Download Introduction to modern dynamics: chaos, networks, space, and time 2nd edition david d. nolt

Page 1


Visit to download the full and correct content document: https://ebookmass.com/product/introduction-to-modern-dynamics-chaos-networks-sp ace-and-time-2nd-edition-david-d-nolte/

More products digital (pdf, epub, mobi) instant download maybe you interests ...

Introduction to Modern Dynamics: Chaos, Networks, Space, and Time 2nd Edition

David D. Nolte

https://ebookmass.com/product/introduction-to-modern-dynamicschaos-networks-space-and-time-2nd-edition-david-d-nolte/

Introduction to Modern Analysis, 2nd Edition

https://ebookmass.com/product/introduction-to-modernanalysis-2nd-edition-kantorovitz/

Introduction to Chaos, Fractals and Dynamical Systems

Phil Laplante

https://ebookmass.com/product/introduction-to-chaos-fractals-anddynamical-systems-phil-laplante/

Foundations of Space Dynamics 1st Edition Ashish Tewari

https://ebookmass.com/product/foundations-of-space-dynamics-1stedition-ashish-tewari/

Introduction to Networks v6 Companion Guide 1st Edition, (Ebook PDF)

https://ebookmass.com/product/introduction-tonetworks-v6-companion-guide-1st-edition-ebook-pdf/

Introduction to Unity ML-Agents: Understand the Interplay of Neural Networks and Simulation Space Using the Unity ML-Agents Package 1st Edition Dylan

Engelbrecht

https://ebookmass.com/product/introduction-to-unity-ml-agentsunderstand-the-interplay-of-neural-networks-and-simulation-spaceusing-the-unity-ml-agents-package-1st-edition-dylanengelbrecht-2/

Introduction to Unity ML-Agents: Understand the Interplay of Neural Networks and Simulation Space Using the Unity ML-Agents Package 1st Edition Dylan Engelbrecht

https://ebookmass.com/product/introduction-to-unity-ml-agentsunderstand-the-interplay-of-neural-networks-and-simulation-spaceusing-the-unity-ml-agents-package-1st-edition-dylan-engelbrecht/

Demanding Energy: Space, Time and Change 1st Edition

Allison Hui

https://ebookmass.com/product/demanding-energy-space-time-andchange-1st-edition-allison-hui/

Interference: The History of Optical Interferometry and the Scientists Who Tamed Light David Nolte

https://ebookmass.com/product/interference-the-history-ofoptical-interferometry-and-the-scientists-who-tamed-light-davidnolte/

INTRODUCTIONTOMODERNDYNAMICS

DavidD.Nolte PurdueUniversity

GreatClarendonStreet,Oxford,OX26DP, UnitedKingdom

OxfordUniversityPressisadepartmentoftheUniversityofOxford. ItfurtherstheUniversity’sobjectiveofexcellenceinresearch,scholarship, andeducationbypublishingworldwide.Oxfordisaregisteredtrademarkof OxfordUniversityPressintheUKandincertainothercountries

©DavidD.Nolte2019

Themoralrightsoftheauthorhavebeenasserted

FirstEditionpublishedin2015

SecondEditionpublishedin2019

Impression:1

Allrightsreserved.Nopartofthispublicationmaybereproduced,storedin aretrievalsystem,ortransmitted,inanyformorbyanymeans,withoutthe priorpermissioninwritingofOxfordUniversityPress,orasexpresslypermitted bylaw,bylicenceorundertermsagreedwiththeappropriatereprographics rightsorganization.Enquiriesconcerningreproductionoutsidethescopeofthe aboveshouldbesenttotheRightsDepartment,OxfordUniversityPress,atthe addressabove

Youmustnotcirculatethisworkinanyotherform andyoumustimposethissameconditiononanyacquirer

PublishedintheUnitedStatesofAmericabyOxfordUniversityPress 198MadisonAvenue,NewYork,NY10016,UnitedStatesofAmerica

BritishLibraryCataloguinginPublicationData Dataavailable

LibraryofCongressControlNumber:2019945041

ISBN978–0–19–884462–4(hbk.) ISBN978–0–19–884463–1(pbk.)

DOI:10.1093/oso/9780198844624.001.0001

Printedandboundby CPIGroup(UK)Ltd,Croydon,CR04YY

LinkstothirdpartywebsitesareprovidedbyOxfordingoodfaithand forinformationonly.Oxforddisclaimsanyresponsibilityforthematerials containedinanythirdpartywebsitereferencedinthiswork.

PrefacetotheSecondEdition

IntroductiontoModernDynamics:Chaos,Networks,SpaceandTime (2015)ispart ofanemergingeffortinphysicseducationtoupdatetheundergraduatephysics curriculum.Conventionaljunior-levelmechanicscourseshaveoverlookedmany moderndynamicstopicsthatphysicsmajorswilluseintheircareers:nonlinearity, chaos,networktheory,econophysics,gametheory,neuralnets,geodesicgeometry, amongothers.Thesearethetopicsattheforefrontofphysicsthatdrive high-techbusinessesandstart-upswheremorethanhalfofphysicistsare employed.Thefirsteditionof IntroductiontoModernDynamics contributed tothiseffortbyintroducingthesetopicsinacoherentprogramthatemphasized commongeometricpropertiesacrossawiderangeofdynamicalsystems.

Thesecondeditionof IntroductiontoModernDynamics continuesthattrend byexpandingchapterstoincludingadditionalmaterialandtopics.Itrearranges severaloftheintroductorychaptersforimprovedlogicalflowandexpandsthem toaddnewsubjectmatter.Thesecondeditionalsohasadditionalhomework problems.

Neworexpandedtopicsinthesecondeditioninclude

• Lagrangianapplications

• Lagrange’sundeterminedmultipliers

• Action-anglevariablesandconservedquantities

• Thevirialtheorem

• Non-autonomousflows

• AnewchapteronHamiltonianchaos

• Rationalresonances

• Synchronizationofchaos

• Diffusionandepidemicsonnetworks

• Replicatordynamics

• Gametheory

• Anextensivelyexpandedchapteroneconomicdynamics

Thegoalofthesecondeditionof IntroductiontoModernDynamics isto strengthenthesectionsonconventionaltopics(whichstudentsneedfortheGRE physicssubjecttest),makingitanidealtextbookforbroaderadoptionatthejunior

level,whilecontinuingtheprogramofupdatingtopicsandapproachesthatare relevantfortherolesthatphysicistswillplayinthetwenty-firstcentury.

Thehistoricaldevelopmentofmoderndynamicsisdescribedin Galileo Unbound:APathAcrossLife,theUniverseandEverything,byD.D.Nolte,published byOxfordUniversityPress(2018).

Preface:TheBestPartsofPhysics

Thebestpartsofphysicsarethelasttopicsthatourstudentseversee.These aretheexcitingnewfrontiersofnonlinearandcomplexsystemsthatareat theforefrontofuniversityresearchandarethebasisofmanyofourhightechbusinesses.TopicssuchastrafficontheWorldWideWeb,thespreadof epidemicsthroughgloballymobilepopulations,orthesynchronizationofglobal economiesaregovernedbyuniversalprinciplesjustasprofoundasNewton’s Laws.Nonetheless,theconventionaluniversityphysicscurriculumreservesmost ofthesetopicsforadvancedgraduatestudy.Twojustificationsaregivenforthis situation:first,thatthemathematicaltoolsneededtounderstandthesetopicsare beyondtheskillsetofundergraduatestudents,andsecond,thatthesearespecialty topicswithnocommonthemeandlittleoverlap.

IntroductiontoModernDynamics:Chaos,Networks,SpaceandTime dispelsthese myths.Thestructureofthisbookcombinesthethreemaintopicsofmodern dynamics—chaostheory,dynamicsoncomplexnetworksandthegeometryof dynamicalspaces—intoacoherentframework.Bytakingageometricviewof physics,concentratingonthetimeevolutionofphysicalsystemsastrajectories throughabstractspaces,thesetopicsshareacommonandsimplemathematical languagewithwhichanystudentcangainaunifiedphysicalintuition.Giventhe growingimportanceofcomplexdynamicalsystemsinmanyareasofscienceand technology,thistextprovidesstudentswithanup-to-datefoundationfortheir futurecareers.

Whilepursuingthisaim, IntroductiontoModernDynamics embedsthetopics ofmoderndynamics—chaos,synchronization,networktheory,neuralnetworks, evolutionarychange,econophysics,andrelativity—withinthecontextoftraditionalapproachestophysicsfoundedonthestationarityprinciplesofvariational calculusandLagrangianandHamiltonianphysics.Asthephysicsstudentexplores thewiderangeofmoderndynamicsinthistext,thefundamentaltoolsthatare neededforaphysicist’scareerinquantitativescienceareprovided,including topicsthestudentneedstoknowfortheGraduateRecordExamination(GRE). Thegoalofthistextbookistomodernizetheteachingofjunior-leveldynamics, responsivetoachangingemploymentlandscape,whileretainingthecoretraditionsandcommonlanguageofdynamicstexts.

Aunifyingconcept:geometryanddynamics

Instructorsorstudentsmaywonderhowanintroductorytextbookcancontain topics,underthesamebookcover,oneconophysicsandevolutionaswellasthe physicsofblackholes.However,itisnotthephysicsofblackholesthatmatters, ratheritisthedescriptionofgeneraldynamicalspacesthatisimportantandthe understandingthatcanbegainedofthegeometricaspectsoftrajectoriesgoverned bythepropertiesofthesespaces.Allchangingsystems,whetherinbiologyor economicsorcomputerscienceorphotonsinorbitaroundablackhole,are understoodastrajectoriesinabstractdynamicalspaces.

Newtontakesabackseatinthistext.Hewillalwaysbeattheheartofdynamics, butthemodernemphasishasshiftedawayfrom F = ma toanewerperspective whereNewton’sLawsarespecialcasesofbroaderconcepts.Thereareeconomic forcesandforcesofnaturalselectionthatarejustasrealastheforceofgravity onpointparticles.Forthatmatter,eventheforceofgravityrecedesintothe backgroundasforce-freemotionincurvedspace-timetakesthefore.

UnlikeNewton,HamiltonandLagrangeretaintheirpositionshere.Thevariationalprincipleandtheminimizationofdynamicalquantitiesarecoreconcepts indynamics.Minimizationoftheactionintegralprovidestrajectoriesinreal space,andminimizationofmetricdistancesprovidestrajectories—geodesics— indynamicalspaces.ConservationlawsarisenaturallyfromLagrangians,and energyconservationenablessimplificationsusingHamiltoniandynamics.Space andgeometryarealmostsynonymousinthiscontext.Definingthespaceofa dynamicalsystemtakesfirstimportance,andthegeometryofthedynamicalspace thendeterminesthesetofalltrajectoriesthatcanexistinit.

Acommontool:dynamicalflowsandthe ODEsolver

Amathematicalflowisasetoffirst-orderdifferentialequationsthataresolved usingasmanyinitialvaluesastherearevariables,whichdefinesthedimensionality ofthedynamicalspace.Mathematicalflowsareoneofthefoundationstonesthat appearscontinuallythroughoutthistextbook.Nearlyallofthesubjectsexplored here—fromevolvingvirusestoorbitaldynamics—canbecapturedasaflow. Therefore,acommontoolusedthroughoutthistextisthenumericalsolution oftheordinarydifferentialequation(ODE).Computerscanbebothaboonand abanetothemodernphysicsstudent.Ontheonehand,theeasyavailabilityof ODEsolversmakeseventhemostobscureequationseasytosimulatenumerically, enablinganystudenttoplotaphaseplaneportraitthatcontainsallmannerof behavior.Ontheotherhand,physicalinsightandanalyticalunderstandingof complexbehaviortendtosufferfromthecomputer-gamenatureofsimulators. Therefore,thistextbookplacesastrongemphasisonanalysis,andonbehavior

underlimitingconditions,withthegoaltoreduceaproblemtoafewsimple principles,whilemakinguseofcomputersimulationstocaptureboththewhole pictureaswellasthedetailsofsystembehavior.

Traditionaljunior-levelphysics:howtouse thisbook

Allthetraditionaltopicsofjunior-levelphysicsarehere.Fromthesimplest descriptionoftheharmonicoscillator,throughLagrangianandHamiltonian physics,torigidbodymotionandorbitaldynamics—thecoretopicsofadvanced undergraduatephysicsareretainedandareinterspersedthroughoutthistextbook.

What’ssimpleincomplexsystems?

Thetraditionaltopicsofmechanicsareintegratedintothebroaderviewofmodern dynamicsthatdrawsfromthetheoryofcomplexsystems.Therangeofsubject matterencompassedbycomplexsystemsisimmense,andacomprehensive coverageofthistopicisoutsidethescopeofthisbook.However,thereisstill asurprisinglywiderangeofcomplexbehaviorthatcanbecapturedusingthe simpleconceptthatthegeometryofadynamicspacedictatesthesetofall possibletrajectoriesinthatspace.Therefore,simpleanalysisoftheassociated flowsprovidesmanyintuitiveinsightsintotheoriginsofcomplexbehavior.The specialtopicscoveredinthistextbookare:

• Chaostheory(Chapter4)

Muchofnonlineardynamicscanbeunderstoodthrough linearization ofthe flow equations(equationsofmotion)aroundspecial fixedpoints.Visualizingthe dynamicsofmulti-parametersystemswithinmultidimensionalspacesismade simplerbyconceptssuchasthe Poincarésection, strangeattractors thathave fractal geometry,and iterativemaps

• Synchronization(Chapter6)

Thenonlinear synchronization oftwoormoreoscillatorsisastartingpointfor understandingmorecomplexsystems.Asthewholecanbegreaterthanthesum oftheparts,globalpropertiesoftenemergefromlocalinteractionsamongthe parts.Synchronizationofoscillatorsissurprisinglycommonandrobust,leading to frequency-entrainment, phase-locking,and fractionalresonance thatallowsmall perturbationstocontrollargenetworksofinteractingsystems.

Preface:TheBestPartsofPhysics

Preface:TheBestPartsofPhysics

• Networktheory(Chapter7)

Everywherewelooktoday,weseenetworks.Theonesweinteractwithdaily aresocialnetworksandrelatednetworksontheWorldWideWeb.Inthis chapter,individualnodesarejoinedintonetworksofvariousgeometries,suchas small-worldnetworks and scale-freenetworks.The diffusion ofdiseaseacrossthese networksisexplored,andthesynchronizationof Poincaréphaseoscillators can inducea Kuramototransition tocompletesynchronicity.

• Evolutionarydynamics(Chapter8)

Someoftheearliestexplorationsofnonlineardynamicscamefromstudiesof populationdynamics.Inamoderncontext,populationsaregovernedbyevolutionary pressuresandbygenetics.Topicssuchasviralmutationandspread,aswellasthe evolutionofspecieswithina fitnesslandscape,areunderstoodassimplebalances within quasispecies equations.

• Neuralnetworks(Chapter9)

Perhapsthemostcomplexofallnetworksisthebrain.Thischapterstartswiththe singleneuron,whichisa limit-cycleoscillator thatcanshowinteresting bistability and bifurcations.Whenneuronsareplacedintosimpleneuralnetworks,suchas perceptrons or feedforwardnetworks,theycandosimpletasksaftertrainingby error back-propagation.Thecomplexityofthetasksincreaseswiththecomplexityof thenetworks,and recurrentnetworks,likethe Hopfieldneuralnet,canperform associatedmemoryoperationsthatchallengeeventhehumanmind.

• Econophysics(Chapter10)

Amostbafflingcomplexsystemthatinfluencesourdailyactivities,aswellas thetrajectoryofourcareers,istheeconomyinthelargeandthesmall.The dynamicsof microeconomics determineswhatandwhywebuy,whilethedynamics of macroeconomics drivesentirenationsupanddowneconomicswings.These forcescanbe(partially)understoodintermsofnonlineardynamicsandflows ineconomicspaces. Businesscycles andthediffusionofpricesonthe stockmarket arenolessunderstandablethanevolutionarydynamics(Chapter8)ornetwork dynamics(Chapter7),andindeeddrawcloselyfromthosetopics.

• Geodesicmotion(Chapter11)

Thischapteristhebridgebetweentheprecedingchaptersoncomplexsystems andthesucceedingchaptersonrelativitytheory(bothspecialandgeneral).This iswherethegeometryofspaceisfirstfullydefinedintermsofa metrictensor,and wheretrajectoriesthrougha dynamicalspace arediscoveredtobepathsof force-

freemotion.The geodesicequation (ageodesicflow)supersedesNewton’sSecond Lawasthefundamentalequationofmotionthatcanbeusedtodefinethepathof massesthroughpotentiallandscapesandthepathoflightthroughspace-time.

• Specialrelativity(Chapter12)

Inadditiontotraditionaltopicsof Lorentztransformations and mass-energy equivalence,thischapterpresentsthebroaderviewoftrajectoriesthroughMinkowski space-time whosegeometricpropertiesaredefinedbythe Minkowskimetric. Relativisticforcesandnoninertial(accelerating)framesconnecttothenext chapterthatgeneralizesallrelativisticbehavior.

• Generalrelativity(Chapter13)

Thephysicsof gravitation,morethananyothertopic,benefitsfromtheoverarchingthemedevelopedthroughoutthisbook—thatthegeometryofaspace definesthepropertiesofalltrajectorieswithinthatspace.Indeed,inthisgeometric viewofphysics,Newton’sforceofgravitydisappearsandisreplacedbyforcefreegeodesicsthrough warped space-time.Mercury’sorbitaroundtheSun, andtrajectoriesoflightpast blackholes,areelementsofgeodesicflowswhose propertiesareeasilyunderstoodusingthetoolsdevelopedinChapter4and expandeduponthroughoutthistextbook.

Preface:TheBestPartsofPhysics

Acknowledgments

Igratefullyacknowledgethemanyhelpfuldiscussionswithmycolleagues EphraimFischbach,AndrewHirsch,SherwinLove,andHisaoNakanishiduring thepreparationofthisbook.Specialthankstomyfamily,LauraandNicholas,for puttingupwithmy“hobby”forsomanyyears,andalsofortheirencouragement andmoralsupport.IalsothanktheeditorsatOxfordUniversityPressforhelpin preparingthemanuscriptandespeciallySonkeAdlungforhelpingmerealizemy vision.

PartI GeometricMechanics

Traditionalapproachestothemechanicsofparticlestendtofocusonindividual trajectories.Incontrast,moderndynamicstakesaglobalviewofdynamical behaviorbystudyingthesetofallpossibletrajectoriesofasystem.Modern dynamicsfurthermorestudiespropertiesindynamicalspacesthatcarrynames like statespace, phasespace, and space–time.Dynamicalspacescanbehighlyabstract andcanhavehighdimensionality.Thisinitialpartofthebookintroducesthe mathematicaltoolsnecessarytostudythegeometryofdynamicalspacesandthe resultingdynamicalbehaviorwithinthosespaces.Centraltomoderndynamics isHamilton’sPrincipleofStationaryActionastheprototypicalminimization principlethatunderliesmuchofdynamics.Thisapproachwillleadultimately (inPartIII)tothegeodesicequationofgeneralrelativity,inwhichmatter warpsMinkowskispace(space–time),andtrajectoriesexecuteforce-freemotion throughthatspace.

PhysicsandGeometry

Moderndynamics,likeclassicaldynamics,isconcernedwithtrajectoriesthrough space—thedescriptionsoftrajectories(kinematics)andthecausesoftrajectories (dynamics).However,unlikeclassicalmechanics,whichemphasizesmotionsof physicalmassesandtheforcesactingonthem,moderndynamicsgeneralizesthe notionoftrajectoriestoencompassabroadrangeoftime-varyingbehaviorthat goesbeyondmaterialparticlestoincludeanimalspeciesinecosystems,market pricesineconomies,andvirusspreadonconnectednetworks.Thespacesthat thesetrajectoriesinhabitareabstract,andcanhaveahighnumberofdimensions. ThesegeneralizedspacesmaynothaveEuclideangeometry,andmaybecurved likethesurfaceofasphereorspace–timewarpedbygravity.Thecentralobject ofinterestindynamicsistheevolvingstateofasystem.Thestatedescriptionof asystemmustbeunambiguous,meaningthatthenextstatetodevelopintimeis

1

1.1Statespaceanddynamical flows4

1.2Coordinaterepresentationof dynamicalsystems10

1.3Coordinatetransformations15

1.4Uniformlyrotatingframes25

1.5Rigid-bodymotion32

1.6Summary48

1.7Bibliography48

1.8Homeworkproblems49

Foucault’sPenduluminthePantheoninParis

1 SeeA.E.Jackson, PerspectivesofNonlinearDynamics (CambridgeUniversity Press,1989).

uniquelydeterminedbythecurrentstate.Thisiscalleddeterministicdynamics, whichincludesdeterministicnonlineardynamicsforwhichchaotictrajectories mayhaveanapparentrandomnesstotheircharacter.

Thischapterlaysthefoundationforthedescriptionofdynamicalsystemsthat movecontinuouslyfromstatetostate.Familiesoftrajectories,calleddynamical flows,arethefundamentalelementsofinterest;theyarethefieldlinesofdynamics. Thesefieldlinesaretodeterministicdynamicswhatelectricandmagneticfield linesaretoelectromagnetism.Onekeydifferenceisthatthereisonlyonesetof Maxwell’sequations,whileeverynonlineardynamicalsystemhasitsownsetof equations,providinganearlylimitlessnumberofpossibilitiesforustostudy.

Thischapterbeginsbyintroducinggeneralideasoftrajectoriesasthesetofall possiblecurvesdefinedbydynamicalflowsinstatespace.Todefinetrajectories, wewillestablishnotationtohelpusdescribehigh-dimensional,abstract,and possiblycurvedspaces.Thisisaccomplishedthroughtheuseofmatrix(actually tensor)indicesthatlookstrangeatfirsttoastudentfamiliaronlywithvectors, butwhichareconvenientdevicesforkeepingtrackofmultiplecoordinates.The nextstepconstructscoordinatetransformationsfromonecoordinatesystem toanother.Forinstance,acentralquestioninmoderndynamicsishowtwo observers,oneineachsystem,describethecommonphenomenathatthey observe.The physics mustbeinvarianttothechoiceofcoordinateframe,butthe descriptionscandifferwidely.

1.1Statespaceanddynamicalflows

Configurationspaceisdefinedbythespatialcoordinatesneededtodescribea dynamicalsystem.Thepaththesystemtakesthroughconfigurationspaceisits trajectory.Eachpointonthetrajectorycapturesthesuccessiveconfigurationsof thesystemasitevolvesintime.However,knowingthecurrentconfigurationof thesystemdoesnotguaranteethatthenextconfigurationcanbedefined.For instance,thetrajectorycanloopbackandcrossitself.Thevelocityvectorthat pointedonedirectionattheearliertimecanpointinadifferentdirectionatalater time.Therefore,avelocityvectormustbeattachedtoeachconfigurationtodefine howitwillevolvenext.

1.1.1Statespace

Byaddingvelocities,associatedwitheachofthecoordinates,totheconfiguration space,anewexpandedspace,called statespace,iscreated.Foragiveninitial condition,thereisonlyasinglesystemtrajectorythroughthismultidimensional space,andeachpointonthetrajectoryuniquelydefinesthenextstateofthe system.1 Thistrajectoryinstatespacecancrossitselfonlyatpointswhereallthe velocitiesvanish,otherwisethefuturestateofthesystemwouldnotbeunique.

Example1.1 Statespaceofthedampedone-dimensionalharmonicoscillator

Thedampedharmonicoscillatorinonecoordinatehasthesinglesecond-orderordinarydifferentialequation2

where m isthemassoftheparticle, γ isthedragcoefficient,and k isthespringconstant.Anysetofsecond-order time-dependentordinarydifferentialequations(e.g.,Newton’ssecondlaw)canbewrittenasalargersetoffirst-order equations.Forinstance,thesinglesecond-orderequation(1.1)canberewrittenastwofirst-orderequations

Itisconventionaltowritethesewithasingletimederivativeontheleftas

inthetwovariables (x, v) with

and

. Statespace forthissystemofequationsconsistsoftwo coordinateaxesinthetwovariables (x, v),andtheright-handsideoftheequationsareexpressedusingonlythesame twovariables.

Tosolvethisequation,assumeasolutionintheformofacomplexexponentialevolvingintimewithanangular frequency ω as(seeAppendixA.1)

InsertthisexpressionintoEq.(1.1)toyield

withthecharacteristicequation

wherethedampingparameteris β = γ/2m,andtheresonantangularfrequencyisgivenby ω 2 0 = k/m.Thesolution ofthequadraticequation(1.6)is

Usingthisexpressionfortheangularfrequencyintheassumedsolution(1.4)gives

Considertheinitialvalues x(0) = A and ˙ x(0) = 0;thenthetwoinitialconditionsimposethevalues

2 The“dot”notationstandsforatimederivative: ˙ x = dx/dt and ¨ x = d 2 x/dt 2 .Itisa modernremnantofNewton’sfluxionnotation.

Example1.1 continued

Thefinalsolutionis

whichisplottedinFig.1.1(a)forthecasewheretheinitialdisplacementisamaximumandtheinitialspeediszero. Theoscillator“ringsdown”withtheexponentialdecayconstant β Theangularfrequencyofthering-downisnot equalto ω0 ,butisreducedtothevalue ω 2 0 β 2 .Hence,thedampingdecreasesthefrequencyoftheoscillatorfrom itsnaturalresonantfrequency.Asystemtrajectoryinstatespacestartsataninitialcondition (x0 , v0 ),anduniquely tracesthetimeevolutionofthesystemasacurveinthestatespace.InFig.1.1(b),onlyonetrajectory(streamline)is drawn,butstreamlinesfillthestatespace,althoughtheynevercross,exceptatsingularpointswhereallvelocitiesvanish. Streamlinesarethefieldlinesofthevectorfield.Muchofthestudyofmoderndynamicsisthestudyofthegeometric propertiesofthevectorfield(tangentstothestreamlines)andfieldlinesassociatedwithadefinedsetofflowequations.

Figure1.1 Trajectoriesofthedampedharmonicoscillator.(a)Configurationpositionversustime.(b)Statespace,everypoint ofwhichhasatangentvectorassociatedwithit.Streamlinesarethefieldlinesofthevectorfieldandaredense.Onlyasingle streamlineisshown.

1.1.2Dynamicalflows

Thisbookworkswithageneralformofsetsofdynamicalequationscalleda dynamicalflow.Theflowforasystemof N variablesisdefinedas

or,moresuccinctly,

whichisasystemof N simultaneousequations,wherethevectorfunction Fa is afunctionofthetime-varyingcoordinatesofthepositionvector.If Fa isnotan explicitfunctionoftime,thenthesystemis autonomous,withan N -dimensional statespace.Ontheotherhand,if Fa isanexplicitfunctionoftime,thenthesystem is non-autonomous,withan(N + 1)-dimensionalstatespace(spaceplustime)The solutionofthesystemofequations(1.12)isasetoftrajectories qa (t ) throughthe statespace.

Inthisbook,thephrase configurationspace isreservedforthedynamicsof systemsofmassiveparticles(withsecond-ordertimederivativesasinExamples 1.1and1.2).Thedimensionofthestatespaceforparticlesystemsisevendimensionalbecausethereisavelocityforeachcoordinate.However,forgeneral dynamicalflows,thedimensionofthestatespacecanbeevenorodd.For dynamicalflows,statespaceandconfigurationspacearethesamething,andthe phrase statespace willbeused.

Example1.2 Anautonomousoscillator

Systemsthatexhibitself-sustainedoscillation,knownasautonomousoscillators,arecentraltomanyofthetopics ofnonlineardynamics.Forinstance,anordinarypendulumclock,drivenbymechanicalweights,isanautonomous oscillatorwithanaturaloscillationfrequencythatissustainedbygravity.Onepossibledescriptionofanautonomous oscillatorisgivenbythedynamicalflowequations

Example1.2 continued

where ω isanangularfrequency.The(x, y)state-spacetrajectoriesofthissystemarespiralsthatrelaxtotheunitcircleas theyapproachadynamicequilibrium,showninFig.1.2.Withoutthesecondtermsontheright-handside,thisissimply anundampedharmonicoscillator.Examplesandproblemsinvolvingautonomousoscillatorswillrecurthroughoutthis bookinChapters4(Chaos),6(Synchronization),7(Networks),8(EvolutionaryDynamics),9(Neurodynamics)and 10(EconomicDynamics).

Figure1.2 Flowlinesofanautonomousoscillatorwithalimitcycle.Alltrajectoriesconvergeonthelimitcycle.

Example1.3 Undampedpoint-masspendulum

Theundampedpoint-masspendulumiscomposedofapointmass m onamasslessrigidrodoflength L.Ithasa two-dimensionalstate-spacedynamicsinthespace(

, ω )describedby

Thestate-spacetrajectoriescanbeobtainedbyintegratingtheseequationsusinganonlinearODEsolver.Alternatively, thestate-spacetrajectoriescanbeobtainedanalyticallyifthereareconstantsofthemotion.Forinstance,becausethe pendulumisundampedandconservative,thetotalenergyofthesystemisaconstantforagiveninitialcondition,

referencedtothebottomofthemotioninconfigurationspace.Ifthemaximumangleofthependulumforagiven trajectoryis θ0 ,then

Example1.3 continued and

whichissolvedfortheinstantaneousangularvelocity ω as

Theseareoscillatorymotionsfor θ0 <π .Forlargerenergies,themotionisrotational(alsoknownaslibration).The solutionsinthiscaseare

wherecos θ 0 isnotaphysicalangle,butisaneffectiveparameterdescribingthetotalenergyas

The(θ , ω )state-spacetrajectoriesoftheundampedpoint-masspendulumareshowninFig.1.3.Whenthestatespace pertainstoaconservativesystem,itisalsocalled phasespace.ConservativesystemsareHamiltoniansystemsandare describedinChapter3.

Figure1.3 Statespaceoftheundampedpoint-masspendulum.Theconfigurationspaceisone-dimensionalalongtheangle θ .Closedorbits(oscillation)areseparatedfromopenorbits(rotation)byacurveknownasaseparatrix.

Example1.4 Athree-variableharmonicoscillator

Asanexampleofanodd-dimensionalstatespace,considerthethree-dimensionalflow

Thismathematicalmodelisequivalenttoathree-variablelinearoscillatorwithnodissipation.Tosolvethisflow,assume asolutionintheformofacomplexexponentialintimeevolvingwithanangularfrequency ω as x(t ) = Xei ω t .Insert thisexpressionintoEq.(1.21)toyield

Solvetheseculardeterminantfortheangularfrequency ω :

Thesolutions,foranyinitialcondition,arethreesinusoidswithidenticalamplitudesandfrequencies,butwithrelative phasesthatdifferby ±2π/3.Adynamicalsystemlikethisis not equivalenttomodelingaparticlewithinertia.Itisa dynamicalflowwithastate-spacedimensionequaltothreethatmightmodelthebehaviorofaneconomicsystem,or anecologicalbalanceamongthreespecies,oracoupledsetofneurons.Inthestudyofmoderndynamicalsystems, theemphasismovesawayfromparticlesactedonbyforcesandbecomesmoreabstract,butalsomoregeneraland versatile.

Thisexamplehaswhatiscalled“neutralstability.”Thismeansthatevenaslightperturbationofthissystemmay causetheoscillationstoeitherdecaytozeroortogrowwithoutbound.InChapter4,astabilityanalysiswillidentify thissystemasa“center.”Thisoscillatorysystemisnotarobustsystem,becauseasmallchangeinparametercan causeamajorchangeinitsqualitativebehavior.However,therearetypesofself-sustainedoscillationsthat are robust, maintainingsteadyoscillatorybehaviorevenasparameters,andevendissipation,change.Theseareautonomous oscillatorsandareinvariablynonlinearoscillators.

1.2Coordinaterepresentationofdynamical systems

Although physics mustbeindependentofanycoordinateframe,thedescription ofwhatwesee does dependonwhichframeweareviewingitfrom.Therefore,it oftenwillbeconvenienttoviewthesamephysicsfromdifferentperspectives.For

thisreason,weneedtofindtransformationlawsthatconvertthedescriptionfrom oneframetoanother.

1.2.1Coordinatenotationandconfigurationspace

Thepositionofafreeparticleinthree-dimensional(3D)spaceisspecifiedby threevaluesthatconventionallycanbeassignedtheCartesiancoordinatevalues x(t ), y(t ),and z(t ).Thesecoordinatesdefinetheinstantaneousconfiguration ofthesystem.Ifasecondparticleisadded,thentherearethreeadditional coordinates,andthe configurationspace ofthesystemisnowsix-dimensional. Ratherthanspecifyingthreenewcoordinatenames,suchas u(t ), v(t ),or w(t ), itismoreconvenienttouseanotationthatisextendedeasilytoanynumberof dimensions.Indexnotionaccomplishesthisbyhavingtheindexspanacrossall thecoordinatevalues.

Vectorcomponentsthroughoutthistextwillbedenotedwithasuperscript.For instance,thepositionvectorofafreeparticlein3DEuclideanspaceisa3-tuple ofvalues

Vectorsarerepresentedbycolumnmatrices(whichisthemeaningofthesuperscriptshere3 ).Itisimportanttorememberthatthesesuperscriptsarenot “powers.”Acoordinatecomponentraisedtoan nthpowerwillbeexpressed as (xa )n .For N freeparticles,asingle3N -dimensionalpositionvectordefines theinstantaneousconfigurationofthesystem.Toabbreviatethecoordinate description,onecanusethenotation

x = xa a = 1, ,3N

(1.26)

wherethecurlybracketsdenotethefullsetofcoordinates.Anevenshorter,and morecommon,notationforavectorissimply

xa (1.27)

wherethefullset a = 1, ,3N isimplied.Caseswhereonlyasinglecoordinate isintendedwillbeclearfromthecontext.Thepositioncoordinatesdevelopin timeas

xa (t )

(1.28)

whichdescribesatrajectoryofthesysteminits3N -dimensionalconfiguration space.

3 Thesuperscriptisapartofthenotationfortensorsandmanifoldsinwhich vectorsdifferfromanothertypeofcomponentcalledacovectorthatisdenoted byasubscript.InCartesiancoordinates, asuperscriptdenotesacolumnvector andasubscriptdenotesarowvector(see AppendixA.3).

1.2.2Trajectoriesin3Dconfigurationspace

Atrajectoryisasetofpositioncoordinatevaluesthatvarycontinuouslywith asingleparameteranddefineasmoothcurveintheconfigurationspace.For instance,

where t isthetimeand s isthepathlengthalongthetrajectory.Oncethetrajectory ofapointhasbeendefinedwithinitsconfigurationspace,itishelpfultodefine propertiesofthetrajectory,likethetangenttothecurveandthenormal.The velocityvectoristangenttothepath.Forasingleparticlein3D,thiswouldbe

wherethe ds/dt termissimplythespeedoftheparticle.Inthesimplifiedindex notation,thisis

where T a isaunittangentvectorinthedirectionofthevelocity:

Eachpointonthetrajectoryhasanassociatedtangentvector.Inadditiontothe tangentvector,anotherimportantvectorpropertyofatrajectoryisthenormalto thetrajectory,definedby

where N a istheunitvectornormaltothecurve,andthecurvatureofthe trajectoryis

where R istheradiusofcurvatureatthespecifiedpointonthetrajectory. Theparameterizationofatrajectoryintermsofitspathlength s isoften amore“natural”wayofdescribingthetrajectory,especiallyundercoordinate transformations.Forinstance,inspecialrelativity,timeisnolongeranabsolute parameter,becauseitistransformedinamannersimilartoposition.Thenitis

possibletodefineapathlengthinterval ds2 inspace–timethatremainsinvariant underLorentztransformation(seeChapter12)andhencecanbeusedtospecify thepaththroughspace–time.4

Example1.5 Parabolictrajectoryinagravitationalfield

Thisisafamiliarproblemthatgoesbacktofreshmanphysics.However,it isseenhereinaslightlydifferentlight.Consideraparticleinaconstant gravitationalfieldthrownwithinitialvelocity v0 inthe x direction.The mathematicaldescriptionofthismotionis

withthesolution,forinitialconditions

givingthespatialtrajectory

Thespeedoftheparticleis

withthearclengthelement

andthetangentvectorcomponents

continued 4 Moregenerally,theinvariantsquared pathlengthinterval ds2 isanessentialpart ofthemetricdescriptionofthegeometry ofspace–timeandotherdynamicalspaces, andisakeyaspectofgeodesicmotionfor bodiesmovingthroughthosespaces(see Chapter11).

Example1.5 continued

Thetrajectoryanditstangentvectoraredescribedasfunctionsofposition— ageometriccurveratherthananexplicitfunctionoftime.Whiletheresults forthisfamiliarproblemmaylookunfamiliar,itissimilartothedescription oftrajectoriesinspecialrelativity,ortogeodesictrajectoriesneargravitating bodiesinspace–timethatwillbetreatedinlaterchapters.

1.2.3Generalizedcoordinates

TheconfigurationcoordinatesconsideredsofarhavebeenCartesiancoordinates (x, y, z).However,thereareabstractcoordinates,called generalizedcoordinates, thatmaybemoreeasilyemployedtosolvedynamicalproblems.Generalized coordinatesariseindifferentways.Theymaybedictatedbythesymmetryof theproblem,likepolarcoordinatesforcircularmotion.Theymaybedefined byconstraintsonthephysicalsystem,likeaparticleconstrainedtomoveona surface.Ortheymaybedefinedbycoupling(functionaldependence)between thecoordinatesofamulticomponentsystem,leadingtogeneralizedcoordinates knownasnormalmodes.Generalizedcoordinatesareoftendenotedby q’s. Theymaybedescribedintermsofothercoordinates,forinstanceCartesian coordinates,as

wherethetransformationsassociatedwitheachindexmayhavedifferentfunctionalformsanddonotneedtobelinearfunctionsoftheirarguments.The generalizedcoordinatesdonotneedtohavethedimensionoflength,andeachcan havedifferentunits.However,itisrequiredthatthetransformationbeinvertible (one-to-one).

Generalizedcoordinatescanbeusedtosimplifythedescriptionofthemotions ofcomplexsystemscomposedoflargenumbersofparticles.Ifthereare N particles,eachwiththreecoordinates,thenthetotaldimensionoftheconfiguration spaceis3N andthereisadensesetofsystemtrajectoriesthatthreadtheirway throughthisconfigurationspace.However,oftenthereareconstraintsonthe physicalsystem,suchastherequirementthatparticlesbeconstrainedtoreside onaphysicalsurfacesuchasthesurfaceofasphere.Inthiscase,thereare equationsthatconnecttwoormoreofthecoordinates.Ifthereare K equationsofconstraints,thenthenumberofindependentgeneralizedcoordinatesis 3N K andthemotionoccursona(3N – K )-dimensionalhypersurfacewithin theconfigurationspace.Thishypersurfaceiscalleda manifold.Inprinciple,it ispossibletofindthe3N – K generalizedcoordinatesthatspanthismanifold,

andthemanifoldbecomesthenewconfigurationspacespannedbythe3N – K generalizedcoordinates.Furthermore,someofthegeneralizedcoordinatesmay notparticipateinthedynamics.Thesearecalled ignorablecoordinates (alsoknown as cycliccoordinates),andtheyariseowingtosymmetriesintheconfigurationspace plusconstraints,andareassociatedwithconservedquantities.Thedimensionality ofthedynamicalmanifoldonwhichthesystemtrajectoryresidesisfurther reducedbyeachoftheseconservedquantities.Ultimately,afteralltheconserved quantitiesandalltheconstraintshavebeenaccountedfor,themanifoldthat containsthesystemtrajectorymayhaveadimensionmuchsmallerthanthe dimensionoftheoriginalCartesianconfigurationspace.

Example1.6 Beadslidingonafrictionlesshelicalwire

Considerabeadslidingwithoutfrictiononahelicalwirewithnogravity.Thetrajectoryisdefinedin3DCartesian coordinatesby

parameterizedbytime t.Therearetwoconstraints

where a isthepitchofthehelixand θ = ω t .Theseconstraintsreducethe3Ddynamicsto1Dmotion(3–2 = 1),and the1Dtrajectoryhasasinglegeneralizedcoordinate

whichisalsoequaltothepathlength s.Thespeedoftheparticleisaconstantandis

1.3Coordinatetransformations

Forageneralcoordinatetransformation,theoriginalCartesiancoordinates x, y, and z arerelatedtocoordinates q1 , q2 ,and q3 bythefunctions

x = x q1 , q2 , q3 y = y q1 , q2 , q3 z = z q1 , q2 , q3

Theseequationscanbeinvertedtoyield

whichmaybegeneralizedcoordinatesthatarechosentosimplifytheequations ofmotionofadynamicalsystem.

1.3.1Jacobianmatrix

TheJacobianmatrixofthetransformationisdefinedfromthecoordinatetransformations(andinversetransformations)as

Thedeterminant |J | iscalledtheJacobian.

TheJacobianmatrixrequirestwoindicestodefineitsindividualelements,just asavectorrequiredoneindex.BecausetheJacobianmatrixisgeneratedusing derivatives,anindexnotationthatdistinguishesbetweenthedifferentialvectorin thenumeratorrelativetothedifferentialvectorinthedenominatoris

row index column index

wherethesuperscriptandsubscriptrelateto xa and qb ,respectively.Thesuperscriptiscalledacontravariantindex,andthesubscriptiscalledacovariantindex. Onewaytorememberthisnomenclatureisthat“co”goes“below.”Thecovariant indexreferstothecolumnsofthematrix,andthecontravariantindexreferstothe rows.Columnvectorshavecontravariantindicesbecausetheyhavemultiplerows, whilerowvectorshavecovariantindicesbecausetheyhavemultiplecolumns.Row vectorsarealsoknownascovariantvectors,orcovectors.

WhentransformingbetweenCartesianandgeneralizedcoordinates,an infinitesimaltransformationisexpressedas

Turn static files into dynamic content formats.

Create a flipbook
Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.
Download Introduction to modern dynamics: chaos, networks, space, and time 2nd edition david d. nolt by Education Libraries - Issuu