简介：Learningiswidelyusedinintelligentplanningtoshortentheplanningprocessorimprovetheplanquality.Thispaperaimsatintroducinglearningandfatigueintotheclassicalhierarchicaltasknetwork(HTN)planningprocesssoastocreatebetterhighqualityplansquickly.TheprocessofHTNplanningismappedduringadepth-firstsearchprocessinaproblem-solvingagent,andthemodelsoflearninginHTNplanningisconductedsimilartothelearningdepth-firstsearch(LDFS).Basedonthemodels,alearningmethodintegratingHTNplanningandLDFSispresented,andafatiguemechanismisintroducedtobalanceexplorationandexploitationinlearning.Finally,experimentsintwoclassicaldomainsarecarriedoutinordertovalidatetheeffectivenessoftheproposedlearningandfatigueinspiredmethod.

简介：Anovelalgorithm,theImmuneQuantum-inspiredGeneticAlgorithm(IQGA),isproposedbyintroducingimmuneconceptsandmethodsintoQuantum-inspiredGeneticAlgorithm(QGA).WiththeconditionofpreservingQGA'sadvantages,IQGAutilizesthecharacteristicsandknowledgeinthependingproblemsforrestrainingtherepeatedandineffectiveoperationsduringevolution,soastoimprovethealgorithmefficiency.TheexperimentalresultsoftheknapsackproblemshowthattheperformanceofIQGAissuperiortotheConventionalGeneticAlgorithm(CGA),theImmuneGeneticAlgorithm(IGA)andQGA.