多选题

    Antcolonyoptimization(ACO)

    Antcolonyoptimization(ACO)isaprobabalistic(stochastic),heuristicoptimizationtechniqueinspiredbythewayantsmake&findpathsfromthecolonytofood.Thetechniqueisusedtosolvediscreteoptimizationproblemsthatcanbereducedtofindinggoodpathsthroughgraphs.ThefirstappearanceofanACOsystemwasinaPhDthesisin1992byMarcoDorigoatPolitecnicodiMilano.ItwascalledAntSystem(AS).Since1995variousotherextendedversionsofAShavebeendeveloped,includingAntColonySystem(ACS)andMAX-MINAntSystem(MMAS).In1999DorigoproposedtheACOmetaheuristicthatbecamethemostsuccessfulandrecognizedalgorithmbasedonantbehavior.
    B. Whenoneantfindsanoptimalpathfromthecolonytoafoodsource,otherantsaremorelikelytofollowthatpath&positivefeedbackeventuallycausesalltheantstofollowthesamepath.TheACOalgorithmmimicsthisbywalkingaroundthegraphrepresentingtheproblemtosolve.Thesealgorithmshavebeenappliedtothesymmetric&asymmetrictravelingsalesmanproblemwithnear-optimalresults.
    C. TherehasbeenaninterestinusingACOfornetworkrouting&urbantransportationsystemssincethealgorithmcanberuncontinuouslygivingittheabilitytoadapttochangesinrealtime.Thisisanadvantageoverthesimulatedannealing&geneticalgorithmapproachessincetheydonotchangedynamically.
    D. Inmachinelearning&dataminingproblems,ACOvariationshavebeenusedtocreateamodelofthewayworkerants"cluster"antcorpsesinantcemetarymaintenance.Thishasbeenappliedtoataskcalledclusteringinmachinelearningwhichinvolvesfindinggroupsofobjectsthataresimilar.Thismethodhasproventohavehigherperformance&accuracythanpreviousclassicalmethods.

    火星搜题