Download Swarm Evolutionary and Memetic Computing 6th International Conference SEMCCO 2015 Hyderabad India December 18 19 2015 Revised Selected Papers 1st Edition Bijaya Ketan Panigrahi ebook All Chapters PDF

Download the Full Version of textbook for Fast Typing at textbookfull.
com
Swarm Evolutionary and Memetic Computing 6th

International Conference SEMCCO 2015 Hyderabad
India December 18 19 2015 Revised Selected Papers
1st Edition Bijaya Ketan Panigrahi
https://textbookfull.com/product/swarm-evolutionary-and-
memetic-computing-6th-international-conference-
semcco-2015-hyderabad-india-december-18-19-2015-revised-
selected-papers-1st-edition-bijaya-ketan-panigrahi/
OR CLICK BUTTON
DOWNLOAD NOW
Download More textbook Instantly Today - Get Yours Now at textbookfull.com

Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.
Swarm Evolutionary and Memetic Computing 4th International

Conference SEMCCO 2013 Chennai India December 19 21 2013
Proceedings Part I 1st Edition M. Fatih Tasgetiren
https://textbookfull.com/product/swarm-evolutionary-and-memetic-
computing-4th-international-conference-semcco-2013-chennai-india-
december-19-21-2013-proceedings-part-i-1st-edition-m-fatih-tasgetiren/
textboxfull.com
Big Data Benchmarking 6th International Workshop WBDB 2015

Toronto ON Canada June 16 17 2015 and 7th International
Workshop WBDB 2015 New Delhi India December 14 15 2015
Revised Selected Papers 1st Edition Tilmann Rabl
https://textbookfull.com/product/big-data-benchmarking-6th-
international-workshop-wbdb-2015-toronto-on-canada-
june-16-17-2015-and-7th-international-workshop-wbdb-2015-new-delhi-
india-december-14-15-2015-revised-selected-papers-1st-ed/
textboxfull.com
Smart Health International Conference ICSH 2015 Phoenix AZ

USA November 17 18 2015 Revised Selected Papers 1st
Edition Xiaolong Zheng
https://textbookfull.com/product/smart-health-international-
conference-icsh-2015-phoenix-az-usa-november-17-18-2015-revised-
selected-papers-1st-edition-xiaolong-zheng/
textboxfull.com
Information and Communications Security 17th International

Conference ICICS 2015 Beijing China December 9 11 2015
Revised Selected Papers 1st Edition Sihan Qing
https://textbookfull.com/product/information-and-communications-
security-17th-international-conference-icics-2015-beijing-china-
december-9-11-2015-revised-selected-papers-1st-edition-sihan-qing/
textboxfull.com
Social Computing in Digital Education First International
Workshop SOCIALEDU 2015 Stanford CA USA August 19 2015
Revised Selected Papers 1st Edition Fernando Koch
https://textbookfull.com/product/social-computing-in-digital-
education-first-international-workshop-socialedu-2015-stanford-ca-usa-
august-19-2015-revised-selected-papers-1st-edition-fernando-koch/
textboxfull.com
Distributed Computer and Communication Networks 18th

International Conference DCCN 2015 Moscow Russia October
19 22 2015 Revised Selected Papers 1st Edition Vladimir
Vishnevsky
https://textbookfull.com/product/distributed-computer-and-
communication-networks-18th-international-conference-dccn-2015-moscow-
russia-october-19-22-2015-revised-selected-papers-1st-edition-
vladimir-vishnevsky/
textboxfull.com
Cloud Computing and Services Science 5th International

Conference CLOSER 2015 Lisbon Portugal May 20 22 2015
Revised Selected Papers 1st Edition Markus Helfert
https://textbookfull.com/product/cloud-computing-and-services-
science-5th-international-conference-closer-2015-lisbon-portugal-
may-20-22-2015-revised-selected-papers-1st-edition-markus-helfert/
textboxfull.com
High Performance Computing and Applications Third

International Conference HPCA 2015 Shanghai China July 26
30 2015 Revised Selected Papers 1st Edition Jiang Xie
https://textbookfull.com/product/high-performance-computing-and-
applications-third-international-conference-hpca-2015-shanghai-china-
july-26-30-2015-revised-selected-papers-1st-edition-jiang-xie/
textboxfull.com
Service Oriented Computing ICSOC 2015 Workshops WESOA

RMSOC ISC DISCO WESE BSCI FOR MOVES Goa India November 16
19 2015 Revised Selected Papers 1st Edition Alex Norta
https://textbookfull.com/product/service-oriented-computing-
icsoc-2015-workshops-wesoa-rmsoc-isc-disco-wese-bsci-for-moves-goa-
india-november-16-19-2015-revised-selected-papers-1st-edition-alex-
norta/
textboxfull.com
Bijaya Ketan Panigrahi
Ponnuthurai Nagaratnam Suganthan
Swagatam Das
Suresh Chandra Satapathy (Eds.)
LNCS 9873
Swarm, Evolutionary,
and Memetic Computing
6th International Conference, SEMCCO 2015
Hyderabad, India, December 18–19, 2015
Revised Selected Papers
123
Lecture Notes in Computer Science 9873
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen
Editorial Board
David Hutchison
Lancaster University, Lancaster, UK
Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA
Josef Kittler
University of Surrey, Guildford, UK
Jon M. Kleinberg
Cornell University, Ithaca, NY, USA
Friedemann Mattern
ETH Zurich, Zurich, Switzerland
John C. Mitchell
Stanford University, Stanford, CA, USA
Moni Naor
Weizmann Institute of Science, Rehovot, Israel
C. Pandu Rangan
Indian Institute of Technology, Madras, India
Bernhard Steffen
TU Dortmund University, Dortmund, Germany
Demetri Terzopoulos
University of California, Los Angeles, CA, USA
Doug Tygar
University of California, Berkeley, CA, USA
Gerhard Weikum
Max Planck Institute for Informatics, Saarbrücken, Germany
More information about this series at http://www.springer.com/series/7407
Bijaya Ketan Panigrahi
Ponnuthurai Nagaratnam Suganthan
Swagatam Das Suresh Chandra Satapathy (Eds.)
•
Swarm, Evolutionary,
and Memetic Computing
6th International Conference, SEMCCO 2015
Hyderabad, India, December 18–19, 2015
Revised Selected Papers
123
Editors
Bijaya Ketan Panigrahi Swagatam Das
IIT Indian Statistical Institute
New Dehli Kolkata
India India
Ponnuthurai Nagaratnam Suganthan Suresh Chandra Satapathy
Nanyang Technological University Department of Computer Science
Singapore Engineering
Singapore Anil Neerukonda Institute of Technology
and Sciences
Visakhapatnam
India
ISSN 0302-9743 ISSN 1611-3349 (electronic)

Lecture Notes in Computer Science
ISBN 978-3-319-48958-2 ISBN 978-3-319-48959-9 (eBook)
DOI 10.1007/978-3-319-48959-9
Library of Congress Control Number: 2016956619
LNCS Sublibrary: SL1 – Theoretical Computer Science and General Issues
© Springer International Publishing AG 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
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
This LNCS volume contains the papers presented at the 6th Swarm, Evolutionary and
Memetic Computing Conference (SEMCCO 2015) held during December 18–19, 2015,
at CMR Technical Campus, Hyderabad, India. SEMCCO is regarded as one of the
prestigious international conference series that aims at bringing together researchers
from academia and industry to report and review the latest progress in cutting-edge
research on swarm, evolutionary, memetic computing, and other novel computing
techniques like neural and fuzzy computing, to explore new application areas, to design
new bio-inspired algorithms for solving specific hard optimization problems, and finally
to raise awareness of these domains in a wider audience of practitioners.
SEMCCO 2015 received 150 paper submissions from 12 countries across the globe.
After a rigorous peer-review process involving 400 reviews in total, 40 full-length
articles were accepted for oral presentation at the conference. This corresponds to an
acceptance rate of 27 % and is intended for maintaining the high standards of the
conference proceedings. The papers included in this LNCS volume cover a wide range
of topics in swarm, evolutionary, memetic, and other intelligent computing algorithms
and their real-world applications in problems selected from diverse domains of science
and engineering.
The conference featured the following distinguished keynote speakers: Dr.
P.N. Suganthan, NTU, Singapore, and Dr. Rammohan Mallipeddi, Kyungpook
National University, South Korea.
We take this opportunity to thank the authors of all submitted papers for their hard
work, adherence to the deadlines, and patience with the review process. The quality of a
refereed volume depends mainly on the expertise and dedication of the reviewers. We
are indebted to the Program Committee/Technical Committee members who not only
produced excellent reviews but also did so in the short time frames that they were given.
We would also like to thank our sponsors for providing all the logistic support and
financial assistance. First, we are indebted to Management and Administrations (faculty
colleagues and administrative personnel) of CMR Technical Campus, Hyderabad. We
thank Prof. Carlos A. Coello Coello, and Prof Nikhil R. Pal, the General Chairs, for
providing valuable guidelines and inspiration to overcome various difficulties in the
process of organizing this conference. We would also like to thank the participants of
this conference. Finally, we would like to thank all the volunteers for their tireless
efforts in meeting the deadlines and arranging every detail to make sure that the
conference could run smoothly. We hope the readers of these proceedings and the
participants of the conference found the papers and conference inspiring and enjoyable.
December 2015 Bijaya Ketan Panigrahi

P.N. Suganthan
Swagatam Das
S.C. Satpathy
Organization
General Chairs
Nikhil R. Pal Indian Statistical Institute, Kolkata, India
Carlos A. Coello Instituto Politécnico Nacional, México
Coello
General Co-chairs
Swagatam Das Indian Statistical Institute, Kolkata, India
B.K. Panigrahi IIT Delhi, New Delhi, India
Program Chair
S.C. Satapathy Anil Neerukonda Institute of Technology and Sciences,
Visakhapatnam, India
Finance Chair
Srujan Raju CMR Technical Campus, Hyderabad, India
Steering Committee Chair

P.N. Suganthan NTU, Singapore
Special Session Chairs

Sanjoy Das Kansas State University, Kansas, USA
Zhihua Cui Taiyuan University of Science and Technology, China
Samuelson Hong Oriental Institute of Technology, Taiwan
International Advisory Committee/Technical Review Committee
Almoataz Youssef Abdelaziz, Egypt Carlos A. Coello Coello, Mexico

Athanasios V. Vasilakos, Athens, Greece Chilukuri K. Mohan, USA
Alex K. Qin, France Delin Luo, China
Amit Konar, India Dipankar Dasgupta, USA
Anupam Shukla, India D.K. Chaturvedi, India
Ashish Anand, India Dipti Srinivasan, Singapore
Boyang Qu, China Fatih M. Tasgetiren, Turkey
VIII Organization
Ferrante Neri, Finland Namrata Khemka, USA

Frank Neumann, Australia N. Puhan, India
Fayzur Rahman, Portugal Oscar Castillo, Mexico
G.K. Venayagamoorthy, USA Pei-Chann Chang, Taiwan
Gerardo Beni, USA Peng Shi, UK
Hai Bin Duan, China Qingfu Zhang, UK
Heitor Silvério Lopes, Brazil Quanke Pan, China
Halina Kwasnicka, Poland Rafael Stubs Parpinelli, Brazil
Hong Yan, Hong Kong, SAR China Rammohan Mallipeddi, Singapore
Javier Del Ser, Spain Roderich Gross, UK
Jane J. Liang, China Ruhul Sarker, Australia
Janez Brest, Slovenia Richa Sing, India
Jeng-Shyang Pan, Taiwan Robert Kozma, USA
Juan Luis Fernández Martínez, Spain Suresh Sundaram, Singapore
Jeng-Shyang Pan, Taiwan S. Baskar, India
Kalyanmoy Deb, India S.K. Udgata, India
K. Parsopoulos, Greece S.S. Dash, India
Kay Chen Tan, Singapore S.S. Pattanaik, India
Ke Tang, China S.G. Ponnambalam, Malaysia
K. Shanti Swarup, India Saeid Nahavandi, Australia
Lakhmi Jain, Australia Saman Halgamuge, Australia
Leandro Dos Santos Coelho, Brazil Shizheng Zhao, Singapore
Ling Wang, China Sachidananda Dehuri, Korea
Lingfeng Wang, China Samuelson W. Hong, Taiwan
M.A. Abido, Saudi Arabia Vincenzo Piuri, Italy
M.K. Tiwari, India X.Z. Gao, Finland
Maurice Clerc, France Yew Soon Ong, Singapore
Meng Joo Er, Singapore Ying Tan, China
Meng-Hiot Lim, Singapore Yucheng Dong, China
M.F. Tasgetiren, Turkey
Contents
Self-adaptive Ensemble Differential Evolution with Sampled Parameter

Values for Unit Commitment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Nandar Lynn, Rammohan Mallipeddi,
and Ponnuthurai Nagaratnam Suganthan
Empirical Assessment of Human Learning Principles Inspired PSO

Algorithms on Continuous Black-Box Optimization Testbed. . . . . . . . . . . . . 17
M.R. Tanweer, Abdullah Al-Dujaili, and S. Suresh
Visual Cryptography Based Lossless Watermarking for Sensitive Images . . . . 29

Surekha Borra, Viswanadha Raju S., and Lakshmi H.R.
Cohort Intelligence and Genetic Algorithm Along with AHP

to Recommend an Ice Cream to a Diabetic Patient . . . . . . . . . . . . . . . . . . . 40
Suhas Machhindra Gaikwad, Rahul Raghvendra Joshi,
and Anand Jayant Kulkarni
Design, Construction and Analysis of Model Dataset for Indian Road

Network and Performing Classification to Estimate Accuracy
of Different Classifier with Its Comparison Summary Evaluation . . . . . . . . . 50
Suwarna Gothane, M.V. Sarode, and K. Srujan Raju
A Hybrid EMD-ANN Model for Stock Price Prediction. . . . . . . . . . . . . . . . 60

Dhanya Jothimani, Ravi Shankar, and Surendra S. Yadav
Development of Back Propagation Neural Network (BPNN) Model

to Predict Combustion Parameters of Diesel Engine. . . . . . . . . . . . . . . . . . . 71
M. Shailaja and A.V. Sita Rama Raju
An Improved Quantum Inspired Immune Clone Optimization Algorithm . . . . 84

Annavarapu Chandra Sekhara Rao, Suresh Dara, and Haider Banka
Diagnosis of Parkinson Disease Patients Using Egyptian Vulture

Optimization Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Aditya Dixit, Alok Sharma, Ankur Singh, and Anupam Shukla
Variance Based Particle Swarm Optimization for Function Optimization

and Feature Selection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Yamuna Prasad, K.K. Biswas, M. Hanmandlu,
and Chakresh Kumar Jain
X Contents
Analysis of Next-Generation Sequencing Data of miRNA for the Prediction

of Breast Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Indrajit Saha, Shib Sankar Bhowmick, Filippo Geraci, Marco Pellegrini,
Debotosh Bhattacharjee, Ujjwal Maulik, and Dariusz Plewczynski
Genetic Algorithm Based Speed Control of Electric Vehicle

with Electronic Differential. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Nair R. Deepthi and J.L. Febin Daya
An Ant Colony Optimization Approach for the Dominating Tree Problem . . . 143
Shyam Sundar, Sachchida Nand Chaurasia, and Alok Singh
Multi-objective Power Dispatch Using Stochastic Fractal Search

Algorithm and TOPSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Hari Mohan Dubey, Manjaree Pandit, B.K. Panigrahi,
and Tushar Tyagi
Particle Swarm Optimization for the Deployment of Directional Sensors . . . . 167

Pankaj Singh, S. Mini, and Ketan Sabale
Region Based Multiple Features for an Effective Content Based Access

Medical Image Retrieval an Integrated with Relevance Feedback Approach . . 176
B. Jyothi, Y. MadhaveeLatha, P.G. Krishna Mohan, and V.S.K. Reddy
Robot Workcell Layout Optimization Using Firefly Algorithm . . . . . . . . . . . 188

Akif Muhtasim Alim, S.G. Ponnambalam, and G. Kanagaraj
Particle Swarm Optimization Based on the Winner’s Strategy. . . . . . . . . . . . 201

Shailendra S. Aote, M.M. Raghuwanshi, and L.G. Malik
Black Hole Artificial Bee Colony Algorithm . . . . . . . . . . . . . . . . . . . . . . . 214

Nirmala Sharma, Harish Sharma, Ajay Sharma,
and Jagdish Chand Bansal
A Gravitational Search Algorithm for Energy Efficient Multi-sink

Placement in Wireless Sensor Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
P.C. Srinivasa Rao, Haider Banka, and Prasanta K. Jana
Optimum Clustering of Active Distribution Networks Using Back Tracking

Search Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
Reham A. Osama, Almoataz Y. Abdelaziz, Rania A. Swief,
Mohamed Ezzat, R.K. Saket, and K.S. Anand Kumar
Energy Efficient Clustering for Wireless Sensor Networks:

A Gravitational Search Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
P.C. Srinivasa Rao, Haider Banka, and Prasanta K. Jana
Contents XI
Hybridizing Cuckoo Search with Bio-inspired Algorithms for Constrained

Optimization Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
G. Kanagaraj, S.G. Ponnambalam, and A.H. Gandomi
A Hybrid Genetic Algorithm Using Dynamic Distance

in Mutation Operator for Solving MSA Problem . . . . . . . . . . . . . . . . . . . . . 274
Rohit Kumar Yadav and Haider Banka
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287

Self-adaptive Ensemble Differential Evolution
with Sampled Parameter Values
for Unit Commitment
Nandar Lynn1, Rammohan Mallipeddi2,

and Ponnuthurai Nagaratnam Suganthan1(&)
1
School of Electrical and Electronics Engineering,
Nanyang Technological University, Singapore, Singapore
{nandar001,epnsugan}@ntu.edu.sg
2
School of Electronics Engineering, Kyungpook National University,
Daegu, South Korea
[email protected]
Abstract. In literature, empirically and theoretically, it has been

well-demonstrated that the performance of differential evolution (DE) is sensi-
tive to the choice of the mutation and crossover strategies and their associated
control parameters. According to the No Free Lunch theorem, a single set of
well-tuned combination of strategies and their associated parameter combination
is not suitable for optimization problems having different characteristics. In
addition, different mutation and crossover strategies with different parameter
settings can be appropriate during different stages of the evolution. Based on this
observation, DE with an ensemble of mutation and crossover strategies and their
associated control parameters referred to as EPSDE was proposed. However, it
has been observed that the fixed discrete parameter values as in EPSDE may not
yield optimal performance. In this paper, we propose self-adaptive DE algorithm
(Sa-EPSDE) with a set of mutation strategies while their associated parameter
values F and CR are sampled using mean and standard deviation values. In
addition, the probability of selecting a combination to produce an offspring at a
particular generation during the evolution process depends on the success of the
combination. The performance of the proposed Sa-EPSDE algorithm is evalu-
ated on a set of 14 bound-constrained problems designed for Conference on
Evolutionary Computation (CEC) 2005. In order to validate the performance of
proposed Sa-EPSDE algorithm on real-world applications, the algorithm is
hybridized with a simple priority listing method and applied to solve unit
commitment problem by considering 10-, 20-, 40-, 60-, 80- and 100-bus systems
for one day scheduling period. The results showed that the proposed method
obtained superior performance against other compared algorithms.
Keywords: Differential evolution Global optimization Parameter

adaptation Ensemble Mutation strategy adaptation Unit commitment
Scheduling
© Springer International Publishing AG 2016

B.K. Panigrahi et al. (Eds.): SEMCCO 2015, LNCS 9873, pp. 1–16, 2016.
DOI: 10.1007/978-3-319-48959-9_1
2 N. Lynn et al.
1 Introduction
Differential Evolution (DE) [1] is a simple and efficient population based stochastic
search technique that is inherently parallel. Due to its ability to handle a variety of
optimization problems, DE is being employed in diverse fields of science and engi-
neering [2–5]. During the last decade, significant research has been done to improve the
search performance of DE. However, to face the challenges posed by the modern
application areas, the performance of DE needs to be enhanced further.
Experimentally [6, 7] and theoretically [8], it has been verified that the performance
of DE is sensitive to the mutation strategy, crossover strategy and intrinsic control
parameters such as population size (NP), crossover rate (CR) and scale factor (F). In
other words, the best combination of strategies and their associated control parameters
can be different for different optimization problems. In addition, for the same opti-
mization problem the best combination can vary depending on the available compu-
tational resources and accuracy requirements [9]. Therefore, to successfully solve a
specific optimization problem, it is necessary to perform trial-and-error search for the
most appropriate combination of strategies and their associated parameter values.
However, the trial-and-error search process is time-consuming and incurs high com-
putational costs. Therefore, to overcome the time consuming trial-and-error procedure,
DE algorithm with different adaptation schemes [10–13] have been proposed in the
literature. In addition, motivated by the observation that during the evolution process
the population of DE may traverse through different regions in the search space, within
which different strategies with different parameter settings may be more effective than
others, in [9, 14] the authors proposed a DE algorithm (EPSDE) based on the idea of
ensemble strategies and parameters. In EPSDE [14], pools of distinct mutation and
crossover strategies along with pools of distinct parameters values for each control
parameter (F and CR) coexist throughout the evolution process and compete to produce
offspring population.
From the literature, it is observed that ensemble of parameters and strategies with
their associated parameters shows significant impact on the performance of the DE
algorithms. However, it has been observed that the performance of the DE algorithm
can be improved by using parameter values sampled from distributions whose mean
values are distributed with the range of the parameters [15, 16]. In addition, during
every generation the number offspring members produced by a combination in the
ensemble depend on its performance during the previous few generations of the evo-
lution. In other words, the probability of a combination producing an offspring in the
current generation depends on its performance during the previous generations of the
evolution. The reminder of this paper is organized as follows: Sect. 2 presents a brief
overview of the differential evolution algorithm and a literature survey on the variants
of DE algorithms. Section 3 presents the proposed Sa-EPSDE algorithm. The perfor-
mance of proposed Sa-EPSDE is evaluated using numerical benchmark problems in
Sect. 4 and unit commitment problem in Sect. 5. Finally, the paper is concluded in
Sect. 6.
Self-adaptive Ensemble Differential Evolution 3
2 Literature Review
2.1 Differential Evolution
Differential Evolution (DE) being a parallel direct search method utilizes NP
D-dimensional parameter
n vectors, oso-called individuals, which encode the candidate
solutions, i.e. Xi;G ¼ x1i;G ; . . .; xD
i;G ; i ¼ 1; . . .; NP. G represents the generation count.
The uniform randomization of the initial population tries to cover the search space
constrained by the prescribed minimum and maximum parameter bounds, Xmin ¼
fx1min ; . . .; xD
min g and Xmax ¼ fxmax ; . . .; xmax g, as much as possible. For example, the
1 D
th th
initial value of the j parameter of the i individual at generation G = 0 is generated by:
j j j j
xi;0 ¼ xmin þ rand j ð0; 1Þ: xmax xmin j ¼ 1; 2; . . .; D ð1Þ
where randj(0,1) is a uniformly distributed random variable in the range [0,1) [19].
After initialization, DE employs mutation operation to produce mutant vector Vi,G
corresponding to each individual Xi,G, so-called target vector, in the current population.
For each target vector Xi,G in generation G, its associated mutant vector can be gen-
erated via mutation strategy. The most frequently used mutation strategies are [9]:
“DE/best/1”:
Vi;G ¼ Xbest;G þ F:ðXr1i ;G Xr2i ;G Þ ð2Þ
“DE/best/2”:
Vi;G ¼ Xbest;G þ F:ðXr1i ;G Xr2i ;G Þ þ F:ðXr3i ;G Xr4i ;G Þ ð3Þ
“DE/rand/1”:
Vi;G = Xr1i ;G þ F:ðXr2i ;G Xr3i ;G Þ ð4Þ
“DE/rand/2”:
Vi;G = Xr1i ;G þ F:ðXr2i ;G Xr3i ;G Þ þ F:ðXr4i ;G Xr5i ;G Þ ð5Þ
“DE/current-to-rand/1”:
Ui;G = Xi;G þ K:ðXr1i ;G Xi;G Þ þ F:ðXr2i ;G Xr3i ;G Þ ð6Þ
The indices r1i ; r2i ; r3i ; r4i ; r5i are mutually exclusive integers randomly generated
anew for each mutant vector within the range [1, NP], which are also different from the
index i. Xbest,G is the best individual vector with the best fitness value in the population
at generation G. K is randomly chosen within the range [0, 1]. In the above equations,
the scale factor F 2 ð0; 1 þ Þ [17] is a positive real number that controls the rate of
4 N. Lynn et al.
population evolution. There are various claims and counter-claims regarding the suit-
able range of the F values [17]. However, F must be above a certain critical value to
avoid premature convergence [6]. In addition, a larger F increases the probability of
escaping from a local optimum [6]. But if F becomes too large, the number of function
evaluations to find the optimum grows very quickly. In [18], it was theoretically proved
that DE could converge to global optimum in the long time limit if F can be trans-
formed into a Gaussian random variable. However, it was later demonstrated [17] that
unless the variance of the randomizing distribution is very small, DE will suffer a
significant performance loss on highly conditioned non-separable functions.
After the mutation, crossover operation is applied to each pair of the target vector
Xi,G and its corresponding mutant vector Vi,G to generate a trial vector:
Ui;G ¼ fu1i;G ; . . .; uD
i;G g. The crossover operation speeds the convergence by a constant
factor [17]. In the basic version, DE employs the binomial crossover defined as follows:
j

j vi;G if ðrand j ð0;1Þ CRÞ or ðj ¼ jrand Þ j ¼ 1; 2 ; ...; D
ui;G ¼ j
xi;G otherwise
ð7Þ
In Eq. (7), the crossover probability,CR 2 ½0; 1 is a user-specified constant that

controls the fraction of parameter values that are copied to the trail vector from the
mutant vector. jrand is a randomly chosen integer in the range [1, D]. The crossover rate
(CR) is a probability (0 ≤ CR ≤ 1) of mixing between trial and target vectors and
controls how many components are mutated in each element of the current population.
In [6], various guidelines to select the appropriate values for a problem-at-hand were
put forward. However, a simple guideline which states that small values are suitable for
separable problems while larger values of CR are suitable for multi-modal, parameter
dependent problems [17] is commonly used.
If generated trial vector exceeds the corresponding upper and lower bounds, it is
randomly and uniformly reinitialized in the range of lower and upper bound. The
objective function values of all trial vectors are evaluated. After crossover operation, a
selection operation is performed. The fitness function value of each trial vector f (Ui,G)
is compared to that of its corresponding target vector f(Xi,G). If the trial vector has less
or equal fitness function value (in a minimization problem), the trial vector will replace
the target vector and enter the population of the next generation. Otherwise, the target
vector will remain in the population for the next generation. The selection operation
can be expressed as follows:

Ui;G ; if f ðUi;G Þ f ðXi;G Þ
Xi;G þ 1 ¼ ð8Þ
Xi;G ; otherwise
The 3 steps (mutation, crossover and selection) are repeated generation after gen-
eration until a termination criterion (reaching the maximum number of function eval-
uations set) is satisfied.
2.2 Literature Review

As DE is sensitive to parameter settings [7], to avoid manual tuning of parameter
settings and to get an optimal performance, several variants of DE based on the adap-
tation and self-adaptation of mutation strategies and control parameters have been
proposed. DE may suffer from stagnation and premature convergence due to improper
selection of control parameters, parameters being kept fixed through the whole search
process [19]. To overcome these problems FADE was proposed which adapts the
control parameters F and CR based on fuzzy logic controllers whose inputs are the
relative function values and individuals of successive generations to adapt the param-
eters for the mutation and crossover operation [19]. FADE enables to choose the initial
control parameters freely and adjust the control parameters on-line to dynamically adapt
to changing situations. A parameter adaptation of DE (ADE) based on controlling the
population diversity and a multi-population approach was proposed [20]. In [10], a
self-adaptation scheme (SDE) in which CR is generated randomly for each individual
using a normal distribution N (0.5, 0.15), while F is adapted analogous to the adaptation
of crossover rate CR in [21].
DE can encompass a number of trial vector generation strategies, each of which
may be effective over certain problems but poorly perform over the others. In [12], a
self-adaptive DE algorithm (SaDE) was proposed in which the mutation strategies and
the respective control parameter are self-adapted based on their previous experiences of
generating promising solutions. F was randomly generated with a mean and standard
deviation of 0.5 and 0.3 respectively. In SaDE, four effective trial vector generation
strategies namely the DE/rand/1/bin, DE/rand-to-best/2/bin, DE/rand/2/bin and finally
DE/current-to-rand/1 were chosen to constitute a strategy candidate pool. In SaDE
algorithm, for each target vector in the current population, one trial vector generation
strategy is selected from the candidate pool according to the probability learned from
its success rate in generating improved solutions within a certain number of previous
generations, called the Learning Period (LP). The selected strategy is subsequently
applied to the corresponding target vector to generate a trial vector. More specifically,
at each generation, the probabilities of choosing each strategy in the candidate pool are
summed to 1. These probabilities are initially equal (1/K for K strategies in the pool)
and are then gradually adapted during evolution, based on the Success and Failure
Memories (nsk,g and nfk,g) over the previous LP generations. The adaptation of the
probabilities take place in such a fashion that, the larger the success rate for the kth
strategy in the pool within the previous LP generations, the larger is the probability of
applying it to generate trial vectors at the current generation. The probability of
selecting the strategies (k = 1, 2… K) in the pool is updated at the generation G using
the following equation:
Sk;G
pk;G ¼ PK ð9Þ
k¼1 Sk;G
6 N. Lynn et al.
where,
PG1
g¼GLP nSk;g
Sk;G ¼ PG1 PG1 þe ð10Þ
g¼GLP nsk;g þ g¼GLP nfk;g
(k = 1, 2, .., K; G > LP)

In Eq. (10), Sk,G is the success rate of trail vectors generated by the kth strategy that
can successfully enter the next generation within learning period LP. In order to avoid
the null success rate, ε = 0.01 is used.
In JADE [22], a new mutation strategy “DE/current-to-pbest”, a generalization of
the classic “DE/current-to-best” was proposed. The reliability of the algorithm is fur-
ther improved by the adaptive parameter control. “DE/current-to-pbest” utilizes not
only the best solution information but also the information of other good solutions.
Vi;G ¼ Xi;G þ Fi;G :ðXpbest;G Xi;G Þ þ Fi;G :ðXr1i ;G Xr2i ;G Þ ð11Þ
p
Xbest;G is randomly chosen as one of the top 100p % individuals in the population with
p 2 ð0; 1. Fi,G is the mutation factor of Xi,G, generated independently at each gener-
ation G, according to a Cauchy distribution as follows:
Fi;G ¼ randci ðlF ; 0:1Þ ð12Þ
SF denotes the set of all successful mutation factors in generation G. The location
parameter µF is initialized to be 0.5 and then updated at the end of each generation as:
lF ¼ ð1 cÞ:lF þ c:meanL ðSF Þ ð13Þ

P
F2
F2SF
meanL ¼ P ð14Þ
F
F2SF
where, meanL(.) is the Leher mean. CRi,G is the crossover probability of Xi,G, generated
independently at each generation G, according to a normal distribution as follows:
CRi;G ¼ randni ðlCR ; 0:1Þ ð15Þ
SCR denotes the set of all successful crossover probabilities in generation G. The mean
µCR is initialized to be 0.5 and then updated at the end of each generation as:
lCR ¼ ð1 cÞ: lCR þ c:meanA ðSCR Þ ð16Þ

where, meanA(.) is the usual arithmetic mean. JADE usually performs better with 1/c 2
[5, 20] and p 2 [5 %, 20 %]; i.e., life span of µCR and µF values ranges from 5 to 20
generations and top 5–20 % high quality solutions are considered for mutation. In [15],
an improved version of JADE called success history based adaptive DE (SHADE)
algorithm was proposed in which history-based parameter adaptation scheme is used.
In SHADE, F and CR values are generated using Eqs. (12) and (15) and mean values
of successful F and CR are stored in historical memory. Linear population size
reduction was introduced to improve the search performance of SHADE in [16].
3 Self-adaptive DE with Ensemble of Mutation Strategies

and Sampled Parameter Values (Sa-EPSDE)
An ensemble of mutation strategies and parameter values for DE (EPSDE) in which a

pool of mutation strategies, along with a pool of values corresponding to each asso-
ciated parameter competes to produce successful offspring population was proposed.
Moreover, EPSDE is incorporated with a self-adaptive framework [23] in which the
mutation and crossover strategies and parameters in the pool are gradually self-adapted
by learning from their previous recorded performance [12]. In Sa-EPSDE, the
parameter values are fixed discrete values. However, the performance of the Sa-EPSDE
algorithm can be further enhanced if the parameter values used during evolution
process are sampled from the distribution with different mean values [15, 16].
Therefore, Sa-EPSDE is modified with the values of F and CR which are sampled from
fixed mean values using Eqs. 12 and 15, respectively. In order to balance the speed and
efficiency while solving problems with different characteristics, the pool of CR values
is taken in the range 0.1 to 0.9 in steps of 0.1. Based on the literature, the pool of
F values is taken in the range 0.4 to 0.9 in steps of 0.1. The mutation strategies are used
as same as in EPSDE [9]. The proposed Sa-EPSDE algorithm is demonstrated below:
8 N. Lynn et al.
STEP 1: Set G=0, and randomly initialize NP individuals PG = {X 1,G ,..., X NP ,G }

with X i ,G = {x 1i ,G ,..., x iD,G }, i = 1,..., NP uniformly distributed in the range [Xmin, Xmax],
where X min = {x 1min ,..., x min
D
} and X max = {x 1max ,..., x max
D
} . Initialize strategy probability (pk,G,
k=1,..., K; K is number of available strategies in the pool) and learning period LP.
STEP 2: Select a pool of mutation strategies and a pool of values for each associated
parameters corresponding to each mutation strategy.
STEP 3: Each population member is randomly assigned with one of the mutation
strategy from the pool and the associated parameter values are chosen randomly from
the corresponding pool of values.
STEP 4: WHILE stopping criterion is not satisfied
DO
FOR i = 1 to NP
STEP 4.1 Calculate strategy probability pk,G and update Success and
Failure Memory (nsk,g and nfk,g)
IF G > LP
FOR k=1: K
update pk,G by equation (9)
remove nsk,G-LP and nfk,G-LP out of Success and Failure
Memory respectively
END FOR
END IF
STEP 4.2 Mutation:
Generate a mutated vector corresponding to the target vector Xi,G.
Sample the F value to be used during the mutation using equation
12, where mean value to be used in equation 12 is the discrete value
selected from the pool.
STEP 4.3 Crossover
Generate a trial vector for each target vector Xi,G. Sample the CR
value to be used during the mutation using equation 15, where mean
value to be used in equation 15 is the discrete value selected from
the pool.
STEP 4.4 Selection
Evaluate the trial vector Ui,G
IF f(Ui,G) ≤ f(Xi,G), THEN Xi,G+1 = Ui,G, f(Xi,G+1) ≤ f(Ui,G) , nsk,G = nsk,G+1
IF f(Ui,G) < f(Xbest,G), THEN Xbest,G = Ui,G, f(Xbest,G) ≤ f(Ui,G)
/* Xbest,G is the best individual in generation G */
END IF
ELSE
Xi,G+1 = Xi,G, f(Xi,G+1) ≤ f(Xi,G), nfk,G = nfk,G+1
END IF
STEP 4.5 Updating
IF f(Ui,G) > f(Xi,G), THEN randomly select a new mutation strategy
and parameter values from the pools or from the
stored successful combinations.
END IF
END FOR
Store nsk,G & nfk,G, k=1,..., K, into Success and Failure Memory respectively.
STEP 4.6 Increment the generation count G = G + 1
END WHILE
4 Experimental Study on Numerical Benchmark Problems
In this paper, CEC 2005 benchmark functions featuring different properties

(uni-modal/multi-modal, shifted, rotated, scalable, separable/non-separable) are used to
evaluate the performance of the proposed Sa-EPSDE with F and CR sampled from the
distribution with fixed mean values using Eqs. 12 and 15. The proposed Sa-EPSDE
algorithm is compared with SaDE and EPSDE and the experiments are conducted on
the first 14 benchmark problems described in [24]. In the experiments, the population
NP is set to 30 and three different dimension D sizes (10D, 30D and 50D) are tested.
The number of function evaluation (FES) is set up 100,000 for 10D problems, 300, 000
for 30D problems and 500,000 for 50D problems. For all the algorithms, the experi-
ment is run 25 times for each problem. The experiment results for 10D, 30D and 50D
are shown in Tables 1, 2 and 3 respectively and the best experimental results are
highlighted in bold in each table.
For 10 dimensional uni-modal problems, all the algorithms perform approximately
equal on function F1. The proposed Sa-EPSDE algorithm outperforms best on func-
tions F2, F4 and F5 and EPSDE offers best performance on function F3. For
multi-modal problems, SaDE performance is best on F7, F11 and F14, EPSDE on F9
and F10 and Sa-EPSDE on all other problems. Overall, compared to SaDE and EPSDE
algorithms, proposed Sa-EPSDE algorithm provides best performance on 8 out of 14
10 dimensional problems.
For 30 dimensional problems in Table 2, compared to SaDE, the proposed
Sa-EPSDE algorithm outperforms on 9 out of 14 functions and performs approximately
equal on test function F8. Compared to EPSDE algorithm, Sa-EPSDE algorithm offers
Table 1. Comparison between Sa-EPSDE, SaDE and EPSDE for 10D problems
Type Error means (F(x) – F(x*)) ± Standard Deviation for 10D problems
F SaDE EPSDE Sa-EPSDE
mean –std mean –std mean –std
Uni-modal F1 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00
F2 1.48E–27 4.22E–27 1.92E–28 1.59E–28 8.94E–29 1.25E–28
F3 2.56E+04 3.44E+04 3.45E+01 1.24E+02 1.12E+03 3.09E+03
F4 1.37E–28 1.82E–28 4.06E–28 4.95E–28 1.11E–28 1.08E–28
F5 3.01E–08 1.36E–07 3.60E+01 4.28E+01 0.00E+00 0.00E+00
Multi-modal F6 5.57E+00 1.45E+01 3.19E–01 1.10E+00 6.78E–26 1.63E–25
F7 5.96E–02 4.57E–02 2.13E–01 2.04E–02 6.84E–02 4.42E–02
F8 2.04E+01 6.72E–02 2.04E+01 8.00E–02 2.02E+01 1.07E–01
F9 3.98E–02 1.99E–01 0.00E+00 0.00E+00 0.00E+00 0.00E+00
F10 7.08E+00 2.74E+00 6.02E+00 2.08E+00 1.24E+01 6.00E+00
F11 2.21E+00 1.09E+00 6.08E+00 9.79E–01 3.94E+00 1.66E+00
F12 1.96E+02 5.13E+02 2.22E+02 1.23E+02 2.27E+01 5.80E+01
F13 4.04E–01 1.25E–01 2.49E–01 3.81E–02 2.15E–01 8.25E–02
F14 2.47E+00 4.66E–01 3.51E+00 3.21E–01 3.06E+00 3.59E–01
10 N. Lynn et al.
Uni-modal F1 3.43E–29 1.03E–28 1.57E–29 4.22E–29 0.00E+00 0.00E+00
F2 9.60E–05 2.85E–04 1.78E–24 4.81E–24 1.27E–25 4.26E–25
F3 6.25E+05 2.12E+05 5.24E+04 3.91E+04 1.19E+05 6.46E+04
F4 6.77E+02 6.21E+02 1.47E+03 2.35E+03 1.15E+00 3.48E+00
F5 4.40E+03 6.74E+02 1.99E+03 7.75E+02 1.89E+03 8.34E+02
Multi-modal F6 4.84E+01 3.96E+01 1.12E+00 1.83E+00 2.69E–23 4.77E–23
F7 2.03E–02 1.52E–02 6.20E+00 4.40E+00 1.78E–02 1.84E–02
F8 2.09E+01 4.78E–02 2.10E+01 4.91E–02 2.07E+01 2.06E–01
F9 1.83E+00 1.74E+00 1.19E–01 3.30E–01 0.00E+00 0.00E+00
F10 5.76E+01 1.39E+01 5.38E+01 1.54E+01 5.97E+01 2.12E+01
F11 2.02E+01 3.17E+00 3.33E+01 4.25E+00 2.42E+01 3.52E+00
F12 3.23E+03 3.03E+03 3.58E+04 2.76E+04 2.80E+03 2.58E+03
F13 2.46E+00 8.81E–01 1.71E+00 4.85E–01 8.78E–01 1.70E–01
F14 1.23E+01 4.81E–01 1.35E+01 4.33E–01 1.25E+01 4.49E–01
Uni-modal F1 1.89E–28 3.78E–28 2.82E–28 2.57E–28 0.00E+00 0.00E+00
F2 7.16E–01 1.60E+00 3.01E–19 1.43E–18 2.40E–13 5.97E–13
F3 1.19E+06 3.79E+05 1.44E+07 3.99E+07 3.81E+05 1.27E+05
F4 1.23E+04 5.24E+03 1.70E+04 1.56E+04 1.36E+03 1.88E+03
F5 1.16E+04 1.62E+03 6.43E+03 1.85E+03 7.37E+03 1.81E+03
Multi-modal F6 1.24E+02 1.10E+02 2.40E+00 2.00E+00 1.60E–19 7.45E–19
F7 4.14E–03 8.14E–03 1.05E+00 1.12E–01 5.31E–03 1.21E–02
F8 2.11E+01 3.23E–02 2.11E+01 4.07E–02 2.11E+01 6.83E–02
F9 1.05E+01 4.77E+00 9.15E–01 2.20E+00 2.39E–01 4.34E–01
F10 1.72E+02 3.21E+01 1.60E+02 3.65E+01 1.15E+02 2.77E+01
F11 4.33E+01 3.34E+00 7.16E+01 3.72E+00 4.89E+01 4.56E+00
F12 1.04E+04 6.20E+03 2.41E+05 4.07E+04 1.67E+04 1.16E+04
F13 4.04E+00 1.48E+00 5.29E+00 4.05E–01 1.50E+00 2.74E–01
F14 2.16E+01 6.68E–01 2.34E+01 3.72E–01 2.22E+01 5.78E–01
better performance 10 out 14 test functions and performs approximately equal on

function F2 and F8. SaDE performs better on function F10, F11 and F14 and EPSDE
does on F3 and F10.
For 50 dimensional problems in Table 3, compared to SaDE, the proposed
Sa-EPSDE algorithm performs better on 9 out of 14 problems. SaDE performs best on
function F7, F11, F12 and F14. Compared to EPSDE algorithm, Sa-EPSDE algorithm
provides better performance on 11 out 14 problems. EPSDE outperforms Sa-EPSDE on
functions, F2 and F5. The three algorithms perform approximately equal on function
F8. Overall, Sa-EPSDE algorithm outperforms best on 7 out of 14 uni-modal and
multi-modal 50D problems.
In all three dimensions, Sa-EPSDE algorithm outperforms SaDE and EPSDE
algorithms by a large margin such as in function F6 and F9. The algorithm performs
only slightly worse than the other two algorithms on few problems such as in function
F11 and F12. Therefore, proposed Sa-EPSDE algorithm offers best performance on all
10D, 30D and 50D shifted rotated benchmark problems.
5 Experimental Study on Unit Commitment Problem
Unit commitment (UC) problem is one of the important optimization problems in the
electrical power system. UC problem deals with producing the optimal schedule of
available power generating units over a scheduling period while satisfying load demand
and spinning reserve requirements at the minimum production cost. UC can be divided
into two subproblems: unit scheduled subproblem and economic dispatch subproblem.
Unit scheduled subproblem determines on/off schedule of generating units while
meeting the system and generating unit constraints. Economic dispatch subproblem
deals with allocating load demand and spinning reserve requirements among the
committed units during each scheduling hour.
Therefore, mathematically, UC has been commonly formulated as a nonlinear,
large scale, mixed-integer combinatorial optimization problem with constraints and the
two subproblems can be solved separately. A large number of deterministic and
metaheuristic optimization methods have been applied to solve the UC problem in the
literature. Deterministic methods are fast yet may miss the optimal solution and
meta-heuristic methods are flexible yet computationally expensive. However, by
combining these two methods, they can benefit from each other and enable us to solve
the large scale UC problems. Thus, in this paper, we hybridized our proposed
Sa-EPSDE with simple priority listing method to solve the UC problem. In the pro-
posed hybrid solution, simple priority listing method is used as a deterministic method
to solve the unit scheduled subproblem and our proposed Sa-EPSDE is used as a
metaheuristic method to handle the economic dispatch subproblem.
5.1 Problem Formulation

Objective Function: Mathematically, overall UC objective function can be described as
follows:
12 N. Lynn et al.
Minimize Production Cost (PC):
N X
X T
PC ¼ Fit ðPit Þ þ STit ð16Þ
i¼1 t¼1
where, Fit(Pit) is fuel cost of unit i at time t which is a quardic function of output power
generation of a unit.
Fit ðPit Þ ¼ ai þ bi Pit þ ci P2it ð17Þ
ai, bi and ci are cost coefficients.

STit is start up cost unit i at time t. The start up cost depends on the time the unit has
been off before start up. The start up cost will be cold start up cost (SCic ) when down
time duration (Toffi ) of unit i exceedes cold start hour (c-s-houri) in excess of minimum
down time (Tidown ) and will be hot start up cost (SCih ) when down time duration is less
than c-s-houri in excess of Tidown.

SCih if Tioff Tidown þ c s houri
SCit ¼ ð18Þ
SCic if Tioff [ Tidown þ c s houri
Constraints: The constraints to be satisfied during the optimization are as follows:

(a) System power balance: Generated power from the committed units must be bal-
anced with the system power load demand Dt at time t.
XN
i¼1
Pit ¼ Dt ð19Þ
(b) System spinning reserve requirement: Spinning reserve Rt is required in the

operation of a power system in order to prevent load interruption from certain
equipment outages. The reserve is usually considered to be 5 % or 10 % of the
forecasted load demand.
XN
I Pmax
i¼1 it i
Dt þ Rt ð20Þ
(c) Generation power limits: each unit has generation range limited by the minimum
Pmin
i and maximum Pmax
i power values as follows:
Pmin
i Pit Pmax
i ð21Þ
(d) Unit minimum up Tiup and down Tidown time: Once each unit is turn on/shut down,
it must be committed/decommitted for a certain predefined time before it is shut
down/brought online.
Tiup Tion ð22Þ
Tidown Tioff ð23Þ
(e) Unit initial status: At the start of the scheduling period, initial status of each unit
must be taken into account.
5.2 Simple Priority Listing Method

Priority listing method is the simple solution which is based on maximum power
generation capacity of each unit. The units with higher maximum power generation
capacity will have higher priority to commit. Thus, a unit with the highest maximum
generation power capacity will be located on the top of the priority list and the other
units will be located in the ascending order of their maximum output power towards the
bottom of the list. For the units of equal maximum power generation capacity, the one
with lower heat rate will have higher priority. The heat rate is calculated according to
the following formula:
Fit ðPit Þ
HRi ¼ ð24Þ
Pit
Based on the priority list, the units are committed until the load demand and the
spinning reserve constraints are satisfied at each time interval.
5.3 Experimental Results

A benchmark system of 10 generators from [25] is used as a power system testbed in
this paper. The experiment is conducted for 10-, 20-, 40-, 60-, 80-, and 100-bus
systems. For the 20-bus system and above, the base 10 units are replicated and the load
demand is multiplied accordingly. The spinning reserve is considered to be 10 % of the
load demand in this experiment. The scheduling period is for one day (24 h) and the
load demand for 24 h is presented in Table 4 [26]. The performance of proposed
hybrid model is evaluated with other hybrid model combined basic DE and PL method.
The results are compared in terms of average production cost averaged over 30 runs
Table 4. Load demand [26]

hr Dt hr Dt hr Dt hr Dt hr Dt hr Dt
1 700 5 1000 9 1300 13 1400 17 1000 21 1300
2 750 6 1100 10 1400 14 1300 18 1100 22 1100
3 850 7 1150 11 1450 15 1200 19 1200 23 900
4 950 8 1200 12 1500 16 1050 20 1400 24 800
14 N. Lynn et al.
Table 5. Performance comparison in terms of production cost ($)

Unit systems Sa-EPSDE with PL DE with PL Cost difference
10 624774.92 624807.60 3.27E+01
20 1349484.31 1349520.18 3.59E+01
40 2752714.27 2753588.75 8.74E+02
60 4152472.31 4153835.67 1.36E+03
80 5545851.90 5548220.52 2.37E+03
100 6291795.18 6293104.69 1.31E+03
and shown in Table 5. As seen in Table 5, the hybrid model of SaEPSDE and PL
performs better than DE+PL hybrid model in all the power systems. Especially
SaEPSDE with PL obtains lower production cost of 1000*2400 per day on the large
power system of 60-, 80- and 100- bus systems.
6 Conclusion
In this paper, self-adaptive DE algorithm with ensemble of strategies and sampled

parameter values (Sa-EPSDE) with F and CR sampled from the distribution of different
fixed mean values is proposed to determine suitable strategy with associated parameter
settings for different stages of the search process. The performance of Sa-EPSDE
algorithm is tested on shifted rotated uni-modal and multi-modal CEC 2005 benchmark
functions and compared with SaDE and EPSDE algorithms on 10D, 30D and 50D
problems. According to mean and standard deviation criteria, the proposed Sa-EPSDE
algorithm performs better than other two DE algorithms on all three dimensions. In
order to evaluate the performance of Sa-EPSDE on real-world applications, Sa-EPSDE
is combined with priority listing method and applied to solve unit commitment power
problem. The performance is compared with basic DE algorithm in terms of mean
production cost averaged over 30 trial runs. The experimental results showed that
proposed Sa-EPSDE offered lower production cost than compared DE algorithm.
Acknowledgement. This work was supported by the Singapore National Research Foundation
(NRF) under its Campus for Research Excellence and Technological Enterprise (CREATE)
programme, and Cambridge Advanced Research Centre in Energy Efficiency in Singapore
(CARES), C4T project.
References
1. Das, S., Suganthan, P.N.: Differential evolution: a survey of the state-of-the-art. IEEE Trans.
Evol. Comput. 15, 4–31 (2011)
2. Das, S., Konar, A.: Automatic image pixel clustering with an improved differential
evolution. Appl. Soft Comput. 9(1), 226–236 (2009)
3. Mallipeddi, R., et al.: Efficient constraint handling for optimal reactive power dispatch
problems. Swarm Evol. Comput. 5, 28–36 (2012)
4. Mallipeddi, R., et al.: Robust adaptive beamforming based on covariance matrix

reconstruction for look direction mismatch. Prog. Electromagn. Res. Lett. 25, 37–46 (2011)
5. Venu, M.K., Mallipeddi, R., Suganthan, P.N.: Fiber Bragg grating sensor array interrogation
using differential evolution. Optoelectron. Adv. Mater. Rapid Commun. 2, 682–685 (2008)
6. Gämperle, R., Müller, S.D., Koumoutsakos, P.: A parameter study for differential evolution.
In: Advances in Intelligent Systems, Fuzzy Systems, Evolutionary Computation, pp. 293–
298. WSEAS Press, Interlaken, Switzerland (2002)
7. Liu, J., Lampinen. J.: On setting the control parameter of the differential evolution method.
In: Proceedings of MENDEL 2002 8th International Conference on Soft Computing (2002)
8. Jingqiao, Z., Sanderson. A.C.: An approximate gaussian model of differential evolution with
spherical fitness functions. In: IEEE Congress on Evolutionary Computation (2007)
9. Mallipeddi, R., Mallipeddi, S., Suganthan, P.N., Tasgetiren, M.F.: Differential evolution
algorithm with ensemble of parameters and mutation strategies. Appl. Soft Comput. 11,
1679–1696 (2011)
10. Omran, Mahamed, G.,H., Salman, A., Engelbrecht, Andries, P.: Self-adaptive differential
evolution. In: Hao, Y., Liu, J., Wang, Y., Cheung, Y.-m., Yin, H., Jiao, L., Ma, J., Jiao, Y.-C.
(eds.) CIS 2005. LNCS (LNAI), vol. 3801, pp. 192–199. Springer, Heidelberg (2005). doi:10.
1007/11596448_28
11. Brest, J., et al.: Self-adapting control parameters in differential evolution: a comparative
study on numerical benchmark problems. IEEE Trans. Evol. Comput. 10, 646–657 (2006)
12. Qin, A.K., Huang, V.L., Suganthan, P.N.: Differential evolution algorithm with strategy
adaptation for global numerical optimization. IEEE Trans. Evol. Comput. 13, 398–417
(2009)
13. Tvrdik, J.: Adaptation in differential evolution: a numerical comparison. Appl. Soft Compu.
9, 1149–1155 (2009)
14. Mallipeddi, R., Suganthan, P.N.: Differential evolution algorithm with ensemble of
parameters and mutation and crossover strategies. In: Panigrahi, B.K., Das, S., Suganthan,
P.N., Dash, S.S. (eds.) SEMCCO 2010. LNCS, vol. 6466, pp. 71–78. Springer, Heidelberg
(2010). doi:10.1007/978-3-642-17563-3_9
15. Tanabe, R., Fukunaga. A.: Evaluating the performance of SHADE on CEC 2013 benchmark
problems. In: IEEE Congress on Evolutionary Computation (2013)
16. Tanabe, R., Fukunaga. A.S.: Improving the search performance of SHADE using linear
population size reduction. In: IEEE Congress on Evolutionary Computation (2014)
17. Price, K.V., Storn, R.M., Lampinen, J.A.: Differential Evolution: A Practical Approach to
Global Optimization. Natural Computing Series. Springer, Berlin (2005)
18. Zaharie, D.: Critical values for the control parameters of differential evolution. In:
Proceedings of 8th International Conference on Soft Computing, MENDEL 2002, Brno,
Czech Republic (2002)
19. Liu, J., Lampinen, J.: A fuzzy adaptive differential evolution algorithm. Soft. Comput. 9,
448–462 (2005)
20. Zaharie, D.: Control of population diversity and adaptation in differential evolution
algorithms. In: Proceedings of the 9th International Conference on Soft Computing, Brno,
pp. 41–46 (2003)
21. Abbass, H.A.: The self-adaptive pareto differential evolution algorithm. In: IEEE Congress
on Evolutionary Computation, pp. 831–836 (2002)
22. Zhang, J.: JADE: adaptive differential evolution with optional external archive. IEEE Trans.
Evol. Comput. 13, 945–958 (2009)
23. Shi-Zheng, Z., Suganthan. P.N.: Comprehensive comparison of convergence performance of
optimization algorithms based on nonparametric statistical tests. In: IEEE Congress on
Evolutionary Computation (2012)
16 N. Lynn et al.
24. Suganthan, P.N., Hansen, N., Liang, J.J., Deb, K., Chen, Y.-P., Auger, A., Tiwari, S.:
Problem definitions and evaluation criteria for the CEC 2005 special session on
real-parameter optimization. In: Proceedings of Congress on Evolutionary Computation,
pp. 1–50 (2005)
25. Kazarlis, S.A., Bakirtzis, A.G., Petridis, V.: A genetic algorithm solution to the unit
commitment problem. IEEE Trans. Power Syst. 11, 83–92 (1996)
26. Juste, K.A., Kita, H., Tanaka, E., Hasegawa, J.: An evolutionary programming solution to
the unit commitment problem. IEEE Trans. Power Syst. 14, 1452–1459 (1999)
Empirical Assessment of Human Learning
Principles Inspired PSO Algorithms
on Continuous Black-Box Optimization Testbed
M.R. Tanweer(B) , Abdullah Al-Dujaili, and S. Suresh
School of Computer Engineering, Nanyang Technological University,

Singapore, Singapore
{muhammad170,aldujail001}@e.ntu.edu.sg, [email protected]
Abstract. This paper benchmarks the performance of one of the recent

research directions in the performance improvement of particle swarm
optimization algorithm; human learning principles inspired PSO vari-
ants. This article discusses and provides performance comparison of nine
different PSO variants. The Comparing Continuous Optimizers (COCO)
methodology has been adopted in comparing these variants on the noise-
less BBOB testbed, providing useful insight regarding their relative effi-
ciency and effectiveness. This study provides the research community a
comprehensive account of suitability of a PSO variant in solving selective
class of problems under different budget settings. Further, certain recti-
fications/extensions have also been suggested for the selected PSO vari-
ants for possible performance enhancement. Overall, it has been observed
that SL-PSO and MePSO are most suited for expensive and moderate
budget settings respectively. Further, iSRPSO and TPLPSO have pro-
vided better solutions under cheap budget settings where iSRPSO has
shown robust behaviour (better solutions over dimensions). We hope this
paper would mark a milestone in assessing the human learning princi-
ples inspired PSO algorithms and used as a baseline for performance
comparison.
Keywords: PSO · Human learning principles inspired PSO variants ·

COCO methodology · Black-box optimization
1 Introduction
Particle Swarm Optimization (PSO) [10] is a population-based search optimiza-
tion algorithm, inspired from the social behavior a swarm of bird. It has been
widely used for solving numerous optimization problem [4,11,22] and has suc-
cessfully provided solutions to the complex real-world optimization problems
[15,19,21]. In the past two decades, its simplicity and computational efficiency
has attracted the researchers. As a result, several research directions have been
studied including parameter tuning, neighbourhood topology, learning strategies
etc. [5,13,27]. The current state-of-the-art research in PSO includes a diverse

c Springer International Publishing AG 2016
B.K. Panigrahi et al. (Eds.): SEMCCO 2015, LNCS 9873, pp. 17–28, 2016.
DOI: 10.1007/978-3-319-48959-9 2
18 M.R. Tanweer et al.
collection of modified PSO variants, with one variant performing better than
other on a class of optimization problems. It has been shown in human learn-
ing psychology that human beings are better planners and possess intelligent
information processing skills. This helps one to perform better self-regulation
of the cognitive strategies and hence enhance the decision making abilities [14].
Therefore, algorithms developed using human learning principles have shown
promising characteristics [17,20]. Inspired from these findings, researchers have
tried to design such PSO variants that can provide better solutions on various
classes of problems by introducing human-like learning principles [18,22,24,25].
This research direction has provided robust and efficient PSO variants capa-
ble of solving more classes of optimization problems. Recently, researchers have
developed several human learning principles inspired PSO variants which has
significantly enhanced the algorithms’ performance. Therefore, it is required to
benchmark the performance of these algorithms to come up with a baseline for
performance comparison in assessing the human learning principles inspired PSO
algorithms.
This paper aims towards comparing eight different human learning principles
inspired PSO variants carefully selected to reflect different self and social princi-
ples applied to the PSO algorithm against the standard PSO algorithm. Through
experimental evaluation, this paper marks the differences among them and inves-
tigates the suitability of the algorithms for solving optimization problems with
different characteristics for different computational requirements. Furthermore,
it draws several concluding points that can help the future development in the
PSO research. To achieve the paper’s goals, the experimental evaluation must
be able to discover and distinguish the good features of the variants over others,
and show their differences over different stages of the search for various goals
and situations. In this paper, the Comparing Continuous Optimizer (COCO)
methodology [7] has been adopted as it meets the requirements. It comes with a
testbed of 24 scalable noiseless functions [8] addressing such real-world difficulties
as ill-conditioning, multi-modality, and dimensionality. The selected PSO vari-
ants have diverse behaviour in terms of solution accuracy, convergence rate and
robustness. Some require more time to converge [9,23] whereas others are com-
putationally efficient [12,21,24]. Therefore, the performance has been assessed
over different settings of function evaluations to investigate the capability of
the algorithms in solving computationally expensive, moderate and cheap bud-
get optimization problems. This evaluation provide the researchers an insight
of selection of an appropriate algorithm depending on what is known about the
optimization problem in terms of evaluation budget, dimensionality and function
structure.
The rest of the paper is organized as follows: Sect. 2 provides a brief descrip-
tion of the selected human learning principles inspired PSO variants. In Sect. 3,
the numerical assessment of the algorithms is presented, including experimen-
tal setup, procedure for evaluating the algorithms’ performance and discussion
of the results. Section 4 summarizes the main conclusions from this study, and
suggests possible extensions for further performance improvements.
Exploring the Variety of Random
Documents with Different Content
So, when the hatter reached the house of his friend, he found him in
great perplexity over this matter, for each of those to whom the
Westland children had been assigned had some special reason for
asking to be excused the service; and when he saw Drayton, he
made up his mind that he had come to him on a similar errand.
"I know what thou past come to say to me, friend, for thou art not the
first visitor I have had concerning this business. Of course, Dame
Drayton is fearful for her own children, and hath sent thee to say she
cannot take these, though—"
"Nay, nay; the word of the Lord to my wife is, that these little ones
should not be separated the one from the other, and she desires me
to say to thee that she would prefer to have them all, an it so please
thee."
"She will take all these children!" exclaimed the Quaker in a tone of
astonishment.
"Even so, friend; for she deems it but adding to their burden of
sorrow at this time to be parted the one from the other."
"And what sayest thou to this?" asked the other, looking keenly at
the hatter; for he was not a wealthy man, but had to work hard for
the maintenance of his family, and to add thus to his burden was no
light matter.
"I can but follow the word of the Lord in me, and that is that I take
these little ones until their parents can claim them at my hand."
"Be it so, then; and the Lord bless thee in thy work, for thou hast
lifted a heavy burden of care from my mind anent this matter. I have
chosen a discreet messenger to bring them from their home, lest one
of us being known should draw the attention of the authorities to
what we were doing, and that might end in our being lodged in gaol
with our brother Westland."
"But how shall I know this messenger?" asked the hatter. "I can go at
once to Triggs' Stairs and meet him."
"Nay, it is a woman who hath chosen this difficult service; and if thou
art in doubt concerning who it is, by reason of other passengers
being near, ask her the way to the Dyers' Garden; for by that signal
was she to know to whom she might deliver the children."
"I will not fail thee," said the hatter. "And when I have taken charge of
these little ones, I will bid her come to thee and give a due account
of how she hath sped on her errand."
And, saying this, Master Drayton bade his friend farewell, and went
at once to the waterside, where he feared the messenger would be
waiting for him.
CHAPTER II.
DAME DRAYTON.
THE Thames in the reign of Charles the Second was the great
highway of traffic for the city of London. There were no steamboats,
it is true, but watermen, duly licensed by the city authorities, and
wearing badges,—much as cabmen do at the present time,—were
always ready with their boats to take passengers wherever they
might want to go; then there were wherries, and splendidly
decorated barges for pleasure parties; so that the river was always a
scene of busy traffic, and especially towards dusk on a summer
evening, for then people would be returning home, or hastening to
embark; so that the time had been well chosen for the coming of the
Westland children, for they were more likely to escape observation
now than earlier in the day.
Triggs' Stairs was a well-known landing-place, not very far from his
own home; and the hatter went by the shortest cuts, through the
busy narrow streets leading to the river, for fear of keeping the
messenger waiting, and thus attracting the attention of watermen
and passengers alike.
But just as Master Drayton reached the top of the landing-stairs a

boat touched the platform below, which the hatter felt sure had
brought those he was seeking. The children were neatly clad, but
there was a sad woe-begone look in their faces, and two of them
seemed to shrink behind the young woman who sat between them.
She too looked anxious, until she caught sight of the hatter, and then
she seemed to gain more confidence, and led the children up the
steps as briskly as their wet and dangerous condition would permit.
"Thee are sent to us by our brother Staples," she said, almost before
the question of identification could be asked.
"Yea, I am here to take charge of the little ones; but thou wilt come
and see my wife, and tell her what is needful to be told," returned
Master Drayton; for he noticed that only a very small bundle had
been brought with them, and this was carried by the elder girl. She
was about thirteen, he judged, and singularly like her father, as he
had seen him a day or two before, when he stood in the court of the
Lord Mayor, and was condemned to lose his ears, and then be
transported as an obstinate schismatic, dangerous to the king and
his authority.
It was not the habit of Quakers to talk in the streets, and so they
walked towards Soper Lane, which was close to the river, without
asking any further questions, for fear of being overheard by some
one passing.
Deborah opened the street door, and received them with a smile of
welcome, as she explained that her mistress was in her own room;
which Master Drayton knew how to interpret, and went himself to tell
her the children had come.
"And none have dared to make them afraid, since the Lord had them
in His keeping," said his wife with a pleasant smile; and she
hastened to the keeping-room to welcome these strangers to their
new home.
"My own little girls, who are to be your sisters, you know, were
obliged to go to bed, they were so sleepy, but you will see them in
the morning;" and as she spoke she kissed each of the shy,
frightened little strangers, putting an arm around each, while she
spoke to the Friend who had brought them.
"They were hiding in the cellar when I reached the house; for it
seems that our brother hath given great offence to his neighbours by
his plainness of speech when he preached and denounced their
wickedness, and so they had revenged themselves upon him, by
well-nigh stripping his dwelling as soon as he and his wife were
taken to prison. Even the clothes seem to have been stolen, for I
could find none but these," she said, touching the little bundle that
had been placed on the table.
"I think the soldiers took some of the things," said the elder girl at this
point; "but mother had said, 'Thee stay with thy sisters in the cellar,'
just before they dragged her away, and Dorothy was so frightened
when we heard the people running up and down stairs, that I could
not go and see what they were doing."
"That was wise of thee, dear child," said Dame Drayton with a sigh;
for she could not help wondering what would happen to her own
darlings if she and her husband should ever fall into similar trouble.
Sometimes it seemed impossible that they could long escape
suspicion, and then anyone might denounce them who happened to
bear them any ill-will.
The messenger who had brought the children did not stay long, for
the streets of London were no fit place for a woman after dusk, even
though she might be staid and discreet; and so, as soon as the
necessary particulars had been given, Master Drayton put on
another hat and coat, to go with her to the Friend who had
undertaken to manage the affair for the committee of suffering.
While her husband was gone, Dame Drayton took the children to the
little bedroom she had prepared for them near her own, and the
nervous, frightened manner of the two younger girls fully justified
what her fears had been concerning them.
They clung to their elder sister, trembling even at the kind attentions
of their friend, lest she should attempt to tear them from this last
protector.
"Thee will let us sleep together," said the little mother, as she took a
hand of each of her younger sisters, and led them upstairs. "We are
not hungry now, only a little tired with the fright," she explained,
when Dame Drayton would have had supper brought into the
keeping-room for them.
"Certainly ye shall sleep together, and to-morrow I hope we shall

learn to know each other better;" and she shut the children in to
themselves, for she could see that it would be kinder to leave them
now, than to press any attentions upon them, or to ask them any
further questions. Before she went to bed, however, she gently
opened the door and looked in, but found to her relief that they were
sound asleep in each other's arms; and they did not rouse the next
morning until all the house was astir, and the sun peeping in at their
windows.
"This is Bessie Westland, and these are her little sisters, Rose and
Dorothy," said Dame Drayton the next morning, introducing the new-
comers to her own children and the family assembled at the
breakfast table.
One of the apprentices had just raised his horn of small ale to drink,
but at the name of Westland he paused, and looked first at the new-
comers, and then at his companion; for the name of Westland had
been heard of a good deal during the last few days, and the lads
were not likely to forget it.
The hatter noticed the look that passed between the two boys, and it
did not tend to make him feel more comfortable; for although it was
known that he was a strict and godly citizen, the fact of his being a
Quaker he desired to keep secret as far as possible, but he feared
now that the coming of these children might be the means of its
discovery.
Dame Drayton had also noticed the surprised looks in the lads'
faces; but she felt sure they might be trusted not to mention what
they had heard out of the house, for they were steady, quiet, reliable
lads, and their occupation kept them out of touch with many of the
more turbulent of their class. Their parents were steady God-fearing
people; and so Dame Drayton put aside all fear of mischief coming
to them through the apprentices.
The children were naturally shy of each other at first; but by degrees
this slipped off like a garment there was no further need to use, and
the first question Bessie asked was about her mother and father.
"When can I go and see them, Martha Drayton?" she asked.
There was no disrespect in the girl's tone; but she came of a more
stern and uncompromising family of Quakers, and would have
looked upon it almost as a sin to use any title of courtesy, however
much she might revere the individual. Dame Drayton knew all this,
but it came upon her with something like a shock, to be addressed
as "Martha Drayton" by this child, and she paused for a minute
before answering her question.
Then she said, slowly and cautiously, "Dear child, thou hast been
placed in our care by the committee of suffering. They will nathless
see to it that ye see your father in due time; but thou must not run
into needless danger, or bring suspicion upon this household."
"Art thou ashamed of being a Quaker, then, as our enemies call us?"
asked the girl rather severely. "To tremble and quake because of sin
was a mark that we were children of the Highest, my father said, and
should we be ashamed of that?"
"Nay, nay, we should be unworthy of our high calling if we were to

despise the work of the Spirit in our hearts; but dost thou not see,
Bessie, that if we were to prate in the streets of these things, we
should bring trouble and sorrow upon those whom God hath given
us to protect?"
"But the trouble and sorrow would be good for them an it came to
them," said Bessie Westland.
"Even so; but if I could not offer thee and thy sisters a safe abiding-
place now, the trouble and sorrow of thy father and mother would be
increased tenfold."
The girl loved her father very dearly, and would have suffered
anything herself to lessen his affliction, and so this view of the matter
touched her a little; and Dame Drayton took this opportunity of
pressing upon her the need of caution.
"We are not called to raise up to ourselves enemies needlessly. It is

only when some truth is to be held firmly and unflinchingly, that we
may thus brave the law and the mob who alike are against it."
"But my father held that it was the duty of a true friend of sinners to
preach the truth to them at all times, whether they would hear, or
whether they would forbear," said Bessie after a minute or two.
"Then, my dear child, if that was the voice of God to him, he could do
no other than obey it, and God hath honoured him in calling him to
witness to that truth. If the same word came to me, I too must obey;
but the voice of the Spirit in my heart was, that I should shield and
protect thee and thy sisters, and thus comfort the heart of those
called to suffer for His name's sake."
"But—but if thee art a true Friend, would not the word of the Lord be
the same to thee as to my father?" said Bessie after a pause.
"Nay, that is where thee makest so grave a mistake," said Dame

Drayton, sitting down by Bessie's side, and drawing little Rose close
to her. "The Lord hath a word of guidance for each if we will but
listen and obey it, without seeking to follow what He may say to
another. See now, He hath made me the mother of tender children,
and given to thee the care of little sisters, which is next in honour to
that of being their mother. Now His word to us will be in accord with
this, to guide and direct us in our duty, how to walk before them in
love."
"But my mother—?" began Bessie.
"Thy mother is a brave and true Friend, following the word of the
Lord, I doubt not," said Dame Drayton quickly; "but because she did
that which the Lord, bade her do, it doth not follow that thee should
do the same, for the voice of the Spirit may have altogether another
word for thee, and thou must listen to that word and follow it, though
it lead thee in the way thou wouldest shun. Just now, thou art longing
to proclaim to all London that thou art of the despised sect of
Quakers, and by this thou wouldest bring grave trouble upon all this
household, for the Lord Mayor would not send to arrest a girl like
thee, but the man and woman who harboured thee, and so we
should be sent to the Bridewell, and thou and my own little ones
become an added burden to our brethren."
"Would they not send me to prison?" said Bessie, in a disappointed

tone.
"I trove not; though King Charles may profess to think men and
women are plotting against his throne, he would scarcely accuse a
child like thee, and so thou and thy sisters would but be cast forth
upon the world again. Wilt thou try to think of this, Bessie; and to
remember that the Lord ever speaks to us of the duty that lies
nearest to our hand, if we will but listen and obey, instead of seeking
to follow the word He may have given to another? This is how so
many mistakes are made, dear child. We think that the word spoken
to another must be for us also, and so our ears are deafened to the
true message that the Spirit is trying to make us hear and
understand."
"But dost thou not think my father obeyed the voice in his heart?"
asked Bessie quickly.
"Yea, verily, dear child. Nought but the strength that God alone can
give can help even a Friend to bear testimony to the truth before
such cruel enemies; but dost thou not see that, while some are
called to be martyrs for the truth, others are commanded to take up
the cross of everyday life, and bear it meekly and patiently, though it
lead not to such honour and renown as the martyr may claim? This
is what we are called to, dear child. Thou and I must take care of the
little ones at home, not denying our faith if any ask us concerning it,
but seeking not to thrust it before the eyes of men; content to be
unnoticed and unknown, but ever listening to the voice that will not
fail to make itself heard in our hearts, if we will but listen with a
simple mind."
Bessie bowed her head, but she was only half convinced of the truth
her friend had spoken. Her father had declared again and again that
they had no right to sit calmly doing the everyday work of life, while
sinners were perishing for lack of the word of life.
He had not scrupled to denounce his neighbours who went to church

as formalists and hypocrites, and even in the church itself had stood
up and warned parson and people alike, telling them that God could
be worshipped in the open fields, in the house or shop, better than in
a steeple-house; and he had gathered crowds around him in the
fields beyond Southwark, and taught them the truth as he had
received it from the lips of George Fox, the founder of their Society.
He was a true and ardent disciple of Fox, counting nothing dear so
that he might proclaim the truth, the whole truth—as he thought—for
in the tenacity with which he held to the little bit he had been able to
grasp, he failed to see that he could not grasp the whole. That those
whom he denounced so unsparingly also held the truth as they
perceived it, or at least another facet of the precious gem, casting its
inspiring light upon them, was dark to him.
This had not been heeded by the authorities at first, and Westland,
like many another earnest man, was allowed to preach and teach
sinners the error of their ways, and warn them of the wrath to come.
For to make men tremble and quake, and cry to God for mercy
through the Lord Jesus Christ, was the object of all the Quakers'
preaching, and the term "Quaker" had been given them in derision
on account of this.
For a time these people had been allowed to follow their own way
without much interference from the authorities; but their unsparing
denunciation of vice and wickedness, whether practised by rich or
poor, doubtless raised the resentment of the king, though a political
reason was the one put forward for their persecution. The safety of
the throne, it was pretended, called for the suppression of these
illegal meetings, as sedition was being taught under cover of religion.
So Westland was an early victim, and suffered the loss of his goods,
for everything he possessed had to be sold to pay the fine inflicted
upon him. But so far from deterring him from doing what he
conceived to be his duty, this did but make him the more determined
to teach and preach upon every occasion possible.
The next time, a short term of imprisonment, and one ear was cut off
by way of punishment. But almost before the place was healed he
was preaching again, and denouncing steeple-houses, and those
who put their trust in them.
This time the authorities were determined to silence him, and so he

had been condemned to lose his other ear, and then be sent as a
slave to one of His Majesty's plantations in America, and all London
was ringing with the name of Westland, and the punishment that had
been dealt out to him as an incorrigible Quaker.
CHAPTER III.
AUDREY LOWE.
"MOTHER, am I truly and verily bound 'prentice to Master Drayton?"

asked one of the lads when he went home that night.
His mother was a widow, and lived a mile or two from Soper Lane,
and moreover was so busily employed in lace-making all day that
she heard very little of what went on around her unless her son
Simon brought home news that he had heard during the day, or on
his way home at night, so that his next question as to whether she
had heard the name of Westland only made the widow shake her
head as she counted the threads of her lace to make sure that she
was not doing it wrong.
"Has Master Drayton taken another apprentice of that name?" asked

his mother, not pausing in her work to look at her boy's face, or she
would have seen a look of horror there as he answered quickly—
"It isn't quite so bad as that, mother."
"What do you mean, Sim?" she asked. "Dame Drayton is our Dame
Lowe's own sister, and a godly woman, I have heard, as well as her
husband—godly and charitable as the parson himself," added the
widow, "or he would not have taken thee to learn the trade and
business of a hatter without price, merely because I was a widow
known to the parson and his wife."
"Oh yes, they are godly and kind; but did I ever tell you, mother, that
one of the rules of the workshop is that we shall not speak more than
is needful?"
"And a very wise rule too, if proper work is to be done," said his
mother quickly.
"That may be; but you have heard of the Quakers, mother?"
"Oh yes, a set of infidel people who speak against the king and the
church, and rebel against all law and order. A most pestilent and
unruly people, I have heard; but surely—"
Sim folded his arms and leaned upon the table, the guttering candle
lighting up his face so that his mother could not fail to see the fright
and horror depicted there as he said—
"I believe Master Drayton is a Quaker, mother."
"Nay, nay, Sim; 'tis a thing impossible. Dame Lowe told me her sister
was a godly Puritan like ourselves; more stiff in her opinions
altogether than she and the parson, for Dame Drayton had
counselled that he should give up the church rather than use the
new prayer-book, since he could not believe and accept all that was
taught in it, and—"
"It would have been a bad case for us if Parson Lowe had refused to
conform to the new rule, like so many did," interrupted Sim at this
point.
"Yes, it would; and a worse case for his wife and children, for they
might have starved by this time instead of living in a comfortable
house, with money to help the poor as well as themselves, and I
must say, since these changes came, parson has been even more
strict and attentive to his duties, though none could complain of him
before."
"But what has that to do with Master Drayton being a Quaker?"

asked the lad, a little impatiently.
"Why, cannot you see, Sim, that all the family are of so godly a sort,
that they would not be likely to take up with all the unruly and wild
notions that these pestilent people teach?"
"I don't know what the Quakers teach, but I know that one fellow
named Westland has had his ears cut off, and now three girls of the
same name have come to live with us in Soper Lane. If they were
not Quakers themselves, would they take in a disgraced Quaker's
children?"
The widow looked at her son for a minute as if she thought this
argument was unanswerable, but after a minute's pause she said—
"They are kind and godly folk, you say, and so it may be they are not
of this pestilent sect that hath been suffered to spring up to speak
evil of dignities, though they succour these children."
But although the widow said this, she decided to go and see her
friend the vicar, and have a word with his wife too if it was possible,
for it would never do to let her son—her only child—become
contaminated, even though he was being taught his trade without the
cost of a penny to herself.
During the rest of the evening she asked the boy a good many
questions about his work in Soper Lane, and the ways of the
household, but there seemed no fault to be found with anything,
though doubtless the household was ruled strictly, as most Puritan
homes were. Still, what Sim had told her about Westland made her
uncomfortable, and before she went to bed she decided to go to the
vicarage the next morning as soon as Sim had started to Soper
Lane, and doubtless the parson or Dame Lowe would be able to
explain everything, and set her fears at rest.
But when she went, she heard from the maid-servant that the vicar
was ill in bed with a bad cold, and that she would have to wait a little
while to see her mistress.
"Then I will wait," said the widow, for she had scarcely been able to
sleep for thinking of the peril to which her boy might be exposed if it
should be true that his master was a Quaker, as he suspected.
Dame Lowe would be able to set her fears at rest, she hoped, and
the moment the lady entered the room the widow began a recital of
her trouble.
At first she was too full of what she had to say, and how frightened
she had become, to pay much attention to the lady herself, but after
a minute or two she noticed that she was trembling, and her face
had become as white as the lace ruffle she wore round her neck.
"I—I am afraid you are ill, madam," said the widow, stopping short in
her recital, and looking hard at the lady.
"Just a little faint. I have been anxious about the vicar, you see—but
go on with your story. What did your boy say was the name of these
children?"
The lady spoke eagerly, and looked almost as frightened and

anxious as her visitor, though she was careful not to let her know
that it arose from the same cause, and spoke of the vicar's illness as
being a little alarming, and having upset her.
"But tell me about those children who have gone to live at the house
in Soper Lane. Who did you say they were?"
"Well, now, Sim couldn't be quite sure, of course; but he is a careful

lad, and he says there was a Quaker of the same name had his ears
cut off for heresy only a day or two ago. Of course, I told Sim that his
master, being a godly and charitable man, might have had
compassion on these witless children without being himself a
Quaker."
"Then it is suspected that Dame Drayton and her husband are both
Quakers. Is that what you mean, Tompkins?"
The lady's mood had changed during the last few moments, and she
looked hard at the widow and spoke in a severe tone, as though
such a charge as she brought was not to be believed.
"I—I don't know what to think," said the widow. "Of course, as
Madam Drayton is your sister she could scarcely be infected with
such heresy as these wild Quakers believe."
"I trow not, indeed. My sister was brought up in a true godly fashion,
but the same charity that moved Master Drayton to take Simon as an
apprentice without fee or reward, because you were a poor widow
known to me to have lost so much by the plague and the great fire,
may have moved her to help these poor children, if no one else
would do it."
"Then you think my Sim would be quite safe there?" said the widow
in a deprecating tone.
Madam Lowe looked surprised at the question. "Why should he not

be safe?" she asked. "You told me the other day that he was learning
his trade very well, and had certainly improved in his manners."
"But—but if his master should be a Quaker it would be little better

than sending him where he would catch the plague, this new plague
of heresy that is abroad, and for my Sim to turn Quaker would be
worse than losing the others by the pestilence." And at the thought of
all the sorrow and suffering she had endured through this scourge,
the Widow Tompkins fairly burst into tears.
"There, don't cry—I am sure you are frightening yourself for nothing.
I know my sister to be a gentle godly woman; no more like the wild
fanatic Fox than you are. She attends her own parish church as you
do, and therefore you may rest content that Sim is safe."
The widow allowed herself to be comforted by this assurance.
"It's all we've got to hold fast by in the way of knowing what to
believe, for there's been so many changes in religion, as well as
other things the last few years, that simple folk like me, who have no
learning, hardly know what they ought to believe sometimes; but to
have Sim turn Quaker would just break my heart, when I was looking
forward to a little comfort after all my trouble."
"Oh, Simon will be a good son, and a comfort to you, I have no

doubt," said the lady, rising to dismiss her visitor. "Take care that he
is at church by seven o'clock next Sunday morning, for the vicar is
going to catechise all the lads and wenches of the parish, and it will
not do for Simon to be absent from his place in the chancel."
"My son will be early. I am glad the vicar is going to give them a
wholesome reminder of what they ought to know and do, as
respectable citizens and members of the Church of England. It will
help to stop this wild Quaker heresy, I trow."
The lady smiled and nodded her assent; but she was too impatient
for her visitor to go to make any verbal reply to this, and as soon as
she had closed the street door she went upstairs to a little room
where a girl sat sewing.
"What is the matter, mother?" she asked as the lady seated herself,
and buried her face in her hands.
For a minute or two the lady sat thus, and when she removed them
she was looking white and anxious.
"Oh, Audrey, I wish I had never persuaded your father to—" But
there she stopped, for the girl's wondering eyes told her she was
speaking of things she had long ago resolved to bury in her own
heart. "My dear, I want you to go and see your Aunt Martha," she
said quickly.
"Aunt Martha?" repeated the girl in a tone of wonder.
"Have you forgotten her, Audrey? It is not so many years since you
saw your aunt."
"But I thought you said she died in the time of the first plague," said
the girl, still looking at her mother with a puzzled expression in her
face, as if trying to recall some memory of the forgotten relative.
"Nathless you will remember her again when you see her," said
Dame Lowe, in answer to her daughter's puzzled look. "I want you to
go to her this afternoon, and say that the Widow Tompkins, who is
the mother of one of her husband's 'prentice lads, hath been here
with a tale about Quakers that is disgraceful to any godly
household."
"The Widow Tompkins is always in a fright about something,"

returned Audrey slightingly. "What did she say about the Quakers
and my Aunt Martha? What has she to do with them?"
"Nothing, I wot; but Master Drayton, her husband, is not always so
discreet as he should be, and Simon hath brought home some tale
to his mother about Quaker children being harboured in the house.
Your aunt ought to be told that this is known, and will soon become
the talk of the town if they are not sent away."
"Would you like me to bring the children here, mother, to save Aunt
Martha the trouble of them?"
The lady looked at her daughter, aghast with horror at the proposal.
"Audrey, you must not speak so lightly of such matters. For us to be

suspected of any touch with these Quakers would mean ruin, and we
might be thrown out of house and home, like so many clergymen's
families have been, for it is known that your father always felt they
were unjustly treated, though he signed the declaration that saved us
from being turned into the streets like beggars. This is why I want
you to go and see your aunt to-day, for if people think the Draytons
are Quakers they may suspect us next. Oh dear! why will people go
wild about religion like this man Fox? It is sure to bring disgrace
upon somebody. As if the fire and the plague had not caused misery
enough in London, they must now begin making fresh trouble about
religion, just as I hoped things were getting more settled and
comfortable."
"Mother dear, do not look so troubled about this. Surely God can
take care of us and of London too. How is my father now?"
"Not much better, and I do not want him to hear about this, or it will
make him anxious and unfit to catechise the children in church on
Sunday morning. Now, Audrey, we shall have dinner at eleven, and
then I should like you to go to your aunt, who lives in Soper Lane,
and you can see for yourself who these Quaker children are, and
find out whether your aunt still goes to the parish church, for I hear
these fanatics call it a steeple-house, and will by no means join in
the prayers as they are set forth in the prayer-book."
The errand in itself was not at all to the taste of a girl like Audrey; but
the dim recollection she had of her aunt made her desirous of seeing
her once more, and she could only wonder how and why it was that
her mother had been silent concerning Dame Drayton, for they had
but few relatives, and Audrey herself was the only child now. Two
had died during the great plague, and she could only suppose that it
was because her aunt lived in the City, and her mother still had a
lingering dread of the plague returning, that she had not heard this
aunt spoken of for so long a time.
Although they lived within easy walking distance of the City, and she
knew her father sometimes went there on business, she did not
remember ever having seen it herself, for they lived in the
fashionable suburbs of St. Martin-in-the-Fields, and generally went to
walk in the westerly direction among the fields and green lanes. The
parish did not wholly belie its name as yet, for they lived in the midst
of the open country, although so near London. Her old nurse was to
walk with her, and call for her at Soper Lane at four o'clock, that they
might reach home before sunset, for although their way lay through
the best and most fashionable thoroughfares in the town, they were
by no means safe from footpads. Although the Strand was the
residence of many of the nobility, and Fleet Street had most of the
best shops lining its footway, these were generally shunned by
travellers after sundown—unless they were on horseback, armed
and attended by two or three stout serving-men.
So Audrey and her nurse set out on their journey about half-past
eleven, and less than a mile from her own home, Audrey was in a
place altogether new to her.
"I wonder why we have not come this way to walk before," said the
girl looking round at the handsome houses in the Strand.
But the 'prentice lads in the front of their masters' shops in Fleet
Street, all eager to press their wares upon their notice, were not a
pleasant feature of the scene to Audrey.
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade
Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.
Let us accompany you on the journey of exploring knowledge and

personal growth!
textbookfull.com

Download Swarm Evolutionary and Memetic Computing 6th International Conference SEMCCO 2015 Hyderabad India December 18 19 2015 Revised Selected Papers 1st Edition Bijaya Ketan Panigrahi ebook All Chapters PDF

Uploaded by

Copyright:

Available Formats

Download Swarm Evolutionary and Memetic Computing 6th International Conference SEMCCO 2015 Hyderabad India December 18 19 2015 Revised Selected Papers 1st Edition Bijaya Ketan Panigrahi ebook All Chapters PDF

Uploaded by

Document Information

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

Download Swarm Evolutionary and Memetic Computing 6th International Conference SEMCCO 2015 Hyderabad India December 18 19 2015 Revised Selected Papers 1st Edition Bijaya Ketan Panigrahi ebook All Chapters PDF

Uploaded by

Copyright:

Available Formats

Download the Full Version of textbook for Fast Typing at textbookfull.

Swarm Evolutionary and Memetic Computing 6th

Download More textbook Instantly Today - Get Yours Now at textbookfull.com

Swarm Evolutionary and Memetic Computing 4th International

Big Data Benchmarking 6th International Workshop WBDB 2015

Smart Health International Conference ICSH 2015 Phoenix AZ

Information and Communications Security 17th International

Distributed Computer and Communication Networks 18th

Cloud Computing and Services Science 5th International

High Performance Computing and Applications Third

Service Oriented Computing ICSOC 2015 Workshops WESOA

ISSN 0302-9743 ISSN 1611-3349 (electronic)

Library of Congress Control Number: 2016956619

LNCS Sublibrary: SL1 – Theoretical Computer Science and General Issues

© Springer International Publishing AG 2016

Printed on acid-free paper

This Springer imprint is published by Springer Nature

December 2015 Bijaya Ketan Panigrahi

Steering Committee Chair

Special Session Chairs

International Advisory Committee/Technical Review Committee

Almoataz Youssef Abdelaziz, Egypt Carlos A. Coello Coello, Mexico

Ferrante Neri, Finland Namrata Khemka, USA

Self-adaptive Ensemble Differential Evolution with Sampled Parameter

Empirical Assessment of Human Learning Principles Inspired PSO

Visual Cryptography Based Lossless Watermarking for Sensitive Images . . . . 29

Cohort Intelligence and Genetic Algorithm Along with AHP

Design, Construction and Analysis of Model Dataset for Indian Road

A Hybrid EMD-ANN Model for Stock Price Prediction. . . . . . . . . . . . . . . . 60

Development of Back Propagation Neural Network (BPNN) Model

An Improved Quantum Inspired Immune Clone Optimization Algorithm . . . . 84

Diagnosis of Parkinson Disease Patients Using Egyptian Vulture

Variance Based Particle Swarm Optimization for Function Optimization

Analysis of Next-Generation Sequencing Data of miRNA for the Prediction

Genetic Algorithm Based Speed Control of Electric Vehicle

Multi-objective Power Dispatch Using Stochastic Fractal Search

Particle Swarm Optimization for the Deployment of Directional Sensors . . . . 167

Region Based Multiple Features for an Effective Content Based Access

Robot Workcell Layout Optimization Using Firefly Algorithm . . . . . . . . . . . 188

Particle Swarm Optimization Based on the Winner’s Strategy. . . . . . . . . . . . 201

Black Hole Artificial Bee Colony Algorithm . . . . . . . . . . . . . . . . . . . . . . . 214

A Gravitational Search Algorithm for Energy Efficient Multi-sink

Optimum Clustering of Active Distribution Networks Using Back Tracking

Energy Efficient Clustering for Wireless Sensor Networks:

Hybridizing Cuckoo Search with Bio-inspired Algorithms for Constrained

A Hybrid Genetic Algorithm Using Dynamic Distance

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287

Nandar Lynn1, Rammohan Mallipeddi2,

Abstract. In literature, empirically and theoretically, it has been

Keywords: Differential evolution Global optimization Parameter

© Springer International Publishing AG 2016

Vi;G ¼ Xbest;G þ F:ðXr1i ;G Xr2i ;G Þ ð2Þ

Vi;G ¼ Xbest;G þ F:ðXr1i ;G Xr2i ;G Þ þ F:ðXr3i ;G Xr4i ;G Þ ð3Þ

Vi;G = Xr1i ;G þ F:ðXr2i ;G Xr3i ;G Þ ð4Þ

Vi;G = Xr1i ;G þ F:ðXr2i ;G Xr3i ;G Þ þ F:ðXr4i ;G Xr5i ;G Þ ð5Þ

Ui;G = Xi;G þ K:ðXr1i ;G Xi;G Þ þ F:ðXr2i ;G Xr3i ;G Þ ð6Þ

In Eq. (7), the crossover probability,CR 2 ½0; 1 is a user-speciﬁed constant that

2.2 Literature Review

(k = 1, 2, .., K; G > LP)

Vi;G ¼ Xi;G þ Fi;G :ðXpbest;G Xi;G Þ þ Fi;G :ðXr1i ;G Xr2i ;G Þ ð11Þ

In Eq. (7), the crossover probability,CR 2 ½0; 1 is a user-speciﬁed constant that