人工智能论文英文版-Decomposability-Guaranteed Cooperative Coevolution for Large-Scale Itinerary Planning.pdfVIP

1

Decomposability-GuaranteedCooperative

CoevolutionforLarge-ScaleItineraryPlanning

ZiyuZhang,PeilanXu,Member,IEEE,YuetongSun,YuhuiShi,WenjianLuo,SeniorMember,IEEE

Abstract—Large-scaleitineraryplanningisavariantoftheautomateditineraryplanninghasgainedincreasingattention,

travelingsalesmanproblem,aimingtodetermineanoptimalofferingtravelersmoreefficientandpersonalizedsolutions[4].

paththatmaximizesthecollectedpointsofinterest(POIs)scoresAutomateditineraryplanningalgorithmscanbebroadly

whileminimizingtraveltimeandcost,subjecttotravelduration

constraints.Thispaperanalyzesthedecomposabilityoflarge-categorizedintoexactandapproximatemethods.Exactal-

scaleitineraryplanning,provingthatstrictdecomposabilitygorithms,suchasmixed-integerlinearprogramming[5]and

5

2isdifficulttosatisfy,andintroducesaweakdecomposabilitygeneralalgebraicmodelingsystems[6],guaranteeoptimal

0definitionbasedonanecessarycondition,derivingthecor-solutionsbutsufferfromhighcomputationalcomplexity.

2respondinggraphstructuresthatfulfillthisproperty.WithGiventhatitineraryplanningisNP-hard[7]–[9],approximate

decomposabilityguaranteed,weproposeanovelmulti-objective

ncooperativecoevolutionaryalgorithmforlarge-scaleitinerarymethodsareoftenpreferredforefficiency.Theseinclude

uplanning,addressingthechallengeso