PDF Nature Inspired Computation in Engineering 1St Edition Xin She Yang Eds Ebook Full Chapter

Nature Inspired Computation in
Engineering 1st Edition Xin-She Yang

(Eds.)
Visit to download the full and correct content document:
https://textbookfull.com/product/nature-inspired-computation-in-engineering-1st-editio
n-xin-she-yang-eds/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Nature-inspired Computation and Swarm Intelligence:

Algorithms, Theory and Applications 1st Edition Xin-She
Yang (Editor)
https://textbookfull.com/product/nature-inspired-computation-and-
swarm-intelligence-algorithms-theory-and-applications-1st-
edition-xin-she-yang-editor/
Nature-Inspired Algorithms and Applied Optimization 1st

Edition Xin-She Yang (Eds.)
https://textbookfull.com/product/nature-inspired-algorithms-and-
applied-optimization-1st-edition-xin-she-yang-eds/
Smart Trends in Systems Security and Sustainability

Proceedings of WS4 2017 1st Edition Xin-She Yang
https://textbookfull.com/product/smart-trends-in-systems-
security-and-sustainability-proceedings-of-ws4-2017-1st-edition-
xin-she-yang/
Brain and Nature-inspired Learning, Computation and

Recognition 1st Edition Licheng Jiao
https://textbookfull.com/product/brain-and-nature-inspired-
learning-computation-and-recognition-1st-edition-licheng-jiao/
Discrete Problems in Nature Inspired Algorithms First
Edition Shukla
https://textbookfull.com/product/discrete-problems-in-nature-
inspired-algorithms-first-edition-shukla/
Discrete Problems in Nature Inspired Algorithms 1st

Edition Anupam Prof. Shukla
https://textbookfull.com/product/discrete-problems-in-nature-
inspired-algorithms-1st-edition-anupam-prof-shukla/
Nature Inspired Design of Hybrid Intelligent Systems

1st Edition Patricia Melin
https://textbookfull.com/product/nature-inspired-design-of-
hybrid-intelligent-systems-1st-edition-patricia-melin/
Nature Inspired Computing and Optimization Theory and

Applications 1st Edition Srikanta Patnaik
https://textbookfull.com/product/nature-inspired-computing-and-
optimization-theory-and-applications-1st-edition-srikanta-
patnaik/
Advances in Nature and Biologically Inspired Computing

Proceedings of the 7th World Congress on Nature and
Biologically Inspired Computing NaBIC2015 in
Pietermaritzburg South Africa held December 01 03 2015
1st Edition Nelishia Pillay
https://textbookfull.com/product/advances-in-nature-and-
biologically-inspired-computing-proceedings-of-the-7th-world-
congress-on-nature-and-biologically-inspired-computing-
Studies in Computational Intelligence 637
Xin-She Yang Editor
Nature-Inspired
Computation
in Engineering
Studies in Computational Intelligence
Volume 637
Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: kacprzyk@ibspan.waw.pl
About this Series
The series “Studies in Computational Intelligence” (SCI) publishes new develop-

ments and advances in the various areas of computational intelligence—quickly and
with a high quality. The intent is to cover the theory, applications, and design
methods of computational intelligence, as embedded in the fields of engineering,
computer science, physics and life sciences, as well as the methodologies behind
them. The series contains monographs, lecture notes and edited volumes in
computational intelligence spanning the areas of neural networks, connectionist
systems, genetic algorithms, evolutionary computation, artificial intelligence,
cellular automata, self-organizing systems, soft computing, fuzzy systems, and
hybrid intelligent systems. Of particular value to both the contributors and the
readership are the short publication timeframe and the worldwide distribution,
which enable both wide and rapid dissemination of research output.
More information about this series at http://www.springer.com/series/7092

Xin-She Yang
Editor
Nature-Inspired Computation
in Engineering
123
Editor
Xin-She Yang
School of Science and Technology
Middlesex University
London
UK
ISSN 1860-949X ISSN 1860-9503 (electronic)

Studies in Computational Intelligence
ISBN 978-3-319-30233-1 ISBN 978-3-319-30235-5 (eBook)
DOI 10.1007/978-3-319-30235-5
Library of Congress Control Number: 2016933797
© Springer International Publishing Switzerland 2016

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG Switzerland
Preface
Nature-inspired computation has become increasingly popular in engineering and is

thus leading to a paradigm shift in problem-solving and design approaches. Many
design problems in engineering are complex with multiple nonlinear constraints,
and their design objectives can often be conflicting under stringent design
requirements. Such problems can be very challenging to solve because of their
complexity, nonlinearity and potentially high-dimensionality. Traditional algo-
rithms, especially gradient-based algorithms, can struggle to cope and the results are
not satisfactory. New and alternative methods are highly needed.
In the last two decades, nature-inspired computation and optimization algorithms
have demonstrated their effectiveness as an alternative set of design tools to deal
with such complex design problems. As a result, the popularity of nature-inspired
optimization algorithms has started to increase significantly in recent years. These
nature-inspired algorithms include particle swarm optimization, differential evolu-
tion, cuckoo search, firefly algorithm, bat algorithm, bee algorithms and ant colony
optimization as well as others. The key features of these algorithms are their
simplicity and flexibility, which enable them to highly adaptable and sufficiently
efficient to deal with a wide range of optimization problems. In practice, they are
also easy to be implemented in almost any programming languages, and all these
factors have contributed to their popularity in engineering.
A majority of nature-inspired optimization algorithms can be said to belong to
the class of swarm intelligence (SI) based algorithms. Swarm intelligence concerns
the high-level behaviour arising from simple interactions among multiple agents.
Though the main characteristics of swarming behaviour may be drawn from dif-
ferent sources of inspiration in nature, the procedures and steps of each algorithm
can be very different. However, they seem to use the self-organized abilities of
complex systems based on simple interaction rules. Since most algorithms treat
problems under consideration as a black box, it is possible for such algorithms to
deal with complex problems with different properties whether continuous, discrete
or mixed. In a broad sense, swarm intelligence is part of evolutionary computation
paradigm and bio-inspired computation is part of nature-inspired computation;
v
vi Preface
however, there is still some confusion about some terminologies in the literature.
For example, there are some overlaps and there are no agreed standard definitions
about bio-inspired computation, nature-inspired computation, metaheuristic and
computational intelligence. Therefore, we will not enter the debate about what the
right terminology or subject fields should be, but we will rather focus our attention
on nature-inspired computation and its applications in engineering.
The diversity and rapid advances in nature-inspired computation have resulted in
a much richer literature. Therefore, a timely review is necessary to summarize the
latest developments in terms of algorithm developments and their applications in
engineering. Algorithms and topics include discrete firefly algorithm, discrete
cuckoo search, plant propagation algorithm, parameter-free bat algorithm, gravi-
tational search, biogeography-based algorithm, differential evolution, particle
swarm optimization and others. State-of-the-art applications and case studies
include vehicle routing, swarming robots, discrete and combinatorial optimization,
clustering of wireless sensor networks, cell formation, economic load dispatch,
metamodeling, surrogate-assisted cooperative co-evolution, data fitting and reverse
engineering as well as other real-world applications.
As a timely review volume, this book can be an ideal reference for researchers,
lecturers, graduates and engineers who are interested in latest developments in
nature-inspired computation, artificial intelligence and computational intelligence.
It can also serve as a reference for relevant courses in computer science, artificial
intelligence, machine learning, natural computation, engineering optimization and
data mining.
I thank our editors, Drs. Thomas Ditzinger and Holger Schaepe, and staff at
Springer for their help and professionalism. Last but not least, I thank my family for
the help and support.
London Xin-She Yang

December 2015
Contents
Nature-Inspired Optimization Algorithms in Engineering:

Overview and Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Xin-She Yang and Xingshi He
An Evolutionary Discrete Firefly Algorithm with Novel Operators
for Solving the Vehicle Routing Problem with Time Windows. . . . . . . . 21
Eneko Osaba, Roberto Carballedo, Xin-She Yang and Fernando Diaz
The Plant Propagation Algorithm for Discrete Optimisation:
The Case of the Travelling Salesman Problem . . . . . . . . . . . . . . . . . . . 43
Birsen İ. Selamoğlu and Abdellah Salhi
Enhancing Cooperative Coevolution with Surrogate-Assisted
Local Search. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Giuseppe A. Trunfio
Cuckoo Search: From Cuckoo Reproduction Strategy
to Combinatorial Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Aziz Ouaarab and Xin-She Yang
Clustering Optimization for WSN Based on Nature-Inspired
Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Marwa Sharawi and Eid Emary
Discrete Firefly Algorithm for Recruiting Task in a Swarm
of Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Nunzia Palmieri and Salvatore Marano
Nature-Inspired Swarm Intelligence for Data Fitting in Reverse
Engineering: Recent Advances and Future Trends . . . . . . . . . . . . . . . . 151
Andrés Iglesias and Akemi Gálvez
A Novel Fast Optimisation Algorithm Using Differential Evolution
Algorithm Optimisation and Meta-Modelling Approach . . . . . . . . . . . . 177
Yang Liu, Alan Kwan, Yacine Rezgui and Haijiang Li
vii
viii Contents
A Hybridisation of Runner-Based and Seed-Based

Plant Propagation Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
Muhammad Sulaiman and Abdellah Salhi
Economic Load Dispatch Using Hybrid MpBBO-SQP Algorithm . . . . . 217
Ali R. Al-Roomi and Mohamed E. El-Hawary
Gravitational Search Algorithm Applied to the Cell Formation
Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
Manal Zettam and Bouazza Elbenani
Parameterless Bat Algorithm and Its Performance Study . . . . . . . . . . . 267
Iztok Fister Jr., Uroš Mlakar, Xin-She Yang and Iztok Fister
Contributors
Ali R. Al-Roomi Dalhousie University, Electrical and Computer Engineering,

Halifax, NS, Canada
Roberto Carballedo Deusto Institute of Technology (DeustoTech), University of
Deusto, Bilbao, Spain
Fernando Diaz Deusto Institute of Technology (DeustoTech), University of
Mohamed E. El-Hawary Dalhousie University, Electrical and Computer
Engineering, Halifax, NS, Canada
Bouazza Elbenani Department of Computer Science, Mohammed V University,
Rabat, Morocco
Eid Emary Faculty of Computers and Information, Cairo University, Giza, Egypt
Iztok Fister Faculty of Electrical Engineering and Computer Science, University
of Maribor, Maribor, Slovenia
Iztok Fister Jr. Faculty of Electrical Engineering and Computer Science,
University of Maribor, Maribor, Slovenia
Akemi Gálvez Department of Applied Mathematics and Computational Sciences,
University of Cantabria, Santander, Spain
Xingshi He College of Science, Xi’an Polytechnic University, Xi’an, People’s
Republic of China
Andrés Iglesias Department of Applied Mathematics and Computational
Sciences, University of Cantabria, Santander, Spain; Department of Information
Science, Faculty of Sciences, Toho University, Funabashi, Japan
Alan Kwan School of Engineering, Cardiff University, Cardiff, UK
Haijiang Li School of Engineering, Cardiff University, Cardiff, UK
ix
x Contributors
Yang Liu School of Engineering, Cardiff University, Cardiff, UK

Salvatore Marano Department of Computer Engineering, Modeling, Electronics,
and Systems Science, University of Calabria, Rende, CS, Italy
Uroš Mlakar Faculty of Electrical Engineering and Computer Science, University
of Maribor, Maribor, Slovenia
Eneko Osaba Deusto Institute of Technology (DeustoTech), University of
Aziz Ouaarab LRIT, Associated Unit to the CNRST (URAC 29), Mohammed
V-Agdal University, Rabat, Morocco
Nunzia Palmieri Department of Computer Engineering, Modeling, Electronics,
and Systems Science, University of Calabria, Rende, CS, Italy; School of Science
and Technology, Middlesex University, London, UK
Yacine Rezgui School of Engineering, Cardiff University, Cardiff, UK
Abdellah Salhi Department of Mathematical Sciences, University of Essex,
Colchester, UK
Birsen İ. Selamoğlu University of Essex, Colchester, UK
Marwa Sharawi Faculty of Computer Studies, Arab Open University, Cairo,
Egypt
Muhammad Sulaiman Department of Mathematics, Abdul Wali Khan University
Mardan, Mardan, Khyber Pakhtunkhwa, Pakistan
Giuseppe A. Trunfio DADU, University of Sassari, Alghero, Italy
Xin-She Yang School of Science and Technology, Middlesex University, London,
UK
Manal Zettam Department of Computer Science, Mohammed V University,
Rabat, Morocco
Nature-Inspired Optimization Algorithms
in Engineering: Overview and Applications
Xin-She Yang and Xingshi He
Abstract Nature-inspired computation has become popular in engineering

applications and nature-inspired algorithms tend to be simple and flexible and yet
sufficiently efficient to deal with highly nonlinear optimization problems. In this
chapter, we first review the brief history of nature-inspired optimization algorithms,
followed by the introduction of a few recent algorithms based on swarm intelligence.
Then, we analyze the key characteristics of optimization algorithms and discuss the
choice of algorithms. Finally, some case studies in engineering are briefly presented.
1 Introduction
Optimization is everywhere and important in many applications such as engineer-

ing, business activities and industrial designs. The aims of optimization can be very
diverse—to minimize the energy consumption and costs, to maximize the profit, out-
puts, performance and efficiency. Thus, optimization concerns almost every appli-
cation from engineering design to business planning and from holiday planning to
vehicle routing. In reality, because resources, time and money are always limited in
real-world applications, the optimal use these valuable resources under various con-
straints should be sought. As most real-world applications are often highly nonlinear,
it requires sophisticated optimization tools so as to find quality and feasible solutions.
Nature-inspired computation is now playing an increasingly important role in
many areas, including artificial intelligence, computational intelligence, data min-
ing, machine learning and optimization [1]. Good examples of such algorithms are bat
algorithm, cuckoo search, particle swarm optimization, firefly algorithm, bee algo-
rithms and others [1–4]. Obviously, their popularity can be attributed to many factors
X.-S. Yang (B)

School of Science and Technology, Middlesex University, The Burroughs,
London NW4 4BT, UK
e-mail: x.yang@mdx.ac.uk
X. He
College of Science, Xi’an Polytechnic University, Jinhua South Road,
Xi’an, People’s Republic of China
© Springer International Publishing Switzerland 2016 1

X.-S. Yang (ed.), Nature-Inspired Computation in Engineering,
Studies in Computational Intelligence 637, DOI 10.1007/978-3-319-30235-5_1
2 X.-S. Yang and X. He
and one key reason is that these algorithms are simple, flexible, efficient and highly
adaptable. In addition, from the implementation point of view, they are very simple
to be implemented in any programming language. As a result, these algorithms have
been applied in a wide spectrum of problems in real-world applications.
The main objective of this chapter is to provide an overview of the history of
the nature-inspired computation and review some of the recent nature-inspired algo-
rithms for optimization. Some applications in engineering will also be highlighted.
Therefore, the chapter is organized as follows. Section 2 outlines the basic formula-
tion of optimization problems and the search for optimality, and then Sect. 3 presents
the brief history of the recent developments in nature-inspired computation. Section 4
highlights the key features of metaheuristic algorithms, followed by the introduction
to some recent algorithms based on swarm intelligence in Sect. 5. Section 6 discusses
the choice of algorithms and then Sect. 7 highlights some recent applications of bio-
inspired computation in engineering.
2 Optimization and Optimality
Many problems in engineering can be formulated as an optimization problem, though

how to formulate and the way of formulations usually depend on the perspective that
we are looking at the problem. Though there are many different ways of formulations,
the exact formulations may also depend on the subject area and thus such formulations
may directly or indirectly affect the choice of the solution techniques. Thus, this may
be a subject matter, however, we will focus on the general context of optimization
and search for optimality.
2.1 Optimization
From a mathematical point of view, all almost optimization problems can be formu-
lated in the following generic form:
minimize f i (x), (i = 1, 2, . . . , M), (1)

x∈d
subject to h j (x) = 0, ( j = 1, 2, . . . , J ), (2)
gk (x) ≤ 0, (k = 1, 2, . . . , K ), (3)
where f i (x), h j (x) and gk (x) are functions of the design vector
x = (x1 , x2 , . . . , xd )T . (4)
Here the components xi of x are called design or decision variables, and they can be
real continuous, discrete or the mixed (partly continuous and partly discrete). Here,
Nature-Inspired Optimization Algorithms in Engineering … 3
the functions f i (x) where i = 1, 2, . . . , M are called the objective functions or

simply cost functions, and in the case of M = 1, there is only a single objective and
the optimization process become a single objective optimization. The space spanned
by the decision variables is called the design space or search space d , while the
space formed by the objective function values is called the solution space or response
space. The equalities for h j and inequalities for gk are called constraints.
It is worth pointing out that we can also write the inequalities in the other
way ≥ 0, and we can also formulate the objectives as a maximization problem. In
a rare but extreme case where there is no objective at all, there are only constraints.
Such a problem is called a feasibility problem because any feasible solution is an
optimal solution. For some difficult design problems with multiple, complex, poten-
tially conflicting constraints, to find even a feasible solution may be a challenging
task.
Obviously, if we try to classify optimization problems according to the number of
objectives, then there are two categories: single objective M = 1 and multiobjective
M > 1. Multiobjective optimization is also referred to as multicriteria or even multi-
attributes optimization in the literature. In real-world problems, most optimization
tasks are multiobjective. Though the algorithms we will discuss in this chapter are
equally applicable to multiobjective optimization with some modifications, we will
mainly focus on single objective optimization problems.
Similarly, we can also classify optimization in terms of number of constraints
J + K . If there is no constraint at all (i.e., J = K = 0), then it is called an
unconstrained optimization problem. If K = 0 and J ≥ 1, it is called an equality-
constrained problem, while the case for J = 0 and K ≥ 1 becomes an inequality-
constrained problem.
It is worth pointing out that equality constraints have special properties, and
require special care. One drawback is that the volume of satisfying an equality is
essentially zero in the search space, thus very difficult to get sampling points that
satisfy the equality exactly. Some tolerance or allowance is used in practice.
Furthermore, we can also use the actual function forms for classifying optimiza-
tion problems. The objective functions can be either linear or nonlinear. If the con-
straints h j and gk are all linear, then the problem becomes a linearly constrained
optimization problem. If both the constraints and the objective functions are all lin-
ear, it becomes a linear programming problem. Here ‘programming’ has nothing to
do with computing programming, it means planning and/or optimization. However,
generally speaking, all f i , h j and gk are nonlinear, we have to deal with a nonlinear
optimization problem.
2.2 Search for Optimality
Once an optimization problem is formulated correctly, the main task is to find

the optimal solutions by a proper solution procedure using the right mathematical
techniques. However, there may not be any right solution procedure for some classes
of problems, at least, it is not easy to find a right technique to solve a particular type
of problem.
In many ways, searching for the optimal solution is very similar to treasure hunt-
ing. Imagine a scenario that we are trying to hunt for a hidden treasure in a hilly
landscape within a time limit. In one extreme, suppose we are blind-fold without any
guidance, the search process is essentially a pure random search, which is usually
not efficient. In another extreme, if we are told the treasure is placed at the highest
peak of a known region, we will then directly climb up to the steepest cliff and try to
reach to the highest peak, and this scenario corresponds to the classical hill-climbing
techniques. In most cases, the search is between these extremes. We are not blind-
fold, and we do not know where to look for. Obviously, it is physically impossible
to search every single square inch of an extremely large hilly region so as to find the
treasure.
In reality, the most likely search scenario is that we will do a random walk,
while looking for some hints; we look at some place almost randomly, then move
to another plausible place, then another and so on. Such a random walk is a main
characteristic of modern search algorithms [1]. Obviously, we can either do the
treasure-hunting alone, so the whole path is a trajectory-based search, and simulated
annealing is such a kind. Alternatively, we can ask a group of people to do the hunting
and share the information (and any treasure found), and this scenario uses the so-
called swarm intelligence and corresponds to the algorithms such as particle swarm
optimization and firefly algorithm, as we will discuss later in detail. If the treasure is
really important and if the area is extremely large, the search process will take a very
long time. If there is no time limit and if any region is accessible (for example, no
islands in a lake), it is theoretically possible to find the ultimate treasure (the global
optimal solution).
Going with this analogy even further, we can refine our search strategy a little bit
further. Some hunters are better than others. We can only keep the better hunters and
recruit new ones, this is something similar to the genetic algorithms or evolutionary
algorithms where the search agents are improving. In fact, as we will see in almost
all modern metaheuristic algorithms, we try to use the best solutions or agents,
and randomize (or replace) the not-so-good ones, while evaluating each individual’s
competence (fitness) in combination with the system history (use of memory). With
such a balance, we intend to design better and efficient optimization algorithms.
3 A Brief History of Nature-Inspired Computation
Humans’ approach to problem-solving has always been heuristic or metaheuristic—

by trial and error, especially at the early periods of human history. Many important
discoveries were done by ‘thinking outside the box’, and often by accident; that
is heuristics. In fact, our daily learning experience during our childhood is domi-
nantly heuristic. Despite its ubiquitous nature, metaheuristics as a scientific method
to problem solving is indeed a modern phenomenon, though it is difficult to pinpoint

when the metaheuristic method was first used. Alan Turing was probably the first
to use heuristic algorithms during the second World War when he was breaking the
Enigma ciphers at Bletchley Park. Turing called his search method heuristic search,
as it could be expected it worked most of time, but there was no guarantee to find
the correct solution; however, it was a tremendous success. In his National Physical
Laboratory report on Intelligent machinery in 1948, Turing outlined his innova-
tive ideas of machine intelligence and learning, neural networks and evolutionary
algorithms [7].
The first golden period is the 1960 s and 1970 s for the development of evolution-
ary algorithms. John Holland and his collaborators at the University of Michigan
first developed the genetic algorithms in 1960 s and 1970s. As far back to 1962,
Holland studied the adaptive system and was the first to use crossover and recombi-
nation manipulations for modeling such system. His seminal book summarizing the
development of genetic algorithms was published in 1975 [8]. In the same year, De
Jong finished his important dissertation showing the potential and power of genetic
algorithms for a wide range of objective functions, either noisy, multimodal or even
discontinuous. The essence of a genetic algorithm (GA) is based on the abstraction
of Darwinian evolution and natural selection of biological systems and representing
them in the mathematical operators: crossover or recombination, mutation, fitness,
and selection of the fittest.
During the same decade, Ingo Rechenberg and Hans-Paul Schwefel both then at
the Technical University of Berlin developed a search technique for solving optimiza-
tion problem in aerospace engineering, called evolutionary strategy, in 1963. There
was no crossover in this technique, only mutation was used to produce an offspring
and an improved solution was kept at each generation. This was essentially a simple
trajectory-style hill-climbing algorithm with randomization. On the other hand, as
early as 1960, Lawrence J. Fogel intended to use simulated evolution as a learning
process as a tool to study artificial intelligence. Then, in 1966, L.J. Fogel, together
A.J. Owen and M.J. Walsh, developed the evolutionary programming technique by
representing solutions as finite-state machines and randomly mutating one of these
machines [9]. The above innovative ideas and methods have evolved into a much
wider discipline, called evolutionary algorithms and/or evolutionary computation
[10, 11].
It may not be entirely clear how bio-inspired computation is also relevant to arti-
ficial neural network, support vector machine, and other many learning techniques
in the context of optimization. However, these methods all intend to minimize their
learning errors and prediction (capability) errors via iterative trials and errors. For
example, artificial neural networks are now routinely used in many applications. In
1943, W. McCulloch and W. Pitts proposed the artificial neurons as simple informa-
tion processing units. The concept of a neural network was probably first proposed
by Alan Turing in his 1948 NPL report concerning ‘intelligent machinery’ [7, 12].
Significant developments were carried out from the 1940 s and 1950 s to the 1990 s
with more than 60 years of history. Another example is the so-called support vec-
tor machine as a classification technique that can date back to the earlier work by
V. Vapnik in 1963 on linear classifiers, and the nonlinear classification with kernel
techniques were developed by V. Vapnik and his collaborators in the 1990s. A sys-
tematical summary in Vapnik’s book on the Nature of Statistical Learning Theory
was published in 1995 [13].
Another important period for metaheuristic algorithms is the two decades of 1980 s
and 1990s. First, the development of simulated annealing (SA) in 1983, an optimiza-
tion technique, pioneered by S. Kirkpatrick, C.D. Gellat and M.P. Vecchi, inspired by
the annealing process of metals. The actual first usage of memory in modern meta-
heuristics is probably due to Fred Glover’s Tabu search in 1986, though his seminal
book on Tabu search was published later in 1997 [14]. Marco Dorigo finished his
PhD thesis in 1992 on optimization and natural algorithms [2], in which he described
his innovative work on ant colony optimization (ACO). This search technique was
inspired by the swarm intelligence of social ants using pheromone as a chemical
messenger. At the same time in 1992, John R. Koza of Stanford University published
a treatise on genetic programming which laid the foundation of a whole new area of
machine learning, revolutionizing computer programming [15]. As early as in 1988,
Koza applied his first patent on genetic programming. The basic idea is to use the
genetic principle to breed computer programs so as to gradually produce the best
programs for a given type of problem.
In 1995, the particle swarm optimization (PSO) was developed by American
social psychologist James Kennedy, and engineer Russell C. Eberhart [3], based
on the swarming behaviour of fish and birds. The multiple agents, called particles,
swarm around the search space starting from some initial random guess. The swarm
communicates the current best and shares the global best so as to focus on the
quality solutions. Since its development, there have been about 20 different variants
of particle swarm optimization techniques, and have been applied to almost all areas
of tough optimization problems.
Then, slightly later in 1996 and 1997, R. Storn and K. Price developed their
vector-based evolutionary algorithm, called differential evolution (DE) [16], and
this algorithm proves more efficient than genetic algorithms in many applications. In
1997, the ‘no free lunch theorems for optimization’ were proved by D.H. Wolpert and
W.G. Macready [17, 18]. Researchers have been always trying to find better algo-
rithms, or even universally robust algorithms, for optimization, especially for tough
NP-hard optimization problems. However, these theorems state that if algorithm A
performs better than algorithm B for some optimization functions, then B will out-
perform A for other functions. That is to say, if averaged over all possible function
space, both algorithms A and B will perform on average equally well. Alternatively,
there is no universally better algorithms exist. However, researchers realized that we
do not need the average over all possible functions for a given optimization problem.
What we want is to find the best solutions, which has nothing to do with average
over all possible function space. In addition, we can accept the fact that there is no
universal or magical tool, but we do know from our experience that some algorithms
indeed outperform others for given types of optimization problems. So the research
may now focus on finding the best and most efficient algorithm(s) for a given set of
problems. The objective is to design better algorithms for most types of problems,
not for all the problems. Therefore, the search is still on.
Another very exciting period for developing metaheuristic algorithms is the first
decade of the 21st century. In 2001, Zong Woo Geem et al. developed the harmony
search (HS) algorithm [19], which has been widely applied in solving various opti-
mization problems such as water distribution, transport modelling and scheduling. In
2004, S. Nakrani and C. Tovey proposed the honey bee algorithm and its application
for optimizing Internet hosting centers [20], which followed by the development of
virtual bee algorithm by Xin-She Yang in 2005. At the same time, the bees algorithm
was developed by D.T. Pham et al. in 2005 and the artificial bee colony (ABC) was
developed by D. Karaboga in 2005. Then, in late 2007 and early 2008, the firefly
algorithm (FA) was developed by Xin-She Yang [1, 4], which has generated a wide
range of interests. In 2009, Xin-She Yang at Cambridge University, UK, and Suash
Deb at Raman College of Engineering, India, proposed an efficient cuckoo search
(CS) algorithm [21, 22], and it has been demonstrated that CS can be far more effec-
tive than some existing metaheuristic algorithms. In 2010, the bat algorithm was
developed by Xin-She Yang for continuous optimization, based on the echolocation
behaviour of microbats [23]. In 2012, the flower pollination algorithm was developed
by Xin-She Yang, and its efficiency is very promising.
As the literature is expanding rapidly, the number of nature-inspired algorithms
has increased dramatically [1, 24] As we can see, more and more metaheuristic
algorithms are being developed. Such a diverse range of algorithms necessitates a
systematic summary of various metaheuristic algorithms, however, we only briefly
outline a few of these recent algorithms in the rest of the chapter. But before we pro-
ceed, let us pause and highlight the key characteristics of nature-inspired algorithms.
4 Main Characteristics of Nature-Inspired Algorithms
The key characteristics of nature-inspired optimization algorithms can be analyzed

from many different perspectives. From the algorithmic development point of view,
we can analyze algorithms in terms of algorithm search behaviour, adaptation and
diverse as well as algorithmic components in terms of mathematical equations for
iterations.
4.1 Exploration and Exploitation
The key aim of an algorithm is to generate new solutions that should be better than
previous solutions. For an ideal algorithm, new solutions should be always better than
existing solutions and it can be expected that the most efficient algorithms are to find
the best solutions with the least minimum efforts (ideally in one move). However,
such ideal algorithms may not exist at all.
For stochastic algorithms, solutions do not always get better. In fact, it can be
advantageous to select not-so-good solutions, which can help the search process
escape from being trapped at any local optima. Though this may be counter-intuitive,
such stochastic nature now forms the essential component of modern metaheuristic
algorithms [4]. Exploitation typically uses any information obtained from the prob-
lem of interest so as to help to generate new solutions that are better than existing
solutions. However, this process is typically local, and information (such as gradi-
ents) is also local. Therefore, it is mainly for local search. For example, hill-climbing
is a method that uses derivative information to guide the search procedure. In fact,
new steps always try to climb up the local gradient. The advantage of exploitation is
that it usually leads to very high convergence rates, but its disadvantage is that it can
get stuck in a local optimum because the final solution point largely depends on the
starting point.
On the other hand, exploration makes it possible to explore the search space in
far away regions more efficiently [25], and it can generate solutions with enough
diversity and far from the current solutions. Therefore, the search is typically on a
global scale. The advantage of exploration is that it is less likely to get stuck in a local
mode, and the global optimality can be more accessible. However, its disadvantages
are slow convergence and waste of lot computational efforts because many new
solutions can be far from global optimality.
However, whether local search or global search, all depends on the actual search
mechanisms within an algorithm. Sometimes, there is no clear cut between local or
global. In addition, a fine balance may be required so that an algorithm can achieve
the good performance. Too much exploitation and too little exploration means the
system may converge more quickly, but the probability of finding the true global
optimality may be low. Conversely, too little exploitation and too much exploration
can cause the search path meander with very slow convergence. The optimal balance
should mean the right amount of exploration and exploitation, which may lead to
the optimal performance of an algorithm. Therefore, a proper balance is crucially
important to ensure the good performance of an algorithm.
4.2 Diversity and Adaptation
Looking from a different perspective, nature-inspired algorithms can have both diver-
sity and adaptation. Adaptation in nature-inspired algorithms can take many forms.
For example, the ways to balance exploration and exploitation are the key form of
adaptation [26]. As diversity can be intrinsically linked with adaptation, it is better
not to discuss these two features separately. For example, in genetic algorithms, rep-
resentations of solutions are usually in binary or real-valued strings [8, 26], while in
swarm-intelligence-based algorithms, representations mostly use real number solu-
tion vectors. As another example, the population size used in an algorithm can be
fixed or varying. Adaptation in this case may mean to vary the population size so as
to maximize the overall performance.
For a given algorithm, adaptation can also occur to adjust its algorithm-dependent
parameters. As the performance of an algorithm can largely depend on its parameters,
the choice of these parameter values can be very important. Similarly, diversity in
metaheuristic algorithms can also take many forms. The simplest diversity is to allow
the variations of solutions in the population by randomization. For example, solution
diversity in genetic algorithms is mainly controlled by the mutation rate and crossover
mechanisms, while in simulated annealing, diversity is achieved by random walks.
In addition, adaptation can also be in the form of self-tuning in terms of parameters
so as to achieve better performance automatically [27].
In most swarm-intelligence-based algorithms, new solutions are generated accord-
ing to a set of deterministic equations, which also include some random variables.
Diversity is represented by the variations, often in terms of the population variance.
Once the population variance is getting smaller (approaching zero), diversity also
decreases, leading to converged solution sets. However, if diversity is reduced too
quickly, premature convergence may occur. Therefore, a right amount of randomness
and the right form of randomization can be crucial.
Both diversity and adaptation are important to ensure the effectiveness of an
algorithm. A good diversity will ensure the population can explore different regions
and can thus maintain a non-zero probability of finding the global optimality of the
problem. In addition, good adaption will enable the algorithm to adapt to suit the
problem and landscape under consideration, and thus may ensure a potentially better
convergence than non-adaptive approaches. However, it is not yet clear what a good
degree of diversity should be and what kind of adaptation mechanisms can be used.
4.3 Key Operators in Algorithms
Algorithms can also be analyzed by studying in detail the key algorithmic operators
used in the construction of the algorithms. For example, in the well-established class
of genetic algorithms [8], genetic operators such as crossover (or recombination),
mutation and selection are used [26].
In genetic algorithms, crossover is the operation of generating two new
solutions (offsprings) from two existing solutions (parents) by swapping rele-
vant/corresponding parts of their solutions. This is similar to the main crossover
feature in the biological systems. Crossover usually provides good mixing of solu-
tion characteristics and can usually generate completely new solutions if the two
parents are different. Obviously, when the two parents are identical, offspring solu-
tions will also be identical, and thus provides a good mechanism to maintain good
convergence. It is worth pointing out that crossover in contemporary algorithms may
take different forms, though its essence remains the same.
Mutation is a mechanism to generate a new solution from a single solution by
changing a single site or multiple sites. As in the evolution in nature, mutation often
generates new characteristics that can adapt to new environments, and thus new
solutions during the search process can also be generated in this way. However, in
many nature-inspired algorithms, mutation typically becomes a vector and thus has
taken very different forms from the original binary mutation in the original genetic
algorithms.
Though both crossover and mutation are ways of generating new solutions, selec-
tion provides a pressure for evolution. In other words, selection provides a measure
or mechanism to determine what is better and selects the fittest. This mimics the
key feature of the Darwinian evolution in terms of the survival of the fittest. Without
selection, new solutions and new characteristics will not be selected properly, which
may lead to a diverse and less convergent system. With too much selection pressure,
many new characteristics will die away quickly. If the environment also changes and
if some solutions/characteristics dominate the population, the system may lead to
premature convergence. Thus, a proper selection pressure is also important to ensure
that only good characteristics or solutions that happen to fit into the new environment
can survive.
If we try to link crossover and mutation with exploration and exploitation, we
have to look at them in greater detail because such links can be very subtle. For
example, crossover can provide both exploration and exploitation capabilities, while
mutation mainly provides exploration. On the other hand, selection provides a good
way of exploitation by selecting good solutions. Thus, in order to provide a good
exploration ability to an algorithm, a higher crossover rate is needed, and that is why
the crossover probability is typically over 0.9 in genetic algorithms. However, new
solutions should not be too far from existing solutions in many cases, and to avoid too
much exploration, the mutation rate should be usually low. For example, in genetic
algorithms, the mutation rate is typically under 0.05 [4]. It is worth pointing out that
these values are based on empirical observations (both numerically and biologically).
In different metaheuristic algorithms, it is usually quite a challenging task to decide
what values are most appropriate, and such choices may need detailed parametric
studies and numerical experiments.
5 Some Swarm-Based Algorithms
As nature-inspired algorithms have become more popular, the literature has expanded
significantly. In fact, the number of nature-inspired algorithms has increased dra-
matically in recent years, thus it is not possible to review a good subset of all these
algorithms here. To demonstrate the main points, we will only briefly introduce a
few algorithms that are among most recent and most popular algorithms.
All almost algorithms use some updating rules, derived from some characteristics
in nature, though a vast majority draw inspiration from swarm-based characteristics,
and thus forming a class of swarm intelligence. It is worth pointing out that the updat-
ing or iterative equations can be either linear or nonlinear, though most algorithms
have linear updating equations. Linear systems may be easier to implement, but the
diversity and richness of the system behaviour may be limited.
On the other hand, for the algorithms with nonlinear updating systems, the char-
acteristics of the algorithm can be richer, which may lead to some advantages over
algorithms with linear updating equations. For example, studies show that firefly algo-
rithm can automatically subdivide the whole population into multiple sub-swarms
due to its nonlinear distance-dependent attraction mechanism. At the moment, it is
still not clear if nonlinear systems are always potentially better. In the rest of this
section, when we describe the key formulations of each algorithm, we will mainly
focus on the key characteristics of its updating equations as an iterative system.
5.1 Particle Swarm Optimization
Particle swarm optimization (PSO), developed by Kennedy and Eberhart in 1995

[3] has two updating/iterative equations for position xi and velocity vi of a particle
system:
vit+1 = vit + α1 [g ∗ − xit ] + β2 [xi∗ − xit ], (5)
xit+1 = xit + vit+1 . (6)
There are two parameters α and β in addition to the population size n. Here, two
random vectors 1 and 2 are uniformly distributed in [0, 1].
As this system is linear, it is quite straightforward to carry out stability analysis
[28]. Though the PSO system usually converges very quickly, it can have premature
convergence. To remedy this, over 20 different variants have been proposed with
moderate success. Among many improvements, the use of an inertia function by Shi
and Eberhart [29] seems to stabilize the system well. Other developments include
the accelerated PSO by Yang et al. [30] and some reasoning techniques with PSO by
Fister Jr. et al. [31]. Hybridization with other algorithms also proves useful.
5.2 Firefly Algorithm
The firefly algorithm (FA) developed by Xin-She Yang in 2008 [4] has a single
nonlinear equation for updating the locations (or solutions) of fireflies:
xit+1 = xit + β0 e−γri j (x tj − xit ) + α it ,

2
(7)
where the second term between any two fireflies (i and j) due to the attraction is highly
nonlinear because the attraction is distance-dependent. Here, β0 is the attractiveness
at r = 0, while α is the randomization parameter. In addition, it is a vector of random
numbers drawn from a Gaussian distribution at time t. The nonlinear nature of the
updating equation means that the local short-distance attraction is much stronger than
the long-range attraction, and consequently, the whole population can automatically
subdivide into multiple subgroups (or multi-swarms). Under the right conditions,
each subgroup can swarm around a local mode, thus, FA can naturally deal with
multimodal problems effectively.
Studies show that firefly algorithm can be very efficient in solving many different
problems such as classifications and clustering problems [32] as well as scheduling
problems [33]. The extensive studies of the firefly algorithm and its variants were
reviewed by Fister et al. [24, 34, 35].
5.3 Cuckoo Search
Cuckoo search (CS), developed in 2009 by Xin-She Yang and Suash Deb [21], is
another nonlinear system [22, 36], enhanced by Lévy flights [37]. This makes CS
potentially more efficient than PSO and genetic algorithms.
The nonlinear equations are also controlled by a switching parameter or proba-
bility pa . The nonlinear system can be written as
xit+1 = xit + αs ⊗ H ( pa − ) ⊗ (x tj − xkt ), (8)
xit+1 = xit + αL(s, λ), (9)
where two different solutions x tj and xkt are randomly selected. In addition, H (u) is
a Heaviside step function, is a random number drawn from a uniform distribution,
and s is the step size. The so-called Lévy flights are realized by drawing random step
sizes L(s, λ) from a Lévy distribution, which can be written as
λΓ (λ) sin(πλ/2) 1
L(s, λ) ∼ , (s > 0), (10)
π s 1+λ
where ∼ denotes the drawing of samples from a probability distribution. In addition,
α > 0 is the step size scaling factor, which should be related to the scales of the
problem of interest.
Some recent reviews have been carried out by Yang and Deb [38] and Fister Jr. et
al. [39]. Mathematical analysis also suggested that CS can have global convergence
[40]. Recent studies also suggests that CS can have autozooming capabilities so
that the search process can automatically focus on the promising areas due to the
combination of self-similar, fractal-like and multiscale search capabilities [1]. By

analyzing the algorithmic system, it can be seen that simulated annealing, differential
evolution and APSO are special cases of CS, and that is one of the reasons why CS
is so efficient.
5.4 Bat Algorithm
The bat algorithm, developed by Xin-She Yang in 2010 [23] based on the echolocation
behavior of microbats, is a linear system with updating equations for velocity vit and
location xit of a bat. At iteration t in a d-dimensional search or solution space, the
linear system can be written as
f i = f min + ( f max − f min )β, (11)
vit = vit−1 + (xit−1 − x∗ ) f i , (12)
xit = xit−1 + vit , (13)
where x∗ is the current best solution and β ∈ [0, 1] is a random vector drawn from
a uniform distribution.
Though variations of the loudness and pulse emission rates are regulated
Ait+1 = α Ait , rit+1 = ri0 [1 − exp(−γt)], (14)
they seems to control the exploration and exploitation components in the algorithm.
Here, 0 < α < 1 and γ > 0 are constants. BA has attracted a lot of interest and
thus the literature is expanding. For example, Yang extended it to multiobjective
optimization [41, 42] and Fister et al. formulated a hybrid bat algorithm [43, 44].
5.5 Differential Evolution
Differential evolution (DE), developed by R. Storn and K. Price in 1996 and 1997
[16, 45], uses a linear mutation equation
u it = xrt + F(x tp − xqt ), (15)
where F is the differential weight in the range of [0, 2]. Here, r, p, q, i are four
different integers generated by random permutation.
The crossover operator in DE is controlled by a crossover probability Cr ∈ [0, 1]
and the actual crossover can be carried out in two ways: binomial and exponential.
Crossover can typically be carried out along each dimension j = 1, 1, . . . , d where

d is the number of dimensions. Thus, we have
⎧ t
⎨u j,i if rand < Cr or j = I,
j,i =
x t+1 (16)
⎩
x tj,i if rand > Cr and j = I,
where I is a random integer from 1 to d, so that u it+1 = xit . It may not easy to
figure out whether such crossover is linear or not, a detailed analysis indicates that
crossover can be treated as a linear operator and thus DE is still a linear system in
terms of updating equations. There are many different variants of DE [46] and many
hybridized version in the literature.
5.6 Flower Pollination Algorithm
Another linear system is the flower pollination algorithm (FPA), developed by Xin-
She Yang in 2012 [47], inspired by the flower pollination process of flowering plants.
The main updating equation for a solution/position xi at any iteration t is
xit+1 = xit + γ L(λ)(g∗ − xit ), (17)
where g∗ is the current best solution found among all solutions and γ is a scaling
factor to control the step size. The step sizes as Lévy flights are drawn from a Levy
distribution, that is
λΓ (λ) sin(πλ/2) 1
L(λ) ∼ , (s s0 > 0). (18)
π s 1+λ
Here Γ (λ) is the standard gamma function, and this distribution is valid for large
steps s > 0.
Another equation is also linear
xit+1 = xit + (x tj − xkt ), (19)
which is mainly for local search, though whether it is local or global in this equation
will depend on the solutions chosen from the population. If two solutions are very
different or far away, the search can be global, while two very similar solutions will
typically lead to local search. Here, x tj and xkt are pollen from different flowers of the
same plant species. Here, from a uniform distribution in [0, 1]. In essence, this is
mostly a local mutation and mixing step, which can help to converge in a subspace.
However, it is worth pointing out that the generation of new moves using Lévy flights
can be tricky. There are efficient ways to draw random steps correctly from a Lévy
distribution [1, 37]. The switch between two search branches are controlled by a
probability ps . Recent studies suggested that flower pollination algorithm is very
efficient for multiobjective optimization [48].
6 Theoretical Analysis and Choice of Algorithms
Though different algorithms can have very different iterative equations, it is still
possible to find some common features among different algorithms. However, to
systematically analyze many different algorithms can be a very challenging task.
On the other hand, there is no free lunch in theory when the performance metrics
averaged over all possible problems [17], but ranking of algorithms may still be
possible for a given set of particular problems because the performance in this case
is no longer concerned with all problems and averaging. Thus, for a given set or type
of problems, one of the main tasks is to find some effective algorithms to use among
many different algorithms and thus the choice of algorithms may not be an easy task.
6.1 Theoretical Aspects of Algorithms
The ways of generating new solutions via updating equations depend on the structure
and topology of the search neighbourhood. New solution can be in either a local
neighborhood such as local random walks, the global domain such as those moves
by Lévy or uniform initialization, or cross-scale (both local and global, e.g., Lévy
flights, exponential, power-law, heavy-tailed). For simplicity, let us focus on the
most fundamental operators such as crossover, mutation and selection. From the
mathematical point of view, selection is easy to understand, we will now focus only
on crossover and mutation.
The crossover operator can be written mathematically as
⎛ ⎞ ⎛ ⎞
xit+1 xit
⎝ ⎠ = C(xit , x tj , pc ) = C( pc )⎝ ⎠, (20)
x t+1
j x tj
where pc is the crossover probability, though the exact form of C( ) depends on the
actual crossover manipulations such as at one site or at multiple sites simultaneously.
The selection of i and j can be by random permutation. Furthermore, the choice of
parents can often be fitness-dependent, based on the relative fitness of the parents in
the population. In this particular case, the functional form for the crossover function
can be even more complex. For example, C( ) can depend on all the individuals in
the population, which may lead to C(x1t , x2t , . . . , xnt , pc ) where n is the population
size.
On the other hand, mutation is the variation or modification of an existing solution

and can be written schematically as
xit+1 = M(xit , pm ), (21)
where pm is the mutation rate. However, the form M( ) depends on the coding and
the number of mutation sites. This can be written in most cases as a random walk
xit+1 (new solution) = xit (old solution) + α(randomization), (22)
where α is a scaling factor controlling how far the random walks can go [4, 23].
However, mutation can also be carried out over a subset of the population or the
mutation operator can also be affected by more than one solution. For example, the
mutation operator in differential evolution takes the form xrt + F(x tp − xqt ), which
involves three different solutions.
In general, the solutions can be generated in parallel by random permutation, and
thus we may have a more generic form for modifications
⎛ ⎞
x1t+1
⎜ .. ⎟
⎝ . ⎠ = G(x1 , x2 , . . . , xn , ε, β),
t t t
(23)
xnt+1
where G( ) can be in a very complex form, involving vectors, matrices, random

variable ε and parameter β. However, it still lacks a unified mathematical framework
for algorithm analysis.
6.2 Choice of Algorithms
A very important but also very practical question is how to choose an algorithm for
a given problem. This may be implicitly linked to another question: what type of
problems can an algorithm solve most effectively. In many applications, the problem
under consideration seem to be fixed, we have to use the right tool or methods to
solve in the most effective way. Therefore, there are two types of choices and thus
two relevant questions:
• For a given type of problems, what is the best algorithm to use?
• For a given algorithm, what kinds of problems can it solve?
The first question is harder than the second question, though the latter is not easy to
answer either. For a given type of problems, there may be a set of efficient algorithms
to solve such problems. However, in many cases, we may not know how efficient an
algorithm can be before we actually try it. In some cases, such algorithms may still
need to be developed. Even for existing algorithms, the choice largely depends on
the expertise of the decision-maker, the resources and the type of problems. Ideally,
the best available algorithms and tools should be used to solve a given problem;
however, the proper use of these tools may still depend on the experience of the user.
In addition, the resources such as computational costs and software availability
and time allowed to produce the solution will also be important factors in deciding
what algorithms and methods to use.
On the other hand, for a given algorithm, the type of problems it can solve can be
explored by using it to solve various kinds of problems and then compare and rank so
as to find out how efficient it may be. In this way, the advantages and disadvantages
can be identified, and such knowledge can be used to guide the choice of algorithm(s)
and the type of problems to tackle. The good thing is that the majority of the literature,
including hundreds of books, have placed tremendous emphasis in answering this
question. Therefore, for traditional algorithms such as gradient-based algorithms and
simplex methods, we know what types of problems they usually can solve. However,
for new algorithms, as in the cases of most nature-inspired algorithms, we have to
carry out extensive studies to validate and test their performance. Obviously, any
specific knowledge about a particular problem is always helpful for the appropriate
choice of the best and most efficient methods for the optimization From the algorithm
development point of view, how to best incorporate problem-specific knowledge is
still an ongoing challenging question.
7 Applications in Engineering
The applications of nature-inspired algorithms in engineering is very diverse and to

cover a good fraction of these applications may take a big volume. In the present
book, some of the latest developments will be reviewed and introduced.
Therefore, in the rest of this section, we will briefly highlight a few applications
to show the diversity and effectiveness of nature-inspired optimization algorithms.
• Structural optimization: Design optimization in civil engineering is essentially

structural optimization because the main tasks to design structures so as to maxi-
mize the performance index and to minimize the costs, subject to complex stringent
design codes. Typical examples are pressure vessel design, speed reducer design,
dome and tower designs [22, 41, 42], as well as combination of optimization with
finite element simulation and other designs [42, 49].
• Scheduling and Routing: Scheduling problems such as airline scheduling and vehi-
cle routing can have important real-world applications. The well-known example
is the so-called travelling salesman problem that requires to visit each city exactly
once so that the overall travelled distance must be minimized. Nature-inspired
algorithms have been used to solve such tough problems with promising results
[50–52].
• Software testing: In software engineering, a substantial amount of the costs
for software developments can be related to software testing to make sure the
software can execute smoothly and to meet the design requirements. This means
that multiple independent test paths should be generated to test the behaviour of
the software, though the generation of truly independent paths can be a difficult
task. Nature-inspired optimization algorithms have also been used in this area with
good results [53].
• Image processing: Image processing is a big area with a huge literature. Image
segmentation and feature selection can often be formulated as an optimization
problem, and thus can be tackled by optimization techniques in combination with
traditional image processing techniques [54]. Image automatic registration and
clustering can be solved using nature-inspired algorithms with good performance
[32, 55].
• Data mining: Data mining is an active research area with diverse applications.
Classification and clustering can be closely related to optimization and many
hybrid techniques have recently been developed by combing traditional data min-
ing methods such as k-mean clustering with nature-inspired optimization algo-
rithms such as the firefly algorithm, cuckoo search and bat algorithm. Recent
studies by Fong et al. showed that such hybrid methods can obtain very good
results [56]. For example, Senthilnath et al. compared over a dozen different clus-
tering algorithms, they concluded that the approach based the firefly algorithm can
obtain the best results with the least amount of computational efforts [32].
Obviously, there are many other applications and case studies, interested readers
can refer to more specialized literature [1, 5, 6, 57] and the later chapters of this
book
References
1. Yang, X.S.: Nature-Inspired Optimization Algorithms. Elsevier, London (2014)

2. Dorigo, M., Di Caro, G., Gambardella, L.M.: Ant algorithms for discrite optimization. Artif.
Life 5(2), 137–172 (1999)
3. Kennedy, J., Eberhart, R.C.: Particle swarm optimization. In: Proceedings of IEEE International
Conference on Neural Networks, pp. 1942–1948. Piscataway, NJ (1995)
4. Yang, X.S.: Nat.-Inspir. Metaheuristic Algorithms. Luniver Press, Bristol (2008)
5. Eiben, A.E., Smit, S.K.: Parameter tuning for configuring and analyzing evolutionary algo-
rithms. Swarm Evol. Comput. 1(1), 19–31 (2011)
6. Passino, K.M.: Bactorial foraging optimization. Int. J. Swarm Intell. Res. 1(1), 1–16 (2010)
7. Copeland, B.J.: The Essential Turing. Oxford University Press, Oxford (2004)
8. Holland, J.: Adaptation in Natural and Artificial Systems. University of Michigan Press, Ann
Anbor (1975)
9. Fogel, L.J., Owens, A.J., Walsh, M.J.: Artificial Intelligence Through Simulated Evolution.
Wiley, New York (1966)
10. Judea, P.: Heuristics. Addison-Wesley, New York (1984)
11. Schrijver, A.: On the history of combinatorial optimization (till 1960). In: Aardal, K.,
Nemhauser, G.L., Weismantel, R. (eds.) Handbook of Discrete Optimization, pp. 1–68. Else-
vier, Amsterdam (2005)
12. Turing, A.M.: Intelligent Machinery. National Physical Laboratory, Technical report (1948)
13. Vapnik, V.: Nat. Stat. Learn. Theory. Springer, New York (1995)
Another random document with
no related content on Scribd:
Disgusted at his failure and the years he had wasted, Columbus
with his son Diego turned his back on Santa Fé. His journey took him
near the little seaside town of Palos, where at the Franciscan convent
of La Rabida he sought food and shelter for the night. Its prior, Fra
Juan Perez, once the Queen’s confessor, was delighted to have first-
hand news from the seat of war, and eagerly welcomed his guest;
with the result that all Christopher’s disappointed hopes came
pouring out in a stream of eloquence that soon made a convert of his
listener.
A secret letter from the prior to the Queen, full of respectful
expostulations, her quick response that Columbus should return at
once to Court, her gift of 20,000 maravedis to provide him with
suitable clothing and a mule,—and Juan Perez could write with
fervent joy:
Our Lord has listened to the prayers of His servant. The wise and virtuous
Isabel, touched by the grace of Heaven, gave a favourable hearing to the words of
this poor monk. All has turned out well.
“All has turned out well!” Face to face, Queen and would-be-
discoverer could realize how much their minds were in tune; even
more now than in the early days of his project; for, to the material
benefits he hoped to reap, Columbus, inspired perhaps by the
crusading character of the Moorish war, had added the burning
desire to carry the light of the Catholic Faith across “the sea of
darkness.” This was no mere pose. Religion to the sailor as to the
Queen was an intrinsic part of daily life, something vital and
overshadowing that in the hour of triumph intensified glory, in days
of depression or danger spread protecting wings. In the foreword of
his journal addressed to the sovereigns, he shows very clearly that he
regarded himself not only as pioneer but missionary:
Your Highnesses, as Catholic sovereigns and princes, loving the Holy Christian
Faith and the spreading of it, and enemies of the sect of Mahomet and of all
idolatries and heresies, decided to send me, Christopher Columbus, to the said
regions of India, to see the said princes and peoples and lands, and learn of their
disposition and of everything, and of the measures that could be taken for their
conversion to our Holy Faith.
Behind and beyond “the spreading of the Catholic Faith” in the far
East was another design of still bolder conception, the employment
of the wealth to be found in Cathay and the territories of the great
Khan towards the recovery of the Holy Sepulchre. The latter was the
crowning enthusiasm of every earnest Christian in mediæval times,
and Christopher believing himself “inspired, elected, foreordained,”
held amongst his cherished visions the glory of a final crusade, to
which he should have contributed the war fund.
Upheld by his inborn sense of power, he had returned to Court far
more a conqueror, ready to grant conditions, than a petitioner oft-
refused and eager to snatch the least morsel of favour. The Crown in
its clemency was now, after its long apathy, willing to confer on him
titles and privileges;—all in moderation of course, for Ferdinand and
Isabel were never unnecessarily lavish; but Christopher, valuing
himself and his task by the measure of his faith in the future, laughed
at their moderation. Either he was great enough to succeed and thus
prove worthy of a great reward, or he would fail and his pretensions
and demands fade away with his dreams. The sovereigns, skilled in
striking bargains, might argue and cajole. The Genoese, though his
fate trembled in the balance, never wavered, until at last in April,
1492, caution yielded to greatness, and the terms that he demanded
were signed and sealed.
Columbus and his heirs were to have the hereditary title of
Admiral of all the islands and continents that he might discover, and
should for ever hold the office of Viceroy and Governor-General over
them. He and his heirs should receive one-tenth of all the wealth,
whether metals, jewels, or spices, that should be acquired from these
territories; and he and they should have a perpetual right of
providing one-eighth of the expenses of every expedition sent to the
West, receiving a corresponding profit from the results. These with
extensive judicial and administrative privileges formed the basis of
the document, in return for which Columbus promised to sail into
the unknown and claim it in the name of Castile and her sovereign.
The actual cost of the expedition was, in comparison to the stakes
at issue, trifling; in all less than a thousand pounds of English
money, of which the Crown contributed some £850, Columbus
himself the rest. Three ships formed his fleet; two provided under
compulsion by the town of Palos as punishment for some public
offence, and as reluctantly manned by its inhabitants who looked on
the proposed voyage with horror. Columbus’s own flagship, the
Santa Maria, was a vessel of some hundred tons burden, by modern
standards ill-fitted for aught but coasting work; while the Pinta and
the Niña, commanded by Martin Alonso Pinzon and his brother
Vicente Yañez, noted navigators of the neighbourhood, were mere
merchant “caravels” of half its size.
The story of this first voyage to the New World has been often told:
the distrust and grumbling of the crew which, beginning before they
left Palos on that morning of the 3d of August, 1492, grew ever in
volume as they journeyed westwards, leaving the friendly Azores far
in their rear; the complaints that the wind steadily driving from the
east would never change and thus make any hope of return
impossible; the extraordinary variations of the compass and the
expanse of sea traversed, far in excess of the Admiral’s calculations,
so that, puzzled and anxious at heart himself, he must yet keep a
cheerful face and, lying skilfully, hold panic at bay by scientific
falsehoods and carefully doctored charts. The many cries of “land!
land!” heralding nought save clouds lying low on the horizon; the
ever-doomed hopes aroused by birds and floating grass; and then the
Sargasso Sea with its leagues of golden gulf-weed lapping against the
ship’s side. Was this the impassable ocean where Atlantis had sunk
to rest? Were they indeed destined to die here for their folly?
Then, when patience and hope were alike exhausted, and only the
Admiral’s faith rose triumphant above the general pessimism,
unmistakable signs of land appeared at last; and, on the 12th of
October the Spanish squadron came to anchor before the little island
of Guanahani, one of the Bahamas.
The details of the landing, the astonishment of the natives, “naked
as when their mothers gave them birth,” at the sight of mail- and
silk-clad warriors and the sound of cannons; the account of various
expeditions made to other islands and of the fort built in Española;—
these like the actual voyages may be read at length in the pages of
Washington Irving. It is with the triumphant home-coming of the
hero not with his adventures that we are here concerned.
Attention you two most wise and venerable men and hear of a new discovery
[wrote Peter Martyr to the Archbishop of Granada and Count of Tendilla]. You
remember Columbus the Ligurian, who persisted, when in the camps with the
sovereigns, that one could pass over by way of the Western Antipodes to a new
hemisphere of the globe.... He is returned safe and declares he has found
wonderful things.
Wonderful things indeed! Brown-skinned Indians, green and

scarlet parrots, golden nuggets and ornaments, cotton fibre and
strange roots and seeds; these that he brought with him were but
proofs and trophies of the still more wonderful adventures he
hastened to relate before the sovereigns and their Court. In his
disembarkation at Palos on the 15th of March, 1493, and still more in
his “solemn and very beautiful reception” by the sovereigns at
Barcelona graphically described by the historian Las Casas, he was to
reap at last the meed of honour and enthusiasm so long denied him.
Kneeling before his King and Queen to kiss the hands that afterwards
raised him in gracious condescension to sit with royalty upon the
dais, flattered and fêted by courtiers who had before patronized or
mocked, riding through the crowded streets by Ferdinand’s side
amid cries of admiration and applause;—in these moments he
reached the climax of worldly glory.
Long years stretched before him, when circumstances, his own
failings, and the envy and spite of others, were to rob him of ease,
popularity, and even royal confidence; but for the time being he was
“Don,” “Admiral,” the honoured of Kings, the most discussed and
admired man in Spain, perhaps in all Europe. In one country at least,
the kingdom of Portugal, the result of his voyage was a subject for
poignant regret; and Ferdinand and Isabel, having obtained from
Alexander VI. papal recognition of their right to the newly discovered
territories, were driven to demand a series of bulls, that would
provide them with a definition of their empire, lest Portuguese rivals,
too slow to forestall them in the discovery, should now rob them of
their gains.
By a bull of May 4, 1493, an imaginary line was drawn through the
north and south poles, cutting the Atlantic at one hundred leagues
distance from the Cape Verde Islands and Azores. To the east of this
line was henceforth to stretch the zone of Portuguese dominion, to
the west that of Castile. Later, by the Treaty of Tordesillas, signed by
the Spanish sovereigns and King John in June, 1494, the boundary
was fixed at three hundred and seventy instead of one hundred
leagues distance; and there for the moment national rivalry was
checked.
In the meanwhile Columbus, having organized a second
expedition, had on September 23, 1493, set sail once more for the
west. Very different in size and character was his new fleet from the
former vessels of Palos with their pressed crews; for more than twice
the number of men required for his fourteen caravels had applied for
leave to sail with him, and not a few of those refused had chosen to
embark as stowaways rather than be left behind. It was a case of
unbalanced enthusiasm succeeding to unbalanced hostility, and, as
often happens, the second state was to prove more dangerous than
the first.
Not patriotism, nor a healthy love of adventure, nor even a cool-
headed trading instinct, animated the majority of that idle,
quarrelsome throng who were destined to turn the lands their
discoverer at first believed the “Earthly Paradise” into a hell of
human misery and wrong. It was lust of gold, no hardly-won reward
of toil and sweat, but the fabulous wealth of Cipango and Cathay, to
be picked in nuggets out of the flowing river, found in the turned
surface of the earth, wrung by brutality if necessary from unwilling
natives, that brought a wastrel nobility disgusted with orderly
government at home, to serve under the standard of a man whom
they secretly despised as an ill-bred foreigner. Not all were of this
type. Amid the fourteen crews were some earnest souls, inspired like
their Admiral by a sense of responsibility; but the prevailing element
was selfish, vicious, and insubordinate.
For this Columbus himself was partly to blame. Blinded to the
limits of his achievement by his faith in the glory and wealth yet to
come, and, anxious at all costs to maintain the support of the
Spanish sovereigns, his eloquence had painted a highly-coloured
picture very likely to deceive those who listened. The small quantity
of gold so far obtained was merged in the glittering accounts given by
natives of kingdoms to the south, where precious metals were to be
had for the asking. These, like Amazon islands and lands whose
tribes had tails, proved ever beyond the distant horizon, vanishing at
the Spanish approach. As they melted into thin air so also did
Christopher’s inflated reputation; and those who had looked on him
as a kind of magician, able to conjure up vast quantities of gold, saw
him instead only as a lying adventurer, who had lured them from
civilization and luxury on a false plea.
“Why hast thou taken us out into the wilderness to die?” It is the
cry that from the time of Moses onwards has assailed the ears of the
pioneer enthusiast. The wilderness may prove a paradise; but in that
it falls short of human desires it will be condemned and despised.
Not all the glory of sunshine and colour, of rich soil, luxurious
vegetation, and flowing river, speaking to honest toilers of a possible
kingdom of God on earth, can compensate with an idle rabble for
shattered dreams of gold mines, of jewels, and of spices.
Murmurings, complaints, secret disobedience, open defiance:
these were the fruits of Columbus’s autocracy. When he landed for
the second time in Española, he found the fort which he had left
well-stored with provisions and ammunition burnt to the ground, its
garrison dead, the Indians, once his trusted allies, fleeing before him
afraid into the woods. Inquiry elicited an all too circumstantial tale of
Spanish profligacy, cruelty, and carelessness, once his governing
hand had been removed. Then had come retribution in the form of
an avenging massacre by a warlike tribe from the interior of the
island. The Indians of the coast denied their participation, even
swore on oath that they had helped the garrison to the best of their
ability; and Columbus, anxious to believe them, tried to restore the
old relations. Mutual suspicion, however, had come to reign. His
followers, angry at the fate of their countrymen, accepted it as a
legitimate excuse for intimidating and oppressing all natives. The
hospitality and gifts once so generously lavished were now withheld
or, proving totally inadequate to meet ever-growing Spanish
necessities, were replaced by an enforced tribute, until the link of
willing service was forged into an iron chain of bondage.
Some form of submission of native to European, of the weaker
many to the stronger and more civilized few, was an inevitable
solution of the racial problem. That it developed into absolute slavery
was due, partly to the custom of the day, partly to the difficulties in
which Columbus and his colonists soon found themselves involved.
They had laid the foundations of the system in the New World when
they carried off their first ten Indians in triumph to parade them
through the streets of Barcelona, though the individuals in question
could boast of generous treatment and a baptism with royal
sponsors.
The principle of personal liberty abandoned, Columbus could
declare, not without truth, that as slaves the natives would have a
better chance of learning the doctrines of the Catholic Faith than in
their own wild freedom. Even on the grounds of mercy and good
government he could at first justify his attitude; since he and his
followers contented themselves for the most part with seizing
“Caribs,” a fierce cannibal tribe that preyed upon their weaker
neighbours.
Among the people who are not cannibals [he wrote home] we shall gain great
credit by their seeing that we can seize and take captive those from whom they are
accustomed to receive injuries, and of whom they are in such terror that they are
frightened by one man alone.
Alas for either pious or kindly intentions! Not these but economic
considerations were really to sink the scales. Columbus had
promised to find precious metals in abundance, and yet seven years
after his discovery Bernaldez, the Curate of Los Palacios, made a note
that the expenses of the various expeditions still continued to exceed
the profits.
“Since everything passed through the Admiral’s hands,” he adds,
“there was much murmuring against him, and he made greater
hindrances and delays than he ought in sending back gold to the
King.”
Gold there was little in these early years of exploration; and
demands for precious metals at home were echoed by demands in his
own colony for horses, cattle, and sheep to stock the new settlement.
In this dilemma the Admiral fell back on the wealth of human life,
for which he could reap a handsome profit in the labour-markets of
the Old World besides pacifying some of the grumbling in the New. It
was no longer the conversion of the heathen nor the civilization of
cannibals, that took the first place in his thoughts, but a momentary
respite from increasing financial strain.
A gift of an Indian apiece to each of his greedy crew; a gang of
some five hundred captives of either sex shipped to Europe, huddled
together “with no more care taken of them than of animals destined
for the slaughter-house.”
These, or tales of a like nature, came to the Queen’s ears. “By what
right does the Admiral give away my vassals?” she demanded
indignantly, and ordered the Indians to be released and re-shipped
to their own land.
It must be remembered to her credit [says Filson Young, referring to her attitude
towards this question,] that in after years, when slavery and an intolerable bloody
and brutish oppression had turned the Paradise of Española into a shambles, she
fought almost single-handed and with an ethical sense far in advance of her day
against the system of slavery practiced in Spain upon the inhabitants of the New
World.
Ferdinand cared little for the sufferings of Indians, but their sale
would not bring him the profits he had been led to expect from his
new dominions, and he was therefore more than willing to listen to
the many complaints of tyranny, favouritism, and deceit, brought
against the Governor by those returning from the West. Here the
crowning offence had been in reality the employment of all able-
bodied Europeans, priests as well as laymen, in the construction of a
city in Española to which Columbus gave the name of “Isabella,” “in
remembrance,” says Las Casas, “of the Queen Doña Isabel whom he
above all held in great reverence; and he was more desirous of
serving and pleasing her than any other person in the world.”
“Columbus,” wrote Peter Martyr, “has begun the building of a city
and the planting of our seeds and the raising of cattle.” His words call
up a picture of peaceful and slow-rewarded toil, little to the taste of
the majority pressed to take their share, their natural dislike of
manual labour stimulated by the ennervating climate and habits of
self-indulgence. The crops grew apace, but so also did fever and
disease; and for all that went wrong the people held their foreign
Admiral responsible.
Indeed there was often sufficient foundation to make the reports
brought home plausible. Columbus was a born leader of men in
action, where a strong personality will always dominate; but he had
few gifts as a governor, and least of all that invaluable instinct for
selecting trustworthy subordinates. His choice of officials was often
betrayed; his government, as a rule too kindly towards the cut-throat
ruffians he commanded, on occasions varied by excessive severity.
Whatever its quality he reaped odium, not only amongst the
colonists, but with their relations and friends in Castile.
Enough was obviously at fault to require inspection; and in 1500,
when Columbus who had sailed from Spain on a third voyage in 1498
was occupied in exploring fresh islands, Francisco de Bobadilla, an
official of the royal household, arrived in Española, charged with the
duty of inquiring into the Admiral’s conduct. His high-handed
action, in immediately arresting Columbus and his brothers
Bartholomew and Diego on their return to headquarters, is one of
the most dramatic episodes in history; and its appeal was felt
throughout the length and breadth of Spain.
Villejo, the officer in command of the prisoners on the voyage
home, offered to remove the fetters in which they had been sent on
board, but Columbus sternly refused. He would wear them, he
declared, until he knelt before his sovereigns, keep them by him till
his dying day. Crippled by gout, his hair whitened by care, he
disembarked at Cadiz, the irons clanking on his wrists and ankles;
and at the sight horror and shame spread from cottage and shop to
castle and palace. Was this the discoverer’s reward for a New World?
“Be assured that your imprisonment weighed heavily upon us,”
wrote the sovereigns some years later, still mindful of the shock the
news had given them; and when Columbus knelt before his Queen
the sobs of pent-up bitterness with which he recounted his troubles
awoke answering tears of regret and understanding in her eyes.
“After they had listened to him,” says Oviedo, “they consoled him
with much kindliness and spake such words that he remained
somewhat comforted.”
Confidence was temporarily restored, but the Admiral’s hour of
glory and triumph had passed never to return. His bad treatment
was acknowledged, but so also was his bad government; for though
he might not have deliberately tyrannized and deceived, yet he had
failed to keep order or fulfil his promises. The Queen was growing
old, and, broken by ill-health and private griefs, took less share than
she was wont in public business. Ferdinand had never liked the
Genoese sailor; moreover he was no longer necessary to royal
schemes and that to the astute King was ever sufficient excuse for
discarding a tool.
Columbus sailed for the fourth time to the lands of his discovery in
1502; but it was to find that Nicholas de Ovando, another royal
protégé, had succeeded Bobadilla in command at Española, while
treachery and ill-luck dogged his own efforts. Bitterness and
suspicion had begun to eat like a canker in his mind, and his letters
are full of querulous reproaches that the bargain he had made was
ill-kept and his due share of the commercial profits denied him. In
1504, he returned home suffering in body and spirit, but no longer to
meet with the sympathy for which he craved. Three weeks after he
arrived at Seville Isabel died and her will, that contained a special
petition for the kindly treatment of the natives, made no mention of
his name.
Writing to his son Diego the Admiral says:
The principal thing is affectionately and with great devotion to commend the
soul of the Queen, Our Lady, to God. Her life was always catholic and holy and
ready for all things of His holy service, and for this reason it may be believed that
she is in His holy glory and beyond the desires of this rough and wearisome world.
In these words lie the confession of his own disillusionment. His

world, once so fair a place of material visions and dreams, had
proved in its essence wearisome; and, clad in the Franciscan habit of
renunciation, he himself, on the 20th of May, 1506, passed
thankfully into the rest of God’s “holy glory.”
“His life,” says Filson Young, “flickered out in the completest
obscurity.” No Peter Martyr eulogized his memory in letters to his
courtly patrons. No grateful country of adoption bestowed on him a
gorgeous funeral. Even the lands he had discovered were destined to
receive their name from another, the Florentine sailor Amerigo
Vespucci, whom he himself had helped on the road to fame.
Posterity is the audience that can alone judge truly the drama of
history, and in the thunder of its applause Columbus has long come
to his own.
“The world,” says Thacher, “did not observe his final exit from the
stage. Yet he was a great character, one of the greatest ever passing
before the eyes of men.”
CHAPTER XI
ISABEL AND HER CHILDREN
If it is true that the trappings of the monk often conceal the wearer’s
individuality, it might be added that so also do royal robes. The
contemporary historian is apt to portray his King or Queen garbed in
a cloak of politics, morality, or pageantry, according to his special
enthusiasm; and, unless to his task he brings also the biographer’s
instinct for personality, his likeness though regal and exemplary will
leave the spectator cold. He has forgotten that the abiding measure
of our interest in others is the very humanity he has neglected or
tried to excel.
In the case of “Isabel of Castile” the conventional atmosphere of a
Court is intensified by her own determination to play a royal part.
She rarely forgot that she was Queen. On one occasion the Admiral of
Castile, Ferdinand’s uncle, had ventured to address the King as
nephew; whereupon she, overhearing, reproved him sharply.
“My Lord, the King has no kindred or friends but servants or
subjects!” A petty snub! Unless in judging it we recall the Court of
Henry IV., where Isabel had seen her brother mocked and bullied by
insolent nobles, amongst them a former Admiral of Castile.
Her lifework in building up the reputation of the monarchy must
be carried out in detail as well as on the broad lines of governmental
reform, and the dignity and magnificence of royalty formed part of
the scheme, that aimed at the exaltation of the Crown, not only in the
eyes of Europe but still more of Spain itself. The Castilian grandee
might be losing his official status, the Admiral be no longer essential
to the Fleet, the Constable to the Army, the Duke or Marquis to the
Royal Council; but in the throne-room and ante-chambers of the
palace etiquette more and more demanded their presence. Silken
chains were binding the unruly in a peaceful servitude.
Pulgar, the historian, commenting on Isabel’s insistence on the
outward forms of state, declared that “it pleased her to be served by
grandees and nobles,” while in another place he mentions her retinue
of the daughters of great families “such that we do not read in the
Chronicles that any Queen had before her.”
A household maintained on this scale and with corresponding
luxury was a costly item in royal expenses and, considering the
chronic deficiencies of the Treasury, was perhaps excessive. Yet
Ferdinand and Isabel were both by nature simple and abstemious in
their tastes, and wont in other matters, as we have seen in the case of
Columbus, to err rather on the side of economy than extravagance.
“A King must outshine his subjects,” says Pulgar with a conviction
born of his intimate knowledge of Spanish character. The easy
familiarity of the Emperor Maximilian, “Max the Penniless,” and his
son might be appreciated in Germany and Flanders; the private thrift
of a Louis XI., or lack of ostentation of a Lorenzo de Medici respected
in France or Florence; but the Castilian nature demanded
magnificence and aloofness in its rulers. Even a Ximenes de Cisneros
had been unable to shake off the outward glory of his office when he
accepted the Archbishopric of Toledo; and Ferdinand and Isabel,
children of their race, were fully alive to the appeal of surroundings
suitable to their rank.
Of the impression made by their magnificence on foreigners we
can gather from the diary of a certain Roger Machado who, in the
capacity of king-at-arms, accompanied an English embassy to the
Spanish Court at Medina del Campo in 1488. “People speak,” he says,
“of the honour done to Ambassadors in England; but it is not to be
compared to the honour which is done to Ambassadors in the
kingdom of Castile.”
The torchlight procession that accompanied them from their
lodgings to their evening reception at the palace; the majesty and
condescension of the sovereigns; the speeches, dances, bull-fights,
tourneys; each in turn arouses his admiration; but it is in his account
of the costumes and jewellery that his diary really reaches its apogee
of enthusiasm. The King is “dressed in a rich robe of cloth-of-gold,
woven entirely of gold, and furred with a rich trimming of fine sable.”
The Queen has “a rich robe of the same woven cloth-of-gold ... and
over the said robe a riding-hood of black velvet, all slashed in large
holes so as to show under the said velvet the cloth-of-gold in which
she is dressed.” She wears “crosswise over her left side ... a short
cloak of fine crimson satin furred with ermine, very handsome in
appearance and very brilliant.”
ISABEL OF CASTILE
CARVED WOODEN STATUE FROM

THE CATHEDRAL AT GRANADA
FROM “A QUEEN OF QUEENS” BY

CHRISTOPHER HARE, PUBLISHED BY
MESSRS. HARPER
Roger Machado, had he lived to-day, would surely have made his
fortune as journalist of some fashion-weekly; but even his facile pen
finds it difficult to express adequately the splendour of the Queen’s
jewellery,—her necklace of gold and jewelled roses,—the ribbon at
her breast adorned with diamonds, rubies, and pearls,—the pouch of
her white leather girdle set “with a large balass ruby the size of a
tennis-ball between five rich diamonds and other stones the size of a
bean.”
“Truly, as I believe,” he comments, “and also as I heard it said at
the time, I estimate the dress she then wore at the value of 200,000
crowns of gold”; while on another occasion he declares her dress so
rich that “there is no man who can well imagine what could be the
value of it.”
At a somewhat similar reception of a French embassy, tales of the
Queen’s magnificence evidently spread through Spain; and Fra
Fernando de Talavera, Archbishop of Granada, though no longer her
official Father-Confessor, felt bound to write and remonstrate. Isabel
was, however, able to offer a good defence, declaring that neither the
dresses of herself nor her ladies had been new,—indeed her own
“made of silk and with three bands of gold as plainly as possible,” she
had worn before in Aragon in the presence of these same Frenchmen.
If some of the men’s garments were costly, it had not been by her
orders, rather she had done her uttermost to discountenance it.
She might have mentioned also how, in the critical stages of the
Moorish war, she had pledged the crown jewels to merchants of
Barcelona, thus showing that for all her appreciation of the luxuries
of dress, they did not rank for a second in her thoughts with more
important considerations.
Her magnificence like her severity was calculated and the same
might be said of her liberality. She had seen money wasted on ne’er-
do-wells and was fully determined that no man, merely because he
was powerful or plausible, should prey on her revenues; while for the
regular type of Court-flatterer hunting for sinecures her contempt
amounted to aversion. Galindez Carvajal tells us in his chronicle that
she and Ferdinand kept a book in which they wrote down the names
of those at Court whom they thought most capable and worthy of
reward, consulting it whenever an office fell vacant, and that they did
not hesitate to prefer prudent men of the middle-class to the highly-
born incompetent. Their actual gifts, though scarcely lavish, were
sufficient to cause satisfaction, and Lucio Marineo declares that
“when between the King and Queen there was discussion as to the
fitting reward of any particular service, she on her part always gave
more than the sum on which the two had determined.”
Ferdinand was not unlike his contemporary, Henry VII. of
England, according to Ayala, the Spanish Ambassador’s description
of that monarch: “If gold once enters his strong boxes it never comes
out again. He pays in depreciated coin.”
Against this criticism may be set Machiavelli’s praise: “If the
present King of Spain had desired to be thought liberal, he would not
have been able to contrive, nor would he have succeeded in so many
undertakings.”
It is pleasant to turn from calculated policy to uncalculated
enthusiasm; and this may be truly said of Isabel’s love of her faith.
Both she and the King were strict in the outward observances of
catholicism, and every morning would find Ferdinand at Mass before
he broke his fast, while we are told that on Maundy Thursday his
servants would seek out twelve of the poorest of his subjects and that
he would serve them at supper and wash their feet. Isabel herself
would recite the hours every day like a priest; and, for all the whirl of
ceremonies and duties in which she found herself involved, she
would make time for special devotions so that it seemed to those
about her that her life was “contemplative rather than active.”
Her marked individuality, and the respect she inspired in
Ferdinand, had completely changed the character of the Court from
the old licentious days of Henry IV.; priests of the type of Ximenes
and Fra Fernando de Talavera thronged her ante-chambers; and
courtiers, when they saw her coming, would walk with eyes cast
down in the hope of establishing a reputation for sanctity.
Their hypocrisy can have brought them little. The Queen might be
a saint in her private life; but those who think saints necessarily fools
stand convicted of folly themselves. She was too shrewd a judge of
character to desire to change her Court into a convent, and her
letters to Fra Fernando de Talavera, while breathing affection and
admiration yet venture occasionally to question the suitability to
herself and her surroundings of his standard of asceticism.
It is my wish that not only in matters of importance but in all that concern these
kingdoms you should give me your advice; ... and this I do most earnestly beg, that
you will not cease from writing your opinion on the ground that these things do not
concern you since you are no longer here; for well I know that although absent
your counsel will be worth more to me than that of another present.
She then goes on to thank him for the reproofs he had
administered on the score of the too-great gaiety at Court and to
assure him that in explaining certain matters she is not seeking to
free herself from blame.
As for the French people supping with the ladies at table, that is a thing they are
accustomed to do. They do not get the custom from us; but, when their great guests
dine with sovereigns, the others in their train dine at tables in the hall with the
ladies and gentlemen; and there are no separate tables for ladies. The
Burgundians, the English, and the Portuguese also follow this custom; and we on
similar occasions to this.... I say this that you may see there was no innovation in
what we did, nor did we think we were doing anything wrong in it.... But if it be
found wrong after the inquiry I will make, it will be better to discontinue it in
future.... As for the bull-fights, I feel with you, though perhaps not quite so
strongly. But after I had consented to them, I had the fullest determination never
to attend them again in my life nor to be where they were held.
One of Queen Isabel’s biographers on the contrary tells us that the

Queen admired this national pastime for the skill and courage it
demanded, a statement it is difficult to reconcile with the avowed
distaste in her own letter. Perhaps her enthusiasm was evoked after
the adoption of her device to place false horns, turned points
inwards, on the horns of the bull, that the frequent loss of human life
might be prevented. It was hardly a suggestion to win her popularity
with her subjects, whose enjoyment of a spectacle was always
proportionate to its risks. Isabel herself did not lack the true sporting
instinct, for the chroniclers record a bear-hunt in the woods near
Madrid, where one of the most ferocious of the beasts fell a trophy to
her javelin.
Courage, the natural heritage of her race, her will and pride
exalted almost to a fetish; and Pulgar tells us that “even in the hour
of childbirth she disguised her sufferings and forced herself neither
to show nor utter the pain that in that hour women are wont to feel
and manifest.”
Her reserve was deep, in all that concerned her innermost
thoughts almost like a curtain veiling some sanctuary, that she felt
would be profaned by other eyes, but now and then torn back for the
moment by the stress of some sudden emotion. Her agony of mind
was obvious after the attempted murder of Ferdinand in Barcelona
in August, 1492. The assassin, a madman who believed the King’s
death would result in his own accession to the throne, had hurled
himself on his victim from behind, as he was descending the palace
stairway, inflicting a deep wound in his neck. This, though not fatal,
was aggravated by fever, and for many days the King’s life hung in
the balance.
And on the seventh day [wrote the Queen to Fra Fernando] the fever reached its
climax, so that we were then in fear greater than all that through which we had
previously passed; and this lasted a day and a night of which I will not say that
which Saint Gregory said in the Office for Holy Saturday, more than that it was a
night of hell; so that you may believe, Father, never was the like seen amongst the
people at any time, for officials ceased their work, and none paused to speak with
another. All was pilgrimages, processions, and almsgiving, and more hearing of
confessions than ever in Holy Week.
Ferdinand was popular in Barcelona, and the Council of Justice

there condemned his assassin to a death of ghastly torment, of which
tearing the flesh with red-hot pincers formed but a part. One is
thankful that Isabel issued a special command ordering the man to
be beheaded before this barbarous sentence was enacted.
Her love for Ferdinand was the strongest of her personal
affections, growing rather than diminishing as the years passed, so
that, dying, she sought that she should not be parted from him for
long.
Let my body be interred in the monastery of San Francisco, which is in the

Alhambra of the city of Granada, ... but I desire and command that, if the King, My
Lord, should choose a sepulchre in any church or monastery in any other part or
place of these my kingdoms, my body be translated thither and buried beside the
body of His Highness.
Her wish is fulfilled, and the Catholic sovereigns lie side by side in
the Royal Chapel of Granada; but the love she gave in such
ungrudging measure was never fully returned. Isabel was fair in her
youth, not beautiful perhaps, but graceful and dignified, with soft
chestnut-coloured hair, and blue-green eyes that looked out candidly
upon the world; and to Ferdinand, arriving in disguise at Valladolid,
his blood set on fire by romance and excitement, she had seemed a
bride very worthy of his chivalrous care. Later he learned to respect
and admire her both as his wife and Queen, to love her even after his
fashion; but he was temperamentally cold and self-centred, and the
age set no high standard of fidelity. The chronicles record that he had
four illegitimate children by different mothers, of whom one,
Alfonso, became Archbishop of Saragossa; and Isabel was destined to
suffer bitterly from a jealousy intensified by her pride and strength of
will.
TOMB OF FERDINAND AND ISABEL
FROM NERVO’S “ISABELLE LA

CATHOLIQUE”
Her private life was not, however, unhappy, at least in those years
when her own children were growing up around her, and she could
find time amid the many cares of state to superintend their education
and build dream kingdoms round their future. Her ambitions and
Ferdinand’s were alike centred on their only son, Prince John, whose
birth in Seville on June 30, 1478, we have mentioned in an earlier
chapter.
“My angel” Isabel would playfully call the boy, alluding to his fair
skin and halo of curls; and she spared no pains in moulding his
character that he might one day satisfy her ideal of kingship. The
retinue that attended the little Prince of Asturias was in miniature a
counterpart of the elaborate household of officials and servants that
surrounded his father and mother; and, while from this environment
he imbibed a sense of the grandeur and aloofness of his position, he
also learned early the lesson of regal responsibility.
As president of a miniature Council of State, he listened to
frequent discussions of the economic and political problems of the
day by men chosen for their ability and experience; but it must not
be imagined that such strong diet was alone provided for his mental
digestion. Youth cries out for the companionship of youth; and
Isabel, recognizing the wisdom of this decree of nature, established a
class of ten boys, five older and five of his own age, against whose
wits the heir to the throne might sharpen his intellect in healthy
competition.
His love of music, inheritance from his grandfather, John II. of
Castile, was encouraged and developed; and often in the evenings the
choir boys of the Royal Chapel would assemble in his room, and he
and they sing together; or on other occasions he would summon his
musicians and play on the organ, or on one of the stringed
instruments of the day. Musical proficiency was a sure road to his
interest and regard.
In his position as heir to the Spanish dominions, it was natural
that Prince John’s life should stand more in the limelight of publicity
than his sisters’: but their education was in fact scarcely less
considered and planned than his. The Queen had always possessed
an intense admiration for classical learning; and it was one of
Ferdinand’s regrets that civil war had called him from the
schoolroom to the camp, when he could do little more than read and
write. He never understood Latin, the common language of cultured
Europe; but Isabel made time to study its grammar and composition
with Beatriz de Galindo, a famous teacher of her own sex, on whom
the Court had bestowed the appropriate nickname “La Latina.”
This course of education the Queen pursued with her usual
thoroughness and determination; and, if she did not achieve the true
scholar’s facility in translation and speech, she was at any rate able to
understand the orations of foreign ambassadors, and to interpret to
her husband the letters of the young Italian diplomat, Peter Martyr,
who took so lively an interest in her student’s career.

PDF Nature Inspired Computation in Engineering 1St Edition Xin She Yang Eds Ebook Full Chapter

Uploaded by

Copyright:

Available Formats

You might also like

PDF Nature Inspired Computation in Engineering 1St Edition Xin She Yang Eds Ebook Full Chapter

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PDF Nature Inspired Computation in Engineering 1St Edition Xin She Yang Eds Ebook Full Chapter

Uploaded by

Copyright:

Available Formats

Nature Inspired Computation in

Engineering 1st Edition Xin-She Yang

Nature-inspired Computation and Swarm Intelligence:

Nature-Inspired Algorithms and Applied Optimization 1st

Smart Trends in Systems Security and Sustainability

Brain and Nature-inspired Learning, Computation and

Discrete Problems in Nature Inspired Algorithms 1st

Nature Inspired Design of Hybrid Intelligent Systems

Nature Inspired Computing and Optimization Theory and

Advances in Nature and Biologically Inspired Computing

Xin-She Yang Editor

The series “Studies in Computational Intelligence” (SCI) publishes new develop-

More information about this series at http://www.springer.com/series/7092

ISSN 1860-949X ISSN 1860-9503 (electronic)

Library of Congress Control Number: 2016933797

© Springer International Publishing Switzerland 2016

Printed on acid-free paper

This Springer imprint is published by Springer Nature

Nature-inspired computation has become increasingly popular in engineering and is

London Xin-She Yang

Nature-Inspired Optimization Algorithms in Engineering:

A Hybridisation of Runner-Based and Seed-Based

Ali R. Al-Roomi Dalhousie University, Electrical and Computer Engineering,

Yang Liu School of Engineering, Cardiff University, Cardiff, UK

Xin-She Yang and Xingshi He

Abstract Nature-inspired computation has become popular in engineering

Optimization is everywhere and important in many applications such as engineer-

X.-S. Yang (B)

© Springer International Publishing Switzerland 2016 1

2 Optimization and Optimality

Many problems in engineering can be formulated as an optimization problem, though

minimize f i (x), (i = 1, 2, . . . , M), (1)

subject to h j (x) = 0, ( j = 1, 2, . . . , J ), (2)

the functions f i (x) where i = 1, 2, . . . , M are called the objective functions or

2.2 Search for Optimality

Once an optimization problem is formulated correctly, the main task is to find

3 A Brief History of Nature-Inspired Computation

Humans’ approach to problem-solving has always been heuristic or metaheuristic—

to problem solving is indeed a modern phenomenon, though it is difficult to pinpoint

4 Main Characteristics of Nature-Inspired Algorithms

The key characteristics of nature-inspired optimization algorithms can be analyzed

4.1 Exploration and Exploitation

4.2 Diversity and Adaptation

4.3 Key Operators in Algorithms

5 Some Swarm-Based Algorithms

5.1 Particle Swarm Optimization

Particle swarm optimization (PSO), developed by Kennedy and Eberhart in 1995

vit+1 = vit + α1 [g ∗ − xit ] + β2 [xi∗ − xit ], (5)

xit+1 = xit + vit+1 . (6)

5.2 Firefly Algorithm

xit+1 = xit + β0 e−γri j (x tj − xit ) + α it ,

5.3 Cuckoo Search

xit+1 = xit + αs ⊗ H ( pa − ) ⊗ (x tj − xkt ), (8)

xit+1 = xit + αL(s, λ), (9)

combination of self-similar, fractal-like and multiscale search capabilities [1]. By

5.4 Bat Algorithm

f i = f min + ( f max − f min )β, (11)

vit = vit−1 + (xit−1 − x∗ ) f i , (12)