Applications of Bat Algorithm and its Variants

182 Pages·2021·6.198 MB·English

by  Nilanjan Dey, V. Rajinikanth

Springer Tracts in Nature-Inspired Computing
Nilanjan Dey
V. Rajinikanth   Editors
Applications
of Bat
Algorithm and
its Variants
Springer Tracts in Nature-Inspired Computing
Series Editors
Xin-SheYang,SchoolofScienceandTechnology,MiddlesexUniversity,London,
UK
Nilanjan Dey, Department of Information Technology, Techno India College of
Technology, Kolkata, India
Simon Fong, Faculty of Science and Technology, University of Macau, Macau,
Macao
The book series is aimed at providing an exchange platform for researchers to
summarize the latest research and developments related to nature-inspired
computing in the most general sense. It includes analysis of nature-inspired
algorithms and techniques, inspiration from natural and biological systems,
computational mechanisms and models that imitate them in various fields, and
the applications to solve real-world problems in different disciplines. The book
series addresses the most recent innovations and developments in nature-inspired
computation,algorithms,modelsandmethods,implementation,tools,architectures,
frameworks, structures, applications associated with bio-inspired methodologies
and other relevant areas.
The book series covers the topics and fields of Nature-Inspired Computing,
Bio-inspired Methods, Swarm Intelligence, Computational Intelligence,
Evolutionary Computation, Nature-Inspired Algorithms, Neural Computing, Data
Mining, Artificial Intelligence, Machine Learning, Theoretical Foundations and
Analysis, and Multi-Agent Systems. In addition, case studies, implementation of
methods and algorithms as well as applications in a diverse range of areas such as
Bioinformatics, Big Data, Computer Science, Signal and Image Processing,
Computer Vision, Biomedical and Health Science, Business Planning, Vehicle
Routing and others are also an important part of this book series.
The series publishes monographs, edited volumes and selected proceedings.
More information about this series at http://www.springer.com/series/16134
Nilanjan Dey V. Rajinikanth
(cid:129)
Editors
Applications of Bat
Algorithm and its Variants
123
Editors
Nilanjan Dey V.Rajinikanth
Department ofInformation Technology Department ofElectronics
Techno India Collegeof Technology andInstrumentation Engineering
Kolkata, West Bengal, India St.Josephs College ofEngineering
Chennai, Tamil Nadu,India
ISSN 2524-552X ISSN 2524-5538 (electronic)
SpringerTracts inNature-Inspired Computing
ISBN978-981-15-5096-6 ISBN978-981-15-5097-3 (eBook)
https://doi.org/10.1007/978-981-15-5097-3
©TheEditor(s)(ifapplicable)andTheAuthor(s),underexclusivelicensetoSpringerNatureSingapore
PteLtd.2021
Thisworkissubjecttocopyright.AllrightsaresolelyandexclusivelylicensedbythePublisher,whether
thewholeorpartofthematerialisconcerned,specificallytherightsoftranslation,reprinting,reuseof
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmissionorinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilar
ordissimilarmethodologynowknownorhereafterdeveloped.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfrom
therelevantprotectivelawsandregulationsandthereforefreeforgeneraluse.
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
authorsortheeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinor
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations.
ThisSpringerimprintispublishedbytheregisteredcompanySpringerNatureSingaporePteLtd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Preface
Swarm intelligence basedmetaheuristicalgorithmsareextensivelyimplementedto
solve a variety of real-world optimization problems due to their adaptability and
robustness. Bat Algorithm (BA) is one of the most successful swarm intelligence
procedures developed in 2010 and extensively used in various optimization tasks
for a decade. The mathematical model of BA is quite straightforward and easy to
understand and enhance, when compared to other swarm approaches. Hence, the
BA has attracted the attention of researchers, who are working to find the optimal
solutions in a variety of domains, such as N-dimensional numerical optimization,
constrained/unconstrainedoptimizationandlinear/nonlinearoptimizationproblems.
Alongwiththetraditional BA,theenhancedversionsofBAarealsoconsideredto
solve a variety of optimization problems in science, engineering and medical
applications. This book highlights the essential concepts of the traditional BA
algorithmand itsrecentvariantsandalsoitsapplicationtofindanoptimalsolution
for a variety of real-world engineering and medical problems.
The reason for this book is to help beginners and researchers in understanding
thebasicconceptsandrecentadvancementsinbatalgorithmstoenhancetheresults
inexistingtechnologicaltrendsanddesignchallenges.Thisbookisconcernedwith
supportingandenhancingtheutilizationofbatalgorithmsisavarietyofreal-world
optimization problems ranging from numerical optimization to medical data anal-
ysis. This work presents a well-standing forum to discuss the characteristics of the
traditional and recent versions of the bat algorithm in various fields. The book is
proposed for professionals, scientists and engineers, who are concerned about the
methods using the bat algorithm. It provides an outstanding foundation for
undergraduate and postgraduate students as well. It has several features, including
an outstanding basis of the bat algorithm analysis, and it includes different appli-
cations and challenges with extensive studies for systems that have used bat
algorithms.
The book is organized as follows:
Chapter1proposesanewhybridbinaryversionofthebatalgorithmtosolvethe
dominant feature selection problems, which is a vital procedure to implement a
classifier unit. In this work, Bat Algorithm (BA) is integrated with an enhanced
v
vi Preface
version of the Differential Evolution (DE) algorithm. In this work, the BA with its
capacityfor echolocationtoexplorethefeaturespace iscombined with DEandits
ability to converge to the best global solution in the search space. The general
performance of the proposed algorithm is investigated by comparing it with the
originaloptimizersandotheroptimizersthathavebeenusedforfeatureselectionin
the literature. The proposed algorithm and its various optimizers are applied over
datasets obtained from the UCI repository. The results prove the ability of the
proposed algorithm to search the feature space for optimal feature combinations.
Key issues and future research directions are also highlighted in this chapter.
Chapter 2 discusses the multi-objective optimization of engineering design
problems through Pareto based bat algorithm. In multi-objective optimization
problems,sinceadifferentobjectivevalueisgeneratedagainsteachdecisionvector,
the superiority of the solutions over each other is determined by considering the
trade-off among the objective values. In this work, one of the recent metaheuristic
optimization methods based on swarm intelligence, that is, the so-called a Pareto
based bat algorithm inspired by the behaviour of determining the direction and
distance ofanobject using theechoof thesoundcalled theecholocationof batsis
used in order to obtain optimum solutions of multi-objective engineering design
problems. In this regard, a four-bar planar truss, a real-sized welded steel beam as
well as a multi-layer radar absorber are selected as multi-objective engineering
designoptimizationproblems. Onobtaining theresults(optimal designs), potency,
and reliability of the proposed multi-objective Pareto based bat algorithm are
reported.
Chapter3proposesastudyontheBatAlgorithm(BA)techniquetoexaminethe
skin melanoma images. Skin melanoma is one of the major types of cancer in
people from Caucasian race. Due to its consequence, a considerable number of
research works are being proposed by the researchers to develop the probable
computer-based assessment technique for the Skin Melanoma Image (SMI). This
workaimstodevelopandimplementacomputerizedtoolfortheassessmentofthe
SMIbasedontherecentmachinelearningtechnique.Intheproposedwork,theBat
Algorithm(BA)assistedexaminationtechniqueisimplementedtoprocesstheSMI.
Inthiswork,adetailedevaluationofthetraditionalBAandtherecentversionofthe
BAareconsideredtoassesstheperformanceoftheproposedtechnique.Thiswork
considersthevariantsofBA,suchasLevy-Flight(LF),Brownian-Walk(BW)and
the Ikeda-Map (IM) to pre-process the skin melanoma pictures. The pre-processed
SMIs are then processed with the DRLS segmentation approach and the perfor-
mance of the considered BAs are validated by computing the essential Image
Performance Metrics (IPM), and the result of this study confirm that the final
outcome attained with the BW-guided BA offered a better result compared to the
LF- and IM-based techniques. This technique is tested and validated using the
benchmark database existing in the literature.
Chapter 4 discusses thresholding of gray/RGB-scale images using Bat
Algorithm (BA). The essential task in image thresholding is to enhance the infor-
mation in raw image by identifying an optimal threshold. The proposed work
executes a multi-threshold procedure for a class of benchmark images using
Preface vii
Kapur’sEntropy(KE).Manualidentificationofappropriatethresholdisacomplex
taskforhigherthresholdvaluesandhence,thisworkemployedBAtofindthefinest
thresholdforthegray/RGBtestimagery.Inthiswork,testimageswithadimension
of 512 x 512 are considered for the experimental evaluation. This work also pre-
sentsaperformancecomparisonofthetraditionalandtheenhancedBAtoidentifya
suitable methodology to attain better outcomes without compromising the quality.
Chapter5presentsacomparativestudyonbat-inspiredcomputingalgorithmand
its variants in search of Near-Optimal Golomb Rulers (OGRs) for Wavelength-
Division Multiplexing (WDM) systems. Near-OGR sequences can be used as a
channel-allocation scheme to reduce one of the important nonlinear crosstalk gen-
eratedviaFour-WaveMixing(FWM)signalsinopticalWDMsystems.TheOGRs
provide unequally spaced channel-allocation, a bandwidth-efficient scheme, and
then the uniformly spaced channel-allocation methods to minimize the FWM
crosstalk. To explore the search space, the bat-inspired computing algorithm is
hybrid in its simple form with Differential Evolution (DE) mutation and random
walk characteristics. The algorithms solve the two parameters, namely, the length
of the Golomb ruler and total unequally spaced channel bandwidth occupied by
OGRsintheopticalWDMsystems.Theresultsrevealthatthepresentedbat-inspired
computing algorithm and its variants are better than other classical computing
methods such as Extended Quadratic Congruence (EQC) and Search Algorithm
(SA)andnature-inspiredcomputingalgorithms,namely,GeneticAlgorithms(GAs),
and simple Big Bang–Big Crunch (BB-BC) computing algorithm to generate
near-OGRsintermsofthelengthoftheruler,thetotaloccupiedchannelbandwidth,
the Bandwidth Expansion Factor (BEF), the CPU time and the computational
complexity.
Chapter 6 presents a model order reduction methodology based on the Bat
Algorithm (BA). Analytical study of large-scale linear time-invariant systems is a
very tedious and complicated task in the category of real-life optimization prob-
lems. So, simplification procedures for these complex problems are needed. In the
solutiontacticofthiscomplex problem,ModelOrderReduction (MOR)isanovel
concept providing a simpler model than the original one based on mathematical
approximation.Intheliterature,severalmetaheuristicsareemployedtosolveMOR
problem.Inthesamelineoforder,thischapterpresentsatechniquetosolveMOR
problem using modified bat algorithm based on Levy-flight and opposition-based
learning.The concept ofLevy-flight random walk andOpposition-Based Learning
(OBL) is embedded to Bat Algorithm (BA) to avoid local optima trapping and to
enhancetheexploitationandexplorationability.Toevaluatetheperformanceofthe
proposed methodology, it is tested over three different MOR problems with dif-
ferent transferfunctions.Thenumericalandstatistical resultsverifythesupremacy
of the proposed variant in terms of stability of reduced-order systems.
Chapter7proposesamethodologytodetectthetumourfrombrainMRI.Cancer
isadiseasecausedbyanabnormalgrowthofcells.Thisisamenacingdiseasethat
can severely affect the quality of life of individuals. It can also take a toll on the
emotional well-being of the patient along with physical repercussions. The cells
whichformmalignanttumourscanoccurinanypartofthebody,butthebrainisan
viii Preface
area where the chance of survival is minimal if not treated accurately in time.
RadiologistsandoncologistsmakeuseofMRIscans,whichprovideimagesofthe
brain. These images can have different appearances depending on the setting of
pulse sequences of the MRI such as T1-W, T2-W, MPR, DWI and FLAIR. This
chapter focuses on the segregation of tumour region based on sequences such as
T1-weightedand T2-weightedimages ofthebrain.Thesegmentation isperformed
by making use of a fusion of bat and Interval Type-2 Fuzzy C Means Clustering
(IT2FCM) algorithms, which is aimed at simplifying the task of a radiologist.
Chapter 8 discusses a detailed review of signal, speech and image processing
basedonBatAlgorithm(BA).BAisaprominentmetaheuristicapproachbasedon
the hunting mechanism of bats in nature. BA applications are rapidly growing in
many engineering research fields since its invention in 2010. This chapter reviews
thedevelopmentsandapplicationsofthebatalgorithminsignal,speechandimage
processing. In particular, this review focuses on five main research areas for BA
applicationsinsignal,speechandimageprocessing:speechenhancement,adaptive
filtering, image compression, enhancement and thresholding in segmentation. On
theotherhand,thischapteralsoreviewsthenewvariantsanddevelopmentsinBA
applied for the aforesaid applications. The detailed implementation of the algo-
rithms and their objective functions is also presented in this chapter.
Chapter 9 implements the Bat Algorithm (BA) assisted methodology to extract
the tumour section from abnormal brain MRI slices. This work implemented a
heuristicalgorithmbasedexaminationprocedure.Inthis,theBatAlgorithm(BA)is
usedtoimprovethevisibilityofthetumour-pixelsusingKapur’sEntropy(KE).The
improved tumour section is then collected using the Watershed Segmentation
Method (WSM). Later, a comparison linking the tumour portion and available
Ground-Truth-Image (GTI) is executed and the Quality Measures (QM) are com-
putedindividuallyforFLAIRandT2.Duringthisinvestigation,theMRIsliceswith
240(cid:1)240(cid:1)1pixelresolutionareconsideredandthistechniqueisimplementedon
400 MRI slices (FLAIR = 200 + T2 = 200). The average result of QM confirmed
that, the BA-based technique helped to get superior result (Accuracy >95% for
FLAIR/T2modalityslices).Inthefuture,thismethodologycanbeimplementedto
analysethe MRI slices obtained from the hospitals.
Editors
Kolkata, India Nilanjan Dey
Chennai, India V. Rajinikanth
Contents
1 ANewHybridBinaryAlgorithmofBatAlgorithmandDifferential
Evolution for Feature Selection and Classification . . . . . . . . . . . . . . 1
Abdelmonem M. Ibrahim and Mohamed A. Tawhid
2 Multi-objective Optimization of Engineering Design Problems
Through Pareto-Based Bat Algorithm . . . . . . . . . . . . . . . . . . . . . . . 19
Deniz Ustun, Serdar Carbas, and Abdurrahim Toktas
3 A Study on the Bat Algorithm Technique to Evaluate the Skin
Melanoma Images. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Nilanjan Dey, V. Rajinikanth, Hong Lin, and Fuqian Shi
4 Multi-thresholding with Kapur’s Entropy—A Study Using Bat
Algorithm with Different Search Operators . . . . . . . . . . . . . . . . . . . 61
V. Rajinikanth, Nilanjan Dey, and S. Kavitha
5 Application of Bat-Inspired Computing Algorithm
and Its Variants in Search of Near-Optimal Golomb Rulers
for WDM Systems: A Comparative Study . . . . . . . . . . . . . . . . . . . . 79
Shonak Bansal, Neena Gupta, and Arun K. Singh
6 Levy Flight Opposition Embedded BAT Algorithm for Model
Order Reduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Shalini Shekhawat, Akash Saxena, Rajesh Kumar,
and Vinay Pratap Singh
7 Application of BAT Algorithm for Detecting Malignant
Brain Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Adit Kotwal, Rishika Bharti, Mansi Pandya, Harshil Jhaveri,
and Ramchandra Mangrulkar
ix