一種基于鯨魚優(yōu)化的多路徑路由發(fā)現(xiàn)算法

一種基于鯨魚優(yōu)化的多路徑路由發(fā)現(xiàn)算法AbstractMultiplepathroutingdiscoveryisachallengingprobleminnetworkcommunicationsduetothevariousparametersinvolved,suchasnetworktopology,trafficloadlevels,andtransmissionreliability.Toovercomethischallenge,anewalgorithmbasedontheWhaleOptimizationAlgorithmisproposedinthispaper.TheWhaleOptimizationAlgorithmisanature-inspiredalgorithmthatborrowsitstechniquesfromthebehaviorofhumpbackwhales'groupmovements.Theproposedalgorithmselectsmultiplepathstoroutedatapackets,seekingtofindtheoptimalsolutionthatbalancesthetrade-offbetweennetworkdelayandtransmissionrobustness.Experimentalresultsshowthattheproposedalgorithmoutperformsexistingworksintermsofreduceddelay,decreasedpacketloss,andenhancednetworkthroughput.IntroductionAsthedemandforhigh-speedandreliablecommunicationservicescontinuestorise,thenetworkarchitecturemustadaptaccordinglytoprovideefficientroutingofdatapackets.Multiplepathroutingdiscoveryisapromisingapproachtocopewiththesedemandssinceitcanbalanceloaddistribution,reducecongestion,improvereliability,andincreaseresilience.Thetraditionalapproachtorouting,whichusessinglepathrouting,isvulnerabletolinkfailures,resultingincommunicationdisruptionsandreducedqualityofservice.Multiplepathroutingishighlyefficientintermsofend-to-enddelay,bandwidthutilization,andpacketlossrate.Severalalgorithmshavebeenproposedtoaddressthemultiplepathroutingproblem,suchasAntColonyOptimization,GeneticAlgorithm,ParticleSwarmOptimization,andArtificialBeeColonyAlgorithm,etc.Amongthesealgorithms,WhaleOptimizationAlgorithm(WOA)isthemostrecentone,whichhasbeenshowntohavesuperiorperformanceinsolvingvariousoptimizationproblems.WOA,asnature-inspiredevolutionaryalgorithm,takesinspirationfromthesocialbehaviorofhumpbackwhaleswhilesearchingfortheoptimalsolution.WOAhasseveralsignificantadvantagescomparedtoothernature-inspiredalgorithmssuchassimplicity,adaptability,andfewertuningparameters.WOAemploysthreesearchmechanisms:exploration,exploitation,andboundaryconstraints,tobalancetheexplorationandexploitationofthesearchspace.ThesemechanismsincreasetheconvergencerateandsolutionaccuracyofWOA.WOAcontinuouslyimprovesitssearchabilitythroughiterationsthatupdatethepositionandvelocityofeachwhaleinthepopulation.ThispaperproposesanewalgorithmbasedontheWOAtosolvethemultiplepathroutingproblem.Theproposedalgorithmselectsdifferentpathsfortransmission,whichcanbalancetheloaddistribution,improvelinkutilization,andenhancenetworkcongestionavoidance.Themaincontributionsofthepapercanbesummarizedasfollows:1.ProposalofanewalgorithmbasedontheWOAtosolvethemultiplepathroutingproblem.2.EvaluationoftheproposedalgorithmagainstseveralexistingalgorithmssuchastheAntColonyAlgorithm,theGeneticAlgorithm,andtheParticleSwarmOptimizationAlgorithm.3.Analysisoftheexperimentalresultsobtainedandcomparisonoftheproposedalgorithm'sperformance.Theremainderofthispaperisorganizedasfollows:Section2summarizesthepreviousresearchonmultiplepathroutingalgorithms.Section3presentsindetailtheproposedalgorithmbasedontheWOAformultiplepathrouting.Section4providestheexperimentalsetupanddataanalysis.Finally,Section5summarizestheresultsandconcludesthepaper.RelatedWorkMultiplepathroutingalgorithmshavebeenwidelystudiedinrecentyears.Themaingoalofthesealgorithmsistobalancetheloaddistribution,mitigatenetworkcongestion,andimprovenetworkperformance.Theearliestworkonmultiplepathroutingwasproposedintheearly1990sandappliedtotheInternet.However,thesealgorithmswerenotwidelyusedduetoslowdatatransmissionrates.Withtheincreasingdemandforhigh-speednetworks,researchershavedevelopedmanyalgorithmstoaddressmultiplepathroutingproblems.Someofthemostpopularalgorithmsarediscussedbelow.TheAntColonyOptimization(ACO)Algorithm,inspiredbytheforagingbehaviorofants,hasbeenusedtosolvemanyoptimizationproblems.TheACOalgorithmappliesaprobabilisticpheromonemodeltoguidetheant'sdecision-makingprocessduringpathselection.ACOhasshownexcellentperformanceincongestionavoidance,loadbalancing,androutingoptimizationincomputernetworks.TheGeneticAlgorithm(GA)isanotherpopularalgorithmthathasbeenwidelyusedinnetworkroutingoptimization.GAusesevolutionarystrategiestooptimizenetworkroutingbyselectingthefittestsolutionsfromapopulationofpotentialsolutions.TheGAalgorithmisefficientinexploringthesearchspaceandcanfindnear-optimalsolutionsinashorttime.TheParticleSwarmOptimization(PSO)Algorithmmodelsthebehaviorofbirdsandfishtosolveoptimizationproblems.PSOishighlysuitableformultiplepathroutingincomputernetworks,asitcanimplementrapiddecisionswhenfacedwithcomplexnetworktopologies,unpredictabletrafficloads,andvariablenetworkperformance.WOAisarelativelynewalgorithmthatwasfirstproposedin2016.WOAismodeledbasedonthesocialbehaviorofhumpbackwhalestosolveoptimizationproblems,suchasthemultiplepathroutingproblem.WOAhasdemonstratedsuperiorperformanceinmanyapplicationsbyusingfewerparametersandrequiringnogradientinformation.ProposedAlgorithmTheWOAalgorithm'sbasicideaistomodelthesocialbehaviorofhumpbackwhalesandapplytheirbehaviortosearchforoptimalsolutions.TheWOAalgorithmemploysthreemechanisms:exploration,exploitation,andboundaryconstraints,tobalancethesolution'saccuracyandconvergencerate.WOAusesthefollowingequationsduringthesearchprocess:X(t+1)=X(t)+A(D(t,X(best))*C(t,best)-X(t))X1(t+1)=X(best)-A*r1*(X(Worst)-X(t))X2(t+1)=X(mean)-A*r2*(X(rand)-X(t))WhereX(t)isthecurrentsearchposition,X(t+1)istheupdatedsearchposition,X(best)representsthebestpositioninthesearch,X(Worst)istheworstsearchposition,X(mean)representsthemeansearchposition,X(rand)isarandomsearchposition,Aisthesearchcontrolparameter,r1andr2arerandomnumbersbetween0and1,andCandDarethetwotransferfunctionsoftheWOAalgorithm.TheproposedalgorithmbasedontheWOAformultiplepathroutingworksasfollows:1.TheWOAalgorithminitializesapopulationofhumpbackwhaleswithrandompositionsandvelocities.2.Thealgorithmdividesthenetworktopologyintoseveralsubnetworksbyperformingnetworkpartitioningusingagraph-theoreticalmethod.3.Thealgorithmselectsasourcenodeandadestinationnodeforeachsubnetwork.4.EachwhaleselectsapathfortransmissionbyapplyingtheWOAalgorithm,whichselectsasetofpathswithminimalend-to-enddelaysandmaximumrobustness.5.ThealgorithmupdatesthepositionandvelocityofeachwhaleandselectsagainthepathusingtheWOAalgorithm.6.Thealgorithmevaluatesthetransmissionpathbasedonseveralqualitymetrics,suchasend-to-enddelayandtransmissionrobustness.7.Thealgorithmselectsthebestsolutionfromthepopulationofhumpbackwhalesandtransmitsthedatapackettoitsdestinationonthatpath.8.Thealgorithmcontinuesthesearchandpathdiscoveryprocessuntilthemaximumnumberofiterationsisreachedorthetargetsolutionisachieved.ExperimentalResultsTheproposedalgorithmwasimplementedandevaluatedonasimulatednetworkenvironmentusingthens-3simulator.Thesimulationenvironmentincludedarandomnetworktopologywith20nodes,8paths,and160links,withauniformpacketgenerationrateof1packet/sec.Theperformanceoftheproposedalgorithmwascomparedwiththreestate-of-the-artalgorithms,namelyACO,GA,andPSO.Themetricsusedforevaluatingtheperformanceofthealgorithmswereend-to-enddelay,packetlossrate,andnetworkthroughput.TheresultsofthesimulationareshowninFig.1-3.Fig.1showstheend-to-enddelayofthedifferentalgorithms.ItcanbeseenthattheproposedalgorithmbasedonWOAhasthelowestend-to-enddelayamongallthetestedalgorithms.TheWOAalgorithmoutperformsACO,GA,andPSOintermsofreduceddelay.Fig.2showsthepacketlossrateforthetestedalgorithms.ItcanbeseenthattheproposedWOAalgorithmhasthelowestpacketlossratecomparedtotheothertested