ }
FIGURE 1.4 Geometrical representation of the set S = x ( x1 − 4 ) + ( x 2 − 4 ) ≤ 9 .
2 2

I. The Basic Concepts

 1.5 Basic terminology and notation 11

1.5.3 Notation for Constraints

Constraints arise naturally in optimum design problems. For example, the material of the
system must not fail, the demand must be met, resources must not be exceeded, and so on. We
shall discuss the constraints in more detail in chapter: Optimum Design Problem Formula-
tion. Here we discuss the terminology and notations for the constraints.
We encountered a constraint in Fig. 1.4 that shows a set S of points within and on the circle
of radius 3. The set S is defined by the following constraint:

( x1 − 4)2 + ( x2 − 4)2 ≤ 9 (1.5)

A constraint of this form is a “less than or equal to type” constraint and is abbreviated as
“≤  type.” Similarly, there are greater than or equal to type constraints, abbreviated as “≥ type.”
Both are called inequality constraints.

1.5.4 Superscripts/Subscripts and Summation Notation

Later we will discuss a set of vectors, components of vectors, and multiplication of matri-
ces and vectors. To write such quantities in a convenient form, consistent and compact nota-
tions must be used. We define these notations here. Superscripts are used to represent different
vectors and matrices. For example, x(i) represents the ith vector of a set and A(k) represents the
kth matrix. Subscripts are used to represent components of vectors and matrices. For example, xj is
the jth component of x and aij is the i–jth element of matrix A. Double subscripts are used to
denote elements of a matrix.
To indicate the range of a subscript or superscript we use the notation
xi ; i = 1 to n (1.6)

This represents the numbers x1, x2, …, xn. Note that “i = 1 to n” represents the range for the
index i and is read, “i goes from 1 to n.” Similarly, a set of k vectors, each having n compo-
nents, is represented by the superscript notation as

x ( j ) ; j = 1 to k (1.7)

This represents the k vectors x(l), x(2), …, x(k). It is important to note that subscript i in Eq. (1.6)
and superscript j in Eq. (1.7) are free indices; that is, they can be replaced by any other variable.
For example, Eq. (1.6) can also be written as xj, j = 1 to n and Eq. (1.7) can be written as x(i), i = 1
to k. Note that the superscript j in Eq. (1.7) does not represent the power of x. It is an index that
represents the jth vector of a set of vectors.
We also use the summation notation quite frequently. For example,
c = x1 y1 + x2 y2 + . . . + xn yn (1.8)

is written as
n
c = ∑ xi yi (1.9)
i=1

I. The Basic Concepts

12 1. Introduction to Design Optimization

Also, multiplication of an n-dimensional vector x by an m × n matrix A to obtain an m-

dimensional vector y is written as
y = Ax (1.10)

Or, in summation notation, the ith component of y is

n
yi = ∑ aij x j = ai 1 x1 + ai 2 x2 + . . . + ain xn ; i = 1 to m (1.11)
j=1

There is another way of writing the matrix multiplication of Eq. (1.10). Let m-dimensional
vectors a(i); i = 1 to n represent columns of the matrix A. Then y = Ax is also written as
n
y = ∑ a( j ) x j = a(1) x1 + a( 2 ) x2 + . . . + a( n) xn (1.12)
j=1

The sum on the right side of Eq. (1.12) is said to be a linear combination of columns of matrix
A with xj, j = 1 to n as its multipliers. Or y is given as a linear combination of columns of A
(refer Appendix A for further discussion of the linear combination of vectors).
Occasionally, we must use the double summation notation. For example, assuming m = n
and substituting yi from Eq. (1.11) into Eq. (1.9), we obtain the double sum as
n  n  n n
c = ∑ xi  ∑ aij x j  = ∑ ∑ aij xi x j (1.13)
i=1  j=1  i=1 j=1

Note that the indices i and j in Eq. (1.13) can be interchanged. This is possible because c is
a scalar quantity, so its value is not affected by whether we sum first on i or on j. Eq. (1.13) can
also be written in the matrix form, as we will see later.

1.5.5 Norm/Length of a Vector

If we let x and y be two n-dimensional vectors, then their dot product is defined as
n
( x • y ) = xT y = ∑ xi yi (1.14)
i=1

Thus, the dot product is a sum of the product of corresponding elements of the vectors x
and y. Two vectors are said to be orthogonal (normal) if their dot product is 0; that is, x and y
are orthogonal if (x • y) = 0. If the vectors are not orthogonal, the angle between them can be
calculated from the definition of the dot product:

(x • y) = x y cos θ , (1.15)

where u is the angle between vectors x and y, and ‖x‖ represents the length of vector x (also
called the norm of the vector). The length of vector x is defined as the square root of the sum of
squares of the components:

I. The Basic Concepts

 1.5 Basic terminology and notation 13
n
x = ∑x 2
i = (x • x ) (1.16)
i=1

The double sum of Eq. (1.13) can be written in the matrix form as follows:
n n n  n 
c = ∑ ∑ aij xi x j = ∑ xi  ∑ aij x j  = x T Ax (1.17)
i=1 j=1 i=1  j=1 

Since Ax represents a vector, the triple product of Eq. (1.17) is also written as a dot product:

c = x T Ax = ( x • Ax ) (1.18)

1.5.6 Functions of Several Variables

Just as a function of a single variable is represented as f(x), a function of n independent
variables x1, x2, …, xn is written as

f ( x ) = f ( x1 , x2 , . . . , xn ) (1.19)

We deal with many functions of vector variables. To distinguish between functions, sub-
scripts are used. Thus, the ith function is written as

gi ( x ) = gi ( x1 , x2 , . . . , xn ) (1.20)

If there are m functions gi(x), i = 1 to m, these are represented in the vector form

 g1 ( x ) 
 
 g (x ) 
g (x) =  2  =  g1 ( x ) g2 ( x ) . . . gm ( x ) 
T
(1.21)
:
 
 gm ( x ) 

Throughout the text it is assumed that all functions are continuous and at least twice continu-
ously differentiable. A function f(x) of n variables is continuous at a point x* if, for any ε > 0,
there is a d > 0 such that

f ( x ) − f ( x *) < ε (1.22)

whenever ‖x − x*‖ < d. Thus, for all points x in a small neighborhood of point x*, a change in
the function value from x* to x is small when the function is continuous. A continuous func-
tion need not be differentiable. Twice-continuous differentiability of a function implies not only
that it is differentiable two times, but also that its second derivative is continuous.
Fig. 1.5a,b shows continuous and discontinuous functions. The function in Fig. 1.5a is
differentiable everywhere, whereas the function in Fig. 1.5b is not differentiable at points
x1, x2, and x3. Fig. 1.5c is an example in which f is not a function because it has infinite values
at x1. Fig. 1.5d is an example of a discontinuous function. As examples, functions f(x) = x3

I. The Basic Concepts

14 1. Introduction to Design Optimization

FIGURE 1.5 Continuous and discontinuous functions. (a) and (b) Continuous functions; (c) not a function; and
(d) discontinuous function.

and f(x) = sinx are continuous everywhere and are also continuously differentiable. However,
function f(x) = |x| is continuous everywhere but not differentiable at x = 0.

1.5.7 Partial Derivatives of Functions

Often in this text we must calculate derivatives of functions of several variables. Here we
introduce some of the basic notations used to represent the partial derivatives of functions of
several variables.

First Partial Derivatives

For a function f(x) of n variables, the first partial derivatives are written as

∂ f (x )
; i = 1 to n (1.23)
∂ xi

The n partial derivatives in Eq. (1.23) are usually arranged in a column vector known as the
gradient of the function f(x). The gradient is written as ∂f/∂x or ∇f(x). Therefore,

 [∂ f (x )]/∂ x1 
 
∂ f (x )  [∂ f (x )]/∂ x2  (1.24)
∇f (x ) = =
∂x   
 [∂ f (x )]/∂ x 
 n


I. The Basic Concepts

 1.5 Basic terminology and notation 15
Note that each component of the gradient in Eqs. (1.23) or (1.24) is a function of vector x.

Second Partial Derivatives

Each component of the gradient vector in Eq. (1.24) can be differentiated again with respect
to a variable to obtain the second partial derivatives for the function f(x):

∂2 f ( x )
; i , j = 1 to n (1.25)
∂ xi ∂ x j

We see that there are n2 partial derivatives in Eq. (1.25). These can be arranged in a matrix
known as the Hessian matrix, written as H(x), or simply the matrix of second partial deriva-
tives of f(x), written as ∇2f(x):

 ∂ 2 f (x ) 
H(x ) = ∇ 2 f (x ) =   (1.26)
 ∂ xi ∂ x j  n× n

Note that if f(x) is continuously differentiable two times, then Hessian matrix H(x) in
Eq. (1.26) is symmetric.

Partial Derivatives of Vector Functions

On several occasions we must differentiate a vector function of n variables, such as the
vector g(x) in Eq. (1.21), with respect to the n variables in vector x. Differentiation of each com-
ponent of the vector g(x) results in a gradient vector, such as ∇gi(x). Each of these gradients
is an n-dimensional vector. They can be arranged as columns of a matrix of dimension n × m,
referred to as the gradient matrix of g(x). This is written as

∂g( x)
∇g ( x ) = = ∇g1 ( x ) ∇g2 ( x ) … ∇gm ( x )  n× m (1.27)
∂x

This gradient matrix is usually written as matrix A:

∂gj
A =  aij  n× m ; aij = ; i = 1 to n; j = 1 to m (1.28)
∂ xi

1.5.8 US–British Versus SI Units

The formulation of the design problem and the methods of optimization do not depend
on the units of measure used. Thus, it does not matter which units are used to formulate
the problem. However, the final form of some of the analytical expressions for the problem
does depend on the units used. In the text, we use both US–British and SI units in examples
and exercises. Readers unfamiliar with either system should not feel at a disadvantage when
reading and understanding the material since it is simple to switch from one system to the
other. To facilitate the conversion from US–British to SI units or vice versa, Table 1.1 gives

I. The Basic Concepts

16 1. Introduction to Design Optimization

TABLE 1.1 Conversion Factors for US–British and SI Units

To convert from US–British To SI units Multiply by
Acceleration
Foot/second2 (ft./s2) Meter/second2 (m/s2) 0.3048*
2 2
Inch/second (in./s ) Meter/second2 (m/s2) 0.0254*
Area
Foot2 (ft.2) Meter2 (m2) 0.09290304*
2 2
Inch (in. ) Meter2 (m2) 6.4516E–04*
Bending moment or torque
Pound force inch (lbf·in.) Newton meter (N·m) 0.1129848
Pound force foot (lbf·ft.) Newton meter (N·m) 1.355818
Density
Pound mass/inch3 (lbm/in.3) Kilogram/meter3 (kg/m3) 27,679.90
Pound mass/foot3 (lbm/ft.3) Kilogram/meter3 (kg/m3) 16.01846
Energy or work
British thermal unit (BTU) Joule (J) 1055.056
Foot pound force (ft.·lbf) Joule (J) 1.355818
Kilowatt-hour (KWh) Joule (J) 3,600,000*
Force
Kip (1000 lbf) Newton (N) 4448.222
Pound force (lbf) Newton (N) 4.448222
Length
Foot (ft.) Meter (m) 0.3048*
Inch (in.) Meter (m) 0.0254*
Inch (in.) Micron (m); micrometer (mm) 25,400*
Mile (mi), US statute Meter (m) 1609.344
Mile (mi), International, nautical Meter (m) 1852*
Mass
Pound mass (lbm) Kilogram (kg) 0.4535924
Ounce Grams 28.3495
Slug (lbf·s2ft.) Kilogram (kg) 14.5939
Ton (short, 2000 lbm) Kilogram (kg) 907.1847
Ton (long, 2240 lbm) Kilogram (kg) 1016.047
Tonne (t, metric ton) Kilogram (kg) 1000*
Power
Foot pound/minute (ft.·lbf/min) Watt (W) 0.02259697
Horsepower (550 ft. lbf/s) Watt (W) 745.6999

I. The Basic Concepts

 1.5 Basic terminology and notation 17
TABLE 1.1 Conversion Factors for US–British and SI Units (cont.)
To convert from US–British To SI units Multiply by
Pressure or stress
Atmosphere (std) (14.7 lbf/in.2) Newton/meter2 (N/m2 or Pa) 101,325*
One bar (b) Newton/meter2 (N/m2 or Pa) 100,000*
Pound/foot2 (lbf/ft.2) Newton/meter2 (N/m2 or Pa) 47.88026
2 2 2 2
Pound/inch (lbf/in. or psi) Newton/meter (N/m or Pa) 6894.757
Velocity
Foot/minute (ft./min) Meter/second (m/s) 0.00508*
Foot/second (ft./s) Meter/second (m/s) 0.3048*
Knot (nautical mi/h), international Meter/second (m/s) 0.5144444
Mile/hour (mi/h), international Meter/second (m/s) 0.44704*
Mile/hour (mi/h), international Kilometer/hour (km/h) 1.609344*
Mile/second (mi/s), international Kilometer/second (km/s) 1.609344*
Volume
Foot3 (ft.3) Meter3 (m3) 0.02831685
3 3 3 3
Inch (in. ) Meter (m ) 1.638706E–05
Gallon (Canadian liquid) Meter3 (m3) 0.004546090
3 3
Gallon (UK liquid) Meter (m ) 0.004546092
Gallon (UK liquid) Liter (L) 4.546092
Gallon (US dry) Meter3 (m3) 0.004404884
Gallon (US liquid) Meter3 (m3) 0.003785412
Gallon (US liquid) Liter (L) 3.785412
3 3
One liter (L) Meter (m ) 0.001*
One liter (L) Centimeter3 (cm3) 1000*
One milliliter (mL) Centimeter3 (cm3) 1*
3 3
Ounce (UK fluid) Meter (m ) 2.841307E–05
Ounce (US fluid) Meter3 (m3) 2.957353E–05
Ounce (US fluid) Liter (L) 2.957353E–02
Ounce (US fluid) Milliliter (mL) 29.57353
Pint (US dry) Meter3 (m3) 5.506105E–04
Pint (US liquid) Liter (L) 4.731765E–01
Pint (US liquid) Meter3 (m3) 4.731765E–04
Quart (US dry) Meter3 (m3) 0.001101221
Quart (US liquid) Meter3 (m3) 9.463529E–04
* Exact conversion factor.

I. The Basic Concepts

18 1. Introduction to Design Optimization

conversion factors for the most commonly used quantities. For a complete list of conversion
factors, consult the IEEE/ASTM (2010) publication.

Reference
IEEE/ASTM, 2010. American National Standard for Metric Practice. SI 10-2010. The Institute of Electrical and
­Electronics Engineers/American Society for Testing of Materials, New York.

I. The Basic Concepts

 C H A P T E R

2
Optimum Design Problem
Formulation

Upon completion of this chapter, you will be able to:

• Translate a descriptive statement of the • Identify and define the design problem’s
design problem into a mathematical constraints
statement for optimization • Transcribe the problem formulation into a
• Identify and define the problem’s design standard model for design optimization
variables
• Identify and define an optimization criterion
for the problem

It is generally accepted that the proper definition and formulation of a problem take more than
50% of the total effort needed to solve it. Therefore, it is critical to follow well-defined proce-
dures for formulating design optimization problems. In this chapter, we describe the process
of transforming the design of a selected system and/or subsystem into an optimum design
problem. Methods for solving the problem will be discussed in subsequent chapters; here we
focus on properly formulating the problem as an optimization problem.
Several simple and moderately complex applications are discussed in this chapter
to illustrate the problem formulation process. More advanced applications are dis-
cussed in Chapters 6 and 7 and 14–19.
The importance of properly formulating a design optimization problem must be stressed be-
cause the optimum solution will be only as good as the formulation. For example, if we for-
get to include a critical constraint in the formulation, the optimum solution will most likely
violate it. Also, if we have too many constraints, or if they are inconsistent, there may be no
solution for the problem. However, once the problem is properly formulated, good software
is usually available to solve it.
It is important to note that the process of developing a proper formulation for optimum de-
sign of practical problems is iterative in itself. Several iterations usually are needed to ­revise
Introduction to Optimum Design. http://dx.doi.org/10.1016/B978-0-12-800806-5.00002-0
Copyright © 2017 Elsevier Inc. All rights reserved.

19
20 2. Optimum Design Problem Formulation

the formulation before an acceptable one is finalized. This iterative process is further dis-
cussed in chapter: Optimum Design: Numerical Solution Process and Excel Solver.
For most design optimization problems, we will use the following five-step procedure to
formulate the problem:
Step 1: Project/problem description
Step 2: Data and information collection
Step 3: Definition of design variables
Step 4: Optimization criterion
Step 5: Formulation of constraints
Formulation of an optimum design problem implies translating a descriptive state-
ment of the problem into a well-defined mathematical statement.

2.1 THE PROBLEM FORMULATION PROCESS

We will describe the tasks to be performed in each of the foregoing five steps to develop
a mathematical formulation for the design optimization problem. These steps are illustrated
with some examples in this section and in later sections.

2.1.1 Step 1: Project/Problem Description

Are the Project Goals Clear?
The formulation process begins by developing a descriptive statement for the project/
problem, usually by the project’s owner/sponsor. The statement describes the overall
objectives of the project and the requirements to be met. This is also called the statement of
work.

EXA M PLE 2.1 DESIGN OF A CANTILEVER BEAM, PROBLEM

DESCRIPTION
Cantilever beams are used in many practical applications in civil, mechanical, and aerospace en-
gineering. To illustrate the step of problem description, we consider the design of a hollow square-
cross-section cantilever beam to support a load of 20 kN at its end. The beam, made of steel, is 2 m
long, as shown in Fig. 2.1. The failure conditions for the beam are as follows: (1) the material should
not fail under the action of the load, and (2) the deflection of the free end should be no more than
1 cm. The width-to-thickness ratio for the beam should be no more than 8 to avoid local buckling
of the walls. A minimum-mass beam is desired. The width and thickness of the beam must be within
the following limits:

60 ≤ width ≤ 300 mm (a)

3 ≤ thickness ≤ 15 mm (b)

I. The Basic Concepts

 2.1 The problem formulation process 21

FIGURE 2.1 Cantilever beam of a hollow square cross-section.

2.1.2 Step 2: Data and Information Collection

Is all the Information Available to Solve the Problem?
To develop a mathematical formulation for the problem, we need to gather information
on material properties, performance requirements, resource limits, cost of raw materials, and
so forth. In addition, most problems require the capability to analyze trial designs. Therefore,
analysis procedures and analysis tools must be identified at this stage. For example, the finite-
element method is commonly used for analysis of structures, so the software tool available
for such an analysis needs to be identified. In many cases, the project statement is vague,
and assumptions about modeling of the problem need to be made in order to formulate and
solve it.

EXA M PLE 2.2 DATA AND INFORMATION COLLECTION

FOR CANTILEVER BEAM
The information needed for the cantilever beam design problem of Example 2.1 includes expressions
for bending and shear stresses, and the expression for the deflection of the free end. The notation
and data for this purpose are defined in Table 2.1.
The following are useful expressions for the beam:

A = w 2 − (w − 2t)2 = 4t(w − t), mm 2 (c)

1 1 1 4 1
I= w × w 3 − ( w − 2 t ) × ( w − 2 t )3 = w − (w − 2t)4, mm 4 (d)
12 12 12 12

1 w 1 (w − 2t) 1 3 1
Q = w 2 × − ( w − 2 t )2 × = w − (w − 2t)3, mm 3 (e)
2 4 2 4 8 8

M = PL, N/mm (f)

V = P, N (g)

I. The Basic Concepts

22 2. Optimum Design Problem Formulation

TABLE 2.1 Notation and Data for Cantilever Beam

Notation Data

A Cross-sectional area, mm2

E Modulus of elasticity of steel, 21 × 104 N/mm2
G Shear modulus of steel, 8 × 104 N/mm2
I Moment of inertia of the cross-section, mm4
L Length of the member, 2000 mm
M Bending moment, N/mm
P Load at the free end, 20,000 N
Q Moment about the neutral axis of the area above the neutral axis, mm3
q Vertical deflection of the free end, mm
qa Allowable vertical deflection of the free end, 10 mm
V Shear force, N
w Width (depth) of the section, mm
t Wall thickness, mm
σ Bending stress, N/mm2
σa Allowable bending stress, 165 N/mm2
τ Shear stress, N/mm2
τa Allowable shear stress, 90 N/mm2

Mw
σ= , N/mm 2 (h)
2l

VQ
τ= , N/mm 2 (i)
2 It

PL3
q= , mm (j)
3EI

2.1.3 Step 3: Definition of Design Variables

What are these Variables?
HOW DO I IDENTIFY THEM?
The next step in the formulation process is to identify a set of variables that describe the
system, called the design variables. In general, these are referred to as optimization variables
or simply variables that are regarded as free because we should be able to assign any value
to them. Different values for the variables produce different designs. The design variables
should be independent of each other as far as possible. If they are dependent, their values
cannot be specified independently because there are constraints between them. The number
of independent design variables gives the design degrees of freedom for the problem.

I. The Basic Concepts

 2.1 The problem formulation process 23
For some problems, different sets of variables can be identified to describe the same sys-
tem. Problem formulation will depend on the selected set. We will present some examples
later in this chapter to elaborate on this point.
Once the design variables are given numerical values, we have a design of the system.
Whether this design satisfies all requirements is another question. We will introduce a number
of concepts to investigate such questions in later chapters.
If proper design variables are not selected for a problem, the formulation will be either
incorrect or not possible. At the initial stage of problem formulation, all options for specifica-
tion of design variables should be investigated. Sometimes it may be desirable to designate
more design variables than apparent design degrees of freedom. This gives added flexibility
to problem formulation. Later, it is possible to assign a fixed numerical value to any variable
and thus eliminate it from the formulation.
At times it is difficult to clearly identify a problem’s design variables. In such a case, a
complete list of all variables may be prepared. Then, by considering each variable individu-
ally, we can determine whether or not it can be treated as an optimization variable. If it is a valid
design variable, the designer should be able to specify a numerical value for it to select a trial
design.
We will use the term “design variables” to indicate all optimization variables for the opti-
mization problem and will represent them in the vector x. To summarize, the following con-
siderations should be given in identifying design variables for a problem:
• Generally, the design variables should be independent of each other. If they are
not, there must be some equality constraints between them (explained later in
several examples).
• A minimum number of design variables is required to properly formulate a design
optimization problem.
• As many independent parameters as possible should be designated as design
variables at the problem formulation phase. Later on, some of these variables can
be assigned fixed numerical values.
• A numerical value should be given to each identified design variable to determine
if a trial design of the system is specified.

EXA M PLE 2.3 DESIGN VARIABLES FOR CANTILEVER BEAM

Only dimensions of the cross-section are identified as design variables for the cantilever beam
design problem of Example 2.1; all other parameters are specified:

w = outside width (depth) of the section, mm

t = wall thickness, mm

Note that the design variables are defined precisely including the units to be used for them.
It is also noted here that an alternate set of design variables can be selected: wo = outer width of the
section, and wi = inner width of the section. The problem can be formulated using these design vari-
ables. However, note that all the expressions given in Eqs. (c)–(j) will have to be re-derived in terms
of wo and wi. Thus the two formulations will look quite different from each other for the same design
problem. However, these two formulations should yield same final solution.

I. The Basic Concepts

Another random document with
no related content on Scribd:
 Fig. 313.—Diodon maculatus.

Fig. 314.—Diodon maculatus, inflated.

C. Molina.—Body compressed, very short; tail extremely short,
truncate. Vertical fins confluent. No pelvic bone.
The “Sun-fishes” (Orthagoriscus) are pelagic fishes, found in
every part of the oceans within the tropical and temperate zones.
The singular shape of their body and the remarkable changes which
they undergo with age, have been noticed above (p. 175, Figs. 93,
94). Their jaws are undivided in the middle, comparatively feeble, but
well adapted for masticating their food, which consists of small
pelagic Crustaceans. Two species are known. The common Sun-
fish, O. mola, which attains to a very large size, measuring seven or
eight feet, and weighing as many hundredweights. It has a rough,
minutely granulated skin. It frequently approaches the southern
coasts of England and the coasts of Ireland, and is seen basking in
calm weather on the surface. The second species, O. truncatus, is
distinguished by its smooth, tessellated skin, and one of the scarcest
fishes in collections. The shortness of the vertebral column of the
Sun-fishes, in which the number of caudal vertebra is reduced to
seven, the total number being seventeen, and the still more reduced
length of the spinal chord have been noticed above (p. 96).

THIRD SUB-CLASS—CYCLOSTOMATA.
Skeleton cartilaginous and notochordal, without ribs and without
real jaws. Skull not separate from the vertebral column. No limbs.
Gills in the form of fixed sacs, without branchial arches, six or seven
in number on each side. One nasal aperture only. Heart without
bulbus arteriosus. Mouth anterior, surrounded by a circular or
subcircular lip, suctorial. Alimentary canal straight, simple, without
coecal appendages, pancreas or spleen. Generative outlet
peritoneal. Vertical fins rayed.
The Cyclostomes are most probably a very ancient type.
Unfortunately the organs of these creatures are too soft to be
preserved, with the exception of the horny denticles with which the
mouth of some of them is armed. And, indeed, dental plates, which
are very similar to those of Myxine, are not uncommon in certain
strata of Devonian and Silurian age (see p. 193). The fishes
belonging to this sub-class may be divided into two families—

First Family—Petromyzontidæ.
Body eel-shaped, naked. Subject to a metamorphosis; in the
perfect stage with a suctorial mouth armed with teeth, simple or
multicuspid, horny, sitting on a soft papilla. Maxillary, mandibulary,
lingual, and suctorial teeth may be distinguished. Eyes present (in
mature animals). External nasal aperture in the middle of the upper
side of the head. The nasal duct terminates without perforating the
palate. Seven branchial sacs and apertures on each side behind the
head; the inner branchial ducts terminate in a separate common
tube. Intestine with a spiral valve. Eggs small. The larvæ without
teeth, and with a single continuous vertical fin.
“Lampreys” are found in the rivers and on the coasts of the
temperate regions of the northern and southern hemispheres. Their
habits are but incompletely known, but so much is certain that at
least some of them ascend rivers periodically, for the purpose of
spawning, and that the young pass several years in rivers, whilst
they undergo a metamorphosis (see p. 170). They feed on other
fishes, to which they suck themselves fast, scraping off the flesh with
their teeth. Whilst thus engaged they are carried about by their
victim; Salmon have been captured in the middle course of the Rhine
with the Marine Lamprey attached to them.

Fig. 315.—Mouth of Larva of

Petromyzon branchialis.
 Fig. 316.—Mouth of Petromyzon
fluviatilis. mx, Maxillary tooth; md,
Mandibulary tooth; l, Lingual tooth;
s, Suctorial teeth.
Petromyzon.—Dorsal fins two, the posterior continuous with the
caudal. The maxillary dentition consists of two teeth placed close
together, or of a transverse bicuspid ridge; lingual teeth serrated.

The Lampreys belonging to this genus are found in the northern

hemisphere only; the British species are the Sea-Lamprey (P.
marinus), exceeding a length of three feet, and not uncommon on
the European and North American coasts; the River-Lamprey or
Lampern (P. fluviatilis), ascending in large numbers the rivers of
Europe, North America, and Japan, and scarcely attaining a length
of two feet; the “Pride” or “Sand-Piper” or Small Lampern (P.
branchialis), scarcely twelve inches long, the larva of which has been
long known under the name of Ammocoetes.
Ichthyomyzon from the western coasts of North America is said
to have a tricuspid maxillary tooth.
Mordacia.—Dorsal fins two, the posterior continuous with the
caudal. The maxillary dentition consists of two triangular groups, each
with three conical acute cusps; two pairs of serrated lingual teeth.
 Fig. 317.—Mouth of Mordacia mordax, closed and
opened.
A Lamprey (M. mordax) from the coasts of Chile and Tasmania.
This fish seems to be provided sometimes with a gular sac, like the
following.[47]

Fig. 318.—Mordacia mordax.

Geotria.—Dorsal fins two, the posterior separate from the caudal.
Maxillary lamina with four sharp flat lobes; a pair of long pointed
lingual teeth.
Two species, one from Chile and one from South Australia. They
grow to a length of two feet, and in some specimens the skin of the
throat is much expanded, forming a large pouch. Its physiological
function is not known. The cavity is in the subcutaneous cellular
tissue, and does not communicate with the buccal or branchial
cavities. Probably it is developed with age, and absent in young
individuals. In all the localities in which these Extra-european
Lampreys are found, Ammocoetes forms occur, so that there is little
doubt that they undergo a similar metamorphosis as P. branchialis.
 Second Family—myxinidæ.
Body eel-shaped, naked. The single nasal aperture is above the
mouth, quite at the extremity of the head, which is provided with four
pairs of barbels. Mouth without lips. Nasal duct without cartilaginous
rings, penetrating the palate. One median tooth on the palate, and
two comb-like series of teeth on the tongue (see Fig. 101). Branchial
apertures at a great distance from the head; the inner branchial
ducts lead into the œsophagus. A series of mucous sacs along each
side of the abdomen. Intestine without spiral valve. Eggs large, with
a horny case provided with threads for adhesion.

Fig. 319.—Ovum of Myxine

glutinosa, enlarged.
The fishes of this family are known by the names of “Hag-Fish,”
“Glutinous Hag,” or “Borer;” they are marine fishes with a similar
distribution as the Gadidæ, being most plentiful in the higher
latitudes of the temperate zones of the northern and southern
hemispheres. They are frequently found buried in the abdominal
cavity of other fishes, especially Gadoids, into which they penetrate
to feed on their flesh. They secrete a thick glutinous slime in
incredible quantities, and are therefore considered by fishermen a
great nuisance, seriously damaging the fisheries and interfering with
the fishing in localities where they abound. Myxine descends to a
depth of 345 fathoms, and is generally met with in the Norwegian
Fjords at 70 fathoms, sometimes in great abundance.
 Myxine.—One external branchial aperture only on each side of the
abdomen, leading by six ducts to six branchial sacs.
Three species from the North Atlantic, Japan, and Magelhæn’s
Straits.

Fig. 320.—Myxine australis. A, Lower aspect of head; a, Nasal aperture; b, Mouth;

g, Branchial aperture; v, Vent.
Bdellostoma.—Six or more external branchial apertures on each
side, each leading by a separate duct to a branchial sac.
Two species from the South Pacific.

FOURTH SUB-CLASS—LEPTOCARDII.
Skeleton membrano-cartilaginous and notochordal, ribless. No
brain. Pulsating sinuses in place of a heart. Blood colourless.
Respiratory cavity confluent with the abdominal cavity; branchial
clefts in great number, the water being expelled by an opening in
front of the vent. Jaws none.
This sub-class is represented by a single family (Cirrostomi) and
by a single genus (Branchiostoma);[48] it is the lowest in the scale of
fishes, and lacks so many characteristics, not only of this class, but
of the vertebrata generally, that Hæckel, with good reason,
separates it into a separate class, that of Acrania. The various parts
of its organisation have been duly noticed in the first part of this
work.
The “Lancelet” (Branchiostoma lanceolatum, see Fig. 28, p. 63),
seems to be almost cosmopolitan within the temperate and tropical
zones. Its small size, its transparency, and the rapidity with which it
is able to bury itself in the sand, are the causes why it escapes so
readily observation, even at localities where it is known to be
common. Shallow, sandy parts of the coasts seem to be the places
on which it may be looked for. It has been found on many localities of
the British, and generally European coasts, in North America, the
West Indies, Brazil, Peru, Tasmania, Australia, and Borneo. It rarely
exceeds a length of three inches. A smaller species, in which the
dorsal fringe is distinctly higher and rayed, and in which the caudal
fringe is absent, has been described under the name of
Epigionichthys pulchellus; it was found in Moreton Bay.
 APPENDIX.
DIRECTIONS FOR COLLECTING AND PRESERVING FISHES.
Whenever practicable fishes ought to be preserved in spirits.
To insure success in preserving specimens the best and
strongest spirits should be procured, which, if necessary, can be
reduced to the strength required during the journey with water or
weaker spirit. Travellers frequently have great difficulties in procuring
spirits during their journey, and therefore it is advisable, especially
during sea voyages, that the traveller should take a sufficient
quantity with him. Pure spirits of wine is best. Methylated spirits may
be recommended on account of their cheapness; however,
specimens do not keep equally well in this fluid, and very valuable
objects, or such as are destined for minute anatomical examination,
should always be kept in pure spirits of wine. If the collector has
exhausted his supply of spirits he may use arrack, cognac, or rum,
provided that the fluids contain a sufficient quantity of alcohol.
Generally speaking, spirits which, without being previously heated,
can be ignited by a match or taper, may be used for the purposes of
conservation. The best method to test the strength of the spirits is
the use of a hydrometer. It is immersed in the fluid to be measured,
and the deeper it sinks the stronger is the spirit. On its scale the
number 0 signifies what is called proof spirit, the lowest degree of
strength which can be used for the conservation of fish for any length
of time. Spirits, in which specimens are packed permanently, should
be from 40 to 60 above proof. If the hydrometers are made of glass
they are easily broken, and therefore the traveller had better provide
himself with three or four of them, their cost being very trifling.
Further, the collector will find a small distilling apparatus very useful.
By its means he is able not only to distil weak and deteriorated spirits
or any other fluid containing alcohol, but also, in case of necessity, to
prepare a small quantity of drinkable spirits.
Of collecting vessels we mention first those which the collector
requires for daily use. Most convenient are four-sided boxes made of
zinc, 18 in. high, 12 in. broad, and 5 in. wide. They have a round
opening at the top of 4 in. diameter, which can be closed by a strong
cover of zinc of 5 in. diameter, the cover being screwed into a raised
rim round the opening. In order to render the cover air-tight, an
indiarubber ring is fixed below its margin. Each of these zinc boxes
fits into a wooden case, the lid of which is provided with hinges and
fastenings, and which on each side has a handle of leather or rope,
so that the box can be easily shifted from one place to another.
These boxes are in fact made from the pattern of the ammunition
cases used in the British army, and extremely convenient, because a
pair can be easily carried strapped over the shoulders of a man or
across the back of a mule. The collector requires at least two, still
better four or six, of these boxes. All those specimens which are
received during the day are deposited in them, in order to allow them
to be thoroughly penetrated by the spirit, which must be renewed
from time to time. They remain there for some time under the
supervision of the collector, and are left in these boxes until they are
hardened and fit for final packing. Of course, other more simple
vessels can be used and substituted for the collecting boxes. For
instance, common earthenware vessels, closed by a cork or an
indiarubber covering, provided they have a wide mouth at the top,
which can be closed so that the spirit does not evaporate, and which
permits of the specimens being inspected at any moment without
trouble. Vessels in which the objects are permanently packed for the
home journey are zinc boxes of various sizes, closely fitting into
wooden cases. Too large a size should be avoided, because the
objects themselves may suffer from the superimposed weight, and
the risk of injury to the case increases with its size. It should hold no
more than 18 cubic feet at most, and what, in accordance with the
size of the specimens, has to be added in length should be deducted
in depth or breadth. The most convenient cases, but not sufficient for
all specimens, are boxes 2 feet in length, 1½ foot broad, and 1 foot
deep. The traveller may provide himself with such cases ready
made, packing in them other articles which he wants during his
journey; or he may find it more convenient to take with him only the
zinc plates cut to the several sizes, and join them into boxes when
they are actually required. The requisite wooden cases can be
procured without much difficulty almost everywhere. No collector
should be without the apparatus and materials for soldering, and he
should be well acquainted with their use. Also a pair of scissors to
cut the zinc plates are useful.
Wooden casks are not suitable for the packing of specimens
preserved in spirits, at least not in tropical climates. They should be
used in cases of necessity only, or for packing of the largest
examples, or for objects preserved in salt or brine.
Very small and delicate specimens should never be packed
together with larger ones, but separately, in small bottles.
Mode of preserving.—All fishes, with the exception of very large
ones (broad kinds exceeding 3–4 feet in length; eel-like kinds more
than 6 feet long), should be preserved in spirits. A deep cut should
be made in the abdomen between the pectoral fins, another in front
of the vent, and one or two more, according to the length of the fish,
along the middle line of the abdomen. These cuts are made partly to
remove the fluid and easily decomposing contents of the intestinal
tract, partly to allow the spirit quickly to penetrate into the interior. In
large fleshy fishes several deep incisions should be made with the
scalpel into the thickest parts of the dorsal and caudal muscles, to
give ready entrance to the spirits. The specimens are then placed in
one of the provisional boxes, in order to extract, by means of the
spirit, the water of which fishes contain a large quantity. After a few
days (in hot climates after 24 or 48 hours) the specimens are
transferred into a second box with stronger spirits, and left therein for
several days. A similar third and, in hot climates sometimes a fourth,
transfer is necessary. This depends entirely on the condition of the
specimens. If, after ten or fourteen days of such treatment the
specimens are firm and in good condition, they may be left in the
spirits last used until they are finally packed. But if they should be
soft, very flexible, and discharge a discoloured bloody mucus, they
must be put back in spirits at least 20° over proof. Specimens
showing distinct signs of decomposition should be thrown away, as
they imperil all other specimens in the same vessel. Neither should
any specimen in which decomposition has commenced when found,
be received for the collecting boxes, unless it be of a very rare
species, when the attempt may be made to preserve it separately in
the strongest spirits available. The fresher the specimens to be
preserved are, the better is the chance of keeping them in a perfect
condition. Specimens which have lost their scales, or are otherwise
much injured, should not be kept. Herring-like fishes, and others with
deciduous scales, are better wrapped in thin paper or linen before
being placed in spirits.
The spirits used during this all-important process of preservation
loses, of course, gradually in strength. As long as it keeps 10° under
proof it may still be used for the first stage of preservation, but
weaker spirits should be re-distilled; or, if the collector cannot do this,
it should be at least filtered through powdered charcoal before it is
mixed with stronger spirits. Many collectors are satisfied with
removing the thick sediment collected at the bottom of the vessel,
and use their spirits over and over again without removing from it by
filtration the decomposing matter with which it has been
impregnated, and which entirely neutralises the preserving property
of the spirits. The result is generally the loss of the collection on its
journey home. The collector can easily detect the vitiated character
of his spirits by its bad smell. He must frequently examine his
specimens; and attention to the rules given, with a little practice and
perseverance, after the possible failure of the first trial, will soon
insure to him the safety of his collected treasures. The trouble of
collecting specimens in spirits is infinitely less than that of preserving
skins or dry specimens of any kind.
When a sufficient number of well-preserved examples have been
brought together, they should be sent home by the earliest
opportunity. Each specimen should be wrapped separately in a piece
of linen, or at least soft paper; the specimens are then packed as
close as herrings in the zinc case, so that no free space is left either
at the top or on the sides. When the case is full, the lid is soldered
on, with a round hole about half an inch in diameter near one of the
corners. This hole is left in order to pour the spirit through it into the
case. Care is taken to drive out the air which may remain between
the specimens, and to surround them completely with spirits, until
the case is quite full. Finally, the hole is closed by a small square lid
of tin being soldered over it. In order to see whether the case keeps
in the spirit perfectly, it is turned upside down and left over night.
When all is found to be securely fastened, the zinc case is placed
into the wooden box and ready for transport.
Now and then it happens in tropical climates that collectors are
unable to keep fishes from decomposition even in the strongest
spirits without being able to detect the cause. In such cases a
remedy will be found in mixing a small quantity of arsenic or
sublimate with the spirits; but the collector ought to inform his
correspondent, or the recipient of the collection, of this admixture
having been made.
In former times fishes of every kind, even those of small size,
were preserved dry as flat skins or stuffed. Specimens thus prepared
admit of a very superficial examination only, and therefore this
method of conservation has been abandoned in all larger museums,
and should be employed exceptionally only, for instance on long
voyages overland, during which, owing to the difficulty of transport,
neither spirits nor vessels can be carried. To make up as much as
possible for the imperfection of such specimens, the collector ought
to sketch the fish before it is skinned, and to colour the sketch if the
species is ornamented with colours likely to disappear in the dry
example. Collectors who have the requisite time and skill, ought to
accompany their collections with drawings coloured from the living
fishes; but at the same time it must be remembered that, valuable as
such drawings are if accompanied by the originals from which they
were made, they can never replace the latter, and possess a
subordinate scientific value only.
Very large fishes can be preserved as skins only; and collectors
are strongly recommended to prepare in this manner the largest
examples obtainable, although it will entail some trouble and
expense. So very few large examples are exhibited in museums, the
majority of the species being known from the young stage only, that
the collector will find himself amply recompensed by attending to
these desiderata.
 Scaly fishes are skinned thus: with a strong pair of scissors an
incision is made along the median line of the abdomen from the
foremost part of the throat, passing on one side of the base of the
ventral and anal fins, to the root of the caudal fin, the cut being
continued upwards to the back of the tail close to the base of the
caudal. The skin of one side of the fish is then severed with the
scalpel from the underlying muscles to the median line of the back;
the bones which support the dorsal and caudal are cut through, so
that these fins remain attached to the skin. The removal of the skin
of the opposite side is easy. More difficult is the preparation of the
head and scapulary region; the two halves of the scapular arch
which have been severed from each other by the first incision are
pressed towards the right and left, and the spine is severed behind
the head, so that now only the head and shoulder bones remain
attached to the skin. These parts have to be cleaned from the inside,
all soft parts, the branchial and hyoid apparatus, and all smaller
bones, being cut away with the scissors or scraped off with the
scalpel. In many fishes, which are provided with a characteristic
dental apparatus in the pharynx (Labroids, Cyprinoids), the
pharyngeal bones ought to be preserved, and tied with a thread to
the specimen. The skin being now prepared so far, its entire inner
surface as well as the inner side of the head are rubbed with
arsenical soap; cotton-wool, or some other soft material is inserted
into any cavities or hollows, and finally a thin layer of the same
material is placed between the two flaps of the skin. The specimen is
then dried under a slight weight to keep it from shrinking.
The scales of some fishes, as for instance of many kinds of
herrings, are so delicate and deciduous that the mere handling
causes them to rub off easily. Such fishes may be covered with thin
paper (tissue-paper is the best), which is allowed to dry on them
before skinning. There is no need for removing the paper before the
specimen has reached its destination.
Scaleless Fishes, as Siluroids and Sturgeons, are skinned in the
same manner, but the skin can be rolled up over the head; such
skins can also be preserved in spirits, in which case the traveller
may save to himself the trouble of cleaning the head.
 Some Sharks are known to attain to a length of 30 feet, and
some Rays to a width of 20 feet. The preservation of such gigantic
specimens is much to be recommended, and although the difficulties
of preserving fishes increase with their size, the operation is
facilitated, because the skins of all Sharks and Rays can easily be
preserved in salt and strong brine. Sharks are skinned much in the
same way as ordinary fishes. In Rays an incision is made not only
from the snout to the end of the fleshy part of the tail, but also a
second across the widest part of the body. When the skin is removed
from the fish, it is placed into a cask with strong brine mixed with
alum, the head occupying the upper part of the cask; this is
necessary, because this part is most likely to show signs of
decomposition, and therefore most requires supervision. When the
preserving fluid has become decidedly weaker from the extracted
blood and water, it is thrown away and replaced by fresh brine. After
a week’s or fortnight’s soaking the skin is taken out of the cask to
allow the fluid to drain off; its inner side is covered with a thin layer of
salt, and after being rolled up (the head being inside) it is packed in a
cask, the bottom of which is covered with salt; all the interstices and
the top are likewise filled with salt. The cask must be perfectly water-
tight.
Of all larger examples of which the skin is prepared, the
measurements should be taken before skinning so as to guide the
taxidermist in stuffing and mounting the specimens.
Skeletons of large osseous fishes are as valuable as their skins.
To preserve them it is only necessary to remove the soft parts of the
abdominal cavity and the larger masses of muscle, the bones being
left in their natural continuity. The remaining flesh is allowed to dry
on the bones, and can be removed by proper maceration at home.
The fins ought to be as carefully attended to as in a skin, and of
scaly fishes so much of the external skin ought to be preserved as is
necessary for the determination of the species, as otherwise it is
generally impossible to determine more than the genus.
A few remarks may be added as regards those Faunæ, which
promise most results to the explorer, with some hints as to desirable
information on the life and economic value of fishes.
 It is surprising to find how small the number is of the freshwater
faunæ which may be regarded as well explored; the rivers of Central
Europe, the Lower Nile, the lower and middle course of the Ganges,
and the lower part of the Amazons are almost the only fresh waters
in which collections made without discrimination would not reward
the naturalist. The oceanic areas are much better known; yet almost
everywhere novel forms can be discovered and new observations
made. Most promising and partly quite unknown are the following
districts:—the Arctic Ocean, all coasts south of 38° lat. S., the Cape
of Good Hope, the Persian Gulf, the coasts of Australia (with the
exception of Tasmania, New South Wales, and New Zealand), many
of the little-visited groups of Pacific islands, the coasts of north-
eastern Asia north of 35° lat. N., and the western coasts of North and
South America.
No opportunity should be lost to obtain pelagic forms, especially
the young larva-like stages of development abounding on the
surface of the open ocean. They can be obtained without difficulty by
means of a small narrow meshed net dragged behind the ship. The
sac of the net is about 3 feet deep, and fastened to a strong brass-
ring 2 or 2½ feet in diameter. The net is suspended by three lines
passing into the strong main line. It can only be used when the
vessel moves very slowly, its speed not exceeding three knots an
hour, or when a current passes the ship whilst at anchor. To keep the
net in a vertical position the ring can be weighted at one point of its
circumference; and by using heavier weights two or three drag-nets
can be used simultaneously at different depths. This kind of fishing
should be tried at night as well as day, as many fishes come to the
surface only after sunset. The net must not be left long in the water,
from 5 to 20 minutes only, as delicate objects would be sure to be
destroyed by the force of the water passing through the meshes.
Objects found floating on the surface, as wood, baskets,
seaweed, etc., deserve the attention of the travellers, as they are
generally surrounded by small fishes or other marine animals.
It is of the greatest importance to note the longitude and latitude
at which the objects were collected in the open ocean.
 Fishing in great depths by means of the dredge, can be practised
only from vessels specially fitted out for the purpose; and the
success which attended the “Challenger,” and North American Deep-
sea explorations, has developed Deep-sea fishing into such a
speciality that the requisite information can be gathered better by
consulting the reports of those expeditions than from a general
account, such as could be given in the present work.
Fishes offer an extraordinary variety with regard to their habits,
growth, etc., so that it is impossible to enumerate in detail the points
of interest to which the travellers should pay particular attention.
However, the following hints may be useful.
Above all, detailed accounts are desirable of all fishes forming
important articles of trade, or capable of becoming more generally
useful than they are at present. Therefore, deserving of special
attention are the Sturgeons, Gadoids, Thyrsites and Chilodactylus,
Salmonoids, Clupeoids. Wherever these fishes are found in sufficient
abundance, new sources may be opened to trade.
Exact observations should be made on the fishes the flesh of
which is poisonous either constantly or at certain times and certain
localities; the cause of the poisonous qualities as well as the nature
of the poison should be ascertained. Likewise the poison of fishes
provided with special poison-organs requires to be experimentally
examined, especially with regard to its effects on other fishes and
animals generally.
All observations directed to sex, mode of propagation, and
development, will have special interest: thus those relating to
secondary sexual characters, hermaphroditism, numeric proportion
of the sexes, time of spawning and migration, mode of spawning,
construction of nests, care of progeny, change of form during growth,
etc.
If the collector is unable to preserve the largest individuals of a
species that may come under his observation he should note at least
their measurements. There are but few species of fishes of which the
limit of growth is known.
 The history of Parasitic Fishes is almost unknown, and any
observations with regard to their relation to their host as well as to
their early life will prove to be valuable; nothing is known of the
propagation of fishes even so common as Echeneis and Fierasfer,
much less of the parasitic Freshwater Siluroids.
The temperature of the blood of the larger freshwater and marine
species should be exactly measured.
Many pelagic and deep-sea fishes are provided with peculiar
small round organs of a mother-of-pearl colour, distributed in series
along the side of the body, especially along the abdomen. Some
zoologists consider these organs as accessory eyes, others (and it
appears to us with better reason) as luminous organs. They deserve
an accurate microscopic examination made on fresh specimens; and
their function should be ascertained from observation of the living
fishes, especially also with regard to the question, whether or not the
luminosity (if such be their function) is subject to the will of the fish.

Fig. 321.—Scopelus boops, a pelagic fish, with luminous organs.