Application of Evolutionary Algorithms for Multi-objective Optimization in VLSI and Embedded Systems

181 Pages·2015·2.95 MB·English

by  M.C. Bhuvaneswari (eds.)

M.C. Bhuvaneswari E ditor
Application of
Evolutionary Algorithms
for Multi-objective
Optimization in VLSI and
Embedded Systems
Application of Evolutionary Algorithms
for Multi-objective Optimization in VLSI
and Embedded Systems

M.C. Bhuvaneswari
Editor
Application of Evolutionary
Algorithms
for Multi-objective
Optimization in VLSI
and Embedded Systems
Editor
M.C.Bhuvaneswari
ElectricalandElectronicsEngineering
PSGCollegeofTechnology
Coimbatore,TamilNadu,India
ISBN978-81-322-1957-6 ISBN978-81-322-1958-3(eBook)
DOI10.1007/978-81-322-1958-3
SpringerNewDelhiHeidelbergNewYorkDordrechtLondon
LibraryofCongressControlNumber:2014947350
©SpringerIndia2015
Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpart
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation,broadcasting,reproductiononmicrofilmsorinanyotherphysicalway,andtransmissionor
informationstorageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilar
methodologynowknownorhereafterdeveloped.Exemptedfromthislegalreservationarebriefexcerpts
inconnectionwithreviewsorscholarlyanalysisormaterialsuppliedspecificallyforthepurposeofbeing
enteredandexecutedonacomputersystem,forexclusiveusebythepurchaserofthework.Duplication
ofthispublicationorpartsthereofispermittedonlyundertheprovisionsoftheCopyrightLawofthe
Publisher’s location, in its current version, and permission for use must always be obtained from
Springer.PermissionsforusemaybeobtainedthroughRightsLinkattheCopyrightClearanceCenter.
ViolationsareliabletoprosecutionundertherespectiveCopyrightLaw.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexempt
fromtherelevantprotectivelawsandregulationsandthereforefreeforgeneraluse.
While the advice and information in this book are believed to be true and accurate at the date of
publication,neithertheauthorsnortheeditorsnorthepublishercanacceptanylegalresponsibilityfor
anyerrorsoromissionsthatmaybemade.Thepublishermakesnowarranty,expressorimplied,with
respecttothematerialcontainedherein.
Printedonacid-freepaper
SpringerispartofSpringerScience+BusinessMedia(www.springer.com)
Preface
Thisbookdescribeshowevolutionaryalgorithms(EA)suchasgeneticalgorithms
(GA) and particle swarm optimization (PSO) can be used for solving multi-
objectiveoptimizationproblemsintheareaofembeddedandVLSIsystemsdesign.
This book is written primarily for practicing CAD engineers and academic
researchers who wish to apply evolutionary techniques and analyze their perfor-
mance in solving multi-objective optimization problems in VLSI and embedded
systems.
Many real engineering problems have multiple objectives, such as minimizing
cost, maximizing performance, and maximizing reliability. Being a population-
basedapproach,EAaresuitableforsolvingmulti-objectiveoptimizationproblems.
ResearchhasbeencarriedouttoascertainthecapabilitiesofGAandPSOinsolving
complexandlargeconstrainedcombinationaloptimizationproblems.Areasonable
solutiontoamulti-objectiveproblemistoconsiderasetofsolutions,eachofwhich
satisfiestheobjectivesatanacceptablelevelwithoutbeingdominatedbyanyother
solution. Graph theoretic approaches and integer/linear programming have been
used to solve problems in embedded and VLSI system design. GA and PSO may
prove to be a general purpose heuristic method for solving a wider class of
engineeringandscientificproblems.
Organization of the Book
This book provides an introduction to multi-objective optimization using meta-
heuristicalgorithms:GAandPSOandtheirapplicationtoproblemslikehardware/
softwarepartitioninginembeddedsystems,circuitpartitioninginVLSI,designof
operational amplifiers in analog VLSI, design space exploration in high level
synthesis, delay fault testing in VLSI testing, and scheduling in heterogeneous
distributed systems. It is shown how in each case the various aspects of the EA,
namely its representation, and operators like crossover, mutation, etc., can be
separatelyformulatedtosolvetheseproblems.
v
vi Preface
The content of this book is divided into 9 chapters, and each chapter focuses
on one aspect or application of EA to multi-objective optimization. Each of the
chapters deals with experimental results as well as an analysis of the problem
athand.
Chapter1providesanintroductiontomulti-objectiveGAandPSOalgorithms.
TheterminologyofGAandPSOisintroducedanditsoperatorsarediscussed.Also,
thevariantsofGAandPSOandtheirhybridswithhillclimbingareexplained.
Chapter2addressestheproblemofhardware/softwarepartitioninginembedded
systems.Itdiscusseshowmulti-objectiveEAcanbeappliedtohardware/software
partitioning.
Chapter 3 focuses on the problem of circuit partitioning in VLSI. It describes
howthechromosomeisrepresentedinGAfortheproblemofcircuitpartitioning.
Chapter4explainshowthedesignofoperationalamplifiersinanalogVLSIcan
berepresentedasmulti-objectiveoptimizationproblemandhowGAcanbeapplied
tosolvethedesignofoperationalamplifiers.
Chapter 5 describes how multi-objective EA can be applied to the problem of
designspaceexplorationinhigh-levelsynthesisofVLSI.
Chapter6presentsthedesignspaceexplorationofdatapath(architecture)inhigh
levelsynthesis(HLS)forcomputationintensiveapplications.
Chapter7 elaborates evolutionary algorithm driven high level synthesis design
flow:AlgorithmtoRTL.
Chapter 8 deals with delay fault testing in VLSI testing. It describes how the
problemofdelayfaulttestingcanbeformulatedasamulti-objectiveproblemand
delineatesexperimentalresultsforthevariouscrossoveroperators.
Chapter9exploreshowtheschedulinginheterogeneousdistributedsystemcan
be formulated as a multi-objective problem and the applicability of EA to the
problem. Experimental results are provided for four variants of multi-objective
GAandPSO.
Coimbatore,India M.C.Bhuvaneswari
Acknowledgements
I would like to acknowledge the significant contributions of Ms. M. Shanthi,
Dr. M. Jagadeeswari, Dr. S. Jayanthi, Dr. D. S. Harish Ram, and
Dr. G. Subashini, who have contributed much to the research on application of
multi-objective optimization in VLSI and embedded systems as part of their
doctoral research. I would like to thank Dr. Anirban Sengupta for providing the
chapters on Design Space Exploration of Datapath (Architecture) in High Level
Synthesis for Computation Intensive Applications and Design Flow from Algo-
rithm to RTL using Evolutionary Exploration Approach. I am grateful to the
reviewers Dr. P. Navaneethan, Dr. Reza Javaheri, and Ms. Jayashree Saxena
whose insightful comments enabled us to improve the quality of the book signif-
icantly.IamthankfultotheeditorialteamatSpringer,especiallySwatiMeherishi
andKamiyaKhatter,fortheirsupportinpublishingthisbook.
Finally, I am thankful to my family and friends for their support and
encouragement.
M.C.Bhuvaneswari
vii

Contents
1 IntroductiontoMulti-objectiveEvolutionaryAlgorithms. . . . . . . . 1
M.C.BhuvaneswariandG.Subashini
2 Hardware/SoftwarePartitioningforEmbeddedSystems. . . . . . . . 21
M.C.BhuvaneswariandM.Jagadeeswari
3 CircuitPartitioningforVLSILayout. . . . . . . . . . . . . . . . . . . . . . . 37
M.C.BhuvaneswariandM.Jagadeeswari
4 DesignofOperationalAmplifier. . . . . . . . . . . . . . . . . . . . . . . . . . . 47
M.C.BhuvaneswariandM.Shanthi
5 DesignSpaceExplorationforSchedulingand
AllocationinHighLevelSynthesisofDatapaths. . . . . . . . . . . . . . . 69
M.C.Bhuvaneswari,D.S.HarishRam,andR.Neelaveni
6 DesignSpaceExplorationofDatapath(Architecture)inHigh-Level
SynthesisforComputationIntensiveApplications. ... ... ... .... . 93
AnirbanSengupta
7 DesignFlowfromAlgorithmtoRTLUsingEvolutionary
ExplorationApproach. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
AnirbanSengupta
8 Cross-TalkDelayFaultTestGeneration. . . . . . . . . . . . . . . . . . . . . 125
M.C.BhuvaneswariandS.Jayanthy
9 SchedulinginHeterogeneousDistributedSystems. . . . . . . . . . . . . 147
M.C.BhuvaneswariandG.Subashini
AuthorIndex. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
ix