Approximate dynamic programming solutions with a single network adaptive critic for a class of nonlinear systems-一种针对一类非线性系统的网络自适应批评的近似动态规划解决方案.pdfVIP

20ll ．，con肋f刀l鲫)，铆f9(3)370_380 DOI10．1007／sl1768m11．0191．3 Approximate solutionSwith dynamicprogramming a network criticforaclassof single adaptiVe nonlinear systemS JieDING．S．N．BALAKRISHNAN ofMecllaIlical明d D印删[IIl即t AerospaceEn酉n∞ring，Missouri 6540l，U．S．A． Unive巧i哆ofScience柚d钕hllology’Rolla，MO wiman Abstract：Appmximatedyn枷cpro掣珈m面ng(ADP)f咖ulationimplementedadaptivecritic(AC)-b勰ed neural hasevolvedasa for me network(NN)stnJcturepowerfultechniquesolving tions．As interestinADPand山eACsolutionsare with isa to time，theredireneedconsider escalating possibIeenabling fktorsforttIeir ACstmctureconsistsof is implementations．Atypical two试teractingNNs，whichcomputationallyexpen． sive．Int}lis new ttle‘cost．function-basedne附ork paper，aarchjtectIl陀，called singIe adaptivecntic(J—SNAC)’ispresen砌。 wtlicheliminatesoneofttlenetworl【sina ACstructure．Ttlis is toawideclassofnonlinear typical appmachapplicable in ordertodemonstratethebenefitsandmecontml withttle systemsengineering．In syn