Continuous State Space Q-Learningfor control of nonlinear system

《Continuous State Space Q-Learningfor control of nonlinear system》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、ContinuousStateSpaceQ-LearningforControlofNonlinearSystems.ContinuousStateSpaceQ-LearningforControlofNonlinearSystemsACADEMISCHPROEFSCHRIFTterverkrijgingvandegraadvandoctoraandeUniversiteitvanAmsterdam,opgezagvandeRectorMagnificusprof.dr.J.J.M.Fransetenoverstaanvaneendoorhetcollegevoorpromot

2、iesingesteldecommissie,inhetopenbaarteverdedigenindeAuladerUniversiteitopwoensdag21februari2001,te12:00uurdoorStephanusHendrikusGerhardustenHagengeborenteNeedePromotor:Prof.dr.ir.F.C.A.GroenCo–promotor:Dr.ir.B.J.A.Kr¨oseCommissie:Prof.dr.H.B.VerbruggenProf.dr.P.AdriaansProf.dr.J.N.KokDr.M.A

3、.WieringFaculteitderNatuurwetenschappen,WiskundeenInformaticaThisworkwaspartofSTWprojectAIF.3595:“RobustControlofNonlinearSystemsusingNeuralNetworks”.ItwascarriedoutingraduateschoolASCI.AdvancedSchoolforComputingandImagingCopyright°c2001byStephanH.G.tenHagen.Allrightsreserved.Contents1Intro

4、duction11.1Control.....................................21.1.1Designingthestatefeedbackcontroller................31.1.2Unknownsystems............................51.2ReinforcementLearning............................71.3ProblemStatement...............................81.4OverviewofthisThesis...

5、..........................92ReinforcementLearning112.1ADiscreteDeterministicOptimalControlTask...............112.1.1Theproblem...............................112.1.2Thesolution...............................132.2TheStochasticOptimizationTasks......................132.2.1TheMarkovDecisionProc

6、ess.....................132.2.2Modelbasedsolutions.........................142.2.3ReinforcementLearningsolutions...................162.2.4Advancedtopics.............................202.2.5Summary................................212.3RLforContinuousStateSpaces........................222.3.1Co

7、ntinuousstatespacerepresentations................222.3.2Learningincontinuousdomains....................242.3.3Summary................................262.4Discussion....................................273LQRusingQ-Learning293.1Introduction................